FINAL_BOOK_PDF_Veronika,_Sophia_and_Clark_

Research Notes

Tracking Radical Opinions in Polls of U.S. Muslims

by Veronika Fajmonová, Sophia Moskalenko and Clark McCauley

Abstract

This Research Note examines two telephone polls (2007, 2011) and three Internet polls (2016) to track opinions of U.S. Muslims relating to the war on terrorism. Results indicate that a small but consistent minority (five to ten percent) justify suicide bombing of civilians in defense of Islam, while those seeing the war on terrorism as a war on Islam have declined from more than half to about a third. This decline coincided with a decline in perception of discrimination against Muslims in the U.S., and correlational results confirm that perceived discrimination is one source of seeing the war on terrorism as a war on Islam. Other results from both the Pew and Internet polls show that disapproval of U.S. foreign policies affecting Muslims also contributes to seeing a war on Islam. Discussion emphasizes the value of Internet polling for tracking shifts in the opinions of U.S. Muslims, but acknowledges that polling has not yet discovered what is different about the small minority who justify suicide bombing.

Keywords: Opinion Polls, U.S. Muslims, war on terror, Internet polling

Introduction

It has long been recognized that effective counter-terrorism programs require more than finding, fixing, and finishing those who have turned to political violence. Beyond radical action, there are radical beliefs that can inspire and justify political violence. These beliefs define the battleground of the ‘war of ideas’. Some indication of the importance attached to this war is the fact that googling “Terrorism ‘war of ideas’” in March 2017 produced over seven million hits.

In this Research Note we first consider the jihadist narrative, then track the elements of this narrative in beliefs of U.S. Muslims as represented in two Pew telephone polls and three smaller Internet polls. We are interested in two questions. First, what changes over time can be discerned in opinions of U.S. Muslims relating to the war on terrorism? Are radical opinions becoming more common, less common, or staying the same? Second, are there any useful predictors of radical opinions related to the jihadist narrative? What do we know about how and why some U.S. Muslims maintain radical opinions?

What is the Jihadist Narrative?

Thomas Johnson identified jihadist narrative themes from analysis of primary Taliban sources: “An appeal to past Afghan struggles against ‘foreign invaders.’” “The battle between the Taliban and the Karzai ‘puppet’ regime and its foreign coalition represents a ‘cosmic conflict’ between the ‘righteous’ and the infidel.” “Afghans have a collective religious responsibility to fight the apostates and invaders.”[1]

David Betz offers a similar summary of the elements of the narrative of Global Jihad: “(1) Islam is under general unjust attack by Western crusaders led by the United States; (2) Jihadis, whom the West refers to as ‘terrorists,’ are defending against this attack; (3) the actions they take in defence of Islam are proportionally just and religiously sanctified; and, therefore (4) it is the duty of good Muslims to support these actions.”[2]

The narrative of Global Jihad is conveniently represented as a pyramid of radicalization (Figure 1) in which the base includes Muslims who currently do not accept any of the narrative. The layer above the base includes those who sympathize with the first element of the jihadist frame: that the West is waging a war on Islam (narrative level 1, pyramid level 2). Next higher in the pyramid are Muslims who believe that jihadis are acting in defense of Islam and that their actions are morally and religiously justified (narrative elements 2 and 3, pyramid level 3). Highest in the pyramid of opinion radicalization are Muslims who believe there is an individual duty to support and participate in the defense of Islam (narrative element 4, pyramid level 4.

Figure 1. Opinion Radicalization Pyramid

Tracking Radical Opinions in the 2007 and 2011 Pew Polls of U.S. Muslims

There is reason to believe that different Muslim communities in the US may have different views of political issues. Such differences might arise initially from the political contexts that moved different groups to emigrate from their homelands. Many Iranians came to the US, for instance, when the Shah was toppled. Somalis came to the US when famine and violence made their homeland a dangerous place. But experience in their new homes must also play a role in determining immigrants’ political and social attitudes. For example, some believe that the European experience of jihadist terrorism has been worse than the US experience because Europe is less accepting of immigrants.[3]

Recognizing that immigrant groups can differ markedly in their cultures and their political viewpoints, McCauley and Scheckter used the 2007 Pew poll of U.S. Muslims to compare seven origin groups defined by country of birth: Pakistanis, Iranians, South Asians, Arabs, sub-Saharan Africans, Europeans, and African-Americans. Results indicated substantial group differences, with Iranian-born and African-Americans standing out from other groups in their political opinions.[4]

In this section we conduct the same kind of origin group comparisons for the 2011 Pew poll of U.S. Muslims, with special attention to possible opinion changes between 2007 and 2011.

