July 3rd, 2007
In social research there’s a name for when people try and squeeze the maximum amount of publications out of a study as they can. It’s called ‘salami slicing’ and it tends to be a derogatory term aimed at those who milked their research for more than it was really worth.
Obviously in any large scale piece of research (or even smaller studies) there may be more than one paper you want to write. But eyebrows tend to be raised when a number of papers saying the same thing are published in different journals, or where a lot of coverage is given to a particular study, product or organisation in the media that appears to be based on just one set of data.
Today we’ve got a great example of this going on within a media/commercial outlet with yet another story from the Durex global sex survey.
Last week they told us the French were lousy lovers in some kind of international sex contest. Today we’re being told the shock headline that 1/4 of virgins aren’t practising safer sex.
Press reports have stated “More than a quarter of people in the UK do not use contraception when they lose their virginity” and again introduced a competative angle by stating the UK is in the bottom half of the table for practising safer sex. So now we’ve got safer sex league tables to measure up to? Setting countries against each other is hardly the best way to achieve an increase in contraceptive use. Although it is a great way to guarantee headlines as whether you’re a country that’s in the bottom or the top of the Durex survey’s league table you’ll have something to shout about.
But if you then look at the data it suggests that nearly 80% of teenagers in the UK used contraception the first time they had sex. So that’s a majority of people being safe, rather than the more scary headline of 1/4 of virgins not using contraception. To be fair I suspect that the 80% figure is actually quite unlikely, but this is coming from an online survey for a condom manufacturer so respondents are a. going to be more literate (and therefore may well be better able to negotiate safer sex), b. are interested in contraception/sex information by virtue of being on Durex’s website, and c. are less likely to reveal in this survey that they have not used contraception when they had sex.
Obviously it makes a better headline – and also provides a better opportunity for product promotion – if you imply that people are not using contraception when they first have sex.
Like anyone working in sexual health I want people to have sex when they are ready, to use contraception when they do have sex, and to enjoy their first and following sexual experiences. Criticising the Durex survey is not an exercise in suggesting we don’t need to worry about condom use. It is simply to point out how dodgy tactics like data fishing, tweaking the reporting of percentages and salami slicing findings are not good research practice. It is also inaccurate to just rely on percentages when reporting findings – statistical analysis may well reveal there are fewer significant differences between countries than currently reported in the media.
Add to that a lack of ethical approval for the research and poor practice in survey design we should be careful how we use the data produced from this survey. Although it impresses journalists because it has a large number of respondents and appears to be a global study, if the data collected isn’t faithfully reported (or analysed) then it becomes redundant.
We should consider this before being willing to plug a survey and in turn a product. After all that’s two lots of free advertising for Durex in the space of a week.Tweet