ArticlePDF Available

Abstract and Figures

Evidence that Muslims support patriarchal values more than non-Muslims is abundant but the nature of this evidence is contested. The ‘cultural’ interpretation suggests that patriarchal values are an inherent element of Muslim identity. The ‘structural’ interpretation holds that patriarchal values reside in structural characteristics and have little to do with Muslim identity. Evidence on these contradictory claims is inconclusive. Neither have advocates of the cultural position shown that Muslim support for patriarchal values remains robust under control of structural characteristics; nor have proponents of the structural position demonstrated that Muslim support for these values vanishes under such controls. Filling this gap, we use multi-level models to test whether Muslim support for patriarchal values vanishes under control of patriarchy's structural underpinnings. We find that Muslim support for patriarchal values is robust against various controls; and we identify mosque attendance as a mechanism to sustain Muslim support for patriarchy in non-Muslim societies. Yet, rising levels of education, labor market participation, and a glacial emancipative trend diminish Muslim support for patriarchy, especially among women.
Content may be subject to copyright.
UNCORRECTED PROOF
PROOF COVER SHEET
Author(s): Christian Welzel and Amy Alexander
Article title: Islam and patriarchy: how robust is Muslim support for
patriarchal values?
Journal acronym: CIRS
Article no: 581801
Enclosures: 1) Query sheet
2) Article proofs
Dear Author,
1. Please check these proofs carefully. It is the responsibility of the
corresponding author to check these and approve or amend them. A
second proof is not normally provided. Taylor & Francis cannot be held
responsible for uncorrected errors, even if introduced during the production
process. Once your corrections have been added to the article, it will be
considered ready for publication.
For detailed guidance on how to check your proofs, please see
http://journalauthors.tandf.co.uk/production/checkingproofs.asp.
2. Please review the table of contributors below and confirm that the first and
last names are structured correctly and that the authors are listed in the correct
order of contribution. This check is to ensure that your name will appear
correctly online and when the article is indexed.
Sequence Prefix Given name(s) Surname Suffix
1 Christian Welzel
2 Amy Alexander
UNCORRECTED PROOF
Queries are marked in the margins of the proofs. Unless advised otherwise,
submit all corrections and answers to the queries using the CATS online
correction form, and then press the ‘‘Submit All Corrections’’ button.
AUTHOR QUERIES
General query: You have warranted that you have secured the necessary
written permission from the appropriate copyright owner for the
reproduction of any text, illustration, or other material in your article.
(Please see http://journalauthors.tandf.co.uk/preparation/permission.asp.)
Please check that any required acknowledgements have been included to
reflect this.
AQ1 Please provide a reference for Huntington (1996)
AQ2 Please provide a reference for Moghadam (2003)
AQ3 Please provide a reference for Inglehart and Welzel 2005
AQ4 Can this be rephrased to say ‘‘an increase in the level ...’’?
AQ5 There is no Figure 8 in this paper. Please amend this.
AQ6 Should this be 2003a or 3003b?
AQ7 Please provide a reference for Gerring et al. 2005
UNCORRECTED PROOF
Taylor & Francis now offer you a choice of opons to help you make correcons to
your proofs. Your PDF proof file has been enabled so that you can edit the proof
directly using Adobe Acrobat. This is the simplest and best way for you to ensure
that your correcons will be incorporated. If you wish to do this, please follow these
instrucons:
1. Save the file to your hard disk.
2. Check which version of Adobe Acrobat you have on your computer. You can do
this by clicking on the “Help” tab, and then “About”.
If Adobe Reader is not installed, you can get the latest version free from
hp://get.adobe.com/reader/.
If you have Adobe Reader 8 (or a later version), go to “Tools”/ “Comments &
Markup”/ “Show Comments & Markup”.
If you have Acrobat Professional 7, go to “Tools”/ “Commenng”/ “Show
Commenng Toolbar”.
3. Click “Text Edits”. You can then select any text and delete it, replace it, or insert
new text as you need to. It is also possible to highlight text and add a note or
comment.
4. Make sure that you save the file when you close the document before uploading it
to CATS using the “Upload File” buon on the online correcon form. A full list of
the comments and edits you have made can be viewed by clicking on the
“Comments” tab in the boom le-hand corner of the PDF.
If you prefer, you can make your correcons using the CATS online correcon form.
Author Services
UNCORRECTED PROOF
RESEARCH ARTICLE
Islam and patriarchy: how robust is Muslim support for patriarchal
values?
Christian Welzel and Amy Alexander*
Leuphana University, Belgium
(Received November 2010; final version received December 2010)
Evidence that Muslims support patriarchal values more than non-Muslims is
abundant but the nature of this evidence is contested. The ‘cultural’ interpretation
suggests that patriarchal values are an inherent element of Muslim identity. The
‘structural’ interpretation holds that patriarchal values reside in structural
characteristics and have little to do with Muslim identity. Evidence on these
contradictory claims is inconclusive. Neither have advocates of the cultural
position shown that Muslim support for patriarchal values remains robust under
control of structural characteristics; nor have proponents of the structural
position demonstrated that Muslim support for these values vanishes under such
controls. Filling this gap, we use multi-level models to test whether Muslim
support for patriarchal values vanishes under control of patriarchy’s structural
underpinnings. We find that Muslim support for patriarchal values is robust
against various controls; and we identify mosque attendance as a mechanism to
sustain Muslim support for patriarchy in non-Muslim societies. Yet, rising levels
of education, labor market participation, and a glacial emancipative trend
diminish Muslim support for patriarchy, especially among women.
Keywords: values; Muslim; patriarchy; multiculturalism; cross-national study
Introduction
In two studies, Inglehart and Norris (2003a,b) examine the connection between Islam
and patriarchal values.
1
Their findings confirm Huntingtons (1996)
AQ1
thesis of a
cultural chasm between Islam and the ‘West’. However, they clarify that the key
difference is not on matters of democracy but on matters of emancipation. It is in
values about ‘eros, not demos’ (2003a, p. 65) where Muslims and Westerners differ.
Indeed, Inglehart and Norris provide the broadest evidence to date for Muslim
support of patriarchal values.
The evidence is echoed in studies of Muslim women’s education and positional
achievement. Fish (2002) and numerous secular feminists describe adherence to
Islamic norms as a barrier to womens advancement (Afshar 1982, Minces 1982,
Tabari 1982, Ghoussoub 1987, Karam 1998, Moghissi 1999). These views suggest
that socialization under Islamic norms ingrains patriarchal values as an inherent
attribute of Muslim identity.
Others, however, attribute patriarchal values to different factors, which dominate
in Muslim societies for reasons other than Islam itself. These reasons are structural in
5
10
15
20
25
30
35
40
*Email: cwelzel@gmail.com
International Review of Sociology Revue Internationale de Sociologie
Vol. 21, No. 2, July 2011, 249275
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:38
ISSN 0390-6701 print/ISSN 1469-9273 online
# 2011 University of Rome ‘La Sapienza’
DOI: 10.1080/03906701.2011.581801
http://www.informaworld.com
UNCORRECTED PROOF
character as they emanate from economic and political power relations. Thus,
Moghadam (2003, pp. 58)
AQ2
claims:
[T]he position of women in the Middle East cannot be attributed to the presumed
intrinsic properties of Islam. It is also my contention that Islam is neither more nor less
patriarchal than other religions [...] [T]o understand the social implications of Islam,
therefore, it is necessary to look at the broader sociopolitical and economic order within
which it is exercised.
A similar position is taken by Ross (2008). His explanation of Muslim patriarchy
emphasizes non-cultural factors, especially oil and gas rents (2008, p. 117):
Women have made less progress in gender equality in the Middle East than in any other
region. Many observers claim this is due to the regions Islamic traditions. I suggest that
oil, not Islam is at fault [....] Oil production reduces the number of women in the labor
force, which in turn reduces their political influence. As a result, oil-producing states are
left with atypically strong patriarchal norms, laws, and political institutions.
