Quantitative data is numerical data, whereas qualitative data has no numbers attached to it. The gender of respondents in a study, dividing light bulbs into categories like "very bright," "somewhat bright" and "dim," or the type of pizza a customer prefers are all examples of qualitative data. If you say, by contrast, that 51 percent of the plants tested grew to 10 inches or more, while 33 percent grew to 5 inches or less, you are looking at quantitative data.

Qualitative data is by definition non-numerical, but qualitative data can sometimes be assembled to provide quantitative data. For example, if customers in a survey describe how they feel about a food item they purchased, the questionnaire would only provide qualitative data. If the individual questionnaire results were compiled to determine how many or what percentage of customers prefer pepperoni to anchovies, however, the survey would now have provided some quantitative data as well.

Some lab tests provide qualitative results and others quantitative. A procedure called a Western blot, for example, typically provides only qualitative data -- whether or not a particular protein was present, but not how much of it was present. Another common test called the enzyme-linked immunosorbent assay (ELISA) can be performed using approaches that will provide either qualitative or quantitative results. The typical pregnancy test is qualitative; it tests for the presence of human chorionic gonadotrophin (HCG) in the patient's urine, but does not quantify the amount present.

Sometimes qualitative data is preferable. If you are doing a pregnancy test, for example, you know that if high levels of HCG are present, it means you're almost undoubtedly pregnant. You're not really trying to find out precisely what the level of HCG is -- you want a yes-or-no answer, not a numerical answer that will be more difficult for you to interpret. Likewise, if you are testing to determine whether blood samples from a patient are HIV-positive, the patient and her physician want a yes-or-no answer and not a numerical one.

In other experiments or lab tests, quantitative data is preferable. If biochemists are working on determining the isoelectric point of an enzyme (the pH at which it has no net charge), they want a quantitative, numerical answer. Likewise, if you were to test positive for HIV and your doctor ordered a viral load test, a test that gives the amount of virus present per given unit of body fluid, she would be trying to obtain quantitative data for use in planning your treatment.