There have been several studies in the last decade focused on evaluating the efficacy of prayer from a scientific standpoint (see here and here for two examples). The studies I am familiar with were conducted in conjunction with medical facilities to evaluate the efficacy of prayer for the sick. The results of these studies vary. Some show a slight improvement in the control group, some show no difference, while others show a decline in health. Apart from the inconclusive nature of the results, I think such studies are misguided in principle, and tell us little, if anything about God and prayer. To understand why we need to consider the scope of science.
There are two types of causes in the world: event causes (impersonal), agent causes (personal). A series of dominoes falling would be an example of an event-cause. Why did domino Z fall? Because domino Y fell (event) onto domino Z. Why did domino Y fall? Because domino X fell on domino Y. The series of event-causes and effects goes on indefinitely. Each effect is caused by a prior physical event, which in turn was the effect of a previous physical event ad infinitum. No event in the chain can do anything other that what it does because event-causes do not decide; they merely react. Event-causes passively receive their action from a prior event, and then pass that action down a causal chain in a mechanistic, deterministic fashion.
While event-causes are instrumental-movers who passively receive and transfer action, agent-causes are first-movers who act as the absolute source of their own actions. In an agent-cause there are no necessary preconditions that necessitate any particular effect. Agents are prime movers who simply decide to cause a particular state of affairs and then act to do so. The effects produced by agents are not determined by prior events, but are freely chosen by acting on their own volition. The person who chose to knock over the first domino in the example above would be an example of an agent-cause.
Science is properly equipped to evaluate event-causes in the physical world, not agent-causes. Science can recognize the past effects of an agent-cause, but it cannot predict when or how a free-will agent will act in the future. While science is good for telling us the conditions under which water will boil, science is powerless to tell us what someone else will eat for dinner tonight, or how they will react to these words. In short, event-causes are, and agent-causes are not predictable. The efficacy of prayer is simply beyond scientific predictability. Science measures the effects of natural, law-like causes. When it comes to rational and free agents there are no materialistic, law-like causes and effects to measure with precision. In the same way science cannot predict what requests little Johnny’s mom will respond affirmatively to and which one’s she will not (because she is a personal and rational agent whose choices do not operate according to physical laws), science cannot predict which prayers a personal God will respond affirmatively to and which ones He won’t.
All attempts to make a scientific analysis of prayer are doomed to failure because prayer is not a mechanistic type of thing like physics. Prayer does not operate on a series of fixed laws. You don’t say two of this and two of that and voila…out comes X. Prayer involves an interaction between two personal agents, each possessing his own volition. For a prayer to be answered God must freely exercise His volition in such a way that He decides to act to answer our prayer. God may choose to answer the prayer, or He may choose not to answer; in the same way a teacher may choose to grant a student’s request for an extension on her paper, or choose not to.
Prayer studies err in that they treat prayer as if it were a law-like mechanism or magical incantation rather than a willing interaction between free agents. If God chooses not to respond to the prayers of those participating in the study it is concluded that prayer is not efficacious for healing. This conclusion, however, is non-sequitar. When dealing with personal agents there are a wide variety of reasons they choose to act or refrain from acting. Maybe the prayers were not answered because God did not want to heal the individuals being prayed for. Maybe the prayers were not answered because the people praying for them were praying to a false god, and the real God knew if He answered their prayers it would wrongly convince them that the god they prayed to was the true God. Maybe God did not answer the prayers because He does not like being put to the test. There are a host of possibilities, all of which preclude scientists from making any definitive judgments regarding the efficacy of prayer.
This is not to say empirical science is unable to shed any light on the issue. If no prayer ever prayed was ever answered that would be good reason to conclude that God is not concerned with our requests, we are making the wrong kind of requests, God is not powerful enough to answer our requests, or there is no God to hear such requests. If even some prayers are answered, however, and there is no natural explanation for the effect in question, that is good reason to be open to the existence of God and the efficacy of prayer. Granted, there would have to be some standards for testing these experiences to make sure they were of divine origin (were the results likely to have occurred without divine intervention, were the results statistically likely or naturally possible, etc.?) but they could be tested.
Personally, my experience has convinced me that God exists and He answers prayer. While He has chosen to answer only a small portion of my prayers, it is clear to me from those examples that God is willing to answer some prayers, including prayers for healing. Not everyone we pray for is healed, but there are those who are. I don’t need science to tell me that!