Lies, damned lies and Philharmonic statistics
NewsThe Royal Philharmonic has put out another of its small-sample surveys.
This one says: ‘More than six million newcomers to classical music planning to give it a go this year.’
It continues: ‘Six million adults in the UK who admit “not knowing much” about classical music intend to get into it this year. Around a quarter of these newcomers aged under 35, equating to around 1.5 million young people.’
How do they know?
The figure is blown up out of all proportion from a polling sample of 2,020 adults across the UK.
The method used is:
– According to the office of national statistics, there are 52,890,044 people aged over 18 in the United Kingdom.
– The proportion of people who admit they “do not know much about classical music” is 36%, which equates to 19,040,416 people.
– Of these people, 33% said they were interested in giving it a go this year, which equates to 6,283,337.
The 6.2 million is designed to get a headline on ClassicFM (and, sure enough, it did).
All polling depends on extrapolating from a small sample to the whole population. The fact that the sample size was some 2000 people isn’t a problem in itself. The sample selection method (of which we know nothing) is open to question, and the extrapolation sorely needs error bars to indicate the uncertainty of the result.
NL, Beware before you pooh-pooh the statistics! There’s a high correlation between music people and numbers people out here among your readership.
If you don’t believe the theory, cool. But try an online sample size calculator just for fun (e.g., https://www.calculator.net/sample-size-calculator.html)
For a population of 53 million, a sample size of 2,020 yields roughly a 2% margin of error with 95% confidence. You don’t have to sample the whole world to see which way the wind blows – you just need a small sample before the results start to converge on a reliable answer, assuming the sampling is truly random. As to whether those 6 millions will actually follow through, see “New Year’s Resolutions.”
He’s not pooh-poohing anything, just quoting Mark Twain!
Twain was no whiz at statistics, either. 😉
Go on to YouTube, type in “Dave Hurwitz – To Hell With Young People” for his take on attracting them to classical music. It’s an interesting take even if you don’t agree with it.
This is terrible! If they all decide to take up the bassoon after getting into classical music, it will be hard for all those bassoonists to get gigs.
… so you want to gatekeep classical music and not encourage more people to take part in it?
Certainly gatekeep to prevent an epidemic pf bassoonists ,
what’s the definition of “classical music”, n “give it a go”? U can stretch these meanings anyway u want to n produce any results u want, i.e. deceive yourselves n make yourselves feel better.
I am a professional social scientist. A polling sample of 2.000 people is way above the required number to provide statistically relevant numbers for the entire population – provided the poll was done with proper regard to methodology. Please read a book on statistics and public opinion polls before mocking the results.
For about $0.35 @ day, YOU could sponsor a viola player…..
Why? Bagpipers and kazooists probably need it more…..
If well-chosen, 2000 is a sample size that’s plenty big enough to give a reliable result. It’s twice the size of the usual sample of political polling.
The poll hasn’t been released by the RPO yet, but if it’s anything like the survey they did last year it will have been done by a reputable company using sound statistical methods.
Sometimes a little knowledge is more useful than a kneejerk cynicism!
The sample size may be adequate but the questions are meaningless and so will be the assumed and unashamedly reworded conclusions.
The statistical precision with which a random sample can represent a population varies primarily with
the size of the sample, not the size of the population. 2000 is a relative big sample.
The point about adequate sample size has been well made by others. My objection is the way the outcome is expressed: “ Six million adults in the UK who admit “not knowing much” about classical music intend to get into it this year”.
This is a factually incorrect statement. Six million adults didn’t admit. It was 240 people out of a sample of 2020. So it would have been quite reasonable to say, “a poll of around 2000 people suggests that there might be….. six million adults…”
Presenting stats badly just makes it worse.