Which of the following statement regarding summarizing predictive power of logistic regression models is true? group of answer choices probability of correct classification always falls between sensitivity and specificity. sensitivity and specificity are only useful measures when there are only categorical predictors in the model. when picking a cut off value to convert estimated probability to a binary value, a larger cut off value will yield smaller specificity. roc curve method is less informative than comparing two pairs of sensitivity and specificity.

Alarge university wanted to study the relationship between completing an internship during college and students' future earning potential. prom the same graduating class, they selected a random sample of 80 students who completed an internship and 100 students who did not complete an internship and examined their salaries five years after graduation. they found that there was a statistically higher mean salary for the internship group than for the noninternship group. which of the following interpretations is the most appropriate? a. there could be a confounding variable, such as student major, that explains the difference in mean salary between the internship and no internship groups.b. we cannot infer anything from these data since the distribution of salaries is likely right skewed.c. you cannot draw any valid conclusions because the sample sizes are different.d. more students should complete internships because having an internship produces a higher salary.