Logistic regression

Sandra Rogne

For this part, you are able to test your skills and knowledge on logistic regression. Below is a quiz with various question on the concepts and important knowledge logistic regression such as preparations, simple logistic regression, multiple logistic regression, model diagnostics, and interaction analysis. After completing the quiz, the correct answers will appear. At the bottom of the page are links to the sections referred to in the quiz for more information and further review.

Test your skills

Logistic regression

1 / 9

Which of the following variables would be appropriate to use as the outcome in a logistic regression analysis?

Income (continuous)

Social support (low, medium, high)

Cognitive ability (low, high)

Nationality (Swedish, Danish, Norweigan, Finnish)

2 / 9

In logistic regression, what does the term “odds ratio” measure?

The probability of the event occurring for each unit change in the independent variable

The log-odds of the event occurring for each unit change in the independent variable

The relative likelihood of the event occurring between two groups defined by the independent variable

The mean change in the dependent variable for each unit increase in the independent variable

3 / 9

In logistic regression, why is interpreting odds ratios (ORs) as percentages not recommended?

Interpreting ORs as percentages is always inaccurate

ORs cannot be expressed as percentages

Interpreting ORs as percentages can lead to inflated or understated values

ORs should always be interpreted as percentages

4 / 9

When interpreting the coefficient of an independent variable in logistic regression, what does a positive coefficient indicate?

An increase in the independent variable increases the odds of the outcome

An increase in the independent variable decreases the odds of the outcome

The independent variable has no effect on the outcome

All of the above

5 / 9

The ROC curve is a graph that shows how well the estimated model predicts cases (sensitivity) and non-cases (specificity). What we are interested in here is the “area under the curve” (AUC).

The figure displays an AUC value of 0.64, suggesting…

… excellent predictive power

… good predictive power

… fair predictive power

… poor predictive power

… failure

6 / 9

What does a non-significant p-value associated with an interaction term imply in logistic regression analysis?

The model needs to be expanded to include more predictors

The interaction term does not provide any additional explanatory value beyond the main effects

The model might have issues with multicollinearity

The moderator influences the variables in a negative direction

7 / 9

If, when doing model diagnostics for logistic regression, the Hosmer and Lemeshow test produces a p-value of 0.48, what does this indicate?

The estimated model has adequate fit

The estimated model does not adequately fit the data

8 / 9

What is the difference between the commands ‘logistic’ and ‘logit’?

logistic produces odds ratio, logit produces log odds

logit produces odds ratio, logistic produces log odds

9 / 9

You have just ran a multiple logistic regression analysis and now you would like to display your results. How would you do this?

Estimates table

Box plot

Frequency table

A coefficients plot

Histogram

Your score is

The average score is 37%

Review of the quiz

Q1 – See more on Logistic regression
Q2 – See more on Introduction
Q3 – See more on Logistic regression in short
Q4 – See more on Logistic regression in short
Q5 – See more on ROC curve
Q6 – See more on Interaction analysis
Q7 – See more on The Hosmer and Lemeshow test
Q8 – See more on Function
Q9 – See more on Multiple logistic regression