Date: Wed, 16 Jun 1999 08:34:14 -0400
Reply-To: John <jschinka@COM1.MED.USF.EDU>
Sender: "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From: John <jschinka@COM1.MED.USF.EDU>
Subject: Re: Scale construction
Content-Type: text/plain; charset=iso-8859-1
A study by Velicer several years ago demonstrated that an even-numbered
(e.g., 8, 10, 12) set of options was slightly more reliable than an
odd-numbered set. The reason for this is that an odd-numbered set
contains a mid, or neutral point, that will often be selected as a
“safe” response. Removal of the neutral point forces the respondent to
respond on either the negative or positive side of the response set,
resulting in an increase in response variance. Of course, elimination
of the neutral point does suggest that the scale builders believe that
there can be no truly neutral response to an item.
John A. Schinka
Lou Sternberg wrote:
> I realize this is a scaling question rather than an SPSS analysis
> question, but I would appreciate any comments you would care to send
> my way. I have some colleagues who would like to use the 10-point,
> asymmetric question format below. Does this format have any advantage
> over an 11-point, symmetric scale of 0 to 10?
>
> "Now I would like to ask you some questions just about **********.
> Using a 10 point scale where 10 means "Completely Satisfied" and 1
> means "Completely Dissatisfied", with 5 being "neither Satisfied nor
> Dissatisfied", please tell me . . ."Thank you in advance, Lou
> ************************************************************
> Lou Sternberg, Ph.D. Phone: 208-331-8862
> Corporate Research FAX: 208-331-8884
> Blue Cross of Idaho email: lsternberg@bcidaho.com
> Post Office Box 7408
> Boise ID 83707
> ************************************************************
| 
