Comparing Levels of Extremism in Muslim and Christian Populations in Pew’s (2009) Sub-Saharan Survey Data

In recent articles, using single-question and multi-question set analyses of Pew’s survey data, I examined the percentages of Muslim adults who hold views or endorse practices that are hardline fundamentalist or extremist. Continuing that research here, I compare Christians and Muslims. I will also examine whether Muslims who favor Islamic law are more or less hardline/extremist than Christians who favor the Bible as the law. These analyses, like the previous ones, make use of freely-available statistical software (e.g., PSPP), common spreadsheets, and publicly-available microdata.[3] I gave some basic instructions for using PSPP to analyze Pew datasets in previous articles (here, here, and here). The results reported here are publicly verifiable for readers who have adequate knowledge and skill.

The Pew Forum’s “Tolerance and Tension” survey [1] of sub-Saharan Africa, conducted between December of 2008 and April of 2009, was fielded in 19 countries and represented about 75% of the population in the sub-Saharan region.[2]  As mentioned in Pew’s (2010) report, sub-Saharan Africa is “among the most religious places in the world,” where the vast majority of respondents say that religion is “very important in their lives” (p. 3, Complete Report pdf). The survey’s data [3] are for the most part adequate for comparing the views of Christians and Muslims. (The “Tolerance and Tension” (2010) Complete Report pdf is cited throughout this article. Readers may find it helpful to have that pdf open and available).

Results for Christians and Muslims will be compared on questions concerning severe punishments such as whippings and cutting off of hands for crimes like theft, stoning of people who commit adultery, use of arms and violence against civilians to defend religion, making sharia (Islamic law) or the Bible the official law of the land, whether women should be allowed in religious leadership positions, and whether or not one has had one’s daughter(s) “circumcised.” A comparison regarding the death penalty for apostasy is not possible because the question was only asked of Muslims.

Pew’s weight variable was applied to obtain the response choice percentages for each question or set of questions in each country. The overall response choice percentages for each question or set reported here are weighted according to the sizes of the adult (age 18 and older) populations of Muslims and Christians, respectively, per country. For subsets, including the subsets who favor religious law, or who have daughter(s), or both, response choice percentages are weighted according to the estimated population size of the subset per country, per religious group. I made the population estimates shown in Table 1, below, for Muslim and Christian adults using the method described in Appendix A of a previous article, based on Pew’s religion-specific age structure estimates for each country for 2010.[4]

Table 1

Sample and Estimated Population Sizes of Muslim and Christian Adults in Pew’s (2009) Survey of Sub-Saharan Africa
	Sample	Population
Muslims	8287	82808400
Subset in Favor of Sharia as the Law	5142	52240936
Christians	15324	210540000
Subset in Favor of Bible as the Law	9002	126421065
Notes. Samples and population estimates are of Christians and Muslims age 18 or older.
Sample sizes displayed here are of raw counts (unweighted numbers).

Christian samples and populations exclude Djibouti, Mali, and Senegal because not enough Christians were sampled in those countries (ns ≤ 103; see the Complete Report pdf, pp. 17, 66). Muslim samples and populations exclude Botswana, Rwanda, South Africa, and Zambia because not enough Muslims were sampled in those countries (ns < 100; see the Complete Report, pp. 17, 66). There remain 16 countries for the Christian sample, 15 for the Muslim sample, and 12 where adequate samples were available for both.

Appropriate margins of error for the samples shown in Table 1, above, and Table 9, below, cannot be calculated based on Pew’s published information. However, the samples analyzed in this article are large enough that margin of error is not a major concern for the comparisons. For more on this topic, and some links to online calculators that can be used to get a rough estimate of the margins of error, see Appendix C in my previous article. According to those online calculators, the margins of error for the smallest sample (2968) in Table 9 would be ± 1.8% and for the largest sample (15324) in Table 1 would be ± 0.8%, assuming 95% confidence and a response choice proportion of .5.

Analyses

Summaries

Table 2: Summary Results for Single-Question Analyses

Population-Weighted Percentages of Muslims and Christians in Sub-Saharan Countries Who Favored, Justified, or Affirmed Extremist Items
Item Favored, Justified, or Affirmed	Muslims		Christians
	General	Relig. Law	General	Relig. Law
	Sample	Subset	Sample	Subset
Whippings and cutting off hands	44	56	16	18
Stoning of adulterers	38	49	12	14
Using arms, violence to defend religion*	22	25	15	16
Only men should be religious leaders	72	76	41	40
Had daughter(s) “circumcised”**	19	19	12	13
Religious law as official law of the land	63		60
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.
*Often or sometimes justified = “justified”.
**Includes only respondents who had one or more daughters.
Item wording in this table is simplified. See text and original Pew source for exact and complete wording.

