I 1111

.!OLRvn OF OFFEVDER REflJBIUTATlOV

Journal of Offender Rehabilitation, Vol. 47(1/2), 2008. Pp. 101-120. Available online at http://jor.haworthpress.com

Stewart, L. M. (2007). Perceived stress. self-efficacy. and depression, hopelessness. and suicidal ideation in a group 0/ incarcerated women, Unpublished doctoral dissertation. Fielding Graduate University, Santa Barbara.

doi: 10,1080/10509670801940680

AUTHORS' NOTES Jeremy F. Mills is affiliated with the Psychology Department. Bath Institution. Bath, Ontario, and the Department of Psychology. Carleton University, Ottawa. Canada. Daryl G. Kroner is affiliated with the Psychology Department. Pittsburgh Institution. Kingston. Canada. Address correspondence {(J Jeremy F. Mills. Psychology Department. Bath Institution. 5775 Bath Road. PO Box 1500. Bath, Ontario KOI-! IGO, Canada (E-mail MillsJf(acsc-scc.gc.ca). The authors gratefully acknowledge Robert Morgan and Ron Holden for their comments on an earlier version of this manuscnpt. The views expressed 111 this paper arc those of the authors and do not necessarily reflect the view, of the Correctional Service or Canada.

(9 2008 by The Haworth Press, Inc. All rights reserved. doL 10.1080/10509670801940920

Response Distortion in Applications of the Minnesota Multiphasic Personality Inventory-2 (MMPI -2) in Offender Rehabilitation EDWARD HELMES

ABSTRACT The MMPI-2 continues to be widely used in many areas of professional forensic psychology. including the evaluation of criminal offenders for rehabilitation purposes. While many possible applications of the MMPI exist, not all are well-supported by strong empirical evi­ dence. The origins of the scale among psychiatric populations suggest some utility in detecting other forms of psychopathology in use with offenders, In many applications with offenders, the utility of the validity scales and indicators of the MMPI/MMPI-2 in the detection of invalid or questionable profiles may rival the interpretation of the clinical scales in value. This commentary summarizes issues and research upon the nat­ ure and utility of the MMPI/MMPI-2 scales and indices that have been developed to detect various forms of response distortion that can affect the validity of the instrument. KEYWORDS Faking, malingering, MMPI-2, response distortion

The Minnesota Multiphasic Personality Inventory-2 (MMPI-2; Butcher, Dahlstrom, Tellegen, & Kaemmer, 1989) is one of the most widely used psychological assessment instrument in forensic psy­ chological assessment in the U. S. (Lees-Haley, 1992) and Australia (Martin, Allan, & Allan, 2001). Originally developed to assess

101

JOURNAL OF OFFENDER REHABIUTA7JON

psychopathology and to perform differential diagnosis (Norman, 1972), the test has been applied to many different populations and purposes over the decades in which it has been used (Butcher, 2(06). Originally developed in the tradition of 1930s and 1940s U .S. mid-Western "dust bowl empiricism, '" items were selected for the scale on the basis of their ability to discriminate diagnosed groups of psychiatric patients from normal samples (usually visitors to the Minnesota hospitals). Items were retained whether or not a clear link between the diagnostic group and the manifest item content was evi­ dent, resulting in several items that remained of questionable utility. Many such items were dropped during the process of revision, but others remained in the form of "subtle" items, which likely were selected because of chance factors in the multiple comparisons that were inevitable in such a process (Jackson, 1971). Scales were developed to assess major forms of psychopathology, including depression. schizophrenia, "psychopathic deviation," and hypochon­ driasis. and also other characteristics, such as stereotypic sexual orientation and introversion. Use of the test spread rapidly. despite early reservations about its actual utility in practice for many of the purposes for which it had been intended (see review by Helmes & Reddon (1993) for a summary of some of the early literature and issues related to the adequacy of the revision of the MMPI to the MMPI-2). One of the strengths of the MMPI-2 that has been demonstrated over time is a range of scales and other measures that can be used to detect response distortion of several types. This feature can clearly increase its utility in the context of offender rehabilitation where many offenders clearly have external incentives to present themselves positively as reformed and/ or remorseful. There may be some, how­ ever. who wish to present a picture of greater disturbance than is actually present for other reasons. A psychologist who is requested to evaluate offenders for parole or early discharge for special treat­ ment considerations needs to be aware of the likelihood of response distortion in its various forms and to be familiar with methods designed to detect it (Berry, 1(95). Here I review the major forms of response distortion and how they arc assessed with the MMPI-2. There is an extensive literature on response distortion in general in personality assessment using the self-report format that is not dealt with here, and only some of the

Edll'([rd Helilles

103

literature on the evaluation of distorted responding on the MMPI-2 can be covered here.

RESPONSE DISTOR TION From its original development the MMPI/MMPT-2 has incorpor­ ated measures to detect distorted responding. It was early recognized that individuals could both distort their responses in order to appear more disturbed than was actually the case (which scale F was intended to detect) and to appear as less disturbed than was actually the case (which scales Land K were intended to detect). The current range of indices to detect aberrant modes of responding is much more extensive. Nichols and Greene ( 1997) provide a useful analysis of the various possible types of response distortion from inconsistency in responding through to overt lying and how they are reflected in the MMPI-2 validity indices. Bagby, Marshall, Bury, Bacchiochi, and Miller (2006) provide a more recent update on the literature on response distortion on the MMPI-2 and conclude that "the scales and indexes on the MMPI-2 arc mostly effective in discriminating fake-good and fake-bad responding from honest responding" (p. 69). It must, however, be remembered that the MMPI-2 validity indicators were developed to assess response distortion in the context of psychiatric services. Therefore the nature of the manipulated responses was to appear to have a greater or lesser extent of psycho­ logical problems in a context in which if not treatment, at least a sym­ pathetic response was hoped (or. in some cases, feared) to be forthcoming. It is much less clear how well they operate in other con­ texts. Most simulation studies use instructions that are framed within a clearly clinical context, with the exception of some of the faking good studies that are modeled upon employment applicant scenarios. Readers should be aware that much of this literature uses the word "faking" to cover any form of manipulated responding. Its use with samples who are instructed to distort their responses is accurate, but its application to individuals undergoing psychological assessment is much less appropriate because of its connotations of intent. There are many factors that can lead to distorted responding that require consideration in the interpretation of test results that do not involve deliberate intentions to deceive. Such factors are dealt with below.

/1)4

JOURNAL OF OFFto'NDLR REHAB/UTA nUN

Since the development of the L F, and K scales in the 1940s, it has become clear that the K scale. which was intended to assess a defens­ ive style of responding and to correct other scales for presumed under-reporting, was not terribly successful as a correction factor. As long ago as 197/1., there were opinions expressed that supported dropping its use (Butcher & Tellegen, 1978). Despite this. the correc­ tion to scales 4, 7. 8, and 9 continues to be printed upon the published profile sheets. Some years later, there was still no body of research in hand, and Butcher. Graham, and Ben-Porath (1995) stated that with regard to scales that were corrected by K or not "it is premature to conclude that one kind of score is superior to the other" (p. 324). Greene, Gwin, and Staal (1997) expressed it another way when they commented "It is remarkable that in over 50 years there has been virtually no research on the K-correction process, and that this correction procedure was not examined nor validated when it was extended to the MMPI-2. Such research clearly is needed" (p. 26). Since then, Archer. Fontaine, and McCrae (1998) found little or no difference between the correlations of uncorrected and K-corrected scores with psychiatric symptoms. Barthlow, Graham, Ben-Porath, Tellegen, and McNulty (2002) found that only two of 88 correlations between MMPI scales and relevant outpatient demographic and symptom data were higher with a K-correction than without it. While the empirical research on the K-correction to date remains limited. what does exist suggests that the correction fac­ tors used may not generalize to samples other than psychiatric ones and that even in that population. the role of the K scale as a correc­ tion factor has little empirical support. This is an important consider­ ation in legal and forensic applications, as the application of the K-correction could lead to changes in the elevation of a scale from values that would be regarded as not significant clinically to a range that would be so regarded. Without solid empirical support for such changes in interpretation, a practitioner could find their interpretation subject to challenge and difficult to defend. The relative absence of such essential research over such an extended period of time might be regarded by the cynical as one example of reasons not to usc the MMPl-2. One of the stated reasons for the retention of the correction factor is the desire to ensure the applicability of all the research done that used the K-correction. Barthlow et al. (2002) rightly point out that this argument does not

