For example, a researcher who is studying happiness and optimism might find that a secondary data set only includes one of these variables, but not both. A related problem is that the variables may have been defined or categorized differently than the researcher would have chosen.