According to statistical analysis god does not exist

In my field statistical analysis is arguably the most demonstrably powerful tool we have. And given our current quality of technology, it’s incredibly easy. There’s almost no math required! All one must do after formulating their research question, collecting data, and coherently determining variables is press a button. Voila! Thank god for SPSS!! (I’m kidding. Thank IBM for SPSS).

Using statistical analysis we can determine correlations between two or more phenomena (variables). Prior to this we’ve looked at the data and come up with a hypothesis. We’ve also come up with a null hypothesis.

In statistics the null hypothesis is the default position. For example, we might hypothesize that female participation in government is negatively correlated with state corruption. Prior to running our analysis, we will also acknowledge the reverse, or the null hypothesis: Female participation in government has no correlation with state corruption. Then we run our regression to determine if our positive hypothesis is demonstrable. If not, we automatically revert back to our default position, the null hypothesis.

We can do this with the existence of god too, but it doesn’t paint a pretty picture for religious folk.

To be fair, however, our research question has to be a little more complicated than merely, “Does god exist?” For example, our research question would have to be something like, “Is prayer positively correlated with the spontaneous regrowth of amputated limbs in humans?” Or “Is attending faith healing seminars positively correlated with the spontaneous healing of terminal cancer?” In either case our null hypotheses would state that there is no correlation. After our regression I’m almost certain we would find no correlation, and our position reverts back to the null hypothesis.

While this doesn’t answer the question of whether or not god exists, it helps to answer any question about god’s involvement with human affairs. Formulate any scientific question for statistical analysis, and we will find that supernatural explanations are always insufficient explanations. If we are constantly reverting back to the null hypothesis, god’s place in human affairs is essentially nonexistent. In statistics we can sufficiently say god does not exist.

Of course, we have to acknowledge the limits of our research. A deist, for example, would almost certainly argue that these questions are outside the scope of the divine because god remains supernatural by not involving him/herself with natural affairs. I’m by no means deist, but this is the route I take. Although I don’t believe in any god, I see no reason to believe that if one exists it meddles with natural phenomena. Therefore, although statistical analysis leaves no room for the existence of god and essentially says, “God doesn’t exist,” and although we will always revert to the null hypothesis in questions relating to god’s existence, we cannot say, “God doesn’t exist,” because statistical analysis can only handle phenomena we can actually observe. And if god exists and remains supernatural, he/she leaves no evidence for us to observe.

About Rayan Zehn

I'm a political scientist.
This entry was posted in Atheism, Uncategorized and tagged , , , , , , , , , , , . Bookmark the permalink.

13 Responses to According to statistical analysis god does not exist

  1. ohioorations says:

    This is pretty annoying. According to statistical analysis, WE should not exist. This is the wrong angle for the theist OR atheist to take.
    The concept of”chance” can be used by either camp. We expect a certain order in the universe without which the concept of chance disintegrates. Yet, chance is not a force in and of itself but the result of rigorous observation. The necessary order of three universe presupposes chance. Theists augur this is insist evidence fir the existence of a designer while atheists attribute some “vital force” property to chance. Either way, make a choice and live your life

  2. Statistical analysis is no substitute for common sense.

    And everyone knows that numbers will tell you anything you want if you torture them with sufficient brutality..

    Statistical analysis in the hands of the agenda-driven atheist is numeric torture as sufficiently brutal as it gets.

    • charles says:

      I can’t let these statements go, even though I know from past experience that I should ignore SOM…
      These kinds of ridiculous statements are made all the time, usually by people who do not understand statistics. I have no idea if you do or not, but what you are saying is false.

      Common sense can lead to all kinds of errors, like thinking the world is flat. Science and the proper use of statistics results in all kinds of knowledge and innovation.

      Everyone knows that all statements other than this one that begin with “everyone knows” should be ignored.

      Yes, people can abuse numbers and lie, as they can with all forms of information. That does not at all mean that statistics and statistical analysis are worthless. And it is all the more reason for people to understand math and statistics, to avoid being duped. The value of statistical analysis is that it quantifies uncertainty. Properly done and properly reported, the assumptions are clear and the limits of the conclusions are clear. It is not a perfect tool, but it is an excellent tool. To devalue statistics because liars can misuse it and people who don’t understand it can misinterpret it is completely unreasonable.

