Translational research focuses on the application of basic scientific findings to clinically relevant mechanisms and from those mechanisms to clinical services. Our research specifically examines the pathways and outcomes of Interpersonal Violence (IPV) within families. By studying IPV through a translational lens we are doing more than just looking at what is most commonly thought of as partner abuse. Instead, we are expanding the scope to look at all the relationship and family factors that that contribute to violent behavior, the influence of IPV on the psychological and physical health of family members, and ways to prevent or treat those factors.
As you will see by reviewing our work, not only have we done what is more typically thought of as IPV research, but we have also investigated, among other things, the impact of IPV on oral health, conflict in teen dating relationships, anger responses and regulation in adult couple conflict, and physiological responses to both parenting and partner stimuli. We have used the knowledge gained from our research to test methods of prevention in a variety of settings including civilian, clinical, and military. We have developed classification systems for defining types of maltreatment that are being disseminated worldwide. Our research continues to expand with each passing year and as the tree below shows it all started from one idea.
In addition, we have established lines of research that intersect with our primary focus on IPV in families. We are interested in how to best disseminate effective, light touch interventions; how to engage communities for effective action; and how providers can most effectively communicate with clients/patients to promote health and wellbeing. We are also examining barriers to engaging in services and how to address those barriers through multidisciplinary approaches (e.g., partnering mental health providers and dentists to reduce dental fear and increase compliance).
The Rapid Marital Interaction Coding System (RMICS; Heyman, 2004) is an event-based system designed to measure frequencies of behavior and behavioral patterns (i.e., sequences) between intimate partners during conflicts. It is adapted from the Marital Interaction Coding System (MICS), the oldest and most widely used couples observational system (Heyman, 2001) until it was phased out in the late 1990s. We built on the established MICS literature and used both empirical and theoretical guidance to create a system (e.g., Heyman, 2001) that would be more reliable and valid despite being faster and easier to train and code. The RMICS has been used in approximately 40 separate investigations with a range of ages (primarily adult married couples, but also preteen siblings, high school dating couples, and engaged couples), populations (e.g., general couples population, couples therapy clients, cancer patients and their partners, families at risk for adolescent drug abuse, Vietnam veterans, friendship pairs), and research purposes.
In creating the RMICS, we first used a factor analysis of all 1,088 couples coded with the MICS over a 5 year period (Heyman, Eddy, Weiss & Vivian, 1995), which indicated that the original 37 microbehavioral MICS codes could be condensed into four "categories" - hostility, constructive problem discussion, humor, and responsibility discussion. The first three factors were used to create codes. The fourth, responsibility discussion, was incorporated into the broader notion of attributions. We used Holtzworth-Munroe and Jacobson's (1988) distillation of attributions into distress-maintaining and relationship-enhancing attribution codes. Further, we added several codes to make the system exhaustive and content valid. We also included two codes added to the original MICS after the factor analysis was conducted - withdrawal and dysphoric affect (Heyman, Weiss, & Eddy, 1995). Two positive codes (self-disclosure and acceptance) were also incorporated from a similar partner coding system, the Kategoriensystem fir partnerschaftliche interaktion (KPI; Hahlweg, Reisner, Kohli, Vollmer, Schindler, & Revenstorf, 1984). A highly negative code, psychological abuse, was added later.
In declining hierarchical importance, the RMICS comprises psychological abuse, distress-maintaining attributions (negative causal explanations); hostility (e.g., angry affect, criticism, combativeness); dysphoric affect (e.g., sad affect); withdrawal (e.g., stonewalling); relationship-enhancing attributions (positive causal explanations); acceptance (e.g., paraphrasing, expressions of caring); self-disclosure ("I" statements that express speaker's feelings, wishes or beliefs; acceptance of responsibility); humor (e.g., joking, laughing); constructive problem discussion (e.g., description of the problem, constructive solutions, questions and agreement); other (statements on something other than a personal or relationship topic; e.g., "Is that the camera?").
The RMICS defines the speaker turn as its basic coding unit. The codes are ordered hierarchically, based on both communication theory and substantial research that demonstrates that negative, followed by positive, followed by neutral codes are of decreasing importance in understanding partner conflict (see Weiss & Heyman, 1997). If someone emits more than one code during a speaker turn, s/he receives the code highest on the hierarchy. To deal with monologues, speaker turns that last more than 30 seconds are interval coded in 30-second segments (i.e., coded as if a new speaker turn occurs every 30 seconds).
Cohen's kappa is calculated on a random subset of couples for each study. Two coders are randomly assigned to code the same tape; they remain blind as to which tapes are being used for reliability testing. Our standard procedure is to assign 25% of the interactions for reliability testing.
The average overall Cohen's kappa per couple for 17 RMICS studies was .59 (SD = .17, N = 469; Heyman, 2004), which is good for complex a coding such as this. To accomplish this, (a) a single confusion matrix was created by collapsing the confusion matrices across all 469 couples, (b) 2 x 2 matrices were calculated for target code versus all other codes, and (c) Cohen's kappa was calculated. Agreement on all codes was good (kappa = .58 to .82), with the possible exception of the most infrequent code (psychological abuse, kappa = .46, which constituted about 0.10% of the observed behavior).
In recent years, agreement statistics that are not overly influenced by distributional characteristics have been added to the reliability calculations. These are V (Spitznagel & Helzer, 1985), G (Holley & Guilford, 1964), and AC1 Gwet (2002).
The RMICS has excellent content, discriminative, convergent, concurrent, and predictive validity (Heyman, 2004).
$100 per 10 minute interaction. (Longer interactions are pro-rated). This includes:
Gwet, K. (2002). Kappa statistic is not satisfactory for assessing the extent of agreement between raters. Statistical Methods for Inter-rater Reliability, 1, 1-5.
Heyman, R. E. (2001). Observation of couple conflicts: Clinical assessment applications, stubborn truths, and shaky foundations. Psychological Assessment, 13, 5-35.
Heyman, R. E. (2004). Rapid Marital Interaction Coding System. In P. K. Kerig & D. H. Baucom (Eds.) Couple observational coding systems (pp. 67-94). Mahwah, NJ: Lawrence Erlbaum Associates.
Heyman, R. E., Brown, P. D., Feldbau, S. R., & Oâ€™Leary, K. D. (1999). Couplesâ€™ Communication Variables as Predictors of Dropout and Treatment Response in Wife Abuse Treatment Programs. Behavior Therapy, 30, 165-190.
Heyman, R. E., Eddy, J. M., Weiss, R. L., & Vivian, D. (1995). Factor analysis of the Marital Interaction Coding System. Journal of Family Psychology, 9, 209-215.
Heyman, R. E., Weiss, R. L., & Eddy, J. M. (1995). Marital Interaction Coding System: Revision and empirical evaluation. Behavioural Research and Therapy, 33, 737-746.
Holley, W., & Guilford, J. P. (1964). A note on the G-index of agreement. Educational and Psychological Measurement, 24, 749â€“754.
Spitznagel, E. L., & Helzer, J. E. (1985). A proposed solution to the baserate problem in the kappa statistic. Archives of General Psychiatry, 42, 725-728.