I think the purpose of the rating system motivates how people rate. Since the App Store appears to give importance to mean average data, I am given motivation in that direction:
If I think the app is good, and other people should download it, my rating makes the most difference if I award a score of 5.
If I think the app is OK but not worth downloading, I can most effectively indicate that by scoring it a 1.
If I want people to read my review (as Turkish reviewers have observed), I should perhaps rate it a 1 as well.
So, if I want my reviews to have an observable effect, I should (perhaps) rate everything 1 or 5.
(Or, I could decide what rating I want the app to have, then rate it a 1 or 5 depending on which direction I need to move it)
The motivation initiated by Apple here is bad - by simplifying everything to a single figure average, we lose all the nuance that you are trying to give with your Norwegian-style ratings.
> (Or, I could decide what rating I want the app to have, then rate it a 1 or 5 depending on which direction I need to move it)
You just re-discovered strategic voting. And that's why the rating system should display the median and not the mean. For the median, you can skip step two: just vote what you want.
If I think the app is good, and other people should download it, my rating makes the most difference if I award a score of 5.
If I think the app is OK but not worth downloading, I can most effectively indicate that by scoring it a 1.
If I want people to read my review (as Turkish reviewers have observed), I should perhaps rate it a 1 as well.
So, if I want my reviews to have an observable effect, I should (perhaps) rate everything 1 or 5.
(Or, I could decide what rating I want the app to have, then rate it a 1 or 5 depending on which direction I need to move it)
The motivation initiated by Apple here is bad - by simplifying everything to a single figure average, we lose all the nuance that you are trying to give with your Norwegian-style ratings.