Foundations of  Psychology 104: Critical Thinking in Science Skill Exercise

Learning Objective: I) To understand the nature of science as a self-correcting activity by seeing how the double movement of reflective thought prevents incorrect hypotheses from becoming accepted as truth.

Definitions: Inductive reasoning is from specific cases to more general principles. An everyday type example might be when a person at a party says, "Your beautiful clothing, meticulous grooming, and birthstone ring show you must be a Libran." He has taken a number of features of this event (namely you) and reached a predicted generalization about it to show  in induction the person goes from some concrete, specific observations to a general categorical statement Astrology is not a science because their hypotheses have not been confirmed; there are no replicated studies that show astrological signs can predict personality traits.

Deductive reasoning is from a general principle to specific predictions so the person might say, "As a Libran, you must like to play chess." As you can see if the person knows your category and has a general principle he can make predictions about other specific behavior you might engage in from the theory.

This shows the double movement of reflective thought as we go from an inductive observation to a general principle which we use to generate a hypothesis from which we design a study (an IV or PlV) to give us data (the DV) to test a prediction from the theory's principle. The problem with astrology is that the predictions do not hold true, you cannot reliably predict a person's personality from their astrological sign. While the rationalistic reasoning aspect of astrology may be formally sound, the empirical confirmation of the predictions has not occurred. If you want more examples read on.

Social Psychology Example

Let's start with inductive reasoning since a common way science develops is when a person notices some regularities in behavior. Imagine Leon Festinger noticed that there seemed to be a relationship between the effort people put into a project or relationship and how positively they feel about it. Imagine he is talking to Jack who has been working a minimum wage job for 10 years and finally saved up enough to buy a Mercedes. Jack says how great it is, how much he likes his Benz. Then Festinger talks to Jim who is the rich president of a bank, he owns exactly the same Mercedes model as Jack but is not nearly as enthusiastic about how great it is. Then he runs into and talks to Tim about his wife who has been an alcoholic, compulsive gambler, and shopoholic and generally caused Tim a lot of trouble, Tim says, "The more trouble she has (causes me) the more I love her. I am taking her to Vincenzo's tonight!" Then he talks to Ernie about his wife who is a model mother and homemaker, Ernie says, "I like her, she's ok I'm going to get the McDonald's 2 for $2 special for a treat for her."

Festinger could inductively reason from these types of specific observations to a general principle, "The more a person "pays" the higher will be their evaluation of what they paid for." Stated more formally, this is one of the hypotheses of Festinger's cognitive dissonance theory; an important concept that appears in most introductory psychology texts . This phrasing of this general hypothesis ("The more a person "pays" the higher will be their evaluation of what they paid for") can then become a general premise which can be tested by designing experiments (see Aronson & Mills, 1959). If the hypothesis's prediction is confirmed in a specific study we have greater confidence in the truth of the hypothesis's premise; if it is disconfirmed we know there is some limitation or problem with the hypothesis or experiment.

This movement (using induction) from specific observations to general premises which are then used to predict further observations that either confirm or disconfirm the original hypothesis is called the double movement of reflective thought. To the extent a theory can generate a lot of predictions that are not contradicted (disconfirmed by the data) it is a strong theory.

Experiments vs. Correlational/Descriptive studies

Experiments are done when the experimenter decided how the participants are treated; for example what type of commercial they will be shown. Correlational/descriptive studies are when the participants "decide" the factor they will be tested on; for example if we group participants by their religion, race, gender, etc. the experimenter can't manipulate those factors.  Our major point here was that you need an experiment to confirm or disconfirm causal statements. Aronson and Mills (1959) did an experiment,  randomly assigning students to get a mild or severe initiation to join a group as a way to really test  Festinger's proposed cognitive dissonance theory. For contrast, consider the correlational research of James Colemen. In the mid 1960s he published research which showed that minority students did better in integrated compared to racially isolated schools. This research was used as a major rationale for aggressively integrating schools since it could be interpreted as: Equalizing educational opportunity by desegregation would tend to equalize achievement

Christopher Jencks and his colleagues (1972) re-analyzed the data and pointed out that Coleman studied already-integrated blacks with already segregated blacks. In other words being in an integrated or segregated school was a subject variable, a pseudo-independent variable, not a randomly assigned school assignment. Jencks argued that here was a PRAH in Coleman's study, instead of the effect of integrated schooling, the "...high test scores of blacks in naturally integrated schools may reflect the greater resources or motivation of black parents who put their children in integrated schools." Since Coleman's research was not an experiment, he did not randomly assign students to integrated or segregated schools, Jencks' PRAH could not be easily disconfirmed. This does not mean that it is not a good idea to integrate schools; it means that it is unsound to claim Coleman's research confirms such action will necessarily raise minority test scores.

As another example consider the Meddock & Osborn  (1968) mouse study. Recall we took Barnett's general premise from his rat studies that: "Domestication causes genetic changes in animals" and tested that in our study. But since it was not an experiment and we had no control over the mice's pre-study environments we could not disconfirm the PRAH, "It was the pre-study environment (the wild mice’s experience living in homes and a dairy barn and the lab mice living in a university mouse colony) rather than genetic differences that may have caused the results of the mouse study."

Aronson, E. & Mills, J. (1959). The effect of severity of initiation on liking for a group. Journal of Abnormal and Social Psychology, 59, 177-181.

Coleman, J. S. (1993). Equality and achievement in education. Boulder, CO: Westview Press Reprint.

Festinger, L. (1957).  A theory of cognitive dissonance. Stanford, CA. : Stanford University Press.

Jencks, C. (1972). Inequality: A reassessment of the effect. New York, NY: Harper Publishers.

Meddock, T. & Osborn. D. (1968) Neophobia in wild and laboratory mice. Psychonomic Science, 12, 223.

Return to Variety of Ways we Know        Return to Home Page

Hypertext tutorial to teach social science experimental design by Don R. Osborn is licensed under a Creative Commons Attribution 3.0 United States License.
Based on a work at cas.bellarmine.edu.
Permissions beyond the scope of this license may be available at drosborn@bellarmine.edu.