On the Limits of Scientific Investigation: Miracles and Intercessory Prayer

Richard L Gorsuch. Miracles: God, Science, and Psychology in the Paranormal. Editor: J Harold Ellens. Volume 2: Medical and Therapeutic Events. Westport, CT: Praeger, 2008.

A common definition of miracles would be an event or condition produced by direct intervention of God. An example of a miracle in the New Testament is in the feeding of the 5,000. It occurred when Jesus was teaching a multitude of people. The disciples noted that it was late and suggested the people be dismissed so that they could find food. Jesus countered by asking them to share the food they had with the people. The amount they had to share was small. Yet after Jesus prayed over the food, not only was everyone fed but there were more baskets of food left over than they had when they started. This was impressive to the New Testament writers as it is reported multiple times in the New Testament (Mt 14:15-21, Mk 6:34-44, Lk 9:10-17, Jn 6:1-14).

The feeding of the 5,000 can be seen as a miracle of the heart. People saw Jesus and the disciples sharing, and so also shared their food. What food they did not absolutely need, they placed in the baskets for others to have. But how do we evaluate this event if we assume that (1) people did not have any significant amount of food to share, and (2) in his giving thanks for the food Jesus prayed that the food would multiply to feed all of the people? Could it have happened as a true miracle, that is, caused directly by God? To answer this question, we need to define further what is meant by natural laws that govern a causal chain of events and the limitations of that paradigm.

Supernatural intervention seems at variance with science, a prime source of truth in the contemporary world. When even the Templetons, who financially support research on religion, see the classical scientific experiment as the best method to establish truth, it would seem that the truth of miracles must be based in scientific experiments. Some experiments have been done on the efficacy of intercessory prayer. Do such studies prove or disprove miracles? Is there evidence that a supernatural force suspends natural laws to cause a miraculous outcome in answer to prayer? Is this the best way to phrase the question?

The purpose of this chapter is to evaluate whether science, and the natural laws that result, and miracles, to which history and experience seem to testify readily, are compatible paradigms. Are science and miracles both possible? If miracles are possible, then evaluating whether or when miracles have happened or are occurring is the task of scholars in the appropriate disciplines within the humanities.

The Science of Science

To understand how science may or may not be able to establish miracles, the nature of science needs to be understood. Is the nature of science such that it can examine miracles, or does its examining miracles remove the miraculous?

The nature of truth, even scientific truth, has been questioned in the current philosophical position of postmodernism (Murphy 1997). Postmodernism arises from the failure of modern philosophy to have found a basis on which truth can be logically built, that is, a secure base that no one can doubt. This skeptical analysis was found already in ancient Greek philosophy. It was dormant throughout the Middle Ages, which ignored the issue or assumed the foundation of all truth to be in divine revelation. Descartes (1701/1990) is the prototype for premodern philosophy in that he assumed God as the foundation of all truth, but he also raised the questions that led to postmodernism.

Descartes sought for a firm foundation for truth by seeking that which he could not doubt. He found that he could doubt virtually everything. Of course, it can be hard to doubt the existence of the chair on which one sits, but some doubt can always be suggested. For example, perhaps the chair is the result of a posthypnotic suggestion and an audience would see that he was just sitting on air. Obviously, this is unlikely but the quest for a firm foundation for knowledge is that it be completely unchallengeable at all levels.

Descartes found there was only one unchallengeable fact, and summarized it in his famous “Cogito, ergo sum.” This is generally translated as “I think, therefore I am.” The one undeniable fact was that he was thinking, for otherwise how could he challenge truths? I believe his equation as stated is too optimistic a position. It would be better translated as “thinking, therefore being,” for “I” has no unchallengeable referent. Descartes quickly dismissed his own skepticism with reference to God and built his philosophy despite his skepticism, as had other medieval philosophers. Unlike Descartes, the postmodernists take this skepticism seriously. For the postmodernist all human intellectual enterprises are suspect because all conclusions can be doubted.

As postmodernism developed, Kuhn (1970) published his work on scientific revolutions. This is a historical treatise which suggests that paradigms—models of thought that attempt to describe reality—are important in science, and determine much of the field’s activity. Our paradigms arise from our culture and from the history of the science and its past paradigms. These interact with the facts identified or data collected, and shape the expression of the facts in theories, thus specifying the range of research seen as legitimate. And the paradigms of science itself change over time, with scientific revolutions arising from the replacement of one paradigm with another.

Combining the skepticism inherent in intellectual enterprises with Kuhn’s notion that science is determined in part by the intellectual enterprises of scientific paradigms raises a serious question about how we define science. Can it be defined sufficiently to encompass the most basic element of all scientific disciplines? What is it that distinctly separates it from nonscientific endeavors? Obviously, the distinction cannot lie in such characteristics as integrity, theory building, or paradigms, nor can it be in publishing scholarly articles. All academic disciplines share these.

