Date:         Thu, 10 Oct 2002 12:13:46 -0400
Reply-To:     "Snider-Lotz, Tom" <TSnider-Lotz@qwiz.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         "Snider-Lotz, Tom" <TSnider-Lotz@qwiz.com>
Subject:      Re: Adjectives Commonly Associated with Correlation Coeficients
Content-Type: text/plain; charset="iso-8859-1"

Yes, it really depends on the situation.  I'm involved in developing employment selection tests.  If the correlation between scores on two forms of the same test was .3, we'd be alarmed; but if the correlation between test scores and a job performance rating was .3, we'd feel OK about it.  As Nels said, Weak-strong would have more general applicability than Poor-Excellent.

Often users reserve judgment until they've taken a step beyond the correlation coefficient.  They might find more value in computing such things as variance accounted for, reduction in error of estimation, or financial implications.

>-----Original Message-----
>From: Nels Tomlinson [mailto:nels_tomlinson@labor.state.ak.us]
>Sent: Thursday, October 10, 2002 11:51 AM
>To: SPSSX-L@LISTSERV.UGA.EDU
>Subject: Re: Adjectives Commonly Associated with Correlation
>Coeficients
>
>
>Strong and weak might be better adjectives for describing
>correlation.  They
>are the ones I normally use.
>
>Are adjectives really the way to go, though?  There are two
>questions one
>can ask about a reported correlation: ``Is it statistically
>significant?''
>and ``Is it meaningful (in terms of the subject discipline)?''.  A
>correlation of 0.3 which is both statistically significant and
>meaningful
>might be described as weak by some scale of adjectives, but
>that might not
>be the best way to describe it.
>
>Hope this helps,
>Nels Tomlinson
>
>-----Original Message-----
>From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]On Behalf Of
>R. Allan Reese
>Sent: Thursday, October 10, 2002 12:23 AM
>To: SPSSX-L@LISTSERV.UGA.EDU
>Subject: Re: Adjectives Commonly Associated with Correlation
>Coeficients
>
>
>On Wed, 9 Oct 2002, Judy Brown wrote:
>> appropriate adjectives to assign to correlation results.
>>    For example, something like...
>> .9 and better might be described as excellent
>> .8 as good
>> .6 -. 7 as moderate
>> .5 as low
>> .4 and below as poor
>
>Whether these descriptions are appropriate depends on a value
>judgment of
>the association.  As a deliberately vivid example, the serial
>sniper will
>know how many bullets s/he has fired and what the hit rate has
>been, but
>s/he and I would differ on whether to call the association good.
>
>R. Allan Reese                       Email: r.a.reese@gri.hull.ac.uk
>Associate Manager                    Direct voice:   +44 1482 466845
>Graduate Research Institute          Voice messages: +44 1482 466844
>Hull University, Hull HU6 7RX, UK.   Fax:            +44 1482 466436
>====================================================================
>Hull University: one of the "Access elite" (THES 18/1/02)
>Widened access / Low drop out / Excellent teaching / Excellent research
>...
>Hull University: one of the hardest funding cuts by HEFCE (THES 3/3/02)
>