The Probability that God Exists is 10%


Consider the proposition “some sort of deistic or theistic God exists”, where “God” need not necessarily be a personal God, but is understood to be more than a transcendent spirit or panentheistic notion of ‘God as nature’. What is my level of confidence that this proposition is true? My best estimate is on the order of 10%. In this piece I will explain how I arrived at this figure, and why I think it is the most reasonable rough indication of humanity’s current state of knowledge on the matter.


When I say that “the probability that God exists is about 10%”, this should be understood within a (loosely) Bayesian framework. In other words, the probability figure is an expression of one’s confidence in the proposition; a statement about how much we know and how much we don’t know. I am not saying God exists in 1 in 10 possible universes, or that the existence of God is literally a random event that would occur one time out of every ten. Many people think that God’s existence is necessary, meaning that if God exists, he necessarily exists – there is no possible way he could have failed to exist. Perhaps that is true, but the question is, how confident can we be that it is true? Unless we assume that all our thoughts and reasoning regarding necessary beings (or similar entities) is infallible, it seems at least possible (perhaps likely) that we could falsely come to believe that something necessarily exists. My probability estimate is thus designed to capture these effects of uncertainty. As such, I do not think it is inconsistent with arguments about the necessary existence of God.

Why 10%?

The particular figure of 10% is fundamentally derived from an excellent survey on philpapers (, which shows that about 15% of philosophers believe in God, while 73% are atheists. These results are broadly comparable to surveys of scientists, which indicate that something like 30% of scientists believe in God (, I have adjusted these percentages down slightly for several reasons. Firstly, these surveys (in particular the Philpapers survey) are disproportionately of American and British philosophers and scientists. Levels of religious belief are substantially higher in the US than in many continental European and Asian countries, for which we do not have comparable data (scroll down to the bottom of this page for some interesting data though – American philosophers are much more religious than their continental counterparts). I think that ideally we should consult a representative survey of thinkers and scientists from across the globe, and that if we had these figures, we would find (for example) significantly lower levels of belief among intellectuals in continental Europe and China. Absent such figures, I have made a downward adjustment from 30% or 15% to about 10%. Secondly, levels of religious belief are lower for more prestigious scientists (, which it seems reasonable to believe correlates at least somewhat with intelligence and careful thinking. Certainly the correlation would be far from perfect, but it seems very plausible that there is at least some positive relationship between knowledge and ability, and the likelihood of holding a carefully considered, informed opinion on this matter.

Why Trust Experts?

Many people, especially theists, will take objection to my approach here. They will question the validity of polling experts as a method of determining the state of knowledge. Truth, so they say, is not a popularity contest. I think, however, that such objections miss the point of this sort of analysis. The fundamental problem is that the arguments and evidence for God’s existence is equivocal. Some people are convinced by them, and some are not. What then should we conclude? Should we simply assume that our subjective analysis of the evidence and arguments is definitive? Should we place ourselves in the position of being ultimate arbiters of truth? “Its compelling to me, therefore it is probably true” is not a reliable way of arriving at accurate beliefs, as we know that most people (even informed people) arrive at many false philosophical and religious beliefs through this method. What we need is some more ‘objective’, more reliable method of analyzing the strength of evidence and the quality of arguments. I propose that the best method we have for this is to take a representative sample of intelligent people who are sufficiently well informed about the evidence and arguments, and determine what proportion of these people find the arguments convincing. If only 30% of informed people find an argument (or set of arguments) compelling, then it seems that this argument is not sufficiently conclusive for one to believe with high confidence. We use this sort of reasoning all the time – if only four or five of the twelve jurors think that that evidence is sufficient to warrant a guilty verdict, then we judge that the evidence is not strong enough for conviction, even though some people think that it is. We weight across many people, in the hope that this will produce a more accurate evaluation of the evidence than would a single person alone.

My fundamental argument here, and I cannot emphasise this strongly enough, is that the mere fact that an opinion is your own does not make it more likely to be true. In other words, if only one out of on hundred informed experts believed a certain fact to be true (and let’s assume there’s no evidence of a conspiracy or the like), then we should be pretty confident that the 99 are right and the one is wrong – even if that one lone expert happens to be you! Unless that lone expert has some very, very compelling reason to think their opinion is privileged (e.g. maybe they have access to a secret document no one else does), this expert should admit that, despite how convincing the case feels to them, it is unreasonable for them to place their own judgment above that of their 99 peers.

