The Pew Research Center and other Pew units and others conducted a survey of Muslims. However, they interpret their results incorrectly. A careful analysis of their appendix like material in their pdf shows that the opposite conclusion should be drawn to their conclusion.
The correct conclusion is that Muslims in the US would stop answering survey questions in the middle when the survey switched from complaining about discrimination against Muslims to whether Muslims, and the survey respondent in particular support al Qaeda or terrorism. Pew reports that response rate was so low they had to up the fee from 25 dollars to 50 dollars to get people to go beyond the middle of the survey.
The survey questions start out slow and then ask about discrimination against Muslims. Then in the middle they start to switch to terrorism by Muslims. At that point, they encountered resistance to answering. They had so many Muslims drop the survey, it was done by phone, they had to change their design to offer 50 dollars. Then Muslims would answer and give the “correct” answer of not supporting terrorism, in large numbers, or al Qaeda.
However, what Pew did was unintentionally create an experiment in experimental economics. For 25 dollars US Muslims won’t criticize al Qaeda or terrorism. For 50 dollars, a few, but still not many, will. Thus the real conclusion is that Muslims do support al Qaeda and terrorism, including against Americans. For 25 dollars, they won’t pretend they oppose them, but for 50 dollars they will. In addition, they had to give their name and address to get the 50 dollars, so they couldn’t get the money without being identified.
Starting from 60,000 or 55,000 (they use both numbers) of contacts, Pew eventually got over half of its 1,000 final survey respondents to say that they didn’t support terrorism or al Qaeda. However, to do this they had to do the following:
- They had to increase the payment for the survey from 25 dollars to 50 dollars after people balked in finishing the survey when asked to answer questions on these issues.
- They had to go back to old survey respondents who were English or Spanish speaking, tended to own a house and be in one place, and tended to go to mosque less.
Thus the Pew survey was really experimental economics. It switched from being a survey to an experiment in economics when they had to increase the payment from 25 dollars to 50 dollars to get people to not just complain about discrimination but to say that they opposed al Qaeda and terrorism. For 25 dollars, they wouldn’t do that.
To get the money, they have to give their true name and true address, to get delivery. They are thus at risk if they say they support al Qaeda and terrorism. So they have to say they don’t support them, in effect. Despite this, some did anyhow. So the experiment was narrowed to a specific issue, how much money do you have to be paid to say you oppose al Qaeda and terrorism. For those responding it was between 25 and 50 dollars.
However, the actual conclusion is slightly more subtle. They started with 55,000 or 60,000 contacts and ended up with 1,000 usable ones. By design, they were only going to get no answers to support al Qaeda and terrorism, since they had to give their name and address to get the money. So the real survey issue was how much do you have to pay to get 1,000 useable responses if you start with 60,000 contacts?
This experiment resulted in the answer that its 50 dollars.
Phrased differently, to find 1,000 Muslims desperate enough
to say they oppose al Qaeda and terrorism against Americans, how much do you have to pay if you start with 60,000 calls. The answer is 50 dollars is enough but 25 is not. So its really helping designers of flawed surveys plan their budget.
From the start of 60,000 calls to 1,000 usable results, and a payment of 50 not 25 to get them, plus an additional recontact of another subgroup they could rely on more, and they admit that, in effect, they allow us to infer that:
Muslims in the US support al Qaeda
Muslims in the US support killing Americans in the US.
These conclusions are strengthened by analysis of some of the other survey questions. Some of this analysis is below, and some at the threads at Jihad Watch and the commentary by Robert Spencer, Hugh Fitzgerald, and posters there.
In particular, 55 percent said they didn’t think the US was sincere in the war on terror (see Question Q.H4), and only 40 percent said that Arabs did 9-11 (Question Q.H3). These were questions they felt safe to answer as they really felt to get the 50 dollars. The answer is not what they believe, its a way to express their animosity against Americans.
But to answer those “safe questions” they had to answer if they supported al Qaeda or supported terrorism. To induce them to say no to those two, which they had to do because the survey took their name and address to pay them, they had to be paid 50 dollars, when paid 25 dollars they would stop the survey and give up the 25 dollars despite having given enough of their time to get half way.
==Here is what Pew said
Muslim Americans: Middle Class and Mostly Mainstream.
“The first-ever, nationwide, random sample survey of Muslim Americans finds them to be largely assimilated, happy with their lives, and moderate with respect to many of the issues that have divided Muslims and Westerners around the world.”
==Comments made by OA at Jihad Watch on this topic
Fitzgerald: A peculiar poll
“To compensate respondents for their time and to make participation in the survey more attractive, an incentive of $50 was offered for completing the interview. The study began with $25 incentive, but this was subsequently increased to $50 to further minimize mid-interview termination. Three-quarters of the respondents provided name and address information for receiving the incentive payment.”
Page 72 of pdf.
