Week 7

Question 1

What do you use to test a random pattern of missing data?

Answer 1

Littles MCR

Question 2

Explain Littles MCR:

Answer 2

Littles compares your data against a matrix which comprises of data that is missing complete at random. It sees how much your data set differs from this matrix. You want your data to differ so you can say that your data is not missing completely at random.

Question 3

If you have a small % of missing data you can use?

Answer 3

Listwise or pairwise deletion.

Question 4

Explain Listwise?

Answer 4

Deletes the whole case for any case that has any missing data. None of it will be used in the analysis.

Question 5

Explain Pairwise

Answer 5

, just removes the case in any analysis that requires the data.
So if you a person filled out question 1 and 2 but not 3. Pairwise would not use that persons data only for the analysis that involves question 3.

Question 6

Explain mean substitution

Answer 6

Replaces the missing data with the mean.

Question 7

Explain estimation by regression

Answer 7

which treats the missing data like a DV. It utilises the info that is known about that person with missing data.

Question 8

Explain Expectation Maximization

Answer 8

It fills in the missing data according to a normal distribution of the data. In order to do this the data must be missing at random.

Question 9

Explain Multiple Imputation

Answer 9

Makes no assumptions about the randomness of the missing data. It is the gold standard of filling in the missing data. However it is more complex than other methods.

Question 10

Violation of Assumptions introduce the following bias into the analysis:

Answer 10

1. Biased parameter estimate
2. Biased standard errors and confidence interval
3. Biased test statistics and p-values

Question 11

What are the 2 main types of outliers?

Answer 11

Univariate and Multivariate

Question 12

Explain Multivariate Outliers

Answer 12

A person who’s pattern of scores on two or more variables is very different from the sample. LeBron, when the combination of the variables are outliers.

Question 13

Explain Univariate Outliers

Answer 13

A persons score on ONE variable is very high or very low compared to the other participants

Question 14

Explain LINEARITY

Answer 14

The assumption is that there is a straight line relationship between two variables. You need to check for nonlinear relationship between the variables. You do this by using scatterplots.

Question 15

Explain NORMALITY

Answer 15

The assumption is not that the distribution of variables in the data set have to be normal

Question 16

What are the two types of normality?

Answer 16

1. Normality in the sampling distribution for a parameter

2. Normality in the residuals

Question 17

Explain HOMOSCEDASTICITY

Answer 17

Estimates the average variance for the sample of both populations. You do this because the means are different (one mean is not represented in one sample) So you pool both data in order to get a better estimate of the standard deviation. This makes it hard to make a sig conclusion

Question 18

Explain Multicollinearity and Singularity

Answer 18

Singularity- when two variable are perfectly correlated

Multicollinearity- Two variables are highly correlated