Date:         Tue, 2 Aug 2005 23:17:10 -0700
Reply-To:     gokhan kaya <kayagk@yahoo.com.ar>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         gokhan kaya <kayagk@yahoo.com.ar>
Subject:      scale development and weghting
Content-Type: text/plain; charset=iso-8859-1

Dear List



I have dealt with scale construction but I did not overcome weighting

I have  20 variables with likert scale

After applying  EFA, I can  obtain 3 factors and save factor scores

Then I try  to obtain weighted relations, equations to calculte unique scale(score) to represent a theoritical concept



In a month ago, I applied categorical regression on  this saved factor scores and variables but their scales are not same

Factor scores are between 0 and 1, but variables are  likert scale I can transform

But I want to construct weighted relations to result in  more valid and reliable outputs



Second question :At this time we are  developing questionaire related with social stratification

and during this process

Can anyone offer me what should I consider to produce efficient scale in the next months after obtaining data set or how can I overcome any problems which may be seem in the future in developing scales ?



 Thanks,



Gokhan KAYA

Candidate Student for Master Program

Department of Sociology

METU, Turkey







---------------------------------
Yahoo! Mail
 Stay connected, organized, and protected. Take the tour