The Impact of Intraclass Correlation on the Effectiveness of Level-Specific Fit Indices in Multilevel Structural Equation Modeling.

Autor/inn/en	Hsu, Hsien-Yuan; Lin, Jr-Hung; Kwok, Oi-Man; Acosta, Sandra; Willson, Victor
Titel	The Impact of Intraclass Correlation on the Effectiveness of Level-Specific Fit Indices in Multilevel Structural Equation Modeling.
Quelle	In: Educational and psychological measurement, (2017) 1, S.5-31Infoseite zur Zeitschrift PDF als Volltext (1); PDF als Volltext (2); PDF als Volltext (3) Verfügbarkeit
Dokumenttyp	gedruckt; online; Zeitschriftenaufsatz
ISSN	0013-1644
DOI	10.1177/0013164416642823
Schlagwörter	Monte Carlo simulation; Studies; Questionnaires; Discriminant analysis
Abstract	Several researchers have recommended that level-specific fit indices should be applied to detect the lack of model fit at any level in multilevel structural equation models. Although we concur with their view, we note that these studies did not sufficiently consider the impact of intraclass correlation (ICC) on the performance of level-specific fit indices. Our study proposed to fill this gap in the methodological literature. A Monte Carlo study was conducted to investigate the performance of (a) level-specific fit indices derived by a partially saturated model method (e.g., CF I PS _ B and CF I PS _ W ) and (b) SRM R W and SRM R B in terms of their performance in multilevel structural equation models across varying ICCs. The design factors included intraclass correlation (ICC: ICC1 = 0.091 to ICC6 = 0.500), numbers of groups in between-level models (NG: 50, 100, 200, and 1,000), group size (GS: 30, 50, and 100), and type of misspecification (no misspecification, between-level misspecification, and within-level misspecification). Our simulation findings raise a concern regarding the performance of between-level-specific partial saturated fit indices in low ICC conditions: the performances of both TL I PS _ B and RMSE A PS _ B were more influenced by ICC compared with CF I PS _ B and SRMRB. However, when traditional cutoff values (RMSEA≤ 0.06; CFI, TLI≥ 0.95; SRMR≤ 0.08) were applied, CF I PS _ B and TL I PS _ B were still able to detect misspecified between-level models even when ICC was as low as 0.091 (ICC1). On the other hand, both RMSE A PS _ B and SRM R B were not recommended under low ICC conditions.
Erfasst von	OLC
Update	2022/1/02