Win a copy of Secure Financial Transactions with Ansible, Terraform, and OpenSCAP this week in the Cloud/Virtualization forum!
  • Post Reply Bookmark Topic Watch Topic
  • New Topic
programming forums Java Mobile Certification Databases Caching Books Engineering Micro Controllers OS Languages Paradigms IDEs Build Tools Frameworks Application Servers Open Source This Site Careers Other Pie Elite all forums
this forum made possible by our volunteer staff, including ...
Marshals:
  • Campbell Ritchie
  • Ron McLeod
  • Paul Clapham
  • Jeanne Boyarsky
  • Bear Bibeault
Sheriffs:
  • Rob Spoor
  • Henry Wong
  • Liutauras Vilda
Saloon Keepers:
  • Tim Moores
  • Carey Brown
  • Stephan van Hulst
  • Tim Holloway
  • Piet Souris
Bartenders:
  • Frits Walraven
  • Himai Minh
  • Jj Roberts

Polls

 
town drunk
( and author)
Posts: 4118
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Max: Polls demonstrate correlative, not causative, relationships. Consider the various ways in which phrasing can affect the answer to a question

Ok.

1999, The Gallup Poll.
"Would you favor or oppose the United States increasing economic aid to Russia?"

Favor43%
Oppose53%
No opinion4%

1999, NBC News/Wall Street Journal Poll:
"As you may know, Russia continues to experience economic difficulties that some fear may threaten the country's stability. Do you think that the United States should or should not participate with other countries to provide financial assistance to help stabilize Russia?"

Should55%
Should not33%
Depends (vol.)6%
Not sure6%
http://www.pollingreport.com/russia.htm

:roll:

But Max, this is just a badly made poll. Are you sure you don’t substitute "a badly made poll" for a poll in general?



Most popular polls are badly made, IMO. Real aren't exciting, because they display that we really don't know very much @ all. How fun is that? Maybe I should have said 'atypical polls?" :roll:


Polls demonstrate correlative, not causative, relationships.

Don't correlations sometimes indicate a cause?



Yes, but from a statistical point of view, that's usually when there's some ridiculously high level of correlation: say 99.995 % percent.




Consider the ways in which phrasing can affect the answer to a question, the discrepancies over what a 'sample population' is, the non-impartial nature of the people conducting the polls, different interpretations of the means on polls: no, polls aren't a worthwhile measure of reality. Much better, IMO, to observe to world around you.

What you advocate here looks simply like "a poor man poll". At least, in "real polls" efforts are made to ensure a random sampling, and this idea is missing in your "observe to world around you".


How is it missing?



"The discrepancies over what a 'sample population' is, the non-impartial nature of the people..." -- Ok, but at least the people who conduct the polls are well aware of these issues and specially trained on them, which cannot be said about us.


I disagree: I happen to know a few professional pollsters, and they are not well-trained @ all, from a scientific point of view. Or is that not a valid polling?



I can ask John Smith what he thinks about, say, this years' elections, but this will be just John Smith's opinion. If I want to generalize what "American people think", then well, I thought this is what the Statistics Science is for! And you seemingly just dismiss it.

It's my opinion that polls, practically speaking, have too many 'moving parts' to allow a single variable to be isolated. As such, their logical usefulness is suspect, @ best.

M
[ June 01, 2004: Message edited by: Max Habibi ]
 
Wanderer
Posts: 18671
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
I tend to agree with Max here. That is, I don't think all polls are without merit, but their usefulness is limited at best, and all too prone to abuse through either ignorance or willful result skewing on the part of pollsters. Perhaps the most meaingingful polls, IMO, are those in which the exact same question is asked at different points in time. E.g. "do you approve of the jobs the President is doing so far?" or "if the election were held tomorrow, who would you vote for?" These may be reasonably accurate in determining whether support is increasing or decreasing, at least. (Assuming the pollsters aren't intentionally trying to skew the results.) Beyond that, things get increasingly fuzzy. Any pollster who begins a question with "as you may know..." or "given that..." should probably be shot on sight. I'd say that in order to be able to take a poll at all seriously, you need to know (a) the exact text of the questions, (b) how were the respondants selected, and (c) who paid for the poll. Any poll that doesn't make this info available should be viewed with deep distrust. For those that do provide this info - well, you should still start with deep distrust, but there's some possibility of improving this opinion.

For anyone who hasn't already seen it: How To Lie With Statistics is a wonderful little read. Highly recommended. Here's where you can find it locally, Map. (BTW your local library sucks, Map - they don't seem to have it.)
 
Leverager of our synergies
Posts: 10065
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Jim: I tend to agree with Max here.

betrayer! Obscurantist! And I counted on your support! :roll:

For anyone who hasn't already seen it: How To Lie With Statistics is a wonderful little read.

