Hi, since I can't leave comments on you nice post, I'm doing it here.
* Thanks for putting up 'long form' posts which I guess what the kids are calling regular articles which take up more than 130 characters.
* Thanks for writing about your work, which I think more people should do.
* For the five star rating, if the mean is not 3 then the distribution can't be normal. It has to be skewed.
* Thanks for making the nice graphs, but please take pity on the color challenged and use much brighter and color blind friendly colors (or simply use gray scale, which would work just fine for your data).
Again, nice work on doing market research and putting up your findings for us to see.
I especially like that you develop a hypothesis at the end of your data description.
You're totally right that ratings datasets will skew positive. I was trying to get really get across that they aren't bimodal at all (unlike what some other people have claimed) and tend to look somewhat normal if you squint hard enough.
Just an observation: it is possible for the data to be normally distributed with a mean higher than 3, that just means that the distribution is truncated closer to the median on the upper tail than the lower one.
However, just from eyeballing his chart the data in this case do appear to be slightly left-skewed.
* Thanks for putting up 'long form' posts which I guess what the kids are calling regular articles which take up more than 130 characters.
* Thanks for writing about your work, which I think more people should do.
* For the five star rating, if the mean is not 3 then the distribution can't be normal. It has to be skewed.
* Thanks for making the nice graphs, but please take pity on the color challenged and use much brighter and color blind friendly colors (or simply use gray scale, which would work just fine for your data).
Again, nice work on doing market research and putting up your findings for us to see.
I especially like that you develop a hypothesis at the end of your data description.