Data Analysis Bootcamp

Question 1

Linear regression can accommodate which kind of indpendent variable(s)?

Accepted Answer

Categorical, binary, continuous

Question 2

For linear regression, which of the following graphs can be used to evaluate homoskedasticity?

Accepted Answer

Scatter plot, residual vs. fitted plot

Question 3

What does the model intercept represent in linear regression?

Accepted Answer

Where the line of best fit cross the Y axis

Question 4

Logistic regression tests the assocation between weight (X, continuous) and diabetes (Y). An odds ratio of 2.5 means…?

Accepted Answer

For every unit increase in weight, odds of diabetes increase by 2.5

Question 5

Which of the following are threat(s) to statistical conclusion validity?

Accepted Answer

Low statistical power, violated test assumptions, unreliable measures

Question 6

Logistic regression tests the assocation between obesity (X) and diabetes (Y). An odds ratio of 2.5 means…?

Accepted Answer

Obese vs. non-obese persons have 2.5 times the odds of diabetes

Question 7

Which of the following is NOT a type of measurement validity?

Accepted Answer

Statistical validity

Question 8

Which statistical test is appropriate for comparing means between two groups?

Accepted Answer

Independent sample t-test

Question 9

A logistic regression model includes an independent variable with four categories. How many dummy coded variables?

Accepted Answer

3

Question 10

When reporting results from a linear regression model, it is best practice to include…

Accepted Answer

Regression coefficient, confidence interval, p-value

Question 11

Which kind of dependent variable(s) is appropriate for logistic regresssion?

Accepted Answer

Binary

Question 12

What is statistical power?

Accepted Answer

Probability of choosing an alternative hypothesis when it is true

Question 13

Which of the following is an assumption of the chi-square test?

Accepted Answer

Each cell has expected frequency ≥ 5

Question 14

What is the advantage of multiple vs. simple linear or logistic regression?

Accepted Answer

To adjust or control for other covariates

Question 15

For logistic regression, which of the following tests can be used to evaluate linearity?

Accepted Answer

Box-Tidwell

Question 16

A linear regression model includes sex (X) and salary (Y). How do you test age as a moderator?

Accepted Answer

Include an interaction term of age*sex

Question 17

Which of the following factors influence statistical power?

Accepted Answer

Sample size, effect size

Question 18

Confidence intervals that do not overlap suggest there is a statistically significant difference between two groups.

Accepted Answer

TRUE

Question 19

Before conducting a statistical test, it is best practice to…

Accepted Answer

All of the above

Question 20

For linear regression, which of the following graphs can be used to evaluate normality?

Accepted Answer

Histogram, Q-Q plot

Question 21

Linear regression tests the association between height (X) and weight (Y). A regression coefficient of 1.5 means...?

Accepted Answer

For a 1-unit increase in height, on average, weight increases by 1.5 units

Question 22

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

Accepted Answer

Variance of Y should be roughly the same for each value of X

Question 23

When reporting results from a logistic regression model, it is best practice to include…

Accepted Answer

Odds ratio, confidence interval, p-value

Question 24

In logistic regression, which of the following graphs can be used to evaluate linearity?

Accepted Answer

Lowess plot

Question 25

The dummy coded variable NOT included in a regression model is…

Accepted Answer

The reference or comparison group

Question 26

Which of the following is NOT an assumption of linear regression?

Accepted Answer

Residuality

Question 27

What does the model intercept represent in logistic regression?

Accepted Answer

Log odds of Y when X=0

Question 28

Which of the following is a strategy for reducing measurement error?

Accepted Answer

Pilot test survey, train interviewers

Question 29

Missing completely at random (MCAR) means there are no systemic differences in persons with missing vs. complete data.

Accepted Answer

TRUE

Data Analysis Bootcamp

Questions 30

Data Analysis Bootcamp

Questions 30