Research Designs: Three Approaches to Studying Behavior
Behavioral scientists agree that their ideas and their theories about human behavior must be backed up by data to be taken seriously. However, although all scientists follow the basic underlying procedures of scientifi c investigation, the research of different scientists is designed with different goals in mind, and the different goals require different approaches to answering the researcher’s questions. These different approaches are known as research designs. A research design is the specifi c method a researcher uses to collect, analyze, and interpret data. Although there are many variants of each, there are only three basic research designs used in behavioral research. These are descriptive research designs, correlational research designs, and experimental research designs. Because these three research designs will form the basis of this entire book, we will consider them in some detail at this point. As we will see, each of the approaches has both strengths and limitations, and therefore all three can contribute to the accumulation of scientifi c knowledge. To fully understand how the research designs work, you need to be aware of the statistical tests that are used to analyze the data. If you are not familiar with statistical procedures (or if you feel that you need a bit of a brushup), you should read Appendix B and Appendix C before you continue.
Descriptive Research: Assessing the Current State of Affairs
The first goal of behavioral research is to describe the thoughts, feelings, and behavior of individuals. Research designed to answer questions about the current state of affairs is known as descriptive research. This type of research provides a “snapshot” of thoughts, feelings, or behaviors at a given place and a given time.
Surveys and Interviews. One type of descriptive research, which we will discuss in Chapter 6, is based on surveys. Millions of dollars are spent yearly by the U.S. Bureau of the Census to describe the characteristics of the U.S. population, including where people work, how much they earn, and with whom they live. Descriptive data in the form of surveys and interviews are regularly found in articles published in newspapers and magazines and are used by politicians to determine what policies are popular or unpopular with their constituents.
Sometimes the data from descriptive research projects are rather mundane, such as “Nine out of ten doctors prefer Tymenocin,” or “The average income in Montgomery County is $36,712.” Yet, other times (particularly in discussions of social behavior), descriptive statistics can be shocking: “Over 40,000 people are killed by gunfi re in the United States every year,” or “Over 45 percent of sixth graders at Madison High School report that they have used marijuana.”
One common type of descriptive research, frequently reported in newspaper and magazine articles, involves surveys of the “current concerns” of the people within a city, state, or nation. The results of such a survey are shown in Figure 1.1. These surveys allow us to get a picture of what people are thinking, feeling, or doing at a given point in time.
Naturalistic Observation. As we will discuss more fully in Chapter 7, another type of descriptive research—known as naturalistic observation—is based on the observation of everyday events. For instance, a developmental psychologist who watches children on a playground and describes what they say to each other while they play is conducting descriptive research, as is a biological psychologist who observes animals in their natural habitats or a sociologist who studies the way in which people use public transportation in a large urban city.
Qualitative Versus Quantitative Research. One distinction that is made in descriptive research concerns whether it is qualitative or quantitative in orientation. Qualitative research is descriptive research that is focused on observing and describing events as they occur, with the goal of capturing all of the richness of everyday behavior and with the hope of discovering and understanding phenomena that might have been missed if only more cursory examinations had been used (Denzin & Lincoln, 2003). The data that form the basis of qualitative research are in their original rich form—for instance, descriptive narratives such as fi eld notes and audio or video recordings. Quantitative research is descriptive research that uses more formal measures of behavior, including questionnaires and systematic observation of behavior, which are designed to be subjected to statistical analysis. The strength of qualitative research is that it vividly describes ongoing behavior in its original form. However, because it does not use statistical analysis, it is generally more subjective and may not fully separate the values of the researcher from the objectivity of the research process. In many cases, however, qualitative data are reported along with quantitative data to provide a fuller description of the observed behavior; this combination of approaches can be very informative.
Strengths and Limitations of Descriptive Research. One advantage of descriptive research is that it attempts to capture the complexity of everyday behavior. For instance, surveys capture the thoughts of a large population of people, and naturalistic observation is designed to study the behavior of people or animals as it occurs naturally. Thus, descriptive research is used to provide a relatively complete understanding of what is currently happening. Nevertheless, descriptive research has a distinct disadvantage in that although it allows us to get an idea of what is currently happening, it is limited to providing static pictures. A study of the current concerns of individuals, for instance, cannot tell us how those concerns developed or what impact they have on people’s voting behavior.