Challenges of Polling U.S. Muslims

Muslims in the US constitute less than one percent of the population and different ethnic and origin subgroups of U.S. Muslims are correspondingly smaller percentages. Standard polling methods, such as random-digit dialing or address sampling, are impractical when the target group will comprise less than one percent of those sampled.

Nevertheless, there have been polls of US Muslims using one or both of two approaches to getting the sample.[5] The first approach is to accumulate Muslim respondents who turn up in national polls, using standard probability sampling techniques. Over many U.S. national polls, a polling company identifies a number of Muslim respondents; these individuals can then be contacted all at once in a poll targeting only Muslims. The second approach is to sample randomly from neighborhoods known to have a relatively high proportion of Muslim residents. The odds of calling a Muslim can thereby be raised, although Muslims living outside ‘Muslim neighborhoods’ will likely be under-represented in this approach.

Polls of U.S. Muslims have typically used a combination of the two approaches, but the samples thus obtained have usually been too small to permit fractionating the sample to compare subgroups defined by birthplace. The 2007 and 2011 Pew polls of U.S. Muslims are exceptional in the size of the sample obtained: each Pew poll had over a thousand respondents.[6] Our study takes advantage of these larger samples to compare groups of U.S. Muslims based on country of birth, and to look for change in the pattern of group results between 2007 and 2011.

There is a special challenge in identifying African-American Muslims in the 2007 and 2011 Pew polls. Pew has estimated that 2.35 million Americans are Muslim, and that 20% of these are African-American (indicating approximately 470,000 African-American Muslims).[7] African-American Muslims are typically members of one of several (typically Sunni) groups, including the American Society of Muslims (the descendent of the original Nation of Islam), the new Nation of Islam (now led by Louis Farrakhan), the Five Percenters, the Dar al-Islam, the Islamic Mission of America, and other smaller organizations. Though some use the term “Black Muslims,” historically this term has had specific connotations in relation to the Nation of Islam; in this Research Note we refer to African-American Muslims.[8]

Unfortunately the 2007 and 2011 Pew polls did not explicitly ask whether respondents were African-American. Thus, rather than looking at all Pew respondents born in the US, our analysis focused on African-American Muslims as defined by three criteria: born in the US, both parents born in the U.S., and self-identified as “Black.”

In sum, we compare origin groups in both the 2007 and 2011 Pew Polls of U.S. Muslims, focusing on three questions tapping radical ideas: opinion of the war on terrorism, opinion of Al-Qaeda, and opinion of suicide bombing in defense of Islam.

2007 and 2011 Pew Polls of U.S. Muslims—Methods

In this section we describe the number of respondents in each origin group for the 2007 and 2011 Pew polls of U.S. Muslims, the treatment of missing data, and the use of unweighted data in our analyses.

Overview of respondents in the 2007 and 2011 Pew Polls of U.S. Muslims. The 2007 Pew Poll had 1050 respondents, including 28 coded missing for place of birth and 15 born in countries coded as ‘other. There were 273 respondents reporting they were born in the US; 135 were identified as African-American Muslims and the remaining 138 respondents do not appear in this report. The total of 2007 Pew Poll respondents represented in this report is thus 869 (1050-28-15-138).

The 2011 Pew Poll had 1033 respondents, including 19 missing place of birth and 95 born in countries coded by Pew as ‘other’. There were 289 respondents reporting they were born in the US; 110 were identified as African-American Muslims and the remaining 179 do not appear in this report. The total of 2011 Pew Poll respondents represented in this report is thus 740 (1033-19-95-179).

Groups defined by birthplace. This report follows the procedures used by McCauley and Scheckter with the 2007 Pew poll to identify six origin groups among respondents of the 2011 Pew poll.[9] In 2011 as in 2007, two countries had large enough numbers of emigrants to form separate groups: Pakistan (158) and Iran (58). South Asia (Afghanistan, Bangladesh, India, Nepal, Sri Lanka) had 125 emigrants, Arab countries had 219 emigrants, and Sub-Saharan Africa had 70 emigrants. In addition, we identified African-American Muslims as those who self-identified as Black and who reported being born in the US with both father and mother also born in the U.S.

In their analysis of the 2007 poll, McCauley and Scheckter identified a seventh origin group: emigrants from European countries (59 respondents).[10] In 2011, however, there were only about 4% of foreign-born U.S. Muslims coming from a European country, and Pew coded such respondents in 2011 as other. European emigrants were thus not included in our comparison of 2007 and 2011 results.

