http://www.bbc.co.uk/news/election-2017-39965925
"In the 2015 general election, the gap between old voters and young voters was massive. Just 43% of 18-24-year-olds went to the polls, compared with 78% of people aged 65 or over, according to Ipsos Mori. That's a huge gap of 35%."
35% seems to me to be understating it a bit. On this reckoning, as far as the proportions actually turning up at the polls were concerned, there was an 81.4% greater likelihood of the oldies getting round to it.
So whose logic is wrong, mine or the BBC's?
BJ
Polling pedants
-
- Lemon Half
- Posts: 9516
- Joined: November 4th, 2016, 8:17 pm
Re: Polling pedants
Yours. But the fault isn't in your numbers, it's in failing to notice that both numbers are correct in their own terms. You are implying a question that was not explicitly asked, but which could be formulated in different ways leading to either result.bungeejumper wrote: So whose logic is wrong, mine or the BBC's?
BJ
-
- Lemon Half
- Posts: 7479
- Joined: November 4th, 2016, 6:11 pm
Re: Polling pedants
You also didn't quote the bit that said "To estimate how many old and young people vote at elections, you have to look at opinion polls".
And we all know how accurate they are
I could postulate that 18-24-year-olds are more likely to lie in polls than the over 65s, and claim the figures back up my theory.
Scott.
And we all know how accurate they are
I could postulate that 18-24-year-olds are more likely to lie in polls than the over 65s, and claim the figures back up my theory.
Scott.
-
- Lemon Half
- Posts: 7157
- Joined: November 8th, 2016, 2:30 pm
Re: Polling pedants
Thank you both, and duly recced.
As a professional idiot at maths (I got a scrape O level pass, but hey, it did the job for university), I often fall for statistical elephant traps like these. But I sometimes have to present statistical materials, and my normal inclination would have been to get round it by saying:
"Just 43% of 18-24-year-olds went to the polls, compared with 78% of people aged 65 or over, according to Ipsos Mori. That's a huge gap of 35 percentage points." Would that have been correct, I wonder?
BJ
As a professional idiot at maths (I got a scrape O level pass, but hey, it did the job for university), I often fall for statistical elephant traps like these. But I sometimes have to present statistical materials, and my normal inclination would have been to get round it by saying:
"Just 43% of 18-24-year-olds went to the polls, compared with 78% of people aged 65 or over, according to Ipsos Mori. That's a huge gap of 35 percentage points." Would that have been correct, I wonder?
BJ
-
- Lemon Half
- Posts: 6203
- Joined: November 4th, 2016, 1:12 pm
Re: Polling pedants
As the BBC article is silent on Postal Voting, I could postulate that, although Polls suggest "43% of 18-24-year-olds went to the polls", the remaining 57% may have availed themselves of the Postal Voting option.
I'd also query the meaning of "The old aren't voting more now - but the young are voting far less.". Is the journalist suggesting that the "old" used to indulge in voting sculduggery?
I'd also query the meaning of "The old aren't voting more now - but the young are voting far less.". Is the journalist suggesting that the "old" used to indulge in voting sculduggery?
-
- Lemon Quarter
- Posts: 3429
- Joined: November 4th, 2016, 10:00 am
Re: Polling pedants
If you want to demonstrate that the gap is small, you correctly call it 35%.
If you want to demonstrate that the gap is large, you correctly call it 81%.
Both are perfectly correct. And hence why statistics are such a wonderful tool to make absolutely any point you want to make. "How to lie with statistics". There's a whole book on the subject, well worth a read. https://www.amazon.co.uk/How-Lie-Statis ... 0140136290
Gryff
If you want to demonstrate that the gap is large, you correctly call it 81%.
Both are perfectly correct. And hence why statistics are such a wonderful tool to make absolutely any point you want to make. "How to lie with statistics". There's a whole book on the subject, well worth a read. https://www.amazon.co.uk/How-Lie-Statis ... 0140136290
Gryff
-
- Lemon Quarter
- Posts: 3934
- Joined: November 8th, 2016, 11:14 pm
Re: Polling pedants
" Hence statistics are..." or "That's why statistics are..."gryffron wrote: And hence why statistics are...
Gryff
GS
-
- Lemon Quarter
- Posts: 2158
- Joined: November 7th, 2016, 2:40 pm
Re: Polling pedants
Indeed; perhaps the over-65s are just being prepared: ipsos mori means "they die".Just 43% of 18-24-year-olds went to the polls, compared with 78% of people aged 65 or over, according to Ipsos Mori.
-
- Lemon Slice
- Posts: 881
- Joined: November 4th, 2016, 1:12 pm
Re: Polling pedants
For values of translation derived from Google. It's junk Latin and a coincidence - there once were two companies - IPSOS and Mori, and they merged -stewamax wrote:Indeed; perhaps the over-65s are just being prepared: ipsos mori means "they die".Just 43% of 18-24-year-olds went to the polls, compared with 78% of people aged 65 or over, according to Ipsos Mori.
https://en.wikipedia.org/wiki/Ipsos_MORI