Date:         Sat, 17 Nov 2007 16:48:44 +0530
Reply-To:     Euclid Associates <euclidassociates@gmail.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Euclid Associates <euclidassociates@gmail.com>
Subject:      reliability question
Content-Type: text/plain; charset=ISO-8859-1

Dear All,
 I will be grateful if you all kindly help me to understand about the
convention.

I have 28 item scale and construct is multidimensional.(I have done
factor analysis). My global reliabiliy is around 0.74. However, when I
calculate factor wise reliability, I find reliability comes down to
0.62, 0.72, 0.54 like this.

I understand Nunnally's cut off criteria is for global. Does it apply
to factor wise reliability too.
To what extent I should consider for confirmatory analysis - otherwise
measurement error will interrupt the confirmatory analysis.

Please help.
Thank you for your time.
With regards
Euclid

=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@LISTSERV.UGA.EDU (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD