Enhancing the Assessment of Verbal Aggression Through Observational Methodology

Abstract

The assessment of verbal aggression in adolescent and young adult dating relationships has largely relied on self-report methodology. We investigated whether information on verbal aggression derived from an observational assessment would enhance the prediction of romantic relationship satisfaction and dissolution in a sample of young adult dating relationships (N = 113). Observationally assessed verbal aggression was moderately associated with self-reported verbal aggression. Consistent with previous findings, neither self-reported nor observationally assessed verbal aggression was associated with relationship dissolution. Observationally assessed verbal aggression and self-reported verbal aggression each uniquely accounted for a substantial amount of variability in romantic relationship satisfaction. The findings of this study provide additional support for conducting multimethod assessments of verbal aggression and incorporating observational methodology in the study of aggression in young adult dating relationships.

Keywords

methodology verbal aggression dating relationships

Previous studies of verbal aggression in dating relationships have largely relied on self-report measures. Common self-report measures on verbal aggression in dating relationships include the Conflict Tactics Scale (Straus, Hamby, Boney-McCoy, & Sugarman, 1996) and the Conflict in Adolescent Dating Relationships Inventory (Wolfe et al., 2001). Yet self-reports of verbal aggression are problematic for several reasons. Given the sensitive nature of dating violence and the stigma attached to such acts, the tendency for individuals to provide self-preserving and socially desirable responses warrants caution when considering information gathered from self-report measures (Bell & Naugle, 2007; Sugarman & Hotaling, 1997). Furthermore, some of the predictors of violence are also associated with underreporting (Szinovacz & Egley, 1995). For example, research with married couples shows that gender may relate to underreporting, as females tend to underreport levels of dating violence by 50% to 56% and males by 60% to 83%. These levels of underreporting—and the fact that they differ between females and males—may introduce systematic error in the extant self-report literature on couple violence (Perry & Fromuth, 2005; Schafer, Caetano, & Clark, 2002; Szinovacz & Egley, 1995). The purpose of the current study was to investigate whether the assessment of verbal aggression in dating relationships could be enhanced by including observational methodology. As experiences of dating violence have been consistently associated with negative individual and romantic couple outcomes (Kaura & Lohman, 2007; O’Leary, 1999), we felt it was important to investigate romantic relationship quality indicators, more specifically, romantic relationship satisfaction and dissolution (Shortt, Capaldi, Kim, & Owen, 2006).

Observational methodology is both objective and capable of providing simultaneous and dynamic assessment of both members of romantic dyads. Observational methodology is widely used in the marital literature. For example, one of the most widely used marital observational coding schemes is the Marital Interaction Coding System (MICS; Hops, Wills, Weiss, & Patterson, 1972). Currently, in its fourth revision (MICS-IV; Heyman, Weiss, & Eddy, 1995), this coding scheme focuses on verbal and nonverbal communicative processes among married couples’ communication. Another well-validated marital coding system is the Couple Interaction Scoring System (Notarius, Markman, & Gottman, 1983), which focuses on problem-solving strategies in couples. Coding systems such as, but not limited to, the aforementioned systems have been used to understand a host of marital processes, such as demand-withdrawal characteristics (Rehman & Holtzworth- Munroe, 2006), negative reciprocity (Cordova, Jacobson, Gottman, Rushe, & Cox, 1993), empathy and self-disclosure (Mitchell et al., 2008), social support (Verhofstadt, Buysse, Ickes, Davis, & Devoldre, 2008), humor (Campbell, Martin, & Ward, 2008), and intimacy, emotional skill, and relationship satisfaction (Mirgain & Cordova, 2007). Thus, observations of marital communication have a long-standing and rich history.

