DEVELOPMENT OF AN INTERPERSONAL AGGRESSION SCALE FOR PEOPLE WITH INTELLECTUAL AND DEVELOPMENTAL DISABILITIES
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University
By Scott T. Matlock, M.A. *****
The Ohio State University 2008 Dissertation Committee
Approved by
Professor Michael Aman, Advisor Professor Betsey Benson Professor Luc Lecavalier Professor Michael Vasey
____________________________________ Advisor Psychology Graduate Program
ABSTRACT
Aggression researchers have made distinctions among forms of aggressive behaviors that seem to be clinically important. Perhaps the most important distinction is that between aggressive behavior which is primarily motivated by the desire to acquire some tangible reward or attain some goal (“cold” aggression), and that which is more impulsive or reactive (“hot” aggression). Reactive aggression is displayed in response to a perceived threat or provocation or possibly in response to frustration. Proactive aggression could be motivated to gain a resource, or could be directed toward a person with the purpose of bullying, domination, or intimidation. The purpose of this project was to develop an informant-based scale of the severity and/or frequency of interpersonal aggression among people with intellectual and developmental disability. I also attempted to incorporate a "reactive/proactive" domain as well, using a “provocation scale” which asked raters whether or not a behavior was more likely to have been provoked or not. Subjects were 512 individuals diagnosed with intellectual disability (an IQ of 70 or lower), and aged 19 to 84 years, inclusive. Raters were 61 staff or family members who were familiar with a subject for at least six months. Factor analysis yielded five subscales: (I) “Verbal Aggression,” (II) “Physical Aggression,” (III) “Hostile Affect,” (IV) “Covert Aggression,” and (V) “Bullying”. These subscales had good to excellent internal consistency and fair to excellent interrater reliability statistics. Concurrent validity was supported by comparisons with the results of the BPI Aggressive/Destructive ii
Behavior subscale. Comparisons of mean scores of groups of subjects divided by demographic, diagnostic, and medication variables yielded results which were supportive of the Problem Scale’s external validity. The Provocation Scale and its subscales had interrater reliability statistics in the fair to good range of clinical significance. Comparisons of mean scores of groups of subjects divided by demographic, diagnostic, and medication variables yielded results which were supportive of the Provocation Scale’s external validity. A few of the t-test results indicate that the Provocation Scale can provide pertinent information not obtainable with the Problem scale alone. These results are encouraging, but it is also clear that important improvements to the scale are needed.
iii
Dedicated to my parents, Ethel and Roger
iv
ACKNOWLEDGMENTS
I thank my advisor, Mike Aman, for his patient guidance and encouragement. I am also grateful to all of the fine faculty and my fellow graduate students at the Nisonger Center for their helpful advice. Thanks go to my wife, Nancy Radcliffe, and all of my family and friends for their loving support. This would not have been possible without them. I am grateful to Johannes Rojahn for getting me through the first leg of this journey. This research was supported by a grant from the Ohio Department of Mental Health, and I am greatly appreciative of their assistance.
v
VITA
August 17, 1962………………………….Born – Seymour, Indiana 1991………………………………………B.S. Psychology, Indiana University 1992 – 1994………………………………Clinical Associate, Muskatatuck State Developmental Center, Butlerville, Indiana 1999………………………………………M.A. Psychology, The Ohio State University 1999 – 2001………………………………Behavior Coordinator, Heinzerling Foundation, Columbus, Ohio 2001 – present……………………………Behavior Support Specialist, Nisonger Behavior Support Services, Columbus, Ohio
PUBLICATIONS Research Publications 1. Rojahn, J., Gerhards, F., Matlock, S.T., & Kroeger, T. L. (2000). Reliability and validity studies of the Facial Discrimination Task for emotion research. Psychiatry Research, 95, 169 – 181. 2. Rojahn, J. Matlock, S. T., & Tasse', M. J. (2000). The Stereotyped Behavior Scale: psychometric properties and norms. Research in Developmental Disabilities, 21, 437-454.
vi
FIELDS OF STUDY Major Field: Psychology
vii
TABLE OF CONTENTS Page Abstract …………………………………………………………………………………...ii Dedication ………………………………..……………………………………………....iv Acknowledgments…………………………………………………………………………v Vita ……………………………………………………………………………………….vi List of Tables........................................................................................................................x Chapters: 1.
Introduction .................................…………………………………………………1
2.
Literature review......................................................................................................4
3.
Method…………………………………………………………………………...35
4.
Data analysis and results…………………………………………………………40
5.
Discussion………………………………………………………………………..56
References..........................................................................................................................80
Appendices Appendix A: Tables.........................................................................................................94 Appendix B: The Scale of Hostility and Aggression, Reactive/Proactive (SHARP) and Aggressive/Destructive Behavior Subscale of the Behavior Problems Inventory (BPI-01)...........................................................................................145 Appendix C: Four factor solution using equamax rotation.............................................151 viii
Appendix D: Six factor solution using equamax rotation...............................................154 Appendix E: Five factor solution using oblimn rotation................................................157 Appendix F: Five factor solution using quartimax rotation............................................160 Appendix G: Five factor solution using varimax rotation..............................................163 Appendix H: Rater recruitment script and consent form................................................166
ix
LIST OF TABLES Page
Table 1
Demographic characteristics of subject sample ………………………………....95
2
Percentages of endorsement for the original 58 SHARP items …………………96
3
Five-factor solution using Maximum Wishart Likelihood extraction with oblique equamax rotation …………………………………………………..99
4
Rotation sums of squared loadings ……………………………..……………...102
5
Correlations between SHARP subscales…………………………………….....103
6
Problem scale total and subscale scores for subjects eighteen to thirty years of age (n = 115)…………...………………………...…104
7
Problem scale total and subscale scores for subjects thirty-one to forty years of age (n = 121)…………………………………….....105
8
Problem scale total and subscale scores for subjects forty-one to fifty years of age (n = 147)………………………………………..106
9
Problem scale total and subscale scores for subjects fifty-one to sixty years of age (n = 90) ..……………………………………....107
10
Problem scale total and subscale scores for subjects aged sixty years or more (n = 31)…………………………………………………….108
11
Problem scale total and subscale scores for male subjects (n = 314)…………..109
12
Problem scale total and subscale scores for female subjects (n = 190)………...110
13
Problem scale total and subscale scores for subjects diagnosed with mild IDD (n = 222)………………………………………….…111
14
Problem scale total and subscale scores for subjects diagnosed with moderate IDD (n = 172)……………………………………….112 x
15
Problem scale total and subscale scores for subjects diagnosed with severe/profound IDD (n = 115)………………………………..113
16
Problem Scale Total and Subscale Scores for All Subjects (n = 512).................114
17
Internal consistency of problem scale total, problem subscales and BPI Aggressive/Destructive Behavior subscale (alpha coefficients)…………...115
18
Interrater reliability of the problem subscales………………………………….116
19
Correlations between the SHARP problem scale subscales and total, and the BPI Aggressive/Destructiveness subscale…………………………...…117
20
Comparison of significant differences in problem scale provocation scores by demographic groups……………………….…………………………118
21
Comparison of Problem Scale validators and Provocation Scale validators (diagnoses)……………………………………………………………………...119
22
T-Scores and Mean Problem Scale Scores for Groups of Subjects Split by Diagnoses...............................................................................................121
23
T-Scores and Mean Provocation Scale Scores for Groups of Subjects Split by Diagnoses...............................................................................................122
24
Comparison of Problem Scale validators and Provocation Scale validators (aggregate diagnostic categories)……………………………………………….123
25
T-Scores and Mean Problem Scale Scores for Groups of Subjects Split by Aggregate Diagnostic Categories...........................................................124
26
Comparison of Problem Scale validators and Provocation Scale validators (medications)……………………………………………………………………125
27
T-Scores and Mean Problem Scale Scores for Groups of Subjects Split by Medications............................................................................................126
28
T-Scores and Mean Provocation Scale Scores for Groups of Subjects Split by Medications............................................................................................127 xi
29
Interrater reliability of the Provocation Total and subscales…………………...128
30
Interrater reliability of the Provocation Quotient and subscale quotients….…...129
31
Interrater reliability of the Provocation/Problem sums for the full scale and subscales………………………………………………….130
32
Interrater reliability of the Provocation/Problem products for the full scale and subscales………………………………………………….131
33
Correlations between the Problem Scale, Provocation Scale, and provocation variations totals…………………………………………...…..132
34
Provocation Scale total and subscale scores for subjects eighteen to thirty years of age (n = 115)……………………………………..…133
35
Provocation Scale total and subscale scores for subjects thirty-one to forty years of age (n = 121)………………………………….........134
36
Provocation Scale total and subscale scores for subjects forty-one to fifty years of age (n = 147)………………………………………..135
37
Provocation Scale total and subscale scores for subjects fifty-one to sixty years of age (n = 90)………………………………………....136
38
Provocation Scale total and subscale scores for subjects aged sixty years or more (n = 31)……………………………………………....137
39
Provocation Scale total and subscale scores for male subjects (n = 314)………………………………………………………....138
40
Provocation Scale total and subscale scores for female subjects (n = 190)……………………………………………………….139
41
Provocation Scale total and subscale scores for subjects diagnosed with mild IDD (n = 222)…………………………………………….140
42
Provocation Scale total and subscale scores for subjects diagnosed with moderate IDD (n = 172)……………………………………….141
43
Provocation Scale total and subscale scores for subjects diagnosed with severe/profound IDD (n = 115)………………………………..142
xii
44
Provocation Scale Total and Subscale Scores for All Subjects (n = 512)...........143
45
Matrix of Correlations between Problem and Provocation Subscales.................144
xiii
CHAPTER 1
INTRODUCTION Aggressive behavior in people with intellectual and developmental disabilities (I/DD) presents many serious difficulties and challenges. Prevalence of aggression in this population has been estimated at 22 percent, and considerable resources are expended in the management and treatment of this problem (National Institute of Health, Consensus Development Panel on Destructive Behaviors in Persons with Developmental Disabilities, 1989). Violent behaviors have been linked to lessened community job opportunities (Block & Rizzo, 1995) and increased social isolation (Golden & Reese, 1996). Aggression is the primary reason for prescription of psychotropic medications (Baumeister, Todd, & Sevin, 1993) and institutional placement referral (Meador & Osborn, 1992) for people with I/DD diagnoses. After reviewing the relevant literature on instruments for assessing psychopathology in intellectual disability, Aman (1991) concluded that there was a shortage of instruments for the assessment of aggressive behavior among people with developmental disabilities. The seriousness of the problem of aggression among people with intellectual disability is well-established (Ross, 1972; Jacobson, 1982; Hill & Bruininks, 1984), and so, therefore, is the importance of having valid and reliable measurement tools for this purpose. 1
Aggression researchers have made certain distinctions among forms of aggressive behaviors that seem to be clinically important. It is likely that one of the most important distinctions is that between aggressive behavior which is primarily motivated by the desire to acquire some tangible reward or attain some goal (“cold” aggression), and that which is more impulsive or reactive (“hot” aggression). Examples of the various labels which have been applied to this dichotomy are Skinner’s “ontogenic” versus “phylogenic” (1969), Zillman’s “incentive-motivated” versus “annoyance-motivated” (1979), and Dodge and Coie's “proactive” (or “instrumental”) versus “reactive” aggression (1987). Reactive aggression is displayed in response to a perceived threat or provocation or possibly in response to frustration. This behavior is generally considered to be impulsive and typically occurs with a strong expression of negative, hostile emotion. Proactive aggression is a non-provoked aggressive act motivated to gain a resource, or could be directed toward a person with the purpose of bullying, domination, or intimidation. While reactive aggression is a hostile response, proactive aggression is planned, instrumental, and more likely to be learned. Cornell et al. (1996) argued, however, that proactive aggression can take place in a moment of passion (as in the case of a man who demands another's money, then beats him to death when he meets resistance), and that reactive aggression can be quite "cold-blooded" (a carefully plotted act of revenge, for example). Dodge and Coie (1987) developed the first empiricallyderived instrument to assess these two constructs. This was later followed by additional tools to achieve the same ends (Brown et al., 1996; Little et al., 2003). The psychometric properties of the original Dodge and Coie (1987), Brown et al. (1996), and Little et al. (2003) instruments lent support to the constructs' validity. 2
A reliable, valid method for differentiating individuals whose aggressive behaviors are more instrumental than impulsive, and vice versa, could be of value to clinicians in the field of intellectual disability and developmental disabilities. For example, researchers have found support for the theory that violent offenders diagnosed with psychopathy engage in more instrumental aggression than non-psychopathic offenders (Cornell et al., 1996). On the other hand, one might expect a high rate of reactive violence to be indicative of an impulse-control disorder, or Cluster B personality disorders which are marked by “disturbance in impulse control and emotional dysregulation” (DSM-IV; American Psychiatric Association, 1994). Diagnosis and treatment planning could perhaps be supplemented by the results of a good proactive/reactive scale. Of course, such an instrument might also be of use to researchers for the purposes of screening, assessing outcome, and so forth.
3
CHAPTER 2
LITERATURE REVIEW
Theoretical Background. Many theoretical explanations for reactive, impulsive, angry aggression are derived from the frustration-aggression model (Dollard, Doob, Miller, Mowrer, & Sears, 1939; Polman, Orobio de Castro, Koops, van Boxtel, & Merk, 2007). Proactive aggression is generally considered to be learned, instrumental behavior which is controlled by contingencies (Dodge & Coie, 1987). The social information processing model (Crick & Dodge, 1996) offers explanations for both types of aggressive behavior. In the following review, I will provide a brief overview of these theories, along with a selection of relevant research. The frustration-aggression model. The concept of "aggressive energy," relabeled as a drive, has long influenced theorists. Most influential among such “aggressive drive” theorists were Dollard et al., who in 1939 introduced the frustration-aggression hypothesis, and Berkowitz (1962) and Feshbach (e.g., 1970), each of whom built upon the Dollard et al. theory. In a monograph entitled Frustration and Aggression (1939), Dollard et al. defined a “frustration” as “an interference with the occurrence of an instigated goal-response at 4
its proper time in the behavior sequence.” By “goal-response” they meant the final reinforcing step in an ongoing chain of behavior. Their initial theory stated that “the occurrence of aggressive behavior always presupposes the existence of frustration and, contrariwise, that the existence of frustration always leads to some form of aggression.” Frustration was, in other words, a necessary and sufficient condition for the occurrence of aggression. Dollard et al. (1939) posited that the strength of motivation for aggression in any given instance is a function of three factors: (a) the reinforcement value of the goal response, (b) the degree to which the goal response is frustrated, (c) and the number of frustrated goal response behavior sequences. This last factor implies that the effects of frustration are cumulative. Smaller frustrations, even if they don’t result in immediate manifestations of overt aggression or hostility, accumulate and increase the strength of some later response to frustration. The authors reconciled the initial failure of these minor frustrations to produce aggressive acts with their assertion that “the existence of frustration always leads to some form of aggression” by considering being “furious inside” as non-overt aggression. Overt acts of aggression can be inhibited by the likelihood of punishment, but the drive to aggress is still present. In the introduction to his 1962 book Aggression: A Social Psychological Analysis, Berkowitz dubbed the Dollard et al. (1939) theory an “overly sketchy” but “extremely sound” foundation. In 1965, Berkowitz proposed that frustration-induced anger creates “only a readiness for aggressive acts.” In addition, aggressive behavior “will not occur, even given this readiness, unless there are suitable cues, stimuli associated with the present or previous anger instigators" [p.308]. 5
Once someone has reached a state of heightened readiness, Berkowitz (1965) wrote, he or she “will not attain completion until the goal object has been aggressively injured.” Berkowitz’ version of the theory stresses the importance of goal achievement (actually injuring the target in some way) over goal response (the aggressive behavior) in drive reduction. It follows then that hostile acts performed against alternative targets merely result in more frustration, and stronger “anger ‘drive state’”(Berkowitz, 1973). Taylor (1986) commented that numerous studies of cathartic resolution of aggressive drive had “produced either equivocal results or a tendency for aggression to increase subsequent to the performance of an aggressive act.” In a thorough review of the literature, Quanty (1976) stated that “results from the studies reviewed cast serious doubt on the traditional aggression catharsis hypothesis…” Since Feshbach’s (1970) model allows for the possibility of reinforcement of the goal response via substitute aggression, it is always possible to attribute results which fail to indicate catharsis to such reinforcement. Thus his catharsis theory eludes falsification. Drive theorists expressed concepts which were, generally speaking, readily operationalized. However, those concepts were still resistant to objective measures. How does one reliably measure what is essentially a nebulous intervening variable, like “frustration”? One can obstruct a subject’s progress toward some goal, but how to gauge the immediate cognitive and emotional effects of that obstruction? Kaufmann (1965) concluded, after a review of the relevant empirical literature, that the concept of frustration was so “protean as to be refractory to operational definition.” Zillman (1979) pointed out that unlike, for instance, the drive to consume food, the drive to aggress is not strengthened by deprivation. Therefore, it is impossible to 6
specify essentially consummatory acts which would eliminate a deficiency (e.g., eating). He reasoned that to characterize infliction of harm upon a target as such an act would be an overextension of a model meant to explain behaviors fundamentally different from aggression. As noted by Feshbach (1997), the authors of the original frustration-aggression hypothesis “readily acknowledged the criticism that the role of learning in the development of aggressive behavior had not been addressed.” He might have added that he and others who had modified the theory had given only slightly more attention to learning than had Dollard and his colleagues (1939). Although, for example, it was implied that one must learn to associate certain objects, settings, etc. with violence in order to establish “suitable cues,” the steady focus of Berkowitz’ work was the mechanism of drive. Learning theories of aggression. Learning theorists can be divided into two general schools of thought regarding human aggression. Some, for example, Skinner (1953), have focused on directly applying what had been learned about classical and operant conditioning of aggressive responses in animals to conspecific human aggression. Others, most prominently Bandura (e.g., 1971c, 1973a), have stressed that, in humans, behavioral mechanisms are strongly influenced by complex cognitive and social structures. There is apparent consensus among learning theorists that, as Scott (1958) put it, “motivation for fighting is strongly increased by success, and that the longer success continues, the stronger the motivation becomes" [p. 126]. However, there has been controversy concerning the definition of said “success” or, in other words, reinforcement. 7
B. F. Skinner (1969), not surprisingly, considered attributing violence in the streets to frustration to be “dangerous psychologizing.” He instead approached aggressive behavior as a product of operant conditioning, with the infliction of injury upon one’s target as its main reinforcer. He proposed that the target’s reactions to being injured were powerfully, even innately, reinforcing stimuli. He wrote, “...stimuli which reinforce aggressive acts are to be found in the behavior of the recipient as he weeps, cries out, cringes, flees, or gives other signs that he has been hurt. (Counter-aggression may be among these behaviors; an aggressive person is reinforced by ‘getting a rise’ out of his opponent)" [p.210]. Skinner considered nearly all human aggression to be at least partially phylogenic, accompanied by autonomic responses (appropriate to “tooth-and-claw competition”) and reinforced by injuring the target. For such acts of violence, inflicting injury was presented by Skinner as a primary reinforcer, like eating or drinking. He pointed to the popularity of violent spectator sports, like boxing, as evidence of the power of such reinforcement. He also described ontogenic aggression, which is learned. This sort of behavior was more instrumental, reinforced via the association of “signs of damage” with secondary reinforcers “which do not otherwise have anything to do with aggression.” Nevertheless, Skinner saw phylogenic autonomic responses as the excitatory component of even this sort of “cooler-headed,” instrumental behavior. He wrote that, in an instance of ontogenic aggression, “an innate capacity to be reinforced by damage to others” can trigger the pattern of autonomic responses seen in phylogenic aggression. “To deal successfully with any specific aggressive act,” he wrote, “we must respect its provenance. Emotional responses, the bodily changes we feel when we are aggressive...may 8
conceivably be the same whether of phylogenic or ontogenic origin; the importance of the distinction is not thereby reduced.)" [p.195]. An obvious reinforcer for violent behavior is the acquisition of tangible rewards. Such rewards serve as secondary reinforcers, as in Skinner’s explanation of ontogenic behavior. A mugger may be reinforced by several stimuli, including perhaps expressions of pain from the victim, but his behavior is likely most strongly reinforced by financial gain. Cowan and Walters (1963) demonstrated the effectiveness of tangible, positive reinforcement of aggression in humans. They rewarded young male subjects with marbles for hitting an inflated doll. When the response-to-reinforcement ratio was increased, the rate of punching increased. In a later study, Walters and Brown (1963) found that boys who had been reinforced for doll-punching on a fixed ratio (FR 6) schedule were more likely to act aggressively in a subsequent frustrating situation (having their candy taken away or a movie shut off) than boys who had been continuously reinforced for punching. Non-tangible rewards, e.g., attention or verbal praise, have also been shown to reinforce aggression. Over the course of a year-long observation of toddlers in small play groups, Fagot (1984) observed that teachers or peers responded to 81 percent of aggressive acts by boys. Although Fagot classified a quarter of those responses as negative (which could nevertheless be reinforcing), he classified 40 percent of the teachers’ responses as positive. He concluded that the contingent attention given contributed to the high frequency of aggressive behaviors observed. Geen and Stonner
9
(1971) found that, when praised by authority figures for having done so, adult subjects delivered higher levels of shock to confederates than those who had received no praise. Bandura’s social-learning theory of aggression (1973a, 1973b) is largely in accord with earlier learning theories, in that it attributes much of aggressive behavior between humans to learning through direct experience. The critical distinction is in his view that humans also acquire some aggressive behaviors through observation of models, with no direct experience whatsoever. Because of our capacity for information processing, our behavior can be modified through observation of the consequences of others’ behavior. So, one’s response rate for a particular behavior should be increased after observing that behavior apparently produces some reward for someone else. Likewise, one should be less inclined to engage in behavior for which one has witnessed others being punished. Bandura proposed two major explanations for this hypothesized effect of vicarious reinforcement and punishment. First, there is the informative function of observed consequences (e.g., Bandura, 1971a). One’s observations can lead to recognition of the prevailing contingencies of reinforcement, thus helping to encourage or inhibit particular responses. Second, he proposed the vicarious conditioning and extinction of emotional arousal (e.g., Bandura, 1971b). He suggested that witnessing the expression of emotion by another can cause an observer to experience a concordant affective reaction (Bandura & Rosenthal, 1966). If a model is rewarded for an act of aggression, the observer experiences vicarious gratification. That gratification then becomes associated with the aggressive act itself. Witnessing an act of aggression being punished would, of course, have an opposite effect.
