Table 5 

Sequential comparison of multinomial logistic models for the prediction of physical activity frequency in girls (n = 3 119) 

Statistics for individual predictors 
Model statistics 



Model 
Age 
Pubic hair stage 
Menarche 
Pubertal timing 
BDS 
Region × body fat 
SES × body fat 
Correct classification 
Pseudo 2 LogLikelihood 
Wald χ^{2 }(df) corr. for model 
Nagelkerke's pseudo R^{2} 



0^{#} 
Wald χ^{2 }(df) corrected 
36.2% 
9393.16 
92.77 (12.99) 
.045 

pvalue*‚ 
< .001 



1 
Wald χ^{2 }(df) corrected 
112.22 (2.90) 
36.7% 
9222.36 
178.55 (15.38) 
.093 

pvalue * 
< .001 
< .001 



2 
Wald χ^{2 }(df) corrected 
41.32 (2.80) 
6.23 (5.74) 
8.71 (5.60) 
37.5% 
9201.17 
181.48 (24.41) 
.099 

pvalue* 
< .001 
.506 
.152 
< .001 



3 
Wald χ^{2 }(df) corrected 
40.28 (2.88) 
5.98 (5.75) 
6.94 (5.61) 
16.70 (5.72) 
38.1% 
9177.39 
185.18 (27.99) 
.106 

pvalue* 
< .001 
.588 
.177 
.480 
< .001 



4 
Wald χ^{2 }(df) corrected 
39.54 (2.90) 
5.65 (5.76) 
7.85 (5.61) 
17.12 (5.71) 
21.38 (8.45) 
38.2% 
9146.57 
189.62 (33.27) 
.114 

pvalue* 
< .001 
.635 
.135 
.414 
.014 
< .001 



5 
Wald χ^{2 }(df) corrected 
38.57 (2.90) 
5.79 (5.75) 
7.70 (5.60) 
17.07 (5.70) 
21.40 (8.43) 
5.75 (2.92) 
8.24 (5.68) 
38.1% 
9128.14 
194.32 (38.08) 
.119 
pvalue* 
< .001 
.616 
.147 
.372 
.012 
.048 
.017 
< .001 



* Adjustment for multiple tests: Šidák sequential ^{# }model 0 = baseline model including body fat percentage and sociodemographic variables: region, SES, migrant background Each row of the table shows the results of one tested model. Lefthand the test statistics for the independent variables are given while righthand information on model fit is displayed. The corrected Wald chisquare test tests if an individual independent variable (individual predictors) or all independent variables together (model statistics) significantly contribute to the prediction of the dependent variable; it is corrected for the sampling plan. Correct classification rate is the proportion of participants for whom the tested model could correctly predict the category of the dependent variable (PA frequency). Pseudo 2 LogLikelihood: In logistic regression models are compared due to their 2 loglikelihood; since for complex samples no likelihood ratio test is available the values are only descriptive; better fitting models have smaller values. Nagelkerke's pseudo R^{2 }is a measure of explained variation in the dependent variable that emulates R^{2 }from linear regression. 

Finne et al. International Journal of Behavioral Nutrition and Physical Activity 2011 8:119 doi:10.1186/147958688119 