Comments

Defining smokers and vapers — 5 Comments

  1. “Have you smoked/vaped in the last 30 days?”.  This is factually meaningless.”

    Surely, it’s specifically designed that way? How else can the weasels prove what they want to prove?

    • Of course it’s designed that way.  The response could be factually correct but it makes no differentiation between an experimenter [just one], a very casual smoker [one or two on Saturday night] and a heavy smoker [60 a day or more]. 

      Equally all these studies and questionnaires rely on two things – an impossibly precise memory and complete honesty.  Unless someone buys one 20-pack every day and smokes every one without fail, the number is going to be inaccurate or at the very least a best-guess.  And how many 40 a day smokers tell their doctor they smoke 20 [or less]?  Hah! 

  2. Having friends who spend their lives with statistical research, the first question is to ask to see the fieldwork of the findings.

    Even sample size is irrelevant if that sample is not a representative cross-section. 

    Anthony Wells of YouGov published a good piece on bad sampling on Tuesday

    http://ukpollingreport.co.uk/blog/archives/9805

    • There are another factors that are common to just about every report that is produced these days – risk and relative risk.

      They run experiments on mice [usually with an extreme level of the “toxin”.  A mouse develops symptoms of a disease and they loudly proclaim that there is a “risk” of the toxin causing the disease, ignoring the levels of exposure and the simple fact that a mouse isn’t a human.

      Now suppose the normal risk of catching this disease is .000001% and their results show that .000002% off the mice succumb, they then scream from the roof tops that the toxin doubles the risk, or it increases risk by 100%.  Technically both statements are correct but are misleading to the extreme.  i.e. there is still no risk of any consequence.

      • Generally, what we receive is so filtered by a determination to find results that fit the thesis proposed that any number pointing in a particular direction will be used.

Hosted by Curratech Blog Hosting