14 Pages

Module 19. Confirmatory Factor Analysis

In the previous module, we considered exploratory factor analysis (EFA) and argued that it is an important statistical tool at the initial development stages when a researcher is uncertain about the underlying structure of his or her test. In fact, EFA methods are not even the first analyses that should be conducted when a scale is proposed. Basic item analysis techniques (Module 13) should be the initial analyses to see if items have acceptable means and variability as well as reasonable item-total correlations. EFA is often conducted after original item checks have been conducted. As one of the last line of analyses conducted in the scale evaluation process, methods in this module are used then to confirm whether the structure of a test conforms with a particular theory or structure hypothesized by the test developers. As a result, the techniques discussed in this module are powerful and are helpful in theoretically evaluating a test.