Table 5.

Table 5 Prediction of Mastery (A grade) at SHIS

Mastery Prediction Model
Omnibus Test	χ2	df	p	Nagelkerke R2
∗ Significance for Spearman Rho Coefficient (Distribution of probability of A is nonnormal).
Statistic	11.58	1	0.001	0.23

Logistic Regression Equation Statistics
							OR 95% CI
Source	B	SE	Wald	df	p	OR	Lower	Upper
GRE-V	0.012	0.004	7.97	1	0.005	1.012	1.004	1.020
Constant	−6.863	2.207	9.67	1	0.002	0.001	—	—

Classification Tables (Mastery Decision Threshold = 0.32.)
Model Development (n = 65)				Model Cross Validation (n = 36)
Mastery	Predicted			Mastery	Predicted
Observed	No A	A	sp %	Observed	No A	A	sp %
			Se %				Se %
No A	38	6	86.4	No A	16	6	72.7
A	11	10	47.6	A	5	9	64.3
NPV, PPV, tot. %	77.6	62.5	73.8	NPV, PPV, tot. %	76.2	60.0	69.4

Predicted Versus Observed Mastery Measures of Association
Association	R	p	r2	Association	r	p	r2
Prob. A versus obs A	0.44	0.002 ∗	0.20	Prob. A versus obs A	0.37	0.024	0.14
Pred. A versus obs A	0.37	0.002	0.14	Pred. A versus obs A	0.37	0.028	0.14
OR pred A/not pred A 95% CI		5.76 (1.71, 19.39)		OR pred A/not pred A 95% CI		4.80 (1.14, 20.27)

CI = confidence interval; GRE-V = Graduate Record Exam verbal score; NPV = negative predictive value; OR = odds ratio; PPV = positive predictive value; SE = standard error; SHIS = School of Health Information Science, UT;

Note: B = Logistic Regression Coefficient; SE = Standard Error; Wald = Wald statistic; OR = odds ratio; df = degrees of freedom. The row headed NPV/PPV % shows, as percents, negative predictive values in columns labeled “No A” and positive predictive values in columns labeled “A”. Columns headed with sp % displayed above Se % show, as percents, specificity in the row labeled “No A”, above sensitivity in the row labeled “A”, above overall accuracy of prediction in the row labeled NPV/PPV %' Prob. A = predicted probability of “A” grade; Obs A = observed binary occurrence of “A” grade; pred A = predicted binary occurrence of “A” grade.