Facebook, Religion, Family, and Drunk Teens: Correlations, Causes, and Policies

We at Terra Nova have been obsessed with avatars drawn from fantasy and science fiction: The Elves and Aliens often found in games. The Facebook page is an avatar too - an idealized creation of self. Facebook avatars represent a much broader swath of the population than game avatars. They're also less restricted, in that you can build them however you want. FB avatars are no less restricted than game avatars in another sense, though: They must make sense within the lore and backstory of the world. In games, it's a fantasy or science lore. With FB, the "backstory" is the context of ordinary daily life. FB avatars are ordinary people living in the 21st century, but they are those lucky people who are always socializing with cool people, always thinking cool thoughts, always going cool places, always doing cool things. Nobody puts up a picture of himself as an exhausted, lonely parent harrying his toddlers into bed. FB avatars are a biased sample of life.

The bias has some known directions, too. For example, FB avatars are biased toward drug consumption. FB avatars are often seen partying. Many is the FB picture of a young man or woman chugging something.

Interesting then that a study shows a large positive correlation among 12-17 year olds between FB time and drinking and smoking and weed. Among teens who use FB at least once a day (70% of the sample), 10 percent smoke cigarettes. Among the other 30%, only 2 percent smoke. Seems significant.

It's not. First, consider the source. The report comes from the National Center for Addiction and Substance Abuse. The chances of such an organization finding no problem is nil. Outfits like that don't explore reality, they hunt for findings that support their world view. (I'm sure they are all wonderful people, very conscientious and concerned. But that's the problem. We want our scientific findings to come from people who don't care what they find.)

Second, buried in the report are some interesting correlations. Eating dinner together as a family is correlated with reduced substance abuse. So is going to a religious service once a week. The effect sizes are of the same order as the purported FB effect. I have seen similar findings for these variables in my own work and in countless other studies of income, crime, out of wedlock birth, you name it. The headline is usually about something else, but family and religiosity almost always have a big and positive effect on the quality of life.

This leads me to suspect that these factors, and not Facebook, are the real cause here. I suspect that kids growing up in chaotic circumstances spend more of their time on the internet than kids who are, you know, doing stuff with their family and going to church youth group. The chaos kids are also more likely to do drugs. So, sure: There's a correlation between FB and drugs. It's there because chaos causes both of them. Eating togther and attending religious services reduces the chaos, reducing both FB time and substance abuse.

You could check this explanation by running a regression of substance abuse on age, sex, and the family factors. If the family situation is causing both drugs and FB, then FB time would have no effect of its own. Very easy to test. (Indeed, so easy that I am suspicious that the authors of this report didn't do it.)

In any case, the authors stuck with the FB correlation and made that the headline. Family and religion effgects? Last page. Now, why is Facebook the headline and not family togetherness and being religious? That's a question worth considering. Is it easier to do something about Facebook than about family cohesion and religiosity? Maybe. People have tried. But you can't regulate media. It's against the constitution, thankfully. Only a couple of months ago, the Supremes routed another video game regulation law.

What about the alternative to blaming media? Maybe its too hard, or too upsetting in some sense, to say something like "youth problems are best addressed by encouraging the formation of stable religious family environments."

Sure it's a hard sentence. But we have to decide. We can keep projecting our cultural problems onto media, or we can actually do something about them. 

Comments

[Andy Havens]

1.

Not sure I agree with you that the FB page is an avatar in any way that connects it to those used in games. My avatar in a game isn't a representation of me; it's a representation of a character I'm playing. Inasmuch as it's a reflection of me playing that particular game... OK. It's connected to me, but it's not meant to *be* or communicate anything about me, necessarily. I'm a nice guy IRL. In WoW, I can play a character that's a total evil ass, do it well, and have friends/co-players respect that I'm doing a good job playing a bad guy; that is, that my avatar represents a construct. The connections that others make between it and me depends on the skill I lavish on the character, not my real self.