Edmml HcllIll's

/IJ5

bear scrutiny. The continued use of the K-correction thus remains difficult to defend in the absence of supportive empirical research. Given the many possible forms of response distortion, some orga­ nizing principles arc useful. Nichols. Greene. and Schrnolck (1989) classified response distortion as either content nonresponsiveness (CNR), in which respondents do not respond in terms of the overt item content, and content-related faking (CRF), in which respon­ dents do respond to the item content, but try to distort their image. whether positively or negatively. Given the extent of the MMPl/ M MPI-2 validity indices, the recent research will be covered under studies of content-related faking first, beginning with the under­ reporting of distress. or faking good, followed by faking bad or over­ reporting of distress (exaggeration). There is less rna terial on conten t nonresponsivcness, such as inconsistency or random responding, and a review of that literature will follow.

UNDER-REPORTING Individuals may have a variety of reasons for presenting them­ selves as well-adjusted when significant psychological problems in reality do exist. Examples include those who are trying to conceal their inner problems in order to obtain a desired benefit, such as an early release. Two scales were developed to assess such tendencies in the original context of the evaluation of psychiatric patients, the K scale discussed above and the Lie (L) scale. The IS items of the Lie scale were based on the work of Hartshorne and May (1930) to assess whether people were trying to present themselves in an improbably positive or negative way (Hathaway & McKinley, 1940). These items remain on the MMPI-2, and represent a less sophisticated method of presenting a false image than was intended with the K scale. One feature of research on the validity indices of the MMPI/ MMPI-2 is that there is little reference to related work in the broader literature. Thus the work of Paulhus (1984; Paulhus & Reid, 1991) on the distinction between self-deception and impression management is rarely noted in the literature on the validity indices of the MMPI/MMPI-2. Paulhus's work on the construct of social desir­ ability has provided a solid theoretical base for the interpretation of much of the literature in this area. In some ways, the lad of a

Inn

.I()[!I 75 has reasonable sensitivity and specificity at levels of yea-saying of 20°/;, and higher. They also note that TRIN appears to detect yea-saying better than nay-saying. Once again, more research using different samples and approaches to acquiescence would refine estimates of the best interpretive practices. Such basic research into applications of the MMPI/MMPl-2 continues to be essential. The history of the MMPI has been one of continued extensions of its use into applications beyond those for which it was originally designed. Efforts are then made by some users to defend and promote the case for these new applications, even though the validity of the interpretations drawn in those applications might be questionable and the relevant resulting literature quite con­ tradictory and inconclusive as to the actual utility of the MMPl-2. Such supportive efforts often seem to make rather tenuous arguments for the continuation of the use of the MMPI-2 in these new areas that lack a strong evidence base, as illustrated in several chapters in Butcher (2006) and other applications that were introduced in recent times in the literature. Forensic applications of the MMPI-2 fall into this area. Psychometrically, the MMPI-2 has satisfactory properties for use in many forensic applications (Weiner, 1995). However, in the area of risk assessment. most of the research has been on the O-H scale or code types involving Scale 4 Pd (Heilbrun & Heilbrun, Jr., 1995). These authors make evident the complex issues in risk asses­ sment and recommend theory-based approaches that go beyond the assessment of "antisociaiity' and low intellectual functioning that

Edward

1/('/11/,'.1"

