'What British Muslims Really Think'.

[ThirtyNineWeeks]

This Ch4 documentary can't possibly be representative of all British Muslims when only 1081 were polled......can it?

AIBU to think that, for some, this kind of documentary will only ever be seen as evidence of 'Islamophobia'?
And the fact that it is made by Channel 4 will immediately turn lots of folks off..

I was just about to ask whether they'd asked all of them?
And whether they had let Ahmedis contribute?

Well it will be interesting to see what Trevor Phillips says I think.

I've never found Trevor Phillips or his thoughts particularly interesting tbh

What about the thoughts of Muslims?

This survey's methodology has been comprehensively torn apart already. It deliberately targeted areas with large Muslim populations and more deprived areas and was deliberately unrepresentative in other ways too. Geared to make anti-Muslim headlines and nothing more. Do not believe a word of it. Bear in mind this is the channel that gave us Benefits Street.

This is the TV version of clickbait. Never mind the content, feel the controversy and the publicity. I bet it's making lots of C4 execs feel all brave. That it might well stoke up hostilities that will be taken out on actual BME people and/or Muslims/white Britons won't concern them.

No one asked me.

I have worked in consumer research for years and I would never go so far as to say 'What British Muslims Think' on a sample size that small. I don't necessarily think you need to have a census of all of them but you need to have a truly representative sample.

Agree it's TV click bait. Awful awful awful.

Never mind the content, feel the controversy

Never mind the quality, feel the width ...

Horrible and inciteful. Especially in this tense and horribly racist climate we are living in at the moment. Why don't they ever poll the thousands of Muslim doctors, lawyers, surgeons, graduates and professionals? I work with lots of Muslims and I feel so sorry for them having to endure what they do.

The title is not What All British Muslims Think. However, I am told that, for a sample to be representative, 5% of the subject has to be polled; that's probably around 136,000 Muslims. I'm just concerned that what could have been an illuminating documentary has been invalidated by the sample size. Surely they'd already considered this?

However, I am told that, for a sample to be representative, 5% of the subject has to be polled; that's probably around 136,000 Muslims.
Um, no. How reliable a poll is depends on the size of the sample. A sample of just over 1,000 would give you a margin of error of about 3.5%, with 95% confidence - in other words, if you repeated the poll, with a hundred more sample groups, 95 of those 100 groups would give you answers within 3.5% of the answers you got the first time round.

The selection of the sample is another matter: I'd want to know if had been normalised by age, sex, educational level, domicile and income to represent a true cross-section. If this poll hasn't done that, I'd be suspicious of its findings.

Thanks, Lumela - the 1/20 idea is ridiculous - on that basis you would need to ask 65m to take part in order to poll the Chinese! opinion polls are often based on around 1000 participants.

The obvious issue for a sample like this is that the people around in the day to fill out the survey/be interviewed, are less likely to, eg, work full-time, and this will skew the results.

Christ I can see the DM article on it already..

I know about sample selection and all that, but this is what made me start the thread and I feel stupid now :

'In statistical hypothesis testing,[1][2] statistical significance (or a statistically significant result) is attained when a p-value is less than the significance level (denoted α, alpha).[3][4][5][6][7][8][9] The p-value is the probability of obtaining at least as extreme results given that the null hypothesis is true whereas the significance level α is the probability of rejecting the null hypothesis given that it is true.[10] Equivalently, when the null hypothesis specifies the value of a parameter, the data are said to be statistically significant at given confidence level γ = 1 − α when the computed confidence interval for that parameter fails to contain the value specified by the null hypothesis.[11][12]

As a matter of good scientific practice, a significance level is chosen before data collection and is often set to 0.05 (5%).[13] Other significance levels (e.g., 0.01) may be used, depending on the field of study.[14] In any experiment or observation that involves drawing a sample from a population, there is always the possibility that an observed effect would have occurred due to sampling error alone.[15][16] But if the p-value is less than the significance level (e.g., p < 0.05), then an investigator may conclude that the observed effect actually reflects the characteristics of the population rather than just sampling error.[1] Investigators may then report that the result attains statistical significance, thereby rejecting the null hypothesis.[17]'

Lumela please explain this to me like a 5yr old.
Surely the size of the sample will dictated by the overall subject total? If as I think 5% is the minimum sample required, then I'm confused by your statement, how reliable a poll is depends on the size of the sample. Aren't you both saying the same thing?

I don't understand the stats stuff but without understanding it, I do understand this is a small snap shot of a huge community.

Therefore the views will not represent the whole community, and that stands if the views of the snap shot all want peace and love.

Aren't you both saying the same thing?
Nope.

Think of flipping a coin: you can get a run of head, head, head, head, head relatively easily, so from a sample of five you could conclude that the coins are weighted to fall heads-up. But if you carry on flipping, you will get more tails and, if the coins are not weighted to fall one way preferentially, you will, fairly soon, have about half heads and half tails.

It's the same when you talk to people. Say you want to ask people's views on gay marriage. By fluke, your first ten people pounced on in the street might all be in favour and you start to think you've hit a very liberal town. But the more people you ask, the more representative your sample will be, and the less prone to chance. Talk to a couple of thousand of them, and you will have a very accurate view of what people in that town think - especially if you have carefully chosen your sample by age group, income, education etc to reflect the population of that town as a whole.

Consequently, it's the absolute size of the sample that matters, not the relative size compared to the population. Various people who can think in numbers and symbols have worked out the margin of error and the confidence level of various sample sizes. See here for info on margin of error. This sort of stuff is used all the time in audit, for example: it means auditors will only test a sample of transactions rather than redoing them all. And when people are enrolled in health studies - look at 10,000 smokers for ten years, and you'll get a very good idea of what goes wrong with their health. If you're a smoker, it won't tell what WILL happen to you, but it will let you judge how likely it is that something MIGHT happen to you.

I hope that's clear... As I said, I was crap at maths.

PS Should have said, if you poll 10,000 people from a population of 100,000, your result will not be far astray from a poll of 10,000 people from an identical population of 100,000,000.

So the statisticians tell us, anyway. And I believe them.

Followed by the programmes
What British people with beards really think
What British people who live at number 18 really think
And
What British people with six toes on each foot really think...

chair , popcorn
There will be fireworks!

Witch I am sure we have already had a slew of - "what so and so " think.

However I think its a relevant and important question in today's climate, it may not be giving the perfect snap shot as discussed, but it is an important question.

I just wonder if the programs title couldn't have been framed more diplomatically and truthfully in light of the study.

It isn't just about sample size, it's about a lot more than that: analysis here.

But of a daft comparison there, Witchend.

*Bit