10
Bandura also wrote that human aggression can be regulated by self-reinforcement or self-punishment (e.g., 1973a). Self-reinforcement would take the form of feelings of personal pride and agency after successful acts of violence. Self-punishment would be exemplified by feelings of regret or self-contempt in the wake of an aggressive incident. Bandura (1973a, 1973b) proposed several means by which individuals circumvent any such feelings of self-contempt. One may “slight aggression by advantageous comparison,” comparing one’s deeds to ones which are much worse. One may attempt to justify aggression by declaring it to be a means to highly noble or necessary ends. Responsibility for aggressive acts may be displaced to some higher authority. Victims can be characterized as subhuman, degenerate, or even responsible for giving the attacker no choice but to act against them. The consequences of one’s actions can be misrepresented as being less harmful than they actually were. And, finally, Bandura described a gradual process of desensitization. As one repeatedly performs acts of aggression, the feelings of self-contempt diminish and the severity and frequency of the aggression gradually increase. It is important to note that Bandura (1973a) held that, although our actions are under cognitive control, “cognitive events” cannot be considered to be the sole causes of behavior. They are always under stimulus and reinforcement control. Therefore, any effective analysis of cognitive control of behavior must include the influences upon a person’s cognitive processes. Social information processing theory. Crick and Dodge (1996) described a social information processing model that included six stages involved in social interaction: encoding, interpretation, selection of goals, response construction and evaluation, 11
response decision, and enactment of response. They theorized that deficits or biases in one or more of these steps are linked to aggressive responding. Reactive aggression, for example, could stem from one's biased interpretation of another's intentions as malicious or threatening. Or, proactive aggression may be linked to a person's evaluation of aggressive behaviors as an acceptable and effective means to produce some desired outcome. Based upon their scores on the Conduct Disorder subscale of the Revised Behavior Problem Checklist (Quay & Petersen, 1987) and the Aggression subscale of the Reiss Screen for Maladaptive Behavior (Reiss, 1988), Fuchs and Benson (1995) divided a sample of 66 borderline-to-moderately intellectually disabled men into two groups, aggressive and non-aggressive. Forty-six men either did not meet criteria for inclusion in either group or were unable to follow the instructions, and were therefore dropped from the study. The remaining subjects were each read a series of eight vignettes describing conflict situations. After each vignette was read, subjects were asked whether they viewed the behavior described to be intentionally negative or accidental. Once this question had been answered, subjects were asked to give up to five ways in which they might respond to such a situation. Next, the vignettes were presented again, but this time a set of three additional response options were given, and the subjects were asked to choose the best. Finally, verbal skills of the subjects were assessed using the Peabody Picture Vocabulary Test - Revised (PPVT-R; Dunn & Dunn, 1981) and the Vocabulary subscale of the Wechsler Intelligence Scale for Children - III (WISC - III; Wechsler, 1991). One-way ANOVA showed no significant difference between groups in receptive or expressive verbal abilities. Two independent raters coded the subjects' initial 12
responses to the stories by (a) number of solutions to each situation, (b) number of aggressive solutions, (c) number of assertive/effective solutions, and (d) number of passive responses. Subjects in the aggressive group generated far more aggressive solutions than their peers in the non-aggressive group. Also, the first responses given by the aggressive group were significantly more likely to be aggressive solutions than those of the non-aggressive group. However, aggressive subjects were not significantly less able to identify hostile intent correctly, nor were they likely to generate multiple solutions, than their non-aggressive counterparts. Schippell, Vasey, Cravens-Brown, and Bretveld (2003) investigated the role of attentional biases at the encoding step of the Crick and Dodge model (1996). Their subjects were 90 typically-developing children, ages 11 to 16 years. The probe detection task (PDT) is a measure of how attention is allocated when two competing stimuli are presented to a subject simultaneously. Schippell et al. gave their subjects a version of the PDT which had been developed by Vasey et al. (1996) for a study which demonstrated that greater attention to threat cues is associated with anxiety. Subjects' teachers and parents were given a battery of questionnaires pertaining to childhood aggression and related variables. The children were presented with 12 vignettes and asked to picture themselves in the situations described, which were ambiguous and could be interpreted as either threatening or non-threatening. They then gave open-ended interpretations and response choices, and also chose one interpretation out of four multiple-choice options (two threatening and two non-threatening). As hypothesized, attentional bias to cues of rejection, ridicule, and/or failure was indeed linked to reactive, but not proactive, aggression. However, the direction of this association was surprising, in that this effect 13
was found for suppressed attention to such cues. The authors suggested that this might be explained if the encoding step is broken down further into early, pre-conscious, automatic processing and later selection for processing after stimuli have entered conscious awareness. It is possible that anxiety-producing stimuli are suppressed at this later point, in spite of a bias to attend to such cues at an earlier, pre-conscious stage. In any case, the association between reactive aggression and an attentional bias for socialthreat cues was viewed as supportive of the Crick and Dodge (1996) model. Schwartz et al. (1998) used a contrived playgroup protocol to examine the socialcognitive and behavioral correlates of aggression in groups of boys. Sixty-six boys (mean age of eight years) were split into 11 playgroups of six. Several play sessions were videotaped, and two independent observers coded events which they judged to be "reactive aggression," "proactive aggression," or "dominance" (non-aggressive, but controlling, behavior). Subjects were interviewed to assess attribution styles, and outcome expectancies for aggressive or assertive behavior. Proactive aggression was associated with positive outcome expectancies for aggression/assertion. Reactive aggression was associated with hostile attributional tendencies and frequent victimization by peers. Two-factor model of reactive and proactive aggression. Several studies employing confirmatory factor analysis have found that reactive and proactive aggression are best conceptualized as two separate constructs, rather than a single continuum (Price & Dodge, 1989; Poulin & Boivin, 2000; Xu & Zhang, 2007). Some (e.g., Bushman & Anderson, 2001) have questioned the scientific and clinical utility of the reactive and proactive aggression constructs, given the high 14
correlation found between them by several studies. This problem was examined by Polman et al. (2007) in a meta-analysis of studies using the reactive/proactive distinction. They reported that most of these studies did find relatively high correlations between the two contructs. However, very low, even negative, correlations between proactive and reactive aggression were found in the very small number of studies which sought to disentangle the forms and functions of the behaviors in question. Little et al. (2003) hypothesized that high correlations between instrumental and reactive aggression were at least in part due to respondents' focusing on the forms, or topographies, of behaviors while having difficulty discerning their functions, such as being "instrumental" or "defensive." Little et al. therefore designed a scale which separated these questions from one another. The development of this instrument is described in greater detail below, in the review of aggression instruments. With aggression's form "disentangled" from its function, Little et al. found instrumental and reactive aggressive behaviors to be only negligibly correlated with one another (disattenuated r = -.10, p < .05). A Review of Aggression Instruments. What follows is a review of a selection of aggression scales which are relevant to the current study because they (a) were developed for the intellectually disabled population, (b) included a hostility domain, or (c) were developed with the intent to gauge reactive and proactive types of aggression. Buss Durkee Hostility Inventory (BDHI). One of the earliest examples of a standardized scale for the assessment of human aggression or hostility is the Buss-Durkee Hostility Inventory (BDHI; Buss & Durkee, 1957). The BDHI is discussed here because 15
of its significant place in the history of aggression measures, and because it appears to be the earliest attempt to assess hostility with a standardized instrument. The authors stated that several subtypes of hostility and aggressiveness had been identified by earlier research (Buss, Durkee, & Baer, 1956). These subtypes were named "Assault," "Indirect Hostility," "Irritability," "Negativism," "Resentment," "Suspicion," and "Verbal Hostility." Citing a lack of any previously-published inventories that attempted anything more than a global estimate of hostility, they set out to develop one which would provide estimates of various subtypes of hostility (e.g., overt versus covert manifestations of hostility). They first selected a pool of items from previous inventories, and then wrote several more items to supplement this pool. Most of the items borrowed from other scales underwent modifications. Item selection criteria were as follows: items should refer to only one subclass of hostility and be worded to minimize defensiveness in responding, the behaviors and attitudes involved should be specific, and the stimulus situations that arouse them should be nearly universal. Because the authors were interested in the relationship between the experience of guilt and the various subtypes of hostility, a "Guilt" subscale was included. The first version of the scale consisted of 105 items. Items intended to gauge each of the various subtypes of hostility were scattered randomly throughout the inventory. It was presented in group fashion to 85 male and 74 female college students. Respondents rated statements such as "I don't seem to get what's coming to me" or "I can think of no good reason for ever hitting anyone" as either true or false. All responses were given anonymously. Two criteria were used to pare down the initial item set, frequency and internal consistency. Only those items for which 15-85% of the sample 16
answered in the direction of hostility were retained. Also, only those items that scored an internal consistency correlation of at least .40 were kept (because the items are scored dichotomously, the biserial correlation coefficient was used). Internal consistency was defined as the correlation of an item with the score of the subscale to which it belonged. Only 60 of the original items met these criteria. Because this left some of the subscales with very few items, therefore vulnerable to a lack of reliability, several new items were written and added. The revised inventory consisted of 94 items. This version was administered to 62 male and 58 female college students, and separate analyses were performed for each gender. The same internal consistency criterion was again used, but the frequency criterion was changed a bit. Because of concern that the original 15-85% range might lead to exclusion of items which differentiated between men and women, the new standard for inclusion was a frequency of 15-85% for either gender. The second item analysis produced an inventory of 66 items for hostility and 9 for guilt. The final form of the scale was given to 85 male and 88 female college students. The eight subscales were scored, and product-moment correlations were computed for each gender. Using Thurstone's centroid method (Thurstone, 1947), two factors were extracted. Considering only factor loadings of .40 and over, the first factor was defined by "Resentment" and "Suspicion" for men, and by "Resentment," "Suspicion," and "Guilt" for women. The second factor was defined by "Assault," "Indirect Hostility," "Irritability," and "Verbal Hostility" for both genders, with the addition of "Guilt" for women. The authors considered these factors to represent an "emotional" or attitudinal" component, and a "motor" component (overt aggression). However, they also noted that
17
factor loadings were not high. The average communality of the eight subscales was .43 for males and .40 for females. Edmunds and Kendrick (1980) performed an extensive validity study of the BDHI. Six groups of subjects, 62 male college students, 57 female college students, 108 "normal" males, 111 "normal" females, 90 delinquent males, and 102 delinquent males, were given the inventory. What is perhaps of greatest relevance here is that they found no significant relationship between scores on the BDHI and observable violent or disruptive behavior. Although such self-report instruments may be useful in assessing non-patient populations, those whose cognitive functioning is impaired by intellectual disability or severe psychopathology cannot be expected reliably to fill out protocols (Yudofsky, Silver, Jackson, Jackson, Endicott & Williams, 1986). Children's Hostility Inventory (CHI). Kazdin, Rodgers, Colbus, and Siegel (1987) considered the delineation of aggression and hostility to be an important feature of the Buss-Durkee Hostility Inventory (BDHI; Buss & Durkee, 1957), as it could have relevance to overt antisocial acts by youngsters and to their underlying cognitive and perceptual processes. Their goal in developing the Children's Hostility Inventory (CHI) was to evaluate aggression and hostility in children with an assessment tool parallel to the BDHI. Items (n = 38) from the Buss-Durkee inventory were reworded to be more applicable to children and adapted to be rated as either true or false by parents. Subjects for the inventory's evaluation were 60 girls and 195 boys, all of whom were inpatients of a psychiatric hospital where children were hospitalized for 2 to 3 months. They were admitted for highly aggressive or destructive behavior, suicidal or homicidal ideation or behavior, and deteriorating family situation. They ranged in age 18
from 6 to 12 years, and in Full Scale IQ from 70 to 133. The respondents were the mothers or maternal guardians of these children. The respondents were also administered the Interview for Antisocial Behavior (IAB; Kazdin & Esveldt-Dawson, 1986), a 30-item structured parent interview designed to gauge a broad range of antisocial behaviors, the Child Behavior Checklist (CBCL; Achenbach & Edelbrock, 1983) and the School Behavior Checklist (SBCL, Form A; Miller, 1977), a set of 96 true/false items to be rated by teachers. The correlation matrix of the CHI subscales (again, based upon those of the BDHI) was subjected to factor analysis. Two factors were found with a minimum eigenvalue of 1.0, which were subjected to varimax rotation. A loading of .40 or more was used to determine assignment to secondary factors. The first factor (Aggression) consisted of subscales labeled Assaultiveness, Indirect Hostility, and Verbal Hostility. The second (Hostility) was made up of subscales called Resentment, Suspicion, and Irritability. Multivariate analyses of variance (MANOVAs) comparing subjects diagnosed with Conduct Disorder (n = 149) with the others (n = 106) found a significant effect across CHI subscales. Males scored significantly higher than females on three subscales: Assaultiveness, F(1, 239) = 10.53, p < .001, Indirect Hostility, F(1, 239) = 5.99, p < .02, and Verbal Hostility, F(1, 239) = 4.36, p < .05. Evaluation of internal consistency across items of the CHI's subscales yielded good results. Pearson product-moment correlations indicated that CHI Aggression subscale scores correlated more highly with other measures of externalizing behavior, the Overt Antisocial Behavior subscale of the IAB (.70) and the Externalizing scale of the CBCL (.53), than did CHI Hostility scores (.49 19
and .43, respectively). CHI Hostility scores correlated more highly with the Internalizing scale of the CBCL (.52), than did CHI Aggression scores (.20). More information is needed regarding the CHI's psychometric properties. For example, test-retest reliability and interrater reliability should be evaluated. More information about the instrument's utility with intellectually disabled children, and perhaps adults, would be helpful. Nisonger Child Behavior Rating Form (NCBRF). In 1996, Aman, Tasse', Rojahn, and Hammer, colleagues at the Nisonger Center for Mental Retardation and Developmental Disabilities at the Ohio State University, modified the Child Behavior Rating Form (CBRF; Edelbrock, 1985). Citing a need for good instruments developed specifically for rating problem behaviors in developmentally disabled children, they introduced the Nisonger Child Behavior Rating Form (NCBRF). The CBRF has 10 Social Competence items and 55 Problem Behavior items. Factor analysis by Kolko in 1988 found that the CBRF's 10 Social Competence items fell into two factors, and the 55 Problem Behavior items made up five factors. The Problem Behavior items are scored on a scale of 0 (behavior did not occur or was not a problem) to 3 (an extremely severe problem). Aman et al. (1996) chose the CBRF because it was brief enough to be completed quickly (7 or 8 minutes), could be completed by teachers and parents, was applicable to a wide range of children, and assessed a broad array of childhood behavior problems. Those problems included stereotypic behavior and selfinjury, which are prominent in mental retardation. Aman et al. (1996) then adapted the CBRF to their purposes. Both severity and frequency of problem behaviors were rated by the new instrument. So, the new four20
point scale went from 0 (behavior did not occur or was not a problem) to 3 (occurred a lot or was a severe problem). Also, four items were rewritten to be more specific and concrete. Finally, 16 new items were added, to describe self-injury, stereotypy, shyness, sensitivity, argumentativeness, etc. The resulting instrument was eventually named the Nisonger Child Behavior Rating Form. A sample of 369 children between the ages of 3 and 16 years who had been referred for assessment at the Nisonger Center were rated on the NCBRF by their parents (usually mothers). The large majority of the children in the sample had IQ scores falling within the range associated with intellectual disability. One hundred eighty-nine of these subjects were also rated on the NCBRF by their teachers. Parents of the first 62 participants selected were also asked to complete the Aberrant Behavior Checklist (Aman, Singh, Stewart, & Field, 1985). The ABC is a 58-item scale with five subscales: (I) Irritability, Agitation, Crying, (II) Lethargy, Social Withdrawal, (III) Stereotypic Behavior, (IV) Hyperactivity/Noncompliance, and (V) Inappropriate Speech. Any items rated for fewer than 5% of the subjects were dropped. Exploratory factor analyses were done using the same methods employed earlier by Kolko (1988). A principal components extraction method was used, followed by varimax rotation. Factors with eigenvalues exceeding 1.0 were kept. For the parent version of the NCBRF, six Problem Behavior factors, accounting for 51.3% of the variance, were designated as subscales and given the following names: Conduct Problem, Insecure/Anxious, Hyperactive, Self-Injury/Stereotypic, Self-Isolated/Ritualistic, and Overly Sensitive. The first subscale, Conduct Problem, is of most interest here. It consists of items describing a variety of acting-out behaviors, including fighting, cruelty to animals, etc. Factor 21
analysis of the teacher Problem Behavior ratings produced a nearly identical factor structure, with the exception of the sixth factor, which was labeled Irritable. The six factors accounted for 49.7% of the variance. Seventy-nine percent of the items in the parent data fell into the same factors as those of the teacher data. Pearson correlations were calculated for parent-teacher agreement on each of the subscales. Correlations ranged from .54 (Self-Injury/Stereotypic) to .22 (not surprisingly, since these are the substantially different sixth factors, Overly Sensitive and Irritable). The Pearson correlation for the Conduct Problem scores was .37. Another measure of reliability, internal consistency, was found to be high for the Problem Behavior subscales. Median alpha values were .85 for the parent ratings and .87 for teacher ratings. As an evaluation of concurrent validity, NCBRF Problem Behavior scores were compared to ABC scores. Pearson correlations (all ps < .01) between parent ratings for the Conduct Problem subscale of the NCBRF and the subscales of the ABC were as follows: Irritability .72, Lethargy/Social Withdrawal .37, Stereotypic Behavior .24, Hyperactivity/Noncompliance .66, and Inappropriate Speech .58. For the teacher ratings, Pearson correlations were as follows: Irritability .60, Lethargy/Social Withdrawal .18, Stereotypic Behavior .28, Hyperactivity/Noncompliance .71, and Inappropriate Speech .40. The Irritability and Hyperactivity/Noncompliance subscales assess behaviors analogous to conduct problems. Thus, this analysis supported the concurrent validity of the NCBRF. Norms, as well as age and gender effects, derived from the results of the study described above were reported by Tasse', Aman, Hammer, and Rojahn (1996). A series of ANOVAs found no significant main effects for age or gender on the Conduct Problem 22
subscales of either version of the NCBRF. A significant interaction effect for age and gender on the Conduct Problem subscale of the parent version was found, however: F (3, 318) = 2.64, p < .05. Two large-scale drug studies used the Conduct Problem subscale of the NCBRF as the primary outcome measure. Both Aman et al. (2002) and Snyder et al. (2002) assessed rispiridone for managing problem behavior in children with disruptive behavior disorders (conduct disorder, oppositional defiant disorder, disruptive behavior disorder not otherwise specified) in children with borderline IQ or mental retardation. Both studies observed a decline in Conduct Problem score of about 48% with rispiridone as compared with about 18% with placebo. As noted by Tasse' et al (1996), the norms they reported did not reflect the average behavior of children with developmental disabilities, because a random sample of the population of interest was not used. Rather, the results represented the behavior of outpatients referred to a University-affiliated program in the Midwest. More evaluation of the NCBRF, using subjects randomly sampled from the population of developmentally disabled children, would help to rectify this. Nevertheless, the NCBRF appears to be a reliable and valid instrument for the assessment of problem behaviors, including aggression toward others, in the intellectually disabled. Behavior Problems Inventory (BPI). The Behavior Problems Inventory (BPI) is a 52-item informant-based rating scale for assessing self-injurious behavior, stereotypy, and aggressive/destructive behaviors in mental retardation and other developmental disabilities. Subscales represent categories of behavior chosen a priori by the authors. Both frequency and severity scales are included. The "Aggressive/Destructive Behavior" 23
subscale has 11 items. The process for developing the BPI has been described by the authors as "incremental," having undergone several revisions since its original inception (in German) in the early 1980's (Edlinger, 1983; Rojahn, 1984). At that time, it featured only items for self-injurious behavior and stereotyped movements. Item selection had been based upon a review of the literature on problem behavior assessment instruments. Several informal field tests followed, helping to reduce the number of items and to hone item definitions. The instrument was eventually translated into English, and a subscale for aggressive/destructive behaviors was added. Again, items were selected on the basis of a review of the relevant literature. Shortly thereafter, the BPI was used in a survey of people with profound mental retardation, developmental disabilities and multiple physical disabilities (Mulick et al., 1988). Finally, a version with an expanded, independently developed stereotypy scale was introduced (referred to by the authors as "BPI-01"; Rojahn, Matlock, & Tasse', 2000). Items are scored on a five-point frequency scale (never = 0, monthly = 1, weekly = 2, daily = 3, hourly = 4) and a four-point severity scale (no problem = 0, slight problem = 1, moderate problem = 2, severe problem = 3). To be scored higher than 0, an item must describe a behavior which has been observed within the 2 months prior to administration of the instrument. Each of the a priori problem behavior categories was given a generic definition that applied to all items within the category. Aggressive/destructive behaviors, in order to be scored greater than 0, had to meet the definition of "abusive, deliberate attacks against other individuals or objects." A thorough psychometric study of the BPI-01 was published by Rojahn, Matson, Lott, Esbensen, and Smalls in 2001. Subjects were 432 individuals functioning in the 24
mild, moderate, severe, and profound ranges of mental retardation living in a developmental center in Louisiana. Four graduate students in clinical psychology conducted BPI-01 interviews with direct care staff members familiar with the subjects about whom they were interviewed for at least six months. To collect retest reliability data, the staff members were re-interviewed a second time approximately one week after the initial interviews for 247 subjects (58% of the full sample). Student-staff membersubject combinations were, of course, held constant during this phase of the study. Data were collected for between-interviewer agreement for 60 subjects (14% of the original sample), with two students present for the interviews, independently scoring the staff members' responses to the BPI-01 items. Pearson correlations were calculated between the severity and frequency scales. The r for the "Aggressive/Destructive Behavior" items was .87 (ranging from .80 to .91). Because of the high correlations found between the frequency and severity scales, the authors excluded the severity data from the other results reported. The goodness of fit of the a priori categories was tested via confirmatory factor analysis. The "Aggressive/Destructive Behavior" factor was found to be the most robust and homogeneous of the three. This was attributed to aggressiveness being a function of social maladaptation, whereas stereotypy and self-injurious behavior would be more properly conceived of as manifestations of personal maladaptation. Internal consistency and test-retest reliability of the Aggressive/Destructive Behavior subscale were both found to be good, according to generally accepted standards (Cicchetti & Sparrow, 1981; Cicchetti, 1994). By these same standards, betweeninterviewer agreement was found to be at the top end of the "fair" range. This isolated 25
weakness may be attributable to the fact that the interviewer-respondent format added an extra source of error to an evaluation of a respondent-based instrument. The 11-item Aggressive/Destructive Behavior subscale of the BPI-01 appears to be a useful tool for clinical assessment of aggression in mentally retarded individuals. The authors suggested that it is well suited to epidemiological studies or for aiding in executive decision-making in regard to behavior problems. More recent studies have found it to be sensitive to the effects of some psychotropic drugs (Aman et al., 1999a, 1999b). However, more items, describing specific behaviors in greater detail, could improve upon this already sound scale. Developmental Behavior Checklist (DBC). The Developmental Behavior Checklist (DBC; Einfeld & Tonge, 1995) is a psychopathology assessment instrument designed for use with children and adolescents with intellectual disability. The 96-item checklist is completed by parents or other primary care givers (Primary Care Givers Version, or DBC-P) or teachers (Teachers Version, DBC-T), reporting problems observed over a six-month period. Its structure was largely modeled on the Child Behavior Checklist (Achenbach & Edelbrock, 1983). To assess internal structure of the DBC-P, data on 1536 Australian and Dutch subjects, representing the full range of intellectual disability, was collected and analyzed. Factor analysis resulted in a five-factor solution. Internal consistency for the subscale that most closely represents the behaviors of interest here, Disruptive/Antisocial, was excellent. One thousand, one hundred fifty-five Australian and Dutch children were rated by their teachers in a similar study of the DBC-T. The authors decided that, because the resulting five-factor solution was so similar to the one found in the earlier study, they 26
would simply use the DBC-P factor structure for both versions. In another study, 1093 children, ages 9 to 13 years and functioning in the mild to severe ranges of ID, were rated by their parents with the DBC-P. Again, internal consistency was excellent. In an evaluation of the interrater reliability of the DBC-T, teachers and teachers' aides of pupils in special classes completed the DBC-T. The teachers and aides had known the subjects at least two months. Teacher-aide agreement was found to be good. Raters for a study of the interrater reliability of the DBC-P were parents of young people with ID attending developmental assessment and psychiatry clinics (42 pairs). In one case, a mother and grandmother living in the same house completed the questionnaire. Two nurses familiar with each subject were also asked to fill out the questionnaire. Excellent levels of agreement were found between parents, and between nurses. DBC-T checklists were completed by teachers and teachers' aides for 16 pupils with moderate to severe ID aged 6-13 years. A second questionnaire was completed by each of the raters after an interval of two weeks. In an evaluation of the DBC-P, 63 pairs of parents, residential workers, and nurses completed the questionnaire twice, again with a two-week interval between ratings. Test-retest reliability for each version was found to be excellent. To gain a measure of criterion group validity, 70 subjects (described as "young people attending developmental assessment clinics") for whom DBC-P checklists had been completed were also assessed by groups of three mental health professionals (two child psychiatrists and one clinical psychologist in each) as to an overall rating of severity, using a 6-point Likert scale. The pool of subjects included individuals who had been rated as "disturbed" as well as "non-disturbed," and the clinicians were blind to 27
these ratings. Those subjects rated by the clinicians as 0, 1, or 2 ("non-cases") or 3 ("possible non-cases") were dubbed Group 1, and those rated 4 through 6 ("cases") made up Group 2. A significant difference was found between the DBC-P scores for each group. Both versions of the Developmental Behavior Checklist appear to be wellresearched, potentially quite valuable tools for clinicians working in the field of intellectual disability. Dodge and Coie's "Teacher Rating Instrument." Dodge and Coie (1987) developed a teacher-rating instrument, consisting of 12 statements to describe examples of proactive, reactive and nonspecific aggressive behaviors. For example, "This child uses physical force in order to dominate other kids" was used as a proactive-aggression item. An example of a reactive-aggression item is, "When this child has been teased or threatened, he or she gets easily upset and strikes back." Each item is rated on a fivepoint scale, ranging from never to almost always to indicate the frequency of the described behavior's occurrence. The initial 12 items were embedded randomly in a survey with an additional 12 statements describing other social behaviors. Teachers of 259 third- through sixth-graders completed the instrument. A principal-components factor analysis, with varimax factor rotation, resulted in a twofactor structure. Only the first factor had an eigenvalue greater than one. The three items loading most strongly on the first factor described reactive aggression. All items on the second factor were proactive aggression descriptions. Based on this finding, two threeitem subscales were created. The internal consistencies of each of these subscales were excellent. The correlation between subscale total scores was .76. The authors concluded 28
that these results, though promising, provided only modest support for their theoretical model. Revised Teacher Rating Scale for Reactive and Proactive Aggression. Brown, Atkins, Osborne, and Milnamow (1996) sought to address some problems with Dodge and Coie's (1987) teacher rating instrument. First, the 1987 instrument defined proactive aggression with hostile (i.e., anger-driven) aggression items only, although the Dodge and Coie model postulated a relationship between proactive and both instrumental and hostile aggression. Second, items were retained which loaded strongly on both factors in the initial factor analysis. Third, Brown et al. questioned the decision to retain Proactive Aggression as a factor, because of an eigenvalue of less than 1.0. A set of 28 items was created with equal numbers of items representing covert antisocial behavior, proactive aggression, reactive aggression, and prosocial behavior drawn from previously-published instruments. The covert antisocial behavior items were included to provide an index of proactive-instrumental behaviors. A 3-point frequency scale was used (0 = never, 1 = sometimes, 2 = very often). The scale was completed by teachers of 186 third- through fifth-grade boys in public schools in Philadelphia. Students in 15 classrooms were asked to name three of their classmates that they liked, and three that they did not. These peer nominations were converted into standardized "negative peer status" scores (negative nominations minus positive nominations). Data were also collected regarding in-school detentions for inappropriate behavior, physical aggression, disrespectful behavior, verbal aggression, and destruction of property.
29
A factor analysis was performed, excluding the seven prosocial behavior items. The authors' criteria for factor independence were: (a) an eigenvalue greater than 1.00, (b) meets the requirements of Cattell's (1966) scree test, (c) contains a minimum of three items with loadings of .40 or greater, and (d) is theoretically meaningful. Two factors met these criteria. The first (named Proactive Aggression) contained five proactive aggression items and five covert antisocial items. The second (named Reactive Aggression) contained six reactive aggression items. Excellent internal consistency coefficients were found for both of these factors. Five items were left unclassified due to failure to meet various criteria. Negative peer status and in-school detentions were correlated with scale scores, controlling for grade. Both factors correlated significantly with both of these variables. These results seemed to represent an improvement on the Dodge and Coie teacher rating scale (1987), and to support the validity of the proactivereactive dimension of human aggression. Little et al. Survey. Seeking to take a step toward consolidating various theories of the structure of aggression (e.g., Coie and Dodge, 1987, or Frick, 1998), Little, Henrich, Jones, and Hawley (2003) introduced a system of measuring aggressive behavior along two major dimensions: the form or topography (appearance) of the behavior, and its motivation or function. Common forms are labeled as "overt,” "physical,” "indirect,” "verbal,” etc. Functional categories are, for example, "proactive," meaning undertaken with an end in mind, or "defensive" or self-protection. Little et al. argued that the study of the functional aspects of agonistic behavior has been hindered by the need to report the form of the behavior as well as the function in the most commonlyused measures. They gave the following sentence as a prototypical scale item: "I threaten 30
others (overt form) to get what I want (instrumental function)." This presents a possible confound, since predictive relations associated with the functional information may be masked by the form information or because effects may indeed be due to the form rather than the function of the behavior. The authors diagrammed a structure with two functional dimensions, the familiar "instrumental" versus "reactive," and two dimensions of form, "overt" and "relational." While overt aggression is intended to harm another directly, relational aggression is more indirect and is meant to damage another's social standing and relationships. Items included on their 36-item self-report survey were written to fit into one of six subscales: "Pure" overt aggression, reactive overt aggression, instrumental overt aggression, "pure" relational aggression, reactive relational aggression, and instrumental relational aggression. Items consisted of single-sentence statements, such as "I'm the kind of person who often fights with others." Participants rated each statement on a four-point scale for how true they thought it was for them, from "not at all true" to "completely true." The authors admitted that, although overt and relational forms of aggression could be measured independently of function, the functional dimensions could not be operationalized outside of context. There are, therefore, no "pure instrumental" or "pure reactive" items in their survey. They pointed out, however, that they could identify the function of the behavior by controlling for the forms. The survey instrument was administered to 1,723 German students in grades 5 through 10. In addition, both self-rated and other-rated measures of frustration intolerance, victimization, hostility, and negative influence were taken for each participant, along with a self-rated measure of social competence and an informant-rated 31
measure of antisocial behavior. Each of these instruments consisted of six items rated on a four-point scale from "not at all true" to "completely true." The model hypothesized by Little et al. showed good soundness of fit, RMSEA = .061. Factor loadings in their model were all significant (all ps < .01). Latent regressions of the results for the additional six-item instruments were estimated onto relational, overt, reactive, and instrumental aggression. The four dimensions were found to have mostly differentiated patterns of relationships with the results of the additional instruments, which the authors held up as support for the criterion validity of the constructs proposed in their model. As a self-report measure, the usefulness of the Little et al. survey with the intellectually disabled population is limited. However, the effort to pull together various categories and dimensions of aggressive behavior into a more integrated whole could serve as a stepping stone for future efforts at developing aggression measures for use with a broad range of populations, including the intellectually disabled. As stated above, the ability to differentiate instances of reactive from instrumental aggression better could be valuable to both researchers and clinicians. An instrument with this ability and that of more clearly discerning a behavior's form from its function could be even more valuable. Summary The frustration-aggression hypothesis, proposed by drive theorists, sought to explain hostile, reactive aggression. Learning theorists posited that instrumental aggression is acquired and maintained as a result of reinforcing consequences of the behavior. The social information processing theory addresses both reactive and proactive 32
aggression, identifying different points in a six-stage model of information processing where problems may give rise to maladaptive behaviors. Many standardized instruments have been successfully developed for the purpose of gauging various topographies of hostile affect and aggressiveness, including some which have attempted to capture reactive and proactive aggression. Objectives and hypotheses of the current study. The purpose of this project was to develop an informant-based scale for the measurement of the severity and/or frequency of interpersonal aggression among people with intellectual and developmental disability. An effort was made to include scale items which reflected a broad range of topographies, including physical and verbal aggression, as well as hostile affect. Importantly, I also attempted to incorporate a "reactive/proactive" domain as well, for the reasons stated above. There were four objectives for this project as follows: (a) Develop a factoranalytically derived instrument for assessing common forms of aggression in adults with intellectual disability; (b) Compute normative data for the subscales that are obtained, broken out by gender, level of functioning, and age group; and (c) Assess the interrater reliability of the subscales that are derived, and (d) Assess the validity of the new scale. I predicted that factor analysis of the SHARP Problem Scale would lead to a multidimensional instrument which would include subscales representing various types of hostility and aggressiveness. Such multidimensional structures have been consistently found by developers of empirically-derived scales of aggression (e.g., Buss & Durkee, 1957; Kazdin, et al., 1987; Miller et al., 1995). I also hypothesized that an association between higher proactive aggression and psychopathy (Cornell et al., 1996), and a link 33
between higher reactive aggression and impulse-control disorders or Cluster B personality disorders, would be found (American Psychiatric Association, 1994). Finally, I predicted that, given that the functions of the behaviors in question would be disentangled from their forms, correlations between reactive and proactive scores would be relatively low.
