Table 4.

Results of multi-level analysis (analysis of explained variance for 10 ACT concepts at unit and nursing home levels, n = 25 nursing homes)

ACT Concept1	% Variance that can be explained by:			% of explained variance gained by:		H0: Variance = 0 (Ha: Variance > 0) (P value)
	Unit 2	Nursing Home3	Unit + Nursing Home4	Aggregating to nursing home5	Aggregating to unit6	Nursing Home = 0 (3 level model)	Unit = 0 (2 level model)
Leadership	7.73%*	6.08%*	8.49%*	0.77%	2.41%	0.0130	0.0002
Culture	11.84%*	10.59%*	11.79%*	-0.05%	1.20%	0.1230	<0.0001
Evaluation	8.98%*	6.42%*	8.34%*	-0.64%	1.91%	0.0726	0.0001
Social Capital	5.08%*	4.18%*	4.90%*	-0.18%	0.72%	0.2399	0.0030
Formal Interactions	3.63%*	2.92%*	3.73%*	0.10%	0.81%	0.2171	0.0140
Informal Interactions	2.15%	1.97%*	2.36%*	0.21%	0.38%	0.3284	0.0557
Resources	9.74%*	10.20%*	11.36%*	1.62%	1.16%	0.1006	<0.0001
OS-staff	32.89%*	28.84%*	30.92%*	-1.97%	2.09%	0.0251	<0.0001
OS-space	34.45%*	31.74%*	35.05%*	0.60%	3.30%	0.0037	<0.0001
OS-time	21.91%*	19.76%*	20.48%*	-1.42%	0.72%	0.2212	<0.0001

¹ n values. Facility (n = 25), Unit (n = 89), Healthcare Aide responses (n = 1243 to 1250 depending on the ACT concept: leadership n = 1247, culture n = 1251, evaluation n = 1245, social capital n = 1244, formal interactions n = 1250, informal interactions n = 1244, structural and electronic resources n = 1243, OS-staff n = 1257, OS-space n = 1254, OS-time n = 1249)

² 2 level model (unit, individual). % of total variance that can be explained by the unit level

³ 2 level model (nursing home, individual) % of total variance that can be explained by the nursing home

⁴ 3 level model (nursing home, unit, individual) % of total variance that can be explained by unit + nursing home

⁵ variance gained by adding nursing home (% unit + nursing home explained variance - % unit explained variance)

⁶ variance gained by adding unit (% unit + nursing home explained variance - % nursing home explained variance)

* p < 0.05 compared to individual only level (Likelihood Ratio test)