115

can be combined into a "dangerousness index" that typify the hulk of MMPl-2 research in the area of risk assessment of dangerous offen­ ders. Much of the research cited in this chapter is quite old, and it is evident from more recent research, for example, Allan, Dawson. and Allan (2006), that much risk assessment of offenders is now based upon specialized instruments and not broad measures of psychopath­ ology such as the MMPI-2. While the evidence from forensic populations is limited, the bal­ ance of research on the validity indicators of the MMPI-2 suggests that they are adequately sensitive to attempts hoth to over- and under­ report levels of psychopathology. This utility is counterbalanced by the relatively high reading level (Grade 8; Butcher et al., 1989) and length of the test, as well as limited utility of many of the scales in forensic applications. This should not be surprising, as the test was not designed for use with offenders and much of the item content greatly predates the introduction of the concept of rehabilitation into the criminal justice system. Thus while the validity scales of the MMPI-2 can be very useful to the forensic psychologist, the remainder of the instrument may be of less utility.

REFERENCES Allan. A., Dawson, D.. & Allan, M. M. (2006). Prediction of the risk of male sexual offending in Australia. Australian Journal of Psychology, 41,60-68. Arbisi, P. A., & Ben-Porath, Y. S. (]995). An MMPI-2 infrequent response scale for use with psychopathological populations: The Infrequency­ Psychopathology Scale, F(p). Psychological Assessment: A Journal of Consulting and Clinical Psychology, 7. 424-431. Arbisi, P. A., & Ben-Porath. Y. S. (1997). Characteristics of the MMPI-2 Ftp) scale as a function of diagnosis in an inpatient sample of veterans. Psychological Assessment, I), ]02-105. Arbisi, P. A.. & Ben-Porath. Y. S. (1998). The ability of Minnesota Multi­ phasic Personality Inventory-2 validity scales to detect fake-bad responses in psychiatric inpatients. Psychological Assessment, 10,221-228. Arbisi, P. A.. & Butcher, J. N. (2004). Failure of the FBS to predict malingering of somatic symptoms: Response to critiq ues by Greve and Bianchini and Lees Haley and Fox. Archives of' Clinical Neuropsl'c!/(Ilogv. 11), 341 345.

!lr,

.lUI R\ If UF OFFl:'Nf)/:R

nrnieu.ncno«

Arbisi, P, A" & Seirne, R, .I, (~()()6), Use of the MMP[-~ n medical settings, l n .J. N, Butcher (Ed,), i'vIMPI-2: A practitianer's guide (pp, 273 ~99), Washington. DC: American Psychological Association, Archer. R, P,. Fontaine. L & Met.rae. R. R. (199~). Effects of two MMP[-~ validity scales on basic scale relations to external criteria. Journal 0/ Persoualit v A ,1,11',1,111 WI/I , 7{), S7 I O~. Bacchiochi . .I. R., & Bagby, R. M. (~006). Development and validation of the Malingering Discriminant Function Index for the MMP[-~. Journal 0/ Pcrsonalitv Assessment, 87, 5161. Baer, R. A" & Miller, .I. (2()02). Underreporting of psychopathology on the MMP[-2: A meta-analytic review. Psvchological AS.I·eSSII/Cl/t , 14, 16 26, Baer. R. A" Wetter. M. W., & Berry, D. T. R. (1992). Detection of under­ reporting of psychopathology on the MMP[: A meta-analysis. Clinical PS.\'c!IO!og,1' RcviclI, t : 509 525. Bagby, R. M., Marshall. M. B.. Bury, A. S.. Bacchiochi, .I. R., & Miller, L S. (~()06). Assessing underreporting and overreporting response styles on the MMP[-2. In J. N. Butcher (Ed.), AIMPI-2: A practitioncr» guidc (pp. 39 69). Washington, DC: American Psychological Association. Barthlow, D. L.. Graham, J. R .. Ben-Porath, Y. S" Tellegcn. A .. & McNulty. J. L. (2002). The appropriateness of the MMP[-2 K correction. AS,I1'SSI//('l/t. o. 219229, Berry, D. T. R, (1995), Detecting distortion in forensic evaluations with the MMP[-2, ln Y, S. Ben-Perth. .I. R. Graham, G. C N. Hall. R. D. Hirschman. & M. S. Zaragoza (Eds.), Forellsic ap plicat ions 0/ the MMPI-2 (pr. ~2 10~) Thousand Oaks. CAl Sage. Berry, D. T. R., BaCT. R. A., & Harris, M . .I. (199[). Detection of malinger­ ing on the MMPI: A meta-analytic review. Clil/icu! Psvrhologv Review, II, 585- 598. Berry. D. T., Wetter. M. W., Baer, R. A., & Widiger, T. A. (1991). Detection of random responding on the MMP[-~: Utility of F, back F, and VRIN scales. Psychological Assessment, 3, 418-423. Bury, A. S" & Bagby, R. M. (200~). The detection of feigned uncoached and coached posttraumatic stress disorder with the MMPI-~ in a sample of workplace accident victims. Psychological Assessment. 14, 47~-484. Butcher, J. N. (Ed.). (2006). MMPI-2: A practitioner's guide. Washington, DC: American Psychological Association. Butcher, J. N .. Arbisi, P. A., Atlis, M. M., & McNulty, J. L. (2003). The construct validity of the Lees-Haley Fake Bad Scale: Does this measure somatic malingering and feigned emotional distress? A rchives 0/ Clinical NCl/rops,1'c!IO!ogl'. 18, 473--485. Butcher.J. N., Dahlstrom. W. G., Graham.J. R" Tellegen, A., & Kacmmer, B. (1989). MUI/I/u! for adininistration and scoring MAIPI-2: Minnesota