I have suggested (Gorsuch 2002a, 2002/2007) that the identifying characteristic of science is seen in the research article. All disciplines that claim to be scientific have a distinctive element not found in papers in other disciplines: a methods section. The criterion for an adequate methods section is that it tells readers how they can replicate the study. Indeed, no finding is taken seriously until it is replicated. (In speaking tours at academic institutions, my audiences have always agreed that this is an essential characteristic of science.) Any conclusion is tentative until it has been replicated. For example, the Festinger, Riecken, and Schachter (1956) book, When Prophecy Fails, examined a group that had prophesied the end of the world. When that did not happen, the leaders decided it was because God had spared the world so that they could preach about a new era. However, others (Dein 2001) could not replicate this result with other such groups and so it has been dropped from many social psychology texts.

When sufficient replications occur, the consistency of the replicable data is referred to as a fact. It is then incorporated into theory, which is then tested in other experiments to see if results are consistent with the theory. Science finds that which can be replicated, and these are considered the norm for that phenomenon, natural laws. Gorsuch (2002a, 2002/2007) has labeled such truth as nomothetic. Only that which is nomothetic, that is, which is replicable and so law-like, is found by science. All nonreplicable events are considered unscientific, and unworthy of further scientific exploration. By the definition of miracles noted above, nomothetic science cannot address miracles, for the latter are violations of the nomothetic. The feeding of the 5,000, as miracle, is a nonreplicable event and so science lacks the tools to investigate whether it was a miracle.

Reductionistic Science

Many psychologists have argued that religion and miracles are a function of nomothetic natural forces. They have sought to explain religion as a function of other variables. Of course religion is impacted by many forces, from the simplest, “how can they be saved if they have not heard,” to the impact of one’s family and community of origin. That is not the issue here. The issue is that of explaining away religion by attributing it to other variables having nothing to do with the truth of religion itself.

Freud is the most famous reductionist. It is questionable in what sense he should be called a scientist, in that he left us no explicit methods by which his work can be replicated. Such methods would need to control for alternative explanations, including, for example, post hoc interpretations. In the record of his work we do not have operational definitions to reduce the interpretation of just any data to support a complex theory. Nor does he give us the necessary checks against the problems of subjective evaluations. We need from him a methods section so there can be formal replication. However, he is accorded the status of a famous scientist by laypersons and professionals, alike.

Freud’s argument in The Future of an Illusion (1927/1964) is that religion is a result of infantile projection of the father figure. Therefore, religion is, as the title suggests, an illusion. Faber (2002) has taken up Freud’s Psychoanalytic approach and his purpose is to demonstrate that “the supernatural is a human fabrication with no basis in reality” (p. 7). He argues that it is rooted entirely in human subjectivity. Others have taken reductionistic positions, but with a scientific base. In multiple studies of motivation, Cattell and Child (1975) considered religion as an attitude based in motivational needs.

There are two major problems with positions such as Freud and Cattell. It is well stated in the introduction to Freud’s book. He assumes that religion is false (Freud 1927/1964), as does Cattell (Gorsuch 2002/2007). If it is false then it must be the function of wish fulfillment, anthropomorphizing, subjective motivation, or other such forces. That does not test the validity of religion. It merely rules it out of the arena of scientific investigations by assuming it to be unreality.

The second major problem with reductionism follows from the first: it fails to consider any evidence that religion is true. Nonetheless, for many of us, the truth of a religious proposition is of prime importance. Faber (2002) states that “The person who attempts to prove or disprove the naturalistic nature of religious experience in a manner similar to that which he would use to prove or disprove the heliocentric theory of the solar system or the molecular theory of chemical bonding will fall flat on his face” (p. 3). Of course this follows when one assumes religion is false. Such lack of concern for the evidence of religions permits Faber to talk about religion as a single class, with only minor reference to variations in religions. The reductionism of Freud and Cattell makes their work irrelevant to the task of understanding how miracles and science relate. Thus Freud, Cattell, Farber, and others in a reductionistic mode are irrelevant to the present task.

The Limit of Science

As science is nomothetic and so addresses that which replicates, all nonreplicable events are not scientific. It is not that science rejects them but rather that they are outside the scope and paradigm of science so that science cannot even discuss them scientifically. To discuss them is to leave the area of replicable events. However, our question is, are there nonreplicate events? If there are, can they be studied? What would be the rules of evidence? Do they apply to miracles?

Gorsuch (2002a, 2002/2007) suggests that there are nonreplicable events. Consider any major event in a person’s life, such as graduation from high school. Can a person replicate that event? No, as the prior experience of graduating would make it a different event for that person. In like manner, one can not replicate reading this book because the first reading changes the nature of the experience. Nor can anyone replicate any major historical event. The experiences of George Washington and his contributions to the founding of the United States cannot be replicated. They are unique to him and his times. Therefore they are outside of science. Indeed it seems that, by the definition of science, many of the most important events in a person’s life are nonreplicable and so outside of the scope of scientific inquiry.

