In my statistics class, I was helping the students understand the difference in one, two, and three or more sample (ANOVA) tests. When we reviewed the matched pairs t-test (tests where we have dependent samples), I decided to use the test scores from their first two exams to see if there were significant differences between the tests.
Now, this year was the first year that I gave a practice exam before the first test. That means that previous years, the first time a student saw what one of my exams looked like, was AT THE FIRST EXAM. The question one might ask is does giving students a practice exam help them.
So during class, I performed a two-sample, matched pairs t-test on the test scores between exam 1 and exam 2 for my Spring 2015 class and my Fall 2015 class. The results are shown below:
As you can see from the p-values, there was a significant difference in exam grades in Spring 2015 between the first and second exam – most likely because the first exam was lower because they had never seen an exam from me, and were unfamiliar with how I give exams.
However, in the Fall 2015 class, there was not a significant difference in exam grades – most likely because they had already seen a practice exam before their first exam, so they were familiar with the way I gave an exam.
I won’t bore you with the other statistical tests, but there was no significant difference in the second exam scores when I compared the Spring 2015 and Fall 2015 classes. Also, there was not statistical difference in the second exam in the Spring of 2015 and the first exam in the Fall of 2015 – meaning, classes performed equally well on any second exam they encountered, whether it was a real exam or a practice exam.
This leads me to conclude that the students performed equally well on the second exam because they already had experience with one of my exams.
So for me, I’m convinced that for Statistical Problem Solving in Geography, giving the students a practice exam helped them perform better.
This was a really fun way to step through the statistical tests for comparing two dependent and independent samples. A number of the students told me now I really understand how this stuff works. And ultimately, that is what it is all about!