| |
In lugnet.admin.general, Todd Lehman writes:
> > > In other words,
> > > a score of -5 (or -50) on a scale of [-10,+10] (or [-100,+100]) much more
> > > accurately communicates negativity than does the equivalent score of 2.5
> > > (or 25) on a scale of [0,10] (or [0,100]), yes?
> >
> > Not necessarily. It's logical, but people are used to things being
> > rated starting at 0. Then again, people are not used to having combined
> > 'good' and 'bad' ratings. Usually, ratings indicate *either* how 'good' or
> > 'bad' something is, not both.
>
> Well, if you grow up in Minnesota, yer usta hearing temperatures like "minus
> forty" (bad) and "ten above" (good). :-)
>
> Most movie rating scales are what?-- 0 to 4, or 0 to 5, sometimes with 1/2's?
>
> Then there's the two-thumbs-up (+2) and the two-thumbs-down (-2).
>
> (I get a big bang out of Mr. Cranky's[1] inverted scale, BTW... :-)
OK, a couple of more thoughts on the [-100,+100] vs. [0,100] choice...
I'm finding it increasingly difficult to defend the position of [-100,+100]
due to a number of reasons. One which I don't think came up yet is how a
new or casual user might feel if his/her message were marked down slightly.
That is, if someone's message is marked down a bit below average, then how
does that appear to them?
[-100,+100] [0,100]
============= =============
-100 0
-80 10
-50 25
-40 30
-20 40
-10 45
0 50
+10 55
+20 60
+40 70
+50 75
+80 90
+100 100
In the [-100,+100] view of things, anything below -50 or so was probably
deserving of that negative of a score -- and the poster might even feel
good about it if the intention was to elicit flamage or ill-will.
But I'm not worried about -50's or -80's. I'm worried about values like
-10 and -20 in the [-100,+100] view of things. Those are probably very
likely to happen from time to time -- especially to newbies -- and it would
be a real shame if someone posted something innocently and had it marked
down in such a way that caused them grief.
Though -10 and 45 are equivalent, and -20 and 40 are equivalent, I think
perceptually -- psychologically -- it "feels" a lot worse to have your post
marked down from 0 to -10 or -20 than from 50 to 45 or 40. (Does that make
any sense?)
If this is true (and I think it is), then this would be a MAJOR shortcoming
and unintentional side-effect of using the [-100,+100] view. We need that
articles can be marked up or down from a neutral position, but not in such a
"painfully obvious" way -- because, after all, the ultimate goal of the
ranking is just to help filter out things at different user-selectable
thresholds -- it's not to penalize or demoralize anyone.
So -- a list of advantages & disadvantages:
[-100,+100]
Advantages:
* Nice big integer domain with 201 possible values
* Midpoint is 0 (very cool)
* Good intuitive quick grokkage for integer math geeks
Disadvantages:
* -10 and -20 sound more demoralizing than 45 and 40
* Hyphen character for minus is hard to discern, especially when plus
is also used
* Probably overly complex for non-geeks
[0,100]
Advantages:
* Nice big integer domain with 101 possible values
* Midpoint is an integer
* 45 and 40 don't sound as demoralizing as -10 and -20
* Simple and familiar range -- looks like percentages
Disadvantages:
* Below average isn't so easy to spot
* Midpoint is 50 (not so bad, really)
* 0 may be confused with neutral when it actually means highly negative
(Other points?)
Most of these points are really "six of one, half dozen of the other" -- but
the thing about -10 == 45 and -20 == 40 really tips the scales in favor of
[0,100], IMHO... It's one thing to score LEGO sets or movies or record
albums on a [-100,+100] scale, but to score what people write about a hobby
that way -- that's just asking for hurt feelings. Yah?
--Todd
|
|
Message has 3 Replies:
 | | Re: Article scoring
|
| (...) <snip> (...) Yes, it makes a lot of sense, and I agree. (...) Right. (...) Meaning...? (...) I wouldn't be too sure it's so complex... then again, I'm a math geek so I wouldn't know ;-) (...) The last two are, IMO, HUGE advantadges for the (...) (27 years ago, 10-Mar-00, to lugnet.admin.general)
| | | Re: Article scoring
|
| On Fri, 10 Mar 2000, Todd Lehman (<Fr6srt.IrF@lugnet.com>) wrote at 03:38:17 (...) I think this has been at the back of my mind all along. I don't use the web i/f, so I've mostly avoided commenting. (27 years ago, 10-Mar-00, to lugnet.admin.general)
| | | Re: Article scoring
|
| Todd: [ ... ] (...) Although I find the [-100,+100] range more intuitive, I think you are right that [0,+100] is the range to use for article scoring. Play well, Jacob (who never claimed to be ordinary ;-) ---...--- -- E-mail: sparre@cats.nbi.dk -- (...) (27 years ago, 12-Mar-00, to lugnet.admin.general)
|
Message is in Reply To:
| | Re: Article scoring
|
| (...) That would rock!! (...) I don't think the HTML would be too heavy, but it could get pretty nasty on the web browser (depending on how many buttons there were per article). I seem to remember the Fibblesnork LEGO Survey pages (which each have (...) (27 years ago, 9-Mar-00, to lugnet.admin.general)
|
52 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
