References in periodicals archive ?
Unlike UnFr-1 and -2, students in the VCUp sample completed the instrument in a much more controlled testing context.
Specifically, one question concerned why there were fewer areas of local misfit when using the VCUp sample compared to the UnFr- and -2 samples.
As expected, the number of standardized covariance residuals was higher than that found using the VCUp sample, and the specific misfit associated with items 2, 3, and 4 was again found, suggesting that its presence was a function of the testing condition (i.e., degree of control) rather than age.
That is, the largest numbers of standardized covariance residuals were found when using data collected in an uncontrolled testing condition (i.e., UnFr-1 and -2 samples), smaller numbers of residuals were found when using data collected in a controlled condition (i.e., CUp and CUp-R samples), and the smallest numbers of residuals were found when using data collected in a very controlled condition (i.e., VCUp).
Given that standard errors can be affected by the sample size (i.e., smaller samples tend to yield larger standard errors and, in turn, may lead to smaller standardized covariance residuals), it was possible that the large residuals in the UnFr-1 and -2 samples were due to their large N, that the moderate residuals in the CUp and CUp-R samples were due to their smaller N, and that the small residuals in the VCUp sample were due its small N.
Acronyms browser ?
Full browser ?