Linking family functioning to dating relationship quality via novelty-seeking and harm-avoidance personality pathways Judith L. Fischer & Jacki Fitzpatrick Texas Tech University

H. Harrington Cleveland Penn State University

ABSTRACT

Organized by the systems theory concepts of equifinality and interdependence (Broderick, 1993) and Huston’s (2000) social ecology model, this study examined two personality-based pathways to dating relationship quality among college students. The first pathway extends from family dysfunction to dating relationship quality via novelty seeking and excessive drinking. The second pathway extends from family dysfunction to relationship quality via harm avoidance and interpersonal competence. Male (n = 64) and female (n = 105) undergraduates completed questionnaires. The findings strongly supported the first hypothesis that was derived from Huston’s social ecology model and the systems principle of equifinality. There were significant associations between adjacent variables in the path model and the presence of both pathways in the tested model provided a good fit to the data. The second hypothesis, based on the systems principle of This article was supported, in part, by Texas Tech University’s Multidisciplinary Seed Grant Program awarded to Judith Fischer, Jacki Fitzpatrick and John Morrow. We thank Bobbi Miller, Paula Burnett, Cat Pause, Chance Ates, Tiffany Gour, Natasha Johnson, Anindida Das, Brittney Schrick, Mitsue Uchida and Jackie Wiersma for their help with this project. A poster version of this article was presented at the National Council on Family Relations Annual Conference, Minneapolis, MN, November, 2006. All correspondence regarding this article should be addressed to Judith L. Fischer, Department of Human Development and Family Studies, Texas Tech University, Lubbock, TX 79409, USA [e-mail: [email protected]]. Journal of Social and Personal Relationships Copyright © 2007 SAGE Publications (www.sagepublications.com), Vol. 24(4): 575–590. DOI: 10.1177/0265407507079257

576

Journal of Social and Personal Relationships 24(4)

interdependence, was not supported when cross-pathway links were included in the tested model. Implications for theory and research are discussed. KEY WORDS: alcohol • dating • drinking • family dysfunction • harm avoidance • interpersonal competence • novelty seeking • relationship quality

College students occupy a life course period that creates a potentially unique nexus for familial, relational and alcohol factors to affect romantic relationships. They date for a variety of reasons and with a varying range of outcomes. College students remain dependent in multiple ways (financially, psychologically) that make it difficult to disengage from their family of origin (e.g., Schultheiss & Blustein, 1994). Without parents’ daily supervision, students have the freedom to explore a variety of experiences, including romantic ones (e.g., Le & Agnew, 2001). Alcohol use also peaks during college years (Bachman, Wadsworth, O’Malley, Johnston, & Schulenberg, 1997), often bringing negative consequences (Wechsler, Lee, Kuo, & Lee, 2000). Many factors that contribute to romantic quality have been identified in the research. These include individual differences (e.g., Chotai, Jonasson, Hagglof, & Adolfsson, 2005), family of origin experiences (Donnellan, Larsen-Rife, & Conger, 2005), and social networks (e.g., Sprecher & Felmlee, 2000). However, identifying individual variables does not address how they function in relation to each other to contribute to romantic quality. It is important to understand the unique processes through which the confluence of factors (romance, family, alcohol) intersects. Guided by systems theory and Huston’s (2000) social ecology model, this study organized and tested ways in which family of origin dysfunction, personality (novelty seeking, harm avoidance), and behaviors (excessive drinking, interpersonal competence) are related to romantic quality among college daters (i.e., dating relationship quality). Theoretical foundation Systems theory is a key theoretical framework in marital studies (White & Klein, 2002). Both married and nonmarried romantic relationships emerge from complex patterns of interaction, so systems principles are equally applicable to marital and nonmarital romantic research. Through interactions with their partners, individuals develop systematic ways of relating that establish the quality (e.g., satisfaction, stability) of their unions. Two systems theory principles relied upon in this study are equifinality (multiple pathways to the same outcome) and interdependence (components of a system are interconnected). Equifinality reflects that individuals select from multiple options the actions that look promising to achieve their goals (Broderick, 1993). Young adults who emerge from similarly functioning families of origin connect to outcomes via multiple pathways (Davies &

Fischer et al.: Pathways to dating quality

577

Cicchetti, 2004). To address equifinality, this study specifies two pathways that link family of origin with dating relationship quality. Interdependence suggests associations among relational dynamics as well as associations among relational and intrapersonal factors. Indeed, Broderick (1993) noted that both individual attributes and interactional factors should be considered in systemic approaches. To address interdependent processes within an actor, this study examines the degree to which the two specified pathways are linked to each other. Huston’s (2000) social ecology model provides an organizing framework that connects different systems levels extending from the macro-environment to the subjective events in individual’s relationships. Huston’s model proposes that more distal levels work directly and through more intermediate and proximate levels on evaluations of the close relationship. One result of these processes is that dating students experience diverse relationship qualities. The model proposed in this study includes a number of system levels. The macro-environment is represented by the distal variable of family of origin dysfunction (deficits in family communication and need fulfillment; Epstein, Baldwin, & Bishop, 1983). Personality is represented by the intermediate variables of novelty seeking (avoiding monotony and seeking the unknown) and harm avoidance (aversion to difficult situations; Cloninger, 1987). The proximal variables are excessive drinking (frequent and heavy consumption of alcohol; Wechsler et al., 2000) and interpersonal competence (prosocial communication behaviors; Buhrmester, Furman, Wittenberg, & Reis, 1988). The subjective evaluation is represented by dating relationship quality. Quality encompasses how one feels about the relationship (e.g., satisfied; Hendrick, 1988, and unambivalent; Walster, Berscheid, & Walster, 1973), future intentions about the relationship (commitment; Sternberg, 1988), and connection with the partner (relational interdependence; Cross, Bacon, & Morris, 2000). Family of origin as a common origin for divergent pathways to relationship quality In their families of origin, future relationship participants are exposed to interpersonal behaviors and norms of interactions that they will carry into their own relationships (e.g., Bradbury & Fincham, 1988). Such influences can be negative when the family of origin is dysfunctional. Previous research has established that aversive family of origin experiences have been associated with the intermediate, proximate and outcome variables of this research. Young adults with unmet family needs vary in their responses to their dysfunctional families. Some respond with novelty seeking (Ravaja & Keltikangas-Jarvinen, 2001) whereas others become more avoidant of interpersonal situations (Reti et al., 2002). Family dysfunction has also been related to the proximate variables of greater alcohol misuse (Mason & Windle, 2002) and interpersonal competence (Bryant & Conger, 2002;