Now... on FB, yes... I can choose what aspects of myself to represent. I can come across as a total jock or playboy; or an incoherent rambling drunk; or a pseudo-intellectual wit; or a self-involved, neurotic artist. I can choose to manifest all those things on different days, in different posts, etc. But, in every case, my FB friends will be reacting to "my avatar" as a representation of my real life me. That is, if I post something funny, they might think, "Andy is funnier than I thought." If I post something tacky, they might decide I'm more of a boor. All those reactions accrue to me. My avatar, in FB, is much closer to an ongoing, ever-flexible advertisement for me than are the ones I create in games.

[Kelly]

2.

Not sure it's fair to credit religiosity as a major factor since that quality in a family may simply be an expression of a more fundamental set of values that often, but not always, leads a family unit to increased religious participation. I think a solid moral foundation in a family is often lumped in with religious habits, but for an agnostic or atheist, it may be easier to see the moral core and related sense of responsibilities as separate from religion. Religion may just be the context for moral teaching within a family, but I think you can easily have a religiously-active family full of hypocrisy and chaos as you can have a religious family that is sincere about values (and has kids whose behavior reflects that environment).

The values seem more important to me than the religious participation, unless the implications are more literal (that a kid at a church youth function is physically unable to be using drugs, drinking, or smoking simply due to supervision and social expectations at those particular hangouts). If it's a deeper observation/trend to values and attitudes, I would think the religious dressing would not matter as much as the sincere moral convictions that might be just as common in non-religious families that demonstrate strong values and a moral consistency to kids.

Anyway, it's interesting to wonder if the simplified and often forced consistency of in-game morality can HELP kids who live in morally ambiguous or chaotic families by giving them a substitute world of rules and values that their home life lacks. Could learning to be a valuable contributing member in a WOW guild make up for the family and community activities you'd get in a stable religiously-active home? Probably not, but given enough time and mental investment in an online world, I don't think it's totally out of the realm of possibilities.

[ecastronova]

3.

Kelly, you are making good points. I think the reason you don't see
studies of spirituality as opposed to religiosity involve the basic
problems of observation. We can't really observe whether a family is
spiritual. Nor can we observe whether a family is providing a solid
moral background. We can only observe concrete behavior, like,
actually getting out of your house and going to a religious service.
Or, actually setting up your schedule so that you sit down once a
day with everyone else in your family. Those things are concrete.
Statistics can only tell us about measurable, observed behavior.

Perhaps the measures are wrong. I was assuming that, while eating
dinner with your family is not a perfect indicator of family
cohesion, it is a *valid* indicator. That is, a family that eats
together is rather more likely to be a strong unit than one that
doesn't. I think that's true, as a general rule. There will be
exceptions, of course, but in social science (unlike physics and
math), an exception does not invalidate a rule. We should not
compare a rule to its exceptions. Rather, in social science, you're
supposed to compare a rule to an opposite rule. So perhaps we don't
really believe that eating together is an indicator of family
cohesion. If not - if we believe the general rule is "families that
eat dinner together are no more or less likely to be strong," then
what explains the findings in this and other papers? What is it
about eating dinner together (if not cohesion and all that
good-family goodness) that has all these positive life outcomes for
kids?

Similarly - as a general rule, can we agree that religious families
are more likely to be spiritual (in the way you describe) than
non-religious families? If not, if instead we believe that religious
families are no more likely to be spiritual in a good way than
anyone else, then what explains the positive relationship between
going to church and kids' life outcomes? If as a general rule,
everyone in church, mosque, or synagogue is sort of a moral
hypocrite, no more or less wholesome or grounded or self-aware or
mature or wholesomely spiritual than anyone else, then what is it
about religiosity in families (as measured by the act of going to
services) that keeps their kids from doing drugs, dropping out,
getting pregnant, and going to jail?

These are interesting questions. When we argue that a particular
statistical measure does not capture something well, this only
raises the question of what it *is* capturing. If eating with your
family and going to services are rather false measures of what they
purport to measure (family cohesion and spirituality), what do they
measure? Why are they correlated with positive life outcomes of
kids?

[D506]

4.

I find it odd. You attribute the negative correlation between FB and drug use as simply being a product of the same cause (chaos), rather than FB being the cause of drug use in itself. The odd part is that you somehow don't make the same conclusion over the positive correlation between drug use and religiosity. The consistent approach would be to assume that these, too, are simply a product of the same cause (stability, in this case), rather than religiosity being the cause of decreased drug use by itself.

