Table 4

Taxonomy of Fitted Multilevel Models for Change in Which Intern Personal Interest Score is Modeled as a Function of Semester in Internship, Gender, and Verbal GRE Score (n of participants = 82; n of waves of longitudinal data = 3)

ParameterModel AModel BModel CModel DModel E
Fixed Effects:       
InitialInterceptγ0024.99***24.99***27.62***24.98***28.99***
Status  (1.20)(1.26)(2.50)(1.22)(2.43)
 FAMALEγ01  -3.33 –5.07~
     (2.73) (0.33)
 GREVγ02   –0.03*–0.04**
      (0.01)(0.48)
Rate ofInterceptγ10–1.25*–1.24*–1.24*–1.22*–1.21*
change  (0.61)(1.59)(0.59)(0.59)(0.09)
Variance Components:       
Level-1:Within-personσ2g50.96***54.97***54.96***55.05***38.31***
   (8.06)(6.17)(6.17)(6.19)(0.98)
Level-2:In initial statusσ2076.23***83.44***81.65***76.17***62.47***
   (19.72)(16.14)(15.85)(15.06)(3.30)
 In rate ofσ213.99    
 change (6.16)    
 covarianceσ012.52    
   (8.17)    
Goodness of fit:       
–2LL  1,787.01,788.61,787.11,782.81,779.2

Note: FEMALE = 1 for females and 0 for males. Verbal GRE score (GREV) is centered on its sample average. The model was estimated using full ML, SAS PROC MIXED.

~p <.1.

*p<.05.

**p<. 01.

***p<.001.

or Create an Account

Close subscription notice
Close access options