Читать книгу Social Work Research Methods - Reginald O. York - Страница 62

Let’s employ our learning about these principles with regard to a research question. This is an example that was presented in another text (York, 1997). We will use one that is simple and may even be fun. Have you ever heard someone say “It must be the full moon” when they witness strange behavior? When this author asked whether the full moon affects the behavior of mental health clients, he often received an affirmative response from social workers and others. Many people are convinced that the full moon has such an effect.

Let’s suppose that we have decided to conduct a study to see if this proposition is true. We will go through the research process to examine it. We will examine the knowledge base currently available to decide where to go from there. If the question has already been well answered by existing research, we will conclude that another study is not necessary; so we will stop our inquiry there.

What is the purpose of our study? Which is the better way to state our purpose given the spirit of scientific inquiry?

1 To prove that the full moon causes unusual behavior

2 To demonstrate that there is more strange behavior during the full moon than when the moon is not full

3 To determine if there is more strange behavior during the full moon than when the moon is not full

4 To prove that the full moon is not related to unusual behavior

What is your choice? If you choose the first one or the last one, you will clearly find yourself outside the bounds of the spirit of scientific inquiry. Remember that science is a method of finding out rather than a method of proving a point. If you choose the second one, you will make the same mistake. If your purpose is to demonstrate that something is true, you are not speaking according to the spirit of scientific inquiry. So this leaves the third option, which is consistent with the spirit of scientific inquiry.

If we seek to prove a point, we will naturally fall into various traps that will hamper our pursuit of knowledge about the subject. Remember that research is a process of discovery, not justification. We should engage in a process of inquiry that is designed to provide an objective appraisal of our research subject. The reduction of the potential of human bias is key to accomplishing this. So our first principle of science is as follows: Scientific research is a process of finding out, not a process of justification.

Another principle that undergirds the scientific method is that you should not reinvent the wheel. This means that you should start your inquiry with a review of what is already known about your research question. In our review of the question about the full moon and strange behavior, we found that there have been a great number of studies undertaken.

So do we need to engage in another one? Maybe but only if the existing research has left some point of the debate uncovered. If you search enough, you will usually find a special aspect of your inquiry that has not yet been fully investigated. The main point here is not to discourage the continued pursuit of a theme but to find what is already known, so that you can couch your inquiry in a manner that is more likely to add to our existing knowledge base.

There have been numerous studies on this question. One such study was conducted on attempted suicides (Mathew, Lindsay, Shanmjganatan, & Eapen, 1991). The records of the Accident and Emergency Department of a large urban hospital were examined to determine if the rate of suicide attempts that came to the attention of this hospital was different during the full moon and at other times. The number of suicides attempted for each day of 1 month were recorded. The full moon fell on Day 15. On that day, a total of 19 suicide attempts were recorded. The largest number of suicide attempts (23) were recorded on Days 3 and 17, one of which was close to the full moon, but the other was at a great distance from it. The number of attempts on the day of the full moon was slightly higher than the average for the entire month, but the difference was determined to be nonsignificant (statistically). When we examine this question, we might also want to compare the 3 days when the moon was at its fullest with the 3 days when it was the least full. If the full moon causes suicide, we would expect the 3 days of the full moon to have a significantly higher suicide rate. For the data from this study, the average number of suicide attempts during the full moon period was 18 as compared with 17 for the new moon period (the 3 days during which the moon is the least full). You would probably conclude that this difference was not significant; therefore, the rate of suicides close to the full moon was not different from the rate when the moon was the least full.

So what do you think? Are you yet convinced that the full moon does not affect behavior, or do you believe that there is a need for further review of the literature?

I hope you said that there is a need for further review of the literature. Why? This is only one study that addressed only one type of behavior (suicide). Maybe the full moon affects mental health in some way other than suicide.

So let’s assume that you said there is a need for further review of the research on this theme. In one such study, a set of researchers examined the records of a psychiatric hospital to determine if dangerous behavior of patients was more prevalent during the full moon than at other times. Dangerous behavior was defined as “erratic behavior which was assessed by qualified mental health professionals as dangerous to self or others to the extent that isolation (seclusion) or restraints were necessary to prevent harm to self or others” (Durm, Terry, & Hammonds, 1986, p. 988). Data for 3 years (1982, 1983, and 1984) were collected. The average number of such incidents of dangerous behavior was actually higher (13.17 per day) during the period when the moon was not full than it was during the period of the full moon (11.61 per day). So these data clearly did not support the idea that the full moon makes people different.

Well, where are you now? Are you ready to conclude that the full moon does not affect behavior? Or are you prepared to review additional studies? I would be inclined to review additional studies because there are so many other aspects of behavior that might be influenced by the full moon.

If you wish for further review, there is a special type of research that could be quite useful. It is a type of research that combines the data from many studies to see the answer to the research question. Such was the case in an article by Byrnes and Kelly (1992). They reviewed 12 studies that examined the relationship between the full moon and things such as crisis calls to police stations, poison centers, and crisis intervention centers. Their conclusion from this review was that “there is no evidence whatsoever for the contention that calls of a more emotional or ‘out-of-control’ nature occur more often at the full moon” (p. 779).

