Antecedents and Effects of Test Anxiety Components: An Intra-Individual Real-Time Approach Including Physiological Measures

• In Event: 67.020 - Promises and Challenges of Studying Physiological Processes in Academic Settings

Abstract

Objectives and Theoretical Framework. As test anxiety is a prevalent emotion in school, knowledge about its antecedents and effects is crucial (Zeidner, 2014). The control-value theory (CVT) refers to control as a central antecedent and performance as a major effect of anxiety (Pekrun, 2006). However, although anxiety constitutes of multiple components (i.e., cognitive, affective, motivational, and physiological components) and some findings suggest that there might be differences regarding their antecedents and effects, previous research (with the exception of rather roughly differentiating just worry and emotionality) largely neglected to examine them separately (Liebert & Morris, 1967). Therefore, this study is investigating the individual anxiety components to identify which component should be primarily addressed in order to increase students’ performance. As the CVT describes intra-individual processes, we aim to extend findings from traditional inter-individual studies by applying an intra-individual approach (within-person measures during a mathematics test; Murayama et al., 2017). Furthermore, in addition to self-reports, the physiological component is examined with physiological measures to provide an objective assessment (i.e., electrodermal activity; EDA - a proxy for emotional arousal; Betancourt et al., 2017). We hypothesize that anxiety will mediate the relationship between control and performance and that the anxiety components will vary in the strength of their mediating effects (details in Figure 1).

Methods. Data from N=50 German 8th graders (Mage=13.72, SD=.53, 50% female) was collected during a mathematics test which consisted of six task-blocks. Students were given five minutes to work on each block. Before they started with each block they completed a state-questionnaire, which included one item for control and three items for each anxiety component from the Academic Emotions Questionnaire (Pekrun et al., 2011). Students’ performance per task-block served as the performance measure. Thus, we had six measurement points for all study variables per student. EDA was continuously recorded with E4 wristbands (Empatica, 2016) and analyzed with Continuous Decomposition Analysis (Benedek & Kaernbach, 2010).

Results. Mediation analysis taking the hierarchical structure (within-person measures) into account was conducted. The model indicates a positive direct effect of control on performance (b=.257, SE=.067, p<.001). Control was significantly negatively related to all of the four anxiety components (self-report). However, higher control was associated with higher EDA. Furthermore, there was a significant indirect effect of control on performance only via the cognitive anxiety component (i.e., mediating effect, b=.104, SE=.045, p<.05).

Significance. As hypothesized, the results indicate that the anxiety components indeed differ with regard to their relative impact on performance in such a way that the cognitive component was the only component for which we found a mediating effect. Therefore, consistent with our assumptions, this component seems to be central for performance. The relation between control and EDA needs further investigation, but might eventually be explained by high control being also associated with positive activating emotions leading to high arousal. Taken together, the results from this multimethod study provide evidence that it is important to distinguish between the components and focus anxiety interventions on the cognitive component in order to increase students’ performance.

Authors

• Anna-Lena Roos, University of Konstanz

• Thomas Goetz, University of Konstanz

• Maike Krannich, Universität Konstanz

• Amanda Jarrell, McGill University