The pattern of results in Table 2 indicates that overall, compared to Muslims, Christians are much less supportive of the harsh punishments, less supportive of violence against civilians to defend religion, more supportive of allowing women (not only men) to be religious leaders, somewhat less likely to have had their daughter(s) “circumcised,” but almost as likely to have favored religious law as the law of the land. Christians who favor making the Bible the official law of the land do not appear to differ much from the general sample of Christians in their support for these items. Compared to Muslims in general, Muslims in favor of making sharia or Islamic law the official law of the land show clearly higher support for the harsh punishments, slightly higher support for using violence against civilians to defend religion and for having only men as religious leaders, but do not appear more likely to have had their daughter(s) “circumcised.”

Table 3: Summary Results for Multi-Question Set Analyses

Population-Weighted Percentages of Muslims and Christians in Sub-Saharan Countries Who Favored at Least One, at Least Two, or Opposed All Extremist Items Per Question Set
Two-, Three-, and Four-Question Sets	Muslims			Christians
	Favor at Least		Oppose	Favor at Least		Oppose
	One	Two	All	One	Two	All
Whippings/cutting, Stoning adulterers	49	33	49	20	8	77
Whippings/cutting; Stoning adulterers; Using violence to defend religion	58	37	30	30	10	51
Whip/cut; Stoning; Violence to defend religion; Only men as religious leaders	88	51	7	59	18	27
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.
Item wording in this table is simplified. See text and original Pew source for exact and complete wording.

The vast majority of Muslims (88%) supports at least one and a slight majority (51%) supports at least two of the set of four extremist items. Among Christians, a majority (59%) supports at least one while a minority (18%) supports at least two of the four extremist items. Comparing the “Favor at least one” results for the three-item set with those for the four-item set, it appears that the additional roughly 30% support among both Muslims and Christians for extremism consists of those who favor only men as religious leaders.

Single-Item Comparisons of the Levels of Extremism/Fundamentalism Among Christians Versus Muslims, and General Samples Versus Religious Law-Favoring Subsets

Making Religious Law the Official Law of the Country

Q94a (p. 285), Q95a (p. 289)

Asked of Christians: Q94 “And do you favor or oppose the following? … a. making the Bible the official law of the land in our country”

Asked of Muslims: Q95 “And do you favor or oppose the following? … a. making sharia, or Islamic law, the official law of the land in our country”

Pew’s response code: 1 = favor, 2 = oppose, 8 = don’t know, 9 = refused.

Table 4

Q95a (Muslims) “…making sharia, or Islamic law, the official law of the land in our country” Q94a (Christians) “…making the Bible the official law of the land in our country”
	Population-Weighted Percentages
	Favor	Oppose	Don’t Know	Refused
Muslims	63	34	2	1
Christians	60	35	4	1
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.

The majority in both groups favors religious law. These majorities are not extremely large and are close to the same size. (They are also consistent with Pew’s median percentages taken across countries, see page 11, Complete Report pdf).

What does each group think is included in religious law? The subsequent analyses will shed some light on this issue.

Punishments Like Whippings and Cutting Off of Hands for Crimes Like Theft and Robbery

Complete Report pdf, Q94c (p. 287), Q95d (p. 292).

Asked of Christians: Q94 “And do you favor or oppose the following? … c. punishments like whippings and cutting off of hands for crimes like theft and robbery”

Q95d, asked of Muslims, is the same as Q94c.

Pew’s response code: 1 = favor, 2 = oppose, 8 = don’t know, 9 = refused.

Table 5

Q94c/Q95d “…punishments like whippings and cutting off of hands for crimes like theft and robbery”
	Population-Weighted Percentages
	Favor	Oppose	Don’t Know	Refused
General Sample
Muslims	44	54	1	1
Christians	16	81	2	1
Subset in Favor
of Religious Law
Muslims	56	42	1	0
Christians	18	80	2	0
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.

The vast majority of Christians opposes these severe punishments. Favoring Biblical law does not make much of a difference. In contrast, a large minority (44%) of Muslims favors these punishments, as does the majority (56%) of the sharia-favoring subset. Clearly, Muslims and Christians differ greatly from each other in their understandings of religious or religiously-based law in regards to these severe punishments. Of course, even the 16% support among Christians for these extremely violent punishments is large enough to be concerning.

Stoning of Adulterers

Complete Report pdf, Q94d (p. 288), Q95e (p. 293).

Asked of Christians: Q94 “And do you favor or oppose the following? … d. stoning people who commit adultery”

Q95e, asked of Muslims, is the same as Q94d.