These views contradict the cultural interpretation; they endorse a structural
interpretation. Patriarchal values are not pronounced among Muslims because of
Islams inherent affinity to patriarchy. Instead, Muslims are socialized under
patriarchal structures that characterize Muslim societies for other reasons than Islam.
One might argue that the two positions are not contradictory: Muslim support for
patriarchal values might be strong for both cultural and structural reasons. But this
conclusion misses the key disagreement. For the structural position takes an extreme
stance: it holds that Muslim support for patriarchal values is solely a derivative of
patriarchal structures and not inherent in Muslim identity. In analytical terms, this
means that in dissociation from patriarchal structures Muslim identity shows no
effect on patriarchal values. By contrast, the cultural interpretation suggests that
Muslim support for patriarchal values is at least partly a property of Muslim identity
and does not vanish when one dissociates Muslims from patriarchal structures.
So far, the evidence is inconclusive. On one hand, Ross provides the broadest
evidence for the structural position but does not analyze patriarchal values. Thus, his
claim that oil-producing states are left with atypically strong patriarchal norms
because oil economies reduce womens access to the labor market remains
undemonstrated. On the other hand, Inglehart and Norris (2003a,b) provide the
broadest evidence for the cultural position but do not take into account oil rents and
other indicators of structural patriarchy. Their claim that the Muslim affinity to
patriarchal values cannot be reduced to structural factors remains undemonstrated,
too. The question of whether Muslim support for patriarchal values holds up in
dissociation from structural factors is open.
This is not a technical question because it touches upon the nature of Muslim
patriarchy. Muslim patriarchy is merely a structural phenomenon if Muslim support
for patriarchal values vanishes in dissociation from structural features that are
typical of Muslim societies for reasons other than Islam. If this is true, Muslims
support patriarchal values more strongly than non-Muslims only if these Muslims
live in Muslim societies. By contrast, Muslim patriarchy is at least partly a cultural
phenomenon if Muslim support for patriarchal values remains robust even in
dissociation from patriarchys structural underpinnings. In this case, patriarchal
45
50
55
60
65
70
75
80
85
250 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:38
UNCORRECTED PROOF
values are an inherent element of Muslim identity. If so, these values travel wherever
Muslims settle and retain their own identity.
We examine this question by analyzing how individual Muslim identification and
Muslim social dominance affect patriarchal values, under control of key structural
factors that have so far been left out of the study of patriarchal values.
Part one of this study describes the research design. Part two introduces the data
and measurements. Using the World Values Survey, we examine the patriarchal
values of about 130,000 respondents from some 80 societies. Part three presents the
findings. The concluding part discusses broader implications.
Research strategy
Inglehart and Norris (2003a,b) present the broadest evidence of a Muslim affinity to
patriarchal values.
2
There are, however, shortcomings in their study. The exclusion of
key structural factors, such as oil rents, is one of them. Another is the unsystematic
distinction between Muslim identity at the individual level and Muslim dominance at
the societal level. Even though the authors demonstrate a tendency towards
patriarchal values for both Muslim self-identification and Muslim social dominance
(2003b, pp. 44, 67), they do not isolate these effects from each other in a multi-level
model. This is not a trivial shortcoming. In the absence of multi-level modeling, it
remains unclear at which level the Muslim affinity to patriarchal values primarily
works: is it the living in a dominantly Muslim society or the self-identification as a
Muslim that favors patriarchal values more strongly?
Existing research has not solved this question. Studies in Europe and the USA
compare the values of Muslim immigrant populations with those of the host
populations. They find that Muslim immigrants support patriarchal values more
strongly even if they are second- or third-generation immigrants (Kosmin and Mayer
2001, Buijs and Rath 2002, Pipes 2002, Fetzer and Soper 2005, Laurence and Vaisse
2006, Morin and Horowitz 2006, Pfaff 2007). However, these studies cannot tell
whether leaving a dominantly Muslim society makes Muslims less patriarchal, because
the studies cannot compare the values of immigrant Muslims with Muslims of similar
age and background in Muslim societies. Vice versa, studies among Muslims in
dominantly Muslim societies lack comparison with immigrant Muslims in Western
societies and, therefore, cannot isolate the effect of living in a Muslim society (Rizzo
et al. 2007, Blaydes and Linzer 2008). In order to separate the effects of living in a
Muslim society and being a Muslim, one needs a sample that allows one to compare
the values of Muslims and non-Muslims in both Muslim and non-Muslim societies.
The World Values Survey offers this possibility. Analyzing these data, we examine
whether Muslim support for patriarchal values vanishes in dissociation from a
number of confounding factors that are not defining elements of Muslim identity.
A Muslim tendency towards patriarchal values can be present on two levels: the
societal level and the individual level. At the societal level, a Muslim tendency
towards patriarchal values means that societies exhibit higher mean levels of
patriarchal values when the Muslim share of the population is larger. At the
individual level, a Muslim tendency towards patriarchal values means that within
any given society, Muslims hold stronger patriarchal values than non-Muslims. If one
can establish that, on both levels, the Muslim tendency towards patriarchal values
90
95
100
105
110
115
120
125
130
International Review of Sociology Revue Internationale de Sociologie 251
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:39
UNCORRECTED PROOF
vanishes under control of confounding factors that are not inherently Muslim, the
patriarchal tendency is not inherently Muslim either.
Confounding factors
Scholarship has claimed a number of factors as responsible for the support of
patriarchal values in a society. Ross (2008) argues that oil rents keep women out of
the workforce and that this effect explains the prevalence in patriarchal values.
Unlike oil economies, knowledge economies involve large proportions of women into
the workforce. If the same logic applies, this should diminish the prevalence of
patriarchal values (Inglehart and Welzel 2005, p. 282).
AQ3
Hence, Muslim societies might
exhibit higher base levels of patriarchal values mostly because the structure of their
economies excludes women, not because they are dominantly Muslim.
Apart from economic structures, political institutions have been found to
influence patriarchy. Specifically, the endurance of democracy decreases a societys
base level of patriarchal values (Paxton and Hughes 2007). Accordingly, Muslim
societies might exhibit higher base levels of patriarchal values because they lack a
democratic tradition, not because of Muslim social dominance.
The effect of individual Muslim identification on patriarchal values, too, might
be accounted for by a number of confounding factors. One factor is religiosity.
A huge literature establishes a link between religious devotion and patriarchal values
(Harkness 1972, Ruether 1974, Peek et al. 1991, Okin 1999, Burn and Busso 2005,
Paxton and Hughes 2007, pp. 109114). Thus, Muslims might be more patriarchal
not because they are Muslim but because more of them are strongly religious.
In order to function as a transmitter of patriarchal values, religion needs a
socializing institution (Beit-Hallahmi and Argyle 1997). Attendance at the mosque or
church for religious service can meet this function. Service attendance exposes people
to the propagated values of their religion and repeats a group experience that fosters
identification with ones religions values (Hood et al. 2005). Hence, Muslims might be
more patriarchal only in as far as they attend religious service more frequently.
Muslims support for patriarchal values may also be confounded with an
individuals level of education. As Inglehart and Welzel (2005, p. 220) demonstrate,
educated people are more tolerant and egalitarian in each of more than 70 societies
surveyed in the WVS. A recent study of fundamentalist patriarchyshows that the
level
AQ4
of education among Muslims significantly decreases their support for
patriarchal values (Blaydes and Linzer 2008). Accordingly, Muslims might not be
more patriarchal because they are Muslims but because they tend to be less educated.