The humanities are disciplines that consider nonreplicable events (Gorsuch 2002). The most unique events are communicated by art and literature. The former communicates unique experiences in a nonverbal language form whereas the latter uses verbal language to communicate. History is the examination of nonreplicable events in other time periods.

Whereas science is referred to as nomothetic, nonreplicable events are referred to as ideothetic events (Gorsuch 2002a, 2002/2007). These include all the areas of the humanities, every discipline which does not use replication as its modus operandum.

Being ideothetic rather than nomothetic means that other evidence is used instead of replicable experiments. The courts are good examples of the search for truth in ideothetic situations. There is no way the crime can be replicated but truth is sought as to who committed the crime. Instead of replication, courts examine multiple sources of evidence, including credible witnesses, physical evidence, and circumstantial evidence. The circumstantial evidence used in ideothetic analysis generally includes the results of nomothetic science. If the evidence supports the defendant having violated the law, then he or she is convicted. As this court example implies, the important decisions that are made about a person’s life may be primarily ideothetic. These include basic questions such as, “Does she or he love me?” “Shall I apply for this graduate school?” and “Am I ready to retire?”

Issues of Evidence

Circumstantial evidence is important in establishing ideothetic truth. Consider the case of the Old (OT) and New Testaments (NT). The life and times reported there are supported by considerable widely accepted circumstantial evidence. While the exact events are not recorded elsewhere, it is known that Jericho existed and had its walls destroyed on multiple occasions. The history of Jerusalem in other reports is consistent with the biblical record, as is the knowledge of Roman practices in the NT era. Jesus’ trial and crucifixion ring true to the historians’ expectation. The Dead Sea Scrolls are consistent with OT and NT. While circumstantial evidence is circumstantial, it is important evidence in building a case for the general accuracy of the Bible. This is in addition to the credible witnesses, those who told others of the events and of the many people involved.

Credible witnesses may include the many authors to whom books of the Bible are attributed. It may also include the canonization process, which involved many more people. While these were not witnesses to the original events, they could witness to the usage of these books as true in their own traditions. Since nothing can be completely proved beyond a doubt, some people accept and some reject the conclusions drawn by Christians from this evidence.

Contrasting with the circumstantial evidence for the Bible is the circumstantial evidence for the religion-like fervor of the German Nazis’ notion of the divine ordination of the Aryan race to establish the thousand-year Reich and bring order and its “true destiny” to the world. There was no phenomenological, heuristic, or circumstantial evidence to vindicate any element of that worldview: no evidence for the existence of Uhr Menschen, no evidence of a divine mandate for Nazi or Prussian conquest, and no evidence for Jews being the counterforce to this divine order or the source of social evil. Circumstantial evidence is important in the search for ideothetic truth.

Miracles fall within the ideothetic realm as they are by definition nonreplicable events, but that does not mean they cannot be studied. As with any ideothetic event, they can be studied through credible witnesses and circumstantial evidence. What makes a witness credible is the experience of the person and how we perceive that experience. Who are credible witnesses for discussing miracles? That will vary depending on one’s paradigm. If one is a Christian, then the NT writings are credible witnesses. The feeding of the 5,000 is accepted as a miracle because credible witnesses have reported it. Note that the scriptures do not discriminate between the two types of miracles that may underlie the feeding of the 5,000. Whether it is a miracle of the heart or an actual physical miracle will depend upon the judgment of the reader and the reader’s religious paradigm.

Might there be conditions under which a nomothetic event has ideographic underpinnings? Actually this is a major position of some who pray for miracles. Murphy (1997) formalizes this by means of a model in which God works at the quantum physics level. The quantum physics level is characterized by its lacking the property of any predictability. In her view, God’s intervention is at this level and it works its way up the causal chain to help the person. The observed cause is just the last step of the original divine action at the quantum level of the ideothetic event.

The view that God may work through nomothetic chains appears common. Casteel’s (1955) classical approach assumes God works through the chain of events: “Even the simplest of our requests may entail the changing of a great many very powerful factors before an answer can come” (p. 117). Perhaps the feeding of the 5,000 was a miracle both in that people had food to bring and that they shared it, and that God multiplied the total quantity to insure that it was more than adequate. In illness, people do pray for the physician and other factors in the perceived causal chain of events that may cure an illness. Ideothetic and nomothetic are not mutually exclusive.

Research on Intercessory Prayer

Nomothetic research on miracles can study many facets of it. Who reports miracles? What kind of miracles do they report? What are the conditions under which miracles are reported? What are the antecedents of reported miracles? What is the impact of the reported miracle on that person and on others? Such questions have been raised in regards to prayer. Research on multiple aspects of prayer can be found in Brown (1994) and Frances and Astley (2001).

