...is the subject of a forthcoming study on which I was lucky enough to be the third author (of three total :P).

An (almost) final draft is here: https://s3.amazonaws.com/wellnamed/e...t-analysis.pdf, we might still make a few tweaks to the discussion section.

Anyway, I thought might be of interest to people here. For me, the most interesting part of this is thinking about what it means to form a "secular" identity, particularly in the context of trends towards desecularization, and thinking about the traditional importance of religious identity in creating social solidarity. Our data is sort of evocative on this topic but doesn't really allow for any complete or general theory. The strongest hint of this process of identity formation is found in that shared beliefs about science and rationality play in people's experiences of deconversion and how they think about what it means to be an "ex-Christian" in our data.

I think our literature review is interesting enough to be worth the time on its own. On identity formation and the internet some of the digital religion literature is intriguing to me (Anderson 1999; Ammerman 2003; Lövheim 2013), probably because I've spent my whole adult life in online communities. The best book on secular identity and deconversion in general is probably Zuckerman's The Non-Religious. On trends towards secularization in Europe and the US Voas (2009) and Brauer (2018) are worthwhile, although our data doesn't fit into that theory at all.

In any case, I thought it would be fun to post this here just because over the years my thinking on a number of issues has been influenced by interesting conversations in this forum, particularly involving regs like Original Position, tame_deuces, uke_master, Aaron W, zumby, and others. My process doesn't look much like the people discussed in the study, probably because I was never attached to a particularly conservative version of Christianity, but the topic is clearly relevant to me personally nevertheless.

All comments and criticisms are welcome. Bear in mind I'm going to just take credit for anything you like and blame the other authors for anything that needs more work :P

Thanks for sharing this. It was a very interesting read.

On the topic of "the social process" (p. 26), by any chance did you happen to come across geographical data on the deconversion stories? I agree with the general concept that community would play a role in the process of identity formation, and that the absence of a "physical" community would create a longing that an internet community could fulfill. From the various statements that are in the article (especially the observation that the most active participants in the forum are ex-fundamentalist), I would suspect that there's overrepresentation from deep red states and underrepresentation from the more cosmopolitan coasts.

Also, are you aware of any research on adult religious conversion? I suspect that the similar triad of intellectual/emotional/social would be present as this is a process of identity formation.

Much like 2+2, the forum allows participants to specify a free-form "location" field in user profiles. About 40% of the posters have that data, and I did scrape it, but I didn't finish normalizing it (e.g. there are plenty of entries like "Midwest USA") because we didn't think we could draw significant conclusions from regional differences within the dataset. For the ones I did process it's about 70% US and the rest pretty evenly split between Australia, Canada, UK, and a couple European countries.

That wouldn't surprise me. We were basically happy enough just to note that people from conservative denominations were strongly represented. There's definitely people who put their location as like "Bible Belt, USA".

I agree, although I haven't read much on adult conversion specifically. I know that the literature exists, although some of it might be older than the deconversion literature. Brinkerhoff and Mackie (1993) use a typology that distinguishes between adult converts and people who stay in the religion they were raised in. IIRC they are using survey data to find differences in attitudes between people following different "religious careers". I don't remember off the top of my head and I don't have the book here but the Streib (2009) book on deconversion also compares their psychological profiles of deconverts to the religious. They might also compare adult converts. It's an interesting book in any case.

The Digital Religion book is internet-specific but the chapters on community and identity are relevant. I think reading those two chapters in particular is what made me want to focus on the idea of identity.

Thanks for the link, it was a good read and congrats on being an author.

I do not know enough to comment on the presented theory. I liked the dataset, it was compelling and felt substantial. A killer combination.

A few minor criticisms, I did them in spoiler tags if you don't want them or if it is too late anyway.



These are minor quibbles, it was a good article. It felt like it dared to handle and discuss the issue. It was also a fairly straightforward read, which in social science is a very good sign of a solid piece.

Thanks. Thanks for the proofreading too.

Yep. That's a typo. Nice.

We are saying that other researchers have focused on those things, and offering Vargas as an example. I expect it's probably not too confusing to the intended audience because it's so typical for articles to begin with literature review in this style.

