Selecting Human Participants for Research
35 Research Methods in Psychology anxiety treatments to different groups of people, experimenters who believe that one treatment will be the best might do a slightly better job with that treatment. When the results are in, this unintentional difference might make the favored treatment look better than the rest. To prevent experimenter bias from confounding results, experimenters often use a double-blind design. In this design, the participants and those giving the treatments are unaware of, or “blind” to, who is receiving a placebo, and they do not know what results are to be expected of the various treatments. Only researchers who have no direct contact with participants have this information, and they do not reveal it until the end of the experiment. Double-blind studies have not yet been conducted with EMDR, which is another reason for caution in drawing conclusions about this treatment. In summary, experiments are vital tools for examining cause-effect relationships between independent and dependent variables, but they are also vulnerable to error. To maximize the value of their experiments, scientists try to eliminate as many confounding variables as possible. They then repeat their work to ensure consistent results and adjust their interpretation of those results to take into account the limitations or problems that remain. Selecting Human Participants for Research double-blind design A research design in which neither the experimenter nor the participants know who is in the experimental group and who is in the control group. sampling The process of selecting participants who are members of the population that the researcher wishes to study. representative sample A sample of research participants chosen from a larger population such that their age, gender, ethnicity, and other characteristics are typical of that larger population. random sample A group of research participants selected from a population each of whose members had an equal chance of being chosen. biased sample A group of research participants selected from a population each of whose members did not have an equal chance of being chosen. Visitors from another planet would be wildly mistaken if they tried to describe the typical earthling after meeting only Arnold Schwarzenegger, Michael Jackson, Madonna, and a trained seal. Psychologists, too, can be led astray if the participants they encounter in their research are not typical of the people or animals about which they want to draw conclusions. Accordingly, one of the most vital steps in scientiﬁc research is the selection of participants, a process called sampling. If they want to make accurate statements about the behavior and mental processes of any large group, psychologists must select a representative sample of participants whose characteristics mirror the rest of that group in terms of age, gender, ethnicity, cultural background, socioeconomic status, sexual orientation, disability, and the like. In theory, psychologists could draw representative samples—of people in general, of Canadians, of Florida college students, or of any other group—by choosing them at random from the entire population of interest. Doing this would require putting hundreds of thousands, perhaps millions, of names into a computer, running a program to randomly select participants, then tracking them down to invite them to take part in the research. This method would result in a truly random sample, because every member of the population to be studied would have an equal chance of being chosen. (Any selection procedure that does not offer this equal chance is said to result in a biased sample.) However, random sampling is usually too expensive and time-consuming to be practical, so psychologists may have to ﬁnd their participants in populations that are more conveniently available. The populations from which these convenience samples are drawn depend to some extent on the size of the researcher’s budget. They might include, for example, the students enrolled in a particular course, the students enrolled on a local campus, the students who are willing to sign up for a study, or visitors to Internet web sites or chat rooms (e.g., Nosek, Banaji, & Greenwald, 2002; Stone & Pennebaker, 2002). Ideally, this selection process will yield a sample that fairly represents the population from which it was drawn. Still, scientiﬁc researchers are obliged to limit the conclusions they draw in light of the samples they draw (Kraut et al., 2004). Because of this obligation, psychologists often conduct additional studies to determine the extent to which their initial conclusions will apply to people who differ in important ways from their original sample (APA Ofﬁce of Ethnic Minority Affairs, 2000; Case & Smith, 2000; Gray-Little & Hafdahl, 2000). For a recap of the strategies psychologists use in their research efforts, see “In Review: Methods of Psychological Research.” 36 Chapter 1 Introduction to the Science of Psychology doing 2 learn by SELECTING RESEARCH PARTICIPANTS Suppose that you want to study people’s willingness to help each other. You have developed a way of measuring helpfulness, but now you need a sample of people to test. Take a minute to think about the steps necessary to select a truly random sample; then ask yourself how you might obtain a representative sample instead. Remember that although the names are similar, random sampling is not the same as random assignment. Random sampling helps ensure that the people studied are representative of some larger group. Random assignment is used in experiments to create equivalence among various groups. Online Study Center in review Improve Your Grade Tutorial: Research Methodologies METHODS OF PSYCHOLOGIC AL RESEARCH Method Features Strengths Pitfalls Naturalistic observation Observation of human or animal behavior in the environment in which it typically occurs Provides descriptive data about behavior presumably uncontaminated by outside inﬂuences Observer bias and participant self-consciousness can distort results. Case studies Intensive examination of the behavior and mental processes associated with a speciﬁc person or situation Provide detailed descriptive analyses of new, complex, or rare phenomena May not provide representative picture of phenomena. Surveys Standard sets of questions asked of a large number of participants Gather large amounts of descriptive data relatively quickly and inexpensively Sampling errors, poorly phrased questions, and response biases can distort results. Correlational studies Examine relationships between research variables Can test predictions, evaluate theories, and suggest new hypotheses Cannot conﬁrm causal relationships between variables. Experiments Manipulation of an independent variable and measurement of its effects on a dependent variable Can establish a cause-effect relationship between independent and dependent variables Confounding variables may prevent valid conclusions. ? 1. The method is most likely to use a double-blind design. 2. Research on a new treatment method is most likely to begin with . 3. Studying language by listening to people in public places is an example of research.