Correlational Research: Seeking Relationships Among Variables
In contrast to descriptive research, which is designed to provide static pictures, correlational research involves the measurement of two or more relevant variables and an assessment of the relationship between or among those variables. A variable is any attribute that can assume different values among different people or across different times or places. Sometimes variables are rather simple—for instance, measures of age, shoe size, or weight. In other cases (and as we will discuss fully in Chapters 4 and 5), variables represent more complex ideas, such as egomania, burnout, sexism, or cognitive development.
As we will see in Chapter 9, the goal of correlational research is to uncover variables that show systematic relationships with each other. For instance, the variables of height and weight are systematically related, because taller people generally weigh more than shorter people. In the same way, study time and memory errors are also related, because the more time a person is given to study a list of words, the fewer errors she or he will make. Of course, a person’s score on one variable is not usually perfectly related to his or her score on the other. Although tall people are likely to weigh more, we cannot perfectly predict how tall someone is merely by knowing that person’s weight.
The Pearson Product–Moment Correlation Coeffi cient. Because the size of the relationships of interest to behavioral scientists is usually very small, statistical procedures are used to detect them. The most common measure of relationships among variables is the Pearson product–moment correlation coeffi cient, which is symbolized by the letter r.
The correlation coeffi cient ranges from r 5 21.00 to r 5 11.00. Positive values indicate positive correlations, in which people who are farther above average on one variable (for instance, height) generally are also farther above average on the other variable (for instance, weight). Negative values of r indicate negative correlations, in which people who are farther above average on one variable (for instance, study time) generally are also farther below average on the other variable (memory errors). Values of the correlation coeffi cient that are farther from zero (either positive or negative) indicate stronger relationships, whereas values closer to zero indicate weaker relationships.
The Use of Correlations to Make Predictions. One type of correlational research involves predicting future events from currently available knowledge. In this case, one or more variables of interest are measured at one time, and other variables are measured at a later time. To the extent that there is a correlation between what we know now and what will occur later, we can use knowledge about the things that we already know to predict what will happen later. For instance, Nettles, Thoeny, and Gosman (1986) used a correlational research design to predict whether college students would stay in school or drop out. They measured characteristics of 4,094 college students at thirty different colleges and universities and assessed the ability of these characteristics to predict the students’ current college grade-point average (GPA). In addition to intellectual variables such as high school GPA and Scholastic Aptitude Test (SAT) scores, they also assessed social variables including socioeconomic status, the students’ reports of interfering social problems such as emotional stress and fi nancial diffi culties, and the students’ perceptions of the quality of faculty– student relations at their university. The last measure was based on responses to questions such as “It is easy to develop close relationships with faculty members,” and “I am satisfi ed with the student–faculty relations at this university.”
As shown in Table 1.2, the researchers found that students’ ratings of the social problems they experienced on campus were as highly predictive of their grade-point average as were the standardized test scores they had taken before entering college. This information allows educators to predict which students will be most likely to fi nish their college education and suggests that campus experiences are important in this regard.
Strengths and Limitations of Correlational Research. One particular advantage of correlational research is that it can be used to assess behavior as it occurs in people’s everyday lives. Imagine, for instance, a researcher who finds a negative correlation between the row in which his students normally sit in his class and their grade on the fi nal exam. This researcher’s data demonstrate a very interesting relationship that occurs naturally for students attending college—those who sit nearer the front of the class get better grades.
Despite the ability of correlational studies to investigate naturally occurring behavior, they also have some inherent limitations. Most important, correlational studies cannot be used to identify causal relationships among the variables. It is just as possible that getting good grades causes students to sit in the front of the class as it is that sitting in the front of the class causes good grades. Furthermore, because only some of all the possible relevant variables are measured in correlational research, it is always possible that neither of the
variables caused the other and that some other variable caused the observed variables to be correlated. For instance, students who are excited by the subject matter or who are highly motivated to succeed in school might both choose to sit in the front of the class and also end up getting good grades. In this case, seating row and grades will be correlated, even though neither one caused the other.