34
CHAPTER 3
METHOD Subjects Subjects were 512 individuals receiving services for developmental disabilities, and aged 19 to 84 years, inclusive. The average age of the sample was 41.5 years, with a standard deviation of 12.3. Most of the subjects were male (61.5%). Twenty-four of them (4.7%) were residents of a developmental center, 93 (18.2%) were residents of supported living homes, 350 (68.4%) were employees of sheltered workshops, 43 (8.4%) were enrollees of a living skills center, and two (0.4%) resided in private homes. Functional levels ranged from mild to severe/profound intellectual disability. See Table 1 for a more detailed demographic breakdown of the sample. Raters Written permission to recruit raters and collect data was obtained from administrators for each site. Raters were recruited, using an IRB-approved script and consent form (see Appendix G). Potential raters were informed that they would be paid seven dollars per completed scale packet. Among the 61 raters were one parent, one psychologist, one behavior specialist, nine supported living supervisors, and 49 workshop instructors. Raters had been familiar with subjects for an average of five and a half years, ranging from six months to 25 years. 35
Instruments The Scale of Hostility and Aggression, Reactive/Proactive (SHARP). The initial version of this new instrument (see Appendix A) contained 58 items, each being a short description of some form of aggressive behavior (e.g., "Makes demeaning comments about others in their absence" or "Pulls others' hair"). For each item, a four-point "Problem Scale" was provided, ranging from zero ("Never happens") to three ("Severe and/or very frequent problem"). Where appropriate, a five-point Proactive/Reactive Scale (later named the “Who Starts It?” Scale, and still later renamed the Provocation Scale) was also provided, ranging from -2 ("Only when provoked, not planned, and/or when s/he ‘just loses it;’”) to +2 ("Always the first to act [i.e., always the instigator; the one who 'starts it']"), with a midpoint of zero to indicate "Equally likely to happen with or without provocation." For the purposes of this instrument, "instigator" was defined as " the one who deliberately initiates the conflict; the first or only one to act in an incident of aggression," and "provocation" as "Any action leading to the aggression, no matter how mild (e.g., verbal teasing or being crowded) or severe (e.g., hitting or kicking), which seems to anger or upset the subject and trigger the behavior." Instructions to raters were as follows: "The questions should be answered based on your knowledge (either from personal observation, reliable records, or the accounts of others familiar with the person being rated [the subject]) of the subject's behavior in the past six months. Please first circle ONE 'Problem Scale' answer for each of the 56 items. If this answer is a '1,' '2,' or '3' (i.e., if it is a problem at all), circle ONE ‘Who Starts It?’ answer for the item if applicable." 36
Items for the preliminary Scale of Hostility and Aggression, Reactive/Proactive (SHARP) were developed through consultation with psychology professionals experienced with clients having disruptive behavior problems. An effort was made to include not only descriptions of the more common (and a few somewhat uncommon) examples of physical violence, but also varieties of hostile affect and verbal behavior. Behavior Problems Inventory (BPI), “Aggressive/Destructive Behavior" Subscale The 11-item “Aggressive/Destructive Behavior” subscale of the Behavior Problems Inventory (BPI; Edlinger, 1983; Rojahn, 1984; Rojahn et al., 2000) was selected as a measure of the congruent validity of the SHARP. This choice was made based upon the BPI subscale’s favorable validity and reliability statistics and relative brevity (Rojahn, 1984). BPI items were scored on a five-point frequency scale (never = 0, monthly = 1, weekly = 2, daily = 3, hourly = 4) and on a four-point severity scale (no problem = 0, slight problem = 1, moderate problem = 2, severe problem = 3). To be scored higher than 0, an item must describe a behavior which was observed within 2 months of administration of the instrument. Aggressive/destructive behaviors, in order to be scored greater than 0, had to meet the definition of "abusive, deliberate attacks against other individuals or objects." For a more detailed description of the BPI and its development, see the Literature Review (Chapter 2). Procedure I recruited staff members at the participating sites to be raters. After a list of volunteers was obtained, copies of the SHARP, along with the 11 BPI Aggressive/Destructive Behavior items, were distributed to the raters. After a brief one37
on-one training, I asked each rater to fill out one of the SHARP forms and the BPI subscale for subjects whom she or he knew for at least six months. I asked raters to assign anonymous ID codes to the subjects, using rater initials followed by the subject’s sequence on the rater's alphabetic roster. I also asked raters to complete a modified version of the SHARP face sheet for each subject, with questions regarding the subject’s identification code, month and year of birth, gender, ethnicity, relationship of rater to the subject, length of time the rater was familiar with the subject, and the rater’s estimation of the severity of the subject’s intellectual disability (“Mild,” “Moderate,” “Severe/profound,” or “Unable to guess”), as well as how confident they were in that estimate (“Certain,” “Very confident,” “Somewhat confident,” or “Complete guess”). I was available to all of the raters to answer follow-up questions throughout the data collection period. To assess interrater reliability, I obtained independent ratings by pairs of qualified raters for forty-four subjects. Five of these pairs were across settings. These subjects were rated by workshop staff and program directors of their supported living homes. Thirty-nine pairs were completed by two instructors sharing a production area at one of the day programs. For one subject, both across- and within-setting rating pairs were obtained. Subjects in the interrater sample were 52.6% male. Most (51.3%) functioned in the moderate range of intellectual disability, while 39.5% had been diagnosed as being in the mild range, and 9.2 had tested in the severe/profound range. Their ages ranged from 22 to 63 years, with an average of 41 years.
38
For 103 subjects, I asked legal guardians to give signed permission (IRB and HIPAA consents) to use information regarding their charges’ medications and diagnoses for validity purposes. I presented these consent packets at the appropriate sites and accessed the relevant data and recorded them onto standard forms. This information was later used to examine the external validity of the SHARP.
39
CHAPTER 4
DATA ANALYSIS AND RESULTS Problem Scale Factor analysis. Prior to doing an exploratory factor analysis, I looked at rates of endorsement for each of the 58 Problem Scale items. Four items were eliminated from further analysis, because they were endorsed by fewer than five percent of raters. These items were numbers 20 (“Uses headlocks or other ‘wrestling’ holds against others”), 24 (“Cuts others with a sharp object”), 29 (“Trips others”), and 41 (“Chokes others/Grabs others’ necks”). Endorsement rates ranged from 1.3% (item 24: “Cuts others with a sharp object”) to 57.2% (item 18: “Reacts suddenly or impulsively to minor provocations”). See Table 2 for endorsement rates for all of the items. I conducted exploratory factor analysis of the inter-item correlation matrix of the 54 remaining SHARP Problem Scale items. Using the Comprehensive Exploratory Factor Analysis program (CEFA, Browne, et al., 2004), I selected the Maximum Wishart Likelihood (MWL) discrepancy function. MWL assumes that the sample is from a multivariate normal distribution. An advantage of MWL over Ordinary Least Squares (OLS) is that MWL provides a significance test, which OLS cannot. It is true that OLS is generally more robust, but this is because it does not make any assumptions regarding distribution. Since I expected any factors found to be correlated with one another (as all 40
of the items described some form of hostile or aggressive behaviors), oblique rotation, specifically, CF (Crawford-Ferguson) Equamax, was chosen. My standard for inclusion of an item in a factor was a loading of at least .35. Items were not included on more than one factor. The five-factor solution (which I ultimately adopted as the best overall fit) is presented in Table 3. Factors were named (I) “Verbal Aggression” (n = 9 items retained), (II) “Physical Aggression” (n = 13), (III) “Hostile Affect” (n = 11), (IV) “Covert Aggression” (n = 8), and (V) “Bullying” (n = 7). Forty-eight items were retained, while six failed to meet the inclusion criterion of a factor loading of at least .35, and were therefore dropped from the scale. Those items were: “Sneers, ‘makes faces’ at others,” “Reacts to insults or teasing by lashing out physically,” “Lashes out at people who are in his/her space,” “Glares at others,” “Doesn't care about others' feelings,” and “Does not seem truly sorry after hurting someone.” Mean factor loadings for factors I through V were 0.62, 0.52, 0.54, 0.57, and 0.55, respectively. The sums of squares of the factor loadings were computed (see Table 4). It should be noted that it would be inappropriate to discuss the amount of variance accounted for in factor analysis (Michael Browne, personal communication, April 17, 2008). “Variance accounted for” is a concept relevant to Principal Components Analysis (PCA), as the purpose of PCA is to explain the variances of manifest variables. Factor analysis, on the other hand, attempts to explain the intercorrelations between manifest variables. While the sum of squared factor loadings for each factor can give the total amount of communality in a factor analysis employing orthogonal rotation, this is not
41
possible in EFA using oblique rotation. However, it does give an indication of the strength of the factor loadings. Rather than weight each item according to its factor loading, the 48 retained items were given an equal weight of 1.0. Therefore, groups of items are henceforth referred to not as “factors” but as “subscales.” Average subscale scores were: (I) Verbal Aggression, 3.07 (range 0 – 27; 0.34 average per item); (II) Physical Aggression, 3.13 (range 0 – 31; 0.24 average per item); (III) Hostile Affect, 8.23 (range 0 – 36; 0.75 average per item); (IV) Covert Aggression, 3.60 (range 0 – 20; 0.45 average per item); and (V) Bullying, 2.60 (range 0 – 21; 0.37 average per item). The average Total Score was 20.20 (range 0 – 119; 0.42 average per item). All Pearson correlations between the Problem subscales were significant at the 0.01 level (two-tailed), ranging from Verbal Aggression and Hostile Affect (.65, p = .0009) to Verbal Aggression and Physical Aggression (.23, p = .0009). See Table 5 for the full subscale correlation matrix. Norms for the SHARP Problem Scale and its subscales, broken down by gender, age group, and level of functioning, are presented in Tables 6 through 15. Norms for the entire sample are provided in Table 16. Although I do not pretend that these averages are standardized based on census or other data to reflect the population of people with developmental disabilities, they are the best "normative" figures I can coffer at this point. The gender, age, and level-of-functioning tables are intended to make it possible for workers to identify a specific group if they feel this is appropriate. Reliability. Internal consistency. I calculated item-total correlations for each of the subscales, as well as for the full scale. Ranges (and means) were as follows: (I) Verbal Aggression: 42
0.56 - 0.89 (0.78); (II) Physical Aggression: 0.54 - 0.80 (0.67); (III) Hostile Affect: 0.65 – 0.81(0.73); (IV) Covert Aggression: 0.54 – 0.79 (0.69); (V) Bullying: 0.62 – 0.71 (0.68); and Total Score: 0.29 – 0.74 (0.54). As another measure of internal consistency, alpha coefficients were computed for the full Problem Scale and for each of its subscales. The alpha coefficient for the full scale was .95. Coefficients for the subscales were as follows: .92 for Verbal Aggression, .89 for Physical Aggression, .92 for Hostile Affect, .82 for Covert Aggression, and .80 for Bullying (see Table 17). For comparison purposes, the BPI Aggressive/Destructive Behavior Subscale was found to have alpha coefficients for its frequency and severity totals of .84 and .85, respectively. Interrater reliability. To assess interrater reliability, intraclass correlations (ICCs) were computed for 44 subjects for whom scales were filled out by two or more raters. Intraclass correlations are appropriate in situations such as this, when there is no clear rationale for assigning the paired raters to the X or the Y axis. Because many, but not all, of the pairs were rated by the same two observers, a two-way mixed effects model was used (Shrout & Fleiss, 1979). Excluding the across-settings data, ICCs for the Problem subscales ranged from 0.59 (Bullying) to 0.78 (Physical Aggression). With the acrosssettings data included (n = 44), ICCs ranged from 0.57 (Covert Aggression) to 0.73 (Hostile Affect). See Table 18 for more details. Concurrent Validity. As an index of concurrent validity, scores for the SHARP Total Score and the subscales were compared to frequency and severity scores for the BPI Aggressive/Destructive Behavior subscale (Table 19). Two-tailed Pearson correlations were calculated, all of which were significant at the .01 level or less. The 43
strongest correlation was between the Physical Aggression subscale and the severity score of the BPI subscale (.86.) The weakest was between the Verbal Aggression subscale and the BPI subscale severity score (.33.) The full Problem Scale correlated with the BPI subscale scores at .70 for severity and .67 for frequency. Demographic Variables. I performed t-tests with an eye toward revealing differences between subscale scores and Total Scores for age and gender. To avoid reporting results based on a faulty assumption of heterogeneous variance between the groups being compared, Levene’s test for equality of variances was performed (Levene, 1960) and the appropriate t scores and degrees of freedom are reported. The current study presented a potential problem with multiple comparisons of groups of subjects defined by several demographic and external variables. One way to control for the inflated Type I error rate (i.e., the chance of incorrectly identifying a difference in means as a significant effect) resulting from several such analyses being made with the same database would have been to use Bonferroni-corrected alpha levels. Current practice is to order comparisons by their expected outcomes before calculating Bonferroni corrections. Thus, no power in alpha is lost until a hypothesis is incorrect. Unfortunately, this step was not taken prior to the t-tests being done. Therefore, in lieu of using Bonferroni-adjusted alpha levels, I decided to consider only those results at the p < .01 to be statistically significant (rather than the more commonly-accepted p < .05 level). However, as this is the first study of the SHARP, I did not want to overlook potentially important results which might provide further validation of the instrument, suggest directions for future research, and so on. For this reason, I also reported results meeting 44
the p < .05 criterion (†), with the caveat that they should only be viewed as exploratory. See Table 20 for more detail. Age: Subjects were divided into two age groups along the median of 41.5 years. The younger group (n = 235) received higher scores than the older subjects (n = 268) for Verbal Aggression (3.62 vs. 2.61, p = .029)†, Physical Aggression (4.10 vs. 2.34, p = .001), Hostile Affect (9.42 vs. 7.30, p = .003), Covert Aggression (4.04 vs. 3.21, p = .035)†, Bullying (3.41 vs. 1.90, p = .0009), and Total Score (24.02 vs. 17.05, respectively, p = .0009). Gender: Females scored non-significantly higher on the Verbal Aggression subscale (3.72 vs. 2.65, p = .026)†. Gender differences on Total Score and the other subscale scores were not significant. Functional level: For the three categories of functional level, I used a one-way analyses of variance with an eye toward revealing differences between subscale scores and Total Scores, and I performed Tukey HSD (“Honest Significant Difference”) tests as post hoc analyses. Mean differences, and levels of significance are reported. See Table 20 for more details. Subjects categorized in the mild range of intellectual disability (n = 220) scored higher on the Verbal Aggression (2.81, p = .0009) and Covert Aggression (1.37, p = .02)† subscales than those in the severe/profound category. Those in the moderate I/DD category (n = 172) had higher mean scores on Physical Aggression (1.64, p = .008) than subjects in the Mild I/DD group. They also were rated higher than the severe/profound ___________________________ †
Exploratory (p ≤ .05) 45
category on Verbal Aggression (2.75, p = .0009), Hostile Affect (2.39, p = .036)†, and Covert Aggression (2.20, p = .0009). Subjects functioning in the Severe/Profound range (n = 115) scored higher on Physical Aggression than the Mild I/DD (3.48, p = .0009) and Moderate I/DD (1.84, p = .01) groups. Ethnicity: For each ethnic category, t-tests were done comparing mean scores on Total Score and each of the subscales, grouping subjects as Caucasian versus all others combined. No significant differences on Total Score or subscale scores were found between ethnic categories. External Validity. As discussed in the Literature Review, certain psychiatric diagnoses are associated with aggressive behaviors or expressions of hostility. To examine the external validity of the SHARP, diagnoses and medications were recorded for those individuals in the subject sample for whom I was given guardian consent. These subjects were dubbed the “validity sub-sample” (n = 103). Roughly equal numbers of individuals functioning at the mild, moderate, and severe/profound levels of I/DD were included. The subsample was 60.2 percent male, with an average age of 41.4 years. T-tests were used to compare mean Total Scores and subscale scores of groups within the “validity sub-sample” (n = 103), and significant results are reported here. See Tables 20 and 21 for details. (a) ADHD. Subjects diagnosed with ADHD (n = 4, 3.9%) were found to have higher scores than all others in the validity sample on Covert Aggression: 10.50 vs. 2.83 (p = .0009); and Bullying: 8.50 vs. 2.56 (p = .002). (b) Autism. Subjects with autistic disorder (n = 9, 8.7%) scored higher on Physical ___________________________ †
Exploratory (p ≤ .05) 46
Aggression: 10.44 vs. 5.01 (p = .01), but lower on Verbal Aggression: 0.00 vs. 3.52 (p = .0009); (c) Psychotic Disorder. Scores for subjects with Psychotic Disorder NOS (n = 14, 13.6%) were higher for Hostile Affect: 14.21 vs. 7.62 (p = .008); (d) Down Syndrome. Those with Down Syndrome (n = 11, 10.7%) scored lower on Physical Aggression: 1.09 vs. 6.01 (p = .0009); Hostile Affect: 2.55 vs. 9.23 (p = .0009); Covert Aggression: 0.36 vs. 3.46 (p = .0009); Bullying: 0.55 vs. 3.05 (p = .0009); and Total Score: 7.09 vs. 23.68 (p = .0009). Data for subjects with certain diagnoses were aggregated into supra-ordinate variables. This was done for the following: (a) Combined Behavior Disorders (a combination of those with Intermittent/Explosive Disorder and Disruptive Behavior Disorder NOS), (b) Autism Spectrum Disorders (Autism plus PDD NOS), (c) Combined Anxiety Disorders (Anxiety Disorder NOS plus Obsessive/Compulsive Disorder), (d) Combined Personality Disorders (Personality Disorder NOS plus Borderline Personality Disorder), and (e) Combined Mood Disorders (Mood Disorder NOS, Bipolar Disorder, and Depression). There were 11 subjects in the Autism Spectrum Disorders category (10.7%). They scored lower than the remainder on Verbal Aggression: 0.45 vs. 3.54 (p = .0009). All other noteworthy results obtained for the supra-ordinate variables were only significant at the exploratory level of significance. Please see Table 23 and 24 for more details about the exploratory results. Next, differences in mean scores for subjects receiving various classes of psychoactive medications and those not receiving these medicines were investigated 47
(Table 25 and 26). Subjects receiving antipsychotic drugs (n = 50, 48.5%) scored higher than the remaining 53 subjects on Physical Aggression: 8.92 vs. 2.25 (p = .0009); Hostile Affect: 11.96 vs. 5.26 (p = .0009); and Total Score: 30.88 vs. 14.35 (p = .0009). Subjects receiving stimulants (n = 6, 5.8%) scored higher on Physical Aggression: 13.00 vs. 5.02 (p = .003). Finally, those taking any psychoactive drug (n = 77, 74.7%) scored higher than the remaining people in the sample on Physical Aggression: 6.69 vs. 1.92 (p = .0009); Hostile Affect: 9.79 vs. 4.73 (p = .002); Covert Aggression: 3.66 vs.1.54 (p = .004); and Total Score: 25.55 vs. 11.15 (p = .0009). Provocation Scale Of the 48 SHARP Problem Scale items, 26 have corresponding Provocation Scale items. This “secondary” scale presents some unusual challenges to analysis. First, the Provocation Scale is conditional (that is, an item is completed only if the corresponding Problem Scale item is endorsed with a one or higher). Therefore, factor analysis would be inappropriate, since raters completed different sets of items. Second, because not all Problem items are paired with a Provocation item, some Provocation subscales included only one or two items. Third, since items may legitimately be left incomplete (if the item is rated “0” on the Problem Scale), missing values had to be dealt with when analyzing responses. For this study, I selected a function of SPSS which replaced all missing values with their series means (i.e., mean of responses to the variable across all subjects). Finally, there is the question of how best to analyze and present results for the Provocation Scale. Several possibilities were considered. For example, should separate totals of negative and positive scores, which would supposedly give one a “reactive” and a “proactive” figure for each subject, be used? Might it be best to examine what, if any, 48
interaction there may be between the Problem and Provocation scores? After much consideration, we decided that the following indices would be investigated: (a) The raw total of all Provocation Scale items endorsed (“Provocation Total”); (b) The raw totals for each of the Provocation subscales; (c) The sums of endorsed Provocation items and their corresponding Problem items (“Provocation/Problem Sum;” PPS); (d) The Provocation/Problem Sums for each of the subscales; (e) The products of endorsed Provocation items and their corresponding Problem items (‘Provocation/Problem Product;” PPP); (f) The Provocation/Problem Products for each of the subscales; (g) A “Provocation Quotient,” which was originally proposed by Cristan Farmer (personal communication, 1/25/2008). This figure was computed by first adding all positive Provocation ratings (i.e., +1 or higher) to find the proactive total (P), then adding all negative Provocation ratings (-1 or lower) for the reactive total (R). The resulting sums were then entered into this equation: (P i + 1/ R i + 1), which yielded Q i , the quotient score. Values of 1.0 were added to the sums to eliminate the possibility of division by zero. Thus, the Q score represents the ratio between proactive and reactive aggression. A Q of 1.0 suggests that a subject who displayed an equal amount of proactive and reactive aggression. A Q greater than 1.0 indicates that the subject had a greater tendency toward proactive aggression, while a Q less than 1.0 indicates someone more likely to engage in reactive aggression. (h) Quotient scores for each of the Provocation subscales. 49
Reliability. As the Provocation Scale is conditionally endorsed, indices of internal consistency such as alpha coefficients or item-total correlations are inappropriate. Such analyses assume that an equal number of items have been completed for each subject. Intraclass correlations were again used to assess interrater reliability. Excluding the handful of across-settings ratings from analysis, ICCs for the raw Provocation Subscales ranged from .54 (Verbal Aggression) to .78 (Physical Aggression). With the across-settings data included, ICCs ranged from .46 (Bullying) to .87 (Physical Aggression). Without the across-settings pairs, an ICC of .55 was found for the full Provocation Scale, and an ICC of .66 was found with the across-setting ratings in the analysis. See Table 29 for more details. Interrater ICCs for the Total Quotient score (Table 30) were -.13 and -.03 (with across-settings pairs excluded and included, respectively). The subscale Quotient scores ranged from -.19 (Verbal Aggression) to .66 (Bullying) with no across-settings pairs. Interrater ICCs ranged for -.32 (Verbal Aggression) to .38 (Physical Aggression) for the comparisons including raters across-settings. Interrater ICCs for Provocation/Problem Sum (Table 31) for the full scale were .33 and .45 (with across-settings pairs excluded and included, respectively). For the subscales, ICCs ranged from -.14 (Bullying) to .74 (Covert Aggression) (single setting only), and from .27 (Verbal Aggression) to .65 (Physical Aggression) in comparisons allowing raters across settings. For the full scale Provocation/Problem Product (Table 32), interrater ICCs were .07 and .04 (with across-settings pairs excluded and included, respectively). The weakest interrater ICC for any of the subscale Provocation/Problem Products was -.29 (Verbal 50
Aggression) and the strongest was .75 (Bullying), both found with the across-settings pairs excluded. When cross-setting ratings were included, the weakest ICC was -.10 (Verbal Aggression) and the strongest was .62 (Bullying). Selection of a Provocation Index. At this point, I simplified matters by choosing one of the Provocation indices before moving on to reporting analyses and results related to the validity of the Provocation data. My criteria for making this selection were: (a) correlations between the Provocation index and Problem Scale totals (see Table 33); (b) reliability statistics of each index; and (c) unique effects of demographic variables and external validators detected by each of the indices (see below). Correlations with the Problem Scale were relevant because they provided an index of whether or not a Provocation index was superfluous, and, if so, to what extent. Because the Provocation Scale items were only endorsed contingent upon the corresponding Problem Likert being rated a “1” or higher, one might expect to find significant correlations between the Problem and Provocation scores. This was indeed so. All correlations between these scores were significant at the .01 level. However, the nearer these correlations approach 1.0, the more an index would appear to be merely “echoing” the information given by the Problem Scale, and the less useful and informative the index would seem to be. This would be similar to the very high correlations usually found between measures of severity and frequency of behaviors when instruments use two separate Likert scales to gauge these dimensions of the same set of items. Havercamp and Reiss (1996) convincingly argued that such high correlations render separate frequency and severity ratings redundant. The raw Provocation Scale total had a notably lower correlation with the Problem Scale total (r = 51
0.17) than did the other Provocation index totals (PPS: r = 0.33; PPP: r = 0.78; and Quotient: r = 0.33). Without question, the raw Provocation Scale was found to have the best interrater reliability of any of the indices (Table 28). This was true not only in a comparison of results for total scores, but also in comparing those for the subscales. In fact, each of the other indices had negative ICCs for at least one of their subscales. Reliability figures for the Total Quotient, PPP Total, and their respective subscales were so low (most ICCs falling below Cicchetti’s [1994] 0.40 criterion of acceptability) that I eliminated these Provocation indices from further consideration. As reported in Tables 20, 21, 24, and 26, when the validity sub-sample was split by demographic variables or the presence or absence of certain diagnoses or medications, each of the Provocation indices yielded some significant t-test results which were unique from the others. Since Provocation/Problem Sum (PPS) was a combination of information from both the Problem Scale and raw Provocation Scale, I was interested in whether it provided results which would not have been found using only the Problem Scale or Provocation Scale data. For seven of the grouping variables, significant differences were found in mean PPS scores which were not found for either the Problem or simple Provocation scores. In the end, I did not determine that PPS was preferable to the simple Provocation Scale. Although it was sensitive to some differences between subjects grouped by demographic, diagnostic, or medication variables, its comparatively poor reliability statistics made it an unattractive choice. From this point forward, this paper will confine
52
itself to the simple, or raw, Provocation scores when discussing this portion of the SHARP. Also, before continuing, please note that negative, or, at least, relatively lower, scores were interpreted as reactive, while higher scores were interpreted as proactive. Tables 34 through 43 provide mean scores for the Provocation Scale and its subscales, broken down by gender, age groups, and level of functioning. Table 44 provides mean scores for the entire sample. Validity. Effects of demographic variables. As I did with the Problem Scale, I employed t-tests to examine differences between Provocation Total and subscale Provocation scores for age and gender, and ANOVA with Tukey HSD post hoc tests for level of functioning (see Table 20). Significant and exploratory results are reported here. Age. Subjects were again split into groups along the median age of 41.5 years. The older group (n = 268) received lower (presumably more reactive) scores than the younger subjects (n = 235) for Verbal Aggression (3.77 vs. 4.50, p = .021)†, Physical Aggression (1.06 vs. 1.75, p = .035)†, Hostile Affect (-4.26 vs. -3.86, p = .01), Bullying (9.49 vs. 9.70, p = .031)†, and Provocation Total Score (6.46 vs. 8.39, respectively, p = .007). Gender. There was a non-significant difference in the mean Covert Aggression subscale scores given to women and men (7.15 vs. 7.24, respectively, p = .048).† Functional Level. Subjects classified in the mild range of intellectual disability ___________________________ †
Exploratory (p ≤ .05) 53
were scored as more reactive on the Verbal Aggression (-1.18, p = .003), and the Hostile Affect (-0.51, p = .01) subscales, as well as Provocation Total Score (-2.29, p = .01), than those in the moderate category. Compared with the severe/profound category, people in the mild group were rated as more reactive on Physical Aggression (-1.21, p = .01), Hostile Affect (-.55, p = .016)†, and Total Score (-.61, p = .01). Table 20 provides more details. External Validation Analyses for Provocation Scores. I again compared the mean scores of groups within the validity sub-sample when grouped by medication regimen or diagnoses (see Tables 21, 23, 24, 26, and 28). Impulse Control Disorder NOS. Those with Impulse Control Disorder NOS (n = 9, 8.7%) were scored as more reactive than all other subjects in the sub-sample on Verbal Aggression (-3.04 vs. -0.08, p = .009), Hostile Affect (-1.44 vs. 0.02, p = .005), and Total Score (-2.24 vs. 8.91, p = .0009). Personality Disorder NOS. Comparatively more reactive scores were assigned to those with Personality Disorder NOS (n = 7, 6.8%). Mean scores for this group were more negative than those for the remaining sub-sample on Physical Aggression (-0.17 vs. 6.70, p = .002), and Total Score (-0.89 vs. 8.58, p = .005). Combined Personality Disorders. This category (n = 9, 8.7%) was more reactive on Physical Aggression than the remainder of the sample (1.18 vs. 6.72, respectively, p = .006), and Total Score (0.61 vs. 8.64, respectively, p = .007). ___________________________ †
Exploratory (p ≤ .05)
54
Medication and Provocation Scores. Findings for comparison between drugs appear in Tables 26 and 28. Subjects taking beta blockers (n = 4, 3.9%) were rated more reactive on Physical Aggression (1.89 vs. 6.57, p = .004), Hostile Affect (-2.28 vs. -0.02, p = .003), and Total Score (4.20 vs. 8.43, p = .004).
