<table cellspacing="0" cellpadding="0" border="0" ><tr><td valign="top" style="font: inherit;"><DIV>The 50-item instrument has 4 subscales where each subscale has 2 to 15 items quantified using 4-point ordinal scale coded 1=strongly disagree to 4=strongly agree. Though I know that the instrument has 4 dimensions, I've tried to test the unidimensionality of the overall 50 items. Using Winsteps, the % of variance explained is somewhat acceptable (>75%) for a unidimensional assumption. Am I wrong if I do Rasch analysis for the entire instrument rather than Rasch analysis for each domain?</DIV>
<hr size=1><a href="http://sg.rd.yahoo.com/ph/search/firefox/mail/signature/*http://ph.downloads.yahoo.com/firefox/"> Firefox 3</a>: Faster, More Secure,Customizable and FREE.