Need these 3 questions below answered:

A. Outliers

How should a researcher define what a "normal" observation is? What are some examples of when being "abnormal" wouldn't justify an observation being deleted.

B. Missing Data

What are the pros and cons of deleting an observation with missing data vs employing an imputation method.

C. Transformations

Regarding the following statement: A researcher doesn't need to worry about transforming data when the assumption of normality is violated. (correct or not? and expand on the answer).