Bias

Question 1

What can bias be detrimental for? (3)

Answer 1

Bias can be detrimental for the parameter estimates, standard errors and confidence intervals and the test statistics and p-values.

Question 2

What is meant by an outlier?

Answer 2

An outlier is a score that is very different from the existing data

Question 3

What can be used in order to spot outliers?

Answer 3

standardized residuals can be used.

Question 4

What are the four ways of correcting problems with data?

Answer 4

Trimming the data (deleting a certain quantity of scores from extremes)
Winsorzing (Substituting outliers with the highest value that is not an outlier)
Applying a robust estimation method (such as bootstrapping)
Transforming the data (applying a mathematical function to data to correct for problems )

Question 5

What can trimming the data be based on? (2)

Answer 5

Percentages of extreme data or based on the amount of standard deviation a value is away from the mean

Question 6

What percentage of the data must be transformed if a transformation is applied?

Answer 6

100%