Sky poll demonstrates support for leaving the single market

[sasguru]

No. I can't explain the problem. You explain to us.

Why not? You told us yesterday you had studied "Pure and Applied Statistics". Did you not do very well at it?
It's a very basic issue.
I'll give you a hint, it's not the size of the sample, but something else that's wrong with it....

[GB9]

Why not? You told us yesterday you had studied "Pure and Applied Statistics". Did you not do very well at it?

Because I can't see any problem with it.

With your advanced skills you clearly do. I'm ready to learn.

[sasguru]

Because I can't see any problem with it.

With your advanced skills you clearly do. I'm ready to learn.

You don't need advanced skills, O-level knowledge would have been sufficient.
But fair enough, see if you can work out the problem.
All you need is here:
https://en.wikipedia.org/wiki/Sampling_(statistics)

[GB9]

You don't need advanced skills, O-level knowledge would have been sufficient.
But fair enough, see if you can work out the problem.
All you need is here:
https://en.wikipedia.org/wiki/Sampling_(statistics)

You have linked to a wiki page about sampling.

So what's the point you are making?

What s the error in this specific poll?

[sasguru]

You have linked to a wiki page about sampling.

So what's the point you are making?

What s the error in this specific poll?

The answer is in the contradiction contained in this statement:

"Sky Data interviewed a nationally representative sample of 1,068 Sky customers by SMS on 17 January 2017"

Do you see what the problem is?

[d000hg]

We can easily prove the UK strongly favours Brexit, leaving the single market, and a Trump presidency by a poll on CUK

[sasguru]

We can easily prove the UK strongly favours Brexit, leaving the single market, and a Trump presidency by a poll on CUK

Well, you at least, understand the technical issue.

The issue is that Sky customers with mobile phones that respond to polls are NOT a representative sample of the population.
That is what is called a biased sample, not necessarily biased towards Brexit, but biased nevertheless.
There is a lot of work being done on a technical level by polling organisations (including a client I work for occasionally) to understand why the polls are going wrong time after time and one of the issues being examined is the biasedness of the samples being used and how to get more representative samples.
But on a technical level, the paying customers of a particular media organisation would certainly not be considered representative of the population at large and some statisticians would also quibble about the use of SMS.

[GB9]

The answer is in the contradiction contained in this statement:

"Sky Data interviewed a nationally representative sample of 1,068 Sky customers by SMS on 17 January 2017"

Do you see what the problem is?

No.

[GB9]

The answer is in the contradiction contained in this statement:

"Sky Data interviewed a nationally representative sample of 1,068 Sky customers by SMS on 17 January 2017"

Do you see what the problem is?

I'm waiting.

I would really appreciate the opportunity to learn so please tell me, (all of us) what is statistically incorrect with the poll.

[sasguru]

I'm waiting.

I would really appreciate the opportunity to learn so please tell me, (all of us) what is statistically incorrect with the poll.

See above.