Observational methodology, however, is seldom used to assess verbal aggression in the context of dating relationships (see, for an exception, Capaldi & Crosby, 1997). There is only one direct investigation of observationally assessed verbal aggression within dating relationships that we know of, namely, Capaldi and Crosby (1997). In a sample of unmarried couples, Capaldi and Crosby found that verbal aggression based on an observational assessment was moderately associated with self-reported verbal aggression in the couple relationship. Using both self-report and observed verbal aggression to create a verbal aggression construct, Capaldi and Crosby concluded that relationship satisfaction was negatively associated with psychological aggression. Therefore, the extant literature suggests that verbal aggression based on observational methodology is associated with verbal aggression based on self-report methodology among dating couples and predicts romantic relationship satisfaction.

What is less clear, however, is whether verbal aggression based on observational methodology uniquely predicts romantic relationship satisfaction, net from the effect of verbal aggression based on self-report methodology. Put another way, if observationally assessed verbal aggression is considered as a separate—from self-reported verbal aggression—predictor of romantic relationship satisfaction, do they account for the effect of one another (i.e., does the effect of one or both become nonsignificant) or do they account for unique prediction of romantic relationship satisfaction? If the predictive validity of both methodologies is unique, then this would suggest the importance of including both methodologies in the assessment of verbal aggression. If the predictive validity of both methodologies is redundant, this would suggest one of these methodologies is superior in predicting romantic relationship functioning.

While observational methodology has a much longer history of employment in the marital violence literature, it is typically conducted with coding schemes that do not mirror self-report items in content. It is our contention that multimethod studies on dating violence would benefit from verbal aggression indicators that are relatively consistent across methods. In other words, coding schemes that are equivalent to questions based on self-report measures may add to the validity of such multimethod constructs. As such, one goal of the current study was to explore the usefulness of an observational coding scheme of verbal aggression that is equivalent—in terms of underlying indicators—to a self-report measure of verbal aggression.

In the current study, we plan to (a) replicate findings by Capaldi and Crosby (1997) on the relationship between self-reported and observationally assessed verbal aggression and (b) extend these findings by investigating if observationally assessed verbal aggression and self-reported verbal aggression uniquely predict romantic relationship satisfaction and dissolution. We hypothesize that observationally assessed and self-reported verbal aggression are positively associated. We also hypothesize that observationally assessed verbal aggression and self-reported verbal aggression are both negatively associated with romantic relationship satisfaction and each uniquely predict romantic relationship satisfaction. Consistent with previous research (Shortt et al., 2006), we predict that neither observationally assessed nor self-reported verbal aggression is associated with romantic relationship dissolution.

Method

Participants

Participants included 120 dating couples from a large Midwestern state university. To be eligible for participation, couples had to be dating for at least 4 months and one of the dyad members had to be a university student. The effective sample size included 113 heterosexual couples, after exclusion of homosexual couples (n = 4) and couples whose observational assessment could not be coded due to technical difficulties during the assessment (n = 3). None of the couples were married or engaged, and 7% of couples were cohabitating at the time of study. Twenty-nine percent of the couples had been involved in their current relationship for 4 to 6 months, 19% had been involved in their current relationship for 7 to 12 months, and 52% had been involved in their current relationship for 12 months or longer. The average age of the participants was 19.78 years. The sample was predominantly White (88%). Human subjects approval was obtained through the Kent State University Institutional Review Board.

Procedure

All participants first completed a battery of survey measures followed by a romantic relationship interview (adapted from the Minnesota Longitudinal Study of Risk and Adaptation, Sroufe, Egeland, Carlson, & Collins, 2005). Next, participants completed an observational assessment. Couples participated in an observational assessment that included a conflict resolution task (Cox, 1991) and a plan-a-weekend getaway task (modified for dating couples from the family vacation task; Cooper & Grotevant, 1987). For the plan-a-weekend getaway task, couples were given the following instructions:

We want you to pretend that you have won a radio contest. The grand prize is an all-expenses paid 3-day weekend. Your job is to plan together what the two of you would do. Remember, you can do anything you want, go anywhere you want, and you have all the money you need. We want you to plan the whole three days together. You will have ten minutes to complete this task.

The observational assessments lasted for approximately 45 min.

