SITAR Wiggins Award Winner: Sarah S. Dermody
Modeling the Complexity of Dynamic, Momentary Interpersonal Behavior: Applying the Time-Varying Effect Model to Test Predictions from Interpersonal Theory
Sarah S. Dermody; Christopher J. Hopwood; Kate M. Thomas; Nicholas Schade; Aidan G.C. Wright
In the past decade there have been considerable advances in the methodology and technology developed to code and capture momentary interpersonal behavior in dyadic transactions (e.g., Sadler et al., 2009). The intensive longitudinal data (ILD) generated by these methods offer new opportunities to examine fine-grained interpersonal processes. Compared to traditional longitudinal data, the total number of observations obtained per participant is substantially larger, which provides opportunities to both describe dynamic processes and hone in on key interpersonal processes in relatively greater detail. Thus, these methodological approaches address key questions raised in the field concerning the description and prediction of dynamic processes in interpersonal behavior.
At the same time, the long strings of continuously sampled data over the course of an interaction come with a host of analytic challenges. The sheer volume of data, with possibly hundreds or thousands of observation points per participant, can initially seem overwhelming to analyze. Further issues may arise when individuals or constructs are not assessed on the same time schedule. These analytic challenges are particularly salient when existing theory does not clearly inform requisite analytic steps to characterize the data. Many traditional longitudinal analysis approaches can adequately account for frequent and/or irregular observations, such has multi-level modeling (MLM) and latent growth curve modeling (LGC). These approaches are valuable insomuch as the associations of interest adhere to assumptions such as conforming to a known functional form (e.g., linear, quadratic) and are consistent over time.
These critical assumptions do not always hold, and as is likely the case for many interpersonal interactions of interest, may necessarily limit our potential to describe and understand key processes that develop or evolve over time. As outlined by Interpersonal Theory, interpersonal phenomena are understood to unfold over time, are not static–but necessarily dynamic, and often adapt to the demands of a situation (Pincus & Ansell, 2013; Sadler et al., 2011). For instance, Interpersonal Theory’s complementarity principle posits a probabilistic set of behavioral patterning in interactions that can be described via the primary dimensions of the interpersonal circumplex. Within this framework and in the context of dyadic interactions, dominance has been shown to pull for submissiveness (and vice versa) while friendliness pulls for friendliness (e.g., Markey et al., 2003; Sadler et al., 2009; Markey et al., 2010; Thomas et al., 2014).
A key component of Interpersonal Theory that remains untested is the extent to which interactions become more “in sync” over time. That is, prior research has focused on overall complementarity that is averaged over the course of an interaction. As previously described, however, as an interaction unfolds, complementarity may necessarily shift as the interacting parties negotiate a transactional rhythm. Thus, this raises a key empirical question of, “Does complementarity increase over the course of interpersonal interactions?”
To address this question, the current investigation provides a methodological demonstration of a recently popularized quantitative method, the Time-Varying Effect Model (TVEM; Tan et al., 2012). TVEM explicitly models relations between variables that are free to change over time. That is to say that TVEM can test the coupling and uncoupling of behavioral associations as time progresses. Furthermore, TVEM is a non-parametric approach, which allows for describing the change over time in constructs of interest as well as their interrelations without being confined to a predetermined functional form. These advantages of the TVEM hold the promise for posing and answering sophisticated questions about time-varying effects expected in interpersonal situations (for substantive examples see Shiyko et al., 2012; Wright et al., 2014). In the context of the question of how complementarity unfolds over time, TVEM can describe the interpersonal behaviors of each participant in an interaction in detail as well as how these interpersonal behaviors relate over time. Specifically, interpersonal theory predictions suggest that individuals increase their synchrony of behavior over the course of an interaction, which would correspond with an effect of confederate behavior on participant responding that increases over time. In the current study, this was tested by having participants come to the laboratory to interact with a confederate. Objective observers rated both the participant and confederate with respect to friendliness- and dominance-related behaviors. It was hypothesized that the strength of the relations between dominance and submissiveness in a dyad as well as between friendliness and friendliness would become stronger over time (i.e., more negative or more positive, respectively), suggestive of increase complementarity over time
Methods
Sample. The present study investigated a sample of unacquainted dyads, which included a sample of undergraduate students (N=67) who interacted with a male confederate. Mean age of the sample was 20 years old. Participants primarily consisted of young woman (60%) and Caucasians (85%). Participants provided informed consent in line with the Michigan State University Institutional Review Board.
