Determining the effect of test anxiety on academic achievement in Singapore

Abstract

This paper assesses the effect of test anxiety on academic performance in Singapore by using data from the PISA 2015 survey. Prior research in educational psychology has concluded that test anxiety has a negative causal relationship with academic performance. Based on a review of existing literature, this study is the first of its kind studying the relationship in Singapore using econometric methods to correct for both measurement error and omitted variable bias. To generate estimates robust to attenuation bias arising from measurement error, we use Error-in-Variables regression. We also address omitted variable bias using the method proposed by Altonji et al. (2000) and Emily Oster (2016) which is based on the use of a general estimator based on the proportional selection of the importance of unobservables relative to observables.

The results show that the OLS estimates derived are robust to attenuation bias but are highly sensitive to the extent of omitted variable bias. This was consistent even after accounting for gender differences. Our research raises concern that the results of the prior research could have been influenced by the nature of the omitted variable bias.