Gharib and Phillips (2013) compared students’ scores on open-book and cheat-sheet examinations. The predictor variable was the type of exam: open-book or cheat-sheet; nominal measurement. The outcome sets variables were:
- students’ grades (scale);
- anxiety scores (presumably scale; 5-point Likert scale was used);
- time of studying for exams (hours, scale). N=225.
The article is relevant for General Psychology because it sheds light on what methods are best for assessing students depending on the desired outcomes of a course.
Assumptions for paired samples t-tests:
- continuous scales of measurement of dependent variables;
- the independent variable consists of two categorical groups;
- no important outliers in differences between groups compared in the test;
- approximately normal distribution of the results.
Assumptions for Pearson’s r:
- continuous scales of measurement of variables;
- each participant has a pair of values;
- no important outliers;
- approximately normal distribution of the results (Warner, 2013).
Clearly, variables (1) and (2) are measured continuously because of the nature of these variables, thus meeting assumptions (a) and (A); this may be untrue for (2), which is based on the Likert-scale measurements.
Assumptions (b) and (B) are met because the same students completed the same quizzes (questionnaires about exams, anxiety scores, time of preparation), and their scores were compared.
Assumptions (c), (C), (d), and (D) are not discussed in the article, creating a limitation.
Questions, Hypotheses, α-level
The research questions were: for open-book and cheat-sheet types exams, (i) do students score differently?; (ii) do they have different levels of anxiety?; (iii) does their time of preparation differ? Also, it was asked: (iv) do students’ anxiety correlate with their grades?
Null hypotheses: for open-book and cheat-sheet exams, H0-i: students do not score differently; H0-ii: students have similar levels of anxiety; H0-iii: the time of preparation is similar. Also, H0-iv: students’ anxiety is uncorrelated with their grades during open-book exams; H0-v: student’s anxiety is uncorrelated with their grades during cheat-sheet exams.
Alternative hypotheses: for open-book and cheat-sheet types of exams: HA-i: there is a significant difference in students’ scores; HA-ii: students have different levels of anxiety; HA-iii: the time of preparation differs. Also, HA-iv: students’ anxiety is correlated with their grades during open-book exams; HA-v: students’ anxiety is correlated with their grades during cheat-sheet exams.
In all cases, α is not specified, but p is compared to.05, so α=.05
- For (1), t(220)=2.94, p<.05. Also, for (1), Pearson’s r(219)=.51, p<.05. Students had slightly but significantly higher grades on open-book exams; the difference in scores was approximately 2%. H0-i is rejected, HA-i is supported.
- For (2), t(209)=1.98, p<.05. Students were slightly but significantly more anxious during cheat-sheet exams. H0-ii is rejected, HA-ii is supported.
- For (3), t(164)=2.00, p<.05; on average, students studied 0.64 hours longer for cheat-sheet exams. H0-iii is rejected, HA-iii is supported.
There was a Pearson’s r(215)=-0.16, p<.05, between anxiety levels and grades for the cheat-sheet exam; for the open-book exam, there was no such correlation. H0-v is rejected, HA-v is supported. Pearson’s r, df, and p are not reported for the open-book exam and anxiety levels, but the authors state that H0-iv should not be rejected.
Effect sizes are not reported.
Therefore, students get somewhat higher grades during open-book exams; are less anxious during open-book exams; prepare less for open-book exams; their anxiety is correlated with a performance during cheat-sheet exams. This can be used to adjust course exams to the desired course outcomes.
No analyses of strengths or limitations of the tests are provided in the article.
Gharib, A., & Phillips, W. (2013). Test anxiety, student preferences and performance on different exam types in introductory psychology. International Journal of e-Education, e-Business, e-Management and e-Learning, 3(1), 1-6. Web.
Warner, R. M. (2013). Applied statistics: From bivariate through multivariate techniques (2nd ed.). Thousand Oaks, CA: SAGE Publications.