578

Journal of Social and Personal Relationships 24(4)

Donnellan et al., 2005; Koesten, 2004). Family of origin experiences have been related to relationship satisfaction as well (Bryant & Conger, 2002; Kelley et al., 2005), possibly indicating an intergenerational transmission of difficulty in interpersonal relationships. Family dysfunction is expected to affect relationship satisfaction (Sabatelli & Bartle-Haring, 2003) through its association with intermediate personality and proximate behavior factors. Novelty-seeking and excessive drinking pathway In Cloninger’s (1987) system, novelty seeking and harm avoidance are described as temperaments, part of the emotional core of personality. Temperaments emerge early in life and are stable across time (Cloninger, Svrakic, & Przybeck, 2006). Novelty seeking leads to frequent exploratory activity, impulsive decision making, and attraction to unfamiliar stimuli (Cloninger, 1987). By encouraging exploration and risk taking, novelty seeking can express itself in both general drinking (Baumrind, 1987) and excessive drinking (e.g., Ball, 2004). Although college students report that drinking is associated with excitement, fun, and sense of belonging (Shim & Maggs, 2005), these potential benefits are often offset by negative consequences for self, others, and romantic relationships (Wechsler et al., 2000). For example, excessive drinking is associated with lower communication quality (Fischer et al., 2005). Thus, the first pathway of the model is family dysfunction (distal variable) → novelty seeking (intermediate variable) → excessive drinking (proximate variable) → dating relationship quality. This novelty-seeking pathway proposes that effects from family of origin to relationship quality are either indirect or mediated. Harm-avoidance and interpersonal competence pathway The second pathway links harm avoidance to relationship quality via interpersonal competence. Harm avoidance reflects pessimism, sensitivity and aversion to difficult situations (Cloninger, 1987). Harm avoidance is correlated with less happiness (Stewart, Ebmeier, & Deary, 2005), more fear, and more unpleasant emotions (Puttonen, Ravaja, & Keltikangas-Jarvinen, 2005). Harm avoidance via amplifying reactivity can be adaptive when it reduces exposure to aversive stimuli, and thus may be adaptive within difficult family environments. However, harm avoidance in response to family dysfunction could leave the person protective toward the self by avoiding communications with others. Indeed, tendencies to be fearful, avoidant, and reactive can lead people to isolate themselves from the interactions necessary to develop interpersonal skills (Tse & Bond, 2005). Communication between partners is a crucial element of systems theory (Broderick, 1993). Interpersonal competence is reflected in communication behaviors such as self-disclosure (Bombar & Littig, 1996) and conflict resolution (Cramer, 2004). Competencies such as conflict-resolution skills can

Fischer et al.: Pathways to dating quality

579

positively affect romantic quality (Givertz & Segrin, 2005) but they can be problematic for individuals who are reactive (Wills, Sandy, Yaeger, & Sinar, 2001). Because interpersonally competent behaviors may engender risk, harm-avoidant people may be less likely to engage in these behaviors. For example, negative assertion can be used to address problems but may also increase the intensity of emotionally valenced interactions. Given these risks, harm-avoidant individuals would experience motivations to engage in fewer interpersonal competence behaviors. Therefore, the second pathway of the model is family dysfunction (distal variable) → harm avoidance (intermediate variable) → interpersonal competence (proximate variable) → relationship quality. Similar to the novelty-seeking pathway, the harmavoidance pathway proposes that effects from family of origin to relationship quality are either indirect or mediated. Additional considerations: The roles of sex and relationship length Prior research has identified sex and length differences in the predictor variables. For example, Gil (2005) reported that men scored higher than women on novelty seeking, whereas Nixon and Parsons (1989) found that men scored lower on harm avoidance. Men have more alcohol problems than women (Roberts & Leonard, 1998). Men and women differ on family of origin experiences, in that women report greater parental happiness (Weigel, Bennett, & Ballard-Reisch, 2003). Koesten (2004) detected sex differences in social competence: Women were more capable of providing support, whereas men were more capable of providing assertion skills with romantic partners. Relationship quality/satisfaction also differs by sex (Bornstein et al., 2004). Therefore, the role of participant sex is based on the central place of gender in social ecology models (Bronfenbrenner, 1989) and empirical findings. Relationship length is associated with romantic variables, such as stronger attachment to partner (Feeney, 2004). Length should be associated positively with relationship quality. Hypotheses • Hypothesis 1 predicts that the model (termed the Independence Model) with a novelty-seeking pathway and a harm-avoidance pathway provides a good fit to the data. Furthermore, reflecting the social ecology model principle that distal effects work through intermediate and proximate effects, there should be significant associations between adjacent variables in the pathways. Intermediate and proximate variables are expected to be significant mediators or indirect effects. • Hypothesis 2 (Interdependent vs. Independent Models). The addition of cross-pathway links in a good fitting Interdependence Model provides for a significantly better fit to the data than a good fitting Independent