55
CHAPTER 5
DISCUSSION Problem Scale of the SHARP As predicted, exploratory factor analysis yielded a multidimensional instrument. The five-factor structure of the SHARP Problem Scale conformed fairly well to my clinical expectations. The model had reasonably good fit, and its factors had good reliability results. The matrix of correlations between the subscales (Table 5) supported the structure’s convergent and divergent validity, with the lowest correlation being that between Verbal Aggression and Physical Aggression (constructs which, while related, need not always occur together) and the highest being that between Verbal Aggression and Hostile Affect (the former so often being the expression of the latter). As stated earlier, I had hoped to develop an instrument which was capable of capturing a variety of types of aggression and expressions of hostility. This goal appears to have been met. The Verbal Aggression and Physical Aggression subscales included varied descriptions of topographies within these two broad categories of behavior. Items in the Hostile Affect subscale reflected many of the more commonly-seen expressions of hostility, as well as impulsivity and attitudes of resentment and defensiveness. Taken together, the items in the Covert Aggression subscale gave a picture of someone who not only seeks to hide, or escape the consequences of, his or her harmful acts, but of one who 56
also attempts to manipulate others by influencing them to act in inappropriate ways. Finally, the Bullying items described someone who seeks to intimidate or take advantage of others. It is noteworthy that Bullying was more strongly correlated with Covert Aggression than with any of the other subscales. Both comprised mostly items which one would associate with manipulative and/or goal-driven (i.e., proactive) behavior, and were perhaps differentiated only by the level of “artfulness” and subtlety the subject was willing or able to employ. A number of different types of oblique rotation (Varimax, Oblimin, Parsimax, and Quartimax) were also tried, but a five-factor solution using CF Equamax rotation was chosen. The choice was made on the basis of interpretability and goodness-of-fit as indicated by Root Mean Square Error of Approximation (RMSEA; Browne & Cudeck, 1992.) The RMSEA point estimate for this five-factor model was .068, considered to be a reasonably good fit (Browne & Cudeck, 1992). Structures of three or fewer factors had RMSEA in the “mediocre” range. Four- and six-factor solutions were found to have similar RMSEA (.072 and .065, respectively). However, the four-factor solution failed to differentiate between items describing covert and bullying behaviors, and the six-factor structure included a superfluous factor consisting of only two items, item 36: “If caught, denies having behaved badly” and item 45: “If caught, makes excuses for bad behavior.” One of the reasons that equamax rotation was selected in the end was that it placed items 36 and 45 in the more appropriate Covert Aggression subscale, rather than Hostile Affect, unlike the other types of oblique rotation that I had tried. It seemed to me that trying to deny or excuse one’s bad behavior was much more specific to being sneaky or covert than to simply displaying negative affect. 57
Looking at the factor structure at the item-level, one finds that most of the items sorted into subscales where they made the most sense, from the standpoint of interpretability, and they had sufficiently robust factor loadings. There were, of course, a few exceptions. Item 52 (“Doesn’t care about others’ feelings”), a description likely to be associated with hostile affect, or manipulative, bullying behavior, loaded most strongly on the Physical Aggression factor. It failed to meet the (admittedly arbitrary) inclusion criterion of a factor loading of at least .35. A similar item, “Does not seem truly sorry after hurting someone,” loaded on the much more appropriate Bullying subscale, but its loading was only .27. This is disappointing, because these items were included in the original pool in hopes of helping to identify subjects who were lacking empathy, and perhaps, therefore, who would be more likely to engage in instrumental aggression. Item 1, “Sneers, ‘makes faces’ at others,” loaded equally on both Verbal Aggression and Bullying, but it too failed to make the arbitrary cut. Three items, “Reacts to insults or teasing by lashing out physically,” “Lashes out at people who are in his/her space,” and “Glares at others,” loaded in the Hostile Affect subscale but fell just short of the inclusion criterion. A children’s version of the SHARP, dubbed the “C-SHARP,” was created by Cristan Farmer (2008). Criteria for inclusion in Farmer’s subject sample (n = 372, out of an original mailing of 1,225; 30.4%) were that a child be between the ages of 5 and 21 years, inclusive, and have evidence of developmental disability. Scales and demographic questionnaires were filled out by the subjects’ parents. The original set of items used for the C-SHARP was identical to that used in the development of the SHARP, and the resulting factor structures were quite similar. In fact, the C-SHARP’s five subscales 58
appeared to represent essentially the same constructs as those measured by the SHARP, and were therefore identically named. The greatest amount of overlap between the two instruments was found in their Physical Aggression subscales, where nine of the ten items (90%) in the SHARP version also appeared in its C-SHARP counterpart. The Bullying subscales shared the fewest items, with only two of the SHARP’s seven items (28%) also being found in the C-SHARP version. The C-SHARP Bullying subscale pulled together most of the items describing harmful acts involving objects, such as Item 17 (“Throws objects at others”) or the two items denoting theft, whereas such items were distributed across the Physical Aggression, Covert Aggression, and Bullying components of the adult version. It is likely that developmental constructs, not yet matured in the Farmer (2008) sample, may account for the differences. Given that the subjects, raters, and sampling strategies used in developing the two instruments were very different, it is encouraging to compare their factor structures. I believe that the degree of overlap tends to support the construct validity of each. Nevertheless, further studies of the SHARP are needed to verify its construct validity and, hopefully, its other psychometric characteristics as well. Reliability. Internal consistency. To address the question of the SHARP’s internal consistency (i.e., whether its items were sufficiently interrelated to support their combination in subscales), Cronbach’s alpha coefficients were computed. Alpha increases when the correlations between items increase. Alpha levels for the subscales were good to excellent (Cicchetti, 1994), ranging from 0.80 for Bullying to 0.93 for
59
Hostile Affect. Alpha for the full scale, .96, was also excellent. Mean item-total correlations for the subscales, ranging from .67 to .78, were high. Interrater reliability. The intraclass correlation coefficient (ICC) is commonly used to measure interrater reliability when there are multiple raters, and it determines how much of the variance between raters is due to variance among the subjects on the attribute being measured and how much is due to measurement error. Considering only the 39 same-setting pairs of raters, ICCs were 0.70 (Verbal Aggression), 0.78 (Physical Aggression), 0.75 (Hostile Affect), 0.72 (Covert Aggression), and 0.59 (Bullying). According to Cicchetti (1994), the level of consistency of the ICCs extended from fair to excellent. With the five across-settings pairs included in the analyses, ICCs were 0.61 (Verbal Aggression), 0.72 (Physical Aggression), 0.73 (Hostile Affect), 0.57 (Covert Aggression), and 0.70 (Bullying), which ranged from fair to good clinical significance. It is unsurprising that Bullying and Covert Aggression performed the most poorly. It seems intuitively compelling that covert behavior is not as visible as the other sorts of acts measured by the SHARP. It is easy to imagine that one of a pair of raters might be more attuned to a subject’s subtler behaviors than the other. Likewise, the items in the Bullying subscale, such as “Tickles or otherwise physically teases others, even after being asked to stop,” might be more open to subjective interpretation than most of those in the other subscales. Also, subjects engaging in bullying (i.e., planned, manipulative, behaviors) may be likely to attempt to hide their actions from our raters, all of whom were in positions of authority over them.
60
Validity. Concurrent validity. Correlations between the SHARP Problem subscales and the Behavior Problems Inventory’s Aggressive/Destructive Behavior subscale were highest with Physical Aggression and lowest with Verbal Aggression. This makes sense, as the BPI subscale focuses almost exclusively on physical acts. However, two of the BPI items were most highly correlated with SHARP subscales other than Physical Aggression. BPI item 9, “Being verbally abusive with others” was most highly correlated with the SHARP Verbal Aggression subscale (0.72 with the severity rating, 0.75 with the frequency rating), and BPI item 11, “Being mean or cruel (e.g., grabbing toys or food from others, bullying others),” correlated most strongly with the SHARP Bullying subscale (0.56 with the severity rating, 0.52 with the frequency rating). Effects of demographic variables. Subjects under the sample’s median age of 41.5 years were given significantly higher scores on Physical Aggression, Hostile Affect, Bullying, and Total Score than the older subjects. At the exploratory level, younger subjects were also given higher scores on the Verbal Aggression and Covert Aggression subscales than the older subjects. This is consistent with several studies of aggressive behavior across the lifespan (e.g., Green & Stacey, 1967; Harris & Knight-Bohnhoff, 1996; Tyrer et al., 2006), which found that people’s tendency to act out violently or express hostility decreased as they “mellowed with age.” It is also possible that effects were related to a cohort difference in the reasons subjects were employed at sheltered workshops (where the majority of data were collected), rather than in the general community. Older people may have been working there due to their having spent much of their lives in institutional settings, and were therefore poorly prepared for less 61
structured work and social environments (Fotheringham et al., 1993). Younger subjects, on the other hand, may have been more likely to be in a sheltered workshop because problems with inappropriate behaviors prevented them from being considered eligible for community employment (Block & Rizzo, 1995). Subjects functioning in the mild and moderate ranges of intellectual disability were given significantly higher scores on the Verbal Aggression subscale and, at the exploratory level, Covert Aggression, than those functioning in the severe and profound levels. The Moderate I/DD group scored significantly higher on Physical Aggression than the Mild I/DD subjects, and higher on Covert Aggression than those functioning in the severe/profound I/DD range. The Severe/Profound I/DD subjects scored significantly higher than the Mild I/DD category on Physical Aggression. Clearly, these results suggested a trend starting with more verbal, expressive behaviors, and “crafty,” planned acts, among subjects with higher IQs and better adaptive behaviors, and moving toward more physical violence among the subjects with more serious disability. Previous researchers also found similar differences between subjects of varying levels of impairment and the severity and/or types of aggression they displayed (Hill & Bruininks, 1984; McClintock et al., 2003; McGrew et al., 1991; Rojahn et al., 2001; Tyrer et al., 2006). Assuming that most subjects with little or no verbal communication skills were in the Severe/Profound I/DD group, it stands to reason that this group would resort more often to physically acting out to make unmet wants or needs known, to express frustration, and so forth. For similar reasons, people with moderate impairment might be more likely than those with milder deficits to physically aggress. Finally, because of their greater verbal skills and recognition of ways to avoid detection and unwanted 62
consequences, the Mild I/DD group was scored more highly than the others in verbal and covert forms of aggression. At the exploratory level of significance, women were found to have higher scores than men on the Verbal Aggression subscale. While this is consistent with some of the literature concerning sex differences in aggressive behaviors (Frodi et al., 1977; Maccoby & Jacklin, 1974), many studies have either found no significant differences between the genders in verbal expression of hostility (e.g., Tasse' et al., 1996), or a higher amount of verbal aggression in men (e.g., Kazdin,1987). Others (Bjorkqvist, 1994; Crick & Grotpeter, 1995) have proposed that many investigators have failed to recognize that girls and women aren’t simply more verbally aggressive, but are actually more likely to engage in “indirect,” or “relational,” aggression, and that certain forms of verbal aggression were a subset of this broader category. Out of curiosity, I added my subjects’ scores on the SHARP items that I considered most representative of relational aggression, items 31 (“Makes insulting comments about others behind their backs”), 40 (“Says ‘I hate [someone not present]’ or makes other hurtful statements in absence of the person concerned”), and 49 (“Encourages others to gang up on someone [physically OR verbally]”) into a single variable. I then conducted another t-test with this sum as the test variable and gender as the grouping variable. Women were indeed rated as significantly higher than men on this makeshift “relational aggression” variable (1.12 vs. 0.74, respectively, p = .015). An analysis of covariance, using the "relational aggression" variable as a covariate (thus removing its effect), resulted in no significant gender difference on Verbal Aggression (p = .571). Of course, items 31, 40, and 49 are not an
63
empirically-derived factor, nor are they even a subset of one, so the result of this exercise should be treated with due skepticism. External Validity Psychiatric Diagnoses and Problem Scale Scores. The external validity of the SHARP Problem Scale was supported by the results of t-tests I performed using various psychiatric diagnoses as grouping variables. ADHD. Subjects with ADHD were found to have significantly higher mean scores in Covert Aggression and Bullying, and, at the exploratory level, Hostile Affect, than those without. While the literature regarding adults with ADHD is far less extensive than that for children, there are studies which have found that adults with ADHD have a greater likelihood than the general population of engaging in antisocial behaviors (McKay & Halperin, 2001; Young et al., 2003). Ingram, Hechtman, and Morgenstern (1999), in a review of research relevant to intellectually disabled adults with ADHD, noted a pattern of higher aggressiveness in adults with ADHD. Autism Spectrum Disorder. Given the language impairment and lack of emotional expression common to autism, it is not surprising that subjects with autism, as well as those in the “Autism Spectrum Disorder” category (i.e., subjects with autism plus those with PDD NOS), had significantly lower scores in Verbal Aggression than the remainder. At the exploratory level, subjects with autism were rated more highly in Physical Aggression than the rest of the sample. This is consistent with the literature (Bhaumik et al., 1997; McClintock et al., 2003). After doing a meta-analysis of prevalence and cohort studies conducted over the previous three decades to identify risk markers for challenging behaviors in people with intellectual disability, McClintock, 64
Hall, and Oliver (2003) found that people with a diagnosis of autism were significantly more likely than others to engage in aggression and other disruptive acts. Researchers have hypothesized that the heightened aggressiveness often found in autistic children and adults may be due to the disorder’s characteristic deficits in communication (Koegel et al., 1994; Rutter, 1978). According to this school of thought, people with autism, lacking effective functional communication skills, turn to expressing themselves in a maladaptive, physical manner. Psychotic Disorder NOS. Subjects with Psychotic Disorder NOS were rated significantly higher for Hostile Affect than the remaining subjects in the sample. They were rated higher on Bullying at the exploratory level. The specific items in the SHARP Hostile Affect and Bullying subscales on which subjects diagnosed with Psychotic Disorder NOS were given particularly high scores were: item 7(“Is quick to anger [‘hotheaded’]”), item 9 (“Broods, pouts, or is sullen”), item 12 (“Crowds others [invades their personal space]”), and item 13 (“Intimidates others”). These were, in fact, the only items in those subscales which were significantly correlated with Psychotic Disorder NOS. Items 9, 12, and 13 could be related to the deficits in social integration which are characteristic of schizophrenia and other psychotic disorders (American Psychiatric Association, 1994). All four items were also compatible with the prevailing literature regarding psychotic individuals’ problems with misreading social cues and anger dysregulation (Novaco, 2007). Delusional symptoms have been linked to increased anger and violence in men with psychotic disorders (Buchanan, 1997). Some have hypothesized that perhaps people with psychotic disorders respond angrily to misperceived threats, the result of poor facial affect recognition (Green et al., 2007; 65
Weiss et al., 2006). Weiss et al. (2006) tested the abilities of schizophrenic inpatients to recognize facial expressions, then reviewed the criminal histories of the people in their sample. They found a link between poor perception of affect and a history of criminal conduct. Several studies (Estoff & Zimmer, 1994; Novaco, 1994; Krakowski & Czobor, 1994) have found that persistent angry and/or violent conduct in schizophrenic patients is predicted by comorbid neurological impairment. In their Synopsis of Psychiatry, Kaplan and Sadock wrote that, “…neurological impairment increases as intellectual impairment increases,” (1972, p.1140). Given the high prevalence of neurological impairment in the intellectually disabled, this could be of particular relevance to the current study. Down Syndrome. Subjects with Down syndrome were rated significantly lower than the rest of the sample on all of the subscales except Verbal Aggression. This result provided more support for the external validity of the SHARP. Tyrer et al. (2006), after interviewing a large sample of caregivers of intellectually disabled adults, found that people with Down syndrome displayed less physical aggression on average than the rest of the sample. Other researchers (Chapman & Hesketh, 2000; Collacott et al., 1998) have found fewer maladaptive behaviors in people with Down syndrome. I also found interesting results for subjects diagnosed with mood disorders, personality disorders, and combined behavior disorders, albeit at the "p = .05" nonsignificant, exploratory level. Mood Disorders. People with a diagnosis of Bipolar Disorder or Mood Disorder NOS, and those in the Combined Mood Disorder category (Bipolar Disorder, Depression, and Mood Disorder NOS combined), were scored higher than the rest of the sample on Hostile Affect. Previous researchers have reported findings supportive of a link between 66
depressed mood and irritability (Charlot et al., 1993). Rudolph and Clark (2001) asked their subjects to complete measures of their general impressions of their relationships with their peers and of their level of depressive symptoms. They found that subjects with higher levels of depressive symptoms expressed significantly more hostile, negative conceptions of their peers than did non-symptomatic subjects. In addition to the higher scores on Hostile Affect, the Bipolar Disorder group was also rated higher than the remainder on Verbal Aggression and Physical Aggression. Goodwin and Jamison (1990) reported that 80% of the bipolar patients they surveyed displayed increased irritability and aggressiveness in manic phases. In a review of the relevant research, Spencer et al. (2001) stressed that clinicians needed to recognize the vulnerability of people with bipolar disorder to comorbid disruptive behavior disorders, such as conduct disorder or intermittent/explosive disorder. Garno, Gunawardane, and Goldberg (2008) found that comorbid borderline personality disorder and current manic and depressive symptoms each significantly predicted aggression in individuals with bipolar disorder. Personality Disorders. Comparatively higher Hostile Affect scores were given to subjects in the Personality Disorder NOS and Combined Personality Disorders (Personality Disorder NOS plus Borderline Personality Disorder) groups. The Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV; American Psychiatric Association, 1994) Cluster B personality disorders, which are typified by disturbance in impulse control and emotional dysregulation, include borderline-, narcissistic-, histrionic- and antisocial personality disorders. Unfortunately, other than the few diagnosed with borderline personality disorder (BPD), the sample did not include any people with a personality disorder diagnosis more specific than “NOS.” Zanarini et 67
al. (1998) found that subjects with BPD scored higher than controls on a scale of dysphoric affective states, particularly for feelings of being betrayed and victimized. Davidson, Jackson, and Kalin (2000) found brain mechanisms underlying the impulsivity, mood instability, aggression, anger, and negative emotion in BPD subjects. The results of another study suggested that people predisposed to impulsive aggression have impaired regulation of the neural circuits that modulate emotion (Davidson, Putnam, & Larson, 2000). Combined Behavior Disorders. Subjects in this category, which combined Intermittent/Explosive Disorder, and Disruptive Behavior Disorder NOS, were rated more highly than the remainder of the sample on Physical Aggression, Hostile Affect, and Bullying. The DSM-IV (American Psychiatric Association, 1994) lists as diagnostic criteria for Intermittent/Explosive Disorder, “…episodes of failure to resist aggressive impulses that result in serious assaultive acts…,” and, “…degree of aggressiveness expressed during episodes is grossly out of proportion to any precipitating psychosocial stressors.” Among the diagnostic criteria for the disruptive behavior disorders are impulsivity, fits of temper, physical violence, bullying, etc. Psychiatric medications and Problem Scale Scores. Subjects receiving antipsychotic drugs were scored significantly higher than the remaining subjects on Physical Aggression, Hostile Affect, and Total Score. An antipsychotic prescription in the sample was most strongly correlated with bipolar disorder, followed by intermittent/explosive disorder and psychotic disorder NOS. To review, people with each of those diagnoses were scored higher on Hostile Affect, and those with
68
intermittent/explosive disorder (as part of Combined Behavior Disorders) and bipolar disorder were also scored higher on Physical Aggression. Subjects receiving stimulants scored significantly higher on Physical Aggression. Most stimulant prescriptions in the sample went to individuals diagnosed with ADHD and/or autism. Again, subjects with autism were rated more highly on Physical Aggressive. Duggal (2007) noted that hyperactivity, short attention span, maladaptive behaviors, and impulsivity are, in addition to being core symptoms or correlates of ADHD, associated with autism (American Psychiatric Association, 2000; Arnold et al. 2006), and this overlap could make medicinal treatment of comorbid autism and ADHD especially challenging. At the exploratory level, those with lithium prescriptions were scored higher on Hostile Affect. The diagnosis most highly correlated with lithium use in my sample was bipolar disorder. Again, the portion of the sample diagnosed with bipolar disorder was rated more highly on Hostile Affect than the remainder. Subjects prescribed any psychoactive drug were scored significantly higher than the remaining people in the sample on Physical Aggression, Hostile Affect, and Covert Aggression. Robertson et al. (2000) reported that the presence of challenging behaviors was the best predictor of psychotropic drug use, especially antipsychotic medications. Holden and Gitlesen (2004) found that over half of the psychoactive prescriptions given to the intellectually disabled adults that made up their sample were indicated for problem behaviors, physical aggression being the most common indication by far.
69
Summary. The Problem Scale portion of the SHARP had a factor structure with reasonably good fit and good face validity. Its subscales had good to excellent internal consistency and fair to excellent interrater reliability statistics. Its concurrent validity was supported by comparisons with the results of the BPI Aggressive/Destructive Behavior subscale. Comparisons of mean scores of groups of subjects divided by demographic, diagnostic, and medication variables yielded results which were supportive of the SHARP’s external validity. Provocation Scale of the SHARP The goal in creating the Provocation Scale was, of course, to attempt to gauge reactive and proactive aggression. Recent research (Polman, Orobio de Castro, Koops, van Boxtel, Merk, 2007) suggested that best results were obtained when the form (topography) and (in this case, instrumental versus reactive) function of aggressive behaviors were measured separately or “disentangled” (Little et al., 2003). To that end, I wanted to ask raters for a separate reactive/proactive rating for each of the applicable items on the Problem Scale. To get at this dimension of the behaviors in question, the presence or absence of provocation was chosen as its most salient defining feature. I thought that this would be less subjective than a rating of the perpetrator’s emotional state upon acting and more clearly recognizable than the presence or absence of some motivating goal. The presence of provocation suggested reactive aggression, while absence of provocation was used as an index of proactive aggression. Twenty-six of the 48 items in the final SHARP Problem Scale factor structure had, in addition to the four-point Problem rating, five-point Likert scales to assess 70
whether or not the behavior described, if displayed, was more or less likely to have been provoked in some way. Because Provocation items were endorsed only if their corresponding Problem items were rated a “1” or above, varying numbers of these items were completed by the raters. This rendered factor analysis and measures of internal consistency inappropriate. Provocation subscales, therefore, were not directly derived from exploratory factor analysis. They were simply those sets of Provocation items which accompanied items in a particular Problem subscale. It should be noted that the Covert Aggression subscale of the Provocation Scale consisted of a single item, item 26: “Steals from others when they aren’t looking.” Table 45 contains a matrix of correlations between the Problem and Provocation subscales. Reliability Interrater reliability. Intraclass correlation coefficients for the Provocation subscales, excluding across-setting pairs, were 0.54 (Verbal Aggression), 0.78 (Physical Aggression), 0.71 (Hostile Affect), 0.65 (Covert Aggression), and 0.59 (Bullying). With the across-setting ratings included, subscale ICCs were 0.52 (Verbal Aggression), 0.87 (Physical Aggression), 0.67 (Hostile Affect), 0.51 (Covert Aggression), and 0.46 (Bullying). ICCs for the Provocation Total Score were 0.55 and 0.66, without and with the across-setting pairs, respectively. These coefficients were considered to be in the fair to good range for rater agreement (Cicchetti, 1994). When one considers that Provocation items were conditionally endorsed, these numbers become a bit more impressive.
