Friday, March 10, 2006

Why Bother Taking Polls?

Riddle me this Batman. Why do the major polling outfits, especially AP/IPSOS even bother to take polls? Rather than spend all that money, they could just call up Pelosi and Rockefeller and get the answers. Then they could just plug in the numbers they want to see.

Webster defines the word poll (in the context we are talking about) as:

“A survey of the public or a sample of the public to record opinion or to acquire information.”

When an organization decides to take an opinion poll to get an idea of what the general public is thinking, they are supposed to pick a statistical population that reflects the general population. Yet, in poll after poll over the past six months (awww…to heck with that – over the past two weeks) we have seen just the opposite. What we have seen is polls that oversample one group’s opinion, and by correlation, undersample another’s. That means that the results from any such poll must immediately be considered invalid. That’s right – invalid.

If statistical sampling of this type were done as a school project (High School or College or Graduate), the student would receive a failing grade. The reason is simple – the statistical population must reflect the real world or the results can not be considered true. If another person then picked the results from that failed, biased poll up, and touted the numbers to others, they would quickly be brought into line.

But of course, that scenario – having the correct statistical population and unbiased results doesn’t belong to the real world. In today’s world, if someone produces a poll with some results – especially results that are against the Bush administration – it is considered valid and honest, no matter what the underlying demographics. And not only that, but immediately the major news outlets pick up the invalid results and publish them as headline material. In essence, they use the poll as a propaganda tool to fit their own agendas. Let’s look at the latest AP/IPSOS poll for an example.

This poll was conducted between March 6 and March 8, 2006 and included 1000 respondents, 828 of which were registered voters.

Before I attempt to see what kinds of questions are asked, I always skip to the end of the poll to pick up the demographics of the respondents. That gives a large clue as to the validity of the poll. In this case, questions 1 and 2 within the party affiliation section tell the tale:


1. Are you currently registered to vote at this address, or not?
ALL ADULTS
Yes................................................... 79
No .................................................... 20
Refused/not sure ........................... 1
2a. Do you consider yourself a Democrat, a Republican, an Independent or none of these? *

INITIAL PARTY IDENTIFICATION REGISTERED VOTERS
Republican ....................................... 29
Democrat ......................................... 35
Independent ..................................... 20
None of these................................... 14
Not sure ......................................... 2

(IF “DEMOCRAT” TO Q.2a, Q.2b ASKED. IF “REPUBLICAN” TO Q.2a, Q.2c ASKED. IF “INDEPENDENT” or “NONE OF THESE” TO Q.2a, Q.2d ASKED. RESULTS SHOWN IN SUMMARY BELOW.)

2b. Do you lean strongly or only moderately toward the Democratic Party?
2c. Do you lean strongly or only moderately toward the Republican Party?
2d. Do your beliefs tend to lean more toward the Democrats or the Republicans?

REGISTERED VOTERS
Strongly Republican.......................... 18
Moderately Republican ..................... 22
Definitely Independent/neither .......... 8
Moderately Democrat........................ 28
Strongly Democrat ............................ 22
Refused/not sure............................. 2
Total Republican ............................ 39
Total Democrat ............................... 51

Any way you look at these numbers you can see that the poll results must be considered invalid. In the last national election the percentage of voters was 37% Republican and 37% Democrat. Any poll which favors one party over another in the statistical population of the respondents does not reflect that real population and will be slanted to the POV of the respondents with the larger number. It is quite disingenuous to call a poll like this anything else – and probably borders on unethical. But let’s take a look at some more:

(ASKED OF HALF THE RESPONDENTS.)
2. How likely is it that civil war will break out in Iraq?
Very likely ......................................... 39
Somewhat likely................................ 39
Not too likely..................................... 12
Not at all likely .................................. 5
A civil war has already broken
out in Iraq (VOL) ............................... 1
Not sure.......................................... 4
Total Likely...................................... 77
Total Not Likely............................... 17

Let me see if I have this correct. The poll was taken during the week of March6 through March 8. Now why would people think that there could be a civil war in Iraq? Let’s go to the newsroom – I’ve picked a sampling of hundreds of articles that were published in the two weeks before the poll was taken. It does not include the evening MSM news broadcasts – many of which led off with the “Iraq Civil War” story:

Iraq on the brink

The crash of civilizations
I've been predicting for some time that Iraq could end up being like Lebanon to the power of 10 if the civil war already underway there should escalate. Last week's bomb attack on the Shiites' Golden Mosque in Samarra may be the trigger for precisely that escalation. The point is that Iraq's "clash" is not between civilizations but within Islamic civilization — between the country's Sunni minority and its Shiite majority.

Remember Beirut? Welcome to Baghdad
Today, with Iraq on the brink of civil war between Shiites and Sunnis — and the Americans looking increasingly helpless, their mission having shifted from "democratization" to peacekeeping — another analogy has been circulating in policy circles: the Lebanese civil war. Like any analogy, it's not perfect, but it's the best we have so far.

