Fifty shades of statistics: Professor David Spiegelhalter on the statistics of our intimate lives

Today sees the publication of Professor David Spiegelhalter’s latest book, Sex by Numbers: What Statistics Can Tell Us About Sexual Behaviour. To coincide with this publication, you can now view the webinar of the lecture which Professor Spiegelhalter gave that was based on this book during last month’s Cambridge Science Festival. If you are not already a registered member, all you need to do is register for free.

Whatever society we live in, and however open-minded we like to think we are, when it comes to our sex lives we all like to keep a few secrets. But this makes the jobs of sexologists – professionals who study sexual behaviour – pretty difficult.

Here Spiegelhalter, Winton Professor for the Understanding of of Risk at Cambridge University, unravels the web of exaggerations, misdirections and downright lies that surround sex in modern society. Drawing on the Natsal survey, the widest survey of sexual behaviour since the Kinsey Report, he answers crucial questions such as what are we all doing? How often? And how has it changed?

Statistics Views interviews Professor Spiegelhalter on what he could tell pre-publication!

 

1. Congratulations on the publication of Sex by Numbers: What Statistics Can Tell Us About Sexual Behaviour. How did the writing process begin? What inspired you to come up with the idea for the book?

The idea was conceived at a lunch with my publisher and the Wellcome Collection, whose exhibition ‘Institute of Sexology’ seemed a great reason for a book about sex statistics. So a good idea, a fine title: the only problem was settling down and writing 300 pages.

2. What were your main objectives when writing the book? What did you wish to achieve in reaching out to your readers?

Tim Harford has cleverly pointed out that this is a book about statistics, disguised as a book about sex. Of course there are a lot of interesting stats, with some extraordinary numbers highlighted in the text to grab readers’ attention, and lots of graphs. But underneath is the running theme of how these numbers were obtained, and how reliable they might be. So my aim is that people will enjoy reading about sex, but also learn something about the reliability of data.

3. For your book, you have drawn much from the Natsal report. In your research, did you also look at past reports, such as Kinsey and Masters & Johnson?

I feature Kinsey in some detail, not that his stats are very reliable, but because he was a pioneer of a particular form of interviewing and data collection, which is now part of history. He was an extraordinary man with some unusual sexual habits of his own, and I have a grudging respect for his contempt for statisticians. I’ve got little about Masters and Johnson – they were laboratory scientists interested in individual sexual response, rather than the behaviour of populations. And their stats were poor.

Investigators of the genetic basis of sexual identity like to study the ‘concordance’ of the sexual identities of identical twins. But many researchers in this area have advertised for people who are an identical twin and consider themselves gay or lesbian, without apparently realising the bias this is building in: sets of twins with concordant sexualities are twice as likely to answer such an advertisement as non-concordant twins. This is quite a subtle, but important, form of selection bias.

4. What was the most bizarre/interesting research you came across?

One of the bizarre experiments investigated whether the ‘disgust’ response was related to sexual arousal. Based on the observation that people are prepared to do things during sex that they might normally balk at doing with another person, the experimenters randomised women to different stimulae, including erotic films, and then set them a number of fairly revolting tasks which I won’t go into here. Those who had been exposed to arousing stimulae managed more of the tasks.

5. Why is this book of particular interest now?

I have the impression that it is becoming easier to discuss sex in a reasonably balanced way, without it being either tabloid sensationalism or bogged down in scientific language. But it still seems as if it’s difficult to discuss sex outside the context of an argument – between men and women, between old and young, for or against legal prostitution and so on.

6. Were there areas of the book that you found more challenging to write, and if so why?

I avoided anything on sexual abuse, but felt I had to cover coercion as it is such an important area, and statistics of coercion and rape are particularly contested.

7. Statistics can be misused and abused and much of your blog, Understanding Uncertainty is about pointing such stats out. What was the worst misuse of statistics that you came across in your research?

Where do I start? ‘Surveys’, which just comprise volunteers who choose to fill in questionnaires on paper or online, are common in this area and are utterly unreliable. But there are more subtle forms of bias. Investigators of the genetic basis of sexual identity like to study the ‘concordance’ of the sexual identities of identical twins. But many researchers in this area have advertised for people who are an identical twin and consider themselves gay or lesbian, without apparently realising the bias this is building in: sets of twins with concordant sexualities are twice as likely to answer such an advertisement as non-concordant twins. This is quite a subtle, but important, form of selection bias.