Question

An investigator finds 5 independent factors, the presence or absence of which correlates with a disease. Which is the next study you will do?

• A. ANOVA
• B. Multiple linear regression
• C. Multiple logistic regression
• D. Kruskal-Wallis test

Hint:

The outcome is binary (disease present or absent), so which regression handles a yes or no outcome?

Answer 1

The Correct Option is C

Step 1: Identify the nature of the variables. There are 5 independent (predictor) factors, and the outcome is a disease that is either present or absent, which is a binary (dichotomous) dependent variable.
Step 2: Choose the regression model based on the type of outcome. When the dependent variable is binary and there are several independent predictors, the appropriate technique is multiple logistic regression, which models the probability of the outcome and yields odds ratios for each factor.
Step 3: Eliminate the wrong choices. Multiple linear regression is used only when the dependent variable is continuous (a numeric quantity), not a yes or no outcome, so it cannot be applied here. ANOVA compares the means of a continuous outcome across groups and does not handle a binary outcome with multiple predictors. The Kruskal-Wallis test is a non-parametric method that compares medians across three or more independent groups and is not a multivariable model for a binary outcome.
Step 4 (recall key correction): The printed answer key marks option B (multiple linear regression). This is statistically incorrect, because a present or absent (binary) outcome cannot be modelled with linear regression. The medically and statistically correct method for a binary outcome with multiple predictors is multiple logistic regression.
Conclusion: The correct next study is multiple logistic regression, so the correct answer is option 3.