This chapter will help you to understand how to evaluate psychological research. It begins by outlining some basic steps in the scientific process, which apply to all scientific research, not just that in psychology.

Once you start with an initial observation or question, you then form a hypothesis, or a tentative prediction or explanation about the phenomenon you are studying. You test the hypothesis by conducting research and analyzing the data, which enables you to determine if your hypothesis may be correct or must be rejected. Further research on the phenomenon then helps to build theory, a set of formal statements about how and why certain events are related. From the theory you derive new hypotheses and test them. People use two major approaches to understanding behavior: hindsight understanding (after-the-fact explanations) and understanding through prediction, control, and theory building (the scientific approach). While hindsight understanding is typical of everyday "common sense," psychologists use the scientific method to study and measure variables, operationally defining them in terms of the procedures used to measure them. Psychologists gather data related to a variable from self-reports, reports by others, physiological measures, or behavioral responses.

Psychologists use several methods of doing research. Descriptive research identifies how humans and animals behave. Case studies, which are in-depth analyses of individuals, groups, or events, are a type of descriptive research. Studies of animals or humans in natural settings are examples of naturalistic observation. In survey research, they administer questionnaires or interview people. Regardless of the method, it is important for the characteristics of the sample to represent the population so that the researchers can make accurate conclusions about the population; this is often accomplished through random sampling. Associations between variables are measured through correlational research. If two variables are correlated, that does not necessarily mean that one variable causes the other variable. The correlation coefficient is a statistic that measures the association between the variables: A positive correlation occurs when higher scores on one variable are associated with higher scores on another variable, while a negative correlation occurs when higher scores on one variable are associated with lower scores on another variable. The plus or minus sign of a correlation coefficient tells you the direction of the relationship, while the absolute value of the statistic tells you the strength of the relationship.

To determine cause-and-effect relationships, researchers must use experiments. In designing an experiment, a researcher intentionally manipulates one or more variables (the independent variable[s]) and measures whether this manipulation produces changes in a second variable or variables (the dependent variable[s]). The researcher attempts to control for other factors (extraneous or confounding variables) that might influence the outcome. In many experiments, participants are randomly assigned to either an experimental group (receives the treatment) or a control group (does not receive the treatment). In a double-blind procedure, the experimenter is unaware ("blind") as to whether a given participant belongs to the experimental group or the control group. To reach conclusions, the researcher compares the two groups statistically to see if there is any difference between them. Another way to design an experiment is to expose each participant to all conditions of the experiment.

Psychologists are particularly interested in assuring others that their research is valid. Internal validity represents the degree to which an experiment supports clear causal conclusions. It is important to try to rule out other factors that may have influenced the results so that the researcher can conclude that it was the manipulation of the independent variable, rather than some other factor, that produced changes in the dependent variable. If the researcher cannot do that, he or she has a problem with the confounding of variables. Some problems that researchers try to control are demand characteristics, or cues that participants pick up about how they are "supposed" to behave; placebo effects, in which participants' expectancies affect their behavior; and experimenter expectancy effects, by which researchers subtly and unintentionally influence the behavior of their participants through their actions. External validity refers to the extent to which the results of a particular study can be generalized to other people, settings, and conditions. Researchers rely on replication, or the ability of a study to be repeated with the same results, to determine the external validity of the findings. Meta-analysis is a statistical procedure for combining the results of different studies that examine a given topic.

Ethical standards are very important in psychological research. The American Psychological Association (APA) guideline of informed consent states that participants should be (1) given a full description of the procedures to be followed, (2) informed of any risks that might be involved, and (3) told that they have the right to withdraw from the study without penalty. Psychologists sometimes use deception studies, in which participants are misled about the nature of the study. These experiments are controversial but may be necessary to obtain natural, spontaneous responses from people. Ethical standards are also applied to research with animals. Although research with animals is controversial, most research psychologists would argue that it is a valuable way to further knowledge of both animal and human behavior.

In today's world, we encounter many oversimplifications, overgeneralizations, and pseudoscientific misinformation. You can avoid being misled by claims such as those in advertisements by applying the critical thinking skills you will learn in this course. As critical thinkers, we need to balance open-mindedness with healthy skepticism and evaluate evidence before deciding what to believe.

NOTE: For more detailed explanations of the statistical procedures outlined in this chapter, turn to the Statistical Appendix following Chapter 15 in your textbook.