26

KEY THEME

The experimental method is used to demonstrate a cause-and-effect relationship between two variables.

KEY QUESTIONS

In this chapter, we’ve noted a number of factors that are associated with higher or lower college grades. But all of these factors—such as time spent sending texts and on Facebook—are correlational, meaning that while they are linked, they do not necessarily indicate that the two factors are causally related.

In contrast to descriptive research and correlational studies, experimental research is used to demonstrate a cause-and-effect relationship between changes in one variable and the effect that is produced on another variable. Conducting an experiment involves deliberately varying one factor, which is called the independent variable, sometimes called the treatment variable. The researcher then measures the changes, if any, that are produced in a second factor, called the dependent variable, also called the outcome variable. The dependent variable is so named because changes in it “depend on” variations in the independent variable.

To the greatest degree possible, all other conditions in the experiment are kept exactly the same for all participants. Thus, when the data are analyzed, any changes that are measured in the dependent variable can be attributed to the deliberate manipulation of the independent variable. In this way, an experiment can provide evidence of a cause-and-effect relationship between the independent and dependent variables.

In designing experiments, psychologists try to anticipate and control for confounding variables. Also called extraneous variables, these factors are not the focus of the experiment. However, confounding variables might produce inaccurate experimental results by influencing changes in the dependent variable. Confounding variables in a psychology experiment could include unwanted variability in such factors as age, gender, ethnic background, race, health, occupation, personal habits, education, and so on.

To illustrate how experimental research works, let’s look at a topic of interest to most college students: What types of study strategies are most effective?

In most educational settings, learning is thought to take place during study, instruction, and practice. Tests, in contrast, are neutral experiences and simply assess what has been learned. But some studies seemed to suggest that being tested on new information helped students learn and remember it better than simply studying it (see Roediger & Butler, 2011). Psychologists Henry Roediger and Jeffrey Karpicke (2006) set out to investigate the effects of testing on learning and memory.

STUDYING THE EFFECTS OF TESTING

How could you design an experiment to show a difference between learning due to studying and learning due to testing? Roediger and Karpicke (2006) designed an experiment that compared the effects of repeated testing with the effects of repeated study periods. They predicted that students who repeatedly took tests after studying would have better long-term memory of the new information than students who repeatedly studied, but were not tested on, the same material. The hypothesis, then, was that “repeated testing improves learning more than repeated studying.”

27

The participants were 60 college undergraduates, aged 18 to 24. The researchers used random assignment to assign participants to one of two groups: either the experimental group or the control group. Random assignment means that all the participants have an equal chance of being assigned to any of the experimental conditions. Random assignment helps ensure that any potential differences among the participants are spread out evenly across all experimental conditions. Random assignment also helps minimize the possibility of bias because the same criteria are used to assign all participants to the different experimental conditions.

In any well-designed experiment, there is at least one control group. The control group serves as a baseline against which changes in the experimental group can be compared. In a typical experiment, the participants assigned to the control group go through all the experimental phases but are not exposed to the independent variable. Only the participants in the experimental group are exposed to the independent variable. In this study, the independent variable was repeated testing. The dependent variable was the score earned on a final test.

Here’s how the experiment was conducted. All of the participants were given a short prose passage to study. Participants in the control group read the passage for five minutes, and took a two-minute break. They then studied the passage again for five minutes before taking another two-minute break. They repeated this process for a total of four consecutive study periods. Note that this is the standard test-preparation method: to repeatedly study the same material until you feel certain that you have mastered it.

Participants in the experimental group were given the same prose passage to learn. They were also allotted five minutes to study the passage, and then took a two-minute break. But rather than restudying the passage, they took a test on the material: They were given a blank sheet of paper and were allowed ten minutes to write down as much information from the prose passage as they could remember. After another two-minute break, without studying the material again, they were given the same test, followed by a two-minute break. This procedure was followed for a total of one study period and three test periods. Figure 1.5 shows the setup of the experiment.

At the end of the session, all of the participants filled out a short questionnaire asking them to predict how well they would remember the material in a week. A week later, all participants were tested on the material.

28

How do you think the two groups would compare on a test of their retention of the material a week later? Conventional wisdom would suggest that the control group members, who studied the material in four periods for a total of twenty minutes, would have learned the material much better than the participants in the experimental group, who, after all, had only studied the material for a total of five minutes.

Figure 1.6 shows the results on the test one week later. Despite having studied the material only one-fourth as long as the control group, the experimental group trounced the control group: They remembered 61% of the material in the passage while the control group remembered only 40%. Interestingly, the control group participants were much more confident of their ability to remember the information than the experimental group.

Were you surprised by the results? In fact, the basic results have been replicated by many different researchers (see Bjork & others, 2013; Rowland, 2014). Multiple studies have supported what has been dubbed the testing effect: the finding that practicing retrieval of information from memory produces better retention than restudying the same information for an equivalent amount of time. In other words, testing—rather than simply being a neutral assessment of what has been learned—is a powerful learning tool in its own right (Carpenter, 2012; Roediger & Nestojiko, 2015).

We’ll discuss some of the reasons for the testing effect in Chapter 6, Memory. In Psych for Your Life at the end of this chapter, we’ll describe some of the other ways in which the testing effect has been explored. And, we’ll discuss additional ways in which you can use psychological research to improve your own memory for new information.

