Missing Data

Question 1

What is a pattern in missing data?

Answer 1

Describes the location of the missing values in a data set

Question 2

Describe a univariate pattern

Answer 2

Data are missing in only one of the variables in the analysis

Question 3

Describe a monotone pattern

Answer 3

Typically associated with longitudinal study where participants drop out and never return

Question 4

Describe an arbitrary/general pattern

Answer 4

Any set of variables may be missing for any subject

Question 5

Describe the missing at completely random (MCAR) assumption

Answer 5

Probability of missing values has nothing to do with what is observed or missing

Question 6

Describe the missing at random (MAR) assumption

Answer 6

Probability of missing values depends only on the observed values

Question 7

Describe the missing not at random (MNAR) assumption

Answer 7

Probability of missing values depends on the missing values themselves or on unmeasured variables, and in addition it can depend on observed values as well

Question 8

Outline the different levels of complication for inference between MCAR, MAR and MNAR

Answer 8

MCAR &laquo_space;MAR «««««< MNAR

Question 9

In terms of probability, define MCAR

Answer 9

Pr( R = r | Yobs, Ymis, ψ) = Pr ( R = r | ψ)

Question 10

In terms of probability, define MAR

Answer 10

Pr( R = r | Yobs, Ymis, ψ) = Pr ( R = r | Yobs, ψ)

Question 11

In terms of probability, define MNAR

Answer 11

Pr( R = r | Yobs, Ymis, ψ) = Pr ( R = r | Ymis, ψ)

Question 12

What is the key feature of MCAR?

Answer 12

Observed can be thought of as a random sample of complete data if no data is missing

Question 13

What is an important implication for the assumption of MAR

Answer 13

Data is MCAR after controlling for Yobs