Excellent article. Only a few weeks ago we had newspapers going mad over a massive % increase in the tax burden on the lowest paid, failing of course to realise that before the USC, the lowest paid paid little or no tax to begin with!

I am 94% confused and 6% bored…….

excellent and timely piece. I have always been of the opinion that journalists should do stats in uni considering how intergral they are to modern journalism and how bad they are understood.

Very interesting article.

George Orwell demonstrated in his novel 1984 how meaningless statistics can manipulate people. Public relation and advertising companies have made fortunes using statistics to help their clients rip off people. Don’t forget that this slot machine will pay out 75% of the time.

It’s a well known fact that 63% of statistics are made up on the spot…

I am 94% confused and 7 % bored, in comparison to the last contributor!

The HIV example assumes there are no false negatives. Is that right?

In all fairness, while scientists and other professionals might get the stats wrong very occasionally in scientific journals, it’s basically irrelevant in context of the ongoing dog’s breakfast which is the reporting of any kind of numerical information by regular journos, most of whom seem hopelessly innumerate.

Im in 110% agreement with this article…

My head hurts

New subject for students……..”statistical engineering”

%ages over 100 have always irked me

I’m corrected. 0.01% is 10,000 times 0.0001 = 1.

My favourite is when new sports boast about their rates of growth.
If you start a sport,and 4 people join that week,your growth rate is a massive 400%/week,allowing the 5 of you to briefly claim yours is the fastest growing sport in the world!

One type of journalistic error I often see is the simplistic reporting of performance against expectation. For example, reporting that tax revenues were slightly better than Department of Finance estimates sounds good, but it doesn’t really tell us much. The reports often don’t tell us whether the estimates themselves were high or low (perhaps deliberately low) to begin with or whether the revenues were higher or lower than last year – which would be much more informative.

Another area the newspapers often confuse is changes in rates of increase or fall. For example, a fall in the rate of GDP growth from 2% to 1% might be reported as a fall in “GDP” when in fact GDP has risen.

Great read.

83.25% of these comments relate to how the above article represented itself.

If they used percentage points in a lot of situations it would help clarify things. ie 3% to 4% is a 1 percentage point increase.

ECB increases mortgage interest increases by 1%.
If the rate goes from 3% to 4% that’s a 25% jump. You WILL pay 25% more on your loan – if only it was 1%.

Thank G Charles I see everything in a different light now…….

Good article, there’s a good TED talk here (“Battling bad science”)about the same thing, although in some cases far more sinister applications:

http://www.ted.com/talks/ben_goldacre_battling_bad_science.html

Excellent article,

Nothing is more annoying than the accidental misapplication of statistics and / or the cynical misuse……..except perhaps the complete lack of understanding in the general public. (of which of course they are not to blame)

It should be included in a standard education, outside of science degrees. Anything that highlights this more is to be welcomed.

Good job.

I’m reminded of Disraeli’s famous quote about lies, attributed to him by Mark Twain.

Great article.

This is a really good example. Statistics can be used as evidence in a way that is quite misleading.

At least two hundred per cent of readers know that this headline isn’t true

If you have two loans for the same amount one at 3% for 30years and the other at 4% for 30years, you will pay 33% more for the loan at 4%.
After 20years the 4% loan has cost 21% more, after one year the 4% loan has cost 0.97% more…..
Boom, the power of compound interest!

Stats are wrong in article HIV test is 99.99% accurate. You test 10,000 people. Test gives 10 people with positive HIV test. HIV Incidence is 1:10,000. Thus 9 out the 10 are free of HIV. So the chances of having HIV Is only 10% WITH a positive test.

Ironic title on the piece (then again, there are probably more readers for “We’re all being misled” than “We’re almost all being misled”:)) but a good point very well made, David.

I know I’m coming to this debate long after the party is over but just seen it now and why not?

Let’s assume that FF calculated Lab received an average of 60 minutes and FF 20 mins. The following are valid percentage calculations and statements:

1. Lab have 200% more air time than FF ((60-20)/20 x 100)
2. FF have 66% less air time than Lab ((60-20/60 x 100)
3. Lab have 300% of the airtime of FF (60/20 x 100)
4. FF have 33% of the airtime of Lab (20/60 x 100)

It is not correct, I believe, to state that the statement from FF ‘Fianna Fáil is now getting more than 100 per cent less access to Prime Time than the Labour Party in the same position.’ is clearly wrong (I know you haven’t done so). It is a contorted and contrived recasting of statement no. 1 (FF have 200% less air time than Lab) that is misleading as that comparison you would normally expect to see would be as per statement no. 2. There are no absolute rules when comparing two numbers as to which must be the denominator in the percentage calculation, however there is custom, practice and expectation and clearly they have not adhered to that here (your reference to a faux pas).

However I disagree with you that the underlying problem here is innumeracy. I believe that the problem here is the cynicism of spin. They wanted a bigger number (200% v 66%) so they went for it as it suited their message. They are relying on the general innumeracy of the wider population to get away with it.