Chapter 2:

Methodology: The Process of Doing Research

Outline

Major Issues

I. Formulating Hypotheses and theories

II. The Observational Method: Describing Social Behavior

III. The Correlational Method: Predicting Social Behavior

IV. The Experimental Method: Answering Causal Questions

V. Basic Versus Applied Research

VI. Ethical Issues in Social Psychology

Methodology: The Process of Doing Research

• Empirical research allows us to test the validity of personal observations and folk wisdom

• Findings from social psychology may appear obvious because of the hindsight bias - the "I knew it all along" phenomenon

• Social psychology relies on 3 types of methods - observational, correlational, & experimental - to provide empirical answers to questions about behavior

I. Formulating Hypotheses and Theories

A. Inspiration from Previous Theories and Research

• Science is cumulative and so people often generate hypotheses based on previous theories & research, e.g., Festinger's cognitive dissonance theory inspired others

B. Hypotheses based on Personal Observations

• Personal experience, current events, & literature can provide sources of testable hypotheses, e.g., the Kitty Genovese case led to testing conditions under which people help

II. The Observational Method: Describing Social Behavior

• Involves systematic observation & measurement of behavior

A. Systematic Observation

• Involves a trained social scientist who observes & codes behavior according to a prearranged set of criteria

B. Participant Observation

• A form of systematic observation whereby the observer interacts with the people being observed, but tries not to change the situation in any way

C. Archival Analysis

• A variant of systematic observation in which the researcher observes social behavior by examining accumulated documents, e.g., newspapers, of a culture

• In archival analysis, well-defined coding categories are created and interjudge reliability is measured

1. Archival Analysis and Pornography

• Observational research can tell a lot about a society's beliefs & values.

• Some studies of pornography reveal high levels of depicted sexual violence by men against women, suggesting its appeal to readers

D. Limits of the Observational Method

• Certain types of behavior are difficult to observe because they are rare or occur in private

• Observational research is confined to one particular group of people, one particular setting, and one particular type of activity, limiting generalizability

• It describes patterns of behavior but does not allow for understanding of relationship or predicting when different kinds of behavior will occur

III. The Correlational Method

• Entails systematically measuring the relationship between two or more variables. Often relies on surveys as well as observations

A. Surveys

• Used when the variable of interest is not easily observable

• The correlation coefficient provides a metric for calculating degrees of association b/w 2 variables. Correlations can range from -1 to +1

• These represent two independent attributes. The sign indicates the direction and the values represent the magnitude or strength of the association

• The validity of survey data depends on using samples that are representative of the population under study

• Random selection can ensure sample representativeness

1. Problems with Survey Data

• Sampling errors, e.g., the 1936 Literary Digest presidential poll fiasco

• Questions that ask people to predict or explain their own behavior are likely to be inaccurate

• Answers to survey questions can be influenced by the way the question is phrased and ordered

B. Correlation Does Not Equal Causation

• Correlation identifies whether 2 variables are associated but not why. Always three possible interpretations

• Confusion of correlation & causality may turn up in media reports

• Only the experimental method can determine causality

IV. The Experimental Method: Answering Causal Questions

A. Independent and Dependent Variables

• The independent variable is that which is manipulated by the researcher - the presumed cause of the change in the other variable

• The dependent variable is the one measured by the researcher to see if changes depend on the level of the independent variable

• Factorial designs use multiple independent variables to study their joint and combined effects on the dependent variable

B. Establishing Internal Validity in Experiments

• An experiment has high internal validity when everything is the same in the different levels of the independent variable, except for the factor of interest

• Internal validity is established by controlling all extraneous variables and by using random assignment of participants to conditions

• Even with random assignment, there is small chance that different characteristics of people may be unevenly distributed across conditions

• To guard against misinterpreting the results in such an event, scientists calculate the probability level (p value) that their experimental results would occur by chance

• By convention, a p-value of .05, i.e., less than or equal to 5 chances in 100 that an event would occur by chance is considered to be statistically significant. E.g., the experiment by Donnerstein (1980) testing the relationship between pornography and aggression

C. Establishing External Validity in Experiments

1. Generalizability Across Situations

• There is always a concern that lab. situations are artificial and therefore don't generalize to real life. There are ways in which an experiment can be realistic

• Mundane realism - the extent to which an experiment is similar to situations encountered in real life

• Psychological/experimental realism - the extent to which the psychological processes triggered are similar to the psychological processes occurring in real life

• Psychological realism often depends deception - a false description of the purpose of the study

2. Generalizability Across People

• The only way to be certain that the results of an experiment represent the beh. of a certain population is to randomly select from that population

• However, this could be impractical and expensive. Social psychologists often assume that the psychological processes studied are basic components of human nature & thus similar across different populations

• To be truly certain, however, studies should be replicated with different populations

3. Replications

• Refer to the repetition of a study, often with populations or in different settings. Provide the ultimate test of an experiment's external validity

• Many findings discussed in the text have been replicated with different populations and/or settings

4. Cross-Cultural Research

• Has 2 main goals: demonstrating the universality of a particular psychological process, and demonstrating differences between cultures

• One challenge in cross-cultural research is determining when differences in results are due to different psychological processes or due to different construals of the situation that lead to different psychological processes operating

D. The Basic Dilemma of the Social Psychologist

• One of the best ways to increase external validity is through field experiments - conducted in real world settings

• However, there is often a trade-off between internal and external validity

• The resolution to this dilemma is the use of replication in both laboratory and field settings

V. Basic Versus Applied Research

• Basic research tries to find the best answer to the question of why people behave the way they do, purely for the sake of curiosity, i.e., knowledge for its own sake

• Applied research is geared toward solving a specific social problem. In practice, this distinction is often fuzzy

VI. Ethical Issues in Social Psychology

• Social psychologists are faced with the problem of trying to conduct experiemnts that realistic and at the same time avoid causing participants unnecessary stress and discomfort

• This dilemma is less problematic when researchers can obtain informed consent specifying the nature of the experiment from the participants before the experiment is conducted

• However, in some cases, full disclosure of the procedures would influence the nature of the results, and in this case, deception in used in which only partial or incomplete information about the procedure is given to the participants in advance

• Ethical principles of psycologists in the conduct of research are summarized in Figure 2.5 (page 60) of your text.

• Among other things, these ethical principles specify that whenever deception is used in experiments, the experimenters conduct a debriefing - an explicit statement to the participants about what deception was used and why it was necessary

• During the debriefing, researchers also try to alleviate any discomfort that might have occurred during the experimental session, and discuss the research with them, which is educational to both parties

• Studies examining the impact of deception in experiments on participants have typically found that participants do not object to the mild discomfort that is typically produced, and that often, they find such experiemnts more interesting than nondeception experiments

  ---

Click here to go back to chapter outlines/notes

Suggested Readings

Brannigan,C.G., & Merrens, M.R. (Eds.) (1995). The social psycholgists: Research adventures. New York: McGraw Hill.

Christensen, L. (1988). Deception in psychological research: when is it justified? Personality and Social Psychology Bulletin, 14, 664-675.

McKenna, R.J. (1995). The undergraduate researcher's handbok: Creative experimentation in social psychology. Boston: Allyn & Bacon.

Pines, A.M., & Malach, C. (1993). Experiencing social psychology: Readings and projects. New York: McGraw Hill.

Back to Homepage