Few people like this idea (“truth isn’t a democracy!”), but I ask how one can possibly justify giving one’s own views epistemic privilege? It is good arguments and quality evidence that are indicative of truth, and the way we attempt to track which arguments are good and which evidence is sufficient is by seeing what proportion of informed persons find them to be so. Obviously there are problems here with cultural presuppositions, biased selection, institutional barriers, etc, but I think it is hard to argue that these problems are greater when we take the average opinions of a large group of people than when we simply considering a single individual’s opinion, with all their unique biases and quirks. Crucially, this argument applies even if that individual happens to be you. Obviously we have direct, immediate access to our own opinions, something we do not have for the opinions of others (though the depth of our insight into our own reasoning processes is quite limited – see Nevertheless, it is not clear why this more immediate access should mean that our own opinions are more likely to be true. Accessibility does not imply truth. It simply means that the ideas and opinions feel more compelling to us, regardless of whether they are right or wrong.

Thus, I think this method of ‘averaging experts’ is the best (albeit imperfect) method we have for trying to determine how the evidence stands on complex and difficult questions like the existence of God. Theists may accuse me of constructing an elaborate justification for a method which ultimately confirms my own bias (since I am an atheist), however I would argue that this method actually yields a much higher probability for God’s existence (10%) than most atheists would generally admit to or feel comfortable with. Thus I think that the confirmation bias argument is at leas, somewhat less plausible than it may initially seem.

Philosophers of Religion

One challenge to my argument derives from the observation that, of philosophers who specialise in ‘Philosophy of Religion’, 72% are theists, compared to the 15% base rate for philosophers in general. This seems potentially to be evidence that, of philosophers who focus specifically on examining the arguments for and against theism, and various relevant philosophical problems, a considerable majority come to be believe in God. There is, however, an obvious problem of causation here. Do philosophically-minded people who are also religious tend to disproportionately specialize in philosophy of religion (so that belief leads to this specialization), or do specialists in philosophy of religion initially more-or-less resemble other philosophers, but later become theists as a result of their exposure to the strong arguments in its favour? Although it is very difficult to say, I think there are good reasons to think that the former explanation plays the dominant role here.

First, it is important to understand that philosophers can select more than one area of specialization for the survey. Most philosophers (if you browse their profiles) have more than one specialization listed, as their work spans a number of different areas. It seems very likely to me that already-religious philosophers are more likely to include ‘Philosophy of Religion’ on their list of specializations, as (regardless of whatever other work they may do), they also have an interest in these matters, precisely because they are religious. Atheistic philosophers are much less likely to do this, resulting in a significant inflation of the relative number of religious philosophers listing ‘philosophy of religion’ as a specialization.

Second, the difference between specialists and non-specialists on the question of theism is very large, far larger than any other such differences. The difference in percentage of theists between specialists and non-specialists in the philosophy of religion is 56%. The next biggest gap is 30% on a rather esoteric question in decision theory, followed by 23% for the B-theory of time. Most specialization effects are much smaller, on the order of 5%-10% or so (see If specialization allows philosophers to focus on the specific arguments surrounding a particular issue and hence arrive at a more reliable, better informed viewpoint than their non-specialist colleagues, it seems that this should apply to a broad number of questions. Perhaps not every question, but still a good number of them. Instead what we see is that the effect is generally fairly small for most questions, and for religion in particular it is dramatically larger (almost twice the size of the second-biggest effect size). I think this is most plausibly explained by the fact that much fewer people specialize in (say) philosophy of time because of a pre-committment to the B-theory of time, whereas that is a real and significant factor in the choice of philosophers to specialise in philosophy of religion.