Iraq's Besieged Sunnis Now Looking to U.S.
The moment of bitter irony for the 52-year-old father of six is emblematic of a sharp shift in Iraqi opinion. Three years after the March 2003 invasion that ousted Saddam Hussein, with the threat of civil war looming, leaders of a nervous Sunni Arab minority have started to drop demands for an immediate U.S. withdrawal.

As Iraq Conflict Changes, Has Bush Kept Up?
Larry Diamond, a senior fellow at the Hoover Institution and former advisor to the U.S.-led occupying authority in Iraq, concisely expressed the evolving view when he wrote in the latest issue of the New Republic: "Iraq is in the midst of a civil war."


Civil War Looms With 68 Killed in Baghdad
The surge of violence deepened the trauma of residents already shaken by fears the country was teetering on the brink of sectarian civil war, threatened talks among Iraqi politicians struggling to form a government and raised questions about U.S. plans to begin drawing down troop strength this summer.

Evictions May Foreshadow Iraq Civil War
BAGHDAD, Iraq -- The sectarian cleansing that drove 68-year-old Abbas al-Saiedi from his home may be as alarming a sign of a country on the brink of civil war as the killings that have swept Iraq in the past week.

Diplomacy Helped To Calm the Chaos
AGHDAD, Feb. 27 -- In the days that followed the bombing of a sacred Shiite shrine, Iraq seemed within a hair's breadth of civil war. But an aggressive U.S. and Kurdish diplomatic campaign appears for now to have coaxed the country back from open conflict between Sunni Arabs and Shiites, according to Iraqi politicians and Western diplomats speaking in interviews on Monday.

30 Killed as Violence Continues in Iraq
Iraq began to tilt seriously toward outright civil war after the Feb. 22 bombing of the revered Shiite Askariya shrine in the mainly Sunni city of Samarra, 60 miles north of Baghdad.

I think it is fairly clear that the reason people think a civil war is looming is that they were bombarded with it for two weeks in the MSM. Great way to take a poll though. Have the talking heads of MSM and Society of Subversiontell people over and over again that the civil war has started, and then ask people what they think. Unethical? You have to answer that for yourselves.

But to help you along the way to your decision, let’s look at the demographics of some of the other polls published in the past few weeks:

From the Washington Post – ABC Poll released on March 6, 2006:

901. Generally speaking, do you usually think of yourself as:
Democrat     32
Republican     28
Independent     36
Other     3
No Op     1

901/904 Leaned Party:
Democrat     50
Republican     42
Independent     7
Other     1
No Op     1

908a. Would you say your views on most political matters are liberal, moderate, or conservative?
Liberal     22
Moderate     42
Conservative     33
Don’t Think in
Those Terms     2
No Op     1

Imagine that. This poll, released just days before the new AP/IPSOS poll also oversampled Democrats. Hard to believe isn’t it? But there is one difference here. This poll had a very different Presidential Approval Rating:

1. Do you approve or disapprove of the way George W. Bush is handling his job as president? Do you approve/disapprove strongly or somewhat?

-----Approve-----
NET     41
Strongly     24
Somewhat     17

-----Disapprove-----
NET     58
Somewhat     14
Strongly     44

No Op     1

Now wait a minute. Is MSM really trying to tell us that in less that 3 days time the President’s approval rating fell 4 points? I suppose it could be, but what would drive that? Maybe we should take a quick look at another MSM poll. On February 27, CBS released a poll also. Let’s take a look at the demographics first:

Total Respondents: 1018

-----Unweighted-----
Total Republicans     272
Total Democrats     409
Total Independents     337

-----Weighted-----
Total Republicans     289
Total Democrats     381
Total Independents     348

Wow! I think I see a pattern here. Three different polls, all conducted and released with ten days of each other, and get this – all of the clearly oversampling people from the democratic party. At least the CBS poll above tries to weight the parties to make it more fair, but even with that it seems that both democrats and independents are over-represented.

But, just like Ronco – There’s More! In the CBS poll we find the President’s numbers to be the following:

Pres. Bush Job Approvals
Now     34%
1/2006     42%
10/2005     35%
11/2004     51%

Let me see if I’m reading the polls correctly. On February 27, 2006, Bush’s approval rating was 34% according to the unbiased CBS poll. On March 5, 2006 it was 41% according to the unbiased Washinton Post/ABC Poll. And today his rating is 37% based on the obviously and completely unbiased AP/IPSOS poll. It is incredible to think that people change their minds so quickly – remember that the polls are supposed to be representing what the true population actually thinks. And according to these wonderfully enlightening polls President Bush saw a magnificent rise in the polls of 21% between February 27 and March 5 (six days) and then plummeted from there by almost 10% over the next four days. Wow – either the public is really, really fickle – or there is something wrong with the polling. I wonder what it could be?