In 2008, the children, who were pupils at 21 public schools in three different regions of Germany, were asked how often they had seen particular ads. These included images for six of the most popular cigarette brands in Germany and eight other products, such as chocolate, clothes, mobile phones, and cars.
In 2011, 30 months later, they were asked the same question, as well as how many cigarettes they had smoked to date, and whether they smoked regularly.
One in three (406) admitted to having tried smoking during the 30 month period, with one in 10 (138) saying that they had smoked within the preceding month.
One in 20 (66) kids said they had smoked more than 100 cigarettes, and were therefore classified as "established" smokers, while a similar proportion (58) said they now smoked every day. A third of the daily smokers were aged 14 or younger; one in four was 16 or older.
Exposure to cigarettes ads was much lower than that for the other non-tobacco products, but an ad for one particular cigarette brand was seen by almost half the kids at least once, and more than 10 times by 13% of the sample.
When a range of well known influences for taking up smoking was factored in, smoking among peers proved the strongest influence, followed closely by exposure to tobacco ads.
The greater the exposure to tobacco ads, the greater was the likelihood that the teen would take up smoking, the analysis showed.
Teens who saw the most tobacco ads (11 to 55) were around twice as likely to become established smokers and daily smokers as those who saw the least (0 to 2.5).
And for each additional 10 sightings of a tobacco ad, a teen was 38% more likely to become an established smoker, and 30% more likely to smoke every day compared with sightings for non-tobacco product ads.
After taking account of other influential factors, the overall risk of becoming an established smoker was between 3% and 7.3% greater, while that of taking up daily smoking was between 3% and 6.4% greater, depending on how many ads the teen had seen.
The authors acknowledge that a large proportion of the original 2300 students involved dropped out, and confirm that as with any observational study, there is always the chance that some as yet unmeasured factor could explain the results.
But they conclude that the data from their study support the content-specific association between tobacco advertising and smoking behaviour and, therefore, the total ban on tobacco advertising advocated by the World Health Organization Framework Convention on Tobacco Control.
"Data from this study support this measure, because only exposure to tobacco advertisements predicted smoking initiation, which cannot be attributed to a general receptiveness to marketing," they write.