Now maybe you are more convinced that there is little evidence of a relationship between human behavior and the full moon. But if you are a tenacious reviewer of the literature, you might go another step and see if there are other reviews of the literature that combined various studies about the full moon. Another review of many studies was undertaken by Rotten and Kelly (1985). A total of 37 published studies were included in their review. They also concluded that there was little evidence to support the theory that the full moon affects people’s behavior.

This pattern of findings of nonsignificant differences between behaviors during the full moon and other periods is found in a review of other sources. But one study was found that claimed to have found a relationship between the full moon and behavior. That study was conducted by Hicks-Casey and Potter (1991). They found that there was more aggressive acting-out misbehavior in a sample of 20 developmentally delayed women during the full moon than at other times. However, the analysis of their data was challenged by Flynn (1991) as having major flaws. And even if you believed that this was a sound study, you should recognize that it is only one study of only 20 people. Perhaps now the score is 1 for the full moon and about 50 against it. Where do you now make your bet about the next study to be undertaken on this theme?

Do we still want to conduct a study of the effect of the full moon on behavior? Our review of the existing literature would suggest that this question has been substantially answered already. Do we want to spend our time reinventing the wheel? We would not, however, be reinventing the wheel if we found a new angle on this topic, such as a type of behavior not yet examined. But, generally speaking, it seems that the evidence clearly fails to support the conclusion that the full moon affects behavior.

Are you convinced? Maybe you are of the opinion that you have seen strange behavior during the full moon yourself, so you don’t need help from research studies. Let’s examine that opinion by considering the nature of the data that would be included in a review of our question. Let’s suppose that we have conducted a study in which we measured people during both the full moon and the new moon (when the moon is least full) and have also measured whether the individuals in our study have exhibited unusual behavior. Exhibit 1.1 reflects hypothetical data that illustrate how this would be examined.

The hypothetical data in Exhibit 1.1 show that the proportion of people with unusual behavior during the full moon was 25%, and this is the same as the proportion of people who demonstrated unusual behavior during the new moon. This would suggest that unusual behavior occurs at the same rate during the full moon and when the moon is not full. Note that it was the proportions rather than the numbers that were the critical facts to review here.

If Jim says that he has seen unusual behavior during the full moon and does not review other data, he is restricting his inquiry to only one cell in this table—the one that reflects unusual behavior during the full moon. He has not examined the number (and proportion) of people who did not exhibit unusual behavior when the moon was full, and more important, he has not compared the behavior during the full moon with behavior when the moon was least full. So he has engaged in an inquiry that is incomplete, something that is not characteristic of the scientific method (or critical thinking).

A third principle that undergirds the scientific method of inquiry suggests that two heads are better than one. If we believe this statement, we will want to assess the dependability of our method of measuring strange behavior. If we believe that two heads are better than one, we would seek evidence of different kinds. If hospital records are not accurate, a review of them would not provide an accurate answer to our question. But if we examine both hospital records and police records, we are in a better position. If fact, the more “heads” we examine, the better. We could also review school records or something else. If we find a consistent pattern, we would feel more comfortable drawing a conclusion. Our review of the evidence regarding the full moon considered a good number of sources of data, not just one. So we adhered to the saying that two heads are better than one.

Exhibit 1.1 Hypothetical Data on the Full Moon and Unusual Behavior

The fourth principle is that things sometimes happen just by chance. What if we found that the proportion of the 217 patients in a mental hospital who were observed to act more strange than usual was 37% during the full moon and 31% when the moon was not full? Could this difference be something that we could write off to chance? If so, we would not conclude that there was a relationship between the full moon and this kind of behavior. This means that we would be unlikely to bet our money on the discovery of a relationship between the full moon and strange behavior if we had the opportunity to repeat this study.

What if we found that 68% of these 217 patients acted more strange than usual during the full moon while only 24% did so when the moon was not full? In this case, we would be less likely to write off these differences to chance. We would be more likely to conclude that we had found a legitimate relationship between the full moon and strange behavior. There are ways to estimate the likelihood that a given set of results would occur by chance. This is the task of statistics. Ruling out chance as a good explanation of our data is based on things like proportions that reveal the strength of the relationship, as noted above, but it is also determined by the number of people from whom you drew the data (i.e., the sample size). In our example, we might be more cautious in our conclusions if the number of people in our data from the mental hospital had been 17 rather than 217. We would normally be more cautious the less the number of people in our study.

The final principle that undergirds the scientific approach is that we should be cautious in drawing conclusions when we are wearing the research hat. For example, we should be very careful to ensure that our study conclusions are consistent with the data we have analyzed. If we had applied a statistical measure to our data and found that the likelihood that our results would occur by chance was high, we would be reluctant to conclude that we had found a legitimate relationship between the variables in our study (e.g., the full moon and strange behavior). In addition, we should be cautious in concluding that our one study has found the final proof of an assertion. We should consider our data to be incomplete to some extent, and we should invite others to continue the research.

When we examine the literature, we will find that different researchers have found different methods for studying our research question. The results are not always consistent. Thus, we should treat the results of any one study with caution. The more studies we find with consistent results, the more confident (and less cautious) we are entitled to be.

We have examined common sense and science though our examination of the literature on whether people act differently when the moon is full. Our literature review has lead us to the conclusion that we do not need to collect and analyze new data on this question because it has been fully reviewed already. To do another study would make us vulnerable to the criticism that we are trying to reinvent the wheel.