There are several systems to classify prayer that have been used in research. Some systems are concerned with types of prayer, as historically classified. Casteel discusses prayer as adoration, confession, thanksgiving, and prayer as asking and receiving. The last divides into petitions for oneself and intercession for others. This is similar to Poloma and Gallup’s (1991) division into ritual, petitionary, conversational, and meditative; or to Hood, Morris, and Harvey’s (1993) contemplative and liturgical. Their model includes two forms of petitionary prayer, in which the goal of one is blessings and the goal of the other, material things. Others categorize prayer as to what it is based on and whether it is inward (self), outward (others), or upward (divine) oriented (Ladd and Spilka 2002, 2006).

While each of these systems is appropriate in its context, this current project needs a somewhat different perspective. It needs to categorize prayer in terms of the source of the effect, and whether that effect can be nomothetic or ideothetic. Only the latter would include miracles.

Prayer may have effects through two sources that would be replicable. Together these two have been referred to as subjective effects, although whether that refers to subject effects or effects only seen subjectively is open to interpretation. This category consists of two subcategories. First are the effects on the person who is being prayed for, the personal effectof prayer. Do they feel the prayer has been effective? This is the only criterion when the intercessory prayer is for inner healing of emotional and spiritual health.

If the conditions under which the person feels the prayer is helpful are replicated, then a personal effect has been found. This may be the placebo effect. The placebo effect is found in research on drugs, which finds that many people feel they are helped even when they are in a control group given a placebo, a physiologically inert substance resembling the pill being tested. These are rather large effects and are nontrivial, but they occur from the psychological impact of believing that one may be receiving a helping drug. In many research designs, these are not distinguished from spontaneous recovery; to separate them would require two control groups for which no one prayed, one believing someone is praying for them and another group who were not told the researchers would pray for them. These effects may occur with all types of prayer, including intercessory prayer.

In addition to personal effects, there may also be social effects. This effect would arise from the impact of the prayers on the person’s social network. Thus a congregation may be more supportive when they are all praying for a person, and that may lead to helpful communication or other support provided by the social network. The support could be psychological or physical. These effects would occur to the extent that the community knows about the praying.

There is one effect that is both social and personal: when a person knows that others are praying for him or her. The effect is social in that communication of the praying is necessary for the effect and personal in that it affects the person psychologically and is open to a placebo interpretation.

Personal and some social effects have been investigated by psychologists for a long time. Brown (1994) and Frances and Astley (2001) are appropriate introductions to that literature. Ladd (2007) has laid out a system for examining prayer psychologically. It encompasses the motivation to prayer, the qualities of prayer including physical postures, the experiential content of prayer, and the ramifications of prayer. While he recognizes a possible supernatural efficacy of prayer, that is external to the psychological model. The psychological model is restricted to nomothetic processes.

With intercessory prayer, there need not be a personal or social effect. The effect may happen without the person knowing it and without involving the person’s social network. Such an effect would be objective. Some have tried to research intercessory prayer for such effects.

Current Intercessory Prayer Research

The purpose of intercessory prayer research is to determine if prayer has an objective effect. It is researchable and several studies have been conducted. However these studies are not without their problems.

Research Design

The purpose of a research design to test for the effect of intercessory prayer is to eliminate alternative explanations for the effect if found. First, control must be made for other conditions that might affect the outcome. For example, Galton (1872/2001) concluded that prayer showed no effects because the sovereigns, who were prayed for by everyone in the nation, lived only 64 years, which is the shortest of 11 groups reported, who were not prayed for by everyone. Of course, there were other differences between the lives of sovereigns and other occupations of that time. It may be that the sovereigns, having often intermarried with each other, had a different genetic base. Or perhaps they ate only the best meats, which in that time would be the fatty meats. Of course, there are a host of other differences that could affect the health of one group compared to another. There are so many alternative explanations for Galton’s results that they cannot be informative. The study is only useful in showing the need for controlled studies.

The problems of Galton’s study are now answered by using randomly assigned intervention and control groups. The control group does not receive the intervention, in this case, intercessory prayer, but, because of the random assignment, will not on the average differ from the intervention group. The random assignment controls for all possible influences that need to be controlled regardless of our knowledge of them.

Other alternative explanations that must be addressed in prayer research concern the fact that research design must control for the alternative explanation of the placebo effect, in this case, an effect on the person from knowing that prayers are being offered for him or her. This can be controlled either by having a group that thinks they may be prayed for when no such prayers are being made or by not telling people in the intervention group that they are being prayed for. They are blind as to whether they are in the intervention group.

Research on intercessory prayer needs to control for someone, including physicians and nurses when health is the issue, treating the patient differently because prayers are being given for them. Hence people involved with the person must be blind as to whether they are in the experimental or control group. Evaluating the outcome also must be controlled. A person reading a file for outcome data may make different assumptions and mistakes if he or she knows which person is in which group. Hence the evaluations must also be blind.

