Ratings based on type of episode

This is an abbreviated version of  a statistics project I did for class.

For my statistics project last semester we were able to pick any data set we wanted (pending professor approval) and found interesting. Was hilarious when all three of the Education graduate students did the same topic. I decided to use the ratings of survivor episodes and was surprised by what I found.

I don’t know about anyone else but I’ve always assumed that premiers and finales have higher ratings that the rest of the season.


Little visual difference between group means.

But apparently not. The visual is backed up by a simple one-way ANOVA analysis, which gives a p-value of .085 and thus does not provide statistically evidence for a difference based on episode type. This was actually not my original question, but surprised me quite a bit.

I’m curious to see if this is true for other shows as well, or maybe this is caused by of some aspect of Survivor.