Hi everyone! Lately I have been trying to improve my typing speed, and have been playing a game called Type Race, where you type various short passages (the passages are selected at random from a text bank) and your score in WPM is recorded. What I want to do is determine whether or not my typing speed has improved as a result of playing the game. In particular, I want to test the hypothesis that my scores have increased since I first started playing, statistically speaking. Right now I have completed 2821 races, with mean of 102.4 wpm and a standard deviation of 11.14 wpm. I have all the results saved in an Excel Spreadsheet I graphed my results with a histogram and noticed that my “population” of scores is quite normal. So, given this model of what I want to, what would be the strongest significance test to run on this data? With my very limited knowledge of statistics, I was thinking I could two random samples, one from the first half of my scores, and another from the second half, and testing the hypothesis that there will be a significant increase in wpm in minute. (In other words, the mean of the first half is larger than the second half). However, I think that there are some issues with this, since the initial scores have a direct influence on the later scores due to improvement from practice. In other words, the post “treatment” results were a direct result of having completing the “pre treatment” results in the first place. Sorry if that made NO SENSE whatsoever, again I’m a statistics newbie. What I’m getting at is, if this is case, wouldn’t it be more appropriate to run some kind of test that takes the pooled variance into account? Granted I would prefer to avoid analysis of variance if possible, but again I’m looking to obtain the most statistically robust results possible. Thanks for any help!