71
Validity Effects of Demographic Variables. Older subjects were given significantly lower, presumably more reactive, scores than younger subjects for Hostile Affect and Provocation Total Score, and, at the exploratory level, Verbal Aggression, Physical Aggression, and Bullying. Subjects classified in the mild range of intellectual disability were scored as more reactive on the Verbal Aggression than those in the moderate category (Table 20 provides more details). From the standpoint of social information processing theory (Crick & Dodge, 1996), these results are difficult to explain, even counterintuitive. I could find nothing in the literature to suggest that people over age 40 might be significantly more likely to misinterpret social cues, or to be more impulsive, than younger people. As for the differences across levels of functioning, one might expect people who are less severely impaired to be less likely to engage in reactive aggression, because they are better able to interpret non-threatening social cues as such. The results made more sense if one took into account the social environments in which the observations were made. Older subjects were scored as less aggressive on the Problem Scale than their younger peers in the sample. Less severely disabled subjects, although scored as more verbally aggressive, were rated as significantly less physically aggressive than the Moderate and Severe/Profound I/DD groups on the Problem Scale. Given that they were living and/or working alongside peers who were significantly more inclined to act out physically, it is understandable that older and Mild I/DD subjects responded with aggression of their own, at least verbally. Schwartz et al. (1998) found that frequent peer victimization was strongly correlated with reactive aggression, and noted that “maltreatment by peers may incite reactive aggression. Proactive aggression, 72
in contrast, is a behavior that can occur in the absence of a peer provocation.” Camodeca (2002), in a longitudinal study of bullying and victimization in a sample of schoolchildren, reported that individuals who were most persistently victimized were also the most reactively aggressive. External Validity Psychiatric Diagnoses and Provocation Scale Scores. Impulse Control Disorder NOS. As hypothesized, subjects diagnosed with Impulse Control Disorder NOS were scored as more reactive than those without the diagnosis. They were rated as significantly more reactive than all other subjects in the sample on Verbal Aggression, Hostile Affect, and Total Score. Impulsivity is, of course, the essential feature of impulse control disorder, as well as being essential to most definitions of reactive aggression (e.g., Dodge & Coie, 1987). Personality Disorders. Mean scores for subjects with Personality Disorder NOS, and those in the Combined Personality Disorders category (Borderline Personality Disorder plus Personality Disorder NOS) were significantly more reactive than those for the remainder of the sample on Physical Aggression and Total Score. Again, the DSMIV lists “disturbance in impulse control and emotional dysregulation” as symptoms of the Cluster B personality disorders, which include borderline personality disorder. At the purely exploratory level of statistical level of significance, I found a few more results of interest: Autism Spectrum Disorders. The portion of the sample with ASD was rated as more proactive on Physical Aggression. The use of aggressive, maladaptive behaviors to
73
communicate one’s needs and wants, as discussed earlier, may be a clear example of instrumental, goal-directed aggression. Combined Behavior Disorders. People with diagnoses in the Combined Behavior Disorders category were more reactive on Physical Aggression. Intermittent/Explosive Disorder, a component of this category, is characterized by, “…episodes of failure to resist aggressive impulses that result in serious assaultive acts…,” the severity of which are “grossly out of proportion to any precipitating psychosocial stressors” (DSM-IV; American Psychiatric Association, 1994). Medication and Provocation Scores. Subjects prescribed beta blockers were rated significantly more reactive on Physical Aggression, Hostile Affect, and Total Score. Beyond their initial use in controlling hypertension or cardiac arrhythmias, beta blockers are often prescribed to aid in dampening the “fight or flight” response in patients with anxiety or anger-control problems. In my sample, use of beta blockers was most strongly correlated with PDD NOS and Impulse Control Disorder NOS. The remaining noteworthy result was significant only at the exploratory level. Those receiving lithium were scored more reactive on Hostile Affect. Lithium is commonly used as a mood stabilizer in patients with bipolar disorder. Bipolar disorder was indeed the diagnosis most strongly correlated with lithium use in my sample. Researchers have identified impulsivity as a prominent characteristic of bipolar disorder (Swann et al., 2001, 2003) along with heightened irritability and aggressiveness in manic phases (Goodwin & Jamison, 1990).
74
Findings for comparisons between groups defined by prescribed medications appear in Tables 26 and 28. Summary. The SHARP Provocation Scale and its subscales had interrater reliability statistics in the fair to good range of clinical significance. Comparisons of mean scores of groups of subjects divided by demographic, diagnostic, and medication variables yielded results which were supportive of the Provocation Scale’s external validity. A few of the t-test results indicated that the Provocation scale can provide pertinent information not obtainable with the Problem scale alone. As predicted, reactive aggression was found to be associated with impulse-control disorders and personality disorders. Also as predicted, the correlation between the positive (proactive) and negative (reactive) totals was low, r = .11 (compared to r = .87, found by Camodeca et al., 2002). These results are encouraging, but it is also clear that important improvements to the scale are needed. Limitations and Directions for Future Research. This study’s sample could have been more demographically representative of the adult intellectually disabled population. Age groups from late adolescence to old age were reasonably well represented, and the gender ratio fell roughly within the range estimated for the I/DD population (30 to 50 percent more males than females; Stevenson, Schwartz, & Schroer, 2000). Subjects functioning in the mild range of impairment were the largest segment of the sample, followed in size by the Moderate and Severe/Profound groups, respectively. However, only a little over 43 percent of the individuals in my sample were diagnosed with mild intellectual disability, compared to an estimated 75 to 90 percent of all people with ID (McDermott, Durkin, Schupf, & Stein, 2007). All data 75
were collected in Ohio, mostly in the Columbus area. More geographic diversity would have been preferable. Also, there was a preponderance of data collected in sheltered workshops. Although there were some subjects in residential and other day program settings, a wider variety of settings likely would have resulted in a better representation of this population. Confirmatory factor analysis with a large, more representative sample of intellectually disabled adults would be a logical next step. Another limitation of the current study is the lack of external validity data for the entire sample. A related weakness was that inclusion in the validity sub-sample was not random but contingent upon guardian consent. Some diagnoses which would have been highly relevant to this study, such as Obsessive/Compulsive Disorder, were either terribly under-represented or completely missing from the sub-sample. The numbers of subjects in some external variable categories were quite small: only four subjects were diagnosed with ADHD, four were receiving beta blockers, and seven were diagnosed with Personality Disorder NOS. Another potential problem was that raters were not necessarily blind to subjects' diagnoses and medication regimens. This introduces the possibility of biased responding. However, it should also be noted that raters were not aware that these external validity data would be collected and analyzed. More work could be done to corroborate the validity of the instrument. Studies with more extensive and diverse external validity data are called for. The concurrent validity of the Provocation Scale should be investigated, using another scale designed to gauge reactive and/or instrumental aggression, such as the Revised Teacher Rating Scale
76
for Reactive and Proactive Aggression (Brown, Atkins, Osborne, & Milnamow, 1996) for comparison. A longitudinal study using both the C-SHARP (Farmer, 2008) and SHARP could address some questions, such as “Which aggressive behaviors are most stable over the years? Which are transitory?”, or, “Which C-SHARP items or subscales, if any, are predictive of aggression problems later in one’s life?” Perhaps more importantly, “Which C-SHARP scores are predictive of particular psychiatric diagnoses later in life?” If such a study were to identify C-SHARP scores predictive of later behavioral and/or psychiatric problems, this could present opportunities for preventative intervention. If the SHARP were shown to be sensitive to changes brought about by behavioral or pharmacological interventions, it could prove very useful for researchers and clinicians. If it is demonstrated that the instrument is reliably sensitive to relevant behavioral change, researchers could use it as a screen, as well as an outcome measure. Clinicians could use it in treatment planning and to measure progress toward treatment goals. Future studies should employ the SHARP and a well-established instrument of aggression, such as the Nisonger Child Behavior Rating Form (NCBRF; Aman, Tasse', Rojahn, & Hammer, 1996), as pre- and post-treatment measures. If the instruments show analogous changes, this could add to information on concurrent validity. Problems with the Provocation Scale and Possible Solutions. The primary goal of characterizing and measuring aggression in intellectually disabled adults appears to have been accomplished. The Problem scale and subscales performed well in measures of reliability and validity, and would appear to be useful in capturing a variety of behavioral topographies and/or expressions of negative affect. 77
I believe that progress was also made toward the secondary goal of assessing the “why” (the motivations or functions) of aggression and hostility. The Provocation Scale holds promise, but more refinement needs to be done. One problem with the current design was the exclusion of several items from the Provocation Scale. This was done because the items were thought to be intrinsically proactive or reactive, or simply unrelated to the question of provocation. I felt, for example, that asking whether or not the item “Is ‘sneaky;’ does things ‘on the sly’” (item 2) was provoked or not would have surely been confusing for raters. However, this left some subscales with only one (Covert Aggression) or two (Hostile Affect) items. It is difficult to obtain good reliability with subscales which have so few items. So, perhaps future versions of the instrument should measure all items on both the Problem and Provocation dimensions. It may be possible to reduce potential confusion by rewording items. For example, the aforementioned item 2 could be changed to “Behaves in a sneaky or sly manner”, or item 51, “Is often grouchy,” might be changed to read, “Often acts grouchy.” I would be reluctant to do this, however, because the reliability and validity statistics for the Problem Scale were obtained for the items as they are currently written. Another difficulty presented by the current design of the Provocation Scale was that items were not to be completed if their corresponding Problem Likert was rated a “0” (to indicate that the behavior was never displayed by the subject). This resulted in varying numbers of items completed across raters, thus ruling out factor analysis and measures of internal consistency.
78
One possible solution to these problems, in my opinion, would be to revise the Provocation Likert. For each item, the rater would be asked if the behavioral or affective description given is true of the subject, and would respond with one of the following: “-2” = “True, but only if provoked,” “-1” = “True, usually provoked,” “0” = “Never true,” “1” = “True, usually unprovoked,” and “2” = “True, always unprovoked.” Note that this is quite similar to the Likert used in the current study, with the important difference that, rather than indicating an equal likelihood of the behavior occurring with or without provocation, “0” would instead indicate an absence of the behavior or emotional state described. Put another way, “0” would point to a behavior or affective state that is equally unlikely, regardless of provocation. The proposed version would be a measure of provocation applicable to all 48 items. Also, if a Problem item is rated a “0” (“Doesn’t happen”), then a “0” on the Provocation dimension for that item would be a valid response. Therefore, if the instrument were to be completed properly, all Provocation items would be marked. With sufficient data obtained, a factor structure for the Provocation Scale could be determined through exploratory factor analysis. Thus, Provocation subscales would no longer need to be derived from EFA done for the Problem Scale. Or, if it is decided that the better option would instead be to continue to link the Provocation subscales to those of the Problem Scale, there would at least be an equal number of items in each dimension of a particular subscale (thus avoiding one- or two-item subscales).
79
REFERENCES
Achenbach, T. M. & Edelbrock, C. (1983). The teacher version of the Child Behavior Profile: I. Boys aged 6-11. Journal of Consulting and Clinical Psychology, 52, 207-217. Aman, M.G. (1991). Assessing psychopathology and behavior problems in persons with mental retardation: A review of the available instruments (DHHS Publication No. [ADM] 91-1712). Rockville, MD: U. S. Department of Health and Human Services. Aman, M. G., Arnold, L. E., & Armstrong, S. C. (1999). Review of serotonergic agents and perseverative behavior in patients with developmental disabilities. Mental Retardation & Developmental Disabilities Research Reviews, 5, 279-289. Aman, M.G., De Smedt, G., Derivan, A., Lyons, B., Findling, R.L., & The Risperidone Disruptive Behavior Study Group. (2002). Risperidone treatment of children with disruptive behavior disorders and subaverage IQ: A double-blind, placebocontrolled study. American Journal of Psychiatry, 159, 1337-1346. Aman, M. G., & Madrid, A. (1999). Atypical antipsychotics in persons with developmental disabilities. Mental Retardation & Developmental Disabilities Research Reviews, 5, 253-263 Aman, M.G., Tasse', M. J., Rojahn, J., & Hammer, D. (1996). The Nisonger CBRF: A child behavior rating form for children with developmental disabilities. Research in Developmental Disabilities, 17, 41-57. Aman, M.G., Singh, N. N., Stewart, A.W., & Field, C. J. (1985). The Aberrant Behavior Checklist: A behavior rating scale for the assessment of treatment effects. American Journal of Mental Deficiency, 89, 485-491. 80
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders, 4th edition. Washington, DC: American Psychiatric Association Bandura, A. (1971). Social learning theory. Morristown, NJ: General Learning Press. Bandura, A. (1971). Vicarious and self-reinforcement processes. In R. Glaser (Ed.), The nature of reinforcement. New York: Academic Press. Bandura, A. (1973). Aggression: A social learning analysis. Englewood Cliffs, N. J.: Prentice-Hall. Bandura, A. (1973). Social learning theory of aggression. In J. F. Knutson (Ed.), The control of aggression: Implications from basic research. Chicago: Aldine. Bandura, A. & Rosenthal, T. L. (1966). Vicarious classical conditioning as a function of arousal level. Journal of Personality and Social Psychology, 3, 54-62. Baumeister, A.A., Todd, M.E., & Sevin, J.A. (1993). Efficacy and specificity of pharmacological therapies for behavioral disorders in persons with mental retardation. Clinical Neuropharmacology, 16(4), 271-294. Berkowitz, L. (1962). Aggression: A social psychological analysis. New York: McGraw-Hill. Berkowitz, L. (1965). The concept of aggressive drive: Some additional considerations. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2). New York: Academic Press. Berkowitz, L. (1973). Words and symbols as stimuli to aggressive responses. In J. F. Knutson (Ed.), The control of aggression: Implications from basic research. Chicago: Aldine. Bhaumik, S., Branford, D., & McGrother, C. (1997). Autistic traits in adults with learning disabilities. British Journal of Psychiatry, 170, 502 -506. 81
Bjorkvist, K., Osterman, K., & Lagerspetz, K.M.J. (1994). Sex differences in covert aggression among adults. Aggressive Behavior, 20, 27-33. Block, M.E., & Rizzo, T.L. (1995). Attitudes and attributes of physical educators Associated with teaching individuals with severe and profound disabilities. Journal of the Association for Persons with Severe Handicaps, 20(1), 80-87. Brown, K., Atkins, M. S., Osborne, M. L., &, Milnamow, M. (1996). A revised teacher rating scale for reactive and proactive aggression. Journal of Abnormal Child Psychology, 24, 473-480. Browne, M.W. & Cudeck, R. (1992). Alternative ways of assessing model fit. Sociological Methods and Research, 21, 230–258. Browne, M. W., Cudeck, R., Tateneni, K., & Mels G. (2004). CEFA: Comprehensive Exploratory Factor Analysis, Version 2.00 [Computer software and manual]. Buchanan, A. (1997). The investigation of acting on delusions as a tool for risk assessment in the mentally disordered. British Journal of Psychiatry, 170 (32), 12-16. Bushman B.J., & Anderson C.A. (2001). Is it time to pull the plug on the hostile versus instrumental aggression dichotomy? Psychology Review, 108, 273–279. Buss, A.H, & Durkee, A. (1957). An inventory for assessing different kinds of hostility. Journal of Consulting Psychology, 4, 343-349. Buss, A. H., Durkee, A., & Baer, M.B. (1956). The measurement of hostility in clinical situations. Journal of Abnormal Social Psychology, 52, 84-86. Carr, E., Newsom, C. D., & Binkoff, J. A. (1980). Escape as a factor in the aggressive behavior of two retarded children. Journal of Applied Behavior Analysis, 13, 101-117. 82
Camodeca, M., Goossens, F.A., Meerum Terwogt, M., & Schuengel, C. (2002). Bullying and victimization among school-age children: Stability and links to proactive and reactive aggression. Social Development, 11, 332-345. Cattell, R. B. (1966). Handbook of multivariate experimental psychology. Rand McNally: Chicago. Chapman, R.S., & Hesketh, L. J. (2000). Behavioral phenotype of individuals with Down syndrome. Mental Retardation and Developmental Disabilities Research Reviews, 6 (2), 84-95. Charlot, L. R., Doucette, A. C., & Mezzacappa, E (1993). Affective symptoms of institutionalized adults with mental retardation. American Journal on Mental Retardation, 98 (3), 408-416. Cicchetti, D.V. (1994). Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychological Assessment, 6, 284-290. Cicchetti, D.G., & Sparrow, S.S. (1981). Developing criteria for establishing interrater reliability of specific items: Applications to assessment of adaptive behavior. American Journal of Mental Deficiency, 86, 127-137. Collacott, R. A., Cooper, S., Branford, D., & McGrother, C. (1998). Behavior phenotype for Down's syndrome. British Journal of Psychiatry,172, 85-89.
Cornell, D.G., Warren, J., Hawk, G., Stafford, E., Oram, G., & Pine, D. (1996). Psychopathy in instrumental and reactive violent offenders. Journal of Consulting and Clinical Psychology, 64, 783-790. Cowan, P.A., & Walters, R. H. (1963). Studies of reinforcement of aggression: I. Effects of scheduling. Child Development, 34, 543-551. Crick, N. R., & Dodge, K.A.(1996). Social information-processing mechanisms on reactive and proactive aggression. Child Development, 67 (3), p. 993-1002. 83
Crick, N. R., & Grotpeter, J. K. (1995). Relational aggression, gender, and socialpsychological adjustment. Child Development, 66(3), 710-722. Davidson, R. J., Jackson, D.C., & Kalin, N.H. (2000). Emotion, elasticity, context, and regulation: Perspectives from affective neuroscience. Psychological Bulletin, 126, 6, 890-909. Davidson, R. J., Putnam, K. M., Larson, C. L. (2000). Dysfunction in the neural circuitry of emotion regulation--a possible prelude to violence. Science, 289 (5479), 591594. Dodge, K. A. & Coie, J. D. (1987). Social-information-processing factors in reactive and proactive aggression in children’s peer groups. Journal of Personality and Social Psychology, 53, 1146-1158. Dollard, J., Doob, L. W., Miller, N. E., Mowrer, O. H., & Sears, R. R. (1939). Frustration and aggression. New Haven, CT: Yale University Press. Duggal, H. S. (2007). Letter to the editor: Ziprasidone for maladaptive behavior and attention-deficit/hyperactivity disorder symptoms in autistic disorder. Journal of Child and Adolescent Psychopharmacology, 17 (2), 261-263. Dunn, L. M., & Dunn, L. (1981). Peabody Picture Vocabulary Test - Revised Manual forForms L and M. Circle Pines, MN: American Guidance Service. Durkin, M.S., Schupf, N. & Stein, Z. (1985) Epidemiology and etiology of mental retardation, In Jacobson, J.W., Mulick, J.A., Rojahn, J. (Eds.) Handbook of Intellectual and Developmental Disabilities, New York: Springer Press. Edelbrock, C. S. (1985). Child Behavior Rating Form. Psychopharmacological Bulletin, 21, 835-837. Edlinger, W. (1983). Selbstverletzungverhalten und Vehaltenskompetenzen bei geistig behindertin Personen (in german). Unpublished Master's Thesis, University of Marburg, Germany. 84
Edmunds, G., & Kendrick, D.C. (1980). The measurement of human aggressiveness. Ellis Horwood: Chichester. Einfeld, S. L., & Tonge, B. J. (1995). The Developmental Behavior Checklist: The development and validation of an instrument to assess behavioral and emotional disturbance in children and adolescents with mental retardation. Journal of Autism & Developmental Disorders, 25, 81-104. Fagot, B. I. (1984). The consequents of problem behavior in toddler children. Journal of Abnormal Child Psychology, 12, 385-396. Farmer, C.A. (2008). Factor structure and psychometric properties of the Children's Scale for Hostility and Aggression: Reactive/Proactive (C-SHARP). Unpublished Master's Thesis, The Ohio State University. Feschbach, S. (1970). Aggression. In P. H. Mussen (Ed.), Carmichael’s manual of child psychology (Vol. 2). New York: Wiley. Feshbach, S. (1997). The psychology of aggression: Insights and issues. In S. Feshbach & J. Zagrodzka (Eds.), Aggression: Biological, developmental,and social perspectives. New York: Plenum Press. Fotheringham, J.B., Abdo, K., Ouellette-Kuntz, H., & Wolfgarth, A.(1993). Survey of community adjustment of previously institutionalized developmentally disabled persons. Canadian Journal of Psychiatry, 38, 641-648. Frodi, A. , Macaulay, J., & Thome, P.R. (1977). Are women always less aggressive than men? Psychological Bulletin 84, 634-660. Frick, P.J. (1998). Conduct disorder and severe antisocial behavior. New York: Plenum Press. Fuchs, C., & Benson, B. A. (1995). Social information processing by aggressive and nonaggressive men with mental retardation. American Journal on Mental Retardation, 100, 244–252. 85
Garno, J. L., Gunawardane, N., & Goldberg, J. F. (2008). Predictors of trait aggression in bipolar disorder. Bipolar Disorders, 10 (2), 285–292.
Geen, R. G., & Stonner, D. (1971). Effects of aggressiveness habit strength on behavior in the presence of aggression-related stimuli. Journal of Personality and Social Psychology, 17, 149-153. Golden, J., & Reese, M. (1996). Focus on communication: Improving interaction between staff and residents who have severe or profound mental retardation, Research in Developmental Disabilities, 17(5), 363-382. Goodwin, F.K., & Jamison, K.R. (1990). Suicide in Manic-Depressive Illness. New York: Oxford University Press. Green, M. J., Waldron, J. H., & Coltheart, M. (2007). Emotional context processing is impaired in schizophrenia. Cognitive Neuropsychiatry, 12 (3), 259-280. Green, R. T., & Stacey, B. G. (1967). The development of a questionnaire measure of hostility and aggression. Acta Psychologica, Amsterdam 26(3)1967 p.265285. Harris, M. B., & Knight-Bohnhoff, K. (1996). Gender and aggression I: Perceptions of aggression. Sex Roles, 35, 1-24. Havercamp, S. M., & Reiss, S. (1996). Composite versus multiple-rating scales in the assessment of psychopathology in people with mental retardation. Journal of Intellectual Disability Research, 40, 167-179. Hill, B. K. & Bruininks, R. H. (1984). Maladaptive behavior of mentally retarded individuals in residential facilities. American Journal of Mental Deficiency, 88, 380-387. Holden, B., & Gitlesen, J. P. (2004). The association between severity of intellectual disability and psychiatric symptomatology. Journal of Intellectual Disability Research, 48 (6), 556 - 562. 86
Ingram, S., Hechtman, L., & Morgenstern, G. (1999). Outcome issues in ADHD: Adolescent and adult long-term outcome. Mental Retardation and Developmental Disabilities Research Reviews, 5, 243-250. Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K. E., & Richman, G. S. (1982). Toward a functional analysis of self-injury. Analysis and Intervention in Developmental Disabilities, 2, 3–20. Jacobson, J. W. (1982). Problem behavior and psychiatric impairment within a developmentally disabled population: II. Behavior severity. Applied Research in Mental Retardation, 3, 369-381. Kaufmann, H. (1965). Definitions and methodology in the study of aggression. Psychological Bulletin, 64, 351-364. Kaplan, H. I, & Sadock B. J. (1972). Kaplan and Sadock's synopsis of psychiatry: behavioral sciences and clinical psychiatry. Philadelphia: Williams and Wilkins. Kazdin, A. E., & Esveldt-Dawson, K. (1986). The Interview for Antisocial Behavior: Psychometric characteristics and concurrent validity with child psychiatric inpatients. Journal of Psychopathology and Behavioral Assessment, 8, 289-303. Kazdin, A. E., Rodgers, A., Colbus, D., & Siegel, T. (1987). Children's Hostility Inventory: Measurement of aggression and hostility in psychiatric inpatient children. Journal of Clinical Child Psychology, 16, 320-328. Koegel, R. L., Camarata, S. M., & Koegel, L. K. (1994). Aggression and noncompliance: Behavior modification through naturalistic language remediation. p.165180, Autism in children and adults: Etiology, assessment, and intervention. Belmont, CA: Thomson Brooks/Cole Publishing Co. Kolko, D. J. (1988). Daily ratings on a child psychiatric unit: Psychometric evaluation of the Child Behavior Rating Form. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 126-132. 87
Krakowski, M.I., & Czobor, P. (1994). Clinical symptoms, neurological impairment, and prediction of violence in psychiatric inpatients. Hospital & Community Psychiatry, 45 (7), 700-705. Levene, H. (1960). In Contributions to Probability and Statistics: Essays in Honor of Harold Hotelling, I. Olkin et al. eds., Stanford University Press, pp. 278-292. Little, T. D., Henrich, C. C., Jones, S. M., & Hawley, P. H. (2003). Disentangling the "whys" from the "whats" of aggressive behaviour. International Journal of Behavioral Development, 27, 122-133. Maccoby, E. E., Jacklin, C. N. (1974). The Psychology of Sex Differences. Stanford, CA: Stanford University Press. McClintock, K., Hall, S., & Oliver, C. (2003). Risk markers associated with challenging behaviours in people with intellectual disabilities: A meta-analytic study. Journal of Intellectual Disability Research, 47, 405-416. McKay, K.E., & Halperin, J.M. (2001). ADHD, aggression, and antisocial behavior across the lifespan: interactions with neurochemical and cognitive function. Annals of the New York Academy of Science, 931, 84-96. McGrew, K.S., Ittenbach, R.F., Bruininks, R.H., Hill, & Bradley, K. (1991). Factor structure of maladaptive behavior across the lifespan of persons with mental retardation. Research in Developmental Disabilities, 12 (2),181-199. Meador, D.M., & Osborn, R.G. (1992). Prevalence of severe behavior disorders in persons with mental retardation and treatment procedures used in community and institutional settings. Behavioral Residential Treatment, 7, 299-314. Miller, L. C. (1977). School Behavior Checklist manual. Los Angeles: Western Psychological Services. Miller, L. S., Klein, R.G., Piacentini, J., Abikoff, H., Shah, M. R., Samoilov, A., & Guardino, M. (1995). The New York Teacher Rating Scale for disruptive and antisocial behavior. Journal of the American Academy of Child and Adolescent Psychiatry, 34, 359-370. 88
Mulick, J. A., Dura, J. R., Rasnake, J. K., & Wisniewski, J. (1988, May). Selfinjurious behavior and stereotypy in nonambulatory children with profound mental retardation. Poster presented at the 2nd. Annual North Coast Regional Conference, Society of Pediatric Psychology, Worthington, OH. National Institute of Health, Consensus Development Panel on Destructive Behaviors in Persons with Developmental Disabilities (1989). Treatment of destructive behaviors in persons with developmental disabilities. (NIH Publication No. 91-2410). Bethesda, MD: U.S. Department of Health and Human Services.
Novaco, R. W. (1994). Anger as a risk factor for violence among the mentally disordered. In Violence and mental disorder: Developments in risk assessment. Chicago: University of Chicago Press.
Novaco, R.W.(2007). Anger dysregulation. In Anger, aggression and interventions for interpersonal violence. Mahwah, NJ, US : Lawrence Erlbaum Associates Publishers. Polman, H., Orobio de Castro, B., Koops, W., van Boxtel, H. W., & Merk, W. W. (2007) A meta-analysis of the distinction between reactive and proactive aggression in children and adolescents.Journal of Abnormal Child Psychology Vol 35(4) p.522-535. Poulin, F. & Boivin, M. (2000). Reactive and proactive aggression: Evidence of a two factor model. Psychological Assessment, 12, 115-122. Price, J. M., & Dodge, K. A. (1989). Reactive and proactive aggression in childhood: Relations to peer status and social context dimensions. Journal of Abnormal Child Psychology, 17, 455-471.
