A Bayesian Approach to God's Existence
Given the total evidence, how probable is God compared with naturalism?
How to use this guide
This guide is built for a 45-60 minute small-group conversation about "A Bayesian Approach to God's Existence." Open with prayer, read the framing aloud, and use the questions below to surface what people actually think before you walk through the case. Aim for honest engagement over consensus.
Facilitator tips
- Read the lesson before the meeting; you do not need to be an expert, just a guide.
- Resist the urge to fill silence. The best discussions follow long pauses.
- When someone raises an objection you cannot answer, write it down and follow up next week.
- Close with a single takeaway from each member, not a doctrinal summary.
What we're studying
Arguments for and against God are rarely deductive proofs. Bayesian reasoning lets us weigh the cumulative force of many pieces of evidence, the way a jury weighs a case, rather than demanding certainty from a single argument.
The case in brief
Richard Swinburne's cumulative Bayesian case asks: compare the prior probability of theism (given its simplicity) with the likelihood of the data on theism vs. on naturalism. For each piece of evidence E, we ask: P(E | theism) vs. P(E | naturalism). If several independent data points are more expected on theism, the posterior probability P(theism | total evidence) rises. Swinburne concludes this total probability is greater than 0.5; the McGrews argue the Bayes factor for the resurrection alone is enormous. The argument is cumulative, not a single silver bullet.
Argument structure
Conclusion: On the total evidence, theism is more probable than naturalism.
- Theism has comparable or better simplicity (one necessary personal being) than rival hypotheses.
- The existence of a contingent universe is far more expected on theism than on naturalism.
- Fine-tuning, consciousness, moral experience, and religious experience all raise the likelihood ratio in theism's favor.
- Historical evidence for the resurrection dramatically boosts the posterior probability of Christian theism specifically.
What if someone says...
"A personal, omniscient being is intuitively more complex than an impersonal universe."
Swinburne argues infinite attributes are actually simpler than arbitrary finite ones (no arbitrary cut-off points). Reasonable minds disagree on how to measure simplicity, but theism is at minimum not disqualified at the prior stage.
"Each step is contestable; small errors compound across a cumulative argument."
Swinburne treats each piece explicitly and grants worst-case likelihoods. The McGrews emphasize that independent evidence multiplies Bayes factors. Even conservative estimates yield a substantial posterior shift.
Discussion questions
- Do you tend to reason about worldviews in yes/no terms or in probability terms?
- What single piece of evidence most shifts your own probability estimate?
- Is there any evidence that would count as disconfirmation of your current view?
- [Small group] Where in your own life does this question feel most pressing?
- [Small group] Who do you know that wrestles with this — and how could you talk with them about it this week?
Going deeper
- The Existence of GodRichard Swinburne · 2004 (2nd ed.) · Bayesian natural theology
- "The Argument from Miracles: A Cumulative Case for the Resurrection of Jesus of Nazareth"Timothy & Lydia McGrew · 2009 · Bayesian historical Jesus
- Reasonable FaithWilliam Lane Craig · 2008 (3rd ed.) · Natural theology