“Labour voters were overwhelmingly supportive (65%).”

Shame the same can’t be said of Labour MPs.

“52% of the public oppose calls by the education secretary Michael Gove asking parents to volunteer to help on Thursday in order to keep schools affected by strike action.”

Depending on the confidence level of the poll, that 52% you wave about could be anywhere between 49% and 55%. So isn’t it better to say “roughly half”? I think it is.

Therefore your headline should be: “Roughly half of parents polled reject Gove call to keep schools open – and roughly half accept.”

I mean, jeez, I agree with you folk on this site – I really do. But the sheer stupidity that comes out with analysis of survey results (particularly over the past few days) makes me want to tear up my maths degree and accept that the world (left and right) just don’t care about accurately reporting numbers.

I expect it from the Daily Mail, but…

Please, please – for my sanity if nothing else – take a remedial stats class!

@ David

“Therefore your headline should be: “Roughly half of parents polled reject Gove call to keep schools open – and roughly half accept.””

Just to prove that I’ll use stats to back Sunny as well as criticise him – you’re right about the first part but wrong about the second. The number of people who “welcomed” the proposal was polled at 37% – so, using your error margin, that’s 34-40%. “Roughly half” would be pushing it.

@Chaise Guevara

Fair point – I apologise – scrub the second part. (I ranted after reading the first sentence – before I’d even gone to read the results. Daft eejit that I am.)

However, my point about the way the headline is written stands. And my point about the general approach to stats on this site.

And let’s not forget that 12% of “don’t knows”. They’re important as well…

Presumably the 37% are having their CRB checks right now ready to cross the picket lines.

I was amazed so many parents had obtained teaching degrees to fill in for absent teachers boy what an educated society we have ?

Parents (everyone in fact, except Labour voters, and even 37% of them agree) also overwhelmingly support calls for strike ballots to have the support of at least 50% of all members.

Sunny is so slapdash today that he even understates his case.

The headline says ‘parents reject’……… etc.

Indeed they do. But the numbers Sunny chooses to cite are those for the total sample – 81% of whom do not have children of school age. The figures for parents who do are somewhat higher. Maybe they’ve looked around the school playground and taken the view they wouldn’t want some of those other parents anywhere near their kids.

@ 4 David

“Fair point – I apologise – scrub the second part. (I ranted after reading the first sentence – before I’d even gone to read the results. Daft eejit that I am.) ”

No worries! I do the same myself with distressing frequency.

“My point about the way the headline is written stands. And my point about the general approach to stats on this site.”

Absolutely agreed. Sunny seems to think that it’s cool to misrepresent data, instead of a good way to immediately lose the argument.

As sample size decreases so does the appropriateness of making inferences about the larger population.

Total sample size: around 2500.
Respondents with children at school: around 500.

Be careful when comparing percentages based on different sample sizes.

@ Chaise Geuvara – I’ll be getting a remedial class in “Reading the Full Thing First” myself. Ahem. But I think misintepretation of stats should be pointed out at all times as it can lead to horrible, horrible ends. (Dr Meadow, are you reading? http://en.wikipedia.org/wiki/Sally_Clark )

David@10: Good point about stats, people often take them at face value and try to build a case around them, only discovering later that a little research, a closer look, would’ve either strengthened their argument, or prevented their embarresment. We all know that politicians manipulate and distort figures from polls and surveys, to suit their own agenda, that’s why we need people who can cut through the fog and shine a light on half-truths and misinformation. Perhaps it should be on the curriculum, part of lessons on how not to get shafted by the system.

@ 10

“But I think misintepretation of stats should be pointed out at all times as it can lead to horrible, horrible ends. ”

@ 11

“Perhaps it should be on the curriculum, part of lessons on how not to get shafted by the system.”

Strongly agree on both counts. If we made sure things like statistics, probability theory and the scientific method were focus points in schools, the population as a whole would be a lot harder to gull.

I’m always willing to put the boot in to Sunny when he misuses polls, but in this case he hasn’t really done anything wrong. The margin of error for a sample of 2,500 is a bit under +/-2, so 52% would most likely indicate a true majority. For a sample of 500, the MoE is about +/-4.5, so 57% clearly indicates over 50% support. The MoEs would increase if YouGov had to use a lot of weighting, but it seems be one of their more balanced samples.

In any case it’s common enough to ignore Don’t Knows when working out support/opposition for something, rather than conscripting them to whatever side suits. Certainly they’re usually ignored in Britain when discussing voting intention. So you could say the poll shows the general population 58% against the volunteers and parents 64% against.

@ 13 Roger

All fair points, but I think the title is a bit out anyway. The indicator of whether parents ‘reject’ this call will be whether they turned up and kept all the schools open today (so I reckon the headline will turn out to be true anyway). You’re right that this shows general disagreement with the government’s approach here.