One Measure Can’t Capture It All : Comparing Different Assessment Methods of Anxiety in Learning and Achievement Situations

The current dissertation is based on three related Research Papers, in which different dimensions of assessing anxiety in learning and achievement situations, namely trait and state self-report assessments of anxiety (Paper I), the different anxiety components (Paper II) and physiological measures of anxiety (Paper III) are examined and compared with each other concerning their results. The first study investigated the discrepancy between generalized trait and situation-specific state self-report measures of anxiety in mathematics in a sample of students with high and low achievement levels. Previous studies have revealed that trait anxiety measures are typically rated higher than state measures. In addition, the academic self-concept has been identified to play a moderating role in the so-called trait-state discrepancy, with a higher academic self-concept leading to a lower discrepancy (i.e., less overestimation of trait anxiety when state assessments reflect actual experience). Therefore, it was predicted in this study that high achievers, due to their higher academic self-concept in mathematics, exhibit a smaller trait-state discrepancy in anxiety than students with low achievement levels. The results were in line with this hypothesis and demonstrated that high achievers even underestimated their anxiety when judging it from a trait as compared to a state perspective. Taken together, this study shows that state and trait self-report measures of anxiety can yield different results and that even high achievers experience some anxiety in class. The second study investigated the different components of test anxiety (i.e., cognitive, affective, physiological, and motivational components) using an intra-individual approach. The first aim of this study was to examine the relative impact of the different components of anxiety in the mediating mechanism that connects control, anxiety and performance - as proposed by the control-value theory. Thereby it sought to identify which component should be primarily addressed by anxiety interventions and instructional techniques in order to improve students’ performance. A second aim of this study was to uncover which anxiety component underlies the commonly used and very convenient single-item anxiety measures. In line with the expectations, the anxiety components indeed differed in their relative impact on test performance, with the cognitive component being central for this outcome. Furthermore, results indicated that behind a single-item anxiety measure we most likely find the affective anxiety component and thus, not the component most relevant for performance. Therefore, it is important to consider and differentiate between the anxiety components and select the assessment method based on the research question at hand. The third study used the methodological approach of a systematic review to investigate the relationship between traditional self-report measures of test anxiety with another type of measure, namely physiological measurement. This assessment method has gained increased attention in pedagogical psychological research in recent years. Results of the systematic review showed that in line with theoretical considerations, self-report measures of test anxiety and physiological measures (e.g., cardiovascular measures, electrodermal measures, cortisol sampling) were significantly correlated in the majority of the reviewed studies. The correlations were moderate, suggesting that although the two measures overlap, they should not be used interchangeably. In conclusion, each of these measures provides information about a person’s anxiety and applying both of the two measures might contribute to a better measurement and thus also a better understanding of this often detrimental emotion. Furthermore, variables that can influence the relation between self-report and physiological measures were identified (e.g., sampling rate, choice of physiological measure and the way the data is analyzed) and are discussed with regard to a successful assessment of physiological data in future research. Taken together, from the results of these three Research Papers it can be concluded that it is important to distinguish between different ways of assessing anxiety and to select the assessment method according to the research focus. Besides its implications for research, the present thesis provides novel perspectives on how to successfully influence anxiety in learning and achievement situations. published