Dissecting a recent poll. CNN makes it up.

[PhillySane]

As part of their latest Public Opinion Poll, CNN posed this question:

"As you may know, a website called Wikileaks has displayed thousands of confidential U.S. government
documents concerning U.S. diplomatic and military policies. Do you approve or disapprove of the
Wikileaks website displaying these documents?"

What's the headline they pull out of this?

CNN Poll: WikiLeaks has few fans in U.S. (with a nice picture of Julian Assange)

Notice how the question is worded. "As you may know" not Do you know? or How much do you know about Wikileaks, just "As you may know". This assumes everyone they ask this question knows this about Wikileaks. Bad research technique. Screen out those who know what they're talking about first, before you ask a loaded question.

Secondly, the respondents simple said whether or not they approve or disapprove of Wikileaks displaying these documents, not that they approve or disapprove of Wikileaks itself, yet the headline reads as if THAT was the question.

What should have been the headline is that the 18-34 demographic doesn't even register on this poll. Why?
Because they weren't even asked the question. Had they been, I'm sure the overall results would have been much different.

Please CNN, stop making news up and report it, thank you.

http://i2.cdn.turner.com/cnn/2010/images/12/30/rel17n.pdf

[Robb]

1. "Dissecting?"

How was this poll conducted? Land lines and cell phones, or one, or the other? How was the sample weighted with regard to gender, age, education, race, or region?

You can neither praise nor pan the results without this information, I believe.

[PhillySane]

3. How would any of that have changed the outcome?

We'll never know. My main point is, CNN has already dissected this poll to pull an anti-Assange headline off of it.
18-34 year olds don't even register. How can you determine Wikileaks' fans in the US when you don't even include this group?

[Robb]

4. What do you mean, "we'll never know"?

The answers to those kinds of questions are exactly how we know whether polls are valid.

For example, if you have fewer 18-34 year olds answer a poll than their representation as a percentage in the population, you use a multiplier to bring some reality to the results. It's called a weighted poll for that reason -- do you know, is the CNN poll weighted? Because it matters a great deal, and this is rarely explained.

[PhillySane]

5. They don't say

but why not just ask 18-34 year olds this question??? And it was a telephone poll.

[Robb]

8. Agreed, it's a crappy-sounding poll.

But there's a lot we don't know. For example, when Princeton Survey Research Associates conduct a poll, they include this explanation of methodology:

"The combined landline and cell phone sample are weighted using an iterative technique that matches gender, age, education, race, Hispanic origin, region, and population density to parameters from the March 2009 Census Bureau's Current Population Survey. The sample is also weighted to match current patterns of telephone status based on extrapolations from the 2009 National Health Interview Survey. The weighting procedure also accounts for the fact that respondents with both landline and cell phones have a greater probability of being included in the combined sample and adjusts for household size within the landline sample. Sampling errors and statistical tests of significance take into account the effect of weighting."

This is how you differentiate between polls and "I asked six guys in my living room what they thought." :)

[muriel_volestrangler]

9. It looks like they could get very few under 35s to answer, and didn't bother asking others

until they got a meaningful sample. The CNN page says the total sample was 1008, but doesn't note that it's heavily skewed towards the older - which we see from the sampling error they quote:
18-34: too big to give figures
35-49: +/- 7.0%
50-64: +/- 5.5%
65+: +/- 5.0%
49-: +/- 5.5%
50+: +/- 3.5%
total: +/- 3.0%

(and this applies to all questions asked in the poll)

So we see that they must have sampled more over-65s than under 50s. Yeah, they may try to weight the poll (but the results I normally see from polls tell you what the weighting they did was). As far as I can tell, they asked about 195 35-49s, 315 under 49s and 315 50-64s, and 380 over 65s. So they asked about 120 under 34s. Given the answers do vary significantly with age (only 9% of over 65s approve), to not even mention that they got very few under 34 answers, or that the result is therefore more inaccurate than the '3%' they claim, is irresponsible. Shoddy work by the pollsters, and by the reporter too, I think.

[PhillySane]

10. Good point

in todays world of cell phones, the under 34 crowd probably doesn't get represented in these telephone polls. Land liners are over repped. Still, if they had any figures at all on the 18-34, they should have reported it along with their weighting technique. That would have been much more accurate. Finally, the wording of the question puts it out there as well known fact by all that Wikileaks has done this horrible thing. So, most people take it as they get it and say they disapprove. How about "Are you aware of the website called Wikileaks?" and move on from there, screening out those who say no. And the headline could have been worded different as well.

[xchrom]

2. Recommend

[leveymg]

6. And, the question contains a falsehood. These aren't military secrets - they're diplomatic cables.

None are classified above Secret. Most are just chatter among State Department employees which reveal little or nothing one couldn't learn by reading the newspapers.

Important military information is classified Top Secret/Sensitive Compartmented Information or Special Access Program - TS/SCI or TS/SAP. None of the Wikileaks we've seen or heard about are classified as such.

Leave it to CNN (Faux Lite News Network) to get that wrong and spread further misunderstanding.

[Junkdrawer]

7. K&R

[rhett o rick]

11. CNN is clearly a propaganda tool of the Corp-Cabal. It is more dangerous than Fox. nm

[FreakinDJ]

12. CNN Poll - Pure Propaganda - ealier polls 60% favor Wikileaks

Was it right for WikiLeaks to leak the Afghan war documents?

Yes (68%, 1,325 Votes)
No (32%, 623 Votes)

http://blogs.reuters.com/ask/2010/08/02/was-war-documents-leak-in-the-publics-best-interest/

So why the HUGE disparity - because the OP is right