I need to run an analysis for an assignment where we have the amount of time someone spent playing a video game, each time they logged into the game and what type of game it was. I need to figure out how long each gamer spent on each game for each login. Any ideas on what sort of statistics I would use to solve this?
2
-
$\begingroup$ I need more info about your assignment, you should be more precise on what do you want to estimate. Do you need to fit a statistical model? If the answer is yes, you probably want to fit a hierarchical model to estimate how long on average each gamer etc. $\endgroup$– Prunus aviumCommented May 11, 2018 at 2:12
-
$\begingroup$ You have the amount of time spend playing and you need to figure out the amount of time spent playing. That is your first and second sentence. It's often very hard to understand the question, and then you'll find the answer quickly. $\endgroup$– GijsCommented May 11, 2018 at 14:30
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Time spent playing a video game is a continuous variable (from 0 to maximum time) and is your outcome.
The type of game played could be your predictor and is a categorical variable e.g. action, platform, RPG, shooter, etc.
Your null hypothesis could be that game type is not associated with time spent playing.
Your alternative hypothesis could be game type is associated with time spent playing.
You can use ANOVA to test the null hypothesis above. If p-value for game type is small enough, let's say < 0.05, then you can say with 95% confidence that game type is significantly associated with time spent playing.
However, you need to check that time spent playing is normally distributed to be able to use ANOVA. Times are often not normally distributed.