Quanty, M.(1976). Aggression catharsis: Experimental investigations and implications. In R. J. Geen & E. I. Donnerstein (Eds.), Aggression: Theoretical and empirical reviews. New York: Academic Press. Quay, H. C., & Peterson, D. R. (1987). Manual for the Behavior Problem Checklist. Coral Gables, FL.: Author. 89
Reiss, S., 1988. In: The Reiss Screen Test manual, International Diagnostic Systems, Inc., Orland Park, IL. Renfrew, J. W. (1997). Aggression and its causes: A biopsychosocial approach. Oxford: Oxford University Press. Robertson, J., Emerson, E., Gregory, N., H. C., T. S., Kessissoglou, S.(2000). Lifestyle related risk factors for poor health in residential settings for people with intellectual disabilities. Research in Developmental Disabilities, 21(6), 469-486. Rojahn, J. (1984). Self-injurious behavior in institutionalized, severely/profoundly retarded adults: Prevalence data and staff agreement. Journal of Behavioral Assessment, 6, 13-27. Rojahn, J., Matlock, S. T., & Tasse, M. J. (2000). The Stereotyped Behavior Scale: Psychometric properties and norms. Research in Developmental Disabilities, 21, 437-454. Rojahn, J., Matson, J.L., Lott, D., Esbensen, A.J., & Smalls, Y. (2001). The Behavior Problems Inventory: An instrument for the assessment of self-injury, stereotyped behavior, and aggression/destruction in individuals with developmental disabilities. Journal of Autism and Developmental Disorders, 31, 577-588. Ross, R. T. (1972). Behavioral correlates of levels of intelligence. American Journal of Mental Deficiency, 76, 545-549. Rudolph, K. D., & Clark, A. G. (2001). Conceptions of relationships in children with depressive and aggressive symptoms: Social-cognitive distortion or reality? Journal of Abnormal Child Psychology, 29 (1), 41-56. Rutter, M. (1978). Brain damage syndromes in childhood: Concepts and findings. Annual Progress in Child Psychiatry & Child Development, 501-527.
90
Schippell, P. L., Vasey, M. W., Cravens-Brown, L. M., & Bretveld, R. A. (2003). Suppressed attention to rejection, ridicule, and failure cues: A unique correlate of reactive but not proactive aggression in youth. Journal of Clinical Child and Adolescent Psychology, 32, 40–55. Schwartz, D., Dodge, K. A., Coie, J. D., Hubbard, J. A., Cillessen, A. H. N., Lemerise, E. A., & Bateman, H. (1998). Social-cognitive and behavioral correlates of aggression and victimization in boys’ play groups. Journal of Abnormal Child Psychology, 26, 431–440. Scott, J. P. (1958). Aggression. Chicago: University of Chicago Press. Shrout, P.E., & Fleiss, J.L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420-427. Skinner, B. F. (1953). Science and human behavior. New York: Macmillan. Skinner, B. F. (1969). Contingencies of reinforcement: A theoretical analysis. New York: Appleton-Century-Crofts. Snyder, R., Turgay, A., Aman, M.G., Binder, C., Fisman, S., Carroll, A., & The Risperidone Conduct Study Group. (2002). Effects of risperidone on conduct and disruptive behavior disorders in children with subaverage IQs. Journal of the American Academy of Child and Adolescent Psychiatry, 41, 1026-1036. Spencer, T. J., Biederman, J., Wozniak, J., Faraone, S.V., Wilens, T.E., & Mick, E. (2001). Parsing pediatric bipolar disorder from its associated comorbidity with the disruptive behavior disorders . Biological Psychiatry ,49 (12) , 1062 - 1070. Stevenson, R.E., Schwartz, C.E., & Schroer, R.J. (2000). X-linked mental retardation. New York: Oxford University Press. Swann A.C., Anderson, J.C., Dougherty, D.M., & Moeller, F.G. (2001) Measurement of inter-episode impulsivity in bipolar disorder. Psychiatry Research, 101(2), 195-197. 91
Tasse', M. J., Aman, M.G.; Hammer, D., & Rojahn, J. (1996) The Nisonger Child Behavior Rating Form: Age and gender effects and norms. Research in Developmental Disabilities, 17, 59-75.
Taylor, S. P. (1986). The regulation of aggressive behavior. In R. J. Blanchard & D. C. Blanchard (Eds.), Advances in the study of aggression (Vol. 2). New York: Academic Press. Thurstone, L. L. (1947). Multiple factor analysis. Chicago: University of Chicago Press. Tyrer, F., McGrother C.W., Thorp, C.F., Donaldson, M., Bhaumik, S., Watson, J.M., & Hollin, C. (2006). Physical aggression towards others in adults with learning disabilities: prevalence and associated factors. Journal of Intellectual Disability Research, 50, 295-304. Walters, R. H., & Brown, M. (1963). Studies of reinforcement of aggression: III. Transfer of responses to an interpersonal situation. Child Development, 34, 563-571. Vasey, M. W., El-Hag, N., & Daleidon, E. L. (1996). Anxiety and the processing of emotionally threatening stimuli: Distinctive patterns of selective attention among high and low-test-anxious children. Child Development, 67, 1173–1185. Wechsler, D. (1991). Wechsler Intelligence Scale for children (3rd ed.). San Antonio, TX: The Psychological Corporation. Weiss, E. M., Kohler, C. G., Nolan, K. A., Czobor, P., Volavka, J., Platt, M. M., Brensinger, C., Loughead, J., Delazer, M., Gur, R. E., & Gur, R. C. (2006). The relationship between history of violent and criminal behavior and recognition of facial expression of emotions in men with schizophrenia and schizoaffective disorder. Aggressive Behavior, 32 (3), 187-194. Xu, Y., & Zhang, Z. (2007). Distinguishing proactive and reactive aggression in chinese children. Journal of Abnormal Child Psychology, 36 (4), 539 -552.
92
Young, S., Toone, B., & Tyson, C. (2003). Cmorbidity and psychosocial profile of adults with Attention Deficit Hyperactivity Disorder. Personality and Individual Differences 35 , 743–755. Yudofsky, S.C, Silver, J.M., Jackson, W., Endicott, J., & Williams, D. (1986). The Overt Aggression Scale for the objective rating of verbal and physical aggression. American Journal of Psychiatry, 143, 35-39. Zanarini, M. C., Frankenburg, F. R., DeLuca, C. J., Hennen, J., Khera, G. S., Gunderson, J. G. (1998). The pain of being borderline: Dysphoric states specific to borderline personality disorder. Harvard Review of Psychiatry, 6 (4), 201-207. Zillmann, D. (1979). Hostility and aggression. New York: Wiley.
93
APPENDIX A TABLES
94
Characteristic Gender: Female
Percentage of Sample 38.5
Male
61.5
18 to 30 years
22.6
31 to 40 years
23.4
41 to 50 years
19.7
51 to 60 years
17.4
61 to 70 years
5.7
71 to 80+ years
1.2
Age:
Level of Functioning: Mild
43.4
Moderate
33.6
Severe/Profound
22.5
Unknown
0.6
Setting: Day Program
76.8
Residential
23.2
Ethnicity: African American
26.0
Caucasian
71.1
Hispanic
1.0
Asian American
0.2
Other
1.4
Table 1: Demographic Characteristics of Subject Sample.
95
Original item assignment
Percent of sample endorsed
1. Sneers, "makes faces" at others
34.1
2. Is "sneaky;" does things "on the sly”
51.2
3. Attempts to touch others in inappropriate,
15.9
sexually suggestive manner 4. Pinches others
10.8
5. Is resentful over seemingly minor issues
45.4
6. Breaks others' things
14.4
7. Is quick to anger ("hot-headed")
44.7
8. Takes others' things by force
13.9
9. Broods, pouts, or is sullen
51.2
10. Calls others insulting names in their absence
23.8
11. Shoves or pushes others
25.9
12. Crowds others (invades their personal space)
35.9
13. Intimidates others
24.7
14. Says, "I hate you" or makes other hurtful statements
25.8
15. Bites others
9.5
16. Insults others to their faces
26.7
17. Throws objects at others
16.4
18. Reacts suddenly or impulsively to minor provocations
57.3
19. Shouts at others in anger
43.1
20. Uses headlocks or other “wrestling” holds against others
2.5
21. Gets mad when caught behaving badly
45.0
22. Is overly argumentative
32.1
23. Uses profanity to shock or offend others
14.4
24. Cuts others with a sharp object
1.3
25. Tickles or otherwise physically teases others,
13.2
even after being asked to stop 26. Steals from others when they aren't looking
22.7
Table 2: Percentages of endorsement for the original 58 SHARP items (continued) 96
Table 2: Continued 27. Reacts to insults or teasing by lashing out physically
16.6
28. Calls others insulting names to their faces
27.6
29. Trips others
3.6
30. Head-butts others
6.8
31. Makes insulting comments about others behind their backs
25.4
32. Breaks own belongings
16.0
33. Makes threatening gestures (like shaking fist)
22.7
34. Charges at others
13.0
35. Verbally teases others, even after being asked to stop
25.0
36. If caught, denies having behaved badly
33.0
37. Pulls others' hair
8.3
38. When angry, is slow to cool off
45.5
39. Spits at others
7.6
40. Says "I hate [someone not present]" or makes other
22.7
hurtful statements in absence of the person concerned 41. Chokes others/Grabs others' necks
4.2
42. Lashes out at people who are in his/her space
33.7
43. Takes offense at things others say or do when s/he shouldn't
43.6
44. Starts trouble by baiting others
23.4
45. If caught, makes excuses for bad behavior
33.5
46. Is hostile, "has a chip on his/her shoulder"
22.3
47. Scratches others with fingernails
13.2
48. Glares at others
29.5
49. Encourages others to gang up on someone
9.2
(physically OR verbally) 50. Hits others with objects
14.8
51. Is often grouchy
35.6
52. Doesn't care about others' feelings
31.9
53. Verbally threatens others with physical harm
17.3
54. Hits or shoves others forcefully
23.6
55. Makes unwanted sexual comments to others
9.0 (Continued)
97
Table 2 Continued 56. Does not seem truly sorry after hurting someone
22.5
57. Gets revenge after some time has passed and
13.7
the other person is not on guard 58. Tries not to get caught while doing harmful things to others
98
19.7
Item Verbal Aggression 1. Sneers, "makes faces" at others 10. Calls others insulting names in their absence 14. Says, "I hate you" or makes other hurtful statements 16. Insults others to their faces 23. Uses profanity to shock or offend others 28. Calls others insulting names to their faces 31. Makes insulting comments about others behind their backs 40. Says "I hate [someone not present]" or makes other hurtful statements in absence of the person concerned 53. Verbally threatens others with physical harm 55. Makes unwanted sexual comments to others 99
Physical Aggression 4. Pinches others 6. Breaks others' things 11. Shoves or pushes others 15. Bites others 17. Throws objects at others 30. Head-butts others 32. Breaks own belongings 34. Charges at others 37. Pulls others' hair 39. Spits at others 47. Scratches others with fingernails 50. Hits others with objects 52. Doesn't care about others' feelings 54. Hits or shoves others forcefully
Factor I
Factor II
Factor III
Factor IV
Factor V
0.31 0.72 0.69 0.79 0.68 0.86 0.67 0.64
0.01 -0.06 0.04 0.02 0.13 0.04 -0.17 -0.04
0.12 0.08 0.20 0.00 -0.04 0.02 0.12 0.24
0.07 0.11 0.01 0.04 0.02 0.10 0.22 0.14
0.31 0.01 0.09 0.15 0.07 0.10 -0.09 -0.10
0.44 0.37
0.08 -0.15
0.18 -0.01
0.03 0.11
0.19 0.24
0.13 -0.02 0.03 -0.07 0.05 -0.09 -0.09 -0.01 0.09 0.21 -0.04 0.00 0.09 -0.06
0.48 0.42 0.44 0.58 0.53 0.56 0.44 0.38 0.57 0.56 0.74 0.51 0.23 0.55
-0.18 0.03 0.10 0.06 0.09 0.05 0.18 0.16 -0.04 -0.03 -0.01 0.18 0.20 0.19
0.15 0.28 -0.10 0.03 0.24 0.07 0.14 0.06 -0.01 0.04 0.10 0.16 0.17 0.03
0.09 0.14 0.40 0.09 0.14 0.02 0.12 0.28 0.00 -0.02 0.01 0.13 0.18 0.28
Table 3: Five-factor solution using Maximum Wishart Likelihood extraction with oblique equamax rotation (Continued) 99
Table 3: Continued
100
Hostile Affect 5. Is resentful over seemingly minor issues 7. Is quick to anger ("hot-headed") 9. Broods, pouts, or is sullen 18. Reacts suddenly or impulsively to minor provocations 19. Shouts at others in anger 21. Gets mad when caught behaving badly 22. Is overly argumentative 27. Reacts to insults or teasing by lashing out physically 38. When angry, is slow to cool off 42. Lashes out at people who are in his/her space 43. Takes offense at things others say, do when s/he shouldn't 46. Is hostile, "has a chip on his/her shoulder" 48. Glares at others 51. Is often grouchy Covert Aggression 2. Is "sneaky;" does things "on the sly" 26. Steals from others when they aren't looking 36. If caught, denies having behaved badly 44. Starts trouble by baiting others 45. If caught, makes excuses for bad behavior 49. Encourages others to gang up on someone (physically OR verbally) 57. Gets revenge after some time has passed and the other person is not on guard 58. Tries not to get caught while doing harmful things to others
-0.03 0.09 0.07 0.03 0.30 -0.02 0.28 0.12 0.08 0.08 0.17 0.17 0.05 0.16
-0.07 0.23 -0.09 0.30 0.09 0.08 -0.21 0.20 0.22 0.27 -0.08 -0.02 0.11 0.09
0.67 0.62 0.59 0.55 0.49 0.48 0.51 0.29 0.54 0.33 0.55 0.43 0.32 0.55
0.14 0.05 0.08 0.06 0.00 0.33 0.18 0.05 0.09 -0.05 0.15 0.10 0.18 0.05
0.13 0.10 0.06 0.12 0.14 0.07 0.00 0.25 0.11 0.29 0.02 0.11 0.24 0.03
-0.01 -0.15 0.11 0.11 0.16 0.12
0.07 0.12 -0.12 -0.10 -0.17 -0.18
0.02 -0.02 0.36 -0.05 0.38 -0.07
0.68 0.61 0.47 0.54 0.42 0.50
0.03 0.02 -0.04 0.26 -0.03 0.07
-0.08
0.24
0.00
0.51
0.17
-0.02
0.04
-0.10
0.80
0.06
(Continued) Table 3: Continued 100
Bullying 3. Attempts to touch others in inappropriate, sexually suggestive manner 8. Takes others' things by force 12. Crowds others (invades their personal space) 13. Intimidates others 25. Tickles or otherwise physically teases others, even after being asked to stop 33. Makes threatening gestures (like shaking fist) 35. Verbally teases others, even after being asked to stop 56. Does not seem truly sorry after hurting someone Italicized items failed to meet criteria for inclusion.
-0.10
-0.11
-0.02
0.11
0.58
-0.01 -0.15 0.13 -0.04
0.15 0.12 0.12 -0.16
-0.11 0.09 0.15 -0.11
0.21 0.10 0.18 0.02
0.49 0.55 0.42 0.78
0.21 0.22 0.05
0.18 -0.30 -0.03
0.13 -0.05 0.19
0.04 0.24 0.18
0.45 0.55 0.27
101 101
Factor
Total
Percent of variance
I
6.334
11.730
11.730
II
5.938
10.996
22.726
III
5.852
10.837
33.563
IV
4.162
7.707
41.270
V
4.013
7.432
48.702
Table 4: Rotation Sums of Squared Loadings.
102
Cumulative Percentage
Verbal Aggression Verbal Aggression
Physical Aggression
Hostile Affect
Covert Aggression
Bullying
.234*
.649*
.440*
.446*
.508*
.421*
.514*
.578*
.539*
Physical Aggression
.234*
Hostile Affect
.649*
.508*
Covert Aggression
.440*
.421*
.578*
Bullying
.446*
.514*
.539*
.576* .576*
103
* Correlation is significant at the 0.01 level (2-tailed)
Table 5: Correlations between SHARP Subscales
103
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.70
0 – 29
6.54
Physical Aggression
6.12
0 – 32
7.45
12.24
0 – 41
11.31
Covert Aggression
2.91
0 – 18
3.52
Bullying
3.47
0 – 21
4.24
29.92
0 – 130
27.48
Hostile Affect
Problem Scale Total
Table 6: Problem Scale Total and Subscale Scores for Subjects Eighteen to Thirty Years of Age (n = 115).
104
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
3.84
0 – 28
5.79
Physical Aggression
3.88
0 – 24
5.73
10.81
0 – 47
9.53
Covert Aggression
2.68
0 – 16
3.57
Bullying
3.35
0 – 21
4.14
24.78
0 – 116
23.35
Hostile Affect
Problem Scale Total
Table 7: Problem Scale Total and Subscale Scores for Subjects Thirty-one to Forty Years of Age (n = 121)
105
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
2.50
0 – 24
4.62
Physical Aggression
2.80
0 – 22
4.68
Hostile Affect
7.59
0 – 35
8.58
Covert Aggression
2.33
0 – 13
3.14
Bullying
1.78
0 – 15
2.65
17.29
0 – 86
19.43
Problem Scale Total
Table 8: Problem Scale Total and Subscale Scores for Subjects Forty-one to Fifty Years of Age (n = 147)
106
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.02
0 – 20
5.28
Physical Aggression
3.12
0 – 21
4.65
10.97
0 – 42
10.41
Covert Aggression
2.32
0 – 16
3.28
Bullying
2.31
0 – 17
3.36
23.06
0 – 100
22.58
Hostile Affect
Problem Scale Total
Table 9: Problem Scale Total and Subscale Scores for Subjects Fifty-one to Sixty Years of Age (n = 90)
107
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
3.06
0 – 21
5.09
Physical Aggression
2.78
0 – 26
4.89
Hostile Affect
7.97
0 – 24
6.76
Covert Aggression
1.41
0 – 11
2.38
Bullying
1.26
0–7
1.77
16.94
0 – 58
16.43
Problem Scale Total
Table 10: Problem Scale Total and Subscale Scores for Subjects Aged Sixty Years or More (n = 31)
108
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
3.11
0 – 29
5.25
Physical Aggression
3.50
0 – 26
5.47
Hostile Affect
9.38
0 – 47
9.57
Covert Aggression
2.54
0 – 16
3.37
Bullying
2.68
0 – 21
3.69
21.48
0 – 130
22.16
Problem Scale Total
Table 11: Problem Scale Total and Subscale Scores for Male Subjects (n = 314).
109
Scale or Subscale Verbal Aggression
Mean Score 4.40
Range 0 – 28
Standard Deviation 5.93
4.23
0 – 30
5.92
10.78
0 – 37
10.08
Covert Aggression
2.39
0 – 18
3.33
Bullying
2.36
0 – 15
3.22
24.55
0 – 101
23.71
Physical Aggression Hostile Affect
Problem Scale Total
Table 12: Problem Scale Total and Subscale Scores for Female Subjects (n = 190)
110
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.30
0 – 29
6.09
Physical Aggression
2.50
0 – 26
4.44
10.11
0 – 42
9.77
Covert Aggression
2.34
0 – 16
3.17
Bullying
2.26
0 – 20
3.38
21.56
0 – 130
23.11
Hostile Affect
Problem Scale Total
Table 13: Problem Scale Total and Subscale Scores for Subjects Diagnosed with Mild IDD (n = 222)
111
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.26
0 – 28
5.83
Physical Aggression
4.16
0 – 32
5.97
11.40
0 – 47
10.72
Covert Aggression
3.05
0 – 18
3.76
Bullying
2.99
0 – 21
3.76
26.15
0 – 124
24.61
Hostile Affect
Problem Scale Total
Table 14: Problem Scale Total and Subscale Scores for Subjects Diagnosed with Moderate IDD (n = 172)
112
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
1.39
0 – 20
3.04
Physical Aggression
5.76
0 – 30
6.80
Hostile Affect
7.56
0 – 35
8.15
Covert Aggression
1.99
0 – 12
3.07
Bullying
2.70
0 – 21
3.77
20.22
0 – 80
20.54
Problem Scale Total
Table 15: Problem Scale Total and Subscale Scores for Subjects Diagnosed with Severe/Profound IDD (n = 115).
113
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
3.1
0 – 27
5.1
Physical Aggression
3.1
0 – 31
5.5
Hostile Affect
8.2
0 – 36
8.0
Covert Aggression
3.6
0 – 20
4.5
Bullying
2.6
0 – 21
3.6
20.2
0 – 119
20.6
Problem Scale Total
Table 16: Problem Scale Total and Subscale Scores for All Subjects (n = 512)
114
Name of Scale
Chronbach’s Alpha
Problem Scale Verbal Aggression
.922
Physical Aggression
.889
Hostile Affect
.915
Covert Aggression
.821
Bullying
.801
Total
.949
Behavior Problem Inventory Aggressive/Destructive Behavior Subscale BPI Frequency Score
.844
BPI Severity Score
.853
Table 17: Internal Consistency of Problem Scale Total, Problem Subscales and BPI Aggressive/Destructive Behavior Subscale (Alpha Coefficients).
115
Scale or subscale
Intraclass Correlation
Intraclass Correlation
(Excluding Across Settings)
(Including Across Settings)
Problem Scale Verbal Aggression
.704
.610
Physical Aggression
.778
.721
Hostile Affect
.750
.734
Covert Aggression
.720
.570
Bullying
.592
.707
.548
.657
Provocation Scale total
Table 18: Interrater reliability of the Problem subscales.
116
Name of Scale or Subscale
Two-tailed Pearson Correlations with BPI Subscale, Frequency and Severity
Verbal Aggression
Frequency Severity
.345* .327*
Physical Aggression
Frequency Severity
.789* .861*
Hostile Affect
Frequency Severity
.512* .538*
Covert Aggression
Frequency Severity
.431* .456*
Bullying
Frequency Severity
.515* .535*
Total
Frequency Severity * Correlation is significant at the 0.01 level (2-tailed)
.669* .700*
Table 19: Correlations between the SHARP Problem Scale Subscales and Total, and the BPI Aggressive/Destructiveness Subscale
117
Demographic Category
Problem Scale
Provocation Scale
Provocation/Problem Sum (PPS)
Provocation/Problem Product (PPP)
Provocation Quotient
Age
(Total)* (Verbal)† (Physical)* (Hostile)* (Covert)† (Bully)* (Physical)* (Bully)†
(Total)* (Verbal)† (Physical)† (Hostile)† (Bully)†
(Total)* (Verbal)* (Physical)* (Hostile)† (Bully)*
(Total)* (Verbal)* (Physical)† (Hostile)* (Bully)*
(Total)* (Verbal)† (Physical)† (Hostile)† (Bully)†
(Total)* (Verbal)* (Hostile)*
(Total)* (Physical)* (Covert)* (Bully)†
(Hostile)* (Covert)*
Level of I/DD: Mild vs. Severe/Profound
Verbal* (Physical)* Covert*
(Total)† Verbal* (Physical)* (Covert)*
(Total)* (Physical)* (Hostile)* (Covert)*
Level of I/DD: Moderate vs. Severe/Profound
Total† Verbal* (Physical)† Hostile† Covert* Verbal†
(Total)* (Verbal)† (Physical)† (Hostile)* (Covert)† nil
(Total)* (Verbal)* (Physical)† (Hostile)* (Covert)* (Total)* (Physical)* (Hostile)† (Covert)*
Verbal* (Physical)† Hostile†
Verbal†
(Physical)*
Verbal† (Covert)†
(Covert)†
(Covert)†
Level of I/DD: Mild vs. Moderate
118
Gender * †
Significant at the 0.01 level (2-tailed) Exploratory (p ≤ 0.05)
(Covert)†
Note: Bold = Difference unique to Provocation Scale (not found on Problem Scale)
Table 20: Comparison of Significant Differences in Problem Scale and Provocation Scores by Demographic Groups
118
Diagnosis
Problem Scale
Provocation Scale
Provocation/Problem Sum (PPS)
Provocation/Problem Product (PPP)
Provocation Quotient
Anxiety Disorder NOS
Nil
nil
nil
nil
nil
ADHD
Hostile† Covert* Bully* (Verbal)* Physical† (Hostile)†
nil
nil
nil
nil
Total† Hostile† Bully† (Verbal)* Physical† (Hostile)†
nil
Total† Physical†
Behavioral Disorder NOS
nil
nil
nil
(Covert)*
(Covert)*
Bipolar Disorder
Total† Verbal† Physical† Hostile† nil
nil
Hostile†
nil
nil
nil
nil
nil
nil
nil
nil
(Physical)*
(Total)* (Physical)* (Hostile)* (Covert)* (Bully)*
nil
(Total)* (Physical)* (Hostile)* (Covert)* (Bully)*
(Total)† (Physical)† (Physical)† (Covert)*
(Total)* (Physical)* (Total)* (Physical)* (Covert)*
Autism
Borderline Personality Disorder Depression Down Syndrome
119
Table 21: Comparison of Problem Scale Validators and Provocation Validators (Diagnoses).
