Table 1.

Multiple regression models assessing the unique predictive effects of first-semester Math Anxiety and Math Ability on university STEM outcomes.

	(A) DV: % STEM Courses					(B) DV: STEM Grades
Predictor	B	SE	t	p	d	B	SE	t	p	d
Math Anxiety	−.109	.026	−4.20	4E − 5	−.637	−2.406	.697	−3.45	7E − 4	−.524
Math Ability	.043	.023	1.91	.058	.290	.919	.602	1.53	.129	.232
% STEM Courses	–	–	–	–	–	-2.121	.652	-3.25	.001	−.493
STEM Grades	−.097	.030	−3.25	.001	−.493	–	–	–	–	–
Trait Anxiety	.026	.024	1.11	.267	.169	1.080	.623	1.74	.085	.263
Verbal Working Memory	−.004	.021	−.20	.841	−.031	.242	.542	.48	.655	.068
Gender	.002	.047	.03	.973	.005	.635	1.23	.52	.606	.078
non-STEM Grades	.157	.029	5.40	2E − 7	.819	7.579	.600	12.64	<2E − 16	1.92
Semesters Absent	.103	.021	4.98	2E − 6	.755	−.623	.585	−1.06	.289	−.161

Note: All predictors are standardized and the DV in each model is in its native units. The B estimates can therefore be interpreted as the change in actual % STEM Courses or STEM Grades associated with a one standard deviation increase in the predictor. d refers to Cohen’s d measure of effect size. For both models, df = 174. Table 1A adjusted R² = .347. Table 1B adjusted R² = .554.