Abstract:

Procrastination affects virtually everyone to some degree. It involves knowing that a task must be performed, yet intentionally failing to motivate oneself to carry it out within the established deadline. [1] described it as “the act of needlessly delaying tasks past the point of discomfort”. But definitions of procrastination vary among authors. As pointed out by [2], whereas a stream uses the term procrastination for dysfunctional forms of delay (e.g., [3]), others also adhere positive forms of delay to this term (active procrastination) (e.g., [4]).

Academic procrastination can be defined as “to voluntarily delay an intended course of study-related action despite expecting to be worse off for the delay” [5]. With the promotion of new teaching and learning methods, where the student is at the centre of learning process, it is important to analyse how and why students delay their academic duties [6]. Many studies have found that academic procrastination is negatively related to academic performance. However, the results are yet inconsistent and other studies fail to find this link. [7] carried out a meta-analysis that revealed that the observed relationship was influenced by the choice of indicators to measure either procrastination (e.g., Procrastination Assessment Scale-Students, Decisional Procrastination Scale or Active Procrastination Scale, among others) or academic performance. Besides, most studies rely on self-reported measurements of procrastination based on students’ responses to a behaviour-like questionnaire. [7] recommend researchers to devise alternatives to self-reported measures for indexing procrastination, because participants cannot evaluate themselves accurately and may bias the results.

Only a few studies have developed more objective instruments to measure procrastination (e.g., [8], [9], [10] and [11] that use delay in submitting mandatory tasks).

Timely detection with reliable data collection instruments can help to diagnose and predict procrastination. Ideally, it could also enable to anticipate corrective actions to avoid dysfunctional outcomes. This work proposes two different, data-driven approaches to measure academic procrastination, leveraging students’ activity reports in Learning Platforms. First, to calculate the average delay per student in accessing online items, as compared to first-access classmates. A considerable amount of materials (files, links, etc.) are available for students in the online platform. Thus, we propose an index that measures the average individual delay, but it is also corrected by the number of items viewed per student. Second, we propose to track the daily access performance during the last week before evaluation activities (midterms, assignments, etc.). The study is carried out in the Faculty of Economics and Business of the Universitat de Barcelona and it makes use of a dataset of 1,060 first-year business students covering the same course and equivalent evaluation system. In a data subsample preliminarily analysed, results show that, first, early-access students outperformed their counterparts in every evaluation activity. Second, students with better performance concentrate their online activity several days before deadlines, whereas lower performers concentrate their activity on the last day.

Keywords:

Academic procrastination, academic performance, learning analytics, educational data mining (EDS), higher education.