Missing data. In this report tabled percentages are calculated without missing responses (‘don’t know’ or ‘refused’) in the denominator.

Weighted vs. unweighted data. In their analysis of the 2007 Pew poll of U.S. Muslims, McCauley and Scheckter used unweighted data.[11] In the present study, we compared results using both weighted and unweighted data for both 2007 and 2011. Results were generally similar but there were some anomalous results using weighting.

Pew recommends using weighted data to correct for sampling biases so that results can be more representative of the population sampled. In the achieved samples of U.S. Muslims in both 2007 and 2011, younger individuals with more education were over-represented and older individuals with less education were under-represented.[12] Higher weights are thus given to older respondents with less education.

One such individual from South Asia reported himself as over fifty years of age, with a high school education, and as a convert to Islam. His responses were weighted x6 and he was alone responsible for a change in the percent of South Asian converts from two percent with unweighted data to eight percent with weighted data. He also had a big influence on the South Asian correlation of convert status with approval of Al-Qaeda: the correlation was negligible with unweighted data but .89 with weighted data.

Although weighting may be useful when results are aggregated across a thousand respondents, weighted data can produce misleading results for analysis of a hundred respondents in a particular origin group. In this Research Note we report results using unweighted data.

2007 and 2011 Pew polls of U.S. Muslims—Results

Here we focus on perceptions of the war on terrorism. A full report of 2007 and 2011 results for items relating to demographics, religiosity, perceived discrimination, and opinions of government policies can be found on the START website.[13] For these items not reported here, 2007 and 2011 results were generally similar except that presidential approval was substantially higher for President Obama (2011) than for President George W. Bush (2007) and satisfaction with how things are going in the country was likewise higher in 2011 than in 2007.

Opinions relating to the war on terrorism. Table 1 shows that in both 2007 and 2011 about half of U.S. Muslims (2007 49-81%; 2011 39-50%) did not believe that the war on terrorism (WOT) is a sincere effort to reduce international terrorism. Two groups showed a substantial decline in doubts about the war on terrorism (African-Americans 81% in 2007 vs. 50% in 2011, Iranians 66% in 2007 vs. 40% in 2011); indeed every origin group showed a numeric decline in doubts about war on terrorism. This decline is likely associated with the massive increases in presidential approval from 2007 to 2011.

Table 1. Opinions related to the war on terrorism. Percent italicized responses.

Favorable opinion of Al-Qaeda (AQ) was low in both polls (1-12% in 2007 and 0-6% in 2011). Similarly, justifying suicide bombing (SB) was low in both polls: 2-10% in 2007 and 2-9% in 2011.

The three terrorism-related items have non-negligible missing rates, raising the possibility that responses may be biased by fear of the consequences of endorsing pro-terrorist opinions. But the pattern of missing does not support this possibility. Table 1 shows that percentages of missing responses for the three terrorism-related items are similar in 2007 and 2011, with 9-22% missing for the WOT item, 9-27% missing for the AQ item, and 0-11% missing for the suicide bombing item. We believe that the most threatening of the three items asks about justifying suicide bombing in defense of Islam, yet this item has lower missing rates than the WOT item and AQ item. Overall, the pattern of missing data suggests that missing rates reflect more honest ignorance than defensiveness.

Consistent with this interpretation is the fact that six of six origin groups showed a decline in percent missing on the AQ item from 2007 to 2011. There is no reason defensiveness would decline, but ignorance might decline as Al-Qaeda continued to be salient in the news during these years. Indeed bin Laden was killed in May 2011 during the Pew poll conducted from 14 April to 22 July 2011.

Predicting radical opinions. As possible predictors of radical opinions we examined eight demographic items and fourteen opinion items (six religiosity items, five discrimination items, three items evaluating the U.S. government and its actions).

Based on substantial correlations between AQ and SB items (.24 in 2007 and .32 in 2011, with missing recoded as described below) we created a RadOpinion scale as the mean of responses to these two items. (The war on terrorism item was not correlated with AQ or SB items in 2007 or 2011). Thus four regression models were calculated: for both 2007 and 2011, a model predicting opinion of the war on terrorism and a model predicting RadOpinion.

To keep the number of respondents constant across the terrorism-related items, missing values for these three items were recoded as mid-scale values. For instance, missing values for the SB items were recoded as 2.5 on the four-point scale for this item (often, sometimes, rarely, never justified). Also, to control for mean opinion differences across origin groups, these groups were coded as dummy variables (African-Americans the uncoded comparison group), and these dummy variables were included in each of the four regression models. With so many predictors, we used a conservative level of significance and present here only predictors with a beta significant at p<.01. Complete results of the regression models are available from the authors.

