Table 2.

Mixed-effects logistic regression model for condition-by-time effects on accuracy (fixed effects).

Accuracy
Predictors	Odds ratios	Confidence interval	P
(Intercept)	2.17	1.23–3.83	0.008
Condition [1]	0.66	0.45–0.97	0.035
Condition [2]	0.74	0.52–1.04	0.085
Condition [3]	0.71	0.50–1.00	0.048
Timepoint	1.00	0.88–1.14	0.956
Relation type [Spatial]	1.09	1.03–1.15	0.002
Premises [2 Premise]	1.54	1.46–1.63	<0.001
Dimensions [1 Dimension]	1.41	1.37–1.45	<0.001
Solution [Indeterminate]	0.76	0.71–0.81	<0.001
Solution [True]	1.10	1.03–1.18	0.006
Age	1.00	0.99–1.01	0.806
Gender [Male]	1.14	0.96–1.36	0.128
Income bracket	1.07	1.03–1.12	0.002
Total education	1.08	0.93–1.26	0.299
Condition [1] * Timepoint	1.46	1.12–1.74	<0.001
Condition [2] * Timepoint	1.31	1.11–1.53	0.001
Condition [3] * Timepoint	1.35	1.11–1.59	<0.001
Random effects
σ2	3.29
τ00 ID	0.51
ICC	0.13
NID	301
Observations	27,563
Marginal R2/Conditional R2	0.046/0.173

Condition 1, adaptive practice; Condition 2, alphabetize spatial tool; Condition 3, mental models training. Condition 0 (no intervention) was the reference level in this model: The following variables were dummy coded. Relation Type: spatial vs. non-spatial; Premises: two premise vs. three premise; Dimensions: one-dimension relations vs. two-dimension relations; Solution: False vs. Indeterminate and False vs. True; Gender: 0, female; 1, male. Bold values indicate significant effects (p < 0.05).