      There. I feel better now.

      To the point of the blog… If the null hypothesis is “no god”, then to reject the null hypothesis, one would have to have data that is unlikely under the null hypothesis. But this is pretty much impossible to do because how does one define these probabilities? What is the the probability that humans exist if there is no god? Nobody can assign a probability to this. All we can do is talk about it and assign personal levels of belief to these sorts of statements.

      I don’t think the author has shown that statistical analysis proves god doesn’t exist, or that prayer doesn’t work… the author has just asserted their belief that there will never be sufficient evidence. In theory, if prayer did work, a hypothesis test could reject the null hypothesis. I’m not sure why the author thinks we would constantly revert back to the null hypothesis, other than an a priori assumption that God does not intervene (a view I agree with, based on my experience). The data could lead us to reject the null, though. Maybe I misunderstood the post.

      • Charles,

        Great point!

        Next time, instead of saying, “Everyone knows…,” I’ll say “Everyone knows except for atheists who deny everything that doesn’t support atheist dogma.”

      • Rayan Zehn says:

        Hey Charles,

        I gotta admit. You summed up my post better than I could. Thanks for clarifying my points.

        To answer your question, in my field (international politics) we generally say we “revert back to the null hypothesis” because the null hypothesis in politics represents the generally accepted state of politics. Using my example of fempar in the post, it used to be generally assumed that female participation would not affect corruption perceptions. I ran a multivariate regression and a moderately strong correlation between female participation in government and decreased corruption perceptions using data from Transparency International. My P-Value was a respectable 0.04

        I ran another regression too. And I honestly thought Islam would contribute to corruption perceptions. I wrote it in my hypothesis. When I ran the regression (for this variable I used a natural log), I got a weak 0.09. To make matters worse, my adjusted R-Square for the entire regression was 0.12. In this case, I wrote “The data does not support my initial hypothesis. Until more data is collected we should revert to the null hypothesis and assume that Islam plays no role in corruption perceptions.” I ran this regression prior to the Arab Spring. I wonder what would happen if I ran it now.

        I showed you some of my work merely to explain why I used the vocabulary I did. I’ll look at other fields to get a good idea of what kind of vernacular they use when addressing the null hypothesis. Anyway, thanks for the comment!

      • ohioorations says:

        I never would assert that stistics are worthless. I LOVE the field of statistics and delight in the idea of mathematically modeling as much of our universe as possible.

        However, one axiom of science states it’s very definition: science is refutable. If I cannot test your hypothesis, your hypothesis is merely a position, an opinion to which you are entitled, but cannot masquerade as science. Atheism makes provocative arguments as does theism. I have heard both camps espouse science when necessary.

        What fascinates and frustrates me is when either camp tries to invalidate the other as pseudoscience.

        Theism, like atheism, it’s a choice to define the universe based on what you KNOW and what you THINK you know. Science is ever changing and expanding, explaining the “how” and “what”. Faith and theism attempts to explain the “why”. Science doesn’t care about the”why”. Religion and theists care.

        Faith is a choice to believe the truth that makes sense to that individual. Atheism is a similar choice, though the most flawed philosophy of all. The premise that there is no god presupposes the atheists have all knowledge and has concluded that there is no god.

        Preposterous! The skeptic or agnostic has stronger legs to stand. The atheist can inject science, math, and any other mechanism to make the case, but can never make the case they are omniscient, therefore god doesn’t exist. That’s foolishness and arrogance. They have made the choice of presumptive omniscience. I don’t buy it.

      • Ohio,

        I, too, love statistics.

        Statistics is math that tells a story which leads to the truth.

        In my present field of biotechnology, we use linear regression to construct a standard curve for a DNA, RNA or protein sample.

        We use linear regression to test the accuracy of our lab results.

        Since religion, the belief in God, or God, Himself, can’t be tested in a controlled lab or with a viable collection of samples, the atheist’s attempt here to use statistical analysis is a rampage of the ridiculous.

      • charles says:

        I currently refer to myself as a skeptic. One reason is that so many people assume “atheism” means a claim to know there is no god. In my experience very few atheists make that claim. If you want to persuade atheists, don’t overgeneralize.
        I’m guilty of it too. I say “Christians believe…” and later I realize I shouldn’t have.

