Factors That Affect Partial Measurement

893 Words Sep 19th, 2016 4 Pages
Despite the number of studies that have been dedicated to investigating the factors that affect partial measurement invariance (Kaplan & George, 1995); there continues to be uncertainty in the proportion of noninvariant items permitted on a scale that will lead to valid statistical conclusions (Donahue, 2006). The number of items ranges from as little as one invariant item (Byrne, Shavelson, & Muthen, 1989; Steenkamp & Baumgartner, 1998) to one-half of the items (Reise, Widaman & Pugh, 1993;) to full scalar invariance (Bollen, 1989; Horn & McArdle, 1992) for meaningful interpretations of latent mean differences.
The effects of partial ME/I on scale use has been examined in two ways. Either applied studies have investigated partial ME/I in terms of establishing the partial measurement invariance of an instrument (Byrne et al., 1989) or methodological researchers have investigated some of the factors thought to affect partial measurement invariance results (Hutchinson, 2002; 7 Hutchinson & Young, 2003; Kaplan, 1989; Kaplan & George, 1995). Methodological researchers have conceptualized the detection of partial measurement invariance in two ways, through the examination of the accuracy of selection of persons within groups (Millsap & Kwok, 2004), and through the power of the likelihood ratio test (Hutchinson, 2002; Hutchinson & Young, 2003; Kaplan, 1989; Kaplan & George, 1995).
Millsap and Kwok (2004) examined selection bias based on a partially invariant instrument. They…
Open Document