Fdll'urd 111'1/111'.1'

1/7

Multiphasic l'ersonalit v lnvcnt ory-Z, Minneapolis, M N: I i'll\ crsity of Minnesota Press. Butcher, .I. N" Graham, .I. R., & Ben-Poraih. Y. S. (1995). Methodological problems and issues in MMP[, MMP[-2. and MMP[-A research. VITc!IO­ iogica] Assessment. A .I01/1'I/1I! of Consulting autl Clinira! PIT C!/O !og l' , 7, 320 3[9. Butcher, .I. N. & Tellegen, A. (197k). Common methodological problems in MMP[ research. Journal o] Consult ing lind Clinical PS.I'c1IO!O~.\'. 46. 620-628. Butcher, .I. N., & Williams, C. L. (200()). Essel/tia!s 0/ AI M rt.: and M M P 1­ A interpretation (2nd ed.). Minneapolis: University of Minnesota Press. Charter, R. A., & Lopez. M. N. (200~). MMPI-2: Confidence intervals for random responding to the F., F Back. and VR[N scales. Journal of Clinical Psvcholog», 59, 985 990. Clark, M. E.. Gir onda, R . .I.. & Young, R. W. (2003). Detection of back random responding: Effectiveness of MM P[-2 and Personality Assessment Inventory validity indices. Psychological Assessment, 15. 223 234. Cohen, .I. (1992). A power primer. American Psrl'!lO!ogist, I!}, 155 159. Crawford. E. F., Greene, R. L., Dupart, T. M., Bongar, 8.. & Childs, H. (2006). MMP[-2 assessment of malingered emotional distress related to a workplace injury: A mixed group validation. Journal o] Pcrsonalit v Assessmcnr. 80. 217 ~21. Fox. D. D .. Gerson, A .. & Lees-Haley, P. R. (1995). Interrelationship of MMP[-2 validity scales in personal injury claims. Journal 0/ Clinical Psych 0 !OiJ,l', 51. 42-47. Gallen, R. T., & Berry, D. T. R. ([ 996). Detection of random responding in MMP[-2 protocols. Assessment, 3. 171-178. Gallen, R. T.. & Berry, D. T. R. ([ 997). Partially random MMP[-2 protocols: When arc they interpretable'? Assessment, 4, 61 68. Gough, H. G. (1950). The F minus K dissimulation index for the MMPI. Journal of Consulting Psychology, 14, 408-41~. Gough, H. G. (1954). Some common misperceptions about neuroticism. JOUri/O! 0/ Consulting Psvchoiogy, 18. 287·292. Greene, R. L. (2000). The M M PI-2: An interpretive manual (2nd cd.). Needham Heights. MA: Allyn & Bacon. Greene, R. L.. Gwin, R., & Staal, M. (1997). Current status of MMP[-2 research: A methodologic overview. Journal of Personality Assessment, Special Issue: Personality assessment instruments: Current status andfuture directions, 68, 20-36. Greve, K. W .. & Bianchini, K. J. (2004). Response to Butcher el al., The construct validity of the Lees-Haley Fake-Bad Scale. Archives o] Clinical Neuropsvcholovv, N, 337 ·339.