580

Journal of Social and Personal Relationships 24(4)

Model. In the Interdependent Model, two links specify associations between the intermediate variable on one pathway and proximate variable on the other (novelty seeking → interpersonal competence (Chotai et al., 2005), harm avoidance → excessive drinking (Rose, 1998)). Two other interdependent links connect the intermediate variables (novelty seeking ← → harm avoidance) and proximate variables (excessive drinking ← → interpersonal competence) to each other. With the Independent Model fully nested within the Interdependent Model, comparing fit between the models provides a direct examination of the systems principle of interdependence. • Hypothesis 3a predicts that sex is related significantly to each of the variables in the study, such that men will report higher family dysfunction, higher novelty seeking, lower harm avoidance, higher excessive drinking, lower interpersonal competence, and lower relationship quality. Mediation/indirect effects of sex to relationship quality are expected via the intermediate/proximate variables in the model. • Hypothesis 3b predicts that length should be significantly and positively related to relationship quality. Method Participants Undergraduates who were presently involved in unmarried romantic relationships participated in the study (N = 169, 105 women and 64 men). Married individuals were not recruited because marriage is associated with downward shifts in alcohol use (Leonard & Roberts, 1996), thus, possibly confounding relationship status and alcohol use. Most respondents (82%) were Anglo/EuroAmerican. The average age was 20 years; 54% of respondents were younger than the legal drinking age. Twenty per cent identified their romantic status as casually dating, 69% identified it as steady dating, 4.1% were cohabiting, and 5.9% were engaged. Of the 201 students who completed the questionnaire, 32 students were excluded from the final sample primarily because they did not meet the recruitment criteria (18 reported on other relationship types, 6 did not meet age requirements, 8 had excessive missing data). Procedures Students were recruited from classes across a variety of majors. Recruiters indicated that participants should be 18–23 years of age, unmarried, live in town, and have a romantic partner who also lived in town and was no more than 5 years older or younger than the participant. Participants attended small group research sessions in college classrooms to complete a questionnaire packet. At the conclusion of the study, respondents were paid $25. Measures Measurement issues. In this study, there are control, distal, intermediate, proximate and outcome variables. In a nonclinical population skewed distributions would be expected, reflecting the fact that the majority experienced functional

Fischer et al.: Pathways to dating quality

581

families of origin, report their behaviors as more positive (e.g., fewer binge drinkers, more interpersonal competencies), and indicate that their relationships are satisfying. Accordingly, significantly skewed variables (tested by critical ratios) were transformed (Kline, 2005) via a log to the base 10 or a square root transformation to reduce skew. After transformation none of the variables were significantly skewed, excessive in kurtosis, nor were there outliers (Kline, 2005). Variables transformed by the log method were: Length, family dysfunction, 2-week binge, satisfaction, and commitment. Variables transformed by the square root method were: Disclosure, emotion support, negative assertion, and unambivalence. Original means and standard deviations, transformed means and standard deviations, and intercorrelations of the measures of the study are available on request. Cronbach’s alpha internal consistency reliability of scales ranged from .74 to .95. Higher scores reflect greater endorsement of the measured variables. Control variables. Gender (coded 1 = female [62%], 2 = male [38%]) and length were control variables. Length was used as a continuous variable, coded: 1 = 0–6 months (31%); 2 = 7–12 months (15%); 3 = 13–18 months (18%); 4 = 19–24 months (11%); 5 = 25–30 months (4%); 6 = 31–36 months (4%); 7 = 37–42 months (8%); 8 = 43–48 months (1%); 9 = more than 48 months (7%). Distal variable. Family dysfunction was assessed via the General Family Functioning (GFF) subscale of the McMaster Family Assessment Device. The family dysfunction 12-item subscale measures the extent to which the family of origin was unaccepting, inarticulate, and a place of bad feelings (Epstein et al., 1983). This subscale discriminates clinician-rated healthy from unhealthy families (Miller, Epstein, Bishop, & Keitner, 1985). Respondents indicated on a 5-point Likert scale (1 = ‘strongly disagree;’ 5 = ‘strongly agree’) the extent to which statements accurately reflected their family of origin (e.g., ‘We avoid discussing our fears and concerns’). Intermediate variables. There were two intermediate variables: Novelty seeking and harm avoidance. Novelty seeking was measured with a 13-item scale from the short form of the Tri-Personality Questionnaire (STPQ; Sher, Wood, Crews, & Vandiver, 1995). The short form was derived from Cloninger’s Tridimensional Personality Questionnaire (TPQ-4; Cloninger, 1987). Respondents indicated whether each item was true (0 = ‘false,’ 1 = ‘true’) about them (e.g., ‘I often do things based on how I feel at the moment without thinking about how they were done in the past’). Harm avoidance was a 22-item scale from the short form of the TPQ-4 (Sher et al., 1995). Respondents indicated whether each item was true (0 = ‘false,’ 1 = ‘true’) about them (‘Usually I am more worried than most people that something might go wrong in the future’). Proximate variables. There were two proximate variables: Excessive drinking, and interpersonal competence. Excessive drinking was a latent variable composed of three continuous indicators chosen to reflect consumption and frequency patterns indicative of excessive drinking. Consumption and frequency measures are commonly used in alcohol studies (Wechsler et al., 2000). The first item measured frequency in terms of the number of days of drinking alcohol in the past month prior to the study (0 = none [26%]; 1 = 1–3 days [23%]; 2 = 4–8 days [22%]; 3 = 9–20 days [22%]; 4 = 21 or more days [7%]). The second indicator was of recent frequency of excessive drinking assessed by the number