Pew’s response code: 1 = favor, 2 = oppose, 8 = don’t know, 9 = refused.

Table 6

Q94d/Q95e “…stoning people who commit adultery”
	Population-Weighted Percentages
	Favor	Oppose	Don’t Know	Refused
General Sample
Muslims	38	59	2	1
Christians	12	85	2	1
Subset in Favor
Of Religious Law
Muslims	49	48	2	1
Christians	14	84	2	0
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.

Thus far, the population-weighted results in Tables 5 and 6 are consistent with what Pew indicated (Complete Report pdf, p. 49), i.e., that support for severe punishments overall is higher among Muslims than among Christians. Here the plurality (49%) of sharia-favoring Muslims and a large minority (38%) of the general sample favor stoning people who commit adultery. Only a relatively small minority of Christians overall (12%) favored the punishment and it was not much bigger (14%) in the Biblical law-favoring subset.

Using Arms and Violence Against Civilians to Defend One’s Religion

Complete Report pdf, Q88 (p. 279)

Q 88. “Some people think that the tactic of using arms and violence against civilians in defense of their religion is justified. Other people believe that, no matter what the reason, this kind of violence is never justified. How about you? Do you personally feel that the tactic of using arms and violence against civilians in defense of your religion can be often justified, sometimes justified, rarely justified, or never justified?”

Pew’s response code: 1 = Often justified, 2 = Sometimes justified, 3 = Rarely justified, 4 = Never justified, 8 = Don’t know, 9 = Refused. 

My classification: 1 or 2 = extremist, 4 = moderate. I deemed “rarely” as not moderate, but also as not sufficiently extremist for the analyses in the present article.

Comment: There are some problems with this question. Briefly, “civilians” is not well-defined, “religion” could be understood as a doctrine or as a group of adherents, “defense” is too broad, and “arms and violence” could range from a limited beating that does not produce serious injury to a use of weapons of mass destruction that kills thousands of people.

Table 7

Q88. “… Do you personally feel that the tactic of using arms and violence against civilians in defense of your religion can be often justified, sometimes justified, rarely justified, or never justified?”
	Population-Weighted Percentages
	Often	Sometimes	Rarely	Never	Don’t Know	Refused
General Sample
Muslims	7	15	17	54	6	1
Christians	5	10	14	64	6	1
Subset in Favor
Of Religious Law
Muslims	9	16	18	51	6	1
Christians	5	11	15	63	6	0
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.

However respondents may have construed the various components of the question, 22% of Muslims indicated that such violence was often or sometimes justified, compared to 15% of Christians.* More Christians said such violence was never justified (64%) as compared to Muslims (54%). There was little difference between the general samples and their respective religious law-supporting subsets.

*Compare with Pew’s median percentages taken across countries, p. 47, Complete Report pdf.

Women Allowed Versus Only Men in Religious Leadership Positions

Complete Report pdf, Q59c (p. 195).

Q59 “Now I’m going to read you two statements. Please tell me whether the FIRST statement or the SECOND statement comes closer to your own views — even if neither is exactly right. …c. 

1 - Women should be allowed to serve in religious leadership roles, such as pastor, priest or imam

OR

2 - Only men should be able to serve in religious leadership roles, such as pastor, priest or imam

…”

Pew’s response coding (with my paraphrasing): 1 = women allowed, 2 = only men, 3 = “Neither/Both equally,” 8 = don’t know, 9 = refused

My classification: 1 = moderate view, 2 = extremist view.

Table 8

Q59c. Response choices: “Women should be allowed to serve in religious leadership roles…” vs. “Only men should be able to serve in religious leadership roles…”
	Population-Weighted Percentages
	Women Allowed	Only Men	Neither/Both	Don’t Know	Refused
General Sample
Muslims	22	72	3	2	0
Christians	52	41	5	2	0
Subset in Favor
of Religious Law
Muslims	19	76	3	2	0
Christians	53	40	5	1	0
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.

The results for this question on the status of women show a large difference between Christians and Muslims (also see p. 55, Complete Report pdf). A substantial majority of Muslims (72%) and a large minority of Christians (41%) prefer that only men should serve, rather than to (also) allow women, in religious leadership roles. However, a slight majority (52%) of Christians prefer to (also) allow women, not only men, as religious leaders, while only 22% of Muslims make that preference. There was little difference between Christians generally and the Biblical law-favoring subset. The sharia-favoring subset showed a slightly greater preference for having only men in these roles compared to Muslims generally.

Female Genital Mutilation (FGM)

Complete Report pdf, Q71 (p. 231) Q72d (pp. 54, 235). 

