The Benefit of a Switch: Answer‐Changing on Multiple‐Choice Exams by First‐Year Dental Students

Abstract

The aim of this study was to determine if dental students would benefit from changing their initial responses to what they have deemed to be more suitable answers during high-stakes multiple-choice examinations. Students are often advised to stay with their first answers despite evidence from other fields suggesting this is not the best course for obtaining optimal final exam scores. Data were collected for 160 first-year DMD students in fall 2013 for three operative dentistry and four biochemistry exams at Tufts University School of Dental Medicine. As students take all of their exams through ExamSoft, a test-taking software application that tracks and records all changes students make during the exam period, the subjective nature of previous studies on answer changing was eliminated. The results showed that all students changed their answers on a minimum of nine questions over the seven exams, with an average of 26.55 (SD=8.8) questions changed per student. Answers changed from an incorrect to a correct response comprised nearly 65% of total answer changes, while changes from a correct to an incorrect answer encompassed slightly above 10% of answer changes. Nearly all students (99.4%) benefitted from answer-changing with a net gain of at least two correct questions, with only one student not increasing the final score. Overall, the students greatly benefitted from changing their answer choice, suggesting that dental students could be advised to change their answers from their first choice if they identify a better option when taking multiple-choice exams.

Adaptive Weight Estimation of Latent Ability: Application to Computerized Adaptive Testing With Response Revision

“…Much of the research on test takers’ response revision behavior in a linear test has focused on investigating how response revisions influence the test performance. Consistent findings have been reported in these studies showing most test takers change responses; test takers typically change a small portion of responses on a test; most of the changes are from wrong to right responses and are likely to improve test scores (e.g., Al-Hamly & Coombe, 2005; Benjamin Jr. et al, 1984; Kruger et al, 2005; Liu et al, 2015; Pagni et al, 2017; Passos et al, 2007). In general, it is believed that response revisions are able to correct errors made by careless or speeded responding, and many studies support response revision when there is a good reason for doing so (e.g., Liu et al, 2015; Pagni et al, 2017).…”

“…Consistent findings have been reported in these studies showing most test takers change responses; test takers typically change a small portion of responses on a test; most of the changes are from wrong to right responses and are likely to improve test scores (e.g., Al-Hamly & Coombe, 2005; Benjamin Jr. et al, 1984; Kruger et al, 2005; Liu et al, 2015; Pagni et al, 2017; Passos et al, 2007). In general, it is believed that response revisions are able to correct errors made by careless or speeded responding, and many studies support response revision when there is a good reason for doing so (e.g., Liu et al, 2015; Pagni et al, 2017). In addition to the findings of the benefit of response revision on improving test performance, many studies across a number of disciplines examined factors that influence the response revision and test performance, such as gender (e.g., Al-Hamly & Coombe, 2005), test taker’s cognitive style (e.g., S.…”

Statistical Foundations for Computerized Adaptive Testing with Response Revision

“…For example, in a conventional paper-pencil test one can review and revise previous answers at any time during the test. Moreover, various empirical studies indicate that the opportunity to revise not only alleviates test-taking anxiety and is appreciated by test takers (e.g., Han 2005;Kruger et al 2015;Liu et al 2015), but it also allows the correction of accidental errors and as a result can improve the validity of test scores (e.g., Al-Hamly and Coombe 2005;Benjamin et al 1984;Pagni et al 2017).…”

Beyond item analysis: Connecting student behaviour and performance using e‐assessment logs

“…To our knowledge, the idea of using log data from an EAP to analyse exam items was first introduced by Neel's 1999 work, presented at the Annual Meeting of AERA (cited in Jung Kim, 2001). To date, exam logs have mostly been used for measuring and modelling exam‐takers' accuracy, speed, revisits and effort (Bezirhan et al, 2021; Klein Entink et al, 2008; Sharma et al, 2020; Wise, 2015; Wise & Gao, 2017); analysing answering and revising behaviour during exams (Costagliola et al, 2008; Pagni et al, 2017); examining and enhancing metacognitive regulation of strategy use and cognitive processing (Dodonova & Dodonov, 2012; Goldhammer et al, 2014; Papamitsiou & Economides, 2015; Thillmann et al, 2013); classifying exam‐takers towards testing services personalisation (Papamitsiou & Economides, 2017); validating the interpretations of test score (Engelhardt & Goldhammer, 2019; Kane & Mislevy, 2017; Kong et al, 2007; Padilla & Benítez, 2014; Toton & Maynes, 2019; van der Linden & Guo, 2008); understanding exam‐takers' performance (Greiff et al, 2016; Kupiainen et al, 2014; Papamitsiou et al, 2014, 2018; Papamitsiou & Economides, 2013, 2014); enhancing item selection in adaptive testing environment (van der Linden, 2008); analysing exam items (Costagliola et al, 2008; Jung Kim, 2001); detecting cheating (Cleophas et al, 2021; Costagliola et al, 2008); and identifying test‐taking strategies (Costagliola et al, 2008). Nonetheless, most of the previous work focused on time‐based behaviours and the interpretation of exam‐taker results; few of them examined the potential of using exam‐taker behaviours to validate or enrich the interpretation of the quality of exam items.…”