SUMMARY OF CHAPTER 4
Correlational research represents a general approach to research that focuses on assessing the covariation among naturally occurring variables. The goal of correlational research is to identify predictive relationships by using correlations or more sophisticated statistical techniques. The results of correlational research also have implications for decision making, as reflected in the appropriate use of actuarial prediction. The greatest limitation of correlational research is the problem of interpreting causal relationships.
Survey research illustrates the principles of correlational research and it provides an accurate and efficient means for describing people's thoughts, opinions, and feelings. Surveys differ in purpose and scope, but they generally involve sampling. Results obtained for a carefully selected sample are used to describe the entire population of interest. Surveys also involve the use of a predetermined set of questions, generally in the form of a questionnaire.
Sampling is a procedure whereby a specified number of elements are drawn from a sampling frame that represents an actual list of the possible elements in the population. Our ability to generalize from the sample to the population depends critically on the representativeness of the sample, the extent to which the sample has the same characteristics as the population. Representativeness is best achieved by using probability sampling rather than nonprobability sampling. In simple random sampling, the most common type of probability sampling, every element is equally likely to be included in the sample. Stratified random sampling is used when analysis of subsamples is of interest.
There are three general survey methods: mail surveys, personal interviews, and telephone interviews. Mail surveys avoid problems of interviewer bias and are especially well suited for examining personal or embarrassing topics. The problem of response bias is a serious limitation of mail surveys. Personal interviews and phone surveys usually have higher response rates and provide greater flexibility. The phone survey is the method of choice for most brief surveys.
Survey research is carried out according to an overall plan called a research design. There are three survey research designs: the cross-sectional design, the successive independent samples design, and the longitudinal design. Cross-sectional designs focus on describing the characteristics of a population or the differences between two or more populations at one point in time. Describing changes in attitudes or opinions over time requires the use of successive independent samples or longitudinal designs. The longitudinal design is generally preferred because it allows the researcher to assess changes for specific individuals and avoids the problem of noncomparable successive samples.
Survey results, like those of other verbal reports, can be accepted at face value unless there is reason to do otherwise, such as pressures on respondents to give socially desirable responses. People's behavior does not always conform to what they say they would do, so survey research will never replace direct observation. However, survey research does provide an excellent way to examine people's attitudes and opinions.