Use of polling as a metric

as well as Malysia, Indonesia and the southern Philippines where societal rules make it difficult to talk to women, difficult for them to give an honest opinion.

Far more importantly and harmfully to polling in the western sense, all respondents are due to innate politeness going to tell you in part what they THINK you want to hear and in pure pragmatism, tell you in part what is beneficial to them or their kin or friends. The percentage of politeness and pragmatism will vary dependent upon who the pollster is and on the subject of the poll -- but those two factors will always skew the poll results.

Recall also Al Taqqiya -- and that it applies based on not to whom one is speaking but to the end recipient...

Recall also the rule of thumb; If given a figure, decrement or enhance it by 80% depending on which way will make the provider look less good. Born hagglers. Neither of those things is a lie in the western sense; the first is dissimulation religiously encouraged (some believe mandated), the second is telling you something that makes the teller look good -- which is required -- or is what he thinks you'd like to hear. Not immoral or wrong, just different. Really different.

Moderator's note: Today Bob's World has posed a question, in a thread on Afghanistan, which deserved this thread's creation - on Post No.4, so start there and return here!

Some hawks like to refer to mysterious polls that allegedly show that most (like 60-90%) of the Afghans (they never seem to ask only Pashtuns these questions) want ISAF to stay and Taliban to be defeated.

Ken,
I think the situation with Taqiya does not apply to polling. Taqqiya as a conscious act is most likely when a particular individual is hiding his real views because of the danger HE or she may face OR when a committed ideologue knows what the party line is and acts to further those aims. I cannot think of any examples where a large population would change answers to suit ideological imperatives...not if questions are prepared a little carefully. Basically, what I am saying is that large populations are not thinking that deeply about the poll you are doing. For example, one could argue that committed islamists being interviewed on CNN may hide their real views about hating infidels because they know what the needs of propaganda are, but the "general population" is unlikely to think that far ahead or that specifically.
Btw, what i have a hard time figuring out is how little people in the West (or its media) seem to know about the aims and ambitions of the committed islamists when they routinely broadcast their views to anyone willing to listen. Their views dont show up in general polling to the same extent because the general population is not part of the "vanguard" party...but they are the views that matter, not the views of the "general public". This is not an election, its a war.

The issue of metrics / performance indicators has been discussed before on at least this thread: http://council.smallwarsjournal.com/...ead.php?t=3895

Today Bob's World has posed a question, in a thread on Afghanistan, which dserves a seperate thread's creation. Alas this is after a few posts commenting, so this post will appear in the midst of the thread and not at the start.

Bob's World's question:

Quote:

---

The key is to use polling of the populace as your primary metrics. Actually, I created a model for for this type of targeting that JSOTF-P has been employing for determing what projects to take on, where, etc for best effect. Coltroup briefed it at the last PASOC. He seemed to be pleased with the results, but you have better data than I do in that regard.

---

On another, earlier thread on Metrics I cannot recall this point being made and it is not my field. So instead I will relate an account by a recognised, local (UK) expert on polling ethnic minority communities - discussed after a frustrating and eventually successful polling of a Muslim community.

The polling was originally specified to be made by phone contact, on grounds of cost; this was rejected and personal interviews chosen. The poll had to include all sections of the community - notably women - and it took repeated visits to establish trust before women could be spoken to. Youth were slightly easier, but more difficult to find at home - as the pollsters hours needed to be adjusted.

On this example alone (dangerous, caution now) how will polling the populace work in "hot spots" like Mindinao and Nuristan?

I have some recollection that in Iraq some polling was done.

davidbfpo

Polling among a citizenry that is completely unaccustomed to the practices and purposes of polling can be dangerously deceptive. The first reaction to an individual asking questions about sensitive matters is likely to be suspicion; it may be difficult or impossible to convince people that their answers are confidential and that you are trying to figure out what "the people" think as opposed to what they, the individuals being questioned, think. Many, often most, conclude that the safest response is to tell the questioner what you think he wants to hear, not what you really think.

When polls in Afghanistan and Pakistan show relatively low levels of sympathy for the Taliban, we have to ask how many people would be willing to identify themselves as Taliban sympathizers to a government that is at war with the Taliban.

All too often polling provides a scientific, statistically significant sampling of what the citizenry thinks are the desired answers to the questions being asked. "Tell 'em what they want to hear" is a time-tested device for getting out of an awkward situation, and a stranger asking delicate questions is an awkward situation.

Quote:

---

Originally Posted by omarali50

I think the situation with Taqiya does not apply to polling. Taqqiya as a conscious act is most likely when a particular individual is hiding his real views because of the danger HE or she may face OR when a committed ideologue knows what the party line is and acts to further those aims.

---

And when the respondent has a sense of humor and like to confuse and befuddle. :wry:

Quote:

---

I cannot think of any examples where a large population would change answers to suit ideological imperatives...not if questions are prepared a little carefully.

---

The last item is often a problem; they frequently are not. Plus, it's not the ideology nearly so much as it is politeness (what the questioner wants to hear) and the pragmatic, personal -- can I benefit from this? The pragmatic aspect which applies to people all over the world, not just the ME and Muslims. Though, as I said, when haggling is a national sport (that's a compliment, not a knock), the folks in the ME and Asia have an edge on any westerner.

Quote:

---

Basically, what I am saying is that large populations are not thinking that deeply about the poll you are doing...

---

That's universally true; my point was that the very important aspect of politeness (NOT a western attribute) and pragmatism (universal but well honed in societies which practice bargaining) can skew results. Taqqiya may or may not apply -- depends on the reason for and wording of the survey. I agree that generally it will not -- though in questions of governance and relations with the west it may.

Dayuhan said it better than I...

Quote:

---

Originally Posted by Ken White

Plus, it's not the ideology nearly so much as it is politeness (what the questioner wants to hear) and the pragmatic, personal -- can I benefit from this?

---

There's also a flip side to benefit... people will be considering whether there may be some risk to them in an honest answer. The safe way out is to say something bland and uncontroversial, get the interview over with, and resume watching your back.

Quote:

---

Originally Posted by Dayuhan

Polling among a citizenry that is completely unaccustomed to the practices and purposes of polling can be dangerously deceptive.

---

This is a key point, and I suspect is one of the largest determinants of the validity and reliability of polling data in conflict-affected countries.

In Afghanistan, I would have serious doubts as to how much weight should be placed on polls. In Iraq, by contrast, polling has become more and more common, is widely reported in the local and regional media, and as a consquence has become increasingly reliable.

In the Palestinian territories, polling takes place (by one firm/NGO or another) every few weeks, is considered a routine part of politics, and has proven to be quite accurate when done well (judging from voter behaviour in elections).