Procedure. The confederate was instructed to cover all dimensions of the interpersonal circumplex in random order. The confederate switched his relative dominance and/or friendliness of his behavior a total of eight times over the course of the interaction. The duration of the interaction was approximately 18 minutes.
Measures. The outcome measures were momentary independent observer assessments of both the participant and confederate’s behavior. Using a computer joystick with labeled axes (horizontal axis: friendliness-related behaviors, vertical axis: dominance-related behaviors), nearly continuous ratings of friendliness and dominance were obtained. The final data included observations at every half second for up to 18 minutes (> 2100 observations per participant).
Analyses
We used TVEM to examine the association between confederate’s interpersonal behaviors (i.e., friendliness, dominance) with the participant’s interpersonal behavior on the same measure (i.e., friendliness and dominance, respectively). This involved first estimating an intercept-only model to describe the participant’s interpersonal behaviors over the course of the interaction. Then, the time-varying association between confederate and participant behavior was tested in two separate models (i.e., confederate friendliness predicting participant friendliness; confederate dominance predicting participant dominance). Analyses controlled for sex of the participant to account for same versus opposite sex dyads.
TVEMs were fit using P-spline functions using the %TVEM SAS macro, which is available for download at methodology.psu.edu (Yang et al., 2012). Following model fit practices outlined in Tan et al. (2012), models were initially fit with a moderate number of knots (K = 10). K was then gradually decreased to examine the impact on model fit. The final model was chosen based on the smallest values of AIC or BIC indices.
Results
Model Selection. The best fitting intercept functions were chosen based on the lowest AIC and BIC. Reducing the number of K from 10 generally led to increases in both AIC and BIC, suggesting that the more complex model with K = 10 should be retained. Results from each model are depicted in corresponding figures. The results in the figure include momentary point estimates depicted by the dark, solid line and corresponding momentary 95% confidence intervals represented by the lighter shaded band.
Intercept Functions: Participant Friendliness and Dominance over Time. The average change over time in participant friendliness level is depicted in Figure 1. On average, controlling for sex, levels of friendliness were elevated and generally stable over the course of the interaction.
Similarly, in Figure 2, the average change over time in participant dominance level is depicted. On average, controlling for sex, participants exhibit average levels of dominance, being neither dominant nor submissive, with a slight decline over the course of the interaction. The spike at the 900 second mark is likely due to the tendency for spline regression to “flare” toward the extremes of the values.
Time-Varying Effect of Confederate Friendliness on Participant Friendliness. The slope function of the effect of confederate friendliness on concurrent participant friendliness is depicted in Figure 3. Overall, there was a positive association between confederate and participant friendliness, as demonstrated by the positive slope function with 95% confidence interval band not inclusive of zero. The magnitude of the association changed over time. Specifically, the slope function generally increased over time, suggestive of a stronger predictive relation between confederate friendliness on participant friendliness over the course of the interaction.
Time-Varying Effect of Confederate Dominance on Participant Dominance. The slope function of the effect of confederate dominance on concurrent participant dominance is depicted in Figure 4. On average, throughout the duration of the interaction, confederate dominance was negatively associated with participant dominance. The strength of the association appeared to increase over time, such that confederate dominance was more strongly and negatively related to participant dominance at towards the end of the interaction than the start of the interaction.