In other words, I would argue that not only do families that do things together and spend time together tend to shy away from Facebook and drug use, they also tend to go to religious services weekly.

The great part is that we can simplify this:

"youth problems are best addressed by encouraging the formation of stable religious family environments."

to simply this:
"youth problems are best addressed by encouraging the formation of stable family environments."

Which has the benefit of consistent application across cultural, religious and non-religious individuals.

[ecastronova]

5.

D506, I think this is also a perfectly reasonable possibility as
well. It may well be that the family is causing everything. Again,
this would be so easy to test. Regress:

Y = a*Dinnertime + b*religion + c*FB

The authors of the report claim a = b = 0, c > 0
I believe a < 0, b < 0, c = 0
D506 asserts a < 0, b = c = 0

[pwnem]

6.

Intressting read, thank you.

~http://pwnem.blogg.se/

[ecastronova]

7.

I wrote to the authors of the study and asked for the data. No response yet. If I get the data, I'll do the stats and report them here.

[ecastronova]

8.

"Dear Professor Castronova,

Thank you for your interest in our work. Unfortunately our policy is to not share this data. If you have additional questions regarding the survey, please let us know.

Sincerely,

Lynn Galligan
Associate Office Administrator
The National Center on Addiction and
Substance Abuse at Columbia University"

[Isaac Knowles]

9.

I'm not surprised. They did a pretty bad job of hiding their bias against social networks in the report summary, I think. I doubt they want anyone to screw with their conclusions. I wonder who funds their research.

[ecastronova]

10.

For closure, I looked into the latest religiosity studies. A good site is here:

http://www.youthandreligion.org/news/preliminary.html

These guys are funded by Lilly, which I assume is neither pro- or anti-religion. They are tenured sociologists, and therefore probably have less of a dog in the fight. (A little dog, maybe. A chihuahua). The theme of the findings is basically that religion is a significant causal factor in good outcomes for families and kids.

What I love about this stuff is the paradox it highlights. If you're a logical person who wants good outcomes for your kids, you should fall in love with your spouse and with God every day. But those commitments are not derivable using pure logic, are they?

[D506]

11.

You went to youthandreligion.org for a non bias view on the effects of religion? I'm not calling BS, but I took a quick look through and the results are largely the same: they claim religious attendance correlates to a healthier life style in a number of areas but I would still argue both religious attendance and say, community service, have the same root cause: a stable family environment.

I didn't poke too much, but all the staff listed on the research team are from the University of Notre Dame (a catholic university) so I very much doubt they could be claimed as "neither pro- or anti-religion".

[ecastronova]

12.

D506, I stand by my assessment: These results do not come from
pro-religious bias. The methods are sound (far sounder than the
Columbia report). Catholic universities in the US have a long
tradition of unbiased teaching and research activity. (I attended
Georgetown as a committed atheist and felt no pressure whatever to
convert or think differently. I converted only in my early 30's.)
Moreover, research from secular institutions has the same findings,
but (as I said earlier) they tend to bury them. If there's bias, and
I think there is, it's in the unconscious tendency of scholars
post-1960 to assume that class, income, media structures, and
identity are the important things to look at, whereas the obvious
things, the things that are simply much closer to the universal
roots of the human tribal living and cultural experience, family and
worship, get overlooked.

I encourage you to poke deeper: I am confident that if you keep your
mind open,  you'll find a long line of research that supports the
findings that are nicely summarized at that site.

[anon]

13.

Really? It is clearly a pro-religion site. Seriously, every single one of their conclusions are pro-religion. Not necessarily untrue, of course, but they are definitely biased.

Please present the results from an _unbiased_ study, not one whose goal is clearly to promote religion.

[ecastronova]

14.

The definition of unbiased is to present the truth, which you admit the site does. Can you point to a pro-religion conclusion on the site that is not true? 

O never mind. I confess! I am pro-religion. I think everyone should attend regular public worship and attempt to open their minds and spirits to formation by great religious teachers. Everyone should believe in God. 

There, I said it. So kill me. 

Mobile