An Introduction to Research Methods in Science
The goals of science map very closely onto the research methods that scientists use. In other words, there are methods that are descriptive in nature, predictive in nature, and explanatory in nature. I will briefly introduce these methods here.
Descriptive Methods
Behavioral scientists use three types of descriptive methods. First is the observational method—simply making observations of human or other animal behavior. Scientists approach observation in two ways. Naturalistic observation involves observing humans or other animals behave in their natural habitat. Observing the mating behavior of chimpanzees in their natural setting would be an example of this approach. Laboratory observation involves observing behavior in a more contrived and controlled situation, usually the laboratory. Bringing children to a laboratory playroom to observe play behavior would be an example of this approach. Observation involves description at its most basic level. One advantage of the observational method, as well as other descriptive methods, is the flexibility to change what one is studying. A disadvantage of descriptive methods is that the researcher has little control. As we use more powerful methods, we gain control but lose flexibility.
observational method Making observations of human or other animal behavior.
A second descriptive method is the case study method. A case study is an in-depth study of one or more individuals. Freud used case studies to develop his theory of personality development. Similarly, Jean Piaget used case studies to develop his theory of cognitive development in children. This method is descriptive in nature, as it involves simply describing the individual(s) being studied.
case study method An in-depth study of one or more individuals.
The third method that relies on description is the survey method—questioning individuals on a topic or topics and describing their responses. Surveys can be administered by mail, over the phone, on the Internet, or as a personal interview. One advantage of the survey method over the other descriptive methods is that it allows researchers to study larger groups of individuals more easily. This method has disadvantages, however. One concern has to do with the wording of questions. Are they easy to understand? Are they written in such a manner that they bias the respondents’ answers? Such concerns relate to the validity of the data collected. Another concern relevant to the survey method (and most other research methods) is whether the group of people who participate in the study (the sample) is representative of all the people about whom the study is meant to generalize (the population). This concern can usually be overcome through random sampling. A random sample is achieved when, through random selection, each member of the population is equally likely to be chosen as part of the sample.
survey method Questioning individuals on a topic or topics and then describing their responses.
sample The group of people who participate in a study.
population All of the people about whom a study is meant to generalize.
Predictive (Relational) Methods
Two methods allow researchers to not only describe behaviors but also predict from one variable to another. The first, the correlational method, assesses the degree of relationship between two measured variables. If two variables are correlated with each other, we can predict from one variable to the other with a certain degree of accuracy. For example, height and weight are correlated. The relationship is such that an increase in one variable (height) is generally accompanied by an increase in the other variable (weight). Knowing this, we can predict an individual’s approximate weight, with a certain degree of accuracy, given the person’s height.
correlational method A method that assesses the degree of relationship between two variables.
One problem with correlational research is that it is often misinterpreted. Frequently, people assume that because two variables are correlated, there must be some sort of causal relationship between the variables. This is not so. Correlation does not imply causation. Remember that a correlation simply means that the two variables are related in some way. For example, being a certain height does not cause you to also be a certain weight. It would be nice if it did, because then we would not have to worry about being either under- or overweight. What if I told you that watching violent TV and displaying aggressive behavior were correlated? What could you conclude based on this correlation? Many people might conclude that watching violent TV causes one to act more aggressively. Based on the evidence given (a correlational study), however, we cannot draw this conclusion. All we can conclude is that those who watch more violent television programs also tend to act more aggressively. It is possible that the violent TV causes aggression, but we cannot draw this conclusion based only on correlational data. It is also possible that those who are aggressive by nature are attracted to more violent television programs, or that some other variable is causing both aggressive behavior and violent TV watching. The point is that observing a correlation between two variables simply means that they are related to each other.
The correlation between height and weight, or violent TV and aggressive behavior, is a positive relationship: As one variable (height) increases, we observe an increase in the second variable (weight). Some correlations indicate a negative relationship: As one variable increases, the other variable systematically decreases. Can you think of an example of a negative relationship between two variables? Consider this: As mountain elevation increases, temperature decreases. Negative correlations also allow us to predict from one variable to another. If I know the mountain elevation, it will help me predict the approximate temperature.
positive relationship A relationship between two variables in which an increase in one variable is accompanied by an increase in the other variable.
negative relationship A relationship between two variables in which an increase in one variable is accompanied by a decrease in the other variable.
Besides the correlational method, a second method that allows us to describe and predict is the quasi-experimental method. Quasi-experimental research allows us to compare naturally occurring groups of individuals. For example, we could examine whether alcohol consumption by students in a fraternity or sorority differs from that of students not in such organizations. You will see in a moment that this method differs from the experimental method, described below, in that the groups studied occur naturally. In other words, we do not assign people to join a Greek organization or not. They have chosen their groups on their own, and we are simply looking for differences (in this case, in the amount of alcohol typically consumed) between these naturally occurring groups. This is often referred to as a subject or participant variable—a characteristic inherent in the participants that cannot be changed. Because we are using groups that occur naturally, any differences that we find may be due to the variable of being a Greek member or not, or the differences may be due to other factors that we were unable to control in this study. For example, maybe those who like to drink more are also more likely to join a Greek organization. Once again, if we find a difference between these groups in amount of alcohol consumed, we can use this finding to predict what type of student (Greek or non-Greek) is likely to drink more. However, we cannot conclude that belonging to a Greek organization causes one to drink more because the participants came to us after choosing to belong to these organizations. In other words, what is missing when we use predictive methods such as the correlational and quasi-experimental methods is control.
quasi-experimental method Research that compares naturally occurring groups of individuals; the variable of interest cannot be manipulated.
