The New Zealand Initiative’s Head of Research hopes a story on Radio New Zealand about youth drinking serves as a sharp wake-up call about the quality of reporting on drinking in New Zealand, and about Alcohol Healthwatch.
Radio New Zealand reported on a new report recycling some older stats from the 2014/15 Attitudes and Behaviour towards Alcohol Survey. Alcohol Healthwatch was reported as hoping the survey serves as a “sharp wake-up call.”
There are a few big problems in this story:
- Without a time trend, there’s no anchor for the numbers reported. Are they big or small? Has it dropped or risen? A rising number might imply more of a policy problem than a falling number, for the same number. The wake-up call framing makes it sound like there’s a rising number rather than a falling one. I blame RNZ on this one.
- The way the stats are presented is misleading. That’s part on RNZ, part on just terribly opaque writing by the Health Promotion Agency.
- Alcohol Healthwatch has to know the numbers forming the basis for this story represent a drop from the prior figures, but they still use it as a hook for pushing the thing they always push – a march towards prohibition. Maybe they told the reporter and the reporter didn’t use it, maybe they didn’t.
I’ll quote from the story now:
About 27 percent of the teens questioned for the survey, carried out for the Health Promotion Agency, said they had at least eight drinks the last time they consumed more than two drinks of alcohol.
More than 50 percent said they had had five or more drinks.
Alcohol Healthwatch director Rebecca Williams said more needed to be done to reduce the availability and cost of liquor, and the marketing of alcohol.
It would be very simple for anyone reading that story quickly to think that more than 50% of kids had five or more drinks and that 27% drank eight or more drinks. But that isn’t it at all. It’s a percentage of the kids who had consumed two or more drinks.
What proportion of the surveyed kids drink at all, or drank two or more drinks sometime in the past 3 months, then matters. And that’s much harder to figure out from the HPA publications than it really should be – so we can forgive RNZ a bit for screwing this part up.
If you want to walk through that part, hit the Appendix bit at the end of this post. Here’s the bottom line on this year’s stats, for clarity:
- 60% did not drink in the past year;
- 3% did not drink more than a single drink per session in the past year;
- 6% drank two or more drinks in a session, but more than three months ago (and we don’t know how much they then had);
- 14% had their last drinking session of two or more drinks in the past three months, and then consumed 2 – 4 drinks;
- 9% had their last drinking session of two or more drinks in the past three months, and then consumed 5 – 7 drinks;
- 8% had their last drinking session of two or more drinks in the past three months, and then consumed 8 or more drinks.
So it isn’t 27% who consumed 8 or more drinks, it’s 8%. And the proportion who consumed more than 5 drinks isn’t over 50%. It isn’t even over 20%. It’s 17%.
The story reads out of something out of a Joel Best book. There’s more work being done by “the last time they consumed more than two drinks of alcohol” than would be obvious to any reader. On the one side, it invites the reader to ignore all the folks who consumed no drinks. On the other, do remember that this is a measure of “most recent session in which two or more drinks were consumed”, not a measure of “any time during the past three months.” If you had two beers yesterday and a bottle of vodka two weekends ago, the survey would count the two-beer session.
Ok. Now that we have that part straight, is it a rising or a falling trend? I hit the numbers on that one over at Offsetting. I had to jump around a bit in the HPA reports – it is just so frustrating that they don’t simply put up a spreadsheet with the by-respondent survey answers, with identifying bits munged out. But here’s the upshot:
- In 2016, 60% of kids aged 15-17 reported being non-drinkers. In 2013, it was 48%;
- In 2016, 29% reported having had a drink in the past month; in 2013, 38% reported having had a drink in the past 4 weeks.
- In 2016, 17% had, within the past three months, consumed 5 or more drinks in their last session. In 2013, 22% of kids had had five or more drinks at least once in the past four weeks. It’s hard to compare those numbers directly: the 2016 numbers would miss your bender 8 weeks ago if you had two drinks last night, but the 2013 numbers wouldn’t notice anything from 8 weeks ago.
Overall it looks like a drop to me: the proportion of non-drinkers is definitely up, the proportion having had a drink in the past month is down, and it looks like the proportion having had 5 or more drinks at a session is down – but that one is harder to tell because of differences in how they set the thing up. If you think kids drink roughly the same amount every time they go out drinking but drink irregularly, then this is likely a fall. If you think that kids have a beer with their folks at dinner during the week and then go out on a bender on the weekend, then it could have been an increase. It’s hard to say from these numbers – so we have to go find some others.
Fortunately, we know from the New Zealand Health Survey that there has been a substantial drop in heavy drinking by teens aged 15-17. In 2006/7, 25% of kids in that cohort consumed six or more drinks in a session at least once per month. In 2013/14, it was 12.5%. In 2014/15, it was 10.7% – less than half of the 2006/7 figure.
But there’s none of that context in the Radio NZ piece. Instead, we have a wake-up call about the need to ban alcohol advertising.
The September paper forming the basis for the story does a terrible job in explaining the method. It first says:
Participants in the 2014/15 Attitudes and Behaviour towards Alcohol Survey (ABAS 2014/15) were asked about the last occasion they consumed two or more alcoholic drinks. Respondents were restricted to those who did so within the last three months, of which there were 193 respondents 15 to 17-years old.
So it sounds like there are 193 who had consumed two or more drinks in the last three months, and an unmentioned number not having consumed that much alcohol over that period, if any.
But then it says:
Respondents who had consumed an alcoholic drink in the past year were asked “How long ago did you have two or more drinks of alcohol on any one occasion?” (n=77). Nearly eight out of ten (79%, 69-87%) reported consuming two or more alcoholic drinks on an occasion within the past three months, 14% (7-23%) more than three months ago and the remaining 7% (3-15%) reported they had not consumed more than one alcoholic drink on an occasion.
If the sample of 193 were drawn from the group who had consumed two or more drinks within the last month, none of the above numbers make sense. It then says that the 60 respondents who had consumed two or more drinks on an occasion within the last three months were asked what they drank on that occasion.
So it looks like the 193 is likely the population of 15-17 year olds surveyed, 77 of that 193 had had at least one drink in the past year, and 60 of that 77 had had at least two drinks in the past 3 months. Of those 60 who consumed two or more drinks in the past 3 months, 45% consumed 2 to 4 drinks on their last drinking occasion (27 respondents), 28% (17 respondents) consumed five to seven drinks, and 27% (16 respondents) consumed eight or more drinks.
In the March report on the same survey, they described it this way:
The 2014/15 Attitudes and Behaviour towards Alcohol Survey (2014/15 ABAS) included questions about their consumption of alcohol and their experiences related to drinking alcohol. There were 193 respondents aged 15 to 17 years.
…Respondents were asked “did you drink any alcohol in the last four weeks?” (n=193). Almost one in three (29%, 23-35%) said yes, one in eight (13%, 9-18%) said they were drinkers but had not consumed alcohol in the past month and the remaining 59% (52-65%) said they were non-drinkers.
And the report on non-drinkers says that 60% of those aged 15-17 were non-drinkers.