Yeah, I think that's a good point. My opinion is that qualitative research (content analysis, in-depth interviews, ethnography) is a good complement to the more representative quantitative surveys that are often used (think of Pew's religious trends series, or the "fuzzy middle" theory). The problem with qualitative content analysis like ours is that it's very difficult to generalize: our sample is probably sui generis in a lot of ways. But the problem with the quantitative stuff is that it filters out a lot of interesting details about what's really happening to create the larger statistically significant trends. We've talked a little about trying to make that point in the conclusion, maybe we ought to try to do it.

Thanks! I feel like it worked out very well. A lot of credit actually needs to go to our reviewers. The version linked here is quite a bit different from the initial draft. They gave us a lot of good advice on additional literature to review and helped us find the "digital religion" theoretical hook. From what I gather from people who do this for a living peer review can be a little hit or miss, but for us I think it was really valuable.

Coincidentally, this Atlantic piece seems like a pretty good example of how qualitative and quantitative methods complement each other: https://www.theatlantic.com/entertai...e-jobs/582175/

This is is the kind of stuff that makes me love sociology

Do you have plans to continue pursuing these questions (the "more research is needed" questions), or are you looking to shift towards something new?

I'm not really sure. We live in a really rural area and have tended to look for projects that are relatively small in scope, don't require any grant money, and are easy to carry out here. Hence internet content analysis. We're pretty opportunistic: this paper happened mostly because the lead author was already a participant in this forum and it just seemed obvious that it would be a great data set. I'm not sure what else we'd do on this topic given those constraints at the moment, but if we found something I think we'd pursue it because it's nice to have already done so much background :P

The next thing queued up is sort of similar but also somewhat different, involving a survey and interviews with people who identify as pagan about what it means to be pagan in a society in which that is a very marginal religious identity that many don't see as legitimate. There's some similarity in that many of the people interviewed and surveyed are former Christians, so maybe we could think of it as a different way of exiting Christianity, in comparison to the secular exits from this study. I'm not sure; we haven't gotten very far with this project yet so it hasn't really taken shape.

Yeah, combination studies are always neat and can really map a subject well.

Sadly a lot of the quantitative research is enamored with highly technical analyses, and those don't always combine to well with qualitative work. Chiefly because very few people fully grasp what they are actually saying when employing them and as such wouldn't have a clue how to compare them to the qualitative work, and secondly because they are often designed to fit the data (which really goes against scientific ethics, but it happens a lot) and qualitative work would make the research look worse.

Truthfully simple analyses combined with qualitative research or experiment would likely have served most subjects a lot better. I have no idea why people think it is a good idea to use 90% of the effort on getting your analyses as complex as possible, when the chief obstacle of quantitative research in social science is to have solid variables.

Sorry for the rant, but we stumbled into one of my pet peeves.

I love a good methods rant

So, last night we were drinking champagne and celebrating shipping it out and we started talking about the next thing, and I mentioned the idea of trying to find something new to do with this. I mentioned finishing coding the geographic data. Then, somewhere in the second bottle we had the much better idea that we ought to look at gender differences. After all, we already know that women describe themselves as more religious than men on average, and men are more likely to identify as atheists. So it would be interesting to see if there are any differences in the extimonies.

Since I already have the datamined data in a form that lets me do comparisons, I ran some queries today to see if it looked suggestive. One thing I did is construct lists of words that are most associated with a particular gender, i.e. by comparing the number of posts by women in which a word appears to the number of posts by men in which that same word appears.You get something like this

Words favored by women:


Words favored by men:


We also have a database that associates specific emotional categories to sets of words, and comparing the emotionality of posts by gender sort of confirms what pops out in those images, i.e. if I look at the most commonly referenced emotional categories across the whole data set each category tends to see more usage by women than men:



Of course it's quite likely that some of these differences just reflect differences in how men and women communicate in general (and the various social factors influencing that), rather than anything specific to deconversion. But considering our analytical distinction between the intellectual, social, and emotional processes of deconversion it's also potentially interesting if men put a bit more emphasis on the intellectual process in their narratives ("evidence", "science", "reality", "contradiction", "rational"), and the women a bit more on the social/emotional side (all the emotion words + "mom", "child", "baby", "home", "friend"...).

