Types of Research Design

Most research in education can be classified as one of two types - descriptive studies and those aimed at discovering causal relationship. Descriptive studies are primarily concerned with finding out "what is". Examples of questions that might be studied by means of a descriptive approach are: Do teachers hold favorable attitudes toward the new mathematics? What kinds of sctivities occur in sixth-grade art classes and how frequently do they occur? Observational and survey methods are frequently used to collect descriptive data.



Although description is an important aspect of the scientific approach in education, most research studies are concerned primarily with discovering causal relationship. Typical causal problems investigated by educational researchers are: What factors determine choice of college major? Does this new instructional strategy lead to increased learning when compared with conventional instructional strategies?

We may distinguish between research designs in terms of their effectiveness in establishing causal links between two or more variables. The causal comparative method is aimed at the discovery of possible causes for the phenomenon being studied by comparing subjects in whom a characteristics is present with similar subjects in whom it is absent or present to a lesser degree. However, this research design can only be used to explore causal relationships, not confirm them.

Suppose that one interested in testing the hypothesis that anxiety impairs performance on timed aptitude tests. If a causal comparative design were used, one might select contrasting groups of high and low anxious students, and then compare their performance on a timed aptitude test. Suppose we found that the high anxious group did indeed have lower test scores on the average than the low anxious group. Although this finding is consistent with the research hypothesis, an alternative causal hypothesis is possible, namely, that poor performance on timed aptitude tests is likely cause anxiety. Thus, the research result do not tell us whether anxiety causes impaired performance or whether impaired performance causes anxiety.

A similar problem occurs in correlational designs. Correlational studies include all research projects in which an attempt is made to discover or clarify relationship through the use of correlational coefficients. The experimental research design is ideally sited to establish causal relationships if proper control are used.To be continued....