Sweet or Sour? The Appeal of E-Cigarette Ads

Once again, another study citing “concerns” reaches the media. This time it is all about flavoured vs non-flavoured e-cigarettes, gateways and smoking all based on exposure to adverts. As David Dorn highlights on his blog post:

So asking kids whether an advert they won’t see (by law), for a thing they can’t buy (by law), in a place they can’t use them (by bye-law) is likely to make them want one is a pointless, fruitless and, frankly bloody idiotic thing to do.

I can’t really argue with that sentiment. So where did this study come from?

Behaviour and Health Research Unit, University of Cambridge, Cambridge, UK

Snappily entitled “Impact of advertisements promoting candy-like flavoured e-cigarettes on appeal of tobacco smoking
among children: an experimental study” 
this study decides to assess the impact on appeal of tobacco smoking after exposure to ads for e-cigarettes with and without “candylike” flavours – by the way, “candy” is usually a generic US term for “sweets”.

Yes you read that right, they were looking specifically at the appeal of tobacco smoking – i.e. how likely are you to smoke – after you’ve seen these e-cig adverts.


“We assigned 598 English school children (aged 11–16 years) to 1 of 3 different conditions corresponding to the adverts to which they were exposed: adverts for flavoured e-cigarettes, adverts for non-flavoured e-cigarettes or a control condition in which no adverts were shown. The primary endpoint was appeal of tobacco smoking. Secondary endpoints were: appeal of using e-cigarettes, susceptibility to tobacco smoking, perceived harm of tobacco, appeal of e-cigarette adverts and interest in buying and trying e-cigarettes.”

Now, forgive me for repetition, but why on earth would you expose an age group that cannot legally buy a product (tobacco cigarettes) to adverts for another product that is vaguely similar? Then based on information gathered, ask similar questions on e-cigarettes to determine the appeal of using them and their adverts.

The group of 598 students were then split into three areas :

  1. Advertisements of candy-like flavoured e-cigarettes (n=206)
  2. Advertisements of non-flavoured e-cigarettes (n=191)
  3. No advertisements (control condition) (n=201)


They did exclude those who were already using either tobacco or e-cigarettes (n=128) resulting in only 470 (though the study actually states they excluded 127 participants who had ever used tobacco or e-cigarettes, leaving 471 – someone can’t do math) for final analysis.

Two of the authors selected twenty-four adverts (split into half & half for flavoured/non-flavoured), each of which were carefully selected to be similar in nature (i.e. “showing an e-cigarette pack with an e-cigarette next to it”), included a person in the image (including three adverts in the flavoured/non-flavoured categories showing the product being used).

Carefully selected adverts showing specific detail to measure the impact of them on the participants to see if a) they will want to smoke and b) how “interested” they are in trying an e-cigarette.

See, the thing with the advert selection is there is a pre-determined theme being shown. If this study is to be as impartial as possible the selection of adverts would be more randomised, after all creating a gallery database to randomly pick x number of images is something that web-coders have been doing for years. Set the search for the total number of images with a certain number for each category, hit search and there you go. That would eliminate any predisposition on the part of the study authors in the advert selection.

How did our Cambridge fellows measure the “appeal” of smoking and e-cigarettes? By using an atypical study reference and asking a graded set of questions (scale 1-5):

‘Please cross the circles that best describe how you feel about smoking tobacco cigarettes’ using attractive/not attractive, cool/not cool, boring/fun. As you would expect, 1 denotes the lowest (not-attractive/not-cool/boring) with 5 meaning the opposite.


The “appeal of tobacco smoking” and “appeal of using e-cigarettes” is consistently low, however it may be that some of the participants chose “socially acceptable” answers to those questions, after all kids aren’t known for being completely truthful are they?

So what do those numbers tell us?

In reality, not a great deal other than perceived interest results. The control group were not shown any adverts, just given the questionnaire, so in one interpretation of the results you could say that “exposure to e-cig ads increases their appeal”, but the devil (as they say) is in the discussion. Interestingly, none of the groups were given a questionnaire to complete prior to gain an understanding of a before and after – known as a baseline reading. Based on a single result, there isn’t much in the way of conclusion that can be gained regarding any “increase”.

“we found no evidence that exposing English children aged 11–16 years to adverts for candy-like flavoured and non-flavoured e-cigarettes increased the low appeal of smoking tobacco, the low appeal of using e-cigarettes, or low susceptibility to tobacco smoking. Nor did it reduce the high perceived harm of tobacco smoking.”

So based on the questionnaire and the responses, there is no evidence to support a supposed “gateway” theory by being exposed to e-cigarette adverts, nor is there any substantial evidence to support that e-cigarette adverts are “targeting” youth. But mention flavours to youth in an advert and their interest increases. Well, flavours are appealing to pretty much every human being. It doesn’t mean that the person is going to buy them.

