Predicting implementation of active learning by tenure-track teaching faculty using robust cluster analysis

International Journal of STEM Education

Table 8 Final logistic regression model for active-learning cluster

	Estimated	95% confidence	Test	p-value
	Odds	Interval	Statistic	p-value
Intercept	9.78	(2.19, 43.69)	2.99	\(<0.001\)*
Faculty type
RG: Teaching faculty
Research faculty	0.28	(0.10, 0.79)	\(-\) 2.39	0.02*
Lecturers	0.47	(0.15, 1.50)	\(-\) 1.28	0.20
Campus
RG: Campus 3
Campus 2	0.19	(0.04, 0.92)	\(-\) 2.06	0.04*
Campus 1	2.56	(0.53, 12.42)	1.17	0.24
Discipline
RG: Biological Sciences
Engineering	0.29	(0.07, 1.23)	\(-\) 1.67	0.09
I &C Sciences	0.56	(0.14, 2.24)	\(-\) 0.81	0.42
Physical sciences	0.13	(0.03, 0.54)	\(-\) 2.84	\(<0.001\)*
Class Size
RG: Small (0–99)
Medium (100–199)	0.14	(0.04, 0.51)	\(-\) 2.98	\(<0.001\)*
Large (200 +)	0.26	(0.08, 0.82)	\(-\) 2.30	0.02*
AIC \(=\) 143.47

The final model was found by using best subsets logistic regression to model the log odds of the active-learning cluster (based on the final cluster assignment) and all possible subsets of the instructor (faculty type, faculty rank, years of teaching, gender) and classroom (campus, discipline, and class size) characteristics. The coefficients represent the increase/decrease in the odds of being in the active-learning cluster for each of the variables of interest (while holding the other variables in the model constant). The reference group (RG) are labeled for each of the categorical variables