Warning labels have no effect on smoking

A recent Ministry of Health survey has shown that there has been no change in smoking rates with the introduction of graphic labels on cigarette packets. 24.3% of the population smoked in 2006, and 23.9% in 2008, with the difference being not statistically significant. Not significant means that any very slight differences could be due to random variation, so the numbers are the same – about 24% in both years.

Now that should tell us that the labels aren’t working. But it isn’t being spun that way. The Press reports

“Warnings credited with smoking fall:
Graphic photographic warnings on cigarette packets are being credited with prompting a dramatic drop in smoking rates.”

That’s the first time I even heard no change being called a “dramatic drop”. But rather than quote the actual figures, they invent a blatant lie:

Two years ago, when the introduction of photographic health warnings on cigarette packets was announced, a quarter of New Zealanders smoked. Now, almost a year after the graphic reminders of the health risks of smoking were placed on every tobacco product sold, that has dropped to one in five.

No need to comment on that nonsense. The official press release at least quotes the actual figures, but also tries to make it sound like there is a change.

The 2008 New Zealand Tobacco Use Survey shows that, after adjusting for age, 23.9% of New Zealand adults, aged 15 to 64, are smokers. The Ministry of Health’s National Director of Tobacco Control, Dr Ashley Bloomfield says, “While the drop on the equivalent 2006 figure of 24.3% is not statistically significant, this result confirms the downward trend in smoking.”

That is ridiculous. If something is not statistically significant AND is obviously a very minor change, THERE IS NO CHANGE. I would expect a reporter to critically analyse the press release and show this – and that would have made a great story by the way. Yet instead they blindly believe it, ignore the stats and invent their own more impressive sounding figure (1/4 to 1/5) to make the stats fit the desired message. Come on – are you a reporter or a Ministry of Health PR spokesperson?

Hat tip: Bernard Darnton

Advertisements

2 Responses to “Warning labels have no effect on smoking”

  1. Sb Says:

    ” If something is not statistically significant AND is obviously a very minor change, THERE IS NO CHANGE. ”

    Very good – so few people seem to understand this. Agree %100

    Sb

  2. Mr Dennis Says:

    Thanks Sb. Statistics are just a different way of looking at data. If there had been a large drop but for some reason (low replication for example) it did not test as statistically significant, you may be justified in still saying it looks like there is a reduction, as they try to in the official press release. But when the change is both minor and not significant, as in this case, there clearly is no change.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: