Academic Achievement g–Diva or a Workhorse? Further Insights From Bifactor Modeling With the Woodcock-Johnson IV

• In Event: 34.079 - Roundtable Session 14
  In Roundtable Session: 34.079-10 - Innovative Research on Factor Analysis and Dimensionality

Abstract

Using confirmatory factor analysis, this study applied different Cattell-Horn-Carroll (CHC) factor models to normative data from the Woodcock-Johnson Tests of Achievement-Fourth Edition (WJ-IV ACH) to determine if academic g (ACH-g) is best conceptualized as a superordinate or breadth factor in its influence on lower-order achievement dimensions. The bifactor model fit the WJ-IV ACH normative data better than the favored higher-order model at ages 9-13 and 14-19. Further, calculation of omega coefficients demonstrated that the domain-specific group factors exhibited poor reliability independent of ACH-g. These results suggest that only the general achievement dimension was sufficiently robust for clinical use and raise questions about the utility of multidimensional achievement tests such as the WJ-IV ACH for high-stakes decision making and treatment planning.

Author

• Ryan McGill, The College of William & Mary