Some experiments involve extra controls to increase the reliability of their findings. One important safeguard is the double-blind technique, which is often used when researchers are testing the effectiveness of a procedure or drug treatment. In a double-blind study, both the participants and the researchers interacting with them are blind, or unaware of the treatment or condition to which the participants have been assigned.

Using a double-blind technique helps guard against the possibility that the researcher inadvertently becomes an extraneous or confounding variable in the study. This can happen when a researcher, without realizing it, displays demand characteristics. These are subtle cues or signals that can bias the outcome of the study by communicating the behavior or response that is expected of the participants. A behavior as subtle as the researcher slightly smiling or frowning when dealing with some participants, but not others, could bias the outcome of a study.

29

Such studies also often involve the use of a placebo, which is a so-called “sugar pill” or other inactive substance or procedure. Although it is inactive, a placebo can produce real effects (see Wager & Atlas, 2015). A placebo effect is any change that can be attributed to beliefs and expectations rather than to an actual drug, treatment, or procedure.

For example, one student in our class asked us whether we believed that the herb ginkgo biloba could improve memory. To test that notion, psychologist Paul Solomon and his colleagues (2002) used a placebo in a double-blind study to test whether ginkgo biloba improved memory and concentration in older adults. Participants in the experimental group took the manufacturer’s recommended daily dosage of ginkgo biloba for six weeks, while those in the control group took an identical dosage of placebo capsules.

The researchers who interacted with the participants did not know which participants received the real and which received the fake ginkgo biloba. The researchers who did know the group assignments did not interact with or evaluate the participants. Memory and other cognitive abilities were assessed at the beginning and end of the six-week study.

Can you predict the results of the ginkgo biloba experiment? At the end of the six-week study, the test scores of both groups rose. However, there were no significant differences between the improvement in the ginkgo biloba and placebo groups. So why did both groups improve? The researchers concluded that it was probably due to the practice effect. The participants’ experience with the tests—the practice they got by simply taking the mental ability tests twice—was the most likely reason that test scores improved in both groups. This experiment illustrates the importance of the control group: Without a control group to compare, the improvement in the experimental group might have been attributed to the drug. But since the control group participants also improved, there must have been another explanation.

A well-designed and carefully executed experiment can provide convincing evidence of a cause-and-effect relationship between the independent and dependent variables. But experiments do have limitations. Because experiments are often conducted in highly controlled laboratory situations, they are sometimes criticized for having little to do with actual behavior. That is, the results may not generalize well, meaning that the results cannot be applied to real-world situations or to more general populations beyond the participants in the study. To minimize this, experiments are sometimes carried out in natural settings, rather than in a laboratory. A second potential limitation is that the phenomena the researchers want to study may be impossible or unethical to control experimentally.

But researchers are sometimes able to take advantage of naturally occurring events or conditions. In a natural experiment, researchers carefully observe and measure the impact of a naturally occurring event or condition on their study participants (Rutter, 2008). Although not true experiments, psychologists can use natural experiments to study the effects of disasters, epidemics, or other events.

30

Sometimes natural experiments involve everyday settings. Psychologists Kandice Kapinos and Olga Yakusheva (2011) were interested in better understanding the relationship between environmental factors and weight gain. Previous research had shown that environmental factors, such as high concentrations of fast-food restaurants, are correlated with the average weight of nearby residents (Inagami & others, 2009). However, since the research was correlational, it wasn’t possible to conclude that proximity of fast-food restaurants caused weight gain. It could be that people with unhealthy eating habits were more likely to choose to live in neighborhoods with easy access to fast food.

Of course, researchers can’t randomly assign large numbers of participants to long-term living situations. But Kapinos and Yakusheva identified a naturally occurring situation in which people are randomly assigned to housing—college dormitories.

College freshmen do tend to gain weight during their first year away from home—the so-called “freshman fifteen”—although the weight gain is typically closer to five pounds (Holm-Denoma & others, 2008). Kapinos and Yakusheva took advantage of the naturally occurring conditions on their own college campus by comparing weight changes in freshmen who lived in dorms with on-site cafeterias and snack bars with weight changes in freshman who lived in dorms that did not have on-site food services. Since freshmen were randomly assigned to the dormitories, the researchers could safely assume that there wasn’t some other factor that might cause differences between the two groups.

What did Kapinos and Yakusheva (2011) find? Female students who were assigned to dormitories with on-site dining facilities gained more weight and exercised less than students who were assigned to dormitories without food services. Male students assigned to dormitories with food services reported eating more meals and more snacks, but did not report gaining weight. Kapinos and Yakusheva (2011) concluded that campus design did play a significant role in influencing healthy—and unhealthy—behaviors in college students.

Before leaving the topic of research methods, one contemporary trend deserves special mention: the increasing use of brain-imaging techniques in virtually every area of psychology. To help highlight the importance of neuroscience, every chapter includes a special “Focus on Neuroscience” feature. This chapter’s Focus on Neuroscience explores brain-imaging techniques and discusses their increasing use in psychological research (see pp. 34–35).

This brief introduction to research methods will give you some idea of how psychologists conduct research. But we hope it also illustrates some of the ways in which scientists—and others—evaluate claims and evidence. In the Critical Thinking box “Think Like a Scientist,” we offer several suggestions to help you evaluate claims that you encounter both inside and outside the classroom.

31

32