Table 5.

Adjusted Correlates of “Positive” and “Negative/Neutral” Interactions.

Variables	Positive Interaction		Negative/neutral interaction
	Estimate (95% CI)	P Value	Estimate (95% CI)	P Value
Age	1.010 (0.978-1.043)	.54	0.993 (0.964-1.023)	.66
Gender, female (Ref = male)	1.547 (0.618-3.871)	.35	0.878 (0.415-1.856)	.73
Race, African American (Ref = Caucasian)	1.173 (0.375-3.672)	.78	0.436 (0.183-1.036)	.06
Marital status, married (Ref = not married)	1.410 (0.545-3.648)	.48	0.602 (0.271-1.336)	.21
Cognitive status, severe impairment (Ref = moderate impairment)	0.699 (0.250-1.957)	.50	0.956 (0.408-2.240)	.92
Interaction location, private area (Ref = not private)	0.137 (0.044-0.431)	<.01	4.521 (1.862-10.981)	<.01
Interaction situation, care related (Ref = not care-related)	1.002 (0.373-2.698)	.99	1.339 (0.583-3.076)	.50
Interpersonal distance, <30 in (Ref = >30 inches)	0.340 (0.151-0.762)	<.01	1.483 (0.745-2.952)	.26
Type of staff/person resident interacting with, nursing (Ref = not nursing)	2.093 (0.771-5.681)	.15	0.507 (0.217-1.186)	.12
Resident participation, active (Ref = passive)	0.052 (0.022-0.123)	<.0001	9.663 (4.555-20.499)	<.0001

Abbreviation: CI, confidence interval.

^aN = 303. Covariates: Age, gender, race, and marital status. Predictors: Cognitive status, interaction location, interaction situation, interpersonal distance, type of staff, and resident participation. Of covariates, race and marital status were associated with “positive” interaction, while only marital status was associated with “negative/neutral” interaction in the bivariate model. While these covariates lost their significance in the multivariable model, the predictors that were significant in bivariate model retained their significance in multivariable model, except interpersonal distance which lost the significance in the multivariable model for “negative/neutral” interactions.