Wednesday, August 6, 2014

Guardian's datablog not buying Gay Times drug/sex survey quoted by BBC Newsbeat, Huffington Post, PinkNews and others

Above: Gay sex and drug use by German cops in Freefall, now streaming on Netflix

The Guardian sneered at the headline's conclusions, referring to "the bone­-shattering dog-bites-man banality of the top line — people have sex when drunk" before getting down to the business of taking apart the survey's numbers, which the newspaper regarded as highly suspect relative to other surveys.
The survey’s conclusion is clear: gay and bisexual men are incorrigible, unrepentant alcoholic, drug-snorting sex dreadnoughts, who wouldn’t be seen dead getting it on without a line of something typically found in a rock star’s dressing room.
      The survey has a fundamental problem, which is that, unlike most professionally-conducted surveys and research, its participants were self-selecting. It was hosted online, advertised to Gay Times readers online, and pitched specifically as a ‘drugs survey’. This can have a big effect on who answers the survey, both in terms of who hears about it, and who feels that it’s relevant to them and so worth taking the time to respond to.
     By way of illustration, a 20-something party animal living in London who uses Twitter, frequents the Vauxhall scene and is comfortable using recreational drugs is both more likely to hear about and more likely to answer the survey than a middle-aged gay man living in the countryside with a dodgy internet connection.
     This may sound quite minor, but we can start to see how this may have a large effect on the results when we compare it to previous research efforts. Similar data can be found not only in a 2012 survey carried out by Stonewall (which had a much larger sample of nearly 7,000), but also in recently published figures from the authoritative Crime Survey for England and Wales (CSEW). Both sets of figures suggest GT has considerably overestimated drug use.

No comments:

Post a Comment