Measures

Self-report assessment of verbal aggression

The Conflict in Adolescent Dating Relationships Inventory (CADRI; Wolfe et al., 2001) assesses dating violence perpetration toward each participant’s current partner over the past year. In the current study, we focused on the 10-item Emotional and Verbal Abuse subscale. Each member of the couples reported on their own behaviors toward partner as well as on their partner’s behaviors toward them on a 4-point Likert-type scale. Previous research has demonstrated adequate reliability and validity for the CADRI and the Emotional and Verbal Abuse subscale (Wolfe et al., 2001). Analyses in this article were limited to a composite of four items (blaming one’s partner for a problem, insulting one’s partner with put-downs, saying something to make one’s partner angry, and bringing up something bad that one’s partner did in the past) that were also observationally coded. The composite scale demonstrated good internal consistency for both females (α = .82) and males (α = .82). Self-reported verbal aggression for males and females was positively correlated (r = .45, p < .001) and therefore averaged to create a dyadic self-report verbal aggression score (M = 1.96, SD = .65).

Observationally assessed verbal aggression.

All assessments were coded for verbal aggression based on a coding scheme adapted from the 10-item Emotional and Verbal Abuse subscale of the CADRI (Wolfe et al., 2001). We chose to not code four items from the original scale as it was unlikely the behaviors would appear during the observational assessment (threatened to end the relationship, ridiculed or made fun of him or her in front of others, kept track of who she was with and where she was) or we had concerns about the potentially subjective nature (spoke to him or her in a hostile or mean tone of voice). Two additional items (did something to make him or her feel jealous, M = 1.05, SD = .31; accused him or her of flirting with another girl or boy, M = 1.06, SD = .23) had so little variability—and adversely affected the internal consistency of the scale—that we did not include them in the final construct. The final construct consisted of the following four items: blaming one’s partner for a problem, insulting one’s partner with put-downs, saying something to make one’s partner angry, and bringing up something bad that one’s partner did in the past. Similar to the self-report CADRI, each of the observational items was rated on a 4-point Likert-type scale, wherein ratings ranged from 1 = never to 4 = often, in regard to the frequency of each abuse’s occurrence over the span of a full observational tape. These observational tapes were coded by trained undergraduate and graduate research assistants. Coders were trained as a group with practice tapes. Tapes were initially coded by seven coders. After all coders were trained up to criterion (80% agreement or higher), they coded in pairs. All disagreements between coders were conference coded in meetings with other coders. The coders watched a tape of the complete observational assessment and coded the interaction throughout the assessment. Coders were naïve of the verbal aggression status on the dyadic member’s self-report.

All tapes were coded by at least two coders. Interrater reliability (intraclass correlations) for all four items was generally good (said something to make partner angry, r = .59; insulting, r = .76; brought up past, r = .78; blaming, r = .81). Agreement was similar across male and female members of dyads (largest difference in r between males and females = .08. The overall scale demonstrated adequate internal consistency for both females (α = .61) and males (α = .62). Observationally coded verbal aggression for males and females was positively correlated (r = .52, p < .001) and therefore averaged to create a dyadic observationally coded verbal aggression score (M = 2.00, SD = .53).

Relationship satisfaction

The seven-item Relationship Adjustment Scale (RAS; Hendrick, 1988) assesses each participant’s romantic relationship satisfaction with their current partner on a 5-point scale (higher scores represent greater romantic relationship satisfaction). Previous research demonstrated good reliability and validity for the RAS (Hendrick et al., 1998). For this study, internal consistency was good for females (α = .88) and males (α = .82). Relationship satisfaction for males and females was positively correlated (r = .49, p < .001) and therefore averaged to create a dyadic relationship satisfaction score.

