So, I was perusing a fairly normal looking derp-fest found on Facebook when I came across this image:
It’s interesting data. Only 30% of atheists have stayed as atheists? Indeed, it was so shocking that said derp-fest of a Facebook group had to caption it with “ha! Where are all you atheists going!” or some such bollocks. Anyway, it wasn’t so much this image that made me take note, as the reaction of some atheists to it. Because, of course, you know, like, it makes Glorious Atheism look bad. Real bad. It’s a big “whoopsie” for Glorious Atheism. So it must be wrong somehow.
So, first port of call in this knee-jerk reaction was to raise issues about the source. The Pew Forum is a religious group! Religious groups lie! Well… there’s really nothing about the Pew Research Center that suggests that it is religiously affiliated, financed or controlled. At least not in a significant way. And even if it was, that alone wouldn’t say anything bad about it. Sure, it can suggest bias, but even those with affiliations can take steps to eradicate or limit the effect bias has on their research – in fact, I’m far more weary of findings coming from explicitly atheist/non-religion groups like the British Humanist Association, who tend to release questionably-acquired survey results all the time. Anyone in any research field will have a bias of one kind or another; a scientist will always be biased towards wanting their experiment to work (we’ll hit things pretty hard before giving up), or a social scientist will have political views (because it’s impossible to be truly neutral on everything), or more generally people will always have a pet theory they want to be correct. There’s nothing wrong with this. The trick is dropping your pet idea if the data says otherwise and making sure you account for real, tangible biases such as selection and confirmation bias. It helps to be neutral, but by no means is it essential. Skepticism and critical thinking is designed to overcome this.
Reacting in a way that assumes an affiliation (that may or may not exist) means the conclusion is necessarily wrong is just a knee-jerk, irrational response. You can make great arguments for it if you’ve done some casual reading on biases and can repeat skeptical mantras but the proof is always in the data and the methodology. Even in the case of medical research trials carried out by a drugs company, the proof is in how the data is collected. To assume otherwise isn’t skepticism, it’s cynicism; and as much of a cynic as I can be, I’d rather not just throw out an idea based on the fact I don’t like its source.
But that’s just a meta-discussion on data acquisition, there’s a much more convoluted problem specific to the question: namely how you go about finding a “retention” rate for atheism.
Being a position of non-belief, a default position, and a null hypothesis with respect to religion, “defining” atheism like this is very difficult. This is why research groups like the Pew Forum split their demographics up into a wide array of different labels. Asking “are you an atheist” and “are you non-religious” is likely to produce different answers even if they are, in fact, the same thing. So, you split it up into “unaffiliated”, and then “religious unaffiliated” and “secular unaffiliated”, then further categories, to try and get a good view of what people actually think. When you do it like this, it turns out that just short of 20% of the US population are effectively non-religious, even though self-defined “atheists” are marginally less numerous than Jews. This is just trying to figure out if people “are” atheists, though. Figuring out how they were raised brings up a whole host of other issues.
This comes back to what I said above: atheism is a position of non-belief, a default position, and a null hypothesis with respect to religion. This makes it quite difficult to really assess if someone was even “raised” as an atheist. Sure, there may be some sad individuals out there who raise their kids on Dawkins and send them to Camp Quest but overall they’re a minority. I’d like to see the “retention” rates for those individuals, but I don’t think we have that data just yet since that sort of “hardcore” non-belief is a relatively recent thing. There’s no litmus test to say that you were “raised” atheist; you could simply not have been raised in a religion, but that says nothing particularly useful. One could easily say I was “raised” atheist because I was never sent to church, but the reality is somewhat convoluted. On the other side of the question, it’s quite easy to say someone is “raised” in a religion – their parents took them to church, they went to church, they outright believed. Raised atheist? Not so easy. What if they were raised in a fairly non-religious manner, but in a non-preachy way? What if their parents were religious but simply didn’t enforce religion in the household? What if parents were atheists but didn’t object to their child going to church as a pre-teen because it wasn’t objectionable? If the question is “what was your parents’ religion”, then what if they never really mentioned it and you went through childhood not knowing? A conversion to a more explicit religion later in life could make it seem like these fairly mild conditions were akin to being actively raised as an atheist, but that would just be in comparison to a later conversion. The reality and what we infer from “being raised atheist” may be very different.
This isn’t to play No True Atheist with the thought of people “de-converting” from non-religion, but it does illustrate some caveats that need to be remembered when considering whether or not someone was actually raised to be non-religious.
Anyway, all this gumph is interesting I’m sure, but the punchline is this:
You can actually type the words “pew forum religious retention” into everyone’s favourite love-or-hate-it web based super-corp, Google. The fact that I didn’t see anyone who was arguing against the chart above being true made me roll my eyes.
Doing this doesn’t take long. It’s easy. And in doing so we can find the actual results from the Pew Forum on Religion and Public Life and it’s survey involving religious retention. It can be found here. The actual report produced by Pew doesn’t include atheist retention or make much of a deal of it, probably for the reasons outlined above and the fact that the explicitly “atheist” contingent is so small (in contrast to the other demographics surveyed) that retention isn’t a significant factor. The survey results that include atheism and make a big point of it come from a different study entirely that merely used Pew’s raw data. That in itself could be a bit of a fudge, for reasons of data mining and so on, but it certainly means that the source attributed in the image is technically incorrect and misleading.
Further from that raw data and Pew’s report of it, we can see a lot of different things. Each vaguely interesting. A very significant number of people shift religion in one way or another, just short of half overall, in fact. We also see that non-Christian religions have a significantly higher retention than others – probably due to ethnic-religious identity (social scientists might want to correct me on that) and the fact that switching between one YHWH-centred religion and another doesn’t require much of a shake-up in your thoughts. We can also see that, in Pew’s words that people “…moving into the unaffiliated category outnumber those moving out of the unaffiliated group by more than a three-to-one margin”. This in itself is telling because it demonstrates an increase in the unaffilliated categories the religious-to-unaffiliated transition is more easily defined than the converse, where identifying a atheist upbringing is difficult and prone to bias.
So the punchline is that the data is fairly sound, with a “but”. And it’s not too difficult to dig into it and find out if it’s any good, which it is. Providing you understand the “but” properly.
Is any of this even something to be actively worried about as a non-believer? No, not at all. Why would it?