Therefore a study needs to randomly assign participants to the intervention group and to at least one control group who believe they may be prayed for just as much as the intervention group. A second random control group would be useful to establish recovery rates for people who have not been told they may have someone praying for them. The study needs to be blind to the condition of the participants in three senses: the participants are blind, others who interact with the participants are blind, and those who enter and process the data are blind to the conditions of each participant.

The Criterion Problem

Prayer research has only investigated areas in which the outcomes would be generally viewed the same by all observers, such as survival of illness and quick recovery. If the research moves to other areas, then who is praying for what may matter. I succumbed to the temptation to pray that the Cowboys, in an important football game, would intercept the opponent’s pass on the next play. To my delight it happened! However, there may have been some mighty prayers on the other side for, a couple of plays later, the Cowboys fumbled the ball and the opponent recovered. Such trivial cases aside, the answer to one person’s prayer may be offset by another’s. An example would be of a prayer for speedy recovery of the patient that is offset by others praying for a nurse, a prayer that God could answer by having the patient stay in the hospital a few more days to help the nurse. These effects are countered by random assignment but may result in lower effect outcomes.

The answer to prayer is assumed in these studies to be a reduced death rate. It is used as the primary criterion in the studies of prayer noted below. However, is this criterion always the best? Consider the case of a child with major trauma to the brain followed by irreversible coma. It was obvious that doctors accepted that the boy would never recover most brain functions. What is the prayer in this case? My prayer was for either a healing miracle or that the boy would “go to be with God.” He died, and I believe that was an appropriate answer to prayer in this case. But prayer research finds that to be evidence against prayer.

The criterion of survival is problematic because Christians may not pray for life in times when others would expect them to do so. Consider Htu Htu Lay, a Christian Karen leader against the oppression of his people by the Burmese military government (FreeBurmaRangers.org ). Given that his life may often be in danger, we on the outside would expect him to pray for safety. That is not what he prays for. Instead he prays “not that I may live a long time, for we all must die. I pray that when I die I shall be doing God’s work.” Long life is not the criterion that is the most relevant to many Christians.

Gorsuch (2002/2007) notes that one Christian perspective is that God is less concerned with long life and more concerned with how that life is lived and with suffering. Jesus in the Garden (Luke 22) prayed for release from the task that led to the crucifixion, but added, “Your will be done.” Suffering on the cross was the result of that prayer. How would this be judged in research on intercessory prayer?

The NT has virtually no passages that tell us how to live longer. Nor does it address this topic at a length suggesting it is a major issue. Yet there are passages about comforting the afflicted and visiting the sick. Healing is a topic and a focus of Jesus’ miracles, but these were mostly with long-term infirmities and only occasionally about critical illness. Even Jesus lived a short life (33 years), and died with minimal suffering for a crucifixion. God seems more concerned with suffering than with long life. Perhaps some deaths should be put in the “God answered prayer” category.

Whether a prayer is answered is always open to the post hoc problem. One may consider a yes or a no to be the answer to a prayer. Such situations are common in ideothetic life. Numerous questions can be considered answered with either response. These include an application to graduate school, taking a driving license test, and proposing marriage. However, these are simple questions with relatively objective answers. For prayers with numerous possible answers, such as “God, what major should I study?” measuring the outcome of the prayer is probably too subjective and ideothetic to be able to use it as a criterion for research at this time.

Present Intercessory Prayer Research

Hodge (2007) has reported a meta-analysis of the literature testing intercessory prayer using 17 studies found through multiple searches of the literature. Meta-analysis is an excellent method of summarizing studies. Instead of subjectively comparing studies, meta-analysis summarizes studies statistically. Effect sizes are entered for each study, along with other information such as the number of participants. The studies are weighted by that number and an overall effect size is computed and tested.

Hodge (2007) reports a significant effect of p <.02. The effect size was .17. However this is based on the published studies. It is possible that studies of prayer are more likely to be published if there are significant results. Journal editors are more receptive to significant studies. Authors are more likely to work to get significant studies published. Fortunately, meta-analysis includes a procedure by which one can determine the number of unpublished studies needed to reduce the observed effect size to nonsignificance, assuming only nonsignificant studies go unreported. Hodges reports that it would require 32 unreported nonsignificant studies to reduce the effect size to nonsignificance.

Note the small effect size requires a large number of participants distributed equally between a prayed-for group and control group(s), in order to be significant. If the number were just enough for a .17 to be significant, half the studies would fall below this mean of .17. Therefore an even larger number is needed. Details of the study would need to be examined in a power analysis to estimate the number needed as well as the investigator’s tolerance for Type II errors, but it would seem that at least 300 are needed to have a chance of finding a significant effect with both Type I and Type II errors held to the .05 level.

The small effect size may be a function of uncontrolled variables, such as low reliability of measures or the impart of other events not reported in the study. With some outcome criteria, the results may be heavily skewed in that almost all may be healed and so the number not healed is small. This also makes effect sizes smaller although meta-analysis can correct for such restrictions of range. Other elements often found in a meta-analysis are not reported by Hodge. Meta-analysis can also correct effect sizes for unreliability, which means that Hodge’s effect sizes are smaller and less significant than if they had been so weighted.