Table 2 shows that in 2007 there are four significant predictors of seeing the war on terrorism as insincere: disapproval of President George W. Bush (beta .19), seeing military force in Afghanistan as a wrong decision (beta .16), seeing discrimination in government surveillance (beta .15), and seeing media unfair to Muslims (beta .13). In 2011 opinions are more crystalized. There is one outstanding predictor: seeing military force in Afghanistan as a wrong decision (beta .34). The adjusted R squares (.20 in 2007; .22 in 2011) indicate that, despite numerous and varied predictors, the level of prediction is only moderate.

Table 2: Regression predicting seeing the war on terrorism as insincere in 2007 and 2011

Beta

Predictor

2007

2011

Media unfair to Muslims

.13**

.06

Discrimination in USG surveillance - yes

.15**

.08

Disapprove president

.19**

.11

Afghanistan was wrong decision

.16**

.34**

R square

.25

.27

Adjusted R square

.20

.22

N

425

381

 

Note. Only predictors with beta significant at p<.01** are tabled, except predictors significant only in 2007 or 2011 show also the non-significant beta for comparison. Missing values were excluded within predictors, but recoded as mid-value for the war on terrorism item predicted.

Table 3 shows that, in 2007 there is one significant predictor of RadOpinion: low education (beta .23). In 2011 there are two significant predictors: low education (beta .20) and feeling physically threatened as a Muslim in the U.S. (beta .23). Again the adjusted R squares (.16 in 2007; .18 in 2011) indicate that the level of prediction is only moderate.

Table 3: Regression predicting RadOpinion Scale in 2007 and 2011

Beta

Predictor

2007

2011

Education - low

.23**

.20**

Being physically threatened

.04

.23**

R square

.21

.24

Adjusted R square

.16

.18

N

425

381

 

Note. Only predictors with beta significant at p<.01** are tabled. Missing values were excluded within predictors, but recoded as mid-value for items averaged in the RadOpinion Scale predicted.

2007 and 2011 Pew polls of U.S. Muslims—Discussion

We used the 2007 and 2011 Pew polls of U.S. Muslims to examine three opinions related to the war on terrorism: seeing the war on terrorism as insincere, favorable opinion of Al-Qaeda, and justifying suicide bombing in defense of Islam. We compared results for these items across time (2007 vs. 2011) and across six origin groups (Pakistan, Iran, South Asian countries except Pakistan, Arab countries, sub-Saharan African countries, and African-Americans).

Stability and change over time. U.S. Muslims maintained their overwhelming disapproval of Al-Qaeda and suicide bombing; across origin groups and years, approval rates ranged from zero to twelve percent. Less reassuring are the results for opinions of the war on terrorism: although every origin group showed at least a small 2007-2011 decline in the percentage seeing the war on terrorism as insincere, in 2011 respondents seeing the war on terrorism as insincere still ranged from 39-50 percent. That is, in 2011 close to half of every origin group continued to see the war on terrorism as insincere.

Predicting opinion that the war on terrorism is insincere. In 2007 there were four significant predictors of seeing the war on terrorism as insincere: disapproval of President Bush, seeing military force in Afghanistan as a wrong decision, seeing discrimination in government surveillance of Muslims, and seeing media unfair to Muslims. In 2011 there was only one significant predictor: seeing military force in Afghanistan as a wrong decision predicts opinion of the war on terrorism as well as all four predictors taken together in 2007.

Our interpretation of this pattern is that disapproval of the war on terrorism in 2007 was part of a broad mix of disapproval of the government and its actions, whereas in 2011, under President Obama, the war on terrorism had crystalized to mean military intervention in Muslim countries.

Predicting RadOpinion. In 2007 only low education was a significant predictor of RadOpinion, that is, the mean of opinion of Al-Qaeda and opinion of suicide bombing. In 2011 there were two significant predictors: low education and reporting being physically threatened or attacked because of being Muslim. The one consistent predictor, low education, is not statistically strong; we do not yet understand why a small minority of U.S. Muslims have this radical opinion.

Group Differences. Broken down by origin group, results for two groups invite special attention: African-Americans and Iran-born.