The poll questions are in categories, Q.A, Q.B, etc. Q.D. (page 95 of pdf) is about discrimination against Muslims. Q.F is where they start to ask about extremism by Muslims . Q.F. is page 95 of the pdf. So when the poll switched from Muslims being discriminated against to Muslims engaged in terrorism, people stopped answering questions when they were paid 25 dollars.
So they had to increase the payment from 25 dollars to 50 dollars so that people would continue to answer questions after Muslims being discriminated against in Q.D. to Muslim terrorism in Q.F and afterwards. Q.H is when they ask the heavy questions on Muslim terrorism, almost the end of the poll.
So at 25 dollars, the RDD survey respondents would all stop answering as soon as the questions stopped being discrimination against Muslims to Muslim terrorism and extremism.
Translation of above. For 25 dollars, people will talk about discrimination after 9-11. But to say that terrorism is bad, i.e. to say other than what they believe, they require 50 dollars. They won’t say al Qaeda is bad for 25 dollars but will for 50 dollars.
This ended up being experimental economics. As experimental economics we discover that to get Muslims to say al Qaeda is bad, we have to pay them 50 dollars, 25 is not enough. This is a different part of economics than survey research, but a valuable contribution nonetheless. Hats off to Pew, they proved that Muslims do support al Qaeda and that it takes 50 dollars to get them to say they don’t.
To summarize, the Pew Study was experimental economics. Their experimental design, by accident, was to determine by experiment how much money it takes to get Muslims to say they are against terrorism or al Qaeda.
Pew discovered by varying the amount of money offered that Muslims in the U.S. would not say terrorism is bad or denounce al Qaeda for 25 dollars but would for 50 dollars.
For 25 dollars, Muslims in the U.S. would complain about discrimination after 9-11, but would not answer questions about whether they supported al Qaeda or terrorism.
The conclusion of the Pew Experimental Design was that it takes 50 dollars to get a Muslim in the U.S. to say they are against terrorism or al Qaeda and 25 dollars is not enough. This is the proper statement of the conclusion of the Pew exercise.
Its reasonable to infer from this that Muslims in the U.S. support al Qaeda and terrorism and killing Americans.
It is reasonable to infer from this that Virgil Goode was right, stop Muslim immigration.
Note all the comment made by J.S. in this thread on providing name and address.
Posted by: J.S. at May 23, 2007 10:38 AM
The table on page 11 of the pdf on criticism of US foreign policy may be the closer representation of actual feelings.
Criticism of U.S. Foreign Policy
US Muslims vs. General public*
War in Iraq % %
Right decision 12 vs. 45
Wrong decision 75 vs. 47
DK/Refused 13 vs. 8
100 vs. 100
War in Afghanistan
Right decision 35 vs. 61
Wrong decision 48 vs. 29
DK/Refused 17 vs. 10
100 vs. 100
U.S. War on Terrorism
Sincere effort 26 vs. 67
Not sincere effort 55 vs. 25
Mixed/DK/Refused 19 vs. 8
100 vs. 100
*General public comparisons were taken
from the following Pew nationwide
surveys, respectively: April 2007,
December 2006, March 2004.
From Jan to April 2007, when these questions were asked, they felt safe to be this critical. The criticism of the war on terror, 55 percent “not sincere” sounds sincere.
“Nearly 60,000 respondents were interviewed to find a representative sample
We might restate the conclusion a little. It takes 50 dollars a person to troll through 60,000 Muslims
to find over 500 out of a 1000 willing to say that they are against al Qaeda and terrorism. Paying 25 dollars isn’t enough to get them to answer questions on terrorism, that will only get them to complain about discrimination after 9-11.
In the end, paying them 50 dollars, they still came up with fewer than 1000 Muslims willing to say they were against al Qaeda and killing Americans.
If we take the under 1000 willing to say they were against terrorism and al Qaeda out of the 60,000, we have less than 2 percent of Muslims in the US out of the sample they started with were willing to denounce al Qaeda and killing Americans.
To get even that many they had to increase the payment for answering questions about terrorism to 50 dollars, because at 25 dollars they stopped answering questions after the discrimination questions.
They also had to recontact an old survey group whose characteristics they knew were much more friendly.
We can look at this as a search for at least 500 US Muslims willing to denounce al Qaeda and terrorism. To get that many they had to up the money for saying specifically that from answering questions on discrimination and reuse an old sample that they knew was more friendly.
Younger Muslims support suicide bombing in greater numbers
“Similarly, a plurality (48 percent) of Muslim Americans also said they believed the war in Afghanistan was a mistake, compared to less than 30 percent of the general public.”
“Only 22 percent (of Arab born Muslims) said they believed that Arabs carried out the 9/11 attacks, compared to 40 percent of all Muslim Americans.”
The response on this question is like the leading indicator in business cycle prediction. The response on “Is suicide bombing ever justified” is a lagging indicator. Also, some Muslims may not accept the term suicide bombing to describe 9-11, 7-7, etc. The Pew survey may not be a culturally valid survey for Muslims, especially non-convert Muslims.