Romantic relationship dissolution

After the initial in-lab assessment, participants completed a 3- and 6-month web-based follow-up assessment. These web-based assessments consisted of up to 14 questions, including a question concerning romantic relationship dissolution. Approximately 84% of participants completed the 3-month follow-up, while 85% of participants completed the 6-month follow-up. Twenty-five percent of couples were no longer dating the same partner at time of follow-up. No statistically significant differences on self-reported verbal aggression, observationally assessed verbal aggression, and relationship satisfaction between the participants who dropped out (N = 11) and participants who completed either the 3-month or 6-month follow-up assessments (N = 102).

Results

First, we wanted to replicate previous findings on the association between observationally assessed and self-reported verbal aggression. As hypothesized, Pearson correlations (see Table 1) indicated observationally coded verbal aggression was positively associated with self-reported verbal aggression (r = .38, p < .001). Next, we investigated the association between verbal aggression and romantic relationship satisfaction. Observationally assessed (r = −.34, p < .001) and self-reported (r = −.42, p < .001) verbal aggression were both negatively associated with romantic relationship satisfaction. We also wanted to understand whether observationally assessed and self-reported verbal aggression uniquely predicted romantic relationship satisfaction. To answer this question, we conducted two linear multiple regression analyses. In the first model, we entered self-reported verbal aggression first, followed by observationally assessed verbal aggression. In the second model, we entered observationally assessed verbal aggression first, followed by self-reported verbal aggression. Results of the first linear regression model (see Table 2) indicated that observationally coded verbal aggression was negatively associated with relationship satisfaction above and beyond self-reported verbal aggression (β = −.21, p = .02), F(3, 112) = 10.65, p < .001, R² = .23, adjusted R² = .20, ΔR² = .21. Results of second regression model indicated self-reported verbal aggression predicted romantic relationship satisfaction above and beyond observationally assessed verbal aggression (β = –.35, p < .001, ΔR² = .35). Together, the findings from these two regression models indicate observationally assessed verbal aggression and self-reported verbal aggression uniquely predict romantic relationship satisfaction.

Table 1.

Descriptive Statistics and Pearson Correlations for Romantic Relationship Indicators and Verbal Aggression

Variable	1	2	3	4
Verbal aggression (self-report)	—
Verbal aggression (observationally assessed)	.38**	—
Romantic relationship dissolution	.28**	−.08	—
Romantic relationship satisfaction	−.42**	−.34**	.02	—
M	1.96	2.00	0.27	4.42
SD	.65	.53	.88	.51

Note: N = 102.

p < .05. **p < .01.

Table 2.

Linear Regression Analysis Results Predicting Relationship Satisfaction From Self-Report and Observationally Assessed Verbal Aggression

Variable	B	SE B	β
Verbal aggression (self-report)	−.27	.07	−.35*
Verbal aggression (observationally assessed)	−.20	.09	−.21*
R ²		.21
F		15.01

p < .05. **p < .01.

Finally, we investigated the association between verbal aggression and romantic relationship dissolution. Results of independent samples t tests indicated that neither observationally assessed, t(100) = .14; p = .88, nor self-reported, t(100) = −.35; p = .72, verbal aggression was significantly associated with relationship dissolution.

Discussion

The current study replicated findings by Capaldi and Crosby (1997) and provides further evidence that self-report and observationally assessed verbal aggression are moderately associated among individuals in dating relationships. We extend these findings, however, by also demonstrating that observationally assessed verbal aggression and self-reported verbal aggression are uniquely associated with romantic relationship satisfaction. Thus, these findings provide further evidence that the assessment of verbal aggression should not solely rely on self-report assessments but can be enhanced by incorporating observational methodology.

Previous studies that use self-report and observational assessments of verbal aggression have primarily focused on enhancing construct validity. Capaldi and Crosby (1997), for example, used both self-report and observational information on verbal aggression as manifest indicators of a latent variable. Although such an aggregated multimethod approach may be useful in some research investigations, it also could potentially obscure some unique effects. That is, we demonstrated observationally assessed verbal aggression predicted romantic relationship satisfaction, net of the effect of self-report psychological abuse. This finding suggests that observationally assessing verbal aggression provides incremental validity above and beyond the information derived from self-report measures of verbal aggression.

