Deck 18: Logistic Regression

Question 1

Which of the following is true about logistic regressions?

A) It is a method to test for causality in correlational designs.
B) It is a method of extracting the strongest predictor variables.
C) It is a method to analyse a binary outcome variable.
D) It is a way to assess interactions in regressions.

Answer 1

C

Question 2

If we wanted to explore the association between gender and having a criminal record, what would be the most appropriate analysis?

A) Logistic regression
B) Chi-square
C) Multiple regression
D) Pearson correlation

Answer 2

B

Question 3

In a chi-square analysis, the residual statistic tells you…

A) the amount of error variance in the model.
B) the number of outliers in the model.
C) the difference between the expected number of cases and actual cases.
D) whether the data is normally distributed.

Answer 3

C

Question 4

If we are looking to explore what predicts being a smoker (0 = non-smoker, 1 = smoker) and the model assigns an individual with the predicted value of 0.73, then what does this mean?

A) The model cannot predict whether this individual is a smoker or a non-smoker.
B) The model predicts that the individual has a 73% chance of being a non-smoker.
C) The model predicts that this individual is a smoker who has a good chance of giving up smoking.
D) The model predicts that the individual has a 73% chance of being a smoker.

Question 5

In a logistic regression output, what does 'block 0' refer to?

A) How well the control variable model predicts the outcome variable
B) How well the predictor variable model predicts the outcome variable
C) Whether the model has met the assumptions
D) How well a random model predicts the outcome variable

Question 6

What does the Cox & Snell statistic tell you?

A) Whether the model is a significant predictor of the outcome variable
B) The amount of estimated variance in the outcome variable that the model explains
C) Whether there is multicollinearity in the model
D) Whether the residuals are normally distributed

Question 7

Which statistic can interpret the direction of any significant predictors?

A) Wald
B) β
C) Cox & Snell
D) p value

Question 8

In the logistic regression output, what does the classification table tell us?

A) The descriptive statistics
B) The assumptions
C) Whether the model is significant
D) How the model allocated cases and the amount of cases it correctly allocated

Question 9

If we wanted to create a scatterplot for the significant finding for our continuous predictor variable, then how should we do this?

A) The predictor variable on the x axis, and the predicted outcome probability on the y axis
B) The predictor variable on the x axis, and the actual outcome variable on the y axis
C) The predicted outcome probability on the x axis, and the actual outcome variable on the y axis
D) The predictor variable on the x axis, and the predicted outcome residual on the y axis

Question 10

When graphing continuous predictors and predicted probabilities, what should a good predicting model look like?

A) A linear line
B) Stepped
C) A u-shaped curve
D) A flat line

Question 11

In the first worksheet exercise, what was the chi square value?

A) .016
B) 1
C) 5.769
D) 4.848

Question 12

In the second worksheet exercise, what is the significance of Block 0?

A) .923
B) .009
C) .006
D) .689

Question 13

In the second worksheet exercise, what is the significance of Block 1?

A) .006
B) .923
C) .009
D) .689

Question 14

How much variance is explained in the final model (Block 1)?

A) 126.147
B) 12.322
C) .116
D) .006

Question 15

What is the significance of the "barrier condition" predictor variable?

A) .011
B) .026
C) .320
D) .078