What are some ways to handle missing data in a study? How should these data be coded in a database?

Question

What will be an ideal response?

Answer 1

Answers to this question will vary but should include reference to the following:
(a) making decisions about whether those who don't answer a particular question are systematically different from those who do; subsequently, the researcher can opt to not do analyses involving that variable or go ahead with those participants who responded to it—keeping in mind that results will only apply to the kinds of people likely to answer such questions; (b) if the missing values seem random, one can substitute the average of everyone else's score on that variable for each missing value; (c) if the study has groups, one can substitute the value for the
average of that group; (d) if a person has a missing value on an item that is part of a multi-item measure, one might decide to substitute that person's average on the items that were completed. It is usually a good idea to create a new variable in your database that includes the substituted values—keeping the original variable as it was. This ensures that the researcher will not lose track of when he or she substituted values.