Well, the first sentence our professor of statistics uttered was "there is lie, outrageous lie, and statistics". Which means it doesn't have to be?

Highly recommended. Here's where you can find it locally, Map.



(BTW your local library sucks, Map - they don't seem to have it.)

 
Mapraputa Is
Leverager of our synergies
Posts: 10065
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Ok, here is something for you, Jim: http://www.amazon.com/exec/obidos/tg/detail/-/1567186106/
 
Ranch Hand
Posts: 2937
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Max's "Russian" poll does make a point, and I also agree with Jim's arguments. However, I do believe that the polls can be useful. I subscribe to the "Consumer Reports" magazine, and polls are one of their criterias (along with the lab tests) when it comes to rating the products. I can tell you that I never regretted buying on their recommendation. If 90% of the consumers opine that product A is a piece of junk, there is hardly a reason to buy it.

Let's take a look at the other poll: Top 250 movies and 100 worst movies. Now, you may think of yourself as a very original individual, but in all likelihood, you would like the top 10 movies better than the bottom 10. It's true that sometimes you will see the highest individual rating for something considered as junk by the 99% of the people, but do you really want to spend your time looking for a gem in the pile of garbage? This is what polls are -- they are the filters of public opinion. Sometimes they will distort the information due to their faulty design, but often times they will save you time and money. Ideally, you want to test drive all 300 different cars that are on the market before you buy one. But why test drive something that costs $50K and will break every 2 months, as the polls indicate?
 
Jim Yingst
Wanderer
Posts: 18671
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
[Max]: Polls demonstrate correlative, not causative, relationships.

[Map]: Don't correlations sometimes indicate a cause?


Sometimes, sure. Quite often in fact. But it's also very often that they indicate an effect. If A correlates with B, did A cause B, or the other way around? Even if A precedes B, it may be that both A and B were caused by C, which wasn't mentioned in the poll. Post hoc, ergo propter hoc is generally dangerous and should always be viewed with suspicion, IMO.

[Map]: And I counted on your support!

Hey, I was busy attacking him over here. It's not my fault that you chose to attack Max in a sitation where he's actually right.

Eugene: I don't have a subscription to Consumer Reports, but I do think that's the sort of poll that is relatively useful. Heck, I get a lot of use out of Amazon reviews, and they're probably far more vulnerable to skewing, IMO. Or rather, at Amazon you get a disproportionate view of the people who were passionately for or against something, and were thus motivated to respond. Still potentially useful, as long as you view each individual report with some skepticism, realizing that the reviewer's priorities may not align with your own.

As for the top 250 vs. bottom 100 movies: well, again I agree this is the sort of poll where there probably is some validity to the results. Usually with polls like this I'm tempted to apply a temporal correction of sorts, as people tend to favor recent movies (or books or songs or whatever) more strongly. But looking at the lists, they didn't do too bad, IMO. Yeah the Lord of the Rings movies had a high showing, but that isn't really undeserved, IMO, and there are a decent number of older movies too. Not bad for what is basically a mass popularity contest. As for the worst 100 - I don't know. I really haven't seen most of these. But I do have a hard time believing there were 84 movies worse than Highlander 2.
 
Max Habibi
town drunk
( and author)
Posts: 4118
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

Originally posted by Jim Yingst:

Hey, I was busy attacking him over here. It's not my fault that you chose to attack Max in a sitation where he's actually right.



Um, Jim, have you check the board lately? You've been mated. May I suggest JavaScript?

M
 
Jim Yingst
Wanderer
Posts: 18671
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Ok, here is something for you, Jim: ]http://www.amazon.com/exec/obidos/tg/detail/-/1567186106/

Hmm, looks like both our libraries suck:

http://nell.boulder.lib.co.us/search/t?SEARCH=fortune+in+a+coffee+cup
http://catalog.portland.lib.me.us/search/t?SEARCH=fortune+in+a+coffee+cup

Or maybe it's not a very good book. Still, it might be worth looking into, for us Java lovers.
[ June 01, 2004: Message edited by: Jim Yingst ]
 
Jim Yingst
Wanderer
Posts: 18671
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Um, Jim, have you check the board lately? You've been mated.

Hey, you could at least buy me a drink first!
 
Jim Yingst
Wanderer
Posts: 18671
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
I came across a listing for another book that looks intriguing:

http://www.amazon.com/exec/obidos/tg/detail/-/0226534219/jranch-20/

Too bad they put that final "s" in the title, or it would have had even more possibilities.
 
