When is a Survey Realistic - ?

[simongt]

According to Road cc, in a recent survey, about 'half the UK population' can't afford to buy a bike for various reasons. I won't bore you with said reasons, but reading the article, the survey was taken from 2000 people across the UK from various jobs / professions.
Hadn't realised that the UK population had dwindled to about 4000 - !

 

[cisamcgu]

Polls done by polling companies are very accurate, polls done by using underpaid students to ask people in the street are b*llocks

If it is a poll then believe it, it is a survey, then ......

 

[Dogtrousers]

Here's the road.cc article.
https://road.cc/content/news/almost-half-british-people-cant-afford-buy-bike-309689
And an article on the reaction to the first article
https://road.cc/content/news/cyclists-react-half-brits-cant-afford-bike-survey-309721

The survey they are reporting on was done by the people behind the Cycle to Work scheme people. I can't find a link to the survey itself. I've no idea whether it was a well done or not. However. given that there is a general perception among the non cycling public that bikes are super expensive, and given that the prices of everything - bikes included - is going up I could imagine that a significant proportion of people might consider bikes too expensive.

It's entirely possible for a well-done survey with a sample of 2,000 to give useful data.

 

[Ming the Merciless]

They definitely won’t be able to afford to buy and run a car in that case…

 

[mustang1]

The cycle to work and similar schemes are yet another reason why bikes have shot up in price. There's inflation, there's fashion, there's golfers converting to cyclists, and there's cycle to work which makes bikes cheaper, so more people buy it and the price goes up.

 

[HLaB]

I wouldn't trust any stated preference survey. People often state one thing and do another.

 

[Drago]

People often state one thing and do another.

I have devoted my entire life to this endeavour.

 

[figbat]

So a body responsible for affordable access to cycle ownership have stated that they are expensive to buy outright?

 

[si_c]

I wouldn't trust any stated preference survey. People often state one thing and do another.

Yes, this is a known problem, however it can be accounted for and more often presents itself where there is a cost to a stated preference. For example in a survey asking for sexual preference can be difficult because young people are often known to have been reluctant to respond accurately if asked in the presence of other family members. Similarly dietary surveys amongst religious groups can get inaccurate results due to the social cost of responding accurately.

This is why as noted above professional polling companies get more accurate results as they have strategies in place to deal with these known problems.

That being said, a well designed survey, even if conducted on the street can still get very accurate results, particularly if they collect enough surveys - and around 2000 seems a reasonable amount - information that allows them to compare the outcome against what might be expected nationally.

I'd hope that there's a technical report which details the survey methodology as well as any benchmarking of the collected results.

Surveys are a fairly specialised field, and whilst I've worked with them a bit, it's been a long time since I've had to work with them professionally.

 

[Emanresu]

It's entirely possible for a well-done survey with a sample of 2,000 to give useful data.

It's the connection between probability (of being right) and sample size. So the larger the sample the higher chance of it being accurate - leaving aside the way the questions have been worded. Seem to remember that 1000 was the magic number for some sort of credibility.

If you look at some adverts now, you'll see "70 people out of 90 thought our product was wonderful". Has little statistical credibility but it keeps the Advertising Standards Authority happy.

 

[Electric_Andy]

Can I afford a bike? Yes
Can I afford the brand new bike I really want? No

Probably most people interpreted the question as the latter

 

[All uphill]

It's the connection between probability (of being right) and sample size. So the larger the sample the higher chance of it being accurate - leaving aside the way the questions have been worded. Seem to remember that 1000 was the magic number for some sort of credibility.

If you look at some adverts now, you'll see "70 people out of 90 thought our product was wonderful". Has little statistical credibility but it keeps the Advertising Standards Authority happy.

And those 90 people have probably been given free samples to try, maybe been the guests of the company at an introductory event in a smart hotel...

 

[PaulSB]

According to Road cc, in a recent survey, about 'half the UK population' can't afford to buy a bike for various reasons. I won't bore you with said reasons, but reading the article, the survey was taken from 2000 people across the UK from various jobs / professions.
Hadn't realised that the UK population had dwindled to about 4000 - !

It's hard to understand your point. Properly conducted polls, surveys etc. give statistically valid results from low numbers of participants.

Are you laughing at the idea half the population can't afford a bike? In which case you need to learn about poverty in the UK.

Or laughing at the poll, survey number? If so read up on polling and survey methods.

 

[Drago]

"Properly conducted" being the salient words here.

Look how few are and how wrong they are when predicting bug national events - over the last 15 years they've been more wrong right at that.

 

[Dogtrousers]

So a body responsible for affordable access to cycle ownership have stated that they are expensive to buy outright?

Amazing. Who'd have thought it? Outcome of study aligns with sponsor's interests shock.

All the same, it doesn't seem that outlandish a conclusion. Maybe worth a raised eyebrow but not much more than that. A lot of people can't afford a lot of stuff.