A better question might have been: Is it ever justified to kill non-Muslims? Is it ever justified to kill non-Muslims in the United States?
The following argues that the Left is attracted to Islam precisely because it is a non-Western religion. Leftist awakenings have a track record of wanting to get rid of the people who embody the old culture. The above survey, to the true leftist, may be good news. It shows that Muslim immigration is working, it is creating irreversible change in culture by changing the people.
One million American Muslims are Muslims first, Americans second
“Q.E12 Do you think of yourself first as an American or first as a Muslim?” Page 93
“14 In April 2006, the question asked Christians living in the U.S., “Do you think of yourself first as American
or first as a Christian?”” Page 93 footnote.
28 == 48 American First
47 == 42 Religion First
18 == 7 Both equally
6 == 1 Neither/Other
1 == 2 Don’t Know Refused
This compares the response for Muslims in the US
to Christians in the US from a 2006 survey.
The symbol == is used to separate the columns, since I couldn’t find a tab and didn’t want to figure out a table in html.
Page 90 (of pdf, all my page numbers refer to pdf page number. Official page number is 84.)
IF BELIEVE KORAN IS WORD OF GOD (Q.E4=1), ASK:
Q.E5 And would you say that [READ, IN ORDER]?
Muslims == General Public (Bible)
86 Koran == 69 The Bible is the word of God, (NET)
50 Koran == 35 The Bible is to be taken literally, word for word,
25 Koran == 28 That not everything in the Bible should be taken literally, word for word.
11 == 6 Other/Don’t know/Refused (VOL. DO NOT READ)
8 Koran == 22 The Bible is a book written by men and is not the word of God (NET)
1 == 2 Other (VOL. DO NOT READ)
5 == 7 Don’t know/Refused (VOL. DO NOT READ)
GP footnote 10 In March 2007, both general public questions asked about “the Bible” instead of the Koran.
The Muslim American study was funded by a generous grant from The Pew Charitable Trusts and was
conducted jointly by two of the Pew Research Center’s projects: The Pew Research Center for the People
& the Press and the Pew Forum on Religion & Public Life, with additional advice and assistance from
staff in the Pew Hispanic Center and the Pew Global Attitudes Project.
Pew Research Center
Andrew Kohut President
Paul Taylor Executive Vice President
Elizabeth Mueller Gross Vice President
Scott Keeter Director of Survey Research
Richard Morin Senior Editor
Vidya Krishnamurthy Communications Manager
Pew Research Center for the People & the Press
Andrew Kohut Director
Carroll Doherty Associate Director, Editorial
Michael Dimock Associate Director, Research
Richard Wike Senior Researcher
Nilanthi Samaranayake Survey and Data Manager
Juliana Horowitz, Rob Suls, Shawn Neidorf Research Associates
James Albrittain Executive Assistant
Pew Forum on Religion & Public Life
Luis Lugo Director
Sandra Stencel Deputy Director
John C. Green Senior Fellow in Religion & American Politics
Gregory Smith Research Fellow
Daniel A. Cox Research Associate
Sahar Chaudhry Program Assistant
==Other comments on this poll
23 May 2007 Vdare
Muslim Fifth Column Polled
One in four Muslims sympathises with motives of terrorists
By Anthony King
The group portrait of British Muslims painted by YouGov’s survey for The Daily Telegraph is at once reassuring and disturbing, in some ways even alarming.
May 23, 2007
Pew propaganda spin
Are Muslim Americans mostly middle class and mostly mainstream, as headlined by the Pew Research Center? Is there any other group which would be surveyed for which the labels “moderate” and “assimilated” would be used when more than one quarter of those under age 30 support suicide bombing attacks, and only 40% believe that the 9/11 attacks were carried out by Muslim terrorists?
Cheerleaders for Pew:
.Survey: U.S. Muslims Assimilated, Opposed to Extremism.
By Alan Cooperman
Washington Post Staff Writer
Wednesday, May 23, 2007; Page A03
“Though socially conservative, Muslims lean toward the Democratic Party, six to one.”
“One of the poll’s most striking findings, Kohut said, is that African American Muslims are considerably more likely than immigrant Muslims to express support for al-Qaeda.”
Will SPLC go on Lou Dobbs to point out the flaws in the Pew Poll conclusions?
In particular, will Richard Cohen and Mark Potok of the Southern Poverty Law Center go on Lou Dobbs to correct the flawed conclusions of the Pew poll? They are good guys, they will want to correct this.
Pew polls tend to be puzzles. Their study on immigration and wages in 2006 was a similar puzzle set to the public. They explicitly said they were not determining if immigration lowered wages overall in the US, just that where immigration was high didn’t result in lower wages there, according to their study. Even if we accepted that result, we would still have the issue that the impact on wages was felt everywhere, which would explain why there is opposition to immigration everywhere. Even if immigration impacted the place of immigration first, immigration started in 1965 so there has been plenty of time for the effects to spread out evenly over the whole country.