Question

CHOOSE ALL THAY APPLY regarding Binomial Logistic Regression

A) the R squared is always less than the adjusted R- squared

B) the outcome variable is usually dichotomous, and always discrete

C) the confidence intervals must not include "1"

D) Fisher's Exact Test may need to be used in indivdual factor analysis

E) It examines prediction based on "best fit linear relationships"

Answer 1

Going through the options one by one :

A. The value of adjusted R-squared or pseudo R-squared lies between 0 and 1. It might be greater than the actual R-squared value. There is no such limitation to it, in that way. However, adjusted R-squared seldom reach 0 or 1 but rather lie between the two integers.

B. The outcome variable is dichotomous and binary(not discrete particularly) in nature.

C. There is no such rule for the confidence interval. Infact, in most of the cases, the confidence interval includes 1.

D. Fischer exact test can be used to identify the proprotion difference of the two variables, but only when the sample size is small (say less than 1000).

E. The prediction is on the basis of the linear relationship between the dependent and the independent variable.

Depending on the above statements we can conclude that, the applicable options are: D and E.

You can comment if you have any doubts. Please rate the answer if it was helpful.