The literature identifies two additional individual-level characteristics that
weaken patriarchal values: female sex and cohort sequence. As documented by
Inglehart and Norris (2003b, p. 44) and Inglehart and Welzel (2005, p. 275), most
societies for which data are available show an increase in emphasis on gender
equality from older to younger generations, following a growing emancipative
zeitgeist. The same analysis shows that women score higher in emancipative values
and, by implication of this, lower in patriarchal values throughout all generations
and in all types of societies. The study by Blaydes and Linzer (2008) confirms these
patterns for Muslim societies. Accordingly, we expect younger Muslims and Muslim
women to exhibit weaker patriarchal values than older Muslims and Muslim men.
135
140
145
150
155
160
165
170
175
252 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:39
UNCORRECTED PROOF
The literature suggests yet another two characteristics to affect patriarchal values,
especially among Muslim women: employment and marital status. With the
advancement of knowledge societies (Lesthaege 2004), more women are mobilized
into the workforce and many of them marry later or choose non-traditional forms of
cohabitation. These developments increase the number of employed and unmarried
women. These women are economically and socially more independent, which has
been shown to lower their support for patriarchal values (Blaydes and Linzer 2008).
How individual-level characteristics affect peoples values is often contextually
shaped by characteristics of the surrounding society. Thus, Muslims support for
patriarchal values at the individual level might vary with the dominance of Muslims
at the societal level. In the extreme case, it might turn out that Muslims support
patriarchal values only in dominantly Muslim societies. In this case, we cannot speak
of a general patriarchal effect of Muslim identification. Instead, the effect is
contingent. To examine this possibility in a conclusive way, one has to combine
societal-level effects, individual-level effects, and cross-level interactions in multi-level
models (Bryk and Raudenbusch 2002). Hence, we use a multi-level approach.
Data and variables
Sample description
To analyze patriarchal values among Muslims and non-Muslims and across societies
with varying percentages of Muslims, we use the two most recent rounds of the
World Values Survey (henceforth: WVS). The two most recent rounds of the WVS
were conducted in 19992001 (round four) and 20052008 (round five). After
excluding missing data on any of the variables included in our multi-level models, we
are left with a sample of 133,295 individuals from 83 societies around the world.
Which societies are included is visible in Figure 8.
3
AQ5
The societies vary in the percentage of Muslims from less than 5% in Australia to
more than 99% in Saudi Arabia. Of these societies, 21 include a sizeable proportion
of Muslims of 10% and more. Twelve of these societies are composed of a clear
majority of Muslims of above 60% and nine societies fill the wide middle range
between 10 and 60%.
4
At the individual level, 33,073 or 25% of our respondents identify themselves as
Muslim; 100,222 or 75% identify themselves with a different or no religious
denomination.
5
Given these figures, the WVS data show enough variation to
separate the effects of Muslim social dominance and individual Muslim self-
identification.
The dependent variable: patriarchal values
We measure patriarchal values using all WVS questions suited to indicate support for
the subordination of women to men. Overall, the WVS includes three such questions
that are fielded in both the fourth and fifth round.
6
The first question reads: Do you agree, disagree or neither agree nor disagree
with the following statements? Then the statement in variable V44 reads: When jobs
are scarce, men should have more right to a job than women. The second and third
questions are taken from the same battery, which read as follows: Do you strongly
180
185
190
195
200
205
210
215
International Review of Sociology Revue Internationale de Sociologie 253
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:39
UNCORRECTED PROOF
agree, agree, disagree, or strongly disagree? Then the item in variable V61 reads: On
the whole men make better political leaders than women do. The item in variable
V62 reads: A university education is more important for a boy than for a girl. Each
of these items offers a rank-ordered four-point response option. We recode responses
in such a way that they all have a scale range from minimum 0 for the least
patriarchal position and 1.0 for the most patriarchal position.
7
The items represent patriarchy in the domains of labor market participation,
education, and political leadership. Analyzing the dimensionality of these items
yields a single principal component with an Eigenvalue above 1.0. On this
component, the three patriarchy items show loadings of .82 for male priority in
political leadership, .75 for male priority in labor market participation, and again .75
for male priority in education. A reliability analysis yields a Cronbachs alpha of .65,
which is above the acceptance threshold of .30 for three items.
The uni-dimensionality of these items justifies a combination into a summary
index of patriarchal values. We add up the scores over the three items and divide the
sum by three. This yields a 12-point index with minimum 0, when there is no support
of patriarchy on any included item, and maximum 1.0, when there is full support of
patriarchy on all included items.
The global mean in patriarchal values falls almost exactly on the midpoint of the
scale, on .49 to be precise, with a standard deviation of .26. The median and mode
are located close to the mean, at .46.
Independent variables: Muslim identification and Muslim dominance
At the individual level, we dichotomize respondents into those who identify
themselves as Muslim (coded 1) and all others who identify with no or another
religious denomination (coded 0). This dichotomy entails no information of how
strongly Muslims identify themselves with their denomination. However, as will be
seen below, the overwhelming majority of people who identify themselves as Muslim
are very strongly religious. It seems obvious that when people are strongly religious,
the identification with their self-reported denomination must also be strong.
To estimate the link between Muslim identification and patriarchal values, we
examine to what extent self-identifying Muslims support patriarchal values more
strongly than non-Muslims with the same characteristics. The alternative approach
would be to ask self-identifying Muslims directly if they think that patriarchal values
are a defining element of their religious identity. The WVS does not ask such a
question for good reason. Such a question would be too sensitive to social
desirability: respondents might not want to admit that patriarchy is a defining
element of their belief even if it is. By contrast, respondents have limited control over
a latent pattern that guides their responses to unconnected questions. Analyzing the
association of responses to unconnected questions uncovers these patterns and thus
evidences a link even if respondents would not admit that link in a direct question.
That respondents did not design their responses from the viewpoint of their
religious denomination can be confidently assumed because the WVS asks
respondents only at the end of the interview to report their religious denomination.
At the societal level, we measure Muslim dominance by the percentage of
Muslims in the residential population, as documented by the UN Demographic
Yearbook for the year 2000 (United Nations Statistical Department 2006) or the
220
225
230
235
240
245
250
255
260
254 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:40
UNCORRECTED PROOF
CIA Factbook otherwise (Central Intelligence Agency 2006). To triangulate these
estimates we compare them with the numbers we obtain when calculating per society
the percentages of Muslims from the individual responses in the WVS. Data from the
two sources correlate at r .95. To take advantage of both data sources, we assign
each society the mean of the percentage of Muslims as calculated from the WVS and
as reported by the UN or CIA, respectively.
8
To simplify things, some analyses use a binary distinction between dominantly
Muslim and dominantly non-Muslim societies: as dominantly Muslim, we
classify societies in which more than 50% of the adults are Muslim (N15); as
dominantly non-Muslim, we classify societies with less than 50% Muslims (N68).
Accordingly, we find that 71,407 respondents of the WVS or 74.5% are non-Muslims in
non-Muslim societies; 4,083 respondents or 4.3% are Muslims in non-Muslim
societies; 3,671 respondents or 3.8% are non-Muslims in Muslim societies; and
16,739 respondents or 17.5% are Muslims in Muslim societies.
Control variables
At the individual level, we control the effect of Muslim identification for the
respondents education, religiosity, religious service attendance, sex and cohort.
These variables are taken from the WVS.
Sex is documented by observation of the interviewer in variable V235. We code
female sex as 1 and male sex as 0.
Cohort is recoded from the question on a respondents age in variable V237.
From information on a respondents age and the year of the survey, we create eight
successive birth cohorts, each covering a 10-year time span, coded from 0 to 1.0, with
values increasing from older to younger cohorts.
9
Education is measured on an eight-point ordinal scale taken from V238 in the
WVS, rescaled from 0 for no formal education to 1.0 for a university degree.
10
Religiosity is measured by question V192, which asks: How important is God in
your life? Please use this scale to indicate. 10 means ‘‘very important’’ and 1 means
‘‘not at all important’’. We change the scale into a 0-to-1.0 format, from 0 for the
lowest to 1.0 for the highest importance of God.
