AN INTRODUCTION TO RESEARCH METHODS
Goal MetResearch MethodsAdvantages/DisadvantagesDescriptionObservational methodDescriptive methods allow description of behavior(s)Case study methodDescriptive methods do not support reliable predictionsSurvey methodDescriptive methods do not support cause-and-effect explanationsPredictionCorrelational methodPredictive methods allow description of behavior(s)Quasi-experimental methodPredictive methods support reliable predictions from one variable to another
Predictive methods do not support cause-and-effect explanationsExplanationExperimental methodAllows description of behavior(s)
Supports reliable predictions from one variable to another
Supports cause-and-effect explanations
1. In a recent study, researchers found a negative correlation between income level and incidence of psychological disorders. Jim thinks this means that being poor leads to psychological disorders. Is he correct in his conclusion? Why or why not?
2. In a study designed to assess the effects of exercise on life satisfaction, participants were assigned to groups based on whether they reported exercising or not. All participants then completed a life satisfaction inventory.
a. What is the independent variable?
b. What is the dependent variable?
c. Is the independent variable a participant variable or a true manipulated variable?
3. What type of method would you recommend researchers use to answer the following questions?
a. What percentage of cars run red lights?
b. Do student athletes spend as much time studying as student nonathletes?
c. Is there a relationship between type of punishment used by parents and aggressiveness in children?
d. Do athletes who are randomly assigned to a group using imagery techniques perform better than those who are randomly assigned to a group not using such techniques?
Doing Science
Although the experimental method can establish a cause-and-effect relationship, most researchers would not wholeheartedly accept a conclusion from only one study. Why is that? Any one of a number of problems can occur in a study. For example, there may be control problems. Researchers may believe they have controlled for everything but miss something, and the uncontrolled factor may affect the results. In other words, a researcher may believe that the manipulated independent variable caused the results when, in reality, it was something else.
Another reason for caution in interpreting experimental results is that a study may be limited by the technical equipment available at the time. For example, in the early part of the 19th century, many scientists believed that studying the bumps on a person’s head allowed them to know something about the internal mind of the individual being studied. This movement, known as phrenology, was popularized through the writings of physician Joseph Gall (1758–1828). At the time that it was popular, phrenology appeared very “scientific” and “technical.” With hindsight and with the technological advances that we have today, the idea of phrenology seems laughable to us now.
Finally, we cannot completely rely on the findings of one study because a single study cannot tell us everything about a theory. The idea of science is that it is not static; the theories generated through science change. For example, we often hear about new findings in the medical field, such as “Eggs are so high in cholesterol that you should eat no more than two a week.” Then, a couple of years later, we might read, “Eggs are not as bad for you as originally thought. New research shows that it is acceptable to eat them every day,” followed a few years later by even more recent research indicating that “two eggs a day are as bad for you as smoking cigarettes every day” (Spence, Jenkins, & Davignon, 2012). You may have heard people confronted with such contradictory findings complain, “Those doctors, they don’t know what they’re talking about. You can’t believe any of them. First they say one thing, and then they say completely the opposite. It’s best to just ignore all of them.” The point is that when testing a theory scientifically, we may obtain contradictory results. These contradictions may lead to new, very valuable information that subsequently leads to a theoretical change. Theories evolve and change over time based on the consensus of the research. Just because a particular idea or theory is supported by data from one study does not mean that the research on that topic ends and that we just accept the theory as it currently stands and never do any more research on that topic.
Proof and Disproof
When scientists test theories, they do not try to prove them true. Theories can be supported based on the data collected, but obtaining support for something does not mean it is true in all instances. Proof of a theory is logically impossible. As an example, consider the following problem, adapted from Griggs and Cox (1982). This is known as the Drinking Age Problem (the reason for the name will become readily apparent).
On this task imagine that you are a police officer responsible for making sure the drinking-age rule is being followed. The four cards below represent information about four people sitting at a table. One side of a card indicates what the person is drinking and the other side of the card indicates the person’s age. The rule is: “If a person is drinking alcohol, then the person is 21 or over.” In order to check that the rule is true or false, which card or cards below would you turn over? Turn over only the card or cards that you need to check to be sure.
Does turning over the beer card and finding that the person is 21 years of age or older prove that the rule is always true? No—the fact that one person is following the rule does not mean that it is always true. How, then, do we test a hypothesis? We test a hypothesis by attempting to falsify or disconfirm it. If it cannot be falsified, then we say we have support for it. Which cards would you choose in an attempt to falsify the rule in the drinking age problem? If you identified the beer card as being able to falsify the rule, then you were correct. If we turn over the beer card and find that the individual is under 21 years of age, then the rule is false. Is there another card that could also falsify the rule? Yes, the 16 years of age card can. How? If we turn that card over and find that the individual is drinking alcohol, then the rule is false. These are the only two cards that can potentially falsify the rule. Thus, they are the only two cards that need to be turned over.
Even though disproof or disconfirmation is logically sound in terms of testing hypotheses, falsifying a hypothesis does not always mean that the hypothesis is false. Why? There may be design problems in the study, as described earlier. Thus, even when a theory is falsified, we need to be cautious in our interpretation. We do not want to completely discount a theory based on a single study.
