The other day I blogged about the garbage work I used to see in Sociology departments amongst the grad students. ‘High School Social Science Projects’, I called them.
Via Christopher Snowdon, here’s a modern equivalent from some ‘health psychologists’ that’s been reported breathlessly in The Telegraph, The Sun, The Daily Mail, The Scotsman, and on Sky News, with a front page on The Metro (although I can’t find the story on the Meto website anymore).
Snowdon has demolished the paper, as he has done with plenty of previous garbage science papers, but still they come, and still they make the news, so I thought I’d add my own 2 cents.
Is this even a scientific paper at all? Let’s look at the ‘Aims’ first, in the authors’ own words:
To describe the marketing messages stated (in text) or depicted (in image) by retailers and producers for low/er strength wine and beer products sold online by the four main supermarkets in the UK during February–March 2016, and to compare these messages with those used for comparable regular strength products.
So their aim, the aim of a supposedly scientific paper, is to look at some supermarket booze ads, describe them, and compare them to other supermarket booze ads. The only thing lacking is the exercise book in which they can paste the ads they’ve cut out.
But wait – this next bit, ‘sample identification’, sounds scientific:
Searches were conducted of the websites of the four main UK retailers (Tesco, ASDA, Sainsbury’s, and Morrisons) between February and March 2016 to identify webpages marketing low/er strength wines and beers. For each webpage marketing low/er strength alcohol products the webpage of a comparable regular strength alcohol product was identified on the same website. Regular strength alcohol products which are as similar as possible to the identified low/er strength alcohol products were selected for analysis
So, no. ‘Sample identification’ just means that they were able to use their finely-trained skills to locate some booze ads on supermarket websites.
Finally, here’s some science:
A series of Chi squared tests were performed to compare differences in marketing messages of low/er vs. regular strength wines and beers. For all analyses, we considered a 5% Type I error rate at the level of the superordinate theme and type of drink (global), with a Holm-Šídák multiplicity correction which takes into account the dependence between the subordinate themes contained within a superordinate theme per drink
Yes, they carried out some entirely pointless statistical analyses, and tried to make it sound all technical. All to get the result that ads for lower-alcohol booze tended to emphasise messages about health, alcohol content and occasions over taste. Seeing as no-one has ever seen an ad for low-alcohol booze that talks about taste, this is hardly a surprise. (Although you can be sure that had the alcohol industry done this, then the same authors would have been screaming about the alcohol industry trying to get everyone, including kids, drinking more because low-alcohol booze tastes better.)
The authors then say:
Low/er compared to regular strength wines were significantly more likely to be marketed as suitable for consumption on any occasion/everyday [e.g., ‘Perfect option for every taste and occasion’] [X 2 (1, n = 172) = 7.75, p = .005]
However, no evidence is provided that any ad said anything about drinking every day. ‘Every occasion’ does not mean ‘every day.’ It doesn’t mean every lunchtime, and every dinner time. It means every time there’s a social gathering, such as Friday night drinks. (Of course, you could take it to mean ‘every possible time you could drink’, but that’s your business. An ad can’t stop you doing that.)
Then we come to the money quote, the one that the newspapers got excited about:
Presenting low/er strength alcohol products as suitable for consumption on a wider range of occasions than regular strength products suggests they may be being marketed to replace soft drinks rather than alcohol products of regular strength.
That’s it. That’s all there is. There is no actual result here. No evidence at all. It all rests on a ‘suggests’. All it’s saying is ‘It kind of looks like the ads say you can drink low-alcohol drinks on occasions where you wouldn’t drink a full-strength one, so maybe these companies are trying to make people drink this stuff instead of a Coke’.
Seriously, that’s all there is to this paper. A dubious Twitter-level opinion, dressed up with some scientific verbiage to make it look credible, but which does nothing to support the claim being made. (Not that I would give a shit whether alcohol companies were doing this anyway. I’m an adult, I can make my own decisions.) This is a Sociology seminar talking-point, not a scientific result.
They also say:
Furthermore, the marketing messages suggesting extended occasions for low/er strength alcohol consumption may be additional to regular strength alcohol consumption
Note again that this all rests on a ‘suggesting.’ I can make suggestions all day, that doesn’t mean they should be published in supposed scientific journals.
while maintaining or extending recognition of the main brand in question.
Do they expect the brands not to do this in their advertising?
Importantly, although the development and sale of low/er strength alcohol alternatives has been portrayed by the alcohol industry as a way of removing units from the market (thereby reducing alcohol consumption in the population), none of the marketing messages captured on retailers’ and producers’ websites marketing low/er strength alcohol wines and beers mentioned drinking less
Why would they talk about drinking less when they’re selling lower-alcohol drinks? The point is the drinks have less alcohol, so you can drink the same amount but still consume less alcohol.
And it’s not like people simply do what ads tell them. We know that they don’t necessarily do that. The government puts “Just say No to drugs” ads everywhere, but not everyone listens, do they? And anyway, why should the producers of a legal commercial product be forced to make ads saying ‘Buy less of our products’?
or reducing alcohol harms
If you saw an ad that said ‘this lower-alcohol beer will damage your liver less than full-strength beer will’, is that any more likely to make you buy it? Psychologists are, after all, always telling us that negative framing doesn’t work as well as positive framing.
Besides, people are already aware of the harms associated with alcohol. It’s not like an ad that went on about it would be giving us new information. We already consider the harm that alcohol can cause when we weigh up when, what and how much to drink. That’s why I don’t drink every lunchtime and every night.
Low/er strength wines and beers appear to be marketed not as substitutes for higher strength products
No evidence has been shown to support this claim.
but as ones that can be consumed on additional occasions
Again, there’s no evidence for this. The authors talk about low alcohol beer being advertised as perfect for having at picnics, barbeques, beaches, sports events, but they’re assuming that drinkers wouldn’t normally drink at such events, when of course they do. There’s no suggestion that you should have these drinks instead of a soft drink. (Of course some people may be tempted to have a low-alcohol drink when perhaps they would have had a soft drink instead of alcohol, but that’s their free choice. Advertisers can’t stop people doing that.)
with an added implication of healthiness…
But lower alcohol drinks are healthy, compared to drinking regular strength alcohol. That’s the context.
Overall, all we have is another High School Social Science Project dressed up as real science, which has fooled newspapers into taking it seriously/given lazy editors some free copy.