11
Frequency of attending religious services is measured by question V186, which
asks: Apart from weddings and funerals, about how often do you attend religious
services these days? Response options are: more than once a week, once a week,
once a month, only on special holy days, once a year, less than once a year
and never, practically never. We recode these options into a seven-point scale from 0
for never, practically never to 1.0 for more than once a week.
12
It is self-evident
that religious service attendance in the case of Muslims means mosque attendance
(and church attendance in case of Christians).
With respect to women, two additional characteristics have been emphasized:
employment status and marital status. Specifically, employed women and unmarried
women are less patriarchal in their values (Inglehart and Norris 2003,
AQ6
Blaydes and
Linzer 2008). Information on a respondents employment status is taken from
question V241 and we code the answers as 0 when no employment is reported and 1.0
when self-employment, part time employment or full time employment is reported.
Information on marital status is taken from question V55. We code the responses as
0 when being married is reported and 1.0 otherwise.
265
270
275
280
285
290
295
300
305
310
International Review of Sociology Revue Internationale de Sociologie 255
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:40
UNCORRECTED PROOF
At the societal level, we control the effect of Muslim social dominance for four
variables that measure key structural aspects of a societys economy and institutions:
a societys per capita oil/gas rent, the size of the female workforce, the endurance of
democracy and a direct measure of patriarchal power structures, namely the inverse
of the Gender Empowerment Measure. These indicators were measured in about
2000, the beginning of the survey period.
To measure per capita oil/gas rents, we use Rosss data, which indicate a countrys
total rents from oil and gas, in thousands of international US dollars, divided by
population size (Ross 2008, p. 111). The highest per capita oil rent among the
societies in our sample is US$3,252 for Saudi Arabia, the lowest is zero dollars for
Singapore and a couple of other countries. Again, this scale range is standardized
into a range from 0 for the lowest and 1.0 for the highest value.
To measure the size of the female workforce, we calculate for each national
sample the percentage of women who report to be employed for pay. The highest
female employment rate is reported in Iceland (70%), the lowest in Saudi Arabia
(10%). Dividing these figures by 100, we standardize them into the 0-to-1.0 scale
format.
Enduring democracy is measured using Gerrings democracy stock variable as
of 2000 (Gerring et al. 2005).
AQ7
This variable adds up the democracy scores a society
has accumulated over time on the Polity IV autocracy-democracy index but
depreciates scores from past years by 1% for each year they precede the reference
year, 2000. This index reflects a societys accumulated experience with democracy
with a premium on recent experience. We standardize the index into a scale range
from minimum 0, which is represented by Saudi Arabia, to maximum 1.0, which is
represented by Australia, Norway, Switzerland and the USA.
Patriarchal structures are directly manifest in the exclusion of women from
positions of power. To measure female exclusion from power, we use the inverse of
the Gender Empowerment Measure provided by the United Nations Development
Program. The gender empowerment measure indicates on a scale range from 0 to 1.0
womens achievement of positions of power (Human Development Report, various
years). Inverted within the same scale range from 0 to 1.0, this indicator measures the
exclusion of women from positions of power.
Findings
The distinctiveness of Muslim identity
Speaking of Muslim identity suggests that self-identifying Muslims represent a
distinct category of people. Evidence for this suggestion is available from question
V185, which asks for peoples religious denomination. From the list of denomina-
tions, we look at those with at least 1,000 entries and compare respondents with these
identifications to respondents identifying themselves as Muslim. Based on the
Iranian and Iraqi samples, we can distinguish Shiite and Sunnite Muslims to check
for major internal divisions among Muslims.
The left-hand diagram of Figure 1 shows for respondents of various religious
denominations the percentages positioning themselves at the very top of the ten-
point religiosity scale. Self-identifying Muslims stick out as the denomination with by
far the largest percentage of strongly religious people: 82%. Even more astounding,
315
320
325
330
335
340
345
350
355
256 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:40
UNCORRECTED PROOF
Orthodox Christians
Hindus
Jews
Catholics
Protestants
Evangelicals
Muslims
Shiites
Sunnis
Percent Strongly Religious
(saying "God is very important")
3,548
3,742
1,382
36,744
16,188
18,365
44,288
3,940
1,790
0
10 20 30 40 50 60 70 80 90
100
0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80
Evangelicals
Catholics
Orthodox Christians
Jews
Protestants
Hindus
Muslims
Sunnis
Shiites
Patriarchal values among the strongly relgious
2,836
1,503
30,154
5,838
9,126
21,712
3,401
1,661
142
1 SD
Figure 1. Strength of religiosity by religious denomination and patriarchal values among the very religious.
International Review of Sociology Revue Internationale de Sociologie 257
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:40
UNCORRECTED PROOF
fully 92% of all self-identifying Muslims place themselves at the two highest scores of
the ten-point religiosity scale. Self-identifying as a Muslim, regardless of the
particular branch of Islam, seems to be almost synonymous with being strongly
religious.
These results are not an artifact of the WVS. The Bertelsmann Foundations
Religion Monitor also finds a uniformly strong religiosity among Muslims in
Indonesia, Morocco, Nigeria and Turkey (Heine and Spielhaus 2009).
The WVS does not ask Muslims how strong their Muslim identity is. But from the
exceptionally strong religiosity of almost all self-identifying Muslims it is evident that
Muslim self-identification means a strong Muslim identity for the large majority of
Muslims. This conclusion is also plausible because Islam does not have a formally
registered membership, which forecloses the possibility that people report Muslim
denomination merely because of formal membership even though they do not really
identify with their religion.
The Muslim/non-Muslim gap in the strength of religiosity nurtures the suspicion
that self-identifying Muslims support patriarchal values simply because they are
exceptionally religious. The right-hand diagram of Figure 1 tests this assumption,
isolating in each religious denomination those people who are at the highest level of
religiosity and comparing their mean scores on the patriarchal values index. Again,
the pattern speaks strikingly to the distinctiveness of Muslims. Even at the highest
level of religiosity, self-identifying Muslims hold much stronger patriarchal values
than people from other denominations, including Hindus, Orthodox Christians and
the main branches of Christianity. These findings justify dichotomizing self-
identifying Muslims against all other religious denominations.
Muslim support for patriarchal values
The right-hand diagram in Figure 1 evidences exceptionally strong support for
patriarchal values among identifying Muslims, even holding religiosity constant. But,
perhaps Muslim support for patriarchal values is entirely a property of Muslim social
dominance and only seems to be a property of individual Muslim identification
because most identifying Muslims live in dominantly Muslim societies. To test this
possibility, Figure 2 compares the patriarchal values of identifying Muslims and non-
Muslims between dominantly Muslim and dominantly non-Muslim societies.
As Figure 2 illustrates, both individual Muslim identification and Muslim social
dominance elevate the mean level of patriarchal values. Muslims in Muslim societies
score at .75 points in patriarchal values, which is .08 scale points above the level of non-
Muslims. In non-Muslim societies, Muslims score at .55 points in patriarchal values,
which is .13 scale points above the level of non-Muslims. Thus, individual Muslim
identification elevates patriarchal values irrespective of Muslim social dominance, and
this tendency is statistically significant at the .001 level.
Yet, Muslim social dominance elevates patriarchal values even more than
individual Muslim identification: in dominantly Muslim societies, Muslims score
.20 scale points higher (at .75) than in non-Muslim societies (.55). The same holds
true for non-Muslims: in dominantly Muslim societies, they score .24 points higher
(at .66) than in non-Muslim societies (.42). Thus, living in a dominantly Muslim
society strengthens patriarchal values more than individual identification as a
Muslim.
360
365
370
375
380
385
390
395
400
258 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:40
UNCORRECTED PROOF
In non-Muslim societies, identifying Muslims are exposed to a less patriarchal
environment. Adjustment to this environment might explain why Muslims in non-
Muslim societies are by .20 scale points less patriarchal than Muslims in Muslim
societies. But even in non-Muslim societies, Muslims retain a by .13 scale points
stronger emphasis on patriarchal values than non-Muslims. A mechanism to retain
this emphasis is to shield Muslim communities from the non-Muslim environment.