582

Journal of Social and Personal Relationships 24(4)

of binges (four drinks on one occasion if a woman, five if a man) in the 2 weeks prior to the study (Wechsler et al., 2000). This variable was labeled 2-week binges (1 = have not [50%]; 2 = once [12%]; 3 = 2 to 4 occasions [24%]; 4 = 5 to 8 occasions [13%]; and 5 = 9 or more occasions [1%]). The third indicator, last binge, measured time since the last excessive drinking episode (0 = not [37%]; 1 = over a year ago [4%]; 2 = within past year [15%]; 3 = within past 2 weeks [44%]). To assess Interpersonal Competence, a modified version of Buhrmester et al.’s (1988) Interpersonal Competence Questionnaire was utilized. Originally, the scale items were written to reflect a variety of relationships (e.g., acquaintance, friend, close partner). Based on our interest in competence with the romantic partner, the items and instructions were specific to that partner. Twenty-four items that addressed the three domains of self-disclosure, emotion support, and negative assertion were used. The respondents indicated on a 5-point Likert scale (0 = ‘I’m poor at this,’ 4 = ‘I’m extremely good at this’) how well they engaged in each behavior (e.g., disclosure – ‘Telling your partner about the things that secretly make you feel anxious or afraid’). Outcome variable: Dating relationship quality. Four scales were used to indicate evaluation, intention and connectedness of relationship quality. Hendrick’s (1988) 7-item satisfaction scale was used to assess satisfaction. Ambivalence was assessed with a 5-item scale developed by Walster et al. (1973; e.g., ‘How ambivalent or unsure are you about continuing in this relationship with your partner?’). This scale was reverse scored so that higher scores reflected unambivalence about the relationship. Commitment was measured with Sternberg’s (1988) 7-item scale. Satisfaction, unambivalence and commitment were answered using a 9-point Likert type format. Romantic couple connectedness was measured with Cross et al.’s (2000) 11-item relational interdependence selfconstrual scale (e.g., ‘My partner is an important reflection of who I am’), using a 7-point Likert type response format. Each item was worded to reflect the relationship with the particular romantic partner.

Results Model fitting and hypotheses evaluation The structural models used to evaluate the hypotheses were examined with several criteria. Global goodness of fit was evaluated with
2/df, CFI, and RMSEA and CMIN/df (the chi-square divided by the degrees of freedom should be < 2.00). A confirmatory factor analysis was used to confirm the factor structure of the latent variables. The results indicated an excellent fit to the data (
2(32) = 40.58, ns, CMIN/df was 1.27, CFI was .99 and RMSEA was .04 with RMSEA 90% confidence interval of .00 to .07). All of the measured indicators were significantly associated (p < .001) with their latent constructs, excessive drinking, interpersonal competence and relationship quality, respectively. Hypothesis 1: An Independence Model with a novelty-seeking pathway and a harm-avoidance pathway provide a good fit to the data. To address this hypothesis a model was fit as shown in Figure 1, including control variables (error terms are omitted to simplify the figure). The results for the Independence Model indicated that this model was a very good fit to the data (
2(79) = 95.30, ns, CMIN/df was 1.21, CFI was .98 and RMSEA was .04 with a 90% confidence

Fischer et al.: Pathways to dating quality

583

FIGURE 1 Predicting relationship quality: Significant unstandardized regression coefficients (standardized coefficients) in the Independence Model Distal Family Variable

Intermediate Personality Variables

Proximate Variables past month 4.46 (.78)

novelty seeking

.03*** (.38)

Relationship Outcome

2 week binge 1.00 (.90)

last binge 5.08 (.84)

EXCESSIVE DRINKING length

–.35** (–.23) 3.86*** (.21)

.29** (.20) .07* (.16)

1.00 (.86)

family dysfunction

RELATIONSHIP QUALITY

73 (.84)

2.28 (.74)

–.89* (–.19)

20.32 (.65)

satisfaction commitment unambivalent interdependence

.25*** (.56) 8.78*** (.28) –.44** (–.28)

INTERPERSONAL COMPETENCE

1.00 (.84)

sex –2.84*** (–.28)

harm avoidance

–.04** (–.29) disclosure

.82 (.68) emotion support

.87 (.68) negative assertion

Note. Nonsignificant paths are represented by dashed lines. *p < .05; **p < .01; ***p < .0001.

interval of .00 to .06). In the Independence Model, 38% of the variance in relationship quality was accounted for. The Independence Model fit indices support the principle of equifinality, that there are multiple paths to the same outcome. Huston’s (2000) social ecology model principle that more distal levels would have effects through intermediate and proximate levels on outcomes is supported by significant adjacent pathway parameters. The pattern of results points to the role of indirect and full mediation of variables in the model; only harm avoidance appears to be a partial mediator (between family dysfunction and interpersonal competence). After the next test of the interdependence model of Hypothesis 2, additional tests assess the significance of the indirect/ mediating effects. Hypothesis 2: Is the Interdependence Model a better fit than the Independence Model? In order to operationalize the systems principle of interdependence, a model fitted with cross-pathway links was hypothesized to provide a better fit to the data than a model without such cross-pathway links. Consistent with the systems theory principle of interdependence, these four cross-pathway links, novelty seeking ← → harm avoidance, novelty seeking → interpersonal competence, harm avoidance → excessive drinking, excessive drinking ← → interpersonal competence, were added to the Independence Model tested in

