How to win fiends and influence people

August 18, 2012 § 1 Comment

Yesterday I was handed a free copy of the Daily Telegraph.
They’re giving them away!

I have read one article and it’s an example of poor statistical reporting. There is a LOT of this in newspapers. I’m picky about the communication of data because I have done a fair bit of it in my line of work.

The article, which takes up less space overall than the accompanying naked torso of James Magnussen, is Facebook no friend of Games athletes. It’s a sloppy piece of reporting and here’s why.

The whole premise of the article is that our athletes’ reliance on social media is at least partly to blame for Australia’s poor performance at the Olympics. Putting aside the issue of whether that (the poor performance) is even true, let’s look at the evidence.

The argument is based on a survey conducted by a third party. There’s no information about sample size or demographics.

The article states up front that ‘A study has revealed that Australians believe social media has become a significant distraction for our top athletes.’ It says that the survey found that this was one of ‘the top issues’. But it doesn’t provide any data on that. At all. There is a quote from an athlete and from her coach and a statement from the survey firm and that is it. We aren’t given any survey data about social media and Olympic performance.

Here’s what we are told though: ‘Just one in six Australians believed elite sport was underfunded’.

Next, the journalist tells us that ‘In contrast, nearly two in five thought throwing extra cash at athletes was not the answer…’

In both statements the writer has chosen to report the exception, not the rule. That’s fine, if that’s the story. But both also express the minority view in connection with a negative statement. The reader has to work really hard to understand what is going on, and be good at fractions.

It’s really saying: ‘Most people thought elite sport was overfunded. In contrast, most people thought extra funding is the answer.’ What?

Note that none of this confusing barrage of data has anything to do with the title or point of the article, which is that SOCIAL MEDIA STOLE OUR MEDAL TALLY! RARRR! Daily Telegraph readers will see the headline and the (irrelevant) data and go away convinced that ‘Facebook is the newest threat to our Olympic success.’ Fiendishly clever.

The article fails another test by using too many different ways of representing the numbers. In just 370 words, we are thrown fractions, percentages and gut feelings. We are told about ‘one in six’, ‘two in five’, ’11 per cent’, ’60 per cent’ and even ‘6.15 out of 10’. We are told that one thing is ‘significant’ and another is a ‘top issue’ without ever finding out what the numbers are. The data that we are given tells a different story to the message we get, which in turn has nothing to do with the headline.

In conclusion, I’m giving this article zero stars. You can be 100 per cent sure of that.


§ One Response to How to win fiends and influence people

  • Anne Powles says:

    You are so right Leonie. Something has gone severely wrong with the way statistics are promulgated at the moment. I will put up my hand to the use of the odd “there’s a tribe” or “60% of people” in a light hearted argument, but when people try to allege “survey” these should be properly done.

    I do not want to be a luddite or a dinasaur and the advent of statistical research programs on computer has been an absolute blessing and time saver. But at least in the old days in the study of Research Methods with (yes speak softly) a slide rule, then a calculator which was a great blessing, we did at least learn what statistic we were using, what we were aiming for and what sized sample we needed to at least make the figure vaguely meaningful.

    There was very little more wonderful that that first time I fed my raw data straight into a main frame computer and the answers came out without my having to do a single sum! But it has become habitual now to allege a survey without providing proper information (such as even sample size to begin with). Some so called scholarly articles do not even put their figures in the article any more, merely their conclusions.

    And if you have ever been the victim of completing a survey on line the standard of questions is appalling. The “yes” and “nos” are often not expressed as logical opposites. There is little room for “grey” answers and if you try to tick more than one box the first tick disappears!

    I do not advocate doing it all by hand again. That was pure torture, but somewhere in the crowded educational curriculum, at some stage, must be the “how to” of statistics, their meaning and their interpretation.

    Anne Powles

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

What’s this?

You are currently reading How to win fiends and influence people at lightlyskipping.


%d bloggers like this: