Optimizing Test Scores

Within this social experiment for my sophomore statistics class, my team approached the question of which factors contributed to the highest test scores. We created a sample test by using questions from practice ACT tests. The explanatory variables of interest that we focused on were Medium (paper or online), Font (print or script), Timing (timed or untimed), and Question Type (multiple choice or short answer) to explain the response variable of test scores. When we implemented our experiment, we also blocked on Gender to eliminate the possibility of it being a confounding variable.

Our results showed that the most statistically significant variables were Time and Gender. In other words, the average test score of students who were able to take the test in as much time as they wanted were significantly higher than the average test score of students who took the test with a time limit.

In addition, the most interesting result was a significant interaction between the variables Font and Question Type. By themselves, these two variables were not statistically significant; however, when they were combined, their interaction was statistically significant.