In short, correlational research is limited to demonstrating relationships between or among variables or to making predictions of future events, but it cannot tell us why those variables are related. For instance, we could use a correlational design to predict the success of a group of trainees on a job from their scores on a battery of tests that they take during a training session. But we cannot use such correlational information to determine whether the training caused better job performance. For that, researchers rely on experiments .
Experimental Research: Understanding the Causes of Behavior
Behavioral scientists are particularly interested in answering questions about the causal relationships among variables. They believe that it is possible, indeed necessary, to determine which variables cause other variables to occur. Consider these questions: “Does watching violent television cause aggressive behavior?”, “Does sleep deprivation cause an increase in memory errors?”, and “Does being in a stressful situation cause heart disease?” Because it is diffi cult to answer such questions about causality using correlational designs, scientists frequently use experimental research. As we will discuss more fully in Chapters 10 and 11, experimental research involves the active creation or manipulation of a given situation or experience for two or more groups of individuals, followed by a measurement of the effect of those experiences on thoughts, feelings, or behavior. Furthermore, experimental research is designed to create equivalence between the individuals in the different groups before the experiment begins, so that any differences found can confi dently
be attributed to the effects of the experimental manipulation.
Elements of Experiments. Let us look, for instance, at an experimental research design used by social psychologists Macrae, Bodenhausen, Milne, and Jetten (1994). The goal of this experiment was to test the hypothesis that suppressing the use of stereotypes may cause an unexpected “rebound” in which those stereotypes are actually used to a greater extent at a later time. In the experiment, college students were shown a picture of a “skinhead” and asked to write a short paragraph describing what they thought he was like. While doing so, half of the students were explicitly told not to let their stereotypes about skinheads infl uence them when writing their descriptions. The other half of the students were just asked to write a description. After the students had fi nished writing their descriptions, they were told that they were going to be meeting with the person they had written about and were taken into a separate room. In the room was a row of nine chairs, with a jean jacket and a book bag sitting on the center one. The experimenter explained that the partner (the skinhead) had evidently left to go to the bathroom but that he would be right back and the students should take a seat and wait. As soon as the students sat down, the experiment was over. The prediction that students who had previously suppressed their stereotypes would sit, on average, farther away from the skinhead’s chair than the students who had not suppressed their stereotypes was confi rmed.
Strengths and Limitations of Experimental Research. This clever experiment nicely demonstrates one advantage of experimental research. The experiment can be interpreted as demonstrating that suppressing stereotypes caused the students to sit farther away from the skinhead because there was only one difference between the two groups of students in this experiment, and that was whether they had suppressed their stereotypical thoughts when writing. It is this ability to draw conclusions about causal relationships that makes experiments so popular.
Although they have the distinct advantage of being able to provide information about causal relationships among variables, experiments, like descriptive and correlational research, also have limitations. In fact, experiments cannot be used to study the most important social questions facing today’s society, including violence, racism, poverty, and homelessness, because the conditions of interest cannot be manipulated by the experimenter. Because it is not possible (for both practical and ethical reasons) to manipulate whether a person is homeless, poor, or abused by her or his parents, these topics cannot be studied experimentally. Thus, descriptive and correlational designs must be used to study these issues. Because experiments have their own limitations, they are no more “scientifi c” than are other approaches to research.
The Selection of an Appropriate Method
The previous sections have described the characteristics of descriptive, correlational, and experimental research designs. Because these three approaches represent fundamentally different ways of studying behavior, they each provide different types of information. As summarized in Table 1.3, each research design has a unique set of advantages and disadvantages. In short, each of the three research designs contributes to the accumulation of scientific knowledge, and thus, each is necessary for a complete study of behavior.
0 comments:
Post a Comment