Skip to main content
. 2016 Feb 16;76(6):954–975. doi: 10.1177/0013164416631100

Table 1.

Testing for Unidimensionality of Data From Two GAT-V Test Forms (Base and New) and GTT.

90% CI for RMSEA
Test data χ2 df CFI TLI WRMRa RMSEA LL UL
GAT-V(B)b 1233.224 170 .966 .962 2.155 .025 .024 .026
GAT-V(N)c 830.437 170 .982 .980 1.740 .020 .019 .021
GTT 33191.931 3002 .906 .903 2.614 .017 .016 .018

Note. GAT-V = General Aptitude Test–Verbal; GTT = General Teacher Test; CFI = comparative fit index, TLI = Tucker–Lewis index, WRMR = weighted root mean square residual, RMSEA = root mean square error of approximation; CI = confidence interval; LL = lower limit; UL = upper limit. A tenable data fit is in place with CFI > .90, TLI > .90, WRMR is close to 1, and RMSEA < .05.

a

In Mplus, WRMR is used with categorical variables, which is the case with the study data (with continuous variables, standardized root mean square residual [SRMR] is used).

b

GAT-V base form.

c

GAT-V new form.