Thu 18.09.2014 - 15:20-17:30 - Risba

Mathematically Highly Able Students’ Time Use in a Computer-Based Reasoning Test  Paper

Presenter: Risto Hotulainen
Author(s): Risto Hotulainen, (University of Helsinki, Finland), Sirkku Kupiainen, (University of Helsinki, Finland), Mari-Pauliina Vainikainen, (University of Helsinki, Finland)

Computer-based assessment (CBA) with the ensuing log data has raised Carroll’s (1963, 1989) concept of time-on-task again to the forefront of assessment research (e.g., Wise, 2006). In their study, Goldhammer et al. (2014) found that higher achieving students need less time for good performance in routine tasks but use more time for better attainment in more complex tasks. In this study, we explore the time use of Finnish mathematically high performing 9th students as compared to their less able peers in a low-stakes CBA measuring verbal and mathematical reasoning. Using CBA log data, two questions were addressed: 1) Do students showing higher ability in a high-stakes curricular mathematics test excel equally in the concurrent low-stakes CBA? and 2) Do the high ability students differ from their peers in their time-on-task in the low-stakes CBA? To answer the questions, a structural equation model (SEM), construed to explain students’ performance in the CBA by prior attainment (GPA), motivational attitudes disclosed in a simultaneously administered self-report questionnaire, and time-on-task as a mediating factor was fitted to a nationally representative data of 4249 Finnish ninth graders (mean age 15.9 years) using the whole cohort (cf. Kupiainen et al., 2014) as a reference point. Preliminary results show that the high ability students spent on average more time on the CBA tasks than their less able peers. The difference was bigger in mathematical than in verbal reasoning (for incremental groups of 10 % in the curricular math test, the effect size was eta2=.084 vs. eta2=.034). Differences between the groups were bigger in the reasoning tests (eta2=.255 vs. eta2=.241), indicating that the more time spent on the tasks paid dividends in the form of better results. The presentation will centre on the fitting of the SEM model on the more able vs. average and weak students.