Modeling Semester-Long Recursive Dynamics of the Expectancy-Value Motivation System Among Undergraduate Biology Students

• In Event: 61.015 - Embracing and Modeling the Complex Dynamics of Motivation and Engagement: Contextual, Temporal, Dynamic, and Systematic

Abstract

Purpose
In the current study, we used 13 weekly waves of data throughout a semester from two samples of undergraduate biology students to investigate motivational constructs from expectancy-value theory (Wigfield & Eccles, 2000) as a dynamic system that operates recursively at the individual student level. Rather than aggregating data from these individuals, we generalize from patterns emerging across individual-focused analyses to theoretical understandings about the principles by which the expectancy-value motivational system behaves in the context. The dual purpose of this study was to demonstrate an analytical approach that models motivational phenomena as dynamic and complex, and to gain robust theoretical understandings of expectancy-value processes.
Theoretical Framework
Expectancy-value theorists contend that students’ investment in schoolwork is the product of their perceived expectancies to do well and their valuing of the domain, with perceived cost hindering this investment. We conceptualized these constructs as mutually influencing each other continuously, and as constituting, together with emotions and investment, an individual’s complex motivation system that iterates along time. Correspondingly, we analyze this complex system with autoregressive mixed effects modeling that models the recursive behavior of the system with “recursive equations” (i.e., yt+1=ƒ[yt]), in which each iteration of the entire system serves as an output of its previous iteration and an input for its next iteration. In the equations, all the variables in the system predict the recursive equations of all other variables.
Method
Semester-long weekly data were collected through brief surveys (scale 0-10) from two samples (N1=145, N2=392) of undergraduate biology students in Fall 2016 and Spring 2017 respectively. Analysis included five constructs of the expectancy-value motivation system—investment, expectancy, value, cost, and an emotion (frustration)—and two contextual variables assumed to influence the system: anticipating within-semester exams and scores on these exams. We specified five autoregressive mixed-effect equations for each participant—one for each systemic construct (Figure 1 presents the equation format). The analysis produces a matrix of 35 coefficients for each participant, which we examined through range, distributions, and correlations for patterns that manifest characteristics of the expectancy-value system across participants in the context.
Results
In both samples, there was marked variability in the valence and magnitude of each and every coefficient, suggesting a substantial idiosyncratic facet of the expectancy-value system. Figure 2 shows expectancy trajectories from sample 1. Also, across participants in both samples, influences on cost trajectories were small to moderate and fluctuated less relative to influences on investment and expectancy trajectories that were moderate to large, and fluctuated greatly. Correlations between coefficients also differed between samples, with very high negative interdependencies (rs=-.60 to -.90) between coefficients of expectancy and of value predicting trajectories of cost, frustration, and investment in sample 1 but not in sample 2.
Significance
Modeling that is compatible with understanding motivation as an individual-level, complex, and dynamic phenomenon revealed more complex processes of the expectancy-value motivational system than those emerging from traditional aggregate, component-dominant, analyses. This analytical approach has potential to capture more authentic characteristics of motivational phenomena and contribute to more robust theoretical understandings.

Authors

• Avi Kaplan, Temple University

• Ting Dai, University of Illinois at Chicago

• Zoran Obradovich, Temple University

• Tony Perez, Old Dominion University

• Jennifer G. Cromley, University of Illinois at Urbana-Champaign

• Kyle Mara

• Michael Balsai