Swimming with the tide....

[Julia9054]

On a serious note, no honestly, what drives someone to partake in a survey?

I've never had the inclination to do one either on-line or by post and as for 'in the street' I avoid all unsolicited contact with 'clipboards ' so just wondering really?

Someone once stopped Al and I in the street and gave us 24 cans of free beer. There were 3 different types and we had to fill in a quite detailed questionnaire afterwards. That's the sort of incentive that works for me!

 

[Bromptonaut]

On a serious note, no honestly, what drives someone to partake in a survey?

I've never had the inclination to do one either on-line or by post and as for 'in the street' I avoid all unsolicited contact with 'clipboards ' so just wondering really?

I was phoned a few years ago by one of the agencies running a public opinion survey. Was at home on my own and bored; thought why not.

Was asked to provide details of my sex, relationship status, housing, work and income bracket. Moved around politics, both my current inclinations and voting history (instinctively left but have voted tactically for Libs), papers I read etc.

I suspect some info was relevant but other bits might be weighting for my value.

 

[Rusty Nails]

I notice a couple of contributors have had the honesty to admit they don’t know enough about the subject and/or don’t have all the facts (questions etc) on this particular survey. Similarly my “O” level statistics is not upto the task of sample sizes etc.

Politicians are frequently accused of being out of touch with ‘public opinion’, perhaps, they rely on surveys too much?

It just occurs to me, how many such surveys actually get an answer which is “right”, whatever “right” may be, is the best measure of their usefulness.

My personal life experience does not match with the survey conclusions, but, I readily accept that, perhaps, my circle of acquaintances is not representative of the Nation.

I doubt whether any one person's acquaintances are representative of the nation. That is why such surveys can be both important and informative.

 

[Craig the cyclist]

I once took part in a survey about the Post Office.

As a result I was recruited to be a 'post-tester'. Every week, sometimes several times a week I had to post things and get things delivered. Sometimes they were easy tasks like post a letter on Tuesday, others were quite complex tasks, like wait for the post to arrive, log what time I received it online, then transfer the clever gadget inside which logged its journey through the system to another type of envelope, write the onward address poorly or with words missing and post a matching letter with the address clearly written in two specific post boxes to be picked up at a specific time.
As you did more the asks increased and I got as far as sending envelopes that were ripped and sending boxes that were strange shapes! I ended up doing it for nearly 2 years.
They paid us in stamps, lots of stamps, and there was a £50 cash draw each month, which I won twice.

 

[BoldonLad]

Someone once stopped Al and I in the street and gave us 24 cans of free beer. There were 3 different types and we had to fill in a quite detailed questionnaire afterwards. That's the sort of incentive that works for me!

You drank 24 cans, between you, then answered some questions……….

 

[BoldonLad]

I doubt whether any one person's acquaintances are representative of the nation. That is why such surveys can be both important and informative.

True, if they (the surveys) are in fact accurate. How would we know?

 

[Julia9054]

You drank 24 cans, between you, then answered some questions……….

We were students at the time. And may have spread it over two evenings!

 

[Rusty Nails]

True, if they (the surveys) are in fact accurate. How would we know?

With reputable surveys all the technical information about the sample including confidence levels in their accuracy should be available. Nothing in life, other than death, is guaranteed, but the information provided can be valuable.

I did a lot of work with the ONS about 12 years ago and was impressed by the effort and seriousness they put into the accuracy of their surveys but cannot vouch for the work of other polling organisations.

 

[BoldonLad]

With reputable surveys all the technical information about the sample including confidence levels in their accuracy should be available. Nothing in life, other than death, is guaranteed, but the information provided can be valuable.

I did a lot of work with the ONS about 12 years ago and was impressed by the effort and seriousness they put into the accuracy of their surveys but cannot vouch for the work of other polling organisations.

I don’t doubt that a lot of effort and seriousness is expended, but, my question was, “are the results accurate?”. I am not saying they aren’t, I am saying, how would we know?

 

[Fab Foodie]

I don’t doubt that a lot of effort and seriousness is expended, but, my question was, “are the results accurate?”. I am not saying they aren’t, I am saying, how would we know?

It depends what you mean by accurate!
Thing is, any given survey is only a snapshot. Some major event could happen the day after the survey was taken (like the Queen dying) and the results could change overnight.
Polls, like other research, are a starting point, a marker an indicator. More often that not other surveys are carried-out that might start to show similar trends. Get enough surveys and you can start to do meta-analyses which improves confidence in the data and sometimes discover new connections.

I don't profess to be a stats master, but know several who are, and it's a bloody complex business!

However, the Govt. has just been pulled-up over Moggs Imperial measures yougov. poll, which even a rank amateur like me can spot is heavily biased, so scepticism is well placed....

 

[Rusty Nails]

I don’t doubt that a lot of effort and seriousness is expended, but, my question was, “are the results accurate?”. I am not saying they aren’t, I am saying, how would we know?

A survey of significantly less than the total population can never be guaranteed as 100% accurate.

Surveys of particular sizes compared to the total population, if they are completely random or appropriately stratified, will all have a defined level of confidence in the accuracy of their results e.g. they could say that they are 95% accurate with a margin of error between + or - 2%.

Either you can accept the theory and science behind the reliability of sampling or you can reject it for whatever reasons you choose, but it is a fact that it is used to underpin work in vital fields such as research in scientific, medical and social attitudes and it has been shown to work.

If the survey does not make available for scrutiny the methodology and statistics behind it then obviously you are less able to know if it is accurate.

 

[Deleted member 49]

Surveys 🙄

 

[winjim]

You can question the results and the methods of surveys but you have to consider how much of the answer you're qualified to understand. In the past few years everyone's suddenly become an expert on international trade, epidemiology, and tactical urban combat. This is why I say it should be the job of the journalist to scrutinise and interpret the survey.

The data are undoubtedly flawed, as all data are, but that doesn't mean they are useless.

 

[All uphill]

I don’t doubt that a lot of effort and seriousness is expended, but, my question was, “are the results accurate?”. I am not saying they aren’t, I am saying, how would we know?

One approach is to repeat the survey using a different sample group. If the two independent surveys results are similar that will increase your confidence in the results.

 

[Fab Foodie]

Some useful picking apart of the report here in this opinion piece...

https://www.theguardian.com/comment...r-more-divided-too?CMP=Share_AndroidApp_Other