Another set of unreported analyses of considerable importance is that of comparing different types of studies for whether they produced the same results. Probably due to the limited number of empirical studies of intercessory prayer, Hodge did not evaluate whether triple blind studies, in which neither the patient, those evaluating the outcome, nor the medical staff knew who was in the prayer or control groups, found different results than single blind or nonblind studies.

We can examine the results for studies that used at least double blind procedures and had a reasonable number of participants. Here they are:

  • Benson, Duseck, Sherwood, Lam, Betha, Caperter, and others (2006), not significant (N = 1,201).
  • Alves, Whelan, Hernke, Williams, Kenny, and O’Fallon (2001), not significant trend (N = 799).
  • Harris, Thoresen, McCullough, and Larson (1999), significant. (N = 990).
  • Byrd (1988), significant (N = 393).
  • Leibovici (2001), significant (N = 3,393).

Hodge did not publish the effect sizes for these but each is large enough to find significant a much smaller correlation than .17. It appears that when the studies are limited to only the better ones, there is limited support for a positive, though small, prayer effect.

So far the effects, if any, are small. In fact they are so small, that giving the patients a drug placebo may be more effective. There could be some increase in effect size upon investigating parameters in which the studies differ. These include the number, type, and nature of the prayers as well as the characteristics of those who are praying.

A problem that future research in this area may wish to address is what other prayers are being offered. In a study with 1,000 patients, random assignment assures that there are few differences between the control and experimental groups. The other prayers by the patient and by those who know the patient will average out to be the same. But yet other prayers, which could be measured by asking the patents about their own prayers and those in their religious support system, could be measured and parceled out statistically. The result would be a more accurate effect size, possibly raising it above a trivial effect.

Another question not examined in Hodge’s meta-analysis was whether there were dosage effects. In medicine, strong doses of a medicine are generally found to have more powerful effects than weaker doses, additional evidence that the intervention is successful. In the studies, what was meant by a prayer varied widely, as did the characteristics of those praying. They were all religiously committed, but the number in Hodge’s (2007) summary of the studies varies from one person praying for the list of patients to 15 Christians praying for individuals, to rotation among a group of 40 that included multiple faiths such as Buddhists, Christians, shamans, and secular people praying. Such a diversity of interventions makes the overall test from the total set of studies difficult to interpret.

The research on prayer needs to consider each person’s religious motivation. That may be Intrinsic Religiosity (Allport and Ross 1967), in which the person prays as a part of his or her religious life, or Extrinsic and Personal (Gorsuch and McPherson 1989; Kirkpatrick 1989). The latter do not pray as a part of their religious life but pray because they have a problem. The problem leads to the prayer. Extrinsic religion generally relates to negatives about one’s life. On the other hand, Intrinsic motivation is associated with positives in one’s life. Just correlating prayer with other personal or social outcome measures is not meaningful, for the correlation will suggest negatives if the sample contains primarily Extrinsics or positives if the sample contains primary Intrinsics. With sizable proportions of each in the sample, they would offset so that the correlation would be zero. Measuring the type of religious motivation clarifies the situation, and one may well find both a positive and a negative correlation with a dependent variable from the two types of motivation.

The nomothetic research on prayer is not yet sufficient to decide whether there is an objective effect. The effect, if any may be stronger if some of the considerations mentioned above are controlled in the study.

Nomothetic Considerations in Intercessory Prayer

The purpose of this chapter is broader than just whether research has or has not found intercessory prayer to be effective. Instead the purpose is to examine the possible meanings of such studies given the nomothetic and ideothetic distinction explained above. The following discussion uses the intercessory prayer research as a starting point for that discussion.

Casteel (1955) includes the following as problems with intercessory prayer: (1) It may be that God only acts through natural law. Then there is no ideothetic action and theology has become deistic with a creator who is no longer in charge of his creation. In this case, miracles as generally defined do not happen and so will not be found in nomothetic research. (2) It can be argued that God, loving his people, already knows and so provides for them, but this becomes a theological determinism taking people out of the decision making. No parent would always do the best for their child by making all the decisions without the child’s input, so why should God? (3) Then there is the criterion problem we noted above: what is an answer to a prayer? If the answer is yes, it is clearly identified. But the answer may be given in multiple ways. The answer could be “no,” “wait,” or “here is a better answer for your need than what you prayed for.” In the eyes of faith, all prayers have an answer even if the answer is, “no comment.” These are all problems for nomothetic intercessory research.