Our data provide no support for the renormalisation hypothesis, since exposure to e-cigarette adverts did not increase the appeal of tobacco smoking in this sample of children.

Don’t think much needs to be said there. The whole “renormalisation” argument is pretty much null and void as various surveys can attest to – wherever e-cigarette use is on the rise, combustible tobacco use is in decline. This isn’t necessarily the “hard evidence” needed, but it is suggestive of a link between the two.

However, our data suggest that certain types of e-cigarette advertising (eg, for candy-like flavoured e-cigarettes) may provide a gateway into tobacco smoking by increasing the appeal of e-cigarette adverts, and increasing interest in buying and trying e-cigarettes.

Um, what? The authors have already suggested that the gateway theory is a myth at the top of the discussion, now they are saying that certain advert types are a gateway to tobacco? To be fair, it is a valid discussion point, but hardly conclusive based on a single study, which of course the authors go on to mention:

Future studies could use the social learning theory and other relevant theories to test the proposed gateway and renormalisation hypotheses for the impact of e-cigarette exposure on tobacco use.

The typical “money shot” call for more research (i.e. funding) into the proposed gateway and renormalisation hypotheses, which their own results indicate doesn’t actually exist. In plain English, “we didn’t find any signs of the gateway or renormalisation theories, but give us more time and we’ll come up with something”.

One limitation pertains to the primary endpoint, which was a measure of attitude and not behaviour. Future studies would benefit from measuring actual smoking behaviour. Similarly, tobacco smoking carries a stigma, therefore, it is possible that children gave socially desirable answers regarding the appeal of tobacco smoking.

Reading between the lines a little here, the authors are suggesting that an attitude towards something is in some way linked to actual behaviour, but they also point out the “socially acceptable” answers – something that many researchers tend to overlook.

Interestingly, the authors proceed to highlight other methods that could be used to provide results, such as showing tobacco adverts (not seen in the UK for years) in parallel to e-cigarette adverts to try to gauge the responses to both; I would suspect that by doing that they’ll identify a pattern of attitudes in youths that like to take more risks.

Future studies should examine the effects of e-cigarette adverts using different formats such as videos. Examining children’s exposure to e-cigarettes via their peers in their naturalistic environments would also complement the current findings.

So vary the advert medium and environment in “future studies”; I suspect that they didn’t do that this time around either due to time constraints (convenient publishing time) or funding constraints.

Future studies could extend the design to incorporate a control condition similar to those in the e-cigarette advert conditions (eg, adverts for neutral products, like stationery).

So the child is happily (?!) browsing through the adverts in the questionnaire to be confronted with a random advert for a Parker pen or some A4 paper as a control? Surely that would prompt the brain to switch off? Lorien expresses her thoughts on this particular aspect here.

While we found the same pattern of results when carrying out sensitivity analyses on the full randomised sample including tobacco smokers and e-cigarette users, it is possible that our results underestimate the effects of e-cigarette adverts on tobacco smoking, since by removing the smokers and e-cigarette users we may have ended up with a sample of the least susceptible children.

In plain English, when we excluded those that already smoked or used e-cigarettes the overall results weren’t significantly different from when we included them and that by removing them, the kids that answered the questions are most likely not bothered by e-cigs or tobacco anyway; but future studies should include existing users.

In a nutshell, this study is aimed at “informing policy” by pointing out a “need for further examination of the rules surrounding e-cigarette advertising”:

our results provide evidence that children find adverts for candy-like flavoured e-cigarettes more appealing than adverts for non-candy-like e-cigarettes, potentially serving as a gateway into tobacco smoking.

No. Nowhere do the results suggest a potential “gateway” into tobacco smoking. While the study itself is limited in scope, the relatively consistent results for the “appeal of tobacco smoking” regardless of whether or not the child saw a “candy flavoured” advert or a “non-flavoured” advert would lend itself to the result that e-cigarettes are not likely to be a gateway to tobacco smoking.

In addition to the gateway concerns, the heightened appeal of the adverts and interest in buying e-cigarettes in children arising from adverts promoting candy-like flavoured e-cigarettes is of concern in and of itself in view of the dangers to the developing brain arising from nicotine exposure and addiction, as well as the unknown long-term physiological effects of using e-cigarettes

So the “gateway” is now a concern rather than a potential, but the next point – heightened interest is of genuine concern, but if the child decides to pick up a flavoured zero-nic e-cig then there’s little to no chance of nicotine exposure and addiction, yet the authors don’t make the point that zero-nic e-cigs even exist, nor do they reference the only study that actually asked about nicotine use.

and secondhand exposure to vaping.

So the entire paper devolves into a moralistic stance on perceptions. Second hand exposure. Good grief.