584

Journal of Social and Personal Relationships 24(4)

Hypothesis 1. First, the results for the Interdependence Model indicated that this model was a very good fit to the data (
2(75) = 90.34, ns, CMIN/df was 1.21, CFI was .98 and RMSEA was .04 with 90% confidence interval of .00 to .06). However, the comparison of the Independence Model as nested within the Interdependence Model provides the test of Hypothesis 2. With nonsignificant improvement (
2(df = 4) = 4.96, ns), this result failed to support Hypothesis 2. Moreover, none of the cross-pathway links were significant. Instead of interdependence, the results supported independence of the two pathways. Hypotheses 3a and 3b: Are sex and length related to the variables of the study? Three of six predicted paths in Hypothesis 3a involving the control variable of sex of the participant were significant. As depicted in Figure 1, men reported significantly greater excessive drinking, less harm avoidance, and less interpersonal competence than women. This mix of significant and nonsignificant results provides only partial support for the effects of participant sex. Effects of sex on relationship quality appeared to be carried via indirect/mediated effects rather than via direct effects. In support of Hypothesis 3b, length was predicted and found to be a significant, positive correlate of relationship quality. Follow-up analyses: Are the indirect/mediating effects significant? The presence of significant paths (Figure 1) from an independent variable to an intermediate or proximal variable and from this intermediate or proximal variable to a dependent variable does not indicate whether these were significant mediating or indirect effects. Thus, mediating and indirect effects were tested for significance using the bootstrap approach of Preacher and Hayes (2004). Biascorrected 95% confidence intervals evaluated the significance of the bootstrap estimate (5000 resamples) of the effect, where nonzero values within the confidence intervals identify a significant effect. This test required the use of measured variables, thus values derived from factor scores for the latent variables were used in these analyses. The tests were based upon interlinked connections contained within the Independence Model (Figure 1). The results of these tests are in Table 1. The indication of indirect, partial mediating, or full mediating effects is determined by the strength of the prior independent-dependent variable association. Mediating effects occur when there is a prior significant association; indirect effects occur when there is not. For example, because the zero-order association between family dysfunction and relationship quality (r = –.14, ns) was nonsignificant, tests involving family dysfunction to relationship quality were for indirect, not mediated effects. The table indicates whether the significant tested findings supported indirect, mediated, or partially mediated effects. These findings provided considerable evidence of indirect/mediating effects. In both the novelty-seeking and harm-avoidance pathways there were significant indirect effects from family dysfunction to relationship quality (Table 1).

Discussion The findings supported the social ecology model’s (Huston, 2000) principle that distal levels influence outcomes through intermediate and proximate levels. The novelty-seeking and harm-avoidance pathways were connected to relationship quality through significant indirect and mediating effects of

Indirect variable Dependent variable

–.08 –.18

–.41  .43 = –.18

.08

.05

.05

.01

–.03

.04

.01

–.02

.18

.23

–.58

–.46

.12

.02

.17

.14

–.09

.40

.11

SE of mean

–.17

Boot-strapped mean indirect effect

.32  –.24 = –.08

–2.77  –.02 = .04

* significant indirect effect; ** significant partial mediation effect; *** significant full mediation effect.

Indirect/mediated effects of control variable–relationship quality association Sex → Harm → Relationship quality avoidance Sex → Excessive → Relationship quality drinking Sex → Interpersonal → Relationship quality competence

Indirect/mediated effects of intermediate variable–relationship quality association Novelty seeking → Excessive → Relationship quality .11  –.21 = –.02 drinking Harm avoidance → Interpersonal → Relationship quality –.06  .45 = –.03 competence

Indirect/mediated effects of family dysfunction–proximate variable association Family dysfunction → Novelty → Excessive drinking 3.45  .12 = .40 seeking Family dysfunction → Harm → Interpersonal competence 8.68  –.05 = –.46 avoidance

Indirect/mediated effects of family dysfunction–relationship quality association Family dysfunction → Novelty → Relationship quality 3.45  –.05 = –.17 seeking Family dysfunction → Harm → Relationship quality 8.68  –.01 = –.09 avoidance Family dysfunction → Excessive → Relationship quality –.12  –.24 = .03 drinking Family dysfunction → Interpersonal → Relationship quality –1.35  .43 = –.57 competence

