Coming up with a grade, be it in tenths or 1-to-5 stars, is not easy. If you lurk around Google Play, you’ll notice that apps rated 4 or better tend to polarize the vast majority of votes on 5, 4, and 1 stars. There’s significantly less people giving 3 or 2 stars.
That is easily explained, because humans tend to see everything as either black or white, that is, all 5 and 1 star votes. 5 stars is “Perfect”. 4 stars are something like “It’s a 5, but there’s this small thing that annoys me, so it’s a 4”. 3 stars is the equivalent of “This is average”. 1 star is “This sucks”. 2 stars, well, it’s of not much use.
On Google Play, in a lot of cases apps have more +1s than 4/5-star reviews. Sometimes, +1s are even more than the total number of reviews.
I think that rating systems based on actual grades will gradually disappear from the web: both Google and Facebook push for “likes”. It’s sad, because a grade conveys more meaning than a mere number of “likes”, but on the other hand it’s harder to come up with.
We’d lose some degree of granularity, but a mixture of the two is probably the best option, and can be easily converted in numbers:
- Like: +2
- Average: +1
- Dislike: +0
Then it’s simply a matter to compute the average. Results above 1 are good, below are bad. I haven’t seen anything like that, but only systems based on thumb-ups/thumb-downs, which lack the Average option. Would this be too complex for the average web user?