Quantitative vs. Qualitative Research

Research can be approached in two ways- through a qualitative study or a quantitative study – depending on the type of problem the researcher needs to research. The researcher’s choice of one of these approaches will shape the procedures to be use in each step of the research process.

In collecting data, the quantitative research uses an instrument to measure the variables in the study. An instrument is a tool for measuring, observing, or documenting quantitative data. It contains specific questions and response possibilities that you establish or develop in advance of the study. Examples of instruments are survey questionnaires, standardized tests, and checklists that you might use to observe behaviors among participants. The researchers administer this instrument to the participants, and collect data in the form of numbers.

For instance, he might collect responses based on students checking boxes on a form, or from checklists you complete as you watch a student perform a task in the classroom. The intent of this process is to apply the results (called generalizing the results) from a small number of people to a large number. The larger the number of individuals studied, the stronger is the case for applying the results to a large number of people. In qualitative research on the other hand, the researcher does not begin data collection with a preestablished instrument to measure distinct variables.

Instead, he seeks to learn from the participants in the study, and develop forms, called protocols, for recording data as the study proceeds. These forms pose general questions so that the participants can provide answers to the questions. Often questions on these forms will change and emerge during data collection. Examples of these forms include an interview protocol, which consists of four or five questions, and an observational protocol, in which the researcher records notes about the behavior of participants. Moreover, you gather text (word) or image (picture) data (Swanborn, 1996).

A major concern that is often raised in research is the issue of reliability, validity and accuracy of a measure.  Reliability is the quality of consistent measurement and the most fundamental test of reliability is “repeatability;” the ability to get the same data values from several measurements made in similar manner (Lincoln and Guba, 1985).  Most researchers would say that a study is reliable if it gives the same measurement to the same individuals or groups, that is, if the test is given twice to the same individuals or groups after the lapse of a certain period.

For example, if a survey question (i.e. shyness surveys from www. Shyness.com) obtained the same response, time after time, from one particular respondent who asked that same question several times in one month that would mean that there was high reliability over time.  If the answers varied in a random pattern, the reliability over time would be low.  If a survey item yielded the same data from one respondent to the next when they did, in fact, hold identical positions on the issue, that would indicate high reliability over respondents, and if their were random differences, reliability over respondents would be low.

For a measure to be reliable, in other words, repeated measurements must produce the same result.  In addition, one strategy for limiting distortion caused by personal values is replication, repetition of research by others in order to assess its accuracy.  If other researchers repeat a study using the same procedures and obtain the same results, everyone gains confidence that the original research (as well as the replication) was conducted objectively.  The need for replication in scientific investigation is probably the reason that the search for knowledge is called research in the first place.

But only a small proportion of social science research is actually subjected to replication (far less than is the case in the natural sciences).  If a replication does not square with the results of an original study, the difference may spark conflict.  So, the general way of approaching the reliability problem is to make as many steps as possible as operational as possible, and for the quantitative researcher to conduct research as if someone were always looking over the shoulders.

But even consistency of results or reliability is no guarantee of validity. Validity means the quality of measuring precisely what one intends to measure.  Say you want to investigate how shy people are, so you ask how often they feel shy being shy.  Notice that, in trying to gauge shyness, what you are actually measuring is fear to do something, which may or may not amount to the same thing.  Thus, even when a measurement yields consistent results (making it reliable), it can still miss the real, intended target (and lack validity).

So a survey is valid to the degree that it measures what and only what it is supposed to measure.  To be valid it must not be affected by the extraneous factors that systematically “push” or “pull” the results in one particular direction.  To the degree that things other than those being measured affect the results by introducing a systematic bias, the results are less valid.

Unfortunately, there are many factors that can bias the result of a survey, including the effect of sampling.  Accuracy on the other hand, is a measure of the degree of trustworthiness of the research. Research which is highly reliable and valid has greater accuracy while research which has low reliability and validity has lesser accuracy.   In sum, research is no better than the quality of measurement that it employs (Lincoln and Guba, 1985).

References

Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Beverly Hills, CA: Sage Publications,         Inc.

Swanborn, Peter G. (1996). A common base for quality control criteria in quantitative and qualitative research. Quality and Quantity, 30:19-35.