About The ARDA | Tools | FAQs | Contact Us

Very few laboratory experiments have been done on the topic of religion, because it is very difficult to manipulate the key variables. In an experiment, the research performs some action and examines its consequences. Typically, in experiments with human beings, a group of research subjects is divided at random into two or more groups. Perhaps one group is a control group, to which nothing unusual is done, but a second group is the treatment group. A hypothetical example is having people fill out a questionnaire about religion, but at random half of the respondents do so while religious music is playing in the background. Results can then be compared to see which religious attitudes become stronger when the music is played. Or, both groups could hear music, but the nature of the music would differ, in which case there is not a control group per se but two comparison treatment groups. There is a long-standing debate about whether experimentation with religion is ethical or even possible, but it has also been argued that without some use of experimentation a science cannot progress (Bateson 1977; Yeatts and Asher 1979). This argument remains unsettled, and few major experimental studies have been done, in part because harmless experimental manipulations may fail to produce interesting results, yet it is possible to include mild experiments in questionnaires.

There are several ways to include experiments in a questionnaire. Sometimes this is done merely as part of the process of pretesting and refining items. For example, one of the most commonly used General Social Survey items in research on religion is POLVIEWS, which asks respondents to say how liberal or conservative their political views are. In 1978, a subset of respondents got a slightly different version, POLVIEWY, which added a response: “Haven’t thought much about this.” Then, in 1983 some respondents got POLVIEWX, which asked political views in terms of a left-right scale in which the intermediate steps were not labeled. In both cases, the original version of POLVIEWS was retained for later surveys. POLVIEWY seemed merely to shift some respondents from the ordinary “don’t know” response to the new one, without affecting other responses. POLVIEWX may have been problematic because the GSS tries to use simple questions that can be understood by every respondent, and clearly labeled responses may do a better job of this than abstract numerical scales.

Experimental methods can be used substantively to accomplish certain research goals. A classical example is Richard Larson’s questionnaire study of how clergy deal with parishioners who may have psychiatric problems. Larson’s respondents included psychiatrists and clergy in a range of denominations, and much of the study was conventional attitudinal research. However one part used experimental methods successfully, to determine how the respondents would judge the severity of emotional problems across the two genders. A major challenge incorporating experiments into questionnaires, is to make the experimental manipulation strong enough to have an effect. Larson’s approach was to use the vignette method, including vivid little stories into the questionnaire that repeated the variable that was being manipulated. Respondents were asked to judge how severe the problem was faced by the person in the story. Here is one (Larson 1968: 252):

“Mr. Thompson has been widowed for five years. He is a respectable person whose wife died when he was just 20. They had no children, and he went home to his parents. He has not had a job in five years and does not seem to want to go out and look for one. Her is very quiet: He does not talk much to anyone — including his parents. He acts as if he is afraid of people, especially young women his own age. He won’t go out with anyone and whenever someone comes to visit his parents, he stays in his own room until the person leaves. He just stays by himself and daydreams about his wife.”

The experimental manipulation is that half of the respondents, selected at random, got a similar paragraph about Mrs. Thompson, with the pronouns changed to feminine, and “wife” changed to :husband.” At the time the study was done, society generally expected greater public activity from men, and thus might judge Mr. Thompson more severely disturbed than Mrs. Thompson, even though their behavior was identical. This was true for the psychiatrists in the same, 53.7 percent calling the male case “severely distributed,” compared with only 48.2 percent of the female case. But the male-female differences were greater for mainstream protestant clergy (45.9 percent versus 32.4), Fundamentalist clergy (46.3 percent versus 27.3 percent), and even more so for Roman Catholic clergy (50.8 percent versus 23.3 percent). If Larson had asked respondent directly whether they held different expectations for males versus females, they might not have admitted the extent to which they did. Here the experimental methods allowed Larson to study the question without revealing to the respondents what his scientific goals were.

Batson, C. Daniel. 1977. “Experimentation in Psychology of Religion: An Impossible Dream.” Journal for the Scientific Study of Religion 16:413-418.

Larson, Richard F. 1968. “The Clergyman’s Role in the Therapeutic Process: Disagreement Between Clergymen and Psychiatrists.” Psychiatry 31:250-263.

Yeatts, John R., and William Asher. 1979. “Can We Afford Not to Do True Experiments in Psychology of Religions?” Journal for the Scientific Study of Religion 18:86-89.

QuickSearch The Knowledge-Base

To search the knowledge-base, enter a term below:

Select a Theory below to learn more:

Select a Concept below to learn more:

TCM Contributors

Would you like to be considered for a position on the Theories, Concepts, and Measures contribution team? If so, click on the link below and complete the TCM Online Request Form.

TCM Online Request Form

If you are already a contributor for Theories, Concepts, and Measures and need assistance in managing the content, please click on the link below for instructions.

Site Administration Instructions