Sunday, 11 December 2016

Lies, Damned Lies...and Statistics

As anyone who follows me on Twitter will already know, most of my tweets these days involve shameless, predictable NQS blog post promotion. Yesterday though, I got a bit out of character. It all started with a BTO tweet that appeared on my Twitter feed, and is reproduced in the screenshot below. I'd like to ask a favour please. Before you plough on through the rest of this post take a few moments to digest the BTO tweet. What does the tweet tell you? How do you read it? What message do you think the BTO is trying to convey here? What is the motive for posting such a tweet?

You get the point. I'm just asking for a bit of analysis. I'd be grateful to hear what you think. Anyway, here it is:


Right then, here's what I thought...

First off, I distrust numbers. 23% eh? How exactly did they measure the 77% that were not brought home? That was my initial reaction. A cynical one. I couldn't access the 'bit.ly/2h5ta1P' link so simply replied to the tweet and asked my question. I received a very quick and helpful reply, thus:


However, before I go on I must mention something else that I thought on reading the original BTO tweet. I thought we were talking about birds. After all, we have a photo of a cat and a bird, the BTO is an organisation centred upon birds, and so surely what this tweet is conveying to us is the concern that for every one bird that is brought home, another three are killed and left. Oh, and I also thought we were talking about British birds (British Trust for Ornithology?).

Yes, I made assumptions. That's why I am curious how others see this tweet. Incidentally Mrs NQS read it exactly the same, without any prompting from me.

Anyway, in another tweet Becky Thomas kindly added that she had given the presentation at the BTO Conference, in case I had any other questions. I replied thanks, but that I'd read the research paper first and come back to her if necessary.

So I read the research paper. Fascinating stuff. And how did they know how many victims the cats brought home vs didn't bring home? By mounting little cameras on their study cats! Of course, why hadn't I worked that out myself?! I read on...

The study took place in the state of Georgia, USA. Not Britain. The cats involved were ordinary pet cats who were allowed outdoors, in various habitats ranging from urban to rural - 55 cats in total. Data were collected over one year. The cats were divided into four groups of 12-15 each, and each group monitored for just one season, i.e. spring, summer, fall or winter. Of those 55 cats, the paper said that 44% displayed hunting activity, i.e. stalking, chasing, etc. As 44% of 55 = 24.2 I'm assuming that 24 of the 55 cats displayed hunting activity. Of those 24, just 16 made one or more successful kills. The total number of creatures killed was 39, all of which were identifiable thanks to the cameras, the vast majority to species. Of those 39 kills, just five were birds. FIVE!

Look again at that original tweet! The tweet is all about birds isn't it?

Ah, but is it? What the research paper does not do is tell you how many of those five birds were brought home; it simply tells you the percentage of prey brought home. Which, in fairness, is exactly what the BTO tweet says too. It's not how I read it though! You?

So off I go again, tippy-tapping away on Twitter like some deranged pedant...


And there we are, the latest storm in my little teacup. I'll be honest, that BTO tweet annoyed me. I thought it was inflammatory. Everyone knows the cat/bird issue is emotive, and I couldn't see any realistic purpose to this tweet beyond stirring things up. Why do that? And after reading the research paper upon which that 23% is based I also feel the tweet is disingenuous.

The BTO tweet was sent on 3 December, and between then and now it has been retweeted 72 times. I wonder what it's achieved. No doubt it has reinforced the views of those who think cats are instruments of Satan, and it's wound up one or two number pedants like me, but what else? We are bombarded every day with statistics designed to uphold a viewpoint or justify some course of action. When those numbers support our own opinion we love them, and when they don't, we point out how they've been twisted and misapplied. Yet nearly all of them are not to be trusted. That tweet does not say reassuring things about other numbers the BTO might publish...

Naughty BTO.

Oh, and three of those 39 confirmed kills were worms. Fact.

19 comments:

  1. Worms!!!! Why didn't you say Gav??? This is far more serious....

    In all honesty, tweets and Facebook posts are diluted by the banal, the misinformed, the delusional and the downright wrong.

    Goodnight...

    ReplyDelete
    Replies
    1. Also a butterfly, 2 dragonflies and 8 Carolina Anoles, in case you were wondering.

      Delete
  2. The way I read it Gav, was that no matter how much we try to fill up kittykins with go-cat. The chaps are predisposed to go out killing things, possibly birds.
    I thought it was about getting rid of cats.
    Funny how such a tweet had it's origins in the USA. After all, I don't think domestic cats had much to do with the demise of the Passenger Pigeon and Eskimo Curlew.