Third, I made an effort to actually discover whether particular philosophers who list themselves as specialising in philosophy of religion came from a religious background, or whether they converted later as a result of exposure to philosophical arguments. Naturally, there is no direct data on this. What I did was to browse through philosopher’s bio pages on the philpapers website, looking for those who listed ‘philosophy of religion’ under their ‘area of specialization’. I then googled their names to find any information available about their religious background. I found 15 scholars with philpapers bios who both specialised in philosophy of religion and also had a stated position (i.e. those who responded to the survey and made their responses public). This may sound low, but remember that only 47 philosophers in total described themselves as specialising in philosophy of religion ( Of these 15, all were theists. I could only find background information on about half of them, but all those I could find information on seem to have been raised as Christians. My methodology here is subject to question, as no one explicitly mentions their upbringing. Instead, I used attendance at a theological seminary or religious college, or completion of a theology degree, as proxies indicating probable pre-existing religious belief. I have included a table of all the scholars I evaluated below. My conclusion from this analysis is that the data are most consistent with the hypothesis that philosophically-minded Christians selectively choose to specialise in philosophy of religion, rather than existing philosophers of religion being led to belief on the basis of the quality of the arguments.

Name Position Background
Garrett DeWeese Theist Dallas Theological Seminary before PhD
Daniel von Wachter Theist Intermediate Exam in Protestant Theology before PhD
Tyler Dalton McNabb Theist B.A. Biblical Studies before PhD
C’zar Bernstein Theist No info
Mark T. Nelson Theist No info
Ben McLean Theist RLDS member, seems for some time
John M. DePoe Theist B.A. Philosophy and Theology
Jonathan Fuqua Theist No info
Kenneth L Pearce Theist No info
Ben Arbour Theist Southwestern Baptist Theological Seminary before PhD
Patrick Toner Theist BA from Franciscan University of Steubenville
Lincoln Stevens Theist BA from Asbury University, Christian liberal arts college
Christopher M. P. Tomaszewski Theist Attended S. Charles Borromeo Seminary before MA
Andrea Ciceri Theist No info
David McNaughton Theist No info

Past Experts

Some have raised the question of why only current experts should be counted. There seems no particular reason why great thinkers of the past should not also have their opinions included in the analysis, and were we to do this we would find the proportion of theists considerably higher than we do currently. I have a few things to say on this matter. First, obviously we do not have the data for past thinkers, so we cannot readily include it in our analysis. Even if we do know the likely direction such an inclusion would have on the probability, we don’t know the magnitude. Second, it must be remembered that history is long and intellectual thought diverse. Many ancient Greek thinkers, and arguably also many Buddhist and other non-western philosophers, would not count as theists in anything resembling the usual modern understanding of the term. In other words, if we are thinking of including past thinkers we cannot restrict ourselves only to medieval and enlightenment thinkers from Europe and the Middle East. Third, it must be remembered that although the proportion of thinkers who were religious in the past was higher, there were also many fewer of them than there are today, meaning that including them in the overall average would have less of an effect than one may naively imagine. Fourth, if one is to include past thinkers, it seems reasonable to include future thinkers as well. Obviously we have even less data on what they believe, but it seems at least plausible that belief in God will continue to remain at relatively low levels. Maybe I am wrong about this, but my point is that if we are to imagine what including thinkers from the next 20 or 40 years would do to the average belief in God, it seems most likely that the percentage would fall. This mitigates, to some extent perhaps, the upward affect of including figures from the past.


My argument here is that the degree of confidence one can place in the claim “God exists” is approximately 10%. Error bars are wide here, so I think one could quite justifiably argue for figures of 20 or maybe 30 percent, or for 5% or less. What I would say, however, is that figures that are ‘dramatically different’ from 10%, say something like 0.5% or 95%, are difficult to justify. I just do not think the degree of honest, intelligent disagreement about these matters merits such strong claims. I also think that theists should take this evidence seriously. The plain fact is that a large majority of philosophers do not believe in God. This obviously is not decisive proof of God’s non-existence (10% is hardly decisive), but it is, I think, more than enough to ‘sit up and take notice’. I think it should lead theists to seriously and critically re-evaluate the strength of their convictions, beliefs which rest ultimately on philosophical positions (even if one thinks that God reveals himself directly to people, that is actually a belief that has very particular philosophical underpinnings and implications). If a theist believes that they have a ‘killer argument’ that allows them to fairly easily and quickly dismiss the majority opinion of philosophers – people who think long and hard about these sorts of questions – I think it is very unlikely indeed that such a retort has not already been advanced (probably in a much more sophisticated form) by some past or present philosopher (for example, if you think belief is primarily a matter of faith and not reason, that is a heavily contested philosophical position called Fideism). The point is, whatever a theist may say about why they believe, their belief system rests upon certain philosophical notions or presuppositions. It is unavoidable. Given that the group of people in the best position to consider the relative merits of these sorts of ideas generally are not religious, I think that is strong reason for the theist to critically re-consider how genuinely confident they can and should be about their religious beliefs.