t

t:

JO( RSAL OF

otn.snt:u

RU!.lBIIJ7A nON

l landcl, R. W .. Arnau, R. C. Archer. R. P.. & Dandy, K. L. (2006). An Evaluation of the MMPI-2 and MMI'I-A True Response Inconsistency (TRIN) Scales . . isscssnient, 13. 9X 106. l lurtshornc. II. & May. H. A. (1930). SIIU!ics 1/1 tlu: nature of character. New York: MacMillan. Hathaway. S. R .. & McKinley. J. C (1940). A multiphasic personality schedule (Minnesota): I. Construction of the schedule. Journal of PITc!IO!lIgr. !O. 249 254. Heilbrun. K. & Heilbrun. A. B. Jr. (199.51. Risk assessment with the MMPI-2 in forensic evaluations. In Y. S. Ben-Perth . .I. R. Graham. G. C N. Hall, R. D. Hirschman. & M. S. Zaragoza (Eds.), Forellsic applicatunts IIj the /'vlMPl-:! (pp. 160 In). Thousand Oaks. CA: Sage. Helmes. E.. & Redden . .I. R. (1993). A perspective on developments in asses­ sing psychopathology: A critical review of the MMPI and MMPI-2. Psychological Bulletin. !13. 4.53471. Iverson. G. L.. & Barton. E. (1999). Iruerscorer reliability of the MMPI-2: Should TRIN and VRIN be computer scored'? Journal 0/ Clinico] Psvchologv; 55. 65-69. Jackson. D, N. (1971). The dynamics of structured personality tests: 1971. Psvchological Review. 78.229 ?.4X. Koss. M. P.. & Burchcr.T. N. (1973). A comparison of psychiatric patients' self-report with other sources of clinical information. JIII/rl/i1! 0/ Research in Pcrsonalit v. 7. 225 236. Koss. M. P.. Butcher. .J. N.. & Hoffman. N. G. (1976). The MMPI critical items: How well do they work? Journal 0/ Consulting and Clinical PITC!tO!og)'. 44. 921 928. Lachar. D.. & Wrobel, T. A. (1979). Validating clinicians' hunches: Construction of a new MMPI critical item set. Journal o] Consulting and Clinical Psychology, 47. 277 284. Larrabee. G. J. (2003a). Detection of symptom exaggeration with the MMPI-2 in litigants with malingered neurocognitive dysfunction. Clinical Neuropsychologist, 17. 54 68. Larrabee. G. J. (20mb). Exaggerated MMPI-2 symptom report in personal injury litigants with malingered neuroeognitive deficit. A rchives 0/ Clinical Neuropsvchology, !8. 673686. Lees-Haley. P. R. (1992). Psychodiagnostic test usage by forensic psycholo­ gists. American Journal 0/ Forensic Psvchologv, 10, 1-25. Lees-Haley. P. R .. English. L. T.. & Glenn. W. J. (1991 ). A Fake Bad Seale on the MMPI-2 for personal injury claimants. Psvcliological Reports, 08. 203210. Lees-Haley, P. R.. & Fox. D. D. (2004). Commentary on Butcher. Arbisi. Atlis, and McNuity (2003) on the Fake Bad Scale. Archives 0/ Clinical Neuropsvchologv, /9. 333-336.

El!lmrd !!,.!IJJ