Why Use Scientific Research Methods?
Psychologists use scientific research methods in order to understand people's feelings, actions and thoughts. We do this for several reasons. Scientific research, if done well, gives us a relatively objective perspective of the object we are studying. Although some people use science to buttress their own personal beliefs and biases, the process of doing scientific research includes several check-points built in to guard against a person's reaching conclusions that are not warranted. For example, when a scientist seeks to publish the results of a research project, the project is reviewed by the editor of a journal and several other experts, who check to see that the work meets the standards of the discipline. They also examine the project to see that people were treated fairly, and that any conclusions that are made are reasonable.
It also allows for what is called "replication." This is when one scientist conducts research, and others repeat it to see if they get the same results. In this way, scientists confirm each others' results. For example, if I conduct a study and find out that religion seems to help people cope with stress, other people can try to "replicate" my results. If they do, then we gain confidence in the idea that religion aids us in coping. If they do not, then we can consider the differences between the various studies--perhaps we measured religiousness in different ways, or maybe people experienced different kinds of stress, or (perish the thought!) perhaps my research was flawed. Replication of research is a very powerful way that scientists can refine their understanding of something.
Goals & Methods of Science
When I was a student at Southern Utah University, one of my favorite professors, Dr. Les Jones, had a knack for explaining the way that science works. Much of what you'll read on this page is adapted from my notes in his classes.
There are different levels of questions that scientists want to answer. One way of thinking of them is by asking ourselves how well we understand the thing that we are studying. The most superficial level of understanding begins with a basic question, such as "What is a religious person like?" or "What are the similarities and differences between religious and non-religious people?" This is basic in the sense that we need to be able to answer "what" before we go on to answer a "when" or "where" question. Likewise, we need to be able to say when and where people are religious before we have a reasonable chance of understanding why people are religious. So, if we list these questions in order of the depth of understanding that they convey, we would have something like this:
Associated with each of these questions are goals that we wish to accomplish. When we ask "What," we are trying to describe the object of our study. "When" and "Where" are the questions we ask when we seek to predict something. These goals all prepare us for the most interesting goal of all, understanding the phenomenon we are studying. Again, we begin with simpler goals, such as description, so that we can later predict and eventually understand the object we are studying. Adding these to our chart, we get a picture something like this:
What are the methods that we use to achieve these goals? Scientists use measurements in order to describe something. If you wanted to describe me, you would measure my height (about 6 feet), my weight (175 lbs.), and maybe count the number of hairs on my head (fewer than most people, unless you count my beard). Similarly, if you we wanted to measure someone's religiousness, you might find out how often they attend church or what kinds of views they have about the Tanakh or Bible or Koran. Of course, these would be simple measurements, but they give you an idea of what the person is like.
Measurement is an important step in psychology because we need to do it well in order to achieve our goals of predicting and understanding the thing that we are studying. In order to predict something, we use a statistical technique called correlation. In correlation, we measure two variables and see how they relate to one another. (For example, tall people tend to weigh more than short people weigh. Height is correlated with weight.) In the psychology of religion one such relationship is that people who attend church frequently tend to have more strict views about scripture than do people who attend church less frequently.
Correlations such as this give us a better understanding of what we are studying, but they don't tell us everything that we want to know. For example, the correlation doesn't tell us whether one thing (church attendance) causes changes in another thing (views about scripture). On one hand, it might be that people attend church because they interpret scriptures strictly. On the other hand, people might interpret scriptures strictly because they attend church. It might even be the case that some third factor is causing both church attendance and strict interpretation of scriptures. Correlational research doesn't help us explain why two things are related--it only tells us that they are related.
In order to understand why things happen, we need to do an experiment. Simply put, an experiment is when you change one thing in order to see how it affects another thing. The key to a good experiment is that there is only one thing that is different between the various groups in your study. For example, if you want to find out how the parable of the good Samaritan might influence people to help someone in need, you might have one half of the people in your experiment think about the parable, and give the other half something else to think about. Then, you might give them a chance to do something helpful, such as donate some time to a homeless shelter. If people who thought about the parable donated more time than the people who did not think about the parable, and if those two groups of people are alike in every other way, then you may conclude that the parable affected the people's behavior. Of course, this is a very simple example, but it illustrates the basic process.
