Feedback ORM Exam January 2015
Feedback ORM Exam January 2015
Overall, students' performance on the Term 1 ORM exam was quite good. The average mark was 56.11. While no marks above 80 were recorded, a sizeable number of students achieved a mark of 70 or above. Students that received a mark below 40 tended to have not meaningfully attempted one or more questions. Indeed, a number of students did not to adequately manage their time in the exam and failed to provide full answers to each question. Feedback from some students after the exam confirmed this. This is surprising as the exam format and length was very similar to those in previous years (indeed the breadth of content was actually narrower this year). However the team will look into whether some adjustments may be made next year. One idea is the publishing the research scenario prior to the exam so that less of the hour in the exam will be spent on reading.
Turning now to specific feedback on the exam questions, the first question asked students to report and interpret some correlation and logistic regression output from SPSS.
Good answers described the various correlations between the study variables and OUTCOME (correlation coefficient, strength, direction and significance), including an explanation of what each association meant in words. Most answers described the logistic regression findings including pseudo R-square and chi-square, identified the significant predictors and their relative importance, and interpreted the odds ratios for the significant predictors. The best answers gave a complete interpretation of the findings and provided some critical commentary on the implications / nature of the tests (e.g. the difficulty in interpreting correlations relating to the dichotomous variable OUTCOME). Poorer answers failed to report the correlation results. Marks were also lost for not reporting the statistics fully or for interpreting the odds ratio incorrectly (e.g. ignoring the recalculated odds ratio or failing to observe
Overall, students' performance on the Term 1 ORM exam was quite good. The average mark was 56.11. While no marks above 80 were recorded, a sizeable number of students achieved a mark of 70 or above. Students that received a mark below 40 tended to have not meaningfully attempted one or more questions. Indeed, a number of students did not to adequately manage their time in the exam and failed to provide full answers to each question. Feedback from some students after the exam confirmed this. This is surprising as the exam format and length was very similar to those in previous years (indeed the breadth of content was actually narrower this year). However the team will look into whether some adjustments may be made next year. One idea is the publishing the research scenario prior to the exam so that less of the hour in the exam will be spent on reading.
Turning now to specific feedback on the exam questions, the first question asked students to report and interpret some correlation and logistic regression output from SPSS.
Good answers described the various correlations between the study variables and OUTCOME (correlation coefficient, strength, direction and significance), including an explanation of what each association meant in words. Most answers described the logistic regression findings including pseudo R-square and chi-square, identified the significant predictors and their relative importance, and interpreted the odds ratios for the significant predictors. The best answers gave a complete interpretation of the findings and provided some critical commentary on the implications / nature of the tests (e.g. the difficulty in interpreting correlations relating to the dichotomous variable OUTCOME). Poorer answers failed to report the correlation results. Marks were also lost for not reporting the statistics fully or for interpreting the odds ratio incorrectly (e.g. ignoring the recalculated odds ratio or failing to observe