Table 3.

MoCA Items That Are Significant Predictors of Lower Education in a Backwards Selection Binomial GLMM

Fixed Effects	Coefficient	SE	Z	P	95% CI
Intercept	−1.262	0.25	−5.03	0.000	−1.78	−0.78
Cube copy correct	−0.581	0.22	−2.67	0.008	−1.01	−0.15
Serial subtraction score	−0.469	0.19	−2.50	0.013	−0.83	−0.09
≥11 F words	−0.800	0.24	−3.36	0.001	−1.26	−0.32
Abstraction score	−0.801	0.17	−4.64	<.001	−1.14	−0.46

Random Effects	SD	χ²	P	95% CI
Study Site	0.475	3.027	0.08	0	0.894

Abbreviations: GLMM, generalized linear mixed model; SD, standard deviation; SE, standard error; CI, confidence interval; MoCA, Montreal Cognitive Assessment.
Note: Model coefficients are in log odds. 95% CIs for model coefficients are in log odds. Serial subtraction and abstraction scores centered at means. There was zero variance for study (i.e., ADNI or NACC) in model; it was excluded as a random effect because the model only converged after it was removed.