I think I understand: Investigating misconceptions regarding hypothesis test concepts among university students

Abstract

Hypothesis testing is an essential tool among researchers and practitioners alike, with its use being being widely taught in many a programme at university level. However, past studies have shown that students hold misconceptions about important statistical concepts. This present study aims to reconfirm past efforts in this area, specifically in a South East Asian higher education institution. To test how well undergraduate university students’ understood key concepts in hypothesis testing, an online multiple choice questionnaire was deployed. The questionnaire also asked for students’ confidence ratings for each question, allowing us to distinguish the confident versus non-confident incorrect responses. A follow-up interview was then conducted to give deeper insights into reasons behind respondents’ errors. The main finding is that there are significantly more confident wrong answers than non-confident ones – highly indicative of the presence of misconceptions among respondents. Among them, students firmly believed that statistical inference procedures provide a direct calculational proof of the null hypothesis. Additionally, students have difficulty formulating correct hypotheses to be tested, and have poor grasp of the role of signficance levels in hypothesis testing. Whether or not students were taking a quantitative-focused programme, or had prior statistics training, had no bearing on their survey score. Despite this, confidence ratings were significantly higher in both groups.

Haziq Jamil

Assistant Professor in Statistics

My research interests include statistical theory, methods and computation, with applications towards the social sciences.