119
(Continued)
Table 21 Continued
120
* †
Impulse Control Disorder
nil
(Total)* (Verbal)* (Hostile)*
nil
Intermittent/Explosive Disorder Mood Disorder NOS
nil
nil
Hostile†
OCD
(Hostile)† (Bully)*
nil
(Total)* (Verbal)† (Hostile)† (Bully)† nil
nil
nil
nil
nil
nil
nil
nil
Physical†
nil
nil
Personality Disorder NOS
nil
(Total)* (Physical)*
(Physical)*
(Total)* (Physical)† (Hostile)†
(Total)* (Physical)†
PDD
nil
nil
nil
nil
nil
PTSD
nil
nil
nil
nil
Total†
Psychotic Disorder NOS
Total† Hostile* Bully†
nil
Total† Verbal† Hostile†
nil
nil
Significant at the 0.01 level (2-tailed) Exploratory (p ≤ 0.05)
Note: Bold = Difference unique to Provocation Scale (not found on Problem Scale)
120
121
External Validator
Verbal Aggression
Physical Aggression
Hostile Affect
Covert Aggression
Bullying
Total Score
ADHD Present (n = 4) Not Present (n = 99) Autism Present (n = 9) Not Present (n = 94) Bipolar Disorder Present (n = 14) Not Present (n = 89) Down Syndrome Present (n = 11) Not Present (n = 92) Mood DO NOS Present (n = 8) Not Present (n = 95) Personality DO NOS Present (n = 7) Not Present (n = 96) Psychotic DO NOS Present (n = 14) Not Present (n = 89)
t(3)= 1.57, p = .21 12.50 2.84 t(93)=-6.08, p = .0009* 0.00 3.52 t(101)=2.15, p = .034† 6.07 2.76 t(101)=-0.14, p = .892 3.00 3.24 t(101)=0.29, p = .774 3.75 3.17 t(101)=1.19, p = .238 5.57 3.04 t(101)=1.8, p = .065 5.71 2.82
t(101)= 1.45, p = .15 10.00 5.30 t(101)=2.50, p = .014† 10.44 5.01 t(101)=2.07, p = .041† 8.71 4.98 t(33)=-6.15, p = .0009* 1.09 6.01 t(101)=1.52, p = .133 8.75 5.21 t(101)=1.08, p = .282 8.00 5.30 t(101)=1.41, p = .161 7.71 5.13
t(101)= 2.44, p = .02 † 18.75 8.10 t(41)= -2.37, p = .023† 5.89 8.77 t(101)=2.23, p = .028† 13.29 7.76 t(38)=-5.16, p =.0009* 2.55 9.23 t(101)=2.36, p = .02† 15.38 7.94 t(101)=2.44, p = .016† 16.14 7.96 t(101)=2.70, p = .008* 14.21 7.62
t(101)=3.73, p =.0009* 10.50 2.83 t(101)= -.0.42, p = .678 2.56 3.18 t(101)=1.43, p = .155 4.64 2.89 t(96)=-6.15, p =.0009* 0.36 3.46 t(101)=0.94, p = .347 4.50 3.01 t(101)=0.47, p = .642 3.86 3.07 t(101)=0.95, p = .342 4.14 2.97
t(101)= 3.17, p = .002* 8.50 2.56 t(101)=-0.55, p = .583 2.11 2.85 t(101)=1.90, p = .061 4.57 2.51 t(77)=-5.21, p =.0009* 0.55 3.05 t(101)=-0.12, p = .902 2.63 2.80 t(101)=2.24, p = .027† 5.86 2.56 t(101)=2.13, p = .035† 4.79 2.47
t(3)= 1.69, p = .19 55.75 20.55 t(25)=-0.59, p = .559 20.00 22.10 t(101)=2.58, p = .011† 35.79 19.73 t(28)=-4.54, p =.0009* 7.09 23.68 t(101)=1.53, p = .13 33.38 20.95 t(101)=1.90, p = .06 37.14 20.80 t(101)=2.35, p = .021† 34.64 19.91
* †
Significant at the 0.01 level (2-tailed) Exploratory (p ≤ 0.05)
Note: "DO" = Disorder
Table 22: T-Scores and Mean Problem Scale Scores for Groups of Subjects Split by Diagnoses
121
External Validator
Verbal Aggression
Physical Aggression
Hostile Affect
Covert Aggression
Bullying
Total Score
Autism Spectrum DO Present (n = 11) Not Present (n = 92) Comb. Behavior DO Present (n = 18) Not Present (n = 85) Comb. Personality Disorder Present (n = 9) Not Present (n = 94) Impulse Control DO Present (n = 9) Not Present (n = 94) Personality DO NOS Present (n = 7) Not Present (n = 96)
t(101)=-0.04, p = .964 -0.39 -0.34 t(101)=-0.75, p = .454 -0.87 -0.23 t(101)=-1.81, p = .074
t(101)=2.16, p =.033† 9.76 5.82 t(19)=-2.50, p = .022† 2.11 7.11 t(101)=-2.82, p = .006*
t(101)=-0.05, p = .957 -0.13 -0.10 t(101)=-0.97, p = .334 -0.42 -0.04 t(101)=-1.37, p = .174
t(101)=0.66, p = .508 1.59 1.51 t(101)=-1.99, p = .049† 1.36 1.55 t(101)=-0.16, p = .872
t(101)=-1.05, p = .297 0.74 1.09 t(101)=-1.43, p = .155 0.74 1.12 t(101)=-1.43, p = .156
t(101)=1.12, p = .264 10.70 7.61 t(19)=-1.98, p = .062 3.00 8.99 t(101)=-2.74, p = .007*
-2.21 -0.16 t(101)=-2.65, p = .009* -3.04 -0.08 t(101)=-1.33, p = .185 -1.94 -0.23
1.18 6.72 t(8)=-2.12, p = .065 0.14 6.82 t(101)=-3.15, p = .002* -0.17 6.70
-0.77 -0.04 t(101)=-3.47, p = .005* -1.44 0.02 t(101)=-1.81, p = .074 -1.10 -0.03
1.50 1.52 t(101)=-1.09, p = .278 1.39 1.53 t(101)=0.38, p = .707 1.57 1.52
0.58 1.10 t(101)=-1.56, p= .122 0.54 1.11 t(101)=-1.70, p = .092 0.41 1.10
0.61 8.64 t(101)=-3.95, p=.0009* -2.24 8.91 t(101)=-2.90, p = .005* -0.89 8.58
122
* †
Significant at the 0.01 level (2-tailed) Exploratory (p ≤ 0.05)
Note: "DO" = Disorder
Table 23: T-Scores and Mean Provocation Scale Scores for Groups of Subjects Split by Diagnoses
122
Diagnostic Category
Problem Scale
Provocation Scale
Provocation/Problem Sum (PPS)
Provocation/Problem Product (PPP)
Autistic Spectrum Disorders
(Verbal)* Physical†
Physical†
(Verbal)* Physical† (Hostile)†
nil
Combined Disruptive Behavior Disorders
Physical† Hostile† Bully† Total†
(Physical)† (Covert)†
nil
(Total)† (Physical)*
(Covert)† (Bully)†
Combined Anxiety Disorders
nil
nil
nil
nil
nil
Combined Mood Disorders
Hostile†
nil
nil
nil
Combined Personality Disorders
Hostile†
(Physical)* (Total)*
(Physical)*
(Verbal)† (Physical)† (Total)†
*
Significant at the 0.01 level (2-tailed) Exploratory (p ≤ 0.05)
†
Provocation Quotient Physical†
nil
(Physical)* (Total)*
Note: Bold = Difference unique to Provocation Scale (not found on Problem Scale)
123
Table 24: Comparison of Problem Scale Validators and Provocation Validators (Aggregate Diagnostic Categories)
123
External Validator
Verbal Aggression
Physical Aggression
Hostile Affect
Covert Aggression
Bullying
Total Score
Autism Spectrum Disorder Present (n = 11) Not Present (n = 92) Comb. Behavior Disorder Present (n = 18) Not Present (n = 85) Comb. Personality Disorder Present (n = 9) Not Present (n = 94) Comb. Mood Disorder Present (n = 9) Not Present (n = 94)
t(74)=-4.44, p = .0009*
t(101)=2.17, p = .032†
t(101)=-0.75, p = .454
t(101)=-0.62, p = .535
t(101)=-0.97, p = .335
t(44)=-0.78, p = .439
0.45 3.54 t(101)=1.49, p = .139
9.36 5.02 t(101)=1.99, p = .049†
6.64 8.74 t(101)=2.39, p = .019†
2.36 3.22 t(101)=0.77, p = .443
1.73 2.91 t(101)=2.27, p = .025†
19.64 22.18 t(101)=2.32, p = .022†
4.94 2.85 t(101)=1.75, p = .083
8.17 4.92 t(101)=0.74, p = .459
12.89 7.59 t(101)=2.29, p = .024†
3.83 2.98 t(101)=0.72, p = .473
4.61 2.40 t(8)=1.30, p = .228
32.72 19.62 t(101)=1.90, p = .06
6.22 2.93 t(101)=1.81, p = .072
7.00 5.34 t(101)=1.20, p = .233
14.78 7.91 t(101)=2.08, p = .04†
4.11 3.03 t(101)=0.82, p = .416
5.44 2.53 t(55)=1.55, p = .126
35.22 20.64 t(55)=1.85, p = .069
4.59 2.54
6.56 4.96
11.03 7.28
3.62 2.88
3.68 2.35
28.06 18.88
124
*
Significant at the 0.01 level (2-tailed) Exploratory (p ≤ 0.05)
†
Table 25: T-Scores and Mean Problem Scale Scores for Groups of Subjects Split by Aggregate Diagnostic Categories
124
Medication
Problem Scale
Provocation Scale
Provocation/Problem Sum (PPS)
Provocation/Problem Product (PPP)
Provocation Quotient
Anticonvulsant
nil
Verbal†
nil
Verbal†
nil
Antidepressant
nil
nil
nil
nil
nil
Antipsychotic
Hostile*
nil
nil
Total* Physical* Hostile* nil
Total* Physical* Hostile† nil
Beta Blocker
Total* Physical* Hostile* Total† Physical† Hostile† nil
(Total)* (Physical)* (Hostile)*
nil
(Total)* (Physical)† (Covert)*
(Hostile)† (Covert)*
Lithium
Hostile†
(Hostile)†
(Physical)* (Covert)*
(Covert)*
Sedative
nil
nil
(Physical)* (Covert)* (Bully)* nil
nil
nil
Stimulant
Total† Physical* Total* Physical* Hostile* Covert*
Verbal†
Total* Physical† Total†
Total† Physical† Total* Physical*
Anxiolytic
125
Any Drug
*
Significant at the 0.01 level (2-tailed) Exploratory (p ≤ 0.05)
†
nil
nil
Total* Physical* Total* Verbal† Physical* Hostile† Bully†
Note: Bold = Difference unique to Provocation Scale (not found on Problem Scale)
Table 26: Comparison of Problem Scale Validators and Provocation Validators (Medications)
125
126
External Validator
Verbal Aggression
Physical Aggression
Hostile Affect
Covert Aggression
Bullying
Total Score
Antipsychotics Present (n = 50) Not Present (n = 53) Anxiolytics Present (n = 22) Not Present (n = 81) Lithium Present (n = 5) Not Present (n = 98) Stimulants Present (n = 6) Not Present (n = 97) Any Drug Present (n = 77) Not Present (n = 26)
t(101)=1.21, p = .23 3.88 2.58 t(101)=0.85, p = .398 4.09 2.98 t(101)=1.34, p = .182 6.40 3.05 t(5)=1.15, p = .300 8.33 2.90 t(101)=1.02, p = .31 3.53 2.27
t(68)=6.08, p = .0009* 8.92 2.25 t(28)=2.15, p = .04† 8.41 4.69 t(101)=0.61, p = .541 7.20 5.40 t(101)=3.09, p = .003* 13.00 5.02 t(89)=4.78, p = .0009* 6.69 1.92
t(89)=4.14, p = .0009* 11.96 5.26 t(101)=2.00, p = .049† 11.77 7.63 t(101)=2.38, p = .019† 17.40 8.06 t(101)=1.39, p = .166 13.33 8.22 t(68)=3.25, p = .002* 9.79 4.73
t(101)=1.90, p = .061 3.94 2.36 t(27)=1.74, p = .094 4.77 2.68 t(101)=-0.81, p = .417 1.60 3.20 t(5)=1.00, p = .377 5.83 2.96 t(80)=2.93, p = .004* 3.66 1.54
t(101)=1.96, p = .052 3.54 2.08 t(101)=0.98, p = .327 3.50 2.59 t(101)=-0.35, p = .728 2.20 2.82 t(101)=1.47, p = .146 5.00 2.65 t(101)=1.82, p = .071 3.18 1.62
t(88)=4.30, p = .0009* 30.88 13.45 t(101)=2.22, p = .02.8† 31.09 19.42 t(101)=1.25, p = .215 34.00 21.30 t(101)=2.33, p = .0.22† 42.00 20.67 t(79)=3.89, p = .0009* 25.55 11.15
*
Significant at the 0.01 level (2-tailed) Exploratory (p ≤ 0.05)
†
Table 27: T-Scores and Mean Problem Scale Scores for Groups of Subjects Split by Medications
126
127
External Validator
Verbal Aggression
Physical Aggression
Hostile Affect
Covert Aggression
Bullying
Total Score
Anticonvulsant Present (n = 43) Not Present (n = 60) Antipsychotics Present (n = 50) Not Present (n = 53) Beta Blockers Present (n = ) Not Present (n = ) Lithium Present (n = 5) Not Present (n = 98) Stimulants Present (n = 6) Not Present (n = 97)
t(101)=2.18, p = .032† 0.47 -0.93 t(101)=2.22, p = .029† 0.38 -1.03 t(101)=-1.37, p = .174 -2.54 -0.25 t(101)=-0.16 , p = .873 -0.57 -0.33 t(101)=2.06, p = .042† 2.29 -0.51
t(101)=-0.14, p = .887 6.14 6.31 t(54)=0.12, p = .903 6.31 6.17 t(101)=-2.96, p = .004* -1.89 6.57 t(101)=-2.66, p = .009* -0.33 6.57 t(101)=1.45, p = .151 9.55 6.03
t(101)=1.26, p = .211 0.12 -0.27 t(91)=2.74, p = .008* 0.30 -0.50 t(101)=-3.01, p = .003* -2.28 -0.02 t(101)=-2.10, p = .038† -1.48 -0.04 t(101)=1.15, p = .254 0.59 -0.15
t(101)=-0.18, p = .86 1.51 1.53 t(101)=0.01, p = .988 1.52 1.52 t(3)=-1.09, p = .354 1.13 1.54 t(101)=-0.12, p = .907 1.50 1.52 t(5)=-0.70, p = .513 1.33 1.53
t(101)=1.00, p = .32 1.18 0.97 t(101)=1.33, p = .185 1.20 0.92 t(101)=-0.82, p = .416 0.64 1.07 t(101)=-0.17, p = .868 0.98 1.06 t(101)=0.64, p = .526 1.32 1.04
t(101)=1.02, p = .311 8.97 7.20 t(68)=1.44, p = .156 9.22 6.73 t(101)=-2.97, p = .004* -4.20 8.43 t(4)=-1.22, p = .289 -0.02 8.35 t(101)=1.93, p = .057 14.46 7.54
*
Significant at the 0.01 level (2-tailed) Exploratory (p ≤ 0.05)
†
Table 28: T-Scores and Mean Provocation Scale Scores for Groups of Subjects Split by Medications
127
Scale or Subscale
Intraclass Correlation
Intraclass Correlation
(Excluding Across Settings)
(Including Across Settings)
Verbal Aggression
.539
.518
Physical Aggression
.776
.874
Hostile Affect
.706
.667
Covert Aggression
.654
.510
Bullying
.592
.458
Total
.548
.657
Table 29: Interrater reliability of the Provocation Total and subscales
128
Scale or Subscale
Intraclass Correlation
Intraclass Correlation
Quotient
(Excluding Across Settings)
(Including Across Settings)
Verbal Aggression
-.189
-.316
Physical Aggression
.066
.385
Hostile Affect
.047
.007
Covert Aggression
.345
.281
Bullying
.657
.375
Full Provocation Scale
-.129
-.029
Table 30: Interrater reliability of the Provocation Quotient and subscale Quotients
129
Scale or Subscale
Intraclass Correlation
Intraclass Correlation
Prov./Prob. Sum
(Excluding Across Settings)
(Including Across Settings)
Verbal Aggression
.230
.269
Physical Aggression
.731
.649
Hostile Affect
.550
.548
Covert Aggression
.743
.621
Bullying
-.137
.548
Full Provocation Scale
.335
.452
Table 31: Interrater reliability of the Provocation/Problem Sums for the full scale and subscales
130
Scale or Subscale
Intraclass Correlation
Intraclass Correlation
Prov./Prob. Product
(Excluding Across Settings)
(Including Across Settings)
Verbal Aggression
-.295
-.103
Physical Aggression
-.070
.424
Hostile Affect
.079
-.007
Covert Aggression
.728
.619
Bullying
.752
.217
Full Provocation Scale
.075
.038
Table 32: Interrater reliability of the Provocation/Problem Products for the full scale and subscales
131
Problem Scale Problem Scale
Provocation Scale
Provocation/ Problem Sum (PPS)
Provocation/ Problem Product (PPP)
Provocation Quotient
1
.17*
.33*
.78*
.33*
Provocation Scale
.17*
1
.71*
.94*
.71*
PPS
.33*
.71*
1
.80*
.69*
PPP
.78*
.94*
.80*
1
.73*
Provocation Quotient
.33*
.71*
.69*
.73*
1
Table 33: Correlations between the Problem Scale, Provocation Scale, and Provocation Index Totals
132
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
5.0
-3 – 20
4.0
Physical Aggression
2.1
-14 – 18
4.5
Hostile Affect
-3.8
-9 – 2
2.0
Covert Aggression
7.3
6–8
0.4
Bullying
9.8
6 – 14
1.1
10.1
-18 – 53
9.9
Provocation Scale Total
Table 34: Provocation Scale Total and Subscale Scores for Subjects Eighteen to Thirty Years of Age (n = 115)
133
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.1
-10– 20
4.0
Physical Aggression
1.3
-11 – 16
3.8
Hostile Affect
-4.0
-8 – 2
1.9
Covert Aggression
7.2
5–8
0.5
Bullying
9.6
4 – 14
1.4
Provocation Scale Total
7.9
-20 – 36
9.1
Table 35: Provocation Scale Total and Subscale Scores for Subjects Thirty-one to Forty Years of Age (n = 121)
134
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
3.7
-8 – 14
2.7
Physical Aggression
1.5
-11 – 15
3.2
Hostile Affect
-4.2
-8 – 2
1.5
Covert Aggression
7.2
5–8
0.5
Bullying
9.5
6 – 13
0.9
Provocation Scale Total
7.5
-14 – 32
6.5
Table 36: Provocation Scale Total and Subscale Scores for Subjects Forty-one to Fifty Years of Age (n = 147).
135
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
3.6
-7 – 16
3.3
Physical Aggression
0.6
-15 – 9
3.2
Hostile Affect
-4.6
-10 – 0
1.7
Covert Aggression
7.1
4–8
0.7
Bullying
9.4
6 – 12
1.1
Provocation Scale Total
5.8
-35 – 30
8.5
Table 37: Provocation Scale Total and Subscale Scores for Subjects Fifty-one to Sixty Years of Age (n = 90)
136
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.5
-5 – 15
2.9
Physical Aggression
0.8
-10 – 7
2.8
Hostile Affect
-3.6
-7 – -1
1.2
Covert Aggression
7.2
6–8
0.3
Bullying
9.6
8 – 11
0.6
Provocation Scale Total
8.3
-7 – 22
5.3
Table 38: Provocation Scale Total and Subscale Scores for Subjects Aged Sixty Years or More (n = 31)
137
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.0
-10 – 20
3.2
Physical Aggression
1.4
-11 – 18
3.6
Hostile Affect
-4.1
-8 – 2
1.6
Covert Aggression
7.2
5–8
0.5
Bullying
9.6
4 – 14
1.1
Provocation Scale Total
8.0
-20 – 53
7.8
Table 39: Provocation Scale Total and Subscale Scores for Male Subjects (n = 314)
138
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.3
-7 – 20
3.9
Physical Aggression
1.3
-15 – 14
3.8
Hostile Affect
-4.1
-10 – 1
1.9
Covert Aggression
7.1
4–8
0.5
Bullying
9.6
6 – 13
1.0
Provocation Scale Total
7.9
-35 – 38
9.1
Table 40: Provocation Scale Total and Subscale Scores for Female Subjects (n = 190)
139
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
3.6
-10 – 20
3.8
Physical Aggression
0.9
-15 – 18
2.9
Hostile Affect
-4.4
-9 – 2
1.7
Covert Aggression
7.2
4–8
0.5
Bullying
9.5
6 – 14
1.0
Provocation Scale Total
6.5
-35 – 53
8.2
Table 41: Provocation Scale Total and Subscale Scores for Subjects Diagnosed with Mild IDD (n = 222)
140
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.7
-6 – 18
3.8
Physical Aggression
1.5
-11 – 14
3.7
Hostile Affect
-3.8
-8 – 2
1.9
Covert Aggression
7.2
5–8
0.5
Bullying
9.6
5 – 13
1.2
Provocation Scale Total
8.9
-14 – 38
8.5
Table 42: Provocation Scale Total and Subscale Scores for Subjects Diagnosed with Moderate IDD (n = 172)
141
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.3
-4 – 11
1.8
Physical Aggression
2.1
-9 – 16
4.7
Hostile Affect
-3.8
-10 – 2
1.5
Covert Aggression
7.3
5–8
0.5
Bullying
9.7
4 – 14
1.2
Provocation Scale Total
9.4
-14 – 32
8.2
Table 43: Provocation Scale Total and Subscale Scores for Subjects Diagnosed with Severe/Profound IDD (n = 115)
142
Scale or Subscale
Mean Score
Range
Standard Deviation
Verbal Aggression
4.1
-10 – 20
3.5
Physical Aggression
1.4
-15 – 18
3.6
Hostile Affect
-4.1
-10 – 2
1.7
Covert Aggression
7.2
4–8
0.5
Bullying
9.6
4 – 14
1.1
Provocation Scale Total
7.4
-34 – 51
7.9
Table 44: Provocation Scale Total and Subscale Scores for All Subjects (n = 512)
143
144
Verbal Aggression Physical Aggression Hostile Affect Covert Aggression Bullying Verbal Agg. Provocation Phys. Agg. Provocation Hostile Aff. Provocation Covert Agg. Provocation Bullying Provocation
Verbal Aggression 1.0
Physical Aggression .23**
Hostile Affect .65**
Covert Aggression .44**
.23**
1.0
.51**
.65**
.51**
.44**
Bullying .45**
Verbal Agg. Provocation .20**
Phys. Agg. Provocation .02
Hostile Aff. Provocation .09*
Covert Agg. Provocation -.10*
Bullying Provocation .02
.42**
.51**
.22**
.25**
.28**
.12**
.16**
1.0
.58**
.54**
.11*
-.002
.12**
-.04
-.02
.42**
.58**
1.0
.58**
.12**
.10*
.14**
.09*
.07
.45**
.51**
.54**
.58**
1.0
.21**
.15**
.24**
.08
.15**
.20**
.22**
.11*
.12**
.21**
1.0
.43**
.64**
.18**
.45**
.02
.25**
-.002
.10*
.15**
.43**
1.0
.53**
.21**
.51**
.09*
.28**
.12**
.14**
.24**
.64**
.53**
1.0
.23**
.42**
-.10*
.12**
-.04
.09*
.08
.18**
.21**
.23**
1.0
.25**
.02
.16**
-.02
.07
.15**
.45**
.51**
.42**
.25**
1.0
Table 45: Matrix of Correlations between Problem and Provocation Subscales
144
APPENDIX B THE SCALE OF HOSTILITY AND AGGRESSION, REACTIVE/PROACTIVE (SHARP) AND AGGRESSION/DESTRUCTIVE BEHAVIOR SUBSCALE OF THE BEHAVIOR PROBLEMS INVENTORY (BPI-01)
145
SCALE of HOSTILITY & AGGRESSION, REACTIVE/PROACTIVE (SHARP) SUBJECT'S ID CODE:___________________________ DATE: ____/____/____ Subject's… Date of Birth: XX /____/____ Gender (please circle): F M Subject is: ____African American ____Caucasian (European origin) ____ Hispanic ____ Asian American ____ Other (if checked, please specify:___________________) IQ/severity of intellectual handicap: __mild, __moderate, __severe/profound. RATER'S NAME:_____________________________ Relationship of rater to subject (e.g., parent, teacher, instructor, etc.): How long have you (the rater) known the subject?: Does the subject have any psychiatric diagnoses? ___NO ____YES If YES, please list all diagnoses below: ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ Does the subject have a behavior support plan for aggressive or hostile behavior(s)?___NO ___YES Does the subject take any medicines for behavioral or emotional problems? ___NO __YES If YES, please list all medicines below: Medicine Dose/day ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ Does the subject have any genetic disorders? Please check any that apply: ____ Down Syndrome ____ Fragile X Syndrome ____ Autistic Disorder, or other PDD Spectrum Disorder ____ Other (if checked, please specify:_______________________________________ 146
INSTRUCTIONS: Based on your experience of the person’s behavior in the past six months, please first circle ONE "Problem Scale" answer for each of the items. Problem Scale: Please circle the number which best reflects the person’s behavior… 0 = Does not happen (NOTE: If circled, the "Who starts it?" rating [next column over] is NOT scored for that item) 1 = Mild or infrequent problem 2 = Moderately serious and/or frequent problem 3 = Severe and/or very frequent problem Next, if this answer is a "1," "2," or "3" (that is, if it is a problem at all), circle ONE "Who starts it?" rating for the item if applicable. "Who starts it?" Rating: Where applicable, please circle the number which best reflects the client's behavior… -2 = Only when provoked, unplanned, or when s/he "just loses it;" -1 = Usually provoked and/or unplanned; 0 = Equally likely to happen with or without provocation; 1 = Usually the first to act, without provocation; 2 = Always the first to act (that is, always the instigator; the one who "starts it"); DEFINITIONS: Instigator: The one who deliberately "starts it;" the first or only one to act in an incident of aggression; Provocation: Any action leading to the aggression, no matter how mild (e.g., verbal teasing or being crowded) or severe (e.g., being hit or kicked), which seems to anger or upset the subject and trigger the behavior; Problem Scale Provoked
0 0
1 1
2 2
3 3
-2
-1
0 N/A
1
2
0
1
2
3
-2
-1
0
1
2
0
1
2
3
-2
-1
0
1
2
0
1
2
3
0 0 0 0
1 1 1 1
2 2 2 2
3 3 3 3
1
2
1
2
147
Not Provoked
Severe/ Frequent
Doesn't happen
1. Sneers, "makes faces" at others 2. Is "sneaky;" does things "on the sly" 3. Attempts to touch others in inappropriate, sexually suggestive manner 4. Pinches others 5. Is resentful over seemingly minor issues 6. Breaks others' things 7. Is quick to anger ("hot-headed") 8. Takes others' things by force 9. Broods, pouts, or is sullen
Who starts it?
N/A -2 -2
-1
0 N/A -1 0 N/A
10. Calls others insulting names in their absence
0
1
2
3
-2
Problem Scale
0
2
Not Provoked
Provoked
0
1
2
3
-2
0
1
2
3
N/A
0
1
2
3
N/A
0
1
2
3
-2
-1
0 0 0
1 1 1
2 2 2
3 3 3
-2 -2 -2
-1 -1 -1
0
1
2
3
0
1
2
3
-2
-1
0
1
2
3
-2
-1
0
1
2
3
N/A
0
1
2
3
N/A
0
1
2
3
-2
-1
0
1
2
3
-2
-1
0
1
2
3
0
1
2
3
0
1
2
3
0
1
2
3
-2
-1
0 0
1 1
2 2
3 3
-2 -2
-1 -1
0
1
2
3
0
1
2
3
-2
-1
0
1
2
3
-2
0
1
2
3
-2
0
1
2
3
N/A
0
1
2
3
N/A
0 0 0
1 1 1
2 2 2
3 3 3
-2
0
1
2
3
148
1
Who starts it?