Independent variable

Unstandardized path coefficients and product

TABLE 1 Bootstrap tests of significance of indirect/mediation effects

–.34 to –.05*

–.20 to –.01*

–.03 to .15

–.04 to –.01*

–.05 to –.01***

–.92 to –.19**

.09 to .78*

–1.06 to –.17*

–.21 to .28

–.40 to .17

–.47 to –.02*

95% bias-corrected confidence interval

Fischer et al.: Pathways to dating quality 585

586

Journal of Social and Personal Relationships 24(4)

intermediate/proximate variables. The findings also supported the systems theory principle of equifinality in the form of two independent pathways from family of origin to relationship quality. The systems principle of interdependence failed to receive support. The role of family dysfunction Similar to other research (e.g., Bryant & Conger, 2002; Donnellan et al., 2005), family dysfunction was treated as the most distal variable. Its contributions to the outcome were via indirect or mediated effects. Moreover, family dysfunction provided a common point of origin for two distinct pathways. When families are dysfunctional, several processes (e.g., unmet needs, low levels of interpersonal competence) may manifest in young adult personality and behaviors. These manifestations then affect romantic outcomes. Although the effects are indirect, family dysfunction plays a consequential role in romance. The novelty-seeking excessive drinking pathway Novelty seeking provided an indirect path between family of origin and relationship quality. Excessive drinking fully mediated the association between novelty seeking and relationship quality. Past research reported links between family functioning and novelty seeking (e.g., Ravaja & Keltikangas-Jarvinen, 2001), novelty seeking and excessive drinking (Cloninger, Sigvardsson, Przybeck, & Svrakic, 1995), and drinking and relationship quality (Marshal, 2003). As noted by Fischer et al. (2005), 69% of college drinking occurred with romantic partners. It may be that drinking motivated by novelty seeking is accompanied by excitement. In comparison to this excitement, the tasks of romantic relationship maintenance might seem mundane or cumbersome. The partner may respond to actor’s excessive drinking with disagreements and disapproval, thus reducing actor’s satisfaction. Through means such as these, excessive drinking could diminish an individual’s perception of dating relationship interactions and quality. The harm-avoidance interpersonal competence pathway Parallel to the novelty-seeking pathway, the harm-avoidance pathway shared a common distal variable of family dysfunction. It is not surprising that individuals who are harm avoidant and have more dysfunctional families are less interpersonally competent. In dysfunctional families it is possible that individuals did not have the opportunity to learn appropriate interpersonal skills, or simply found them to be ineffective with their family members. More avoidant students might have had fewer romantic relationships prior to college, which resulted in attenuated skill development. Alternatively, harm-avoidant individuals might have adequate social skills but choose not to utilize these skills with their romantic partners. Romantic relationships are potentially rewarding, but they also carry risks and uncertainties (e.g., Knobloch & Carpenter-Theune, 2004) that could be especially salient to the harm avoidant. Thus, individuals withhold interactional

Fischer et al.: Pathways to dating quality

587

behaviors if they fear that their behaviors will be misunderstood. The premise fits with Gable and Shean’s (2000) findings that depressed individuals misinterpret communication signals and underestimate their abilities to interact with others. Interpersonal competence was the strongest predictor of relationship quality in this study. Such a linkage is consistent with past research (e.g., Meeks, Hendrick, & Hendrick, 1998). By comparison, excessive drinking was less strongly associated with romantic quality. Although past research has indicated that excessive drinking is a large part of college life (Wechsler et al., 2000), it may play a weaker role in romantic quality for young adults than does the use of effective interpersonal skills. Limitations and strengths of the study and future directions There were several limitations that should be considered. First, the sample was drawn from one context. Although the college context provides a setting for both excessive drinking and relationship experiences (Bachman et al., 1997), other young adult environments could reveal different patterns of association among the variables. Second, only a small number of personality and behavior variables were studied. Intermediate variables could include different personality factors. For example, the five-factor approach to personality assessment suggests other pathways to relationship quality than those seen in this study with harm avoidance and novelty seeking (cf., Frisbie, Fitzpatrick, Feng, & Crawford, 2000). Moderating effects on the associations that were found in this study should be examined (cf., Fischer & Wampler, 1994). Third, measures were obtained from only one partner. Data from both dating partners would allow examination of mutual partner effects and the interdependence between actor and partner (Thibaut & Kelley, 1959). In a study of spouses, Roberts and Leonard (1998) found that the impact of drinking on marital quality is dependent on both partners’ drinking patterns. Fourth, only one assessment method, a student questionnaire, was used. A multimethod study that utilizes diary, questionnaire, and observer measures of both partners would further illuminate romantic processes. Fifth, the direction of effects could be different than those tested. For example, it is possible that poorer relationship quality increases drinking or decreases interpersonal behaviors. However, prior research does not support this direction of effects. Fischer et al. (2005) tested direction of effects with diary data and reported that drinking affected conversation quality, but quality did not affect drinking. Similarly, Asendorpf and Wilpers’s (1998) longitudinal study found that personality influenced social relationships but relationships did not influence personality. Additional longitudinal research would allow testing of bidirectional and mutual effects consistent with systems and social ecology principles (Broderick, 1993; Huston, 2000). Strengths of the study include a theoretically derived model, consideration of dual pathways to relationship quality, and testing of distal, intermediate and proximate variables in a single model. Similar to Surra and Hughes’s (1997) consideration of multiple pathways to relationship commitment, the

