Browsing by Author "Kwok, Oi-Man"

Now showing 1 - 5 of 5

Dynamics of Teacher-Student Relationships: Stability and Change Across Elementary School and the Influence on Children's Academic Success
(2012-07-01)
Spilt, Jantine L.
;
Hughes, Jan N.
;
Wu, Jiun-Yu
;
Kwok, Oi-Man
;
交大名義發表
"This study modeled teacher-student relationship trajectories throughout elementary school to predict gains in achievement in an ethnic-diverse sample of 657 academically at-risk students (mean age = 6.57 years, SD = .39). Teacher reports of warmth and conflict were collected in Grades 1-5. Achievement was tested in Grades 1 and 6. For conflict, low-stable (normative), low-increasing, high-declining, and high-stable trajectories were found. For warmth, high-declining (normative) and low-increasing patterns were found. Children with early behavioral, academic, or social risks were underrepresented in the normative trajectory groups. Chronic conflict was most strongly associated with underachievement. Rising conflict but not declining Conflict coincided with underachievement. The probability of school failure increased as a function of the timing and length of time children were exposed to relational adversity."
Editorial: Recent Advancements in Structural Equation Modeling (SEM): From Both Methodological and Application Perspectives
(2018-10-09)
Kwok, Oi-Man
;
Cheung, Mike W. L.
;
Jak, Suzanne
;
Ryu, Ehri
;
Wu, Jiun-Yu
The Impact of Intraclass Correlation on the Effectiveness of Level-Specific Fit Indices in Multilevel Structural Equation Modeling: A Monte Carlo Study
(2017-01)
Hsu, Hsien-Yuan
;
Lin, Jr-Hung
;
Kwok, Oi-Man
;
Acosta, Sandra
;
Willson, Victor
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., CFIPS B and CFIPS W) and (b) SRMRW and SRMRB 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 TLIPS B and RMSEAPS B were more influenced by ICC compared with CFIPS B and SRMRB. However, when traditional cutoff values (RMSEA <= 0.06; CFI, TLI <= 0.95; SRMR <= 0.08) were applied, CFIPS B and TLIPS 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 RMSEAPS B and SRMRB were not recommended under low ICC conditions.
The Impact of Intraclass Correlation on the Effectiveness of Level-Specific Fit Indices in Multilevel Structural Equation Modeling: A Monte Carlo Study
(2017-01)
Hsu, Hsien-Yuan
;
Lin, Jr-Hung
;
Kwok, Oi-Man
;
Acosta, Sandra
;
Willson, Victor
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., CFIPS B and CFIPS W) and (b) SRMRW and SRMRB 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 TLIPS B and RMSEAPS B were more influenced by ICC compared with CFIPS B and SRMRB. However, when traditional cutoff values (RMSEA <= 0.06; CFI, TLI <= 0.95; SRMR <= 0.08) were applied, CFIPS B and TLIPS 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 RMSEAPS B and SRMRB were not recommended under low ICC conditions.
Using Design-Based Latent Growth Curve Modeling With Cluster-Level Predictor to Address Dependency
(2014)
Wu, Jiun-Yu
;
Kwok, Oi-Man
;
Willson, Victor L.
;
教育研究所
;
Institute of Education
The authors compared the effects of using the true Multilevel Latent Growth Curve Model (MLGCM) with single-level regular and design-based Latent Growth Curve Models (LGCM) with or without the higher-level predictor on various criterion variables for multilevel longitudinal data. They found that random effect estimates were biased when the higher-level predictor was not included and that standard errors of the regression coefficients from the higher-level were underestimated when a regular LGCM was used. Nevertheless, random effect estimates, regression coefficients, and standard error estimates were consistent with those from the true MLGCM when the design-based LGCM included the higher-level predictor. They discussed implication for the study with empirical data illustration.