  3. charles says:

    Hi Rayan,

    I wasn’t so much referring to the words “revert back to the null” as to your reasoning. Maybe this isn’t what you meant, but when I read your post it sounded like you were saying that statistical analysis *cannot* reject the null in cases of the supernatural. I agree that this is true for cases where it is difficult to calculate the probabilities under the null, but it is not true for all cases. Prayer for example. Under the null, prayer would do nothing. If you get a result that is unlikely just by chance, then you can reject the null that prayer does nothing. That would not necessarily imply the existence of God, as there might be some other mechanism, but further experiments could be done to clarify what is really going on.

    I am always happy when people use and see the value in statistical analysis. I like to give a warning though that it is vital to understand the underlying assumptions behind the methods you use. For example, in regression, the usual assumptions are independence, linearity, and normality (of the residuals). You also have to be very aware of how the data were obtained in order to know how far you can generalize (was it a random sample?, a convenience sample?). So keep using statistics, but keep the assumptions in mind at all times.

    The other advice I always give is to plot your data. That helps enormously in avoiding errors and in evaluating the assumptions. While you can just “press a button” and get results, there is so much more to it than that. I’m sorry if you already know all these things and I am preaching to the choir… Thanks for bearing with my unasked for advice =)

    • Rayan Zehn says:

      Sorry for taking so long to reply. I slept in today. Woke up, saw your response and went back to bed. Then forgot to respond myself.

      I think there’s been some miscommunication. I meant to say that we generally should reject the null in terms of supernatural explanation. If you were referring to my paragraph about deists, I stand by that. In order for something to be supernatural, it can’t interact with a natural world. Otherwise we’d be able to observe it, collect data from its actions, and form hypotheses — meaning it would be natural, not supernatural. If we can collect data, etc., then it is not supernatural, but natural.

      I think (or at least I hope) we’re on the same page. And I hope this clarifies my point.

      Also, (I’m sure you’re familiar with this, but here goes) there’s a rather humorous study (humorous because it was funded by the Templeton Foundation) about prayer’s effects that, surprisingly, in one part of the study, suggested a negative correlation between prayer and complication-free recovery after medical procedures.

      • charles says:

        Thanks for the clarification, and yes I think we are on the same page. I do not see any evidence for the supernatural. Yes, I’ve seen that study. Very interesting.

        About the idea that the supernatural not being able to interact with the natural world. Shouldn’t there be a distinction between being able to observe a supernatural being and the effect of a supernatural being? I suppose it comes down to semantics, but it seems to me that if a supernatural being existed, there is no reason to think it impossible for that being to interact with things in the natural world. That being said, I don’t think we have any evidence of such a thing happening.

        One of my reasons for becoming a skeptic was the lack of experience of the supernatural in my life, which seemed inconsistent with what the Bible claims the life of a Christian should be like. But as I’ve thought about it, I’m not sure what evidence would ever count as “supernatural” for me for reasons similar to what you are talking about. I guess whatever word I use for it, the bottom line is that what I experienced as a Christian was not like what I read about in the Bible – no divine intervention, no guidance, no interaction. Just talking to a brick wall. Deism seems at least possible, but irrelevant.

      • Rayan Zehn says:

        Absolutely!! I think you hit the nail on the head. That was precisely my deism argument. Supernatural beings might exist, but it’s unlikely they interact with the natural world, if you accept my previous argument.

        As for what evidence of the supernatural would look like, I don’t think anything would count. Let’s say, hypothetically, I was able to sufficiently demonstrate a “supernatural” ability to use telekinesis in such a way that it was accepted by modern standards of science. While this ability currently defies the laws of physics and what we understand about the human brain, a demonstration of telekinesis would probably eventually be understood in scientific terms (either in a way that supports physics or in a way that rewrites it). In this case, what we think is “supernatural” would eventually be understood as a natural phenomenon.

        Maybe I’m completely wrong though. This idea literally just popped into my head after reading your comment.

        Lastly, I had to chuckle at the “Deism seems at least possible, but irrelevant” comment because it’s so true.

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 )

Facebook photo

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

Connecting to %s