When using predictive methods, we do not systematically manipulate the variables of interest; we only measure them. This means that, although we may observe a relationship between variables (such as that described between drinking and Greek membership), we cannot conclude that it is a causal relationship. Why? Because there could be other, alternative explanations for this relationship. An alternative explanation is the idea that it is possible that some other, uncontrolled, extraneous variable may be responsible for the observed relationship. For example, maybe those who choose to join Greek organizations come from higher-income families and have more money to spend on such things as alcohol. Or maybe those who choose to join Greek organizations are more interested in socialization and drinking alcohol before they even join the organization. Thus, because these methods leave the possibility for alternative explanations, we cannot use them to establish cause-and-effect relationships.
alternative explanation The idea that it is possible that some other, uncontrolled, extraneous variable may be responsible for the observed relationship.
Explanatory Method
When using the experimental method, researchers pay a great deal of attention to eliminating alternative explanations by using the proper controls. Because of this, the experimental method allows researchers not only to describe and predict but also to determine whether there is a cause-and-effect relationship between the variables of interest. In other words, this method enables researchers to know when and why a behavior occurs. Many preconditions must be met in order for a study to be experimental in nature. Here, we will simply consider the basics—the minimum requirements needed for an experiment.
experimental method A research method that allows a researcher to establish a cause-and-effect relationship through manipulation of a variable and control of the situation.
The basic premise of experimentation is that the researcher controls as much as possible in order to determine whether there is a cause-and-effect relationship between the variables being studied. Let’s say, for example, that a researcher is interested in whether cell phone use while driving affects driving performance. The idea behind experimentation is that the researcher manipulates at least one variable (known as the independent variable) and measures at least one variable (known as the dependent variable). In our study, what should the researcher manipulate? If you identified the use of cell phones while driving, then you are correct. If cell phone use while driving is the independent variable, then driving performance is the dependent variable. For comparative purposes, the independent variable has to have at least two groups or conditions. We typically refer to these two groups or conditions as the control group and the experimental group. The control group is the group that serves as the baseline or “standard” condition. In our study of cell phone use while driving, the control group is the group that does not use a cell phone use while driving. The experimental group is the group that receives the treatment—in this case, those who use cell phones while driving. Thus, in an experiment, one thing that we control is the level of the independent variable that participants receive.
independent variable The variable in a study that is manipulated by the researcher.
dependent variable The variable in a study that is measured by the researcher.
control group The group of participants that does not receive any level of the independent variable and serves as the baseline in a study.
experimental group The group of participants that receives some level of the independent variable.
What else should we control to help eliminate alternative explanations? Well, we need to control the type of subjects in each of the treatment conditions. We should begin by drawing a random sample of subjects from the population. Once we have our sample of subjects, we have to decide who will serve in the control group versus the experimental group. In order to gain as much control as possible, and eliminate as many alternative explanations as possible, we should use random assignment—assigning participants to conditions in such a way that every subject has an equal probability of being placed in any condition. How does random assignment help us to gain control and eliminate alternative explanations? By using random assignment we should minimize or eliminate differences between the groups. In other words, we want the two groups of participants to be as alike as possible. The only difference we want between the groups is that of the independent variable we are manipulating—either using or not using cell phones while driving. Once participants are assigned to conditions, we measure driving performance for subjects in each condition using a driving simulator (the dependent variable). Studies such as this one have already been completed by researchers. What researchers have found is that cell phone use while driving has a negative effect on driving performance (Beede & Kass, 2006; Dula, Martin, Fox, & Leonard, 2011).
random assignment Assigning participants to conditions in such a way that every participant has an equal probability of being placed in any condition.
Let’s review some of the controls we have used in the present study. We have controlled who is in the study (we want a sample representative of the population about whom we are trying to generalize), who participates in each group (we should randomly assign participants to the two conditions), and the treatment each group receives as part of the study (some drive while using a cell phone and some do not). Can you identify other variables that we might need to consider controlling in the present study? How about past driving record, how long subjects have driven, age, and their proficiency with cell phones? There are undoubtedly other variables we would need to control if we were to complete this study. The basic idea is that when using the experimental method, we try to control as much as possible by manipulating the independent variable and controlling any other extraneous variables that could affect the results of the study. Randomly assigning participants also helps to control for subject differences between the groups. What does all of this control gain us? If, after completing this study with the proper controls, we find that those in the experimental group (those who drove while using a cell phone) did in fact have lower driving performance scores than those in the control group, we would have evidence supporting a cause-and-effect relationship between these variables. In other words, we could conclude that driving while using a cell phone negatively impacts driving performance.
control Manipulating the independent variable in an experiment or any other extraneous variables that could affect the results of a study.