Skip to main content
Large-scale Assessments in Education

An IERI – International Educational Research Institute Journal

Table 3 Results of G-DINA Model Fit

From: Using country-specific Q-matrices for cognitive diagnostic assessments with international large-scale data

Country

Q-matrix

AIC

BIC

RMSEA2

SRMSR

Finland

Original

120,390.5

124,375.4

0.0761

0.0862

Stepwise

117,980.4

122,176.9

0.0706

0.0745

Chiu

120,915.6

135,526.8

0.0711

0.0773

United States

Original

292,238.3

296,929.5

0.0846

0.1099

Stepwise

276,331.3

282,622.1

0.0691

0.0735

Chiu

277,763.4

285,927.0

0.0705

0.0811

Singapore

Original

137,450.0

141,628.3

0.0881

0.1205

Stepwise

125,654.6

131,715.5

0.0663

0.0795

Chiu

126,544.2

133,276.5

0.0651

0.0889

Australia

Original

210,511.1

214,942.2

0.0851

0.1203

Stepwise

198,039.7

203,403.4

0.0689

0.0766

Chiu

197,796.4

206,006.8

0.0706

0.0784

Tunisia

Original

138,141.2

142,270.6

0.0814

0.0935

Stepwise

133,381.0

138,004.0

0.0733

0.0796

Chiu

133,287.1

140,533.2

0.0712

0.0793

Five-countries data

Original

948,582.3

954,166.9

0.0643

0.0973

Stepwise

890,001.8

901,176.7

0.0380

0.0427

Chiu

914,767.5

927,584.2

0.0461

0.0668

  1. AIC Akaike Information Criterion, BIC Bayesian Information Criterion, RMSEA2 the limited-information version of Root Mean Square Error of Approximation, SRMSR Standardized Root Mean Square Residual. The best fitting model is in boldface
Back to article page