Another question can be raised about intercessory prayer. Does God wish to pay attention to the researcher’s prayers? Somehow adding one person or even a set of people to pray for a stranger seems less important than the many Christians already praying for that person who know them well. Moreover, there are the thousands of general prayers said each day, praying for all those who are sick. Only if one sees the research’s people as special mediators between the divine and the human would the effect be large enough to not be drowned in the sea of other prayers. The only other option is to assume that God just adds up the number of prayers and then helps those with the most so that adding the research team prayers would add sufficiently to all those being said for a few people that it would put them over the top. This is in keeping with theologies that sponsor prayer wheels so that the prayers are made sufficiently often to manipulate a spiritual force into doing what we desire, but this is a poor fit with Christian theology.

Brown (1994) raises the question of who is to blame if an action fails. If an experiment fails to find an established effect, then Brown suggests the experimenter may be blamed, whether it be the fault of failed apparatus, the experimental design, too few subjects, or other such causes. The ineffectiveness of prayer research may be caused by experimenter-linked phenomena. There are answers that are too readily given for the perceived failure of prayer, “you did not have enough faith,” but Brown’s point is that this is a consideration from within a religious community and tradition.

There are those who hold that intercessory prayer is not testable but for other reasons than the nomotheticideothetic distinction of this chapter. Brown (1994) presents several of these from multiple sources. The most prominent argument is to suggest that the purpose of prayer is less petitions that treat prayer as magic, than prayer as a relationship with God. This, however, changes it from an intercessory prayer to another type of prayer.

Would establishing intercessory prayer on a nomothetic base place prayer in the realm of magic? “Any technology is magic to those whose culture does not understand the technology” is a phrase that could be used to illustrate the problem. How does magic differ from technology? General usage suggests several differences. First, magic is based in arbitrary words and movements, such as “abra cadabra” and waving a wand over a top hat. Another difference is the source of the power that underlies the action. Magic is considered to be a force that has powers over events that are not directly related to the words said or motions made. But these do not seem sufficient critiques, for there are devices such as computers that can be programmed to respond to verbal commands and motions. These devices can have considerable power, such as sending missiles with nuclear bombs.

For magic, other than the illusions of a magician, the difference appears to be having powers over a spiritual force, a spirit that understands and can hear language even though it shows no physical presence. So far, science has not needed spirits as an explanatory principle, and our paradigms are marked by action through a chain of physical events rather than the action of invisible spirits. This may change as more Third World people become involved in scholarship, for the nonbelief in spirits is not universal.

Inclusion of the Leibovici (2001) study along with other intercessory prayer research may be controversial because it was retrospective. A person gave a prayer for half the patients after treatment was completed but before the outcomes were examined. This is unusual because one of the conditions normally expected in assuming that an intervention caused the results is that the cause be prior to the outcome. Leibovici was following a different line of thought: God is outside of space and time, and so can act in prior times as well as current ones. The limitation is that God does not change an outcome after we have observed it.

Note that God could only perform a miracle post hoc before those praying know the results. Hence one can pray post hoc for events that are completed until one knows the outcome. This begins to feel like a good basis for a science fiction story. What, for example, might happen if we pray for a completed event for which we already knew the outcome? If God is independent of time, he could change it; this would be an answer to the prayer. While God may be able to track parallel time lines, we cannot. So if he answered such a prayer, we would remember it as it has been changed, and have no memory of our prayer that changed it. Then there would be no prayer to influence God to change it, and so no answer to any prayer. It is best to go with Leibovici and keep the discussion to those results that we do not yet know.

Is retrospective intercessory prayer legitimate science? By the definition in this chapter that science examines that which replicates, the answer lies in others applying this easy methodology to more cases. If it replicates, science should take it seriously. Before we reject post hoc prayer, we need to realize that physics is not quite as certain about cause and effect, at least at the quantum level, as often implied. In quantum physics, Schroeder’s cat may not be said to be alive or dead until someone checks to see, which is after the quantum event. This is the same line of thinking as in Leibovici.

A nomothetic definition of science contains no assumption that a cause precedes an event. The only problem is the interpretation of cause when the cause comes after the event. But attempting to solve a problem before there is replication of the results, to assure that there is a problem, is not a good use of our time. Therefore judgment on the interpretation of post hoc prayer needs to await replication of Leibovici (2001).

Science normally accepts replicated data consistencies as facts when they fit within a paradigm providing a physical vector for action. Light produces action by the light streaming from the source, which can be followed and measured at any point. Of course, the major question about a replicable prayer effect is the interpretation of the causal agent. Religious people immediately say the cause is God, while nonreligious people posit a hidden unknown cause or something more general. James (1902/1985) posited an unknown force whereas Cattell (1938) laid out a theory of a theopsyche, a psychological force that operates as if there were a supernatural force. Not having a known vector for the action is not an unknown problem in science. Despite the discussions of graviton waves, there is no known physical vector by which gravity operates. In the last volume of his treatises, Newton makes exactly this point and suggests that it shows God acting (Simpson 1992). If there is a replicable nomothetic effect in prayer studies, then a search for a physical vector or action would be warranted. This would be part of the effect of the function of God as creator, and not a miracle as normally defined.