So, now we have completed the table illustrating the questions, goals and methods that scientists use:
QUESTION GOAL METHOD
Let's summarize this. Scientists ask several levels of questions. We do this because we have different goals in mind, and we use various methods in order to achieve these goals. We begin by asking the most simple questions, describing the object we are studying. Next, we measure several things so that we can see how they are associated with one another. This tells us if we can predict one thing by knowing another. After we are able to predict, we conduct experiments so that we can better understand why things are the way they are.
This process should be guided by a theory, a kind of map of the way things work. One theory might be that religion leads people to help others. With this as our theory, we make a prediction that people who read the parable of the good Samaritan will be more likely to donate their time to the local homeless shelter than do people who do not read the parable. We conduct the experiment, carefully recording the amount of time that each person donates to the homeless shelter. Next we compare the amount of time people in each group donated. We use statistics to help us determine whether one group, on average, donated more time than the other. And finally, we evaluate our theory in light of the experiment. If the results of the experiment are consistent with the theory, we gain confidence in the theory. If they are inconsistent, we look carefully for flaws in the experiment, and for flaws in the theory. This might lead us to modify the theory, and repeat the experiment to see if we get consistent results.
Experiments are a powerful way of understanding how one thing affects another. Unfortunately, they are not used very often in the psychology of religion. Here are several reasons why I think that experiments are not as common as they ought to be.
1. Too much research focuses on the measurement of religion. There are several good instruments for measuring religious attitudes and thoughts, yet much research continues to focus on the most superficial level of understanding: description. I believe that the core reason for this is that many people have their own view of what religion is and what distinguishes "good" (or desirable) religiousness from "bad" (undesirable) religiousness. Consequently, a researcher often devotes time to develop yet another scale to measure religion in a way that he or she finds acceptable. (This relates to the issue of coordination of efforts, addressed in my page on the future of the psychology of religion.) The end result is that people spend too much time debating the relative merits of different measurements. Yes, we need to have good measurements in order to do good correlational and experimental studies. But there already are many excellent measurement tools available, and we can use these to answer theoretically meaningful questions. Until we do this, progress will be slower than necessary.
2. It is easier to do correlational research than it is to conduct a good experiment. In order to design a good experiment, we must have a theory suggesting that one thing will cause a change in another. In some cases, our theories have not yet developed to that point. But in other cases, researchers find it too difficult to meet the technical demands of experimental design. Although the basic logic of an experiment is straightforward, it can be challenging to conduct a true experiment. Let's take an extreme case as an example. Suppose we conceptualize religiousness as an unchanging characteristic residing within the individual, like the person's sex. If religion is viewed this way, you might wonder how we could do an experiment involving religion, manipulating a person's religiousness in some way. After all, isn't this almost like assigning people to be male or female? But a little ingenuity can pay big dividends. For example, we can compare religious and nonreligious people on other variables that we are able to manipulate. How do they react to different kinds of information? How do they interact with people in different contexts? Answers to these kinds of questions will improve our understanding of how religion affects people much more than will answers to simply correlational questions. Experiments remain feasible if we consider them thoughtfully, and they tell us much about the merit of our theories.
3. Finally, there sometimes are ethical objections to conducting experiments in the psychology of religion. These are dilemmas that face every researcher in psychology, but sometimes seem more pronounced in studies of religiousness. The matter comes down to this: what gives me (or you or anyone else) the right to do something that might affect someone else's religious beliefs?
There are different ways of answering this question; here's mine. Suppose that I seek to find the relationship between religiousness and prejudice. A common way to do research is to, say, measure the person's belief about scripture, and then measure the amount of prejudice they show. Next, I would calculate the correlation between religiousness and prejudice. After doing this, I might think that I have not affected the participants' beliefs. But I would be wrong! I am misleading myself if I believe that by simply measuring the person's attitudes toward scripture and their prejudice toward their fellow beings that I am not affecting that person. Measuring something can affect it. For example, when I measure the air pressure in my bicycle tire, I also change the amount of pressure in that tire because a little bit of air escapes. Likewise, social psychologists have found that when researchers ask someone about their beliefs, they also impact those very beliefs. The person may not have thought about religion quite that way, or may not have thought about religion for years, until I ask them my questions. Descriptive and correlational research do indeed affect the people who participate in our studies.
The effect of a research project on the people who participate in the project is mainly a matter of degree. As I said before, it is an issue that each researcher must confront. But in my opinion, if you wish not to impact another person's religious beliefs or attitudes, you wish to do the impossible.
Support the Psychology of Religion Site by buying books at Amazon.com.