```Date: Wed, 21 May 1997 18:07:22 GMT Reply-To: "William F. Panak" Sender: "SPSSX(r) Discussion" From: "William F. Panak" Organization: Iowa State University, Ames, Iowa, USA Subject: Re: square root of reliability -- ? jelliott@rs5.tcs.tulane.edu (Jodi Elliott) wrote: >I am doing a lisrel analysis that includes a few single-item indicators >of latent constructs. These single-item indicators are actually scale >scores (means of several items). From various sources I understand that >it is recommended (required?) that THE PATH FROM THE LATENT VARIABLE TO >ITS CORRESPONDING OBSERVED VARIABLE (LAMBDA) BE FIXED TO A VALUE EQUAL >TO THE SQUARE ROOT OF THE RELIABILITY OF THE OBSERVED SCORE. [I'm >slightly paraphrasing a passage from an article by Mathieu, Tannenbaum, and >Salas published in the Academy of Management Journal, 1992, p. 837.] >Anyway, my question is this: What is the rationale and/or the justification >for this procedure? I assume the square root of the reliability has some >meaning derived from an equation relating true scores to observed scores. >But I don't know the exact reasoning and I need to be able to explain my >use of this value in my analysis. Can anyone help with a direct answer or >a reference that I can go read? Thanks very much! >-Jodi Callahan > Tulane University I've conducted analyses where I've fixed the lx loading to a value, such as .80, based on the assumption that 80 percent of observed score variance was actual true score variance. Then, the corresponding td value can be set to .36 (1 minus the squared loading). In essence I was doing what these authors suggest, but just not squaring the reliability estimate. ------------------------------------------------------------------------------ William F. Panak Phone: 515-294-9662 A foolish consistency is the Dept. of Psychology Fax: 515-294-6424 hobgoblin of little minds... W112 Lagomarcino Hall wpanak@iastate.edu Emerson Iowa State University Visit the Infertility Internet Chatsite: Ames, IA 50011-3180 http://www.public.iastate.edu/~wpanak ```

Back to: Top of message | Previous page | Main SPSSX-L page