Nomothetic research can only establish replicable data consistencies and so prayer only as a natural phenomenon. This does not mean it is not, as in Christian theology, from God, but that it is a consistent effect. However, being natural, it would operate by some mechanism, which would be the object of further research, and would not be considered a miracle by our standard definition.

Ideothetic Considerations in Intercessory Prayer

The traditional definition for a miracle is always open to a god-of-the-gaps effect. When little was known of the nomothetic principles of medicine, any healing could have been attributed to a unique event outside of natural law, and so a miracle. Thus a person’s not getting cholera in the medieval era may have been seen as an answer to prayer but might be explained today by that person following the custom of only drinking tea brewed the old-fashioned way, heat the water until it is at a rolling boil, which would have had the then-unknown effect of killing the cholera germs in the water.

On the other hand, few areas of medicine are so well developed that only the nomothetic seems to apply. People respond differently to illness and medications when physicians treat by general nomothetic principles. “The research shows successful results in 82 percent of the cases, so let’s try it,” which may leave God room to intervene in the other 18 percent. A conversation with any medical doctor will turn up many cases of unusual results. This may be interpreted as possible miracles by the religious or as variations on undocumented nomothetic principles in the case of the nonreligious, both of which are, of course, equally post hoc explanations. Whether there is an ideothetic event that can be labeled a miracle would need to be evaluated by ideothetic standards. These are set by the theologians of each faith that wishes to address the issue of miraculous answers to prayers.

Interaction between Nomothetic and Ideographic

There is no reason for an event being just nomothetic or just ideothetic. Joint events happen in everyday life. A unique event such as a successful birth has nomothetic aspects; the course of the pregnancy is well known for the average woman, as are the effects of medical interventions to control a wide range of possible problems. Yet, for that mother, the birth is a very ideographic event and may also be ideothetic for the physician if unusual situations arise. Events can be both nomothetic and ideothetic. Therefore it is possible to have miracles combining with nomothetic events. Every science-based clinical treatment, whether by physicians or psychologists, is based on bringing the nomothetic of science into the ideothetic of a person’s life. God may also bring the ideothetic into a nomothetic event.

If there is an obvious cause for the answer to the prayer, that may or may not be considered an answer to prayer. Some Christians consider a prayer for healing to be answered if the doctor changes the patient’s prescription and the new one is more effective. Indeed, it is common for people to pray for the doctor to make the best decisions for the patient. Few limit an answer to prayer to just those events that cannot be explained any other way.

Interpreting normally nomothetic events to be answers for a prayer distinguishes between direct causation and sanctioning. Smith and Gorsuch (1987) note that courts use this distinction. The physical cause is that which is the direct physical agent. The sanctioning agent is that which guarantees the outcome by this physical agent or, if that is not effective, by another physical agent. When one prays for improved health, the physical cause may be from the taking of a medicine, from a surgeon’s skill, or from a set of radiation treatments. The sanctioning agent is the doctor who selects between these physical causal agents. If the first physical causal agent selected by the doctor is ineffective, the doctor selects another physical agent. The physician may try several approaches until one works. Surely God can do the same. The prayer is for God to sanction the healing by assuring that some physical agent produces the healing. Is it any less of a miracle if God works through a physician? Indeed, is that not the task of a Christian physician, to heal people in response to God’s desires for all to be healed?

This approach means the ideothetic intervention can take place in any of the events leading up to the answer to a prayer. It may be, as Murphy suggests, in the randomness of quantum physics. Or it could be in helping a physician recall a critical article at the most appropriate time. Or it could be in having a cab handy to take a woman suddenly going into labor to the hospital. All of these look nomothetic, only when we look just to the immediate cause, and forget that God as a sanctioning agent may work through dozens of physical agents.

From an ideothetic perspective, God can work miracles at all levels of causal chains and outside of causal chains. This means that God can work along with natural law, and that it may be difficult to separate the nomothetic from ideothetic interventions.


Science only works with replicable events; and miracles, being nonreplicable, are outside the scope or domain of what have been called the exact sciences. However, the normal definition of miracle makes the mistake of creating a false dichotomy. It implies things must result from either an immediate natural cause or a completely ideothetic intervention. With our distinction between nomothetic and ideothetic, there is ample reason to suggest that both are widely involved in our lives. Both scientific interventions and God’s actions flowing from sanctioning a given result may well be the correct understanding; and our definition of miracle may need to shift to take that into account.

What happened at the feeding of the 5,000? God may have been a direct cause by multiplying the loaves and fish; or he may also been a sanctioning cause by reminding a number of the people at the start of the day to bring plenty of food, and motivating them to share it. Indeed, God may have done all those things at once. The appropriate disciplines need to define miracles so that both are recognized as acts of God, one by direct cause and one through sanctioning. Who of us is ready to limit God as to the natural and transcendental channels through which he can and may operate? This is his world and all the channels, presumably, are open to his action.