Q:About user profile ratings in sci-math google.
in [Math]
|
Prev: A Predicart Test for Hypothesis / Proof to follow posted by another author/machine I do not control
Next: Building the confidence inteval for the ratio of 2 measures
From: dan73 on 4 Oct 2009 03:24 What gives with these 5 star profile ratings where any one rating for a post can be 1 too 5! Who are they rated by? Are they rated by the value of the post or the average value of all posts'? ?? Dan
From: hagman on 4 Oct 2009 07:51 On 4 Okt., 13:24, dan73 <fasttrac...(a)att.net> wrote: > What gives with these 5 star profile ratings > where any one rating for a post can be 1 too 5! > > Who are they rated by? > > Are they rated by the value of the post or > the average value of all posts'? > > ?? > > Dan Obviously, this question should rather be posted in a newsgroup about Google Groups.
From: dan73 on 4 Oct 2009 04:45 >> What gives with these 5 star profile ratings >> where any one rating for a post can be 1 too 5! > >> Who are they rated by? > >> Are they rated by the value of the post or >> the average value of all posts'? > >> ?? > >> Dan Hagman wrote: >Obviously, this question should rather be posted in a >newsgroup about Google Groups. Sorry, you are right. My original thought it was math related because to rate a profile about math related posts, the person(s) rating the profiles would have to be experts in that field, but I get your point. Dan
From: Doug Freyburger on 4 Oct 2009 18:53 dan73 wrote: > What gives with these 5 star profile ratings > where any one rating for a post can be 1 too 5! Because a one star rating is common for trolls and spammers, there is less meaning to the numbers 2-4. Thus 5 ends up suggesting value and it's nearly binary much of the time. > Who are they rated by? Anyone who reads using google, so a random sampling of readers not posters. For the years I read on google I rated plenty of posts. My account remains active so I might end up rating more. > Are they rated by the value of the post or > the average value of all posts'? Google displays the average of the ratings of that posting, but few posts get rated by many readers. If you agree with the rating there's little point in repeating it with the same polarizing effect I mentioned. So in the end the ratings of any one post are too random to matter much. Whether the final average for a poster who's received hundred of ratings is more valuable is a matter of tastes - Do you care if hundreds of google readers give you some number of stars? It's definitely easy to identify trolls and spammers in the few cases they behave themselves on a single group but if that poster is behaving on that group it hardly matters.
From: Mensanator on 4 Oct 2009 20:23
On Oct 4, 5:53�pm, Doug Freyburger <dfrey...(a)yahoo.com> wrote: > dan73 wrote: > > What gives with these 5 star profile ratings > > where any one rating for a post can be 1 too 5! > > Because a one star rating is common for trolls and > spammers, there is less meaning to the numbers 2-4. > Thus 5 ends up suggesting value and it's nearly > binary much of the time. > > > Who are they rated by? > > Anyone who reads using google, so a random sampling > of readers not posters. �For the years I read on > google I rated plenty of posts. �My account remains > active so I might end up rating more. > > > Are they rated by the value of the post or > > the average value of all posts'? > > Google displays the average of the ratings of that > posting, but few posts get rated by many readers. > If you agree with the rating there's little point > in repeating it with the same polarizing effect I > mentioned. Little point? Don't know any mathematics, eh? Sure, a post that's already rated 3 won't change if you also rate it 3, but it will become deeper entrenched. > > So in the end the ratings of any one post are too > random to matter much. �Whether the final average > for a poster who's received hundred of ratings is > more valuable is a matter of tastes - Do you care if > hundreds of google readers give you some number of > stars? �It's definitely easy to identify trolls and > spammers in the few cases they behave themselves on > a single group but if that poster is behaving on > that group it hardly matters. |