Logistic models among 192 same sex twin pairs of questionnaire based information in predicting DNA-determined zygosity.

a. PPZ was the question about previously perceived zygosity, CBS was the question whether strangers confused by twins’ appearance, Formal zygosity test-PPZ-CBS was a complex strategy for zygosity determination (Fig 1).
b. p would be the probability of dizygotic (DZ) twins. IPPZ1, IPPZ2, IS1, IS2, IF1, IF1 referred to the dummy variables of question PPZ, CBS and Formal zygosity test-PPZ-CBS respectively, a subscript “1” indicated monozygotic (MZ) twins versus Do not know\Hard to say, subscript “2” indicated DZ versus Do not know\Hard to say. IAge1-Age4 were the dummy variables of age group 30–39, 40–49, 50–59 and ≥60 versus ≤29. IG was the dummy variable of Male versus Female. Bolds represented P-value <0.05.
c. Cut-off point would be the average point of probability (of DZ) maximizing both sensitivity and specificity. Consistency rate would be the proportion of correctly diagnosed MZ and DZ pairs by the logistic model among 192 pairs.
d. Prediction Errors were the raw estimates calculated by the Leave-One-Out Cross Validation (LOOCV) method.
Logistic models among 192 same sex twin pairs of questionnaire based information in predicting DNA-determined zygosity.