TABLE 4.

Student outcomes before and after implementation of the math prerequisite^a

	All students prediction: after prereq > before prereq	Did pass Math pre vs. all post 2013 Prediction: after prereq = before prereq	Did not pass Math pre vs. all post 2013 Prediction: after prereq > before prereq
Grade in Intro Bio
After Prereq (ref: before prereq)	0.124 (0.089) p = 0.163	−0.0574 (0.097) p = 0.552	0.483 (0.112) p = 1.7e−5
ΔAICcb	3.196	4.559	−11.346
Probability of Passing Intro Bio
After Prereq (ref: before prereq)	0.462 (0.163) p = 0.0047	0.145 (0.194) p = 0.452	1.047 (0.200) p = 1.67e−7
ΔAICcb	−5.583	1.469	−21.491

^aStudent outcomes were more favorable after the implementation of the math prereq. Estimates of effects are on a linear 4.0 scale (Grade in Intro Bio) or on the logodds scale (Probability of Passing). Standard error of the estimate shown in parenthesis, boldface indicates significance to a <0.05. Boldface and p values shown for corro­boration with backward selection and change in AICc. (Note that backward model selection was performed using AICc to identify best-fitting model, so p values should not be interpreted; Burnham and Anderson, 2002; Theobald, 2018.)

^bΔAICc comparing the model with and without an indicator for before/after implementation of the prereq (with – without, so negative values indicate that the model with the before/after indicator has superior fit); all models control for prior GPA and include a varying intercept for instructor.