David Spiegelhalter, Sex by numbers, London, Profile books, 2015

For many, sex is more about quality than quantity, so David Spiegelhalter’s Sex by numbers may put off many potential readers.  Yet, in sex quantity seems to have a certain quality of its own –or so it claims Brooke Magnanti, invoking the intellectual authority of Stalin. This book has indeed a quality of its own: accuracy. Our folk understanding of sex is full of made up numbers to which we inadvertently stick without further reflection. Checking them out is more complicated that it seems: there are competing sources and we need a certain degree of statistical (and methodological) literacy to assess them properly. Hence, we can only be grateful to have Spiegelhalter, a world-leading statistician, spelling out for us what we really know about the numbers of sex in a clear and accessible manner.

Spiegelhalter is a Bayesian: for him, probabilities measure how strong our beliefs about random events are. This is a technicality that readers may safely ignore, but it explains why the book starts with a credibility ranking of the available figures about sex: numbers we can believe, numbers that are reasonably accurate, numbers that could be out by quite a long way, numbers that are unreliable, and numbers that have just been made up. Evidence in the two first categories will improve our statistical understanding of sex, whereas the remaining three scores will probably mislead us.

Ranking evidence depends crucially on its sources and half of this book is about how social research on sexuality can be properly carried out. Spiegelhalter's paradigm for reliable data is the British National Survey of Sexual Attitudes and Lifestyles (Natsal-3, 2010-2012). This is based on a random sample of face-to-face interviews funded by a private charity, the Wellcome trust -incidentally, the same trust commissioning this book. Spigelhalter spends time discussing the methodology and comparative reliability of his many sources (devoting an entire appendix to Natsal methods) and the crucial choices on which they all depend (e.g., what counts as a sexual partner). He also uses Natsal data in the first few chapters to introduce a number of handy statistical concepts: the mean and the median are important when we wonder about how much sex we are having.

With all these methodological caveats in sight, Spiegelhalter proceeds to inform you about everything you thought you knew about sex, despite not having a reliable source to check. Statistics about partners, heterosexual and homosexual activity, masturbation, reproduction etc. Most of it illustrated with graphics, about which I will make my only formal complaint about the book: in the epub version, sometimes they were not easy to read (despite trying the graphics on various readers). The text instead is delightful to read. Spigelhalter excels at both clarity and wit, both in the best British tradition, even if (or perhaps because) the topic is sex. 

Spigelhalter is cautious, but not shy, in appraising causality through data. Sometimes the evidence makes more likely some explanations of why sex happens the way it does. But often the data are far from conclusive regarding causation and, at best, they just describe what we do (and how often we do it).  Spiegelhalter adopts a good old positivist stance regarding the science of sex and admits at all points what we do not know, keeping it separate from any normative judgment. The most opinionated readers may be displeased by such a sober discourse on sex. The rest of us will be surely enlightened by the quality in the quantity.

{August, 2016}
{Metapsychology}