Thank you for your comments,  Garry and Rich, My friend didn't anticipate that Cronbach alpha is not the appropriate for his forced ranking questionnaire.  The data were already gathered and no other choice but to use the data.  Can you suggest an alternative? Thank you.

--- On Fri, 7/22/11, Rich Ulrich <rich-ulrich@live.com> wrote:

From: Rich Ulrich <rich-ulrich@live.com> Subject: Re: Reliability coefficients for continuous scores ranging between 0 and 100 To: SPSSX-L@LISTSERV.UGA.EDU Date: Friday, 22 July, 2011, 7:49 PM

 - and that is a secondary reason why it is generally a bad idea to elicit data scores as forced rankings, or other methods that add up to a fixed sum.

(The main reason is that you don't ask for any anchor of absolute good or bad, which you usually want to know.)

-- Rich Ulrich

Date: Fri, 22 Jul 2011 11:09:15 +0100 From: garry@business-analytic.co.uk Subject: Re: Reliability coefficients for continuous scores ranging between 0 and 100 To: SPSSX-L@LISTSERV.UGA.EDU

#yiv1273072921 .yiv1273072921ExternalClass p.yiv1273072921ecxMsoNormal, #yiv1273072921 .yiv1273072921ExternalClass li.yiv1273072921ecxMsoNormal, #yiv1273072921 .yiv1273072921ExternalClass div.yiv1273072921ecxMsoNormal {margin-bottom:.0001pt;font-size:12.0pt;font-family:'serif';} #yiv1273072921 .yiv1273072921ExternalClass a:link, #yiv1273072921 .yiv1273072921ExternalClass span.yiv1273072921ecxMsoHyperlink {color:blue;text-decoration:underline;} #yiv1273072921 .yiv1273072921ExternalClass a:visited, #yiv1273072921 .yiv1273072921ExternalClass span.yiv1273072921ecxMsoHyperlinkFollowed {color:purple;text-decoration:underline;} #yiv1273072921 .yiv1273072921ExternalClass span.yiv1273072921ecxEmailStyle17 {font-family:'sans-serif';color:#1F497D;} #yiv1273072921 .yiv1273072921ExternalClass .yiv1273072921ecxMsoChpDefault {} _filtered #yiv1273072921 {} #yiv1273072921 .yiv1273072921ExternalClass div.yiv1273072921ecxWordSection1 {} Eins Your four items are not independent because they sum to 100.  You are possibly getting a negative average covariance because a high average score on any three items necessarily leads to a low score on the remaining item.  Cronbach’s alpha is not appropriate for this kind of data. Garry From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of Eins Bernardo Sent: 22 July 2011 10:36 To: SPSSX-L@LISTSERV.UGA.EDU Subject: Reliability coefficients for continuous scores ranging between 0 and 100 Dear all, I am posting a question regarding reliability analysis in behalf of my colleague.  His instrument has four subscales and each subscale has four items. The respondents are asked to use 0 to 100 to rate each item within each subscale such that the sum of the four items is 100.  Can we use the Cronbach alpha for the reliability at each subscale level?   For your further info we tried to compute the cronbach alpha of the first subscale with n=42 using spss.  The alpha is negative and out of range.  Please see spss output below: Thank you. Eins. [snip, table]