As far as we know, the current study is the first to develop an observational coding scheme that closely parallels items on self-report measures of verbal aggression. Considering observational coding schemes that conceptually and psychometrically parallel self-report measures is an important next step in advancing multimethod studies on interpersonal violence. Previous studies on verbal aggression that included information based on self-report and observational assessments have enabled us to understand to what degree the residual variance (i.e., lack of overlap between the two methods) was due to (a) differences in self-report versus observational methodology or (b) differences in operationalization of the specific indicators for the construct within each method. The findings from the current study help to further disentangle these differences by having a similar operationalization of the specific indicators and thus obtaining a more accurate picture of the degree to what information on verbal aggression overlaps across self-report and observational methodology. This is an important methodological extension because self-report assessments of verbal aggression rely on participant memory and include information on verbal aggression across a specific period of time (e.g., 6 months or 1 year). However, information on verbal aggression through observational methodology is limited to the occurrence of verbal aggression in a period of several minutes. Furthermore, observationally assessed verbal aggression is context specific. Combining these methodological approaches thus provides a combination of information on short and longer term verbal aggression that reflects context-specific and context-independent behavior. We think this approach holds promise for other investigations and a logical next step includes extending this work into samples of married couples.

Limitations of Current Study

Consistent with the findings from Shortt and colleagues (2006), we found that verbal aggression—either self-report or observationally assessed—was not associated with romantic relationship dissolution. Although this finding is consistent with previous empirical work, we may have been somewhat underpowered in our analyses, given that the 6-month dissolution rate in the current sample was 25%. The low base rate of dissolution in the current sample combined with the fact that most couples had been together for at least 1 year are important limitations of the current study. Therefore, an important next step is to conduct multimethod studies of the association between verbal aggression and romantic relationship dissolution in the initial stages of romantic relationships.

To our knowledge, this study is the first to introduce a “plan-a-weekend getaway” task in observational research with nonmarital couples. What is not clear from the current study is whether this task elicited less verbal aggression than the conflict resolution task. In other words, to what degree average levels of verbal aggression differed across the different tasks. An important next step will be to evaluate the psychometric properties of the “plan-a-weekend getaway” task in other research studies and examine to what degree average levels of verbal aggression—as well as other aspects of couple interactions—differ between this task and others used in research with marital and nonmarital couples.

Footnotes

Acknowledgements

The authors would like to thank Alex Bea, Emily Buehler, Jessica Ross, Megan McElroy, Joseph Pochedly, and Jessica Ross for help with the coding of data and the Transitions into Adulthood and Romantic Relationships Research Team for assistance with data collection.

Portions of this article were previously presented at the 2009 annual meeting of the Midwestern Psychological Association (Chicago, IL).

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was partially funded by the Center for the Treatment and Study of Traumatic Stress Summa Health Systems/Kent State University.

References

Bell

K. M.

Naugle

A. E.

(2007). Effects of social desirability on students’ self-reporting of partner abuse perpetration and victimization. Violence and Victims, 22, 243-256.

Campbell

Martin

R. A.

Ward

J. R.

(2008). An observational study of humor use while resolving conflict in dating couples. Personal Relationships, 15, 41-55.

Capaldi

D. M.

Crosby

(1997). Observed and reported psychological and physical aggression in young, at-risk couples. Social Development, 6, 184-206.

Cooper

C. R.

Grotevant

H. D.

(1987). Gender issues in the interface of family experience and adolescents’ friendship and dating identity. Journal of Youth and Adolescence, 16, 247-264.

Cordova

J. V.

Jacobson

N. S.

Gottman

J. M.

Rushe

Cox

(1993). Negative reciprocity and communication in couples with a violent husband. Journal of Abnormal Psychology, 102, 559-564.

Cox

M. J.

(1991). Marital and Parent–Child Relationships Study. Unpublished manuscript, University of North Carolina at Chapel Hill.