    ReplyDelete
    Replies
    1. Oh yes, Ric, you've reminded me - the very first thing I thought about the tweet was that it was unashamed anti-cat propaganda.

      Delete
  3. I agree, very naughty. Ornithology is a long word and implies expertise. This is yet another reason why we should not trust experts. And on top of all the other small reasons, is why people are distrustful of experts and shows why groups that genuinely have something important to say have unfortunately buggered it up and allowed people who are somewhat less expert [ie cretinous pillocks] to rule the proverbial roost. Not impressed. But if you decide to play fast and loose with facts then in some sense you deserve all you get.

    ReplyDelete
    Replies
    1. Thanks Jono, disappointing isn't it.

      Delete
  4. The thing that stands out to me is that if 55 cats killed 39 prey items over 12 months, it means any one 'average' cat is responsible for 0.7 kills a year. Truly horrifying!

    ReplyDelete
    Replies
    1. To be fair, Karen, each cat was monitored only for a three-month 'season', so we're actually talking 0.7 x 4 = 2.8 kills a year, of which 0.5 were birds, 0.3 were worms, and 0.8 were a North American lizard thing. Clearly these are stats which can readily be applied to UK cats.

      Delete
  5. I think it's simply a case of the wording and the pic being used craftily as 'clickbait' to entice people to use their link. People who like cats would perhaps click it. People who don't like cats would perhaps click it. People that wanted to see what the 23% was specifically about would perhaps click it. People who wanted to know how much of this percentage was birds would perhaps click it (after all it was a tweet from the BTO as you rightly pointed out). The Tweet could potentially 'appeal' to a wide variety of personal interests.
    Personally I didn't take it as an inference that the stats were purely or even primarily about birds as that isn't what it says. However I can see why people might assume that; provocation/misleading is sadly very common in media (something which we probably wouldn't want to see from the BTO, as you pointed out) as it results in media engagement. Perhaps on a more important note though, the stats have been generated from a tiny sample size (I'll ignore the location difference which is also of significance of course). A sample size of 55 cats when the total population is many millions is obviously going to produce a VERY unrepresentative set of data in terms of what the whole population may hunt like. So I definitely agree with you here; why the BTO would want to associate itself with such data is a little concerning.

    ReplyDelete
    Replies
    1. 'Clickbait'...I hadn't thought of that. I cannot access the link, so don't know where it leads. Tim, thanks for your comment - you've picked up a couple of things I hadn't really thought about, including the tiny sample size. In addition, that 23% is derived from a sample of only 39 dead things; even my schoolboy stats knowledge tells me that statistical significance requires a minimum of 40 data points. Not clever.

      Presumably the BTO Twitter account is in the hands of just a few individuals. I wonder if they realise the responsibility that goes with handling that. Surely they don't really want to be winding up and putting off potential recruits like me with such ill-judged output?

      Delete
  6. Oh dear, I see this as a somewhat misguided rantette typical of the blogosphere. You manage to hammer the plank to smithereens and completely miss the nail. There are estimated to be 8 million "domestic" cats in the UK, 30 times the number of the much vilified fox. Most of these cats are urban and we really should try to understand their effects on the already pressurised wildlife in our towns and cities. The only other predators I see in London are raptors/owls/gulls and a very few foxes. Dogs rush around and disturb nesting birds and birders but don't predate. Most of the foxes seem to live out of bins.
    Now BTO tweet a factoid which is accurate but carries a cat looking at a widespread urban bird through a window. Maybe it was the bald 23% that aroused your curiosity. Very good rounding up might have been sensible in view of the numbers, but that would have been manipulating the data. You then discover the research was done in the US using a rather novel method, no problem surely, it is a world wide problem after all (I could ref. Oz here). Not only that but it isn't just birds cats predate but other beasts too - shock, horror! Not a problem again, but you felt misled? I don't think this invalidates the BTO using the tweet - clickbait in a good cause?
    You then attempt to belittle what is a perfectly honest piece of research by that much maligned sector of society who others like to term term experts. This is a real attempt to collect data and use it to understand the foraging behaviour of no doubt well feed pet cats. Good on them, it is a much needed area of research and shame on you for knocking it. Good on Becky Thomas for responding in such a helpful manner. She is doing good work, as an expert!!
    Although personally I don't like animals being patronised as pets by humans I can see it does some good for some. It is balancing that against the potential harm that we need to assay. We are going to need all the help we can get in protecting the environment against the current yahoos, the more facts the better. Well done BTO.

