Table 13.

Mardia’s test of multivariate normality for core continuous variables in APIM.

Model	Mardia skewness	χ²	df	p	Mardia kurtosis	Z	P
Self-care maintenance	8.086	195.416	120	1.65×10⁻⁵	82.805	1.335	0.182
Self-care management	9.434	227.979	120	1.04×10⁻⁸	84.228	2.012	0.044
Self-care confidence	8.177	197.609	120	1.05×10⁻⁵	87.399	3.522	4.29×10⁻⁴

APIM, Actor-Partner Interdependence Model. All three models departed from multivariate skewness, while the management and confidence models also deviated in multivariate kurtosis. Hence, 5000-sample dyadic bootstrap was used in the final APIM to derive path confidence intervals.