So I think we'll probably pursue this. The current plan is to select a random sample of extimonies from both genders for more qualitative analysis and see what that looks like.

That's certainly some interesting data.

There's clearly a strong connection with what's happening with one's spouse, though I don't know if that's "social" or "emotional." It makes a lot of sense, but it would be interesting to see in what ways that role is positive or negative. Given the word choices for women being more fear-based, there could well be patterns of abuse or manipulation on that side. I'm not sure what I would guess the men are thinking about their wives, but my gut reflex is that a lot of them probably got to a point where they only went to church because of their wives. Maybe it's just all negative?

(And apparently women think more of their boyfriend in their extimony than men think of their girlfriend... Not quite sure what to say about that one...)

Anecdotally, it's rare that dads go to church without their wives and children, and it's more common that wives and children will often go to church without dad. This kind of fits in with the speculation of men and their wives above.

I think the time-oriented wording ("anymore" for women and "eventually" for men) is kind of interesting. The first one sounds more like a sudden giving up ("I can't take it anymore") whereas the other one sounds much more like a gradual shift ("Eventually, I stopped going.") At first glance, that kind of points towards an emotional vs intellectual dichotomy (though second glance says maybe not).

On a more data-science-y front, I suspect that if you replaced the gendered terms (replace husband/wife with spouse, boyfriend/girlfriend with significant other, and maybe things like mother/father with parent), based on what I'm seeing here, it seems you should be able to predict male/female authorship with moderately high levels of accuracy. I'm not sure if there are any other types of authorship studies like that out there.

Anyway, at this point I'm just rambling. I look forward to seeing where this goes.

What exactly is the goal of the paper? I understand what the goals are from an emotional standpoint, but I don't understand what the goal was from an analytical standpoint. It just seems like the bulk of the paper is a list of quoted testimonies.

A much more interesting study would be to find out how many apostates folded into other spiritual beliefs. This is more interesting because it shines light on the glaring need of the human mind to find spiritual meaning.

Also, why does the apostate community feel the need to create a fake word like "extimony" to describe what is clearly just a testimony? The word "testimony" has an etymology behind it. "extimony" is just some concatenation of familiar syllables.

Isn't quantitative analysis the only thing that can make the Social "Sciences" a science?

Really, You got all that from the words in those queries? How did you filter out confirmation bias?

No, but it is a common misconception.

The word "analysis" is also misplaced. An analysis alone would hardly make anything a science.

Nope. I'm combining together general observations about human behavior and relating them to observations about the data presented.

The statements were *explicitly* anecdotal and some of my thoughts were a "gut reflex." I don't know how else you would expect me to acknowledge the possibility of confirmation bias (or other types of intellectual biases).

But since you've posted exactly 3 times all in this thread and none of them actually posting a meaningful criticism, I'm going to relegate you to troll territory, at least for now. Maybe you will prove yourself otherwise in future postings.

You don't really know that much about etymology, do you?

Reading between the lines of your comments a bit, my first thought is to be clear that the goals (and data, and methods, and conclusion...) are modest. You seem likely to criticize them for being too modest, and that's fair enough in my view. This was a project carried out with zero budget by 3 people (really mostly just two) and we always intended it to be small in scope.

That said, the first goal was descriptive. We thought the way that this group of ex-Christians used the website to share their deconversion stories and offer each other support was interesting and we wanted to try to give a thorough description of what was happening there. The bulk of the paper is content analysis of things written on the forum for that reason, and the quotations are intended primarily to be illustrative. So the simplest goal is just that the reader get a good feel for the actual content being analyzed.

The second goal, in my view, is to draw attention to different aspects of the deconversion process evident in the narratives, e.g. that they have an intellectual component but also an interesting social component. One of the things journals and reviewers want in an article is for it to fit somehow into the existing theoretical discussion around a topic.