Table 1 shows a small but consistent tendency for African-American Muslims to have more radical opinions than other origin groups. In 2007 and again in 2011, African-American respondents were most negative toward the war on terrorism, most positive toward Al Qaeda, and most positive toward suicide bombing. This consistency suggests that U.S. Muslims may not be a homogenous population and that generalizations about “U.S. Muslims” may be misleading.[14]

Demographically, Iran-born Muslims stand out in both 2007 and 2011. They are older than other origin groups, less religious than other groups on five of our six measures of religiosity, and they are Shi’a rather than Sunni. Despite these differences, their opinions of U.S. government policies look like the opinions of other origin groups. Notably, they are no less likely than other groups to see the war on terrorism as insincere, and no different in their opinions of Al-Qaeda and suicide bombing.

These results are important in showing just how broadly U.S. Muslims disapprove of the war on terrorism. Even Shi’a Muslims in the U.S.–who have sectarian reasons for disliking Sunni-based Al-Qaeda–even Shi’a Muslims show substantial disapproval of the war on terrorism (insincere 66 percent in 2007, 40 percent in 2011).

In brief, results from the Pew 2007 and 2011 polls of U.S. Muslims have provided one strong predictor of opinion of the war on terrorism—disapproval of U.S. troops in Afghanistan–but only weak and inconsistent predictors of opinions of Al Qaeda and suicide bombing. Seeking better predictors and more recent assessment of the opinions of U.S. Muslims, we turn now to three Internet polls of U.S. Muslims.

Tracking Radical Opinions in Internet Polls of U.S. Muslims

The research presented in this section is part of a multi-year research project exploring the use of Internet polling as a tool to access opinions and attitudes of U.S. Muslims.

Internet polls of U.S. Muslims--Methods

Panel recruitment. Samples reported here were recruited and data collected by Knowledge Networks (KN), a division of international market research corporation GfK. Sampling begins from the KnowledgePanel, whose members are recruited through national random samples of the U.S., originally by telephone and now almost entirely by postal mail. Households are provided with access to the Internet and a netbook computer, if needed. KnowledgePanel is thus representative of the U.S. adult population, with panel members from listed and unlisted telephone numbers, from telephone, non-telephone and cell-phone-only households, as well as households with and without Internet access.

The target population for our research consists of non-institutionalized adults age 18 and over residing in the United States who identify as Muslim. To sample this population, KN identified Muslim households from the KnowledgePanel and solicited their participation in our study, which has so far included six waves of polling. The results reported here from Waves 4, 5, and 6 had respectively 90, 88, and 87 respondents from KnowledgePanel (about a 50 percent return rate from the panel samples).

In addition, KN recruited ‘opt-in” (volunteer) Muslim respondents from other polling vendors that included respectively 121, 124, and 127 respondents (about a one percent return rate from the opt-in solicitation). The opt-in sample was included to raise the number of respondents to 200, to give more stability to percentages and correlations in our analyses. The cost of this sample enlargement is the threat of unrepresentative results from the opt-in sample. Comparison of results for KnowledgePanel participants with results from the combined panel and opt-in participants suggests that this threat is minimal. This comparison and additional details about KnowledgePanel procedures are available in an overview of Waves 1-3 available on the START website.[15]

For a multi-wave tracking poll, stability of sample characteristics is perhaps as important as representativeness. Even a sample that is not fully representative can capture change in opinion if the characteristics of the sample do not change over time. Our goal was a sampling procedure that could be repeated to produce stable sample characteristics such that at least large changes in opinion over time could be detected. Our design goal was a stability that would permit confidence in measuring changes of opinion of more than 15 percentage points. This kind of stability is evident in the Results reported here for three waves of polling of U.S. Muslims.

The three polls were fielded as follows: Wave 4 28 January–17 February 2016, Wave 5 26 May–15 June 2016, Wave 6 28 October–8 November 2016. These dates are abbreviated to the closest month in Table 4.

Poll items. As with the Pew polls, we focus on two items assessing opinion radicalization of U.S. Muslims: a question about the war on terrorism and a question about suicide bombing and other attacks on civilians in defense of Islam.

Internet Polls of U.S. Muslims—Results

We first look for change over time in opinions about the war on terrorism and suicide bombing, then look for predictors of opinion on these two items. In comparisons across waves we focus on substantial differences, differences of fifteen percentage points or more (statistically, for two groups each with n=210, differences of 15 percentage points are significant at p<.01 two tailed).

Tracking opinions over time. Rather than asking about whether the war on terrorism is insincere, as the Pew polls did, we asked more directly Do you feel the war on terrorism is a war against Islam? (Yes; No; Not sure/Don’t know) Table 4 shows that there was a substantial decrease in the percent of U.S. Muslims saying ‘yes’ to this question: from 47 percent in January 2016 to 30 percent in June 2016 and 32 percent in October 2016.