Hendrick

S. S.

(1988). A generic measure of relationship satisfaction. Journal of Marriage and the Family, 50, 93-98.

Hendrick

S. S.

Dicke

Hendrick

(1998). The relationship assessment scale. Journal of Social and Personal Relationships, 15, 137-142.

Heyman

R. E.

Weiss

R. L.

Eddy

M. J.

(1995). Marital Interaction Coding System: Revision and empirical evaluation. Behavior Research and Therapy, 33, 737-746.

10.

Hops

Wills

T. A.

Weiss

R. L.

Patterson

G. R.

(1972). Marital Interaction Coding System. Eugene: University of Oregon and Oregon Research Institute (National Auxiliary Publication Service No. 02077).

11.

Kaura

S. A.

Lohman

B. J.

(2007). Dating violence victimization, relationship satisfaction, mental health problems, and acceptability of violence: A comparison of men and women. Journal of Family Violence, 22, 367-381.

12.

Mirgain

S. A.

Cordova

J. V.

(2007). Emotion skills and marital health: The association between observed and self-reported emotion skills, intimacy, and marital satisfaction. Journal of Social and Clinical Psychology, 26, 983-1009.

13.

Mitchell

A. E.

Castellani

A. M.

Herrington

R. L.

Joseph

J. I.

Doss

B. D.

Snyder

D. K.

(2008). Predictors of intimacy in couples’ discussions of relationship injuries: An observational study. Journal of Family Psychology, 22(1), 21-29.

14.

Notarius

C. I.

Markman

H. J.

Gottman

J. M.

(1983). Couples interaction scoring system: Clinical implications. In Filsinger

E. E.

(Ed. ), Marriage and family assessment (pp. 117-136). Beverly Hills, CA: SAGE.

15.

O’Leary

K. D.

(1999). Psychological abuse: A variable deserving critical attention in domestic violence. Violence and Victims, 14, 3-24.

16.

Perry

A. R.

Fromuth

M. E.

(2005). Courtship violence using couple data: Characteristics and perceptions. Journal of Interpersonal Violence, 20, 1078-1095.

17.

Rehman

U. S.

Holtzworth-Munroe

(2006). A cross-cultural analysis of the demand–withdraw marital interaction: Observing couples from a developing country. Journal of Consulting and Clinical Psychology, 74, 755-766.

18.

Schafer

Caetano

Clark

C. L.

(2002). Agreement about violence in U.S. couples. Journal of Interpersonal Violence, 17, 457-470.

19.

Shortt

J. W.

Capaldi

D. M.

Kim

H. K.

Owen

L. D.

(2006). Separation for young, at-risk couples: Prediction form dyadic aggression. Journal of Family Psychology, 20, 624-631.

20.

Sroufe

L. A.

Egeland

Carlson

E. A.

Collins

W. A.

(2005). The development of the person: The Minnesota study of risk and adaptation from birth to adulthood. New York, NY: Guilford.

21.

Straus

M. A.

Hamby

S. L.

Boney-McCoy

Sugarman

D. B.

(1996). The revised Conflict Tactics Scales (CTS2): Development and preliminary psychometric data. Journal of Family Issues, 17, 283-316.

22.

Sugarman

D. B.

Hotaling

G. I.

(1997). Intimate violence and social desirability: A meta-analytic review. Journal of Interpersonal Violence, 12, 275-290.

23.

Szinovacz

M. E.

Egley

L. C.

(1995). Comparing one-partner and couple data on sensitive marital behaviors: The case of marital violence. Journal of Marriage and the Family, 57, 995-1010.

24.

Verhofstadt

L. L.

Buysse

Ickes

Davis

Devoldre

(2008). Support provision in marriage: The role of emotional similarity and empathic accuracy. Emotion, 8, 792-802.

25.

Wolfe

D. A.

Scott

Reitzel-Jaffe

Wekerle

Grasley

Straatman

(2001). Development and validation of the Conflict in Adolescent Dating Relationships Inventory. Psychological Assessment, 13, 277-293.