3 thoughts on “The Probability that God Exists is 10%

  1. James 2 questions:

    1. Have surveys of experts in philosophy produced reliable results in the past? If not what is the basis for relying on them? I could buy a poll of experts as a reasonable Bayesian prior but we need to start updating based on some data.

    2. Do theologians and philosophers have any expertise? When you are not part of a field, it can be hard to tell if someone (or a group of self-appointed experts) has expertise. I apply these three tests a) Can they make reliable predictions on matters that are not already known? b) Can they offer excellent and parsimonious explanations for things that otherwise do not make sense? c) Can they give you designs/recipes for achieving certain objectives, recipes that were not already known by induction?

    Arguably theologians fail on all three, with philosophers only slightly ahead of them.

    Take theologians. They confidently predicted the second coming of Jesus within a short time. Their views on cosmology, history of the universe, and the origin of mankind were all utterly wrong (now revised as only metaphorical, but fervently believed before they were disproved). What we get instead is “explaining away” eg ‘sin’ as a pseudo-explanation for suffering. As for designing things, think about the design of the priesthood and in particular the design of the “repent and sin no more” model for reforming child abusers – which was an epic failure. Religious traditions do embody some knowledge, mixed with large admixtures of nonsense, which knowledge is usually inductive truths derived from experience.

    Of course philosophy can be useful for clarifying conceptual confusions per Wittgenstein. But the history of philosophy is a history of embarrassing error for any area in which clear factual information later became available. Thus I doubt the expertise of philosophers in any matter of fact.


    • untimelyreflections, your comments do embody some knowledge, mixed with large admixtures of nonsense. Contrary to your belief that a metaphorical reading of Genesis is a modern revision, this has been an accepted interpretation for millennia (try reading up on the Jewish philosophers Philo and Maimonides, or the Christian theologians Clement, Ambrose and Augustine). It’s no novel suggestion that the purpose of the myth genre was not to present scientific/historical truths about the universe, but philosophical ones (which, dare I say, have even greater implications). In any case, the views of theologians are not represented in select opinions of first century Christians. Certainly we should not dismiss theologians across the ages for these gaps; nor should we dismiss the medical field for Hippocrates’ gaps in knowledge (particularly those outside his field).

      James, an excellent and thought-provoking essay. Thank you for publishing it. It would certainly be interesting to know figures for the number of trained philosophers who have renounced theism contra those trained philosophers who have renounced atheism. But I can’t imagine we’ll ever see such findings (unless they come from yourself in future years…)


  2. It’s a worthwhile exercise to look at expert opinion, even though it’s an area (like mathematics) where in the end you need to consult your own intuition. Since scientists and philosophers in general don’t mostly bother to think much about arguments for the existence of God, taking scientists or philosophers in general as the reference class of experts is not a good idea. So the class of philosophers of religion ought to be the most relevant. Therefore, as James rightly says, it comes down to inquiring into selection biases in that class, which might explain away the pro-God results of that group. It’s true that a number of people become philosophers of religion because they’re committed to religion in the first place (though the opposite phenomenon is not unknown – in some places, e.g. Melbourne, people have become philosophers of religion because they don’t like religion and have a commitment to freeing the populace from the chains of priestcraft and superstition; is James an instance?) But wouldn’t you expect that, if the arguments for God were in fact weak, those who came to look at them professionally would accept that and become atheists? In the end, if you want to go with the experts, you’ll have to go where the experts actually go.
    Past and future “experts: is a tricky issue. “Philosophers are atheists” was a received idea in Roman times, then in medieval and early modern times it was the opposite. With the worldwide increase in religion in recent decades (some of it not very pleasant religion), it’s just hard to say where things are going.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s