[kenberg]

Phil, on 2016-April-28, 22:08, said:

There is a set of words that I use but get tripped up by my spellchecker.

- Agendize
- Incentivize

Am I a neanderthal?

You get tripized by your spell checker maybe? It is trying to un-neanderthize you. Just disablize it, establize who is boss.

[mycroft]

Phil, on 2016-April-28, 22:08, said:

There is a set of words that I use but get tripped up by my spellchecker.

- Agendize
- Incentivize

Am I a neanderthal?

No, you're a Suit. Try installing the en-Manglement language pack - it will synergize your dynamic and orientate your vision. :-)

[gwnn]

Replying to a survey with "oh yea? they asked only 1000 people out of millions! that's ridiculous!"

This is somehow even more annoying because it is a superficially very convincing argument and you can't dismiss it without explaining sampling theory and confidence intervals in detail. By the time I'm finished, the other party will have moved on to some other topic.

[shyams]

gwnn, on 2016-May-16, 01:42, said:

Replying to a survey with "oh yea? they asked only 1000 people out of millions! that's ridiculous!"

This is somehow even more annoying because it is a superficially very convincing argument and you can't dismiss it without explaining sampling theory and confidence intervals in detail. By the time I'm finished, the other party will have moved on to some other topic.

Would you really attempt to explain to the other person in such a situation? I'm impressed by your patience and tolerance!
If I were in the situation, I'd simply brand the other person a duffer and move on.

[gwnn]

Yes. Well, at least normally I do so because I used to be one of those people - at least in the sense that I had the intuition that asking just 1000 people from a million can't be enough. In fact (not writing to you shyams in particular, just for people who might not realize this), if we ask 1000 out of 1M and got 50% of yeses, the confidence interval is somewhere between 47-53% or 46-54% at 95 or 99% level. The funny thing is that (essentially) the same confidence interval applies if the total population were 100M or any larger number {of course in real life larger populations will be more difficult to sample representatively}.

[Zelandakh]

The best way of explaining it to the public is to tell the story of George Gallup versus The Literary Digest for the 1936 Landon-Roosevelt election. The problem with poling is rarely sample size but rather constructing the sample to be representative of the entire population.

[gwnn]

Zelandakh, on 2016-May-16, 04:49, said:

The best way of explaining it to the public is to tell the story of George Gallup versus The Literary Digest for the 1936 Landon-Roosevelt election. The problem with poling is rarely sample size but rather constructing the sample to be representative of the entire population.

How does this story (which, as far as I can gather, is about random polling vs asking readers of a particular magazine) explain anything about apparently small but sufficient samples? Of course representativeness is important but I was talking about the misconception that one needs to ask a significant number of people (compared to the total population) before drawing conclusions about the general public.

edit: oh, the Literary Digest survey was based on 2.3 million (whoa) people while Gallup on 50,000. OK then I kind of understand what you mean, although I don't know if it's totally convincing to outsiders, since it covers why large samples aren't guaranteed to work in practice but not why very small samples could work in principle. Well, in any case, I know my way of explaining is completely unconvincing so I might try yours next time.

This post has been edited by gwnn: 2016-May-16, 08:53

[Zelandakh]

gwnn, on 2016-May-16, 08:48, said:

edit: oh, the Literary Digest survey was based on 2.3 million (whoa) people while Gallup on 50,000.

Exactly, this was the turning point in polling methodology. Before Gallup it was assumed that more respondents meant a more accurate poll and a great deal of time and expense went into creating samples as large as possible. Afterwards it was realised that a small but well-constructed sample offered a much more accurate representation of the general population. It is simply a practical example illustrating what the maths says and shows in a simple way that the viewpoint of the "audience", despite being seemingly logical, is wrong.

[Fluffy]

This reminds me of those silly polls on TV programs asking silly questions, getting 80%+ responses because people who vote are all sheeps watching and trusting the biased program.

People who lose money to vote on those polls is one of those things that makes me feel appart from humanity.

[Phil]

mycroft, on 2016-April-29, 09:53, said:

No, you're a Suit.

....feeling the burn

[gwnn]

Pronouncing determine as deter+mine, which is the de facto standard at scientific conferences by non-natives. Before you say it, yes, I know it's petty and I mispronounce everything too. I also know that it would be more logical if pronunciation were additive like that!

[kenberg]

gwnn, on 2016-May-28, 03:27, said:

Pronouncing determine as deter+mine, which is the de facto standard at scientific conferences by non-natives. Before you say it, yes, I know it's petty and I mispronounce everything too. I also know that it would be more logical if pronunciation were additive like that!

I was recently in Oregon and discovered that the last syllable is pronounced like gun rather than gone. I also discovered that Oregonians are quite insistent about this. See https://www.facebook...80546898870520/ among many other sites.

One of the newspapers addressed how the various presidential candidates pronounced, or mispronounced, the name of their state. Hillary did not much campaign there, so they described her pronunciation as ignoregon.

I am happy to pronounce it as the natives wish, as long as I do not have to agree that their choice is in any way the natural one.

I don't think I have ever heard someone pronounce determine so that it rhymes with deter swine. I just go to the wrong conferences.

[Vampyr]

Fluffy, on 2016-May-17, 08:13, said:

This reminds me of those silly polls on TV programs asking silly questions, getting 80%+ responses because people who vote are all sheeps watching and trusting the biased program.

People who lose money to vote on those polls is one of those things that makes me feel appart from humanity.

These "polls" are lotteries, offering prizes. That is why people are charged for voting (I do not see how they can lose money actually, but they can spend it). What is not immediately obvious is how much money goes into the prize fund and how much to the organisers.

[Vampyr]

I have heard epi-tome and crew-dights from native speakers. And of course nearly everyone pronounces for-tay when they mean fort (forte, as in strength. It is French for strong, not Italian for loud).

[barmar]

kenberg, on 2016-May-28, 04:29, said:

I was recently in Oregon and discovered that the last syllable is pronounced like gun rather than gone. I also discovered that Oregonians are quite insistent about this. See https://www.facebook...80546898870520/ among many other sites.

One of the newspapers addressed how the various presidential candidates pronounced, or mispronounced, the name of their state. Hillary did not much campaign there, so they described her pronunciation as ignoregon.

I am happy to pronounce it as the natives wish, as long as I do not have to agree that their choice is in any way the natural one.

It wasn't until I had a friend in college from Missouri that I learned that they pronounce the last syllable as "ruh" rather than "ree".

But I'll also forgive them if they don't know how to pronounce Worcester. And only native Bostonians pronounce that city as Bahstahn instead of Bawstun -- I don't think any of them expect the rest of the country to adopt their accent.

But face it, spelling and pronunciation in English is a mess, because the language is such a melting pot of other languages.

[barmar]

On The Daily Show this week (I think Thursday night's show) one of the correspondents was making fun of Trevor Noah (who is from South Africa, but I suspect most Americans would mistake his accent for British) for his pronunciation of "controversy" -- he says "con-TROV-er-sy" in stead of "CON-truh-ver-sy" like normal Americans.

[helene_t]

Funny, I would say con-tro-VER-sy.

[barmar]

helene_t, on 2016-May-28, 12:50, said:

Funny, I would say con-tro-VER-sy.

I just checked several online dictionaries, they agree with my pronunciation. The primary emphasis is on "con", the secondary emphasis is on "ver". But they also say that the British pronunciation emphasizes "tro".

The primary emphasis switches to "ver" when you say "controversial". Maybe that's what you're thinking of?

[gwnn]

Did you check the dictionary for Danish-German-Dutch accent in English?

[helene_t]

In Esperanto it is kon-tro-ver-SI-o I think.