Over the last few decades, several scientific studies have attempted to measure the effectiveness of prayer in bringing about healing. But can we really test the therapeutic effect of prayer in the same way that we test conventional medicinal treatments? On the contrary, in this article, I present three arguments in support of the conclusion that such attempts to measure the effectiveness of prayer are based on flawed assumptions. One implication of this is that the failure to draw a positive statistical relationship between prayer and healing does not constitute good evidence against the value of prayer.