Abstract
Studies that use structural equation modeling (SEM) techniques are increasingly encountered in the language assessment literature. This popularity has created the need for a set of guidelines that can indicate what should be included in a research report and make it possible for research consumers to judge the appropriateness of the interpretations made from a reported study. This article attempts to fill this void by providing a set of reporting guidelines appropriate for language assessment researchers.
Notes
1 It is not clear how many of these articles were found in the language assessment literature because the researchers did not distinguish between language learning and language assessment. Moreover, the researchers did not include books in this review, and it is in books that we would expect the space for detailed information about the procedures used.
2 These are informal and heavily simplified conditions. Bollen (Citation1989b) and Kline (Citation2011) provide a comprehensive yet accessible discussion of model identification in SEM.
3 We suggest interested readers consult Allison (Citation2003) who provides a clear and accessible review of missing data techniques for SEM and Rubin (Citation1976) for the definition of missing data mechanism.
4 Although multiple imputation is a general strategy that is applicable to a number of analysis situations, its use in SEM contexts is not without complications and is still an active area of methodological research (Allison, Citation2003; Lee & Cai, Citation2012).
5 This classification system is not consistent in the SEM literature. For instance, sometimes absolute and corrected for parsimony indices are collapsed into one category. Moreover, the three approaches are not always exclusive.