Q71: “Do you have any daughters?”

Pew’s Q71 response code: 1 = Yes, 2 = No, 8 = Don’t know, 9 = Refused.

Table 9

Sample and Estimated Population Sizes of the Subsets of Christians and Muslims Who Have At Least One Daughter
	Sample	Population
Muslims With Daughter(s)	4695	46912713
Subset in Favor of Sharia as the Law	2968	30297416
Christians With Daughter(s)	7613	106116787
Subset in Favor of Bible as the Law	4538	64507539
Notes. Samples and population estimates are of Christians and Muslims age 18 or older.
Sample sizes displayed here are of raw counts (unweighted numbers).

The subset of respondents who answered “Yes” to Q71 were asked (Q72) “And thinking about your child or children, please tell me if you ever do or ever did any of the following things with them. Did you or do you (d). have any of your daughters circumcised?”

Pew’s Q72d response code: 1 = Yes, 2 = No, 3 = Not applicable/Children too young (volunteered), 8 = Don’t know, 9 = Refused.

Table 10

Q72. “Did you … d. have any of your daughters circumcised?”
	Population-Weighted Percentages
	Yes	No	N/A /too young	Don’t Know	Refused
With Daughter(s)
Muslims	19	77	1	0	2
Christians	12	82	4	1	2
Subset in Favor
of Religious Law*
Muslims	19	77	1	1	2
Christians	13	82	3	1	1
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.
*Respondents in this subset have daughter(s) (Q71 = 1) and favor religious law (Q94a = 1) or (Q95a = 1).

Among those with daughter(s), 19% of Muslims and 12% of Christians said they’d had their daughter(s) “circumcised”. There were practically no differences between the general sample and the subset who favored making their religious law the official law of their country. As a point of comparison to the population-weighted results for the general samples with daughter(s), unweighted means of 25% of Muslims (in their 15-country sample) and 11% of Christians (in their 16-country sample) reportedly had their daughter(s) “circumcised.”

Pew did not ask whether the respondent currently approves of FGM. In the region, attitudes toward FGM have been changing. Saying that one has had the practice done to one’s daughters at some point, perhaps many years prior to the survey, doesn’t necessarily mean that one still accepted the practice at the time of the survey. In addition, those who do not have daughters were not asked the question but may approve of the practice, and may have it done at some point in the future when they do have daughters.

Note that these percentages differ somewhat from those reported by UNICEF due at least in part to the different methods.

Multi-Question Comparisons of the Levels of Extremism/Fundamentalism Among Christians Versus Muslims

Two-Question Set: Q94c/Q95d Whippings and cutting off of hands for crimes like theft and robbery, Q94d/Q95e Stoning of people who commit adultery.

Table 11

Two-question set: Q94c/Q95d whippings and cutting off hands, Q94d/Q95e stoning of adulterers.
	Population-Weighted Percentages
	Number of Extremist Items Favored			Opposed Both	Remainder
	Zero	One	Two
Muslims	51	16	33	49	2
Christians	80	12	8	77	4
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.

Forty-nine percent (49%) of Muslims versus 20% of Christians favored at least one of these two severe punishments. Thirty-three percent (33%) of Muslims but only 8% of Christians favored both. The Remainder = Favored Zero – Opposed Both.

Three-Question Set: Q94c/Q95d Whippings and cutting off of hands for crimes like theft and robbery, Q94d/Q95e Stoning of people who commit adultery, Q88 Using arms and violence against civilians to defend your religion.

Table 12

Three-question set: Q94c/Q95d whippings and cutting off hands, Q94d/Q95e stoning of adulterers, Q88 using arms and violence to defend religion.
	Population-Weighted Percentages
	Number of Extremist Items Endorsed				Opposed All	Remainder
	Zero	One	Two	Three
Muslims	42	21	27	10	30	12
Christians	70	20	8	2	51	18
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.

Fifty-eight percent (58%) of Muslims endorsed at least one and 37% endorsed at least two, while 30% of Christians endorsed at least one and 10% endorsed at least two out of the three extremist items. The Remainder = Favored Zero – Opposed All.

Four-Question Set: Q94c/Q95d Whippings and cutting off of hands for crimes like theft and robbery, Q94d/Q95e Stoning of people who commit adultery, Q88 Using arms and violence against civilians to defend your religion, Q59c Whether women should be allowed to be religious leaders.