    ReplyDelete
    Replies
    1. Well, I did rather stick my chin out didn't I.

      Trouble is, I've been home ill for a few days, bored, with time to fritter on getting righteously indignant about nothing and turning it into a blog post. Is it just 'a somewhat misguided rantette'? Perhaps you're right. But at least it has elicited your delightfully robust rejoinder. Thank you, Bob, for taking the time to comment.

      Yes, I realise cat predation has got to be a massive conservation concern, but that wasn't my issue. And I'm disappointed you thought I was attempting to belittle the research carried out in Georgia. On the contrary, I thought it was fascinating and am surprised there wasn't a similar, ideally more extensive study from the UK that could have been referenced. To have misled you so suggests I could have worded my post a lot better. Sorry about that.

      No, my issue was with the BTO's spurious (as I saw it, anyway) use of the results of that research to embellish what I can still only see as a somewhat inflammatory tweet. As I said in the post, clearly I made various incorrect assumptions about the tweet, but that fact does not mean I didn't feel hoodwinked. The disingenuous use of numbers to influence thinking is an everyday part of life which annoys us all no doubt, but I think the BTO should be conspicuously above all that.

      I am not an 'expert' in any field. Most of us aren't. But we are the agents whom the 'experts' need to recruit in order to realise the fruits of their labours. The BTO is just one of countless bodies needing that support. Stuff like this tweet might seem to you like trivial nonsense, but I'm afraid it just puts me off...

      Delete
    2. Maybe the problem may be that 140 characters isn’t really enough room to say you all might ideally want to say, particularly in reference to facts and caveats? Then again, being extravagant with facts (or even with complete lies!) seems to be a part of modern life that for whatever reason seems to be increasingly mainstream and acceptable, and it’s this which I have a problem with. I agree with Gavin that honourable organisations like the BTO should be above all that, they should be the last bastion. It’s unfortunately far too easy to bash out a tweet and think nothing more of it – fire and forget – and there is rarely any kind of meaningful come-back option.

      Now had it not been for this tweet, retweet and blog, I would have been none the wiser as to any of this, but I would still be thinking that domestic cats kill a lot of birds but without really having a great deal to back that up. So I think the robocat approach to try and get some good data is an excellent idea, but I also think it is difficult for the BTO to legitimately use this particular set of data to make any kind of point in the UK. Now you might say that a cat is a cat is a cat (unless it is in a box of course, then who knows?) but I’m certain having been there that urban USA is physically very different to urban UK, and I actually suspect that it is much easier for a cat to knobble a bird in the UK than it is in Georgia – larger open spaces, overall “cleaner” gardens and property boundaries, fatter less nimble cats that drive everywhere.…

      Delete
  7. Come on Bob, reel your neck in.
    Gav simply asked us to offer our first impressions on some unsolicited twitterins.
    Just because you don't like what you have imagined to be the reasoning behind the comments, there's no need to go flying off the handle.
    I've no idea why you feel this way, then again, I'm not a doctor.
    Btw Gav, how's that lovely great ginger tom cat of yours. Last time I saw him he was having a snooze on your sofa.

    ReplyDelete
    Replies
    1. Ric, for several reasons Bob's comment was one of the best I've received in a long time (no disrespect to others intended here - I am grateful for them all) and I welcome any like it. As long as they're not peppered with profanities I couldn't care less about the tone either.

      Billy died getting on for three years ago; we've been catless for a while. I think Sandra is still in mourning...

      Delete
  8. Fair enough Gav, but shame about Billy.

    ReplyDelete
  9. Quick update. Interestingly the BTO has been in touch. They wished to comment on this post but it seems the 'comment' facility on Blogger has a character limit. So the comment has been emailed to me. I'll turn it into a full-blown post, hopefully within the next couple of days.

    ReplyDelete
  10. Certainly thought it was pretty interesting to get a number like that, 23%, and worth knowing, from what I recall when I read the original tweet I didn't make the link to birds myself and suspected many of prey would be small rodents, I guess partly because of seeing cats abandon them over the years (as well as dead birds)...

    ReplyDelete
    Replies
    1. It's quite likely I am in a minority in thinking the tweet implied the 23% referred to birds. Certainly I read it in a way that wasn't intended. However, a photo of a cat looking at a mouse, say, would easily have prevented me from making that assumption. And this post would never have been written...

      Thanks for your comment.

      Delete