Lecture 2: Missing Data (Alternate)

Question 1

What can happen if missing data is not handled properly in psychological research?

Answer 1

It can distort research findings.

Question 2

What example illustrates stigma-driven nonresponse affecting data accuracy?

Answer 2

Reporting of lifetime sexual partners.

Question 3

What does missing data refer to?

Answer 3

Omitted responses on specific variables.

Question 4

What does selection bias refer to?

Answer 4

Systematic differences in who participates in the study.

Question 5

What is a consequence of selection bias in research?

Answer 5

Spurious correlations.

Question 6

What causes spurious correlations in selection bias examples?

Answer 6

Joint selection on two variables.

Question 7

What are common sources of missing data in surveys?

Answer 7

Incomplete questionnaires or selective nonresponse.

Question 8

What two things are important to diagnose in missing data?

Answer 8

Extent and type.

Question 9

What does MCAR stand for?

Answer 9

Missing Completely at Random.

Question 10

What is an example of MCAR?

Answer 10

Data lost due to a random technical glitch.

Question 11

Does MCAR bias estimates?

Answer 11

No.

Question 12

What does MAR stand for?

Answer 12

Missing at Random.

Question 13

What is an example of MAR?

Answer 13

Older participants omitting sexuality questions.

Question 14

What does MNAR stand for?

Answer 14

Missing Not at Random.

Question 15

What is an example of MNAR?

Answer 15

High sexual activity participants omitting partner questions.

Question 16

Which type of missing data biases parameter estimates?

Answer 16

MNAR.

Question 17

What statistical test can assess randomness in missing data?

Answer 17

Little’s MCAR test.

Question 18

What does a significant result in Little’s MCAR test imply?

Answer 18

Systematic missingness.

Question 19

What can variance t-tests help identify in missing data?

Answer 19

Predictors of missingness.

Question 20

What is listwise deletion?

Answer 20

Removing any case with missing data from all analyses.

Question 21

When is listwise deletion considered acceptable?

Answer 21

When less than 5% of data is missing.

Question 22

What is pairwise deletion?

Answer 22

Removing cases only from analyses involving the missing variable.

Question 23

What is a risk of pairwise deletion?

Answer 23

Inconsistent results across analyses.

Question 24

What is mean substitution?

Answer 24

Replacing missing values with the variable mean.

Question 25

What is a problem with mean substitution?

Answer 25

It underestimates variability.

Question 26

What is regression substitution?

Answer 26

Predicting missing values using other variables.

Question 27

What is a limitation of regression substitution?

Answer 27

It underestimates standard errors.

Question 28

What does EM use to estimate missing values?

Answer 28

Known means, variances, and covariances.

Question 29

What does EM generate through regression modelling?

Answer 29

Imputations.

Question 30

What marks the end of EM iteration?

Answer 30

Stable estimates matching original values.

Question 31

What does MI add to EM’s process?

Answer 31

Random error and multiple datasets.

Question 32

How are final estimates in MI calculated?

Answer 32

Averaged across imputed datasets.

Question 33

What is a major benefit of MI?

Answer 33

Preserves standard errors and reflects uncertainty.

Question 34

What is considered the gold standard for handling missing data?

Answer 34

Multiple Imputation.

Question 35

Can statistical methods fully correct MNAR bias?

Answer 35

No.

Question 36

What are three key practices for MNAR data?

Answer 36

Theoretical justification, sensitivity analysis, and transparency.

Question 37

What is the difference between bias and precision loss?

Answer 37

Bias is systematic error; precision loss is reduced power.

Question 38

What should researchers do when reporting missing data handling?

Answer 38

Document methods, decisions, and assumptions clearly.