588

Journal of Social and Personal Relationships 24(4)

present model promotes a better understanding of the linkages among and between the variables. The associations of historical (family of origin dysfunction), personality (novelty seeking, harm avoidance) and behavioral (drinking, competence) factors constituted distinct predictors of quality. The theoretically derived model allowed us to conceptualize the role of personality in complex ways, organizing and testing models that simultaneously included multiple pathways involving personality. In the search for explanations of college student variations in relationship quality, we proposed, tested and found support for two distinct pathways. Future research can expand the populations, variables, systems theory and social ecology axioms that are examined. Such research would broaden knowledge of personality–relationship linkages. REFERENCES Asendorpf, J. B., & Wilpers, S. (1998). Personality effects on social relationships. Journal of Personality and Social Psychology, 74, 1531–1544. Bachman, J. G., Wadsworth, K. N., O’Malley, P. M., Johnston, L. D., & Schulenberg, J. E. (1997). Smoking, drinking, and drug use in young adulthood: The impacts of new freedoms and new responsibilities. Mahwah, NJ: Erlbaum. Ball, S. A. (2004). Personality traits, disorders, and substance abuse. In R. M. Stelmack (Ed.), On the psychobiology of personality: Essays in honor of Marvin Zuckerman (pp. 203–222). New York: Elsevier Science. Baumrind, D. (1987). A developmental perspective on adolescent risk taking in contemporary America. In C. E. Irwin, Jr. (Ed.), Adolescent social behavior and health (pp. 93–125). San Francisco, CA: Jossey-Bass. Bombar, M., & Littig, L. (1996). Babytalk as a communication of intimate attachment: An initial study in adult romances and friendships. Personal Relationships, 3, 137–158. Bornstein, R. F., Geiselman, K. J., Gallagher, H. E., Ng, H. M., Hughes, E. E., & Languirand, M. A. (2004). Construct validity of the Relationship Profile Test: Impact of gender, gender role, and gender role stereotype. Journal of Personality Assessment, 82, 104–113. Bradbury, T., & Fincham, F. (1988). Individual difference variables in close relationships: A contextual model of marriage as an integrative framework. Journal of Personality and Social Psychology, 54, 713–721. Broderick, C. B. (1993). Understanding family process: Basics of family systems theory. Newbury Park, CA: SAGE Publications. Bronfenbrenner, U. (1989). Ecological systems theory. Annals of Child Development, 6, 187–249. Bryant, C. M., & Conger, R. D. (2002). An intergenerational model of romantic relationship development. In A. L. Vangelisti, H. T. Reis, & M. A. Fitzpatrick (Eds.), Stability and change in relationships (pp. 57–82). Cambridge: Cambridge University Press. Buhrmester, D., Furman, W., Wittenberg, M., & Reis, H. (1988). Five domains of interpersonal competence in peer relationships. Journal of Personality and Social Psychology, 55, 991–1008. Chotai, J., Jonasson, M., Hagglof, B., & Adolfsson, R. (2005). Adolescent attachment styles and their relation to the temperament and character traits of personality in a general population. European Psychiatry, 20, 251–259. Cloninger, C. R., (1987). Neurogenetic adaptive mechanisms in alcoholism. Science, 236, 410–416. Cloninger, C. R., Sigvardsson, S., Przybeck, T. R., & Svrakic, D. M. (1995). Personality antecedents of alcoholism in a national area probability sample. European Archives of Psychiatry and Clinical Neuroscience, 245, 239–244. Cloninger, C. R., Svrakic, D. M., & Przybeck, T. R. (2006). Can personality assessment predict

Fischer et al.: Pathways to dating quality

589

future depression? A twelve-month follow-up of 631 subjects. Journal of Affective Disorders, 92, 35–44. Cramer, D. (2004). Satisfaction with a romantic relationship, depression, support and conflict. Psychology and Psychotherapy: Theory, Research and Practice, 77, 449–461. Cross, S., Bacon, P., & Morris, M. (2000). The relational-interdependent self-construal and relationships. Journal of Personality and Social Psychology, 78, 791–808. Davies, P. T., & Cicchetti, D. (2004). Editorial: Toward an integration of family systems and developmental psychopathology approaches. Development and Psychopathology, 16, 477–481. Donnellan, M. B, Larsen-Rife, D., & Conger, R. D. (2005). Personality, family history, and competence in early adult romantic relationships. Journal of Personality and Social Psychology, 88, 562–576. Epstein, N. B., Baldwin, L. M., & Bishop, D. S. (1983). The McMaster family assessment device. Journal of Marital and Family Therapy, 9, 171–180. Feeney, J. A. (2004). Transfer of attachment from parents to romantic partners: Effects of individual and relationship variables. Journal of Family Studies, 10, 220–238. Fischer, J. L., Fitzpatrick, J., Cleveland, B., Lee, J.-M., McKnight, A., & Miller, B. (2005). Binge drinking in the context of romantic relationships. Addictive Behaviors, 30, 1496–1516. Fischer, J. L., & Wampler, R. S. (1994). Abusive drinking in young adults: Personality type and family role as moderators of family-of-origin influences. Journal of Marriage and the Family, 56, 469–479. Frisbie, S. H., Fitzpatrick, H., Feng, D., & Crawford, D. (2000). Women’s personality traits, interpersonal competence and affection for dating partners: A test of the contextual model. Social Behavior and Personality, 28, 585–594. Gable, S. L., & Shean, G. D. (2000). Perceived social competence and depression. Journal of Social and Personal Relationships, 17, 139–150. Gil, S. (2005). Personality traits and coping styles as mediators in risky sexual behavior: A comparison of male and female undergraduate students. Social Behavior and Personality, 33, 149–158. Givertz, M., & Segrin, C. (2005). Explaining personal and constraint commitment in close relationships: The role of satisfaction, closeness, conflict responses, and relational bond. Journal of Social and Personal Relationships, 22, 757–775. Hendrick, S. (1988). A generic measure of relationship satisfaction. Journal of Marriage and the Family, 50, 93–98. Huston, T. L. (2000). The social ecology of marriage and other unions. Journal of Marriage and the Family, 62, 298–320. Kelley, M. L., Nair, V., Rawlings, T., Cash, T. F., Steer, K., & Fals-Stewart, W. (2005). Retrospective reports of parenting received in their families of origin: Relationships to adult attachment in adult children of alcoholics. Addictive Behaviors, 30, 1479–1495. Kline, R. B. (2005). Principles and practice of structural equation modeling (2nd ed.). New York: Guilford Press. Knobloch, L., & Carpenter-Theune, K. (2004). Topic avoidance in developing romantic relationships: Associations with intimacy and relational uncertainty. Communication Research, 31, 173–205. Koesten, J. (2004). Family communication patterns, sex of subject and communication competence. Communication Monographs, 72, 226–244. Le, B., & Agnew, C. R. (2001). Need fulfillment and emotional experience in interdependent romantic relationships. Journal of Social and Personal Relationships, 18, 423–440. Leonard, K. E., & Roberts, L. J. (1996). Alcohol in the early years of marriage. Alcohol Health & Research World, 20, 192–196. Marshal, M. P. (2003). For better or worse? The effects of alcohol use on marital functioning. Clinical Psychology Review, 23, 959–997. Mason, W. A., & Windle, M. (2002). Reciprocal relations between adolescent substance use and delinquency: A longitudinal latent variable analysis. Journal of Abnormal Psychology, 111, 63–76.