Severe/ Frequent
Doesn't Happen
11. Shoves or pushes others 12. Crowds others (invades their personal space) 13. Intimidates others 14. Says, "I hate you" or makes other hurtful statements 15. Bites others 16. Insults others to their faces 17. Throws objects at others 18. Reacts suddenly or impulsively to minor provocations 19. Shouts at others in anger 20. Uses headlocks or other “wrestling” holds against others 21. Gets mad when caught behaving badly 22. Is overly argumentative 23. Uses profanity to shock or offend others 24. Cuts others with a sharp object 25. Tickles or otherwise physically teases others, even after being asked to stop 26. Steals from others when they aren't looking 27. Reacts to insults or teasing by lashing out physically 28. Calls others insulting names to their faces 29. Trips others 30. Head-butts others 31. Makes insulting comments about others behind their backs 32. Breaks own belongings 33. Makes threatening gestures (like shaking fist) 34. Charges at others 35. Verbally teases others, even after being asked to stop 36. If caught, denies having behaved badly 37. Pulls others' hair 38. When angry, is slow to cool off 39. Spits at others 40. Says "I hate [someone not present]" or makes other hurtful statements in absence of the person concerned
-1
-1
0
1
2
0
1
2
0 0 0
1 1 1
2 2 2
0
1
2
0
1
2
0
1
2
0
1
2
1
2
0
1
2
0 0
1 1
2 2
0
1
2
-1
0
1
2
-1
0
1
2
1
2
-2
0 N/A -1 0
1
2
-2
-1
1
2
N/A
N/A -2
-1
0 N/A
N/A
-1
0
41. Chokes others/Grabs others' necks 42. Lashes out at people who are in his/her space 43. Takes offense at things others say or do when s/he shouldn't 44. Starts trouble by baiting others 45. If caught, makes excuses for bad behavior 46. Is hostile, "has a chip on his/her shoulder" 47. Scratches others with fingernails 48. Glares at others 49. Encourages others to gang up on someone (physically OR verbally) 50. Hits others with objects 51. Is often grouchy 52. Doesn't care about others' feelings 53. Verbally threatens others with physical harm 54. Hits or shoves others forcefully 55. Makes unwanted sexual comments to others 56. Does not seem truly sorry after hurting someone 57. Gets revenge after some time has passed and the other person is not on guard 58. Tries not to get caught while doing harmful things to others
0
1
2
3
-2
-1
0
1
2
0
1
2
3
0
1
2
3
1
2
0
1
2
3
N/A
0
1
2
3
N/A
0
1
2
3
N/A
0 0
1 1
2 2
3 3
1 1
2 2
0
1
2
3
0 0 0
1 1 1
2 2 2
3 3 3
-2
-1
0 N/A N/A
1
2
0
1
2
3
-2
-1
0
1
2
0
1
2
3
-2
-1
0
1
2
0
1
2
3
N/A
0
1
2
3
N/A
0
1
2
3
N/A
0
1
2
3
N/A
N/A -2
-2 -2
-1
-1 -1
0
0 0 N/A
Not Provoked
Provoked
Severe/ Frequent
Doesn't Happen
Problem Scale: Please circle the number which best reflects the subject's behavior… 0 = Does not happen (NOTE: If circled, Reactive/Proactive is NOT scored for that item) 1 = Mild or infrequent problem 2 = Moderately serious and/or frequent problem 3 = Severe and/or very frequent problem "Who starts it?" Rating: Where applicable, please circle the number which best reflects the subject's behavior… -2 = Only when provoked, not planned, and/or when s/he "just loses it;" -1 = Usually with provocation and/or unplanned; 0 = Equally likely to happen with or without provocation; 1 = Usually the first to act, without provocation; 2 = Always the first to act (i.e., always the instigator; the one who "starts it");
149
Aggression/Destructive Behavior subscale of the Behavior Problems Inventory (BPI-01) SUBJECT’S ID CODE:__________________________________ INSTRUCTIONS: Based on your experience of the person’s behavior in the past month, please first circle ONE "Frequency Scale" answer for each of the items, either “never”, “monthly”, “weekly”, “daily” or “hourly”. If the answer is “never”, move on to the next item. If the answer is any other than “never”, circle ONE severity rating for the item, either “slight”, “moderate”, or “severe”. GENERIC DEFINITION: Aggressive or destructive behaviors are offensive actions or deliberate overt attacks directed towards other individuals or objects. They occur repeatedly in the same way over and over again, and they are characteristic for that person. Frequency Scale Never
Monthly
Weekly
Daily
Hourly
Slight
Moderate
Severe
1. Hitting others 2. Kicking others 3. Pushing others 4. Biting others 5. Grabbing and pulling others 6. Scratching others 7. Pinching others 8. Spitting on others 9. Being verbally abusive with others 10. Destroying things (e.g., rips clothes, throws chairs, smashes tables) 11. Being mean or cruel (e.g., grabbing toys or food from others, bullying others)
Severity Scale
0 0 0 0 0 0 0 0 0
1 1 1 1 1 1 1 1 1
2 2 2 2 2 2 2 2 2
3 3 3 3 3 3 3 3 3
4 4 4 4 4 4 4 4 4
1 1 1 1 1 1 1 1 1
2 2 2 2 2 2 2 2 2
3 3 3 3 3 3 3 3 3
0
1
2
3
4
1
2
3
0
1
2
3
4
1
2
3
150
APPENDIX C FOUR FACTOR SOLUTION USING EQUAMAX ROTATION
151
4 Factor Solution using Equamax Rotation: 1. Sneers, "makes faces" at others 2. Is "sneaky;" does things "on the sly" 3. Attempts to touch others in inappropriate, sexually suggestive manner 4. Pinches others 5. Is resentful over seemingly minor issues 6. Breaks others' things 7. Is quick to anger ("hot-headed") 8. Takes others' things by force 9. Broods, pouts, or is sullen 10. Calls others insulting names in their absence 11. Shoves or pushes others 12. Crowds others 13. Intimidates others 14. Says, "I hate you" or makes other hurtful statements 15. Bites others 16. Insults others to their faces 17. Throws objects at others 18. Reacts suddenly or impulsively to minor provocations 19. Shouts at others in anger 21. Gets mad when caught behaving badly 22. Is overly argumentative 23. Uses profanity to shock or offend others 25. Tickles or otherwise physically teases others, even after being asked to stop 26. Steals from others when they aren't looking 27. Reacts to insults or teasing by lashing out physically 28. Calls others insulting names to their faces 30. Head-butts others 31. Makes insulting comments about others behind their backs 32. Breaks own belongings 33. Makes threatening gestures 34. Charges at others 35. Verbally teases others, even after being asked to stop 36. If caught, denies having behaved badly 37. Pulls others' hair 38. When angry, is slow to cool off 39. Spits at others 40. Says "I hate [someone not present]" in absence of the person concerned 42. Lashes out at people who are in his/her space 43. Takes offense at things others say or do when s/he shouldn't 44. Starts trouble by baiting others 45. If caught, makes excuses for bad behavior 46. Is hostile, "has a chip on his/her shoulder" 47. Scratches others with fingernails 48. Glares at others
.181 .137
.429 .122
.224 .288
.272 .568
.164 .476 .148 .500 .406 .365 .073 -.046 .636 .389 .352
.079 .108 .198 .059 .269 .125 .253 .743 .140 .027 .299
.044 -.038 .681 .177 .641 .027 .598 .265 .146 .112 .269
.456 .173 .210 .330 .113 .492 .110 .135 .261 .362 .439
.110 .623 .080 .619
.744 -.047 .820 .121
.339 .108 .189 .236
.097 .077 .169 .284
.473 .259 .237 -.089 .141
.203 .449 .174 .440 .668
.578 .531 .579 .590 .119
.135 .108 .327 .176 .099
.175 .165
.150 -.043
-.035 .190
.500 .499
.372 .073 .568
.261 .874 -.088
.340 .205 .097
.220 .116 .067
-.176 .537 .408 .540
.697 -.006 .360 .100
.324 .240 .221 .211
.162 .197 .336 .244
-.038 -.027 .542 .393 .524
.395 .270 .045 .255 .169
.116 .528 .022 .579 .068
.562 .370 -.017 .151 .030
-.036
.683
.401
.082
.464
.219
.339
.161
.055 .045 -.075 .132 .716 .299 152
.341 .260 .320 .323 -.062 .225
.609 .197 .539 .478 .079 .406
.162 .627 .341 .168 .079 .320
(Continued) Continued 49. Encourages others to gang up on someone (physically OR verbally) 50. Hits others with objects 51. Is often grouchy 52. Doesn't care about others' feelings 53. Verbally threatens others with physical harm 54. Hits or shoves others forcefully 55. Makes unwanted sexual comments to others 56. Does not seem truly sorry after hurting someone 57. Gets revenge after some time has passed and the other person is not on guard 58. Tries not to get caught while doing harmful things to others
-.126 .610 .220 .365 .198 .717 -.036
.212 .087 .309 .205 .530 .042 .441
.146 .279 .573 .298 .283 .227 .118
.469 .213 .071 .274 .177 .209 .270
.150
.200
.270
.334
.351
.043
.213
.519
.104
.112
.222
.675
153
APPENDIX D SIX FACTOR SOLUTION USING EQUAMAX ROTATION
154
6 Factor Solution using Equamax Rotation: 1. Sneers, "makes faces" at others 2. Is "sneaky;" does things "on the sly" 3. Attempts to touch others in inappropriate, sexually suggestive manner 4. Pinches others 5. Is resentful over seemingly minor issues 6. Breaks others' things 7. Is quick to anger ("hot-headed") 8. Takes others' things by force 9. Broods, pouts, or is sullen 10. Calls others insulting names in their absence 11. Shoves or pushes others 12. Crowds others 13. Intimidates others 14. Says, "I hate you" or makes other hurtful statements 15. Bites others 16. Insults others to their faces 17. Throws objects at others 18. Reacts suddenly or impulsively to minor provocations 19. Shouts at others in anger 21. Gets mad when caught behaving badly 22. Is overly argumentative 23. Uses profanity to shock or offend others 25. Tickles or otherwise physically teases others, even after being asked to stop 26. Steals from others when they aren't looking 27. Reacts to insults or teasing by lashing out physically 28. Calls others insulting names to their faces 30. Head-butts others 31. Makes insulting comments about others behind their backs 32. Breaks own belongings 33. Makes threatening gestures 34. Charges at others 35. Verbally teases others, even after being asked to stop 36. If caught, denies having behaved badly 37. Pulls others' hair 38. When angry, is slow to cool off 39. Spits at others 40. Says "I hate [someone not present]" in absence of the person concerned 42. Lashes out at people who are in his/her space 43. Takes offense at things others say or do when s/he shouldn't 44. Starts trouble by baiting others 45. If caught, makes excuses for bad behavior 46. Is hostile, "has a chip on his/her shoulder" 47. Scratches others with fingernails
.385 .100
.070 .138
.282 .097
.095 .302
.187 .571
.314 .170
.013 .104 .134 .032 .213 .077 .202 .734 .080 -.047 .236
.033 .461 .035 .477 .300 .247 -.010 -.067 .528 .257 .230
.046 .049 .588 .135 .629 .096 .509 .231 .235 .144 .281
.100 -.076 .396 .166 .309 .007 .356 .184 .083 .127 .191
.191 .222 .183 .283 .145 .315 .097 .167 .135 .107 .283
.540 .129 .172 .226 .146 .494 .100 .061 .394 .521 .433
.706 -.064 .784 .089
.074 .584 .047 .595
.290 .196 .166 .206
.278 -.007 .206 .185
.073 .111 .127 .267
.132 .128 .183 .220
.147 .391 .116 .389 .648
.372 .160 .186 -.122 .152
.571 .522 .473 .506 .063
.291 .290 .392 .374 .206
.145 .115 .290 .137 .058
.171 .157 .166 .076 .118
.074 -.056
.000 .166
.009 .060
.040 .186
.155 .518
.710 .129
.200 .845 -.097
.269 .049 .573
.319 .192 .107
.309 .194 .064
.091 .097 .081
.281 .139 .083
.696 -.041 .292 .055
-.178 .515 .292 .441
.235 .184 .245 .299
.235 .220 .185 .068
.231 .132 .174 .187
-.021 .196 .436 .288
.329 .208 .041 .203 .163
-.152 .014 .556 .293 .563
.005 .110 .062 .572 .044
.261 .746 .021 .279 .119
.275 .288 .019 .180 .040
.556 .081 .046 .152 .060
.670
-.032
.285
.317
.144
-.022
.163
.325
.472
.067
.136
.272
.292 .227 .254 .280 -.067 155
.002 -.012 -.034 .031 .706
.452 .090 .097 .489 .173
.442 .209 .800 .205 -.032
.150 .542 .227 .194 .164
.084 .318 .085 .128 .069
(Continued) Continued 48. Glares at others 49. Encourages others to gang up on someone (physically OR verbally) 50. Hits others with objects 51. Is often grouchy 52. Doesn't care about others' feelings 53. Verbally threatens others with physical harm 54. Hits or shoves others forcefully 55. Makes unwanted sexual comments to others 56. Does not seem truly sorry after hurting someone 57. Gets revenge after some time has passed and the other person is not on guard 58. Tries not to get caught while doing harmful things to others
.176
.187
.432
.172
.282
.266
.204 .045 .265 .164 .482 -.010 .413
-.138 .611 .120 .288 .166 .645 -.086
.027 .165 .599 .307 .186 .260 .072
.158 .324 .222 .165 .325 .159 .152
.465 .134 .152 .250 .068 .115 .151
.133 .216 .059 .220 .224 .316 .258
.151
.066
.224
.214
.237
.274
.018
.276
.240
.031
.615
.211
.098
.079
.095
.153
.807
.155
156
APPENDIX E FIVE FACTOR SOLUTION USING OBLIMIN ROTATION
157
5 Factor Solution using Oblimin Rotation: 1. Sneers, "makes faces" at others 2. Is "sneaky;" does things "on the sly" 3. Attempts to touch others in inappropriate, sexually suggestive manner 4. Pinches others 5. Is resentful over seemingly minor issues 6. Breaks others' things 7. Is quick to anger ("hot-headed") 8. Takes others' things by force 9. Broods, pouts, or is sullen 10. Calls others insulting names in their absence 11. Shoves or pushes others 12. Crowds others 13. Intimidates others 14. Says, "I hate you" or makes other hurtful statements 15. Bites others 16. Insults others to their faces 17. Throws objects at others 18. Reacts suddenly or impulsively to minor provocations 19. Shouts at others in anger 21. Gets mad when caught behaving badly 22. Is overly argumentative 23. Uses profanity to shock or offend others 25. Tickles or otherwise physically teases others, even after being asked to stop 26. Steals from others when they aren't looking 27. Reacts to insults or teasing by lashing out physically 28. Calls others insulting names to their faces 30. Head-butts others 31. Makes insulting comments about others behind their backs 32. Breaks own belongings 33. Makes threatening gestures 34. Charges at others 35. Verbally teases others, even after being asked to stop 36. If caught, denies having behaved badly 37. Pulls others' hair 38. When angry, is slow to cool off 39. Spits at others 40. Says "I hate [someone not present]" in absence of the person concerned 42. Lashes out at people who are in his/her space 43. Takes offense at things others say or do when s/he shouldn't 44. Starts trouble by baiting others 45. If caught, makes excuses for bad behavior 46. Is hostile, "has a chip on his/her shoulder" 47. Scratches others with fingernails
.429 .388
.210 .270
.306 .697
.492 .266
.427 .353
.199 .120 .753 .334 .778 .235 .659 .464 .382 .287 .500
.158 .510 .219 .563 .467 .386 .132 .023 .636 .376 .394
.305 .204 .367 .378 .305 .391 .276 .315 .224 .233 .431
.132 .122 .362 .143 .409 .184 .384 .801 .189 .082 .397
.574 .247 .328 .372 .336 .591 .224 .195 .528 .575 .582
.573 .233 .464 .426
.163 .643 .129 .682
.275 .114 .305 .371
.802 -.011 .853 .208
.271 .246 .317 .397
.722 .705 .678 .671 .353
.528 .308 .336 -.002 .182
.293 .275 .492 .379 .219
.335 .556 .337 .576 .686
.357 .330 .347 .205 .226
.163 .249
.142 .275
.287 .578
.187 .074
.709 .281
.522 .482 .193
.406 .124 .591
.276 .277 .114
.347 .902 -.051
.407 .275 .180
.470 .366 .477 .397
-.082 .579 .425 .562
.374 .250 .324 .247
.778 .073 .428 .167
.118 .315 .573 .433
.318 .593 .144 .722 .224
-.008 .099 .556 .453 .557
.478 .564 .042 .321 .123
.475 .432 .051 .388 .188
.610 .239 .132 .342 .164
.560
.048
.321
.767
.127
.526
.479
.202
.295
.425
.692 .354 .605 .594 .212
.136 .148 .046 .188 .755
.352 .647 .532 .304 .144
.483 .233 .382 .478 .479 .228 .433 .281 -.026 .212 (Continued)
158
Continued 48. Glares at others 49. Encourages others to gang up on someone (physically OR verbally) 50. Hits others with objects 51. Is often grouchy 52. Doesn't care about others' feelings 53. Verbally threatens others with physical harm 54. Hits or shoves others forcefully 55. Makes unwanted sexual comments to others 56. Does not seem truly sorry after hurting someone 57. Gets revenge after some time has passed and the other person is not on guard 58. Tries not to get caught while doing harmful things to others
.563
.358
.392
.343
.430
.219 .446 .677 .459 .499 .428 .278
-.028 .661 .279 .419 .237 .731 -.003
.523 .301 .262 .344 .274 .224 .288
.308 .175 .427 .297 .592 .110 .487
.252 .361 .233 .377 .342 .458 .320
.403
.192
.349
.291
.391
.350
.448
.573
.162
.433
.327
.249
.791
.259
.392
159
APPENDIX F FIVE FACTOR SOLUTION USING QUARTIMAX ROTATION
160
5 Factor Solution using Quartimax Rotation: 1. Sneers, "makes faces" at others 2. Is "sneaky;" does things "on the sly" 3. Attempts to touch others in inappropriate, sexually suggestive manner 4. Pinches others 5. Is resentful over seemingly minor issues 6. Breaks others' things 7. Is quick to anger ("hot-headed") 8. Takes others' things by force 9. Broods, pouts, or is sullen 10. Calls others insulting names in their absence 11. Shoves or pushes others 12. Crowds others 13. Intimidates others 14. Says, "I hate you" or makes other hurtful statements 15. Bites others 16. Insults others to their faces 17. Throws objects at others 18. Reacts suddenly or impulsively to minor provocations 19. Shouts at others in anger 21. Gets mad when caught behaving badly 22. Is overly argumentative 23. Uses profanity to shock or offend others 25. Tickles or otherwise physically teases others, even after being asked to stop 26. Steals from others when they aren't looking 27. Reacts to insults or teasing by lashing out physically 28. Calls others insulting names to their faces 30. Head-butts others 31. Makes insulting comments about others behind their backs 32. Breaks own belongings 33. Makes threatening gestures 34. Charges at others 35. Verbally teases others, even after being asked to stop 36. If caught, denies having behaved badly 37. Pulls others' hair 38. When angry, is slow to cool off 39. Spits at others 40. Says "I hate [someone not present]" in absence of the person concerned 42. Lashes out at people who are in his/her space 43. Takes offense at things others say or do when s/he shouldn't 44. Starts trouble by baiting others 45. If caught, makes excuses for bad behavior 46. Is hostile, "has a chip on his/her shoulder" 47. Scratches others with fingernails
.526 .409
.155 .247
.007 .528
.203 .076
.130 -.032
.224 .087 .669 .247 .665 .244 .619 .711 .285 .217 .517
.174 .497 .168 .551 .403 .390 .074 -.091 .624 .392 .360
.126 .094 .016 .189 -.085 .189 -.035 .000 -.049 .016 .105
.490 .034 .003 .073 -.075 .404 -.050 -.042 .231 .408 .290
-.050 .181 -.324 -.003 -.207 .054 -.229 .370 .008 -.111 .015
.761 .068 .726 .319
.048 .643 .014 .657
-.104 -.024 -.052 .127
-.027 -.017 .057 .030
.307 -.025 .454 .015
.584 .709 .602 .746 .555
.477 .227 .291 -.086 .086
-.070 -.117 .181 .049 -.062
-.047 -.019 -.010 -.064 -.004
-.207 -.038 -.229 -.085 .417
.236 .214
.160 .278
.078 .493
.657 .069
.025 -.087
.482 .757 .028
.364 -.003 .595
-.032 -.092 .013
.108 .005 -.059
-.043 .489 -.041
.720 .218 .503 .292
-.195 .571 .384 .548
.092 .058 -.016 .001
-.102 .019 .288 .142
.308 -.101 .082 -.049
.505 .629 .042 .622 .166
-.038 .041 .542 .395 .524
.214 .305 -.062 -.043 -.030
.493 -.050 -.088 -.048 -.100
.127 -.121 .106 -.178 .169
.749
-.068
-.004
-.152
.253
.444
.443
-.114
.119
-.076
.694 .465 .664 .598 .030
.064 .120 -.019 .127 .754
.019 .437 .260 -.006 .022
-.072 -.156 .265 .066 -.049 -.104 .005 -.095 -.090 .025 (Continued)
161
Continued 48. Glares at others 49. Encourages others to gang up on someone (physically OR verbally) 50. Hits others with objects 51. Is often grouchy 52. Doesn't care about others' feelings 53. Verbally threatens others with physical harm 54. Hits or shoves others forcefully .258 55. Makes unwanted sexual comments to others 56. Does not seem truly sorry after hurting someone 57. Gets revenge after some time has passed and the other person is not on guard 58. Tries not to get caught while doing harmful things to others
.525
.319
.085
.122
-.107
.348 .312 .628 .420 .607 .723 .452
-.055 .638 .212 .385 .158 -.037 -.056
.411 .055 -.074 .083 -.056 .114 .069
.130 .004 -.098 .082 .072 -.080 .198
.076 -.047 -.147 -.023 .180
.417
.166
.114
.190
-.059
.303
.442
.396
.147
-.046
.381
.233
.650
.132
.011
162
.206
APPENDIX G FIVE FACTOR SOLUTION USING VARIMAX ROTATION
163
5 Factor Solution Using Varimax Rotation: 1. Sneers, "makes faces" at others 2. Is "sneaky;" does things "on the sly" 3. Attempts to touch others in inappropriate, sexually suggestive manner 4. Pinches others 5. Is resentful over seemingly minor issues 6. Breaks others' things 7. Is quick to anger ("hot-headed") 8. Takes others' things by force 9. Broods, pouts, or is sullen 10. Calls others insulting names in their absence 11. Shoves or pushes others 12. Crowds others 13. Intimidates others 14. Says, "I hate you" or makes other hurtful statements 15. Bites others 16. Insults others to their faces 17. Throws objects at others 18. Reacts suddenly or impulsively to minor provocations 19. Shouts at others in anger 21. Gets mad when caught behaving badly .274 22. Is overly argumentative 23. Uses profanity to shock or offend others .162 25. Tickles or otherwise physically teases others, even after being asked to stop 26. Steals from others when they aren't looking 27. Reacts to insults or teasing by lashing out physically 28. Calls others insulting names to their faces 30. Head-butts others 31. Makes insulting comments about others behind their backs 32. Breaks own belongings 33. Makes threatening gestures 34. Charges at others 35. Verbally teases others, even after being asked to stop 36. If caught, denies having behaved badly 37. Pulls others' hair 38. When angry, is slow to cool off 39. Spits at others 40. Says "I hate [someone not present]" in absence of the person concerned 42. Lashes out at people who are in his/her space 43. Takes offense at things others say or do when s/he shouldn't 44. Starts trouble by baiting others 45. If caught, makes excuses for bad behavior 46. Is hostile, "has a chip on his/her shoulder" 47. Scratches others with fingernails
.169 .223
.407 .143
.253 .222
.122 .612
.292 .143
.119 .508 .150 .536 .409 .352 .075 -.011 .607 .339 .343
.044 .110 .177 .060 .246 .102 .235 .751 .103 -.016 .272
.091 -.064 .691 .151 .647 .049 .607 .261 .183 .165 .299
.181 .129 .165 .260 .073 .254 .099 .137 .039 .081 .226
.531 .080 .151 .157 .095 .461 .078 .045 .336 .480 .399
.124 .636 .096 .653
.731 -.054 .809 .116
.356 .092 .209 .211
.053 .018 .094 .217
.091 .061 .152 .139
.475 .254 .167 -.062 .665
.180 .422 .560 .435 .127
.585 .556 .316 .585 .052
.074 .042 .125 .199 .074
.114 .130
.101 .241
.105 -.024
.028 .129
.137 .539
.692 .112
.363 .092 .587
.235 .865 -.088
.367 .222 .074
.084 .060 .043
.225 .104 .007
-.120 .554 .381 .531
.717 -.014 .324 .075
.301 .227 .265 .231
.225 .127 .106 .086
-.022 .115 .399 .244
-.060 .040 .560 .398 .556
.374 .284 .046 .233 .173
.151 .487 .001 .586 .044
.314 .431 -.031 .104 .023
.545 .053 -.030 .109 -.029
.009
.694
.385
.143
-.046
.438
.184
.376
.000
.244
.077 .098 -.013 .137 .754
.333 .269 .332 .306 -.058
.606 .173 .505 .490 .041
.166 .061 .529 .323 .391 .055 .124 .124 .059 -.012 (Continued)
164
.057
Continued 48. Glares at others 49. Encourages others to gang up on someone (physically OR verbally) 50. Hits others with objects 51. Is often grouchy 52. Doesn't care about others' feelings 53. Verbally threatens others with physical harm 54. Hits or shoves others forcefully 55. Makes unwanted sexual comments to others 56. Does not seem truly sorry after hurting someone 57. Gets revenge after some time has passed and the other person is not on guard 58. Tries not to get caught while doing harmful things to others
.306
.206
.417
.207
.241
-.063 .631 .228 .381 .198 .703 -.035
.233 .079 .291 .193 .511 .014 .435
.106 .262 .577 .297 .310 .244 .135
.471 .146 .067 .184 .075 .049 .157
.155 .119 .043 .185 .178 .232 .247
.148
.184
.288
.207
.270
.408
.050
.171
.469
.221
.202
.140
.143
.724
.183
165
APPENDIX H RATER RECRUITMENT SCRIPT AND CONSENT FORM
166
RATER RECRUITMENT SCRIPT Greetings, I am asking for your help in evaluating a new instrument for gauging aggressive behavior in people with developmental disabilities. Your employers have agreed in writing to allow me to collect data at this site. Your participation would be purely voluntary and would not be used in any evaluation of your job performance. In fact, whether you agree to help or not will not be made known to your employers. You may stop your participation in the study for any reason, and at any time, without question. If you agree to help, you would be asked to do the following things: 1) In no particular order, list those clients under your care and/or supervision with whom you have been familiar for at least six months. Then number those on the list from top to bottom. From that point on, the number beside each client's name will serve as their identification code. This code will be used in any all communications, forms, etc. with me regarding this study. There will be some clients for whom I have guardian consent to do more extensive research. I will tell you if any of these people are on your list. If so, those individuals will be identified by name. 2) I will give you rating forms to complete for each eligible client. I estimate that you will need about 15 to 20 minutes per client to complete the forms. 3) If feasible, I will come to you to collect the completed forms. Otherwise, I will provide you with self-addressed stamped envelopes for the purpose of returning the forms to me at my address at the Nisonger Center on the OSU campus. Thanks for your consideration. I hope you'll agree to take part in what should be a very worthwhile study!
Scott Matlock, M.A.
167
CONSENT FOR PARTICIPATION IN SOCIAL AND BEHAVIORAL RESEARCH Protocol title: Development of an Interpersonal Aggression Scale for People with Intellectual Disability/ Developmental Disabilities Protocol number: 2005B0143 Principal Investigator: Dr. Michael Aman I consent to my participation in research being conducted by Dr. Michael Aman of The Ohio State University and his assistants and associates. The investigator(s) has explained the purpose of the study, the procedures that will be followed, and the amount of time it will take. I understand the possible benefits, if any, of my participation. I know that I can choose not to participate without penalty to me. If I agree to participate, I can withdraw from the study at any time, and there will be no penalty. I have had a chance to ask questions and to obtain answers to my questions. I can contact the investigators at (614) 688-5617. If I have questions about my rights as a research participant, I can call the Office of Research Risks Protection at (614) 688-4792. I have read this form or I have had it read to me. I sign it freely and voluntarily. A copy has been given to me.
Print the name of the participant: ______________________________________________________
Date: _________________________________
Signed: ___________________________________ (Participant)
Signed: ________________________________ (Principal Investigator or his/her authorized representative)
Signed: ___________________________________ (Person authorized to consent for participant, if required)
Witness: _______________________________ (When required)
HS-027 (Rev. 05/01)
(To be used only in connection with social and behavioral research.)
168