Tuesday, November 27, 2007

Lies, damn lies and statistics

I didn’t take the tiffin into work yesterday, as I knew that C would not be in and she would never forgive me for taking in anything that contained chocolate on a day when she wasn’t in. It might help to give us a bit of energy for our journey down to the south coast tomorrow evening, as we are on a course tomorrow in Portsmouth which should be interesting. This means no blogging tomorrow, but you get an extra long post today anyway.

C has got a job interview on Thursday, so we are going to spend some of our train journey preparing for that. I have no idea why I am helping her with this, as it might mean that she leaves, but it seems a bit harsh not to help her. We already spent an hour last week going through some bits, so hopefully she will have worked on those and we’ll build on that a bit.

Anyway, I have been reading Freakonomics, which has given me much food for thought. The basic premise of the book is that the author, an economist, looks at a range of things that we perhaps take as a given – the statistics we hear on the news, the assertions about crime (and its rise or fall) in the media etc - and analyses it from an economists perspective. An interesting and amusing economist I might add before your heart sinks.

I am certainly not an economist, but I am a sociologist (of sorts). My degree is in sociology and a lot of my interests in life are connected with what makes society tick, what the influencing factors are and how you can break that down in order to address at least some of those issues. As you can tell, I am *very* interesting at parties…

Anyway, Freakonomics is therefore a book that greatly appeals to me, not necessarily because I always agree with it (or think that it even makes sense in places), but because it challenges assumptions and attempts to put the alternative case and I like that quality. Many a time I read the newspaper and there is so much supposition and extrapolation based on so little evidence that it is why I would rarely bother to pay for a newspaper. I do, however, no someone who deliberately buys the Sunday Telegraph because she *knows* she will disagree with it and therefore enjoys the mental stimulation. Each to their own.

A whole range of subjects are covered in the book from the Ku Klux Klan to estate agents to whether teachers help their students to cheat. However, there are three things in particular that I wanted to comment on about this book. One is the assertions about internet dating and the other is about the link between abortion and a decrease in crime and the final one is what seems to be a total misinterpretation of some statistics about the number of deaths caused by swimming pools and guns.

I’ll start with the least controversial of these – that being internet dating, in case you were wondering. Basically what the author is discussing is the relationship between what people say on their internet dating profile compared to a) its accuracy and b) the likely response i.e. how many people respond to a particular profile. It makes interesting reading and looks at some of the factors that are likely to increase or decrease your appeal to potential suitors. The main reason you’re unlikely to get responses is if you fail to post a photo apparently.

Anyway, the book talks about how people’s profiles compare to the demographics of the wider population e.g. if the US 20% of the population are blonde then does that percentage stay the same when people fill out their dating profiles. It seems not. The book says “[Internet daters] were also a lot richer, taller, skinnier and better looking than the average. That, at least, is what they wrote about themselves. More than 4 percent of the online daters claimed to earn more than $200,000 a year, whereas fewer than 1 percent of typical internet users actually earn that much, suggesting that three of the four big earners were exaggerating. Male and female users typically reported they are about an inch taller than the national average. As for weight, the men were in line with the national average, but the women typically said they weighed about twenty pounds less than the national average”

Don’t get me wrong, I’m not going to suggest that I think that people who fill out internet dating profiles are necessarily either honest or accurate – although you do perhaps have to raise an eyebrow at those who describe themselves as “happily married” on an internet *dating* site. What I would question though is whether you can directly compare the general population with internet users. First of all there is only a certain percentage of the population who have a computer with internet access (a minimum requirement for internet dating surely?) – are people who are on very low incomes likely to own a computer at the same ratio as though who earn an average or above average income. I think not – for a combination of economic and aspirational reasons. I’m not saying it doesn’t happen but it is less likely. Those on poor incomes, so we are told and therefore have to rely on the accuracy of this data of course, are often the ones with a poorer diet. This could lead to two things, first is that they might weigh more than the average and second is that due to poor diet their height may be stunted (I’m not suggesting they are midgets, just a bit shorter…). They are also perhaps not used to going to interviews where you have to really sell yourself. If you have a low-skilled manual job, your need to be able to describe all the different options you consider when you are carrying out that role is probably rather less than if you are a professor of economics, for example. So maybe those who use internet dating sites understand the need to highlight their assets and brush over those thngs that don't show them at their best - maybe that's why they are willing to 'market' themselves in this way.

So, what I am arguing is that unless you look at the demographics of those who use the internet (and those who don’t) and see how that skews any of the statistics, you’re not going to have the complete picture. If the people who use dating sites tend not to be drawn from those on a low income then perhaps that does change the averages and mean those who use such sites weigh less and are slightly taller than the average population. Not quite so revelatory then though is it…

To be honest, I have absolutely no idea about the truth of any of this and am playing devil’s advocate really. I was just frustrated that a book that was meant to be demystifying the things we hear reported in the media etc, jumped a step in the process in this instance and compared like for like, when they weren’t necessarily directly comparable.

Good for discussion, if only because I disagreed with it, but not necessarily good to base your hopes of finding the live of your life.


Tom Foolery (TF) said...

Food for thought..... never have trusted stats myself. TFX

Random Reflections said...

TF - the book is actually well worth a read, but it maybe doesn't live up to its own claims on occasion.