Along those lines, one point of interest in my view is that a lot of pre-existing discussion about deconversion focuses either on intellectual reasons why people deconvert or the psychology of individuals who deconvert. To put it another way, there's historically been a lot of posts in this forum about whether God exists or not, mostly containing arguments about whether it's intellectually reasonable to believe that God exists. A lot of academic literature also focuses on loss of faith from that angle. I think all of that is interesting, but sociologically it also seems important to us that how/when/why people deconvert involves grappling with social issues that aren't directly intellectual: how do I come out to my friends/family, how do I cope with losing a lot of social ties, and so on. It seems to me that a lot of the intellectual arguments have been available to people for a long time, but we see faster trends towards secularization more recently. It makes sense to me that some of the explanation around that will incorporate facts about social context: it's the idea that some social contexts either inhibit or allow more easily for people to deconvert. In Voas' "fuzzy middle" data the relevant social context is the religiosity of your parents and generational cohort. In our data the social context is an internet forum that provides social support for people who have trouble finding it in their "real" lives.

So, from a theoretical standpoint part of the goal was to situate the descriptive content amongst other existing literature on deconversion, secularization, and religion on the internet.

There is existing research on this topic. Here for example is some Pew survey data. My rough understanding is that the majority of deconverts end up in the "unaffiliated" group, although it's important to remember that this doesn't mean necessarily that they are atheist, only that they don't affiliate with a specific religious group. A relatively smaller number affiliate with other religions. I think the Zuckerman book The Non-religious should also have some good data on this. In any case, I agree that it's an interesting question, it just wasn't possible to investigate with our data, and has been done by others anyway.

I think you're taking the term too seriously. It's mostly just for fun.

Falsifiable conclusions bring it within the realm of science. Statistical analysis exposes falsifiability. Quantitative methods allow statistical analysis to be brought to bear. No matter how many impressive words are flung around qualitative analysis does not meet the measure of science. I'll agree it can be interesting and can provide leads for further research, but science? No.

I have no interest in proving anything to you. "Troll" is a phony word Internet culture invented to allow mob ad hominem while pretending to be righteous. It describes nothing the English language doesn't already succinctly describe.

If you are making a claim back it up, please.

I'll be honest I thought the paper was well done, and I also think your goal of understanding the social context is right on track. It should be obvious though. Social context, perceived or real, have been known to be the most effective mass persuasion mechanism since the birth of the concept. Marketing and propaganda are inexorable.

Observations are not interesting when they agree with intuition.

Thanks.

Inside anti-Christian jokes detract from the paper, unless of course the targeted publishers "dig it" then it just detracts from the publishers.

I've been told that a sociologist is someone who for $10,000 will tell you what your cab driver would have told you for free. I like that description. But I also think that there's some value to trying to document empirical evidence for intuitive ideas. Thanks for saying you thought the paper was well done, I appreciate that.

I don't consider the use of that term to be particularly "anti-Christian", per se. In any case, I do not find it surprising that recent ex-Christian deconverts have a negative attitude towards Christianity. Sometimes people on this forum express negative opinions of Christianity which I disagree with. Nevertheless I think if it's an important aspect of their self-understanding and we would be doing something wrong if we tried too hard to file down the rough edges. So even if you think "extimony" is inherently pejorative I don't find it problematic to reproduce the language, given the descriptive aims of the paper.

I disagree with your (implicit) claim here that qualitative methods are inherently unfalsifiable. For example, for this particular study I could provide a 3rd party with the exact and complete data set we analyzed and a complete description of the methods used (including software), and they could form their own opinions about our conclusions. Or, I think we've given enough information that other researchers could compare our results to data they collect independently, and see if their data supports similar conclusions. I think that amounts to roughly the same level of falsifiability that is typical in quantitative sociological articles.

I do think that there are questions that are not easily answered by qualitative studies alone, particularly questions about representativeness. On the other hand, I think there are questions that are very hard to answer through purely quantitative studies, because you give up a lot of rich data to boil things down into a form that is amenable to statistical analysis. And statistical analysis is subject to its own problems as well, particularly related to how one goes about measuring things. I think the two types of research are complementary, but then this is a debate that is very old in social sciences.

My favorite recent book on social science methods is probably Howard Becker's Evidence, which does a good job criticizing common problems in sociological research, both quantitative and qualitative.