Abstract
<p>Statistics is regarded as one of the most difficult and frightening subjects by psychologystudents. The present study develops a Bayesian framework for a more comprehensiveunderstanding of students´ attitudes and affect towards statistics, their development andpredictors. Furthermore, an extensive collection of related constructs was raised and analysedpsychometrically. In a longitudinal design, data of first year psychology students was collectedvia online surveys at the beginning of the lectures, midst-semester and after course completion.Four Bayesian models were fitted to inspect response patterns, pairwise item correlations andconstruct change over time. In a final multilevel generalized linear varying effects model, theeffects of Statistics Anxiety on Affective Behaviours towards statistics were analysed.Psychometric analysis yielded surprising results, as most instruments revealed insufficientproperties in the present sample. Comparative analyses revealed that Affective Behavioursincreased over the course of the semester, and Statistics Anxiety decreased. The relationshipbetween Affective Behaviours and Statistics Anxiety was complex, as some subscales werecorrelated positively against expectations. In conclusion, the present study developed aBayesian framework for comprehensive data analysis, presents preliminary results of AffectiveBehaviours and Statistics Anxiety and discusses a potential reliability risk in the field due todeficient reporting of psychometric analyses.</p>