99.873% of statistics are made up

Today I was involved in a Twitter argument with two Prominent Liberal Democrat Bloggers. I’ll leave their names and the precise details of the argument out, because it’s not germane (and also because I may inadvertantly misrepresent one of them in the very abbreviated precis that follows), although anyone who really wishes can look it up on Twitter. But the argument went something along the lines of:

Prominent Liberal Democrat Blogger 1: Sign this petition banning the distimming of doshes!
PLDB2 : But that says that studies show that distimming causes gostaks to go blind. In fact all the studies show it causes them to grow an extra foot!
PLDB1: That doesn’t matter! Just sign the damn petition! Distimming is wrong!
PLDB2: I’m not signing a petition with things in it that are demonstrably untrue!
PLDB1: But you can never be 100% accurate, so just sign the damn thing! Anyway, you can prove anything with statistics!
Me: Are you seriously saying that just because absolute inaccuracy is not possible, you shouldn’t make any effort to remove obvious falsehoods?
PLDB1: Don’t sidetrack the discussion! This is about distimming! Anyway, people will argue over anything, no matter what you do.

And then on, for many more 140-character responses, essentially going round in circles.

Now, this isn’t the first time I’ve seen this attitude recently – a couple of weeks ago there was a storm in a teacup over a famous campaigning organisation running a campaign for an excellent cause, but with a headline figure that was not very accurate. I won’t link to anything about that (although most politically-minded people reading this will have a good idea what I’m talking about), because I don’t want to give ammunition to the kind of people who will use the inaccuracy against the cause itself.

But the thing is, I shouldn’t have to do that. I shouldn’t have to choose between telling the truth and discrediting a worthy campaign. Using misleading or outright wrong facts is the kind of thing we excoriate the Mail or Express about, and we shouldn’t be doing it ourselves. Were the Mail to headline “75% of people think immigrants should be hanged!” then we’d be all over the article, tearing it to shreds, but the same people would be silent if they saw something saying “75% say ID cards are wrong”.

Citations of studies and statistics can be very useful, as can using raw numbers. Amnesty are currently campaigning to stop 128 executions in Iraq., for example, and that’s an important campaign. But it stands or falls on the 128 number, so they’ve ensured they’ve got it right. If it turned out there were only five people being executed, and the other 123 were being given free chocolate instead, Amnesty would quite rightly argue that the death penalty is still wrong, and that any executions are too many. But they would look idiotic. (Sadly, this is not a case where the numbers are wrong…)

We need, as ‘progressives’ (whatever that very devalued word still means) to be at least as strict with ourselves as we are with the other side. In particular, we need to acknowledge unpleasant evidence. We can’t say, for example “Cannabis should be legal, as it’s harmless” – it’s clearly *not* harmless, as the many people suffering from cannabis psychosis would attest. But we *can* say “Cannabis should be legal, *even though it can cause harm*, as the harm it causes is less than the harm caused by denying adults the right to do as they wish with their own brains”. Saying “the minimum wage doesn’t have any negative effect on jobs” is wrong – the minimum wage clearly prevents the creation of some small number of very low-paid jobs. But saying “the overall positive effect of the minimum wage – which prevents workers from living on starvation-level incomes – more than offsets its small negative effect” is truthful.

If our arguments are right, we don’t need spurious pseudo-evidence to back them up, and if they’re wrong we shouldn’t be making those arguments in the first place. Using factoids, rather than facts, is one of the things that makes people think ‘they’re all the same’ – because sooner or later one of those factoids will contradict the listener’s personal experience, and s/he will write the source off as a liar.

We can’t get everything right, but it’s not difficult to find a reliable source for any statement of fact you make (if it can be done for Wikipedia it can be done for a political campaign or petition) and if you do find such a source, at least you can then say “It was in reliable newspaper X or peer-reviewed journal Y”, rather than imitating Reagan and saying “facts are stupid things”.

This entry was posted in politics and tagged , , . Bookmark the permalink.

10 Responses to 99.873% of statistics are made up

  1. Dave Page says:

    I’d say that the same applies to adopting policies which are justifiable rather than popularist…

  2. Jennie says:

    LOL prominent lib dem bloggers! I love you dearly, I do!

  3. Andrew, you have so misrepresented what I actually said in the first part of your blog that it’s almost funny! I don’t have the energy to point out to you yet again, what I actually said rather than your precis; but it just goes to show my point that people read and see what they want to see and therefore throwing stats at each other all day can be a colossal waste of time.

    You don’t need to keep me anonymous, I’m not ashamed of what I was saying – I just don’t think you actually understand what I was saying. I would perhaps counsel against doing précis repeats of arguments you’ve been in again – they tend to be inaccurate at best and defamatory at worst.

