Methods for significance testing of categorical covariates in logistic regression models after multiple imputation: power and applicability analysis

Table 3 P-values from complete data analysis, pooling methods and listwise deletion

	Full data	RR	MR	CHI	VAR	MPR_out	Listwise
Group	0.0515	^a	0.0498	0.0583	0.0643	0.0549	0.3234
Age	0.9602	0.9245	0.9283	0.8780	0.9244	0.8898	0.8245
Gender	0.2250	0.3040	0.2854	0.3017	0.3041	0.2862	0.8836
BMI	0.0764	0.0172	0.0222	0.0137	0.0173	0.0049	0.0103
Education	0.5108	^a	0.7546	0.7235	0.7468	0.4579	0.6141
Sitting	0.0195	^a	0.0396	0.0355	0.0498	0.0306	0.1196
Lifting	0.9830	^a	0.9485	0.8755	0.9484	0.7605	0.9289
Vibration Tools	0.0090	^a	0.0115	0.0130	0.0236	0.0109	0.0833
Pain baseline	0.0000	0.0001	0.0000	0.0000	0.0001	0.0000	0.0008
Physical Functioning	0.0913	0.0970	0.1095	0.0943	0.0970	0.0532	0.0608
Disability	0.0049	0.0032	0.0009	0.0027	0.0032	0.0022	0.0595
Kinesiophobia	0.1730	0.2115	0.2312	0.2084	0.2115	0.2018	0.0438

^aFor the categorical variables the overall p-value could not be obtained by RR. Full data complete data, RR Rubin’s Rules, MR Meng and Rubin pooling, CHI chi-square test with multiple degrees of freedom, VAR pooled sampling variance method; MPR _out Median P Rule with the outcome excluded from model, Listwise analysis after excluding cases with missings

ISSN: 1471-2288