Table 4: Radical opinions in three waves of internet polling of U.S. Muslims

Note. Items not included in the Wave 5 poll are indicated as --.

Our internet polls used the same question about justifying suicide bombing and other attacks on civilians used in the Pew polls (See Table 1). Table 4 shows no change over time in response to this question: the percent seeing suicide bombing as often or sometimes justified was 10 percent, 8 percent, and 9 percent across Waves 4-6. It is worth noting that this kind of consistency over time indicates that our sample characteristics are stable over time, and adds confidence to the demonstration of substantial change for opinions about the war on terrorism noted in the preceding paragraph.

The 2015 jihadist attacks in Paris and San Bernardino prompted us to include in Waves 4 and 5 three questions about experience of discrimination. The first question is general. Since the 2015 terrorist attacks in Paris and San Bernardino, has it become more difficult to be a Muslim in the U.S., or hasn’t it changed very much? (Has become more difficult to be a Muslim in the U.S.; Hasn’t changed very much; Not sure/Don’t know).

The second and third questions are about personal experience. Since the 2015 terrorist attacks in Paris and San Bernardino, have people acted as if they are suspicious of you? (Yes, has happened; No, has not happened; Not sure/Don’t know). Since the 2015 terrorist attacks in Paris and San Bernardino, have you been called offensive names? (Yes, has happened; No, has not happened; Not sure/Don’t know).

Table 4 shows substantial change in opinion for two of the three discrimination items. From Wave 4 to Wave 5, reports of suspicion and offensive names declined respectively from 54 percent to 27 percent and from 73 percent to 19 percent. These declines occurred despite the same time reference: “Since the 2015 terrorist attacks in Paris and San Bernardino.” Of course respondents may be reporting feelings of being discriminated against, rather than recalling specific incidents of discrimination. But the question remains, what explains the substantial changes of report on these two items?

It seems possible that the perception of suspicion and hostility toward U.S. Muslims peaked shortly after the San Bernardino attack on the first of December 2015. Our January 2016 poll could catch this peak, but by the time of our June 2016 poll U.S. Muslims had recovered much of their confidence that life in the U.S. was not so threatening after all.

Finally, we included in Waves 4-6 an item about U.S. foreign policies. Some people say that U.S. foreign policies are dictated by Jewish interests. How do you feel about this? (Agree; Disagree; Not sure). Table 4 shows substantial change in opinion on this question. From 21 percent in January 2016, agreement increased to 43 percent and 47 percent in June and October 2016.

This opinion change may be related to the rise of Donald Trump, who was not seen as a serious candidate in January 2016 but by June was reaching for the Republican nomination he won in July and by October was reaching for the presidency he won in November. In March 2016 both Donald Trump and Hillary Clinton appeared before the American Israel Public Affairs Committee; both promised security for Israel but Trump went further in promising to move the capital of Israel to Jerusalem. He ended with the thought that his daughter Ivanka was about to have a beautiful Jewish baby. It is possible then that the political ascendance of Donald Trump led U.S. Muslims to see increasing Jewish influence in U.S. foreign policy.

To summarize what we have learned about radical opinions of U.S. Muslims: perception of war on Islam decreased between January and June 2016 (47 percent to 30 percent), justification of suicide bombing was unchanged (about ten percent justifying).

In the next two sections we look for correlates of opinions on these two items. Our Internet polls had much smaller samples (around 200 respondents) than the Pew polls (around 1000 respondents). Regression analyses using dozens of predictors were therefore not attempted with the Internet polls, but simple correlations were calculated.

Predicting perception of a war on Islam. In order to test whether perceived discrimination might be related to seeing a war on Islam, we combined the three discrimination items into a scale. For the general item, Has become more difficult to be a Muslim in the U.S. was coded 3, Hasn’t changed very much was coded 1, and Not sure/Don’t know was coded 2. For the two personal experience items, Yes was coded 3, No was coded 1, and Not sure/Don’t know was coded 2. For Wave 4 the discrimination scale had M(211)=2.47 with SD=.47 and alpha=.63. For Wave 5 the discrimination scale had M(212)=1.76 with SD=.67 and alpha=.66.

Numerical values were similarly assigned to responses to the questions about a war on Islam, suicide bombing, and seeing U.S. foreign policies dictated by Jewish interests.

The discrimination scale was significantly related to seeing the war on terrorism as a war on Islam, r(209)=.26 p<.01 in Wave 4 and r(210)=.41 p<.01 in Wave 5.

