Table 3.

Regression Coefficients for the ITERS Subscales Under Different Model Specifications

Model	Subscale predictor	Model
		Est.	Std. Err.
Traditional linear model for ITERS subscales: R2 = 0.29	Furnishings and display	0.52	0.51
	Personal care routines	−0.29	0.41
	Listening and talking	0.05	0.34
	Learning activities	0.48	0.59
	Interactions	1.21 **	0.43
	Program structure	−0.57	0.52
Piecewise Regression model R2 = 0.29	Interactions subscale: slope estimates in quality groups
	Lower-quality group: 1 ≤ ITERS ≤ 4.2	0.63	0.67
	Upper-quality group: 4.2 < ITERS ≤ 7	1.55 **	0.53
Group Category Regression model R2 = 0.30	Interactions subscale: Lower-quality as comparison group
	Lower-quality group: 1 ≤ Interactions ≤ 4.2	NA	NA
	Upper-quality group: 4.2 < Interactions ≤ 7	5.13 **	1.60

Notes.

These models controlled for all the covariates used in Table 1.

We conducted piecewise regression and group category regression analyses only on the Interactions subscale because this was the only subscale that was significantly related to the ITERS under the traditional linear model.

^*p < .05.

^**p < 0.01.