A social space that may sustain such a shield is the mosque: frequent mosque
attendance could reinforce a distinct community experience that helps sustain
Muslim support of patriarchal values (Heine and Spielhaus 2009).
Figure 3 supports this interpretation. From the left-hand diagram we see that, in
non-Muslim societies, only 10% of Muslim men say they never attend the mosque.
But close to 50% say that they attend the mosque more than once a week. This is
five times the proportion of non-Muslim men who say they attend religious service
more than once a week. Interestingly, Muslim men attend the mosque slightly more
frequently in non-Muslim societies than in Muslim societies.
As the right-hand diagram of Figure 3 illustrates, this pattern is even more
pronounced among Muslim women. Mosque attendance is less of a requirement for
Muslim women than for men, so attendance rates differ by sex in most Muslim
societies (Heine and Spielhaus 2009). Figure 3 confirms this regularity. But the key
point is that female mosque attendance is significantly higher in non-Muslim than in
Muslim societies. To be precise, in Muslim societies 29% of Muslim women say they
never attend the mosque while this proportion is only 17% in non-Muslim societies.
Likewise, 25% of Muslim women in Muslim societies say they attend the mosque
more than once a week, compared to 36% in non-Muslim societies. Possibly,
attendance of religious service is more needed in non-Muslim than in Muslim
405
410
415
420
425
Figure 2. The patriarchal effects of Muslim social dominance and individual Muslim
identification.
International Review of Sociology Revue Internationale de Sociologie 259
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:40
UNCORRECTED PROOF
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
Frequency (in percent)
Mosque/Church Attendance
Muslims in Mus-
lim Societies
Non-muslims in non-
muslim societies
Muslims in non-
muslim societies
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
Never Less than
once
a year
Once a
year
On holidays
only
Once a
month
Once a
week
More than
once
a week
Never Less than
once
a year
Once a
year
On holidays
only
Once a
month
Once a
week
More than
once
a week
Mosque/Church Attendance
Frequency (in percent)
Non-muslims in non-
muslim societies
Muslims in mus-
lim societies
Muslims in non-
muslim societies
Women
Men
Figure 3. Frequency of mosque/church attendance among women and men by Muslim identication and Muslim social dominance.
260 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:40
UNCORRECTED PROOF
societies to foster Muslim identity. If so, patriarchal values depend more strongly on
service attendance in non-Muslim than in Muslim societies. We examine this
question below.
Muslim support for patriarchal values under key demographic controls
Is the Muslim affinity to patriarchal values so sweeping that it holds across group
divisions by sex, education, and cohort? The two diagrams in Figure 4 illustrate how
education varies the mean level of patriarchal values among four groups: Muslim
women, Muslim men, non-Muslim women and non-Muslim men. These four groups
are separated depending on whether we find them in Muslim societies (left-hand
diagram) or non-Muslim societies (right-hand diagram). The mean levels of
patriarchal values of these eight groups are shown on the vertical axes but this is
done separately for eight educational categories, ranging from no formal education
on the left to a university degree on the right.
13
For each group, trend lines fall from left to right. This means that higher
education is associated with weaker patriarchal values. The anti-patriarchal effect
of education is uniform because it is visible in all eight groups. Female sex also
shows a uniformly anti-patriarchal effect: in all group categories, women emphasize
patriarchal values less than men. But even though the anti-patriarchal effects of
formal education and female sex are uniform in direction, their strengths vary in
interaction with each other. This is obvious from the fact that patriarchal values fall
on a steeper slope among women than among men with rising education. Education
widens the gender gap over patriarchal values.
Individual Muslim identification still makes a difference, irrespective of divisions
by education, gender, and Muslim social dominance. Muslims always emphasize
patriarchal values more than non-Muslims of the same category.
Muslim social dominance still matters, too. This is evident from the fact that in
each group be it women or men, Muslim or non-Muslim, highly educated or not so
highly educated the base level of patriarchal values is about .15 to .20 scale points
higher when the group is located in a Muslim societies.
Figure 5 replaces level of education by cohort, with cohorts being younger to the
right. As does education, cohort sequence diminishes patriarchal values in all groups,
yet in all groups the base level of patriarchal values is higher when the group is
located in a Muslim society. And, irrespective of sex and Muslim social dominance,
identifying Muslims always emphasize patriarchal values more strongly than non-
Muslims of the same category.
Female sex enhances the anti-patriarchal effect of cohort sequence, and does so
quite pronouncedly, widening the gender gap from an almost unnoticeable difference
among the earliest cohorts to a sizable gap of .10 to .12 scale points among the most
recent cohorts.
Education and cohort pull in the same direction. They both lower patriarchal
values and both do so in a way that increases the gender gap over patriarchal values.
This also holds for self-identifying Muslims and for Muslim societies. Thus, both
Muslim individuals and Muslim societies are susceptible to anti-patriarchal forces.
But none of the anti-patriarchal forces closes the Muslim/non-Muslim gap over
patriarchal values. Even if they are highly educated, belong to a younger cohort, live
430
435
440
445
450
455
460
465
470
International Review of Sociology Revue Internationale de Sociologie 261
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:41
UNCORRECTED PROOF
Dominantly muslim societies
Dominantly non-muslim societies
0.20
0.22
0.24
0.26
0.28
0.30
0.32
0.34
0.36
0.38
0.40
0.42
0.44
0.46
0.48
0.50
0.52
0.54
0.56
0.58
0.60
0.62
0.64
0.66
0.68
0.70
0.72
0.74
0.76
0.78
0.80
Complete
primary
Incomplete
secondary,
vocational
Complete
secondary,
vocational
Incomplete
secondary,
academic
Complete
secondary,
academic
Tertiary level
(university)
Patriarchal values 2000-06
Formal education level
Mean range of 1 SD
per category
Location of group's modus category
0.20
0.22
0.24
0.26
0.28
0.30
0.32
0.34
0.36
0.38
0.40
0.42
0.44
0.46
0.48
0.50
0.52
0.54
0.56
0.58
0.60
0.62
0.64
0.66
0.68
0.70
0.72
0.74
0.76
0.78
0.80
None Incomplete
primary
None Incomplete
primary
Complete
primary
Incomplete
secondary,
vocational
Complete
secondary,
vocational
Incomplete
secondary,
academic
Complete
secondary,
academic
Tertiary level
(university)
Patriarchal values 2000-06
Formal education level
Location of group's modus category
Mean range of 1 SD
per category
Figure 4. The anti-patriarchal effect of education by sex, individual Muslim identication, and Muslim social dominance.
262 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:41
UNCORRECTED PROOF
Dominantly muslim societies Dominantly non-muslim societies
0.20
0.22
0.24
0.26
0.28
0.30
0.32
0.34
0.36
0.38
0.40
0.42
0.44
0.46
0.48
0.50
0.52
0.54
0.56
0.58
0.60
0.62
0.64
0.66
0.68
0.70
0.72
0.74
0.76
0.78
0.80
Patriarchal Values 2000-06
Birth cohorts
Mean range of 1 SD
per category
Location of group's modus category
0.20
0.22
0.24
0.26
0.28
0.30
0.32
0.34
0.36
0.38
0.40
0.42
0.44
0.46
0.48
0.50
0.52
0.54
0.56
0.58
0.60
0.62
0.64
0.66
0.68
0.70
0.72
0.74
0.76
0.78
0.80
Before 1921 1921-30 1931-40 1941-50 1951-60 1961-70 1971-80 After 1980
Before 1921 1921-30 1931-40 1941-50 1951-60 1961-70 1971-80 After 1980
Patriarchal Values 2000-06
Birth cohorts
Mean range of 1 SD
per category
Location of group's modus category
Figure 5. The anti-patriarchal effect of cohort sequence by sex, individual Muslim identication, and Muslim social dominance.
International Review of Sociology Revue Internationale de Sociologie 263
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:41
UNCORRECTED PROOF
in a non-Muslim society and are female, Muslims remain more patriarchal than
non-Muslims of the same category.
How do Muslims in non-Muslim societies manage to retain a stronger emphasis
on patriarchal values than the non-Muslims of their surrounding society? We
hypothesized that creating a space that shields Muslims from the non-Muslim
environment might be instrumental in this respect. Mosques could provide a space to
encapsulate Muslims in non-Muslim societies. If this is the case, one would expect
mosque attendance to associate more strongly with patriarchal values among
Muslims in non-Muslim societies.
The two diagrams of Figure 6 show separately for women and men how the
frequency of attending religious service affects patriarchal values. The pattern is
obvious. In Muslim societies, patriarchal values are so dominant that the mosque is
not needed as a space to sustain patriarchal values. Hence, there is no effect of
mosque attendance on patriarchal values. In non-Muslim societies, we see a different
picture: among both sexes, frequency of mosque attendance increases patriarchal
values from a low of .41 (women) or .47 (men) among non-attendants to a high of .55
(women) or .63 (men) among very frequent attendants. Whether this means that
patriarchal values are actively transmitted in the mosque or whether patriarchal-
minded Muslims self-select into frequent attendants is not clear. Yet in either case,
the mosque is a space that retains patriarchal values in a non-Muslim environment.
Muslim support for patriarchal values under further controls
The multi-level models in Table 1 test the previous findings for statistical robustness.
Looking at the individual-level effects, all five models indicate that more religious
people are more patriarchal. To be concrete, religiosity adds a .05-fraction of its given
score to patriarchal values. So if religiosity is at its maximum score of 1.0 (the
situation when a respondent considers God as very important), this adds exactly .05
scale points to the mean level of patriarchal values (which is .49). When mosque or
church attendance is at its maximum of 1.0 (the situation when a respondent attends
mosque or church more than once a week), this adds another .04 scale points to the
mean in patriarchal values. When formal education is at its maximum of 1.0 (the
situation when a respondent holds a university degree), this reduces the mean level of
patriarchal values by.12 scale points. When cohort sequences is at its maximum score
of 1.0 (the situation when a respondent is born after 1980), this reduces the mean
patriarchy level by another .11 scale points. And when the respondent is a woman, yet
another .07 scale points are to be subtracted from the mean level of patriarchal values.
These are additive effects under mutual control, which are generally weaker than
these variables zero-order effects on patriarchal values. For instance, the zero-order
effect of Muslim identification on patriarchal values is b.11, indicating that if a
respondent is a Muslim, .11 scale points are to be added to the mean level of
patriarchal values. However, Muslim identity is confounded with high religiosity,
frequent mosque attendance, and lower levels of education and when we control for
these confounding characteristics the effect of Muslim identification is cut in half:
holding everything else constant, being a Muslim still adds another .05 scale points
to the mean level of patriarchal values.
Yet, small as this effect is, it is robust against quite a number of confounding
variables at the individual level. In addition, we see from the insignificance of the
475
480
485
490
495
500
505
510
515
264 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:41
UNCORRECTED PROOF
Men
Women
0.20
0.22
0.24
0.26
0.28
0.30
0.32
0.34
0.36
0.38
0.40
0.42
0.44
0.46
0.48
0.50
0.52
0.54
0.56
0.58
0.60
0.62
0.64
0.66
0.68
0.70
0.72
0.74
0.76
0.78
0.80
Mosque/church attendance
Patriarchal values 2000-06
0.20
0.22
0.24
0.26
0.28
0.30
0.32
0.34
0.36
0.38
0.40
0.42
0.44
0.46
0.48
0.50
0.52
0.54
0.56
0.58
0.60
0.62
0.64
0.66
0.68
0.70
0.72
0.74
0.76
0.78
0.80
Never Less than
once
a Year
Once a year On holidays
only
Once a
month
Once a
week
More than
once
a week
Mosque/church attendance
Never Less than
once
a Year
Once a year On holidays
only
Once a
month
Once a
week
More than
once
a week
Patriarchal values 2000-06
Figure 6. The patriarchal effect of mosque/church attendance among women and men by Muslim identication and Muslim social dominance.
International Review of Sociology Revue Internationale de Sociologie 265
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:41
UNCORRECTED PROOF
Table 1. Multi-level models explaining within-societal and between-societal variation in patriarchal values (series 1)
Dependent variable: patriarchal values
Effects
Model 1-1 Model 1-2 Model 1-3 Model 1-4 Model 1-5
Intercept .49 (44.0)*** .49 (49.9)*** .49 (44.1)*** .49 (48.0)*** .49 (48.4)***
Societal-level effects:
Muslim social dominance .25 (7.1)*** .20 (6.2)*** .25 (7.0)*** .21 (6.0)*** .22 (6.7)***
Base religiosity level .27 (4.2)**
Oil economy Insignificant
Female workforce .32 ( 4.0)***
Enduring democracy .14 ( 4.3)***
Fixed individual-level effects:
Strength of religiosity .05 (6.9)*** .05 (6.9)*** .05 (6.9)*** .05 (6.9)*** .05 (6.9)***
Mosque/church attendance .04 (6.1)*** .04 (6.1)*** .04 (6.1)*** .04 (6.1)*** .04 (6.1)***
Formal education .12 ( 13.6)*** .12 (13.6)*** .12 (13.6)*** .12 (13.6)*** .12 ( 13.6)***
Cohort sequence .11 (8.9)*** .11 (8.9)*** .11 (8.9)*** .11 (8.9)*** .11 (8.9)***
Female sex .07 ( 18.6)*** .07 (18.6)*** .07 (18.6)*** .07 (18.6)*** .07 ( 18.6)***
Randomized individual-level effects:
Muslim identification .05 (5.1)*** .05 (3.2)*** .05 (3.3)*** .05 (3.4)*** .05 (2.9)***
Muslim social dominance Insignificant Insignificant Insignificant Insignificant
Base religiosity level Insignificant
Oil economy Insignificant
Female workforce Insignificant
Enduring democracy Insignificant
Explained variances:
Within-society variation of DV 9.8% 9.7% 9.7% 9.7% 9.7%
Between-society variation of DV 46.4% 55.5% 45.9% 55.6% 53.7%
Variation in effect of Muslim
identification
0% 0% 0% 0% 0%
Notes: Entries are unstandardized regression coefficient s with T-values in parentheses (coefficients standardized to first decimals). N (number of observations) is 129,543
respondents at the individual level and 83 countries at the societal level. Individual-level variables (except dummies) are country-mean centered; societal-level variables are
global-mean centered. Explained variances calculated from change in random variance component relative to base model. Estimates calculated with HLM 6.02.
Significance level: insignificant p .090; * p B.090; ** pB.050; *** pB.010.
266 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:42
UNCORRECTED PROOF
interaction effects in Models 1-2 to 1-5 that societal-level characteristics do not vary
the individual Muslims support for patriarchal values. Regardless of how strongly
Muslims dominate a society, how religious the society is on average, how strongly the
economy depends on oil, how large the female workforce is or for how long
democracy has endured, identifying Muslims are a bit more patriarchal than non-
Muslims under all these conditions.
The effect of Muslim identification on support for patriarchal values is robust but
small. For Muslim social dominance, the picture is different. Here, the effect on
patriarchal values is not only significant but quite sizeable. To be precise, if Muslim
social dominance comes close to its maximum score of 1.0 (the situation where
almost 100% of the adult population is Muslim), this adds .25 scale points to the
mean level of patriarchal values. And variation in Muslim social dominance explains
almost 50% of the cross-national variation in patriarchal values. Moreover, the
patriarchal effect of Muslim dominance remains highly significant against controls
of a societys base level of religiosity and key structural factors, including female
workforce participation and the endurance of democracy. The latter two significantly
diminish patriarchal values at the societal level, yet neither of them absorbs the
patriarchal effect of Muslim social dominance.
Muslim support for patriarchal values and patriarchal power structures
So far, we have not tested Muslim support for patriarchal values against a direct
indicator of patriarchal structures. To measure patriarchal structures directly, one
has to measure the extent to which women are excluded from positions of decision-
making power. Such a measure is available by inverting the UNDPs Gender
Empowerment Measure. This index averages womens advancement to positions of
decision-making power in politics, administration, and business on an interval scale
with minimum 0 and maximum 1.0. Inverted, this indicator measures the exclusion
of women from positions of decision-making power.
The rigidly structural position held by prominent scholars suggests that Islam
affects patriarchal values only in as far as it is linked with structural patriarchy. If this
position is correct, neither individual Muslim identification nor Muslim social
dominance will show a significant effect on patriarchal values once patriarchal
structures are taken into account.
This is not the case, as all models in Table 2 evidence. Patriarchal structures
neither diminish nor change the direction or make less significant the patriarchal
effect of Muslim identification. Visually, this is evident from the two partial
regression plots in Figure 7, which show the effects of Muslim dominance and of
patriarchal structures on patriarchal values under mutual control. Mutually
controlled, patriarchal structures and Muslim social dominance explain, respectively,
48 and 35% of the cross-national variation in patriarchal values.
A societys base level of patriarchal values is a reflection of both patriarchal
structures and Muslim dominance. This finding supports a combined structural and
cultural interpretation of patriarchal values and thereby argues against a rigid
formulation of the structural position.
Some additional findings are noteworthy, especially with respect to interactions.
According to Model 2-2, if the respondent is a woman, this reduces the mean level of
patriarchal values by .09 scale points. If the respondent belongs to the youngest
520
525
530
535
540
545
550
555
560
International Review of Sociology Revue Internationale de Sociologie 267
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:42
UNCORRECTED PROOF
Table 2. Multi-level models explaining within-societal and between-societal variation in patriarchal values (series 2)
Dependent variable: patriarchal values
Effects
Model 2-1 Model 2-2 Model 2-3 Model 2-4
Intercept .49 (56.7)*** .49 (56.4)*** .48 (55.3)*** .49 (56.5)***
Societal-level effects:
Muslim social dominance .17 (5.0)*** .16 (4.9)*** .17 (5.1)*** .16 (4.9)***
Patriarchal power structures .41 (8.4)*** .47 (9.6)*** .46 (9.5)*** .46(9.4)***
Fixed individual-level effects:
Strength of religiosity .05 (5.6)*** .05 (5.3)*** .05 (5.5)*** .05 (5.4)***
Formal education .12 ( 13.0)*** .11 ( 11.0)*** .11 ( 12.9)*** .12 ( 13.0)***
Cohort sequence .11 (8.4)*** .05 (4.4)*** .10 (8.5)*** .10 (7.5)***
Female Sex .07 ( 17.7)*** .09 ( 10.1)*** .05 ( 11.7)*** .07 ( 18.5)***
Female sex education level .02 ( 3.0)**
Female sex cohort sequence .09 (7.9)***
Employment (dummy) .01 ( 2.3)* .02 ( 7.2)***
Unmarried (dummy) .01 (2.9)** .01 (3.8)***
Female sex employed .02 ( 5.8)***
Female sex unmarried .04 ( 11.3)***
Randomized individual-level effects:
Muslim identification .05 (5.3)*** .05 (2.9)** .05 (2.6)** .05 (3.0)***
Muslim social dominance Insignificant Insignificant Insignificant
Patriarchal power structures Insignificant Insignificant Insignificant
Mosque/church attendance .05 (9.6)*** .05 (11.2)*** .05 (9.4)*** .05 (11.7)***
Muslim social dominance Insignificant Insignificant Insignificant
Patriarchal power structures .14 ( 4.4)*** .14 ( 4.1)*** .14 ( 4.3)***
Female sex employed unmarried .02 ( 5.7)***
Muslim social dominance Insignificant
Patriarchal power structures Insignificant
268 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:42
UNCORRECTED PROOF
Table 2 (Continued )
Dependent variable: patriarchal values
Effects
Model 2-1 Model 2-2 Model 2-3 Model 2-4
Intercept .49 (56.7)*** .49 (56.4)*** .48 (55.3)*** .49 (56.5)***
Explained variances:
Within-society variation of DV 10.0% 10.3% 10.6% 10.6%
Between-society variation of DV 70.8% 71.4% 72.1% 71.5%
Variation in effect of Muslim identification 0% 0% 0% 0%
Variation in effect of mosque/church attend 0% 35.7% 35.2% 37.2%
Notes: Entries are unstandardized regression coefficients with T-values in parentheses. N (number of observations) is 122,544 respondents at the individual level and 76
countries at the societal level. Individual-level variables (except dummies) are country-mean centered; societal-level variables are global-mean centered. Explained
variances calculated from change in random variance component relative to base-model. Estimates calculated with HLM 6.02.
International Review of Sociology Revue Internationale de Sociologie 269
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:42
UNCORRECTED PROOF
Zimbabwe
Zambia
Vietnam
Venezuela
Uruguay
U.S.A.
U.K.
Ukraine
Turkey
Trinidad
Thailand
Tanzania
Switzerland
Sweden
Spain
S.Africa
Slovenia
Slovakia
Singapore
Saudi
Arabia
Russia
Romania
Portugal
Poland
Philippines
Peru
Pakistan
Norway
Nigeria
NL
Morocco
Moldova
Mexico
Malta
Mali
Macedonia
Luxemb.
Lithuania
Latvia
Kyrgyztan
Jordan
Japan
Italy
Ireland
Iraq
Iran
Indonesia
India
Iceland
Hungary
Guatemala
Greece
Ghana
Germany(W.)
Germany
(E.)
France
Finland
Estonia
Egypt
Denmark
CzechR.
Cyprus
Croatia
Colombia
China
Chile
Canada
BurkinaF.
Bulgaria
Brazil
Belgium
Bangladesh
Austria
Australia
Argentina
-0.22
-0.20
-0.18
-0.16
-0.14
-0.12
-0.10
-0.08
-0.06
-0.04
-0.02
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
0.22
Patriarchal
values
Muslim
dominance
Y = 2E-17 + .19 * X
Partial R sq.: .35
Stronger than predicted
Stronger than predicted
Weaker than predicted
Weaker than predicted
S.Korea
Zimbabwe
Zambia
Vietnam
Venezuela
Uruguay
U.S.A.
U.K.
Ukraine
Turkey
Trinidad
Thailand
Tanzania
Switzerland
Sweden
Spain
S.Africa
Slovenia
Slovakia
Singapore
SaudiArabia
Russia
Romania
Portugal
Poland
Philippines
Peru
Pakistan
Nigeria
NL
Morocco
Moldova
Mexico
Malta
Mali
Macedonia
Luxemb.
Lithuania
Latvia
Kyrgyztan
Jordan
Japan
Italy
Ireland
Iraq
Iran
Indonesia
India
Iceland
Hungary
Guatemala
Greece
Ghana
Germany(W.)
Germany(E.)
France
Finland
Estonia
Egypt
Denmark
CzechR.
Cyprus
Croatia
Colombia
China
Chile
Canada
BurkinaF.
Bulgaria
Brazil
Belgium
Bangladesh
Austria
Australia
Argentina
-0.22
-0.20
-0.18
-0.16
-0.14
-0.12
-0.10
-0.08
-0.06
-0.04
-0.02
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
0.22
-0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10 0.20 0.30 0.40 0.50
-0.32 -0.28 -0.24 -0.20 -0.16 -0.12 -0.08 -0.04 0.00 0.04 0.08 0.12 0.16 0.20 0.24 0.28 0.32
Structural
patriarchy
Patriarchal
values
Y = 8E-18 + .48 * X
Partial R sq.: .48
Stronger than predictedWeaker than predicted
Stronger than predictedWeaker than predicted
S.
