Tuesday, June 26, 2007

Family Dinners and Adolescent Drug Use?

So you’ve probably heard about the guy who managed to sit down for dinner with his wife and children everyday for a year. Then just happened to write a book. That just happened to hit the bookstores in time for Father’s Day - Dinner with Dad: How I Found My Way Back to the Family Table. Okay, this is where some of us, me for one, sigh/roll our eyes/grimace or otherwise display some mixture of discomfort and possible disbelief. Maybe because we don’t have all members of the family at the table most nights of the week. Or we don’t cook so much. Or we don’t have a dad in the house. Or a mom. Or kids who will eat much beyond carbs. But let’s move on.

So, Stracher, the media’s newbie expert on fatherhood and parenting referenced that two-year old study linking family dinners with reduced substance abuse in a recent New York Times editorial. The 2005 report by Colulmbia University’s National Center on Addiction and Substance Abuse (NCASA). The one reporting “the more often teens have dinner with their families, the less likely they are to smoke, drink, or use drugs.” Remember the media blitz? The public service ads featuring NCASA board member Jamie Lee Curtis?

So family dinners reduce substance abuse. Right? Is this true? You know here’s where we take a closer look at the research. We have a phone survey with about 1,000 twelve to seventeen year olds (and 829 of their parents). Sounds impressive. But wait. Researchers started with a potential pool of about 24,000 households. A third either didn’t answer or had a language issue preventing participation. About 9,000 kids who answered gave no information, in other words, didn’t want to participate and about 1,000 interviews were excluded because they were cut short, some by parents not wanting their children to participate.

Not a random sample by any stretch of the imagination. Limited to the 1,000 or so households with teens who happened to be at home where someone answered the phone, wanted to participate and had the permission to do so by BOTH parents. Yes, both parents had to give consent. So of the 24,000 households with teens – only 4% participated. That’s a very, very low response rate. Any survey researcher would be extremely reluctant to make bold conclusions about a population based on such a paltry sample. Do you think this sample represents most households? It underrepresented Black and Latino household for sure.

How about asking teens, with both parents home, possibly within earshot, about smoking, drinking, worse yet, doing drugs? Can you imagine copping to any? Course not. So now we have to worry about kids minimizing their substance use. As did the researchers. So they used mostly indirect questions to assess their “risk score”. You can imagine the kind of questions. But hey, let’s look at some of the actual questions.

Do you know a friend or classmate who uses acid, cocaine, or heroin?
How many of your friends drink alcoholic beverages?
How likely is it that you will try illegal drugs in the future?

Ahhh. That awkward language. Ever any yes to any question including the word “illegal?’ Doesn’t take a degree in psychology or statistics to realize this measure underestimates “risk” not to mention actual drug/cigarette/alcohol use. Underestimating may not be such a large problem. But we’re not even sure if they’ve measured the participant’s behavior, their intentions or their friends. Does it matter? Maybe. We just don’t know.

So did teens having more family dinners report less substance abuse? The authors seem to think so:

“The number of family dinners a teen has in a typical week is a powerful indicator of substance abuse risk. The average risk score of teens having dinner with their families five to seven nights in a typical week is half that of teens having dinner as a family two nights of less per week.”

Sounds definitive. Powerful? Half the risk? Yeah, right. First, let’s look at the risk score. The researchers made the risk scores based on an average of 1.00. Kids who have never tried cigarettes, alcohol or pot are 0.37. So, even kids who have never tried a thing still have some risk. Kind of odd. But let’s move on. Teens who admit to using one of the three are 1.14. Tried all three? Risk jumps to 3.09. So, kids who eat 2 or fewer meals with family have a risk score of 01.49 compared to the kids eating 5 or more – .79. The authors never say in the report, but I assume this is a statistically significant finding – a significant correlation. We’ll assume so. And because we’ve gone over this so often here before, I’ll just say briefly – correlation is not the same as causation.

Eating dinner with family makes kids less likely to take a walk on the wild side? Maybe but I can think of another factor that might be correlated with drug and alcohol use - and with eating at home. How about age. Being older? Older teens are much less likely to eat dinner at home. And much more likely to sample illegal substances. The survey data clearly shows these two relationships. So can age explain both phenomenon – risky behavior and family dinners?

The authors bring this up in the final paragraph of the report. The final paragraph. In a very round-about manner they suggest they have tested to make sure age isn’t a confounding variable in their analyses. That after controlling for age family dinners was still significantly related to drinking, smoking, and using drugs. What does this mean? Controlling for age? Let’s take 15 year-olds. This means that among 15 year-olds, those eating 5 or more FD were less at risk than those eating 2 or less. Fine. But when Carl Bialik, The Wall Street Numbers Guy, asked the authors to look at age again – he reported that age was a stronger predictor of substance use than family dinners. So it’s hard to know what’s going on with these findings.

As for the powerfulness of the finding – it doesn’t look so good. According to the Numbers Guy, age is more powerful. Although I can’t find the numbers myself, he also says that watching R-rated movies and their friend’s sexual activities were also more strongly related to drugs and such than eating din-din with the ‘rents.

Do you sense I’m a little irritated with this report? It shows a lot of bar graphs, easy to read. Very user-friendly. But lacking lots of numbers. Like the average risk by number of family dinners. The key results show numbers for teen eating 5 or more dinners, and those eating less than 2. We don’t see any results for those eating between 3 and 4. Would be nice to put the results into perspective. I’m always suspicious that those omitted numbers might somehow, well, compromise the results. Can’t help but think the authors have tidied up the numbers into a pretty little package. Rather, a simplistic media-ready package for TV Land and Nick at Night. Not a joke. They sponsored the study. Check out the press release.

You should know the study was written up by people working on the behalf of the NCASA. It didn’t appear in a peer-reviewed journal. Means it probably wasn’t scrutinized by other experts in the field (i.e. not affiliated with the NCASA). At least not before it was published.


Do I think that eating side by side with loved or at least related folk reduces drug use? Probably not. Not the actual sitting down and eating together. That we do sit down together probably marks other things - like family cohesiveness, maybe parental involvement. But there are other ways of measuring it besides how often a family dines together. And those who don't dine together can probably be as cohesive. And warm. And non-drug inducing. Do I doubt that family dinners provide some benefits. No. Surely there are all sorts of tangible and intangible benefits. But I'm not the expert of the moment...

Now Cameron Stracher, the Family Dinner guy? What does he have to say? Did his presence have any effect on them? Ah, no. Not really. Well, maybe.

“The real question is not whether Dad matters – he may or may not – but what has Dad lost? In missing family dinner, we fathers are missing a large portion of our children’s lives…They may recover just fine, but we will not.”

So the Family Dinner Demi-God doesn’t even know if his presence mattered to his offspring? Okay so I haven’t read the book. Yet. But could someone who has recommend it?

The Bottom Line: Not entirely convincing look at family dinners and drug use. Extremely small sample. Insufficient analyses. Puffed up conclusions. User-friendly graphs. Little substantive data.


Read the larger version of the report titled "National Survey of American Attitudes on Substance Abuse X: Teens and Parents" by The National Center on Addiction and Substance Abuse at Columbia University.

Or the briefer version, The Importance of Family Dinners II. Both are free even though it looks like you might have to pay. Ignore the dollar signs. Hit the hot pink "Download for free" button.

No comments: