David Gerard wrote:

Ray Saintonge wrote:

...

Magnus Manske wrote:

...

I still prefer a 0-10 range of ratings. I think a decimal normalization would be easier to work with in any subsequent analysis of results.

One can set the range for each topic individually.

Mmm. See discussion at [[m:En validation topics]] and its archive - too many choices of rating is probably a bad thing, because it's hard to agree what a given value means. The test plan so far includes probably far more variables than we'd want in any case ...

The only variable is the individual subjective valuation of the article. The number of points in the range are not different variables, but different values for the same variable. To a large extent the size of the range is arbitrary because one range can be changed to another by a simple multiplication with a constant. What's the difference between a value of 3 on a 1-5 scale and a value of 5 on a 0-10 range?

Since the value which an individual gives to an article is essentially subjective there is no reliable way to define the ratings that can be given. Statistical results tend to be self-normalizing.

Ec