Unlike exploratory qualitative studies, quantitative research requires that the variables posed in research questions be preselected. If earlier research suggested that family size, birth order, gender and age affected aggression in play, then the researcher should measure all these variables. In this way, qualitative, observational research is useful in identifying variables to be studied in a rigorous, quantitative manner in later research.
In designing research questions, the researcher must ensure that definitions of the variables to be measured are precise. For example, in the research design above, researchers must decide how to define an "aggressive act." The definition should be clear, unbiased and able to be accurately measured. This ensures that the research is consistent and therefore reliable.
Research questions must not only be consistently scored the same by different researchers, but the measure must also be valid. For example, different researchers may all score an act as aggressive, but may misunderstand the intent of the child or of the game being played; therefore, the measure, while consistent, is not valid. Some gestures, such as high-fives, may look like physical slaps but may have a different meaning altogether for children at play. In this example, labeling this action aggressive is not valid.
Quantitative research often involves contacting large groups of people to ensure representative samples. A good researcher designs questions for surveys that elicit high response rates to avoid bias. Questionnaires must not be too long or cumbersome to answer. Closed-ended questions that direct responders to choose an answer are popular, but may prove to be invalid if not carefully worded.