Logistic Regression

Question 1

How do we fix the problem of using categorical data?

Answer 1

We predict the probability of Y not just Y.

Question 2

What do we have to do next when we are predicting the probability of Y?

Answer 2

We have to linearise the S-Shape curve so that we have a straight line and not a curve!
This transformation makes interpreting the results differently.

Question 3

What does ML stand for?

Answer 3

Maximum likelihood.

Question 4

What is Y in the logistic regression equation?

Answer 4

It is log(Odds) OR log(p/1-p).

Question 5

How do you conduct a logistic regression in SPSS?

Answer 5

We don’t need to know this for the exam.

Question 6

What is the first step in interpreting the SPSS results?

Answer 6

We check the null model (block 0).

Question 7

What is the null model?

Answer 7

It is the model without including any predictor.

This value has to be beat when we include the predictors.

Question 8

What are the two Pseudo R-squareds that we use for the model fit and what are their cut-off points.

Answer 8

CoxSnell - 0 to 0.75.
Nagelkerke - 0 to 1.

How well you can predict the DV based on the IVs.

Question 9

What is the other model fit we can use and how do we know if the model fits or not?

Answer 9

Goodness of fit - Hosmer-Lemeshow.

P-value must be non-significant.

Question 10

What must Block 1 overall percentage be for the results to be significant?

Answer 10

It must be larger than Block 0 overall percentage.

Question 11

How do we test for multicollinearity in logistic regression?

Answer 11

We can’t… Have to run the original model as linear regression model and check VIF.

Question 12

How do we test for linearity in logistic regression?

Answer 12

Compute natural log of predictor and test the interaction.
Interaction significant = linearity issue.
Look at significance in the table of predictor by log transformed predictor.

Question 13

What is B in the SPSS table?

Answer 13

Log(Odds).

Question 14

How do we interpret B?

Answer 14

It is the change in Log(Odds) for one unit increase in our predictor variable.
Not very intuitive.

Question 15

What should we interpret instead of B?

Answer 15

It is better to interpret this in probabilities.

Question 16

What is OddsRatio?

Answer 16

It is the change in odds resulting in a unit change in the predictor.

Question 17

How do we interpret OddsRatio?

Answer 17

If greater than 1 - predictor increases, odds of being in target group increases (coded as 1). P significant = increase is significant.
If smaller than 1 - predictor increases, odds of being in target group decreases. P significant = decrease is significant.

Question 18

What is the OddsRatio if the probability is the same to be in any of the groups?

Answer 18

1.

Question 19

What number must not be between the lower and upper CI?

Answer 19

1 (all other sessions is 0!!!).

Question 20

What is OddsRatio in the SPSS table?

Answer 20

Exp(B).

Question 21

How do you calculate Odds?

Answer 21

p/1-p.

Question 22

How do you calculate OddsRatio?

Answer 22

e to the power of B.

Question 23

What can Logistic Regression be used for?

Answer 23

To form deep learning networks + study neural pathways and AI.

Question 24

What two assumptions aren’t necessary in Logistic Regression?

Answer 24

The error term doesn’t need to be normally distributed.

Homoscedasticity is not required.

Question 25

When do we use a Logistic Regression?

Answer 25

When we have categorical data.