Seeing U.S. policies dictated by Jewish interests was also significantly related to seeing the war on terrorism as a war on Islam: r(209)=.32 p<.01 in Wave 4 and r(210)=.23, p<.01 in Wave 5.

The two predictors were not consistently correlated: discrimination with policies dictated by Jewish interests was r(209)=.07 ns in Wave 4 and r(210)=.22, p<.01 in Wave 5.

Taken together these results indicate that there are two different sources of the opinion that the war on terrorism is a war on Islam: perceived discrimination and opposition to U.S. foreign policies.

Predicting justification of suicide bombing. Results were disappointing. For the Pew polls, low education was a consistent if weak predictor of the RadOpinion scale that averaged opinion of suicide bombing and opinion of Al-Qaeda. But for the January, June, and October 2016 Internet polls, correlations of education with opinion of suicide bombing were inconsistent and small, never reaching the .01 level of significance. Younger age was significantly correlated with justifying suicide bombing in the June and October 2016 polls: r(210)=-.25 p<.01 and r(214)=-.18 p<.01. But the correlation of age and opinion of suicide bombing in the January 2016 poll was not significant at r(209)=-.09.

Thus, across both the Pew polls and the Internet polls, there is no strong and consistent correlate of justifying suicide bombing. Although perceived discrimination and seeing U.S. foreign policy as dictated by Jewish interests predicted seeing the war on terrorism as a war on Islam, they did not predict justifying suicide bombing.

General Discussion

This report examined two kinds of polling data—telephone polls and internet polls—in order to track and understand radical opinions held by some U.S. Muslims.

We began by examining opinions about the war on terrorism, suicide bombing, and al Qaeda as assessed in the 2007 and 2011 Pew telephone polls of U.S. Muslims. We found some small differences by origin groups, especially for African-American Muslims, but in general results indicated change over time only for opinions of the war on terrorism: a majority of U.S. Muslims saw this war as ‘insincere’ in 2007 but only about half saw it as insincere in 2011. Opinions of al Qaeda and suicide bombing showed no change: in both 2007 and 2011, a small but persistent minority of less than ten percent of respondents had favorable views of al Qaeda or justified suicide bombing.

We used correlational analyses to try to understand what was different about respondents with radical opinions. For opinions of the war on terrorism, results indicated that a consistent and strong predictor was negative opinion of U.S. foreign policy as represented by disapproval of U.S. troops in Afghanistan. For opinions of al Qaeda and suicide bombing, correlational analyses were less successful: low education was the only consistent predictor of radical opinions on these items, and was only a weak predictor.

We turned then to examining opinions of the war on terrorism and of suicide bombing in three internet polls of U.S. Muslims conducted in 2016. Results showed substantial change in opinion of the war on terrorism: in January 2016 about half of U.S. Muslims saw a war on Islam but in June and October 2016 only about a third saw a war on Islam. This change in opinion coincided with a decrease in perceived discrimination, and correlational analyses within both the January and June 2016 polls showed that indeed individuals feeling less discrimination were less likely to see the war on terrorism as a war on Islam. We conclude that at least part of seeing a war on Islam comes from feeling discrimination in the U.S.

But another part seems to come from disapproval of U.S. foreign policies. Seeing U.S. foreign policies as dictated by Jewish interests was correlated with seeing a war on Islam in both the January and June 2016 internet polls. Confirming the relation between foreign policy and seeing a war on Islam is the observation from the Pew polls that disapproval of U.S. troops in Afghanistan was the best predictor of seeing a war on Islam.

It appears then that, for U.S. Muslims, there are two sources of seeing the war on terrorism as a war on Islam: perception of discrimination against Muslims in the U.S. and disapproval of U.S. policies relating to Muslims abroad. If this conclusion is correct, then the war of ideas might try to attack perception of a war on Islam in one or both of two directions. One direction would be to reduce the reality or at least the perception of discrimination against Muslims in the U.S. The other direction would be to reduce U.S. interventions in predominantly Muslim countries, or at least to explain better the need for such interventions.

Numerous polls of U.S. Muslims have assessed opinions representing the sympathizer and justifier levels of the opinion radicalization pyramid, and these opinions—about the war on terrorism and about suicide bombing and other attacks on civilians in defense of Islam—were the focus of the current study. We examine first the 2007 and 2011 Pew telephone polls of Opinions about suicide bombing against civilians in defense of Islam remain opaque. There has not been any change in opinions justifying suicide bombing: Internet polls in January, June, and October 2016 are like the Pew polls in 2007 and 2011 in showing a persistent minority, ten percent or less of U.S. Muslims, justifying suicide bombing. And we have not yet identified a strong or consistent predictor of opinions about suicide bombing. Better understanding of how a small minority of U.S. Muslims persist in justifying suicide bombing should be a priority for future research.

