binary logistic regression

Question 1

what is a binary logistic regression for?

Answer 1

Binary logistic regression is a statistical analysis technique used to model the relationship between one or more independent variables and a binary dependent variable. The dependent variable in this case is categorical and has two possible outcomes, often represented as 0 or 1 (e.g., “yes” or “no,” “success” or “failure”)

Question 2

How to meet the assumptions?

Answer 2

assumption of multicollinearity- look at coefficients table- Vif needs to be below 10 and tolerance needs to be above 0.1

assumption of linearity-look at variables in the equation- sig values (interaction ones) need to be non signifcant

Question 3

How to report the percentage of the cases the model can classify before we include the predictors

Answer 3

look at block 0: beginning block-its a percentage

Question 4

Does the inclusion of predictors improve the model?

Answer 4

Look at omnibus test of model coefficients- if signifcant it improves the model

report the chi squared

x2 (s)= chi 2= p< sig value

Question 5

is the model, including predictors good at predicting the DV?

Answer 5

Look at model summary and report the Nagelkere R2 result-

If Negelkereke R2 is .35, the model predicts 3.1% of the DV

Question 6

Fit of the data- Is there misfit of the data?

Answer 6

Look at the hosmer and lemeshow test-
should be non significant for no misfit

Question 7

When the predictors were added, what percentage of cases can it correctly classify-

Answer 7

Look at classification table- compare it to block 0
again you get a percentage

Question 8

Calculate and report sensitivity, specificity, positive predicted value

Answer 8

look at classification table-
use picture-

calculate positive predicted value- Total, example=
43 participants in total were predicted by the model as
passing the exam.
Out of these, 37 were correctly classified.
(37/43) x 100 = 86.04% positive predictive value

Question 9

What predictor values contributed significantly to the logistic regression?

Answer 9

Look at the variables in the equation, look at the sig values- those that are signifcant are predictors that contributed signifcantly to the model

Question 10

Odds ratio

Answer 10

In exam all will be 1- exposure does not affect the odds of the outcome