590

Journal of Social and Personal Relationships 24(4)

Meeks, B., Hendrick, S., & Hendrick, C. (1998). Communication, love, and relationship satisfaction. Journal of Social and Personal Relationships, 15, 755–773. Miller, I. W., Epstein, N. B., Bishop, D. S., & Keitner, G. I. (1985). The McMaster family assessment device: Reliability and validity. Journal of Marital and Family Therapy, 11, 345–356. Nixon, S. J., & Parsons, O. A. (1989). Cloninger’s tridimensional theory of personality: Construct validity in a sample of college students. Personality and Individual Differences, 10, 1261–1267. Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, & Computers, 36, 717–731. Puttonen, S., Ravaja, N., & Keltikangas-Jarvinen, L. (2005). Cloninger’s temperament dimensions and affective responses to different challenges. Comprehensive Psychiatry, 46, 128–134. Ravaja, N., & Keltikangas-Jarvinen, L. (2001). Clonginger’s temperament and character dimensions in young adulthood and their relation to characteristics of parental alcohol use and smoking. Journal of Studies on Alcohol, 62, 98–104. Reti, I. M., Samuels, J. F., Eaton, W. W., Bienvenu, O. J. III, Costa, P. T. Jr., & Nestadt, G. (2002). Influences of parenting on normal personality traits. Psychiatry Research, 111, 55–64. Roberts, L. J., & Leonard, K. E. (1998). An empirical typology of drinking partnerships and their relationship to marital functioning and drinking consequences. Journal of Marriage and Family, 60, 515–526. Rose, R. J. (1998). A developmental behavioral-genetic perspective on alcoholism risk. Alcohol Health and Research World, 22, 131–143. Sabatelli, R. M., & Bartle-Haring, S. (2003). Family-of-origin experiences and adjustment in married couples. Journal of Marriage and Family, 65, 159–169. Schultheiss, D. P., & Blustein, D. L. (1994). Contributions of family relationship factors to the identity formation process. Journal of Counseling and Development, 73, 159–166. Sher, K. J., Wood, M. D., Crews, T. M., & Vandiver, P. A. (1995). The Tridimensional Personality Questionnaire: Reliability and validity studies and derivation of a short form. Psychological Assessment, 7, 195–208. Shim, S., & Maggs, J. (2005). A cognitive and behavioral hierarchical decision-making model of college students’ alcohol consumption. Psychology and Marketing, 22, 649–668. Sprecher, S., & Felmlee, D. (2000). Romantic partners’ perceptions of social network attributes with the passage of time and relationship transitions. Personal Relationships, 7, 325–340. Sternberg, R. (1988). Triangulating love. In R. Sternberg & M. Barnes (Eds.), The psychology of love (pp. 119–138). New Haven, CT: Yale University Press. Stewart, M. E., Ebmeier, K. P., & Deary, I.J. (2005). Personality correlates of happiness and sadness: EPQ-R and TPQ compared. Personality and Individual Differences, 38, 1085–1096. Surra, C. A., & Hughes, D. K. (1997). Commitment processes in accounts of the development of premarital relationships. Journal of Marriage and the Family, 59, 5–21. Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of groups. New York: Wiley. Tse, W. S., & Bond, A. J. (2005). The application of the temperament and character inventory (TCI) in predicting general social adaptation and specific social behaviors in a dyadic interaction. Journal of Applied Social Psychology, 35, 1571–1586. Walster, E., Berscheid, E., & Walster, G. (1973). New directions in equity research. Journal of Personality and Social Psychology, 25, 151–176. Wechsler, H., Lee, J. E., Kuo, M., & Lee, H. (2000). College binge drinking in the 1990s: A continuing problem. Results of the Harvard School of Public Health 1999 College Alcohol Study. Journal of American College Health, 48, 199–210. White, J. M., & Klein, D. M. (2002). Family theories. Thousand Oaks, CA: SAGE. Weigel, D. J., Bennett, K. K., & Ballard-Reisch, S. (2003). Family influences on commitment: Examining the family of origin correlates of relationship commitment attitudes. Personal Relations, 10, 453–474. Wills, T. A., Sandy, J. M., Yaeger, A., & Sinar, O. (2001). Family risk factors and adolescent substance use: Moderation effects for temperament dimensions. Developmental Psychology, 37, 283–297.