Table 13

Four-question set: Q94c/Q95d whippings and cutting off hands, Q94d/Q95e stoning of adulterers, Q88 using arms and violence to defend religion, Q59c whether women should be allowed to be religious leaders.
	Population-Weighted Percentages
	Number of Extremist Items Endorsed					Opposed All	Rmdr.
	Zero	One	Two	Three	Four
Muslims	12	36	21	23	7	7	5
Christians	41	41	13	5	1	27	14
Notes. Displayed numbers are rounded. See text for details about the population-weighting of percentages.
Dataset: Pew Sub-Saharan Africa Survey 2009.

A population-weighted 88% of Muslims endorsed at least one, 51% endorsed at least two, and 30% endorsed at least three out of the four extremist items. Among Christians, a population-weighted 59% endorsed at least one, 18% endorsed at least two, and 5% endorsed at least three out of the four extremist items. The Remainder = Favored Zero – Opposed All.

The population-weighted mean extremism score (range: 0 – 4) for Muslims = 1.77 and for Christians = .84. (Pew made a partly similar measure which they called a religious jurisprudence index).[5]

Looking across Table 13, we can see that as the number of extremist items in the set increases, the relatively more extremist group (Muslims), in greater percentages, supports higher numbers of items as compared to the less extremist group (Christians). A similar pattern for more extremist versus less extremist countries was shown in a 7-item set in Table 11.1 of the previous article.

Summary

The single-question analyses show that, overall, among adults across numerous sub-Saharan countries, Muslims’ support for hardline fundamentalist/ extremist items was higher than that of Christians. (Differences ranged from about 7% to 31%). Compared to the general sample of Muslims, those who supported sharia as the official law of the land were more likely to support the severe punishments, but were only slightly more likely to justify militant violence against civilians to defend religion, slightly more likely to deny women equal status in the area of religious leadership, and no more likely to have had their daughter(s) “circumcised.” Compared to the general sample of Christians, those favoring the Bible as the official law of the land were not substantially more likely to support or affirm the hardline fundamentalist or extremist items assessed here. 

Multi-question set analyses of these same data showed that, compared to Christians, Muslims favored more extremist/fundamentalist items and were less likely to oppose all of them. As the number of extremist items per set increased, the difference between Muslims and Christians in levels of extremism/fundamentalism became more apparent. These results highlight the usefulness of multi-question set analyses for comparing religious groups according to the same measure of extremism/fundamentalism.

Disclaimer

Pew Research is not responsible for my analyses or my interpretation of their data. From the Pew instructions for downloading data sets: “All manuscripts, articles, books, and other papers and publications using our data should reference the Pew Research Center’s Religion & Public Life Project as the source of the data and should acknowledge that the Pew Research Center bears no responsibility for the interpretations presented or conclusions reached based on analysis of the data.”

Notes and References

[1] Tolerance and Tension: Islam and Christianity in Sub-Saharan Africa, Pew Forum on Religion & Public Life (April, 2010), Complete Report pdf. http://www.pewresearch.org/wp-content/uploads/sites/7/2010/04/sub-saharan-africa-full-report.pdf

[2] “Tolerance and Tension,” Preface, p. ii, Complete report pdf.

[3] The data set analyzed here is the Pew Forum on Religion & Public Life’s Sub-Saharan Africa Survey. The survey was fielded in 2008-2009. The data set package contains files in addition to the data file, including a Codebook and detailed Questionnaire.

[4] Pew’s religion-specific age structure estimates for 2010 can be obtained through the following interactive website: http://globalreligiousfutures.org/countries. A detailed report for these estimates is in The Global Religious Landscape: A Report on the Size and Distribution of Major Religious Groups as of 2010, The Pew Forum on Religion and Public Life (December, 2012) http://assets.pewresearch.org/wp-content/uploads/sites/11/2014/01/global-religion-full.pdf

[5] On pages 50-51 of the Complete Report pdf, Pew describes a partly similar score which they called a “religious jurisprudence” index. It had a 4-point scale based on responses to four items, namely, support for (i) religious law (Biblical or sharia) as the law of the land, (ii) whippings and cutting off hands, (iii) stoning adulterers, and (iv) “allowing leaders and judges to use their religious beliefs when they decide family or property disputes” (Christians, Q94b, p. 286) or “giving Muslim leaders and religious judges the power to decide family or property disputes” (Muslims, Q95b, p. 290). Pew reported a median of 12% across 16 countries for Christians supporting 3 or 4 of those items, and a median of 40% across 15 countries for Muslims supporting 3 or 4 of those items.

As the analyses in the present article showed, however, support for using the Bible as the law of the land was not an indicator of support for the two specific punishments. It is not clear how well support for the Bible as the law indicates support for real examples of religious jurisprudence. It also does not seem to be much of an indicator of extremism or hardline fundamentalism, at least as assessed here with this set of items.