Finally, our results indicate that small-sample internet polling can track at least large shifts in the opinions of U.S. Muslims, that is, shifts of fifteen percentage points or more. The pattern of stability in opinions about suicide bombing gives confidence to observed changes in opinions about the war on terrorism, about discrimination, and about U.S. foreign policy. As an alternative to the traditional telephone polling, Internet polling offers the advantages of lower cost and faster turn-around. These advantages can be especially useful for assessing reactions to unfolding political events, or when longer-term changes require repeated assessment to observe trends.

About the Authors:

Veronika Fajmonová is a PhD candidate focusing on perception of terrorism threat at the Jan Masaryk Center of International Studies at the University of Economics in Prague.

Sophia Moskalenko is a postdoctoral research fellow at the National Consortium for Study of Terrorism and Responses to Terrorism (START).

Clark McCauley is Research Professor of Psychology at Bryn Mawr College and a Principal Investigator with the National Consortium for Study of Terrorism and Responses to Terrorism (START).

Acknowledgement

This research was supported by the Science and Technology Directorate of the U.S. Department of Homeland Security through awards made to the National Consortium for Study of Terrorism and Responses to Terrorism (START). The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the U.S. Department of Homeland Security or START

Notes

[1]Johnson, T. (2007). “The Taliban Insurgency and an Analysis of Shabnamah (‘Night Letters’). Small Wars and Insurgencies, 18(3), 339.

[2] Betz, D. (2008). “The Virtual Dimension of Contemporary Insurgency and Counterinsurgency.” Small Wars and Insurgencies, 19(4), 520.

[3] Fukuyama, F. (2006). “Europe vs. Radical Islam: Alarmist Americans Have Mostly Bad Advice for Europeans.” Slate, February 27. URL: www.slate.com/id/2136964 Also Sageman, M. (2008). Leaderless Jihad. Philadelphia: University of Pennsylvania Press.

[4] McCauley, C. & Scheckter, S. (2011). “Reactions to the War on Terrorism: Origin – Group Differences in the 2007 Pew Poll of U.S. Muslims.” Perspectives on Terrorism, 5 (1), 38-54.

[5] McCauley, C. & Stellar, J. (2009). “U.S. Muslims after 9/11: Poll trends 2001-2007.” Perspectives on Terrorism, 3 (3), 35-47.

[6] Pew Research Center (2007). Muslim Americans: Middle Class and Mostly Mainstream. Research report, 22 May; URL: www.pewresearch.org/assets/pdf/muslim-americans.pdf Pew Research Center (2011). Muslim Americans: No Signs of Growth in Alienation or Support for Extremism. Research Report, 30 August. URL: http://www.people-press.org/2011/08/30/muslim-americans-no-signs-of-growth-in-alienation-or-support-for-extremism/

[7] Pew Research Center (2007).

[8] The Pluralism Project. African-American Islam Reformed: ‘Black Muslims’ and the Universal Ummah. URL: http://pluralism.org/resources/tradition/essays/islam8.php

[9] McCauley & Scheckter (2011).

[10] McCauley & Scheckter (2011).

[11] McCauley & Scheckter (2011).

[12]Fajmonova, V., McCauley, C. (2016). Origin-group differences in the 2007 and 2011 Pew Polls of U.S. Muslims: Reactions to the War on Terrorism. Final Report to the Office of University Programs, Science and Technology Directorate, U.S. Department of Homeland Security, College Park, MD.

URL: www.start.umd.edu/pubs/START_CSTAB_ReactionsWaronTerrorism_Feb2017.pdf

[13] Fajmonova, V., McCauley, C. (2016).

[14] McCauley & Scheckter (2011).

[15] McCauley, C. and Moskalenko, S. (2016). U.S. Muslim Opinion Over Time. Final Report to the Office of University Programs, Science and Technology Directorate, U.S. Department of Homeland Security, College Park, MD. URL: www.start.umd.edu/sites/default/files/publications/local_attachments/START_CSTAB_2.12_MuslimOpinionOverTimeFinalReport_May2016.pdf



Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.

ENHANCING SECURITY THROUGH COLLABORATIVE RESEARCH

Perspectives on Terrorism is  a journal of the Terrorism Research Initiative and the Center for Terrorism and Security Studies

ISSN  2334-3745 (Online)

Disclaimer, Terms and Conditions