Deck 20: Categorical Outcomes: Logistic Regression

Question 1

Which of the following is not a research question which could be investigated with logistic regression?

A) Prediction of group membership
B) Strength of association between criterion and predictors
C) Differences between groups
D) Interaction between predictors

Answer 1

Differences between groups

Question 2

What does the following graph illustrate? (PD is peritoneal dialysis.)

Answer 2

C

Question 3

The odds ratio is

A) The natural log to the bth power, where b is the unstandardized parameter estimate.
B) The natural log to the bth power, where b is the standardized parameter estimate.
C) The base 10 log to the bth power where b is the unstandardized parameter estimate.
D) The base 10 log to the bth power where b is the standardized parameter estimate.

Answer 3

The natural log to the bth power, where b is the unstandardized parameter estimate.

Question 4

What does the log-likelihood measure?

A) The number of degrees of freedom in relation to the parameter set for the model
B) The way in which observed and predicted values fit the model of regression*
C) The value of the correlation coefficient R and its corresponding R² value
D) None of the above

Question 5

What does the Wald statistic represent?

A) The predicted value of the correlation coefficient
B) The proportionate change in the odds ratio
C) The odds ratio of the regression
D) Whether a variable is a significant predictor of the outcome

Question 6

What is an alternative to logistic regression when the dependent variable has more than two classes?

A) Binomial regression
B) Logit regression
C) Legit regression
D) Multinomial regression

Question 7

Large log-likelihood values do not show

A) The extent of the relationship between independent and dependent variables.
B) That there are a large number of unexplained observations in the data.
C) That there is a large measure of deviance.
D) That the model is a good fit.

Question 8

Logistic regression

A) Estimates values in the outcome variable.
B) Estimates a probability that the outcome variable assumes a certain value.
C) Estimates probabilities in the outcome variable.
D) Estimates the probability that the predictor variables assume certain values.

Question 9

In logistic regression, the dependent variable is called a logit, which is

A) The natural log of the odds
B) The rolling mean
C) A log rolling
D) A mean log

Question 10

How would you request a logistic regression analysis in SPSS?

A) Analyze → Nonparametric → Binomial
B) Analyze → Regression → Ordinal
C) Analyze → Regression → Binary Logistic
D) Analyze → Loglinear → General

Question 11

Logistic regression is used when you want to

A) Predict a continuous variable from dichotomous ones.
B) Predict a dichotomous variable from continuous or dichotomous variables.
C) Predict any categorical variable from other categorical variables.
D) Predict a continuous variable from dichotomous or continuous variables.

Question 12

What does logistic regression not do?

A) Predict a dependent variable on the basis of several independent variables.
B) Determine the variance in the dependent variable explained by the independent variables.
C) Rank the relative importance of independent variables.
D) Assess the impact of covariate control variables.

Question 13

What is suggested by very high standard errors in parameter estimates in logistic regression?

A) Shrinkage
B) Multicollinearity
C) An outlier
D) A dummy variable

Question 14

Which of the following is not an assumption of logistic regression?

A) Linearity
B) Independence of errors
C) Multicollinearity
D) None of the above

Question 15

In the above example, the mean is also equal to

A) The mode
B) The probability of randomly drawing a 1
C) The median
D) The probability of drawing a 0

Question 16

Which of the following statements is not true about the Wald statistic?

A) The Wald statistic assesses the individual contribution of a predictor to a logistic regression model.
B) The Wald statistic tends to be biased when the regression coefficient is large.
C) If the Wald statistic is equivalent to the t-statistic in multiple regression.
D) The Wald statistic has a t-distribution.

Question 17

The odds of an event are

A) The ratio of the probability of an event not happening to the probability of the event happening.
B) The ratio of the probability of an event happening to the probability of the event not happening.
C) The probability of an event occurring.
D) None of the above.

Question 18

In determining the odds that a specific type of antibiotic preparation will cure otitis media in children, the odds ratio value was calculated to be greater than 1. With increases in the predictor, what will this mean for the outcome of the result?

A) A greater chance of a cure being achieved than if not treated.
B) A lesser chance of a cure being achieved.
C) A lesser chance of treatment being effective.
D) None of the above.

Question 19

A researcher was interested in predicting whether a person would attempt to commit suicide (score = 1) or not (score = 0) from their depression scores. They found that the value of exp b was 2.56. How should this value be interpreted?

A) If two people have depression scores that differ by 1 unit, then the odds of the person with the higher score attempting suicide are 2.56 higher than for the other person.
B) 2.56 times more people who attempted suicide scored highly on depression.
C) If two people have depression scores that differ by 1 unit, then the odds of the person with the higher score attempting suicide are 2.56 lower than for the other person.
D) Suicide rates are 2.56 times as high in depressed people.

Question 20

With conventional coding in logistic regression of 0 and 1, if there are 100 cases and 40 are coded as 1, what is the mean of the variable?

A) 40
B) .4
C) 2.5
D) 25

Question 21

Which of the following is not used in logistic regression?

A) Forced entry
B) Forward method
C) Predictor method
D) Backward method

Question 22

Collinearity diagnostics can be used for what purpose?

A) To determine which predictor is causing the lowest value.
B) To determine which predictor is causing the highest value.
C) To determine the accuracy of the regression model.
D) To determine which predictor is causing the data to be skewed.

Question 23

Multinomial logistic regression might be used to predict which of the following examples

A) Men prefer to use non-scented ointment in the treatment of fungal infection, whilst women prefer to use scented powder varieties in preference to scented ointment.
B) Men use non-scented ointment to treat fungal infection.
C) Women use non-scented powder to treat fungal infection.
D) Women use scented powder to treat fungal infection whilst men use non-scented powder.

Question 24

Which of the following reasons best describes why overdispersion is a problem in logistic regression?

A) Incorrect correlations may be assumed.
B) The standard error will be limited and may cause the test statistic to be too large and more likely to become significant.
C) The outliers will not significantly affect the results.
D) None of the above.