Ranch Hand
Posts: 1140
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

Originally posted by Jim Yingst:
But I do have a hard time believing there were 84 movies worse than Highlander 2





Originally posted by Jim Yingst:
Too bad they put that final "s" in the title, or it would have had even more possibilities.



 
mister krabs
Posts: 13974
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

Originally posted by Jim Yingst:
Usually with polls like this I'm tempted to apply a temporal correction of sorts, as people tend to favor recent movies (or books or songs or whatever) more strongly.

You see that in polls such as "who is the greatest baseball player of all time". Pople will pick a more recent second baseman like Joe Morgan over someone like Napoleon Lajoie (hall of famer, lifetime .339 hitter, batted .422 one season) just because it has been almost 90 years since Nap hit a baseball in anger and not too many people remember him.
 
Mapraputa Is
Leverager of our synergies
Posts: 10065
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Jim: It's not my fault that you chose to attack Max in a sitation where he's actually right.

Is not!
 
Mapraputa Is
Leverager of our synergies
Posts: 10065
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
I came across a listing for another book that looks intriguing:

]http://www.amazon.com/exec/obidos/tg/detail/-/0226534219/jranch-20/



Grr...
[ June 02, 2004: Message edited by: Mapraputa Is ]
 
Max Habibi
town drunk
( and author)
Posts: 4118
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator

Originally posted by Mapraputa Is:
Jim: It's not my fault that you chose to attack Max in a sitation where he's actually right.

Is not!




Hm...wanna take a poll?
 
Mapraputa Is
Leverager of our synergies
Posts: 10065
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Eugene: Max's "Russian" poll does make a point

Hey, it was me helping Max to make his point! :roll: As for Max, he needs to work on his quoting skills harder.

Max: Hm...wanna take a poll?

Sure thing!

To all who can read it:

Do you think polls are useful or useless?


2:2 so far.

Max & Jim vs. Map & Eugene...
[ June 03, 2004: Message edited by: Mapraputa Is ]
 
John Smith
Ranch Hand
Posts: 2937
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Right, the "Russian" poll was posted by you. There is no need to raise your voice, though -- the best emphasis is made by the choice of the words, not the choice of the font.
 
Jim Yingst
Wanderer
Posts: 18671
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
2:2 so far.

Max & Jim vs. Map & Eugene...


Hey, don't count me. I'm not responding to a poll on this topic.
 
Mapraputa Is
Leverager of our synergies
Posts: 10065
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
I looked up the book in Powells. In fact it's so small I almost read it all There isn't much traffic in "Math" section, but finally somebody did want to look at the books, so I crawled to the nearby "Wedding" section. For some reason I expected it to be a quiet corner. Stupid me, turned out Americans get married much more often they study Math...

The book is good. Thank you for recommendation, Jim. I think, we need to make it a required reading here, in Meaningless Drivel. In addition to "a document about fallacies". Then Max can write a certification guide for a certified Meaningless Driveller.

When I was in high school, they taught us about Gauss distribution. We were told that leaves on a tree make Gauss distribution. We were given a bunch of leaves and we had to measure them to prove that they make Gauss distribution. They did not. I exercised all my brain and finally wrote that it's because the sampling was too small. A half of my class copied my essay, another half copied our other guru's essay. He wrote something similar. The results didn't bother our teacher.

Later I remembered that I read that Gauss distribution is usually observed when some feature is being formed by a number of small factors, neither of which is stronger than the other. The leaves we were given weren't chosen randomly, they all belonged to the same branch. There was a clear perfectly visible law: leaves closer to the stem were bigger. As simple as that, why anybody would expect Gauss distribution in this case is beyond me.

I guess, since I figured out this thing, I looked up at the people who calculated statistics. They must be so much smarter than me, and judging by a few of them I met, they really are. Well, maybe it's different here.
[ June 20, 2004: Message edited by: Mapraputa Is ]
 
blacksmith
Posts: 1332
2
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Mapraputa Is:

I guess, since I figured out this thing, I looked up at the people who calculated statistics... Well, maybe it's different here.

Yes. A lot of people here cite statistics without bothering with the "calculate" part first. It gives statistics a bad name, so people tend to ignore them even when they have something to offer.
 
Ranch Hand
Posts: 5093
  • Mark post as helpful
  • send pies
    Number of slices to send:
    Optional 'thank-you' note:
  • Quote
  • Report post to moderator
Polls are useless except to reinforce the ideas of the person hosting the poll.

If you word the questions right you can force people to either ignore the poll or give the answer you want to show as popular opinion...
 
They weren't very bright, but they were very, very big. Ad contrast:
SKIP - a book about connecting industrious people with elderly land owners
https://coderanch.com/t/skip-book
reply
    Bookmark Topic Watch Topic
  • New Topic