Discussion
The primary aim of this investigation was to use TVEM to examine how complementarity within a dyadic interaction changes over time for unacquainted participants. Consistent with the existing literature, as expected, we found evidence for complementarity over the course of the interaction. On average, friendliness-like behaviors by the confederate corresponded with friendliness-like behaviors by the participants. Similarly, dominance-like behaviors by the confederate appeared to elicit relatively submissive-like behaviors from the participant.
The results of the present study also build upon the existing literature by demonstrating that the degree of complementarity may not be fixed, but rather unfolds over the course of an interaction. We found evidence that individuals increased in their complementarity over time. For example, over the course of the interaction, there was a modest increase in the strength of association between confederate and participant friendliness, as well as the association between confederate and participant dominance. Thus, through the application of TVEM, the present investigation provides initial support that complementarity in the context of unacquainted dyadic interactions is dynamic.
The findings of the present study should be considered in the context of some limitations. The confederate was not fully participating in the interaction as they were instructed to randomly shift between each of the octants of the interpersonal circumplex. It would be important to replicate the present findings in freely interacting dyads to determine generalizability to day-to-day interactions. It is also important to replicate the findings in additional datasets as non-parametric methods, such as TVEM, can yield results specific to particular data.
Nonetheless, it appears that TVEM can provide valuable insight into the nature of interpersonal interactions over time. The application of TVEM in the field opens the door for investigating key questions underlying interpersonal theory, such as those concerning how interpersonal interactions and corresponding complementarity unfold over time. This research could include further investigation of moderating variables of degree of complementarity over the course of an interaction, as well as time-varying effects of complementarity on individual’s response (e.g., emotional, physiological) to the interaction. Of note, these investigations using TVEM could make use of a variety of ILD sources that have gained popularity in the interpersonal research field (e.g., ecological momentary assessment, joystick data, daily diary, GPS data, physical movement, facial action coding). Taken together, we anticipate that TVEM could offer researchers a new tool make use of ILD to investigate the nature of interpersonal interactions over time.
References
Markey, P., Lowmaster, S., & Eichler, W. (2010). A real‐time assessment of interpersonal complementarity. Personal Relationships, 17(1), 13-25.
Markey, P. M., Funder, D. C., & Ozer, D. J. (2003). Complementarity of interpersonal behaviors in dyadic interactions. Personality and Social Psychology Bulletin, 29(9), 1082-1090.
Pincus, A., & Ansell, E. (2013). Interpersonal theory of personality. Handbook of psychology: Personality and social psychology, 5, 141-159.
Sadler, P., Ethier, N., Gunn, G. R., Duong, D., & Woody, E. (2009). Are we on the same wavelength? Interpersonal complementarity as shared cyclical patterns during interactions. Journal of Personality and Social Psychology, 97(6), 1005.
Sadler, P., Ethier, N., & Woody, W. (2011). Interpersonal complementarity//Handbook of interpersonal psychology: theory, research, assessment and therapeutic interventions: New Jersey: Wiley.
Shiyko, M. P., Lanza, S. T., Tan, X., Li, R., & Shiffman, S. (2012). Using the time-varying effect model (TVEM) to examine dynamic associations between negative affect and self confidence on smoking urges: Differences between successful quitters and relapsers. Prevention Science, 13(3), 288-299.
Tan, X., Shiyko, M. P., Li, R., Li, Y., & Dierker, L. (2012). A time-varying effect model for intensive longitudinal data. Psychological methods, 17(1), 61.
Thomas, K. M., Hopwood, C. J., Woody, E., Ethier, N., & Sadler, P. (2014). Momentary assessment of interpersonal process in psychotherapy. Journal of counseling psychology, 61(1), 1.
Wright, A. G., Hallquist, M. N., Swartz, H. A., Frank, E., & Cyranowski, J. M. (2014). Treating co-occurring depression and anxiety: Modeling the dynamics of psychopathology and psychotherapy using the time-varying effect model. Journal of consulting and clinical psychology, 82(5), 839.
Yang, J., Tan, X., Li, R., & Wagner, A. (2012). TVEM (time-varying effect model) SAS macro suite users’ guide.