On 2017 Oct 11, Clive Bates commented:

The most misleading thing about this paper is not merely the paper itself, but the way authors chose to present the findings to the media. As so often happens in academic work in this field, association became causation, uncertainty became confidence and very weak methods became the foundations of an assertive press release.

The press notice headline is:

Vaping doubles risk of smoking cigarettes for teens.

Of course, that is a made-for-media headline but deeply misleading about causation and grossly over-confident, given the study's limitations.

Much more likely is that the characteristics (personality, social and family context) that cause people to smoke also cause them to vape and the effects measured are the result of uncorrected confounding or reverse causation. It is quite possible that vaping acts to prevent smoking in people who would otherwise be susceptible to taking up smoking - an obvious possibility given the decline in teenage smoking in the US that coincided with the rise of vaping (albeit mostly experimentation and mostly without nicotine) - trends charts here.

My more detailed review of this study is here: Review of controversial Canadian ‘gateway effect’ study

On 2017 Sep 18, Neil McKeganey commented:

This paper will be interpreted by some as proof that using e-cigarettes acts as a gateway to smoking. That misleading impression is actually encouraged by one of the lead authors of the paper who in an interview for the Business New under the headline "E-cigarettes may double risk of tobacco smoking in teenagers" comments that the study provides a "yes" answer to the question of whether vaping lead to smoking. That statement is however in direct conflict with the stated limitations included within the published paper to the effect that "..the current study makes no claim of cause and effect". In fact this study is not even about the relationship between e-cig use and actual smoking but about e-cig use and a constructed variable for individual's susceptibility to smoke with the latter defined in terms of adolescents indication that they had not taken a firm decision not to smoke. So not having decided not to smoke is interpreted here as an indication that the individual is susceptible to future smoking. The authors rightly acknowledge that not all of those who are judged to be susceptible to smoking will actually go on to smoke in the future. Clearly that definition is problematic since it may well have been that many of the younger pupils had not found themselves in a situation where they felt they felt compelled to establish a clear view for themselves on the possibility of their future smoking. What the researchers have found here is that those pupils who in their definition were susceptible to smoking were more likely to have used e-cigarettes. As with other similar studies there is an entirely different possible explanation for that finding which has to do with the possibility that those pupils who reported past e-cig use were indeed more likely to smoke in the future because of the influence of a different variable e.g. willingness to experiment with different substances with that variable influencing both the likelihood of smoking and the likelihood of e-cig use rather than the latter influencing the former. As the authors acknowledge they have no measure here of intensity of e-cig use; rather they used a measure of any past use (of whatever actual frequency) and any use over the last 30 days (again of whatever actual frequency). As a result the e-cig using group could contain individuals who had used e-cigs on a single occasion in the past or a single occasion in the last 30 days. The only real conclusion which the authors can draw from this study is as they say that "use of e cigarettes among never smokers may increase the risk of future cigarette use.." It could equally be pointed out, though the authors choose not to say this, that "e-cigarette use amongst non-smokers may not increase future cigarette use."

On 2017 Sep 18, Neil McKeganey commented:

This paper will be interpreted by some as proof that using e-cigarettes acts as a gateway to smoking. That misleading impression is actually encouraged by one of the lead authors of the paper who in an interview for the Business New under the headline "E-cigarettes may double risk of tobacco smoking in teenagers" comments that the study provides a "yes" answer to the question of whether vaping lead to smoking. That statement is however in direct conflict with the stated limitations included within the published paper to the effect that "..the current study makes no claim of cause and effect". In fact this study is not even about the relationship between e-cig use and actual smoking but about e-cig use and a constructed variable for individual's susceptibility to smoke with the latter defined in terms of adolescents indication that they had not taken a firm decision not to smoke. So not having decided not to smoke is interpreted here as an indication that the individual is susceptible to future smoking. The authors rightly acknowledge that not all of those who are judged to be susceptible to smoking will actually go on to smoke in the future. Clearly that definition is problematic since it may well have been that many of the younger pupils had not found themselves in a situation where they felt they felt compelled to establish a clear view for themselves on the possibility of their future smoking. What the researchers have found here is that those pupils who in their definition were susceptible to smoking were more likely to have used e-cigarettes. As with other similar studies there is an entirely different possible explanation for that finding which has to do with the possibility that those pupils who reported past e-cig use were indeed more likely to smoke in the future because of the influence of a different variable e.g. willingness to experiment with different substances with that variable influencing both the likelihood of smoking and the likelihood of e-cig use rather than the latter influencing the former. As the authors acknowledge they have no measure here of intensity of e-cig use; rather they used a measure of any past use (of whatever actual frequency) and any use over the last 30 days (again of whatever actual frequency). As a result the e-cig using group could contain individuals who had used e-cigs on a single occasion in the past or a single occasion in the last 30 days. The only real conclusion which the authors can draw from this study is as they say that "use of e cigarettes among never smokers may increase the risk of future cigarette use.." It could equally be pointed out, though the authors choose not to say this, that "e-cigarette use amongst non-smokers may not increase future cigarette use."

On 2017 Oct 11, Clive Bates commented:

The most misleading thing about this paper is not merely the paper itself, but the way authors chose to present the findings to the media. As so often happens in academic work in this field, association became causation, uncertainty became confidence and very weak methods became the foundations of an assertive press release.

The press notice headline is:

Vaping doubles risk of smoking cigarettes for teens.

Of course, that is a made-for-media headline but deeply misleading about causation and grossly over-confident, given the study's limitations.

Much more likely is that the characteristics (personality, social and family context) that cause people to smoke also cause them to vape and the effects measured are the result of uncorrected confounding or reverse causation. It is quite possible that vaping acts to prevent smoking in people who would otherwise be susceptible to taking up smoking - an obvious possibility given the decline in teenage smoking in the US that coincided with the rise of vaping (albeit mostly experimentation and mostly without nicotine) - trends charts here.

My more detailed review of this study is here: Review of controversial Canadian ‘gateway effect’ study