Question

3. Which of the following examples is/are a sample application of Logistic Regression? (select all that apply)
The probability that a person has a heart attack within a specified time period using person's age and sex.
Customer's propensity to purchase a product or halt a subscription in marketing applications.
Likelihood of a homeowner defaulting on a mortgage.
Estimating the blood pressure of a patient based on her symptoms and biographical data.​

Answer 1

Answer:

I am using logistic regression. When I included the covariates in the model with each predictor (P<.05). I was not confident about the way I interpreted the results.

My understanding is when I include covariate, I am controlling the effect of these covariates on the outcome. Then I will see if the predictor would still affect the outcome or not.

However, I was a little confused when I was interpreting the results. Because some of these covariates (I added in the model) were also significant predictors.

So, does that mean that the insignificant covariates in the model were held constant / controlled?

I hope my question is clear! thanks

Relevant answer

David Morse

Oct 10, 2020

Answer

Hello Sarah,

When you evaluate a LR model, the effect of _any_ independent variable in the model is interpreted as follows:

For each unit change in (this IV), we estimate that the log-odds of cases being in the target outcome group (e.g., 1 instead of 0) increases by B (the regression coefficient) units, when all other IVs in the model are held constant.

Or, for those who love odds or odds ratios:

For each unit change in (this IV), we estimate that odds of cases being in the target outcome group (e.g., 1 instead of 0) increases by exp(B) (the exponentiated value of the regression coefficient) units, when all other IVs in the model are held constant.

In either case, the phrase, "when all other IVs in the model are held constant" implies that, yes, you are "controlling" for the other IV/covariates you have included.

Good luck with your work.