    Just as a side shoot, if you actually looked at the petition, you would there were no stats actually used (FFS) but rather statements about the effect that lapdancing and the clubs where it takes place (and the character in the story line was emulating and promoting) have on the local community and the safety of women. There are plenty of qualitative & quantitative studies to support those statements – I went online last night and found them very quickly.

    There are probably plenty of qualitative and quantitative statements that support the opposing view. I didn’t find any on the web, when I went looking for them last night but I’m sure someone could find some.

    If that is the case then you will end up just having very long arguments and throwing stats at each other and never actually move on.

    My argument is, don’t get distracted by a stats row; if you don’t like the specific stats used then go and find some you do like and reword your own petition so that you are happy with it – if, if you broadly agree that promotion of the sexual industry on a programme with a large child audience is not the best way to halt the sexualisation of young people, and particularly young girls, in this country. I really don’t think that it supports the gender equality duty that the BBC has.

    It is the easiest thing in the world to diss an idea you don’t like by casting scorn on the stats – and normally, the people who do this, like Jennie did yesterday, don’t bother to provide any evidence of why those stats or statements should have scorn poured on them (albeit a difficult job on twitter when you may not be near a computer).

    I see it happen on CiF all the time – those who don’t agree the the fundamental argument neatly sidetracking it into a minute discussion about stats – its pointless, and allows them to get off addressing the meat of the argument.

    Jennie, did to be fair, address the meat of the argument and we had more or less all agreed that we didn’t think that pre watershed lap dancing was really suitable on EastEnders , especially without any ‘discussion’ of the issues around lap and pole dancing – we had moved beyond a playground game of attack the stats and on to some basis of agreement about the substantive issue.

    You can make stats say whatever you want them to say and they don’t provide the answer to everything. There is no right or wrong way to make an argument about something – you may have your preference but the world and life is full of ambiguity and your preference will not be everybody’s and may not even be the most effective.

    • Andrew Hickey says:

      Jo, it was precisely to avoid being ‘defamatory’ that I *didn’t* name you, or the issue in question. I would then have had either to repost the entire tedious discussion or to actually misrepresent what you personally were saying, rather than making a more general point about a whole line of argument. This is specifically *not* about the actual issue you brought up, but about a general principle of which it appears to me your arguments were a special case. I posted the way I did so I could safely generalise while knowing that anyone who wanted to check could look at Twitter and see everything that everyone said, and see your own words without me having to dump a ton of irrelevant stuff into this post.

      “there were no stats actually used (FFS)” – and you were the one who brought ‘stats’ into it. Jennie talked about studies and made the claim that the petition said studies show one thing, when in fact (according to her) studies show the opposite. I neither know nor care who’s actually right on the issue. What I was talking about was an attitude which I believe you’re also displaying in this comment (but people can judge for themselves), but which even if I’m somehow persistently misreading you *has* been displayed by other campaigners, of saying it doesn’t matter what the facts are so long as the soundbite is good.

  4. Zom says:

    I have no idea about the specifics of this debate but this …

    There is no right or wrong way to make an argument about something

    Assuming I’m understanding the intent correctly, is bloody infuriating. As far as I’m concerned there are absolutely right and wrong ways to make an argument. Constructing arguments on lies, isn’t okay for example, or building arguments on knee jerk emotional reactions like disgust, or ad hominen arguments, or arguments built on logical fallacies, or arguing in bad faith… really, I could go on and on and on and on.

    All of the above might help you win in a fight, but that’s not enough in my book, and it shouldn’t be enough for anyone with anything approaching intellectual and ethical standards.

    • Andrew Hickey says:

      Exactly. I sincerely hope that’s *not* what Ms Christie-Smith meant, but if it is then she’s very, very wrong.

  5. Simon Jerram says:

    Well put. I like this post.

    I will treat Zom’s addition as a caveat, because you make an excellent point.

    I am afraid to confirm your summary while lightweight does accurately convey the substance of that argument.

    • Andrew Hickey says:

      I’m glad you thought so – I *thought* it did, but I was also terrified of misrepresenting Ms Christie-Smith (one reason I filed the serial numbers off), and since she thought I *did* misrepresent her I was beginning to think that maybe I was being unbelievably dense.

      (And I agree with Zom, though I hope she didn’t mean actually *mean* that. Zom’s one of the more on-the-ball people I know, and you can generally assume I agree with him unless I specifically say otherwise.)

  6. even if I’m somehow persistently misreading you *has* been displayed by other campaigners, of saying it doesn’t matter what the facts are so long as the sound bite is good.

    That’s it – you are persistently misreading me. I am not saying that the sound bite is the most important thing but not to get sidetracked by a my ‘study/stats are more right that yours debate’, which one can rarely win. Deal with the substantive argument – not a parallel one about studies.

    But enough – you can keep on thinking what you like about what I’m saying, I’m clearly not communicating what I want to say effectively and its time to move on.

Comments are closed.