Korea
Figure 7. The partial effects of Muslim social dominance and patriarchal power structures on patriarchal values.
270 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:42
UNCORRECTED PROOF
cohort, this reduces patriarchal values by another .05 scale points. But if the
respondent belongs both to the youngest cohort and is a woman, patriarchal values
fall by yet another .09 scale points.
As is evident from Model 2-3, employment status and marital status have
negligible main effects on patriarchal values. In interaction with female sex, the two
variables affect patriarchal values more strongly but the effect sizes remain small, even
though they point in the expected direction. To be precise, when the respondent is
both a woman and employed, this reduces patriarchal values by .02 scale points. This
adds to reduction of .04 scale points when the respondent is female and unmarried.
More important is the interaction of the individual respondents religious service
attendance with patriarchal structures at the societal level. On average, service
attendance has a weakly positive effect on patriarchal values but this effect varies
pronouncedly with a societys structural patriarchy, becoming stronger in less
patriarchal societies. Thus, service attendance contributes more to patriarchal
values when the surrounding society is less patriarchal. This confirms the previous
interpretation: the social experience of religious service is a more important anchor for
patriarchal values in societies that are less patriarchal.
Conclusion
Scholars disagree on whether Muslim support for patriarchal values is an inherent
element of Muslim identity and, thus, a defining property of Muslim culture. So far,
the evidence has been inconclusive for two reasons. For one, area-specific studies do
not integrate into a coherent body of evidence because differences in operationaliza-
tion, methodology, and model specification defy generalizations across studies.
Second, among the few broadly comparative studies, none tested Muslim support for
patriarchal values against key structural aspects of patriarchy that might account for
the link between Muslim identity and patriarchal values. And no study did this in a
multi-level design that allows one to decide whether Muslim social dominance,
individual Muslim identification or both nurture patriarchal values. To the best of
our knowledge, this study is the first broadly cross-national and multi-level test of
whether Muslim support for patriarchal values proves robust throughout a multitude
of comparative contrasts, comparing Muslims with non-Muslims over inner-societal
divisions by sex, age, education, and religiosity and over between-societal differences
in the percentage of Muslims, the oil base of the economy, the size of the female
workforce, the endurance of democracy, and the prevalence of patriarchal power
structures. What did we find?
No matter what social sub-group we look at, Muslims in this group are always
more patriarchal than non-Muslims. We find this pattern to hold for each group, be
it defined by sex, cohort, religiosity or education. And we find it to hold in each type
of society, regardless of whether the societys economy is oil-based, mobilizes women
into the workforce, is democratic, religious or patriarchal in its power structures.
Even if we strip Muslim identification from confounding individual-level character-
istics, such as religiosity, mosque attendance, and education, its effect on patriarchal
values remains intact, though on a smaller scale. The tendency of Muslims to
support patriarchal values more strongly than non-Muslims of the same category is a
remarkably robust tendency.
565
570
575
580
585
590
595
600
605
International Review of Sociology Revue Internationale de Sociologie 271
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:43
UNCORRECTED PROOF
Evidence for the robustness of this tendency is new in the comparative scope
provided here. Also new is the evidence of how Muslim identity affects patriarchal
values simultaneously at the individual level and the societal level. At the individual
level, Muslim self-identification results in a modest but consistent tendency to
emphasize patriarchal values more than non-Muslims of the same reference category.
At the societal level, a larger proportion of Muslims elevates a societys base level of
patriarchal values. This affects all groups, including non-Muslims, who support
patriarchal values more strongly than elsewhere when they live among more
Muslims. This pattern is largely unnoticed in the literature.
On the other hand, Muslims are by no means a homogeneous social category
with respect to patriarchal values. Instead, Muslim support for patriarchal values
varies considerably over inner-societal group divisions by sex, cohort, religiosity, and
education and over between-societal differences in the percentage of Muslims and
other characteristics. Important in this context, we find that the group characteristics
that vary the patriarchal values of people in general also vary the patriarchal values
of Muslims and do so in the same direction. This makes Muslims susceptible to anti-
patriarchal group tendencies an insight that has not been emphasized enough in
the literature.
This conclusion is supported by another distinctive pattern: Muslim support for
patriarchal values is a strictly relative phenomenon; it is only visible relative to a
given groups reference level of patriarchal values. The relative nature of Muslim
patriarchy suggests that social comparison is a key mechanism involved here, as
reference group theory assumes. It seems that Muslims adjust the strength by which
they emphasize patriarchal values to the strength of these values in their wider social
environment. Such flexibility is an important indication that patriarchal values are
susceptible to emancipative forces also among Muslims.
In this vein, there is evidence for a disruption of the nexus between Muslim
identity and patriarchal values. Confirming the results of smaller studies on a broader
basis, our findings support the view that education and employment erode patriarchal
values more rapidly among Muslim women than Muslim men. This is an important
insight, as it suggests that educating Muslim girls and fostering employment
opportunities for Muslim women widen the gender gap over patriarchal values
among Muslims. In the long run, strategies that create these opportunities for Muslim
women are key treatmentsto diminish Muslim support for patriarchal values.
Notes
1. Patriarchy is the systematic subordination of women to men. This subordination has
structural and cultural facets. The structural facet is evident in organizational patterns
that enforce female subordination. The cultural facet is manifest in values that legitimize
female subordination (cf. Walby 1989, p. 227). In the following, we use the term culture
in reference to subjective beliefs and structure in the sense of objective patterns of
organization, be those social, economic or political.
2. More precisely, they evidence a Muslim repulsion of gender-egalitarian values. But since
gender-egalitarian values are the opposite of patriarchal values, the repulsion of gender-
egalitarian values is the same as an afnity to patriarchal values.
3. Documentation of sample sizes, sampling methods, questionnaire wording and data access
can be found at the WVS website at www.worldvaluessurvey.org.
4. See the Internet appendix for the list of these countries.
610
615
620
625
630
635
640
645
650
655
272 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:43
UNCORRECTED PROOF
5. Religious identication is asked in WVS question V185: Do you belong to a religion or
religious denomination? If yes, which one? (Code answer due to list below.) In different
countries, country-specic lists of denominations were used. WVS variable V185
summarizes them in a list of 90 different denominations. Code 49 represents Muslim
denomination.
6. We are interested in patriarchal values as such and in religiosity as a potential inuence.
Therefore we keep the two apart and do not follow the practice of Blaydes and Linzer
(2008) to summarize patriarchal values and religiosity into an overall measure of
fundamentalist values.
7. See the Appendix for the exact coding scheme.
8. To bring the distribution on the percentage scale of Muslims closer to a normal
distribution, we logged the percentages. For details, see the Appendix.
9. See the Appendix for the exact coding of this variable.
10. See the Appendix for the exact coding.
11. See the Appendix for the exact coding.
12. See the Appendix for the exact coding.
13. School systems differ in content and quality, which lowers the equivalence of educational
categories across countries. But it does not eradicate equivalence entirely. In all countries,
higher levels of formal education imply more knowledge and a better understanding of the
world and higher cognitive skills. Thus, inequivalence is partial and not complete.
A regularity that surfaces even under partial inequivalence is actually more likely to be
true because partial inequivalence obscures regularities.
References
Afshar, H., 1982. Khomenis teachings and their implications for Iranian women. In Tabari
and Yeganeh, eds. In the shadow of Islam. London: Zed Books, 7590.
Beit-Hallahmi, B. and Argyle, M., 1997. The psychology of religious behavior, belief and
experience. London: Sage.
Blaydes, L. and Linzer, D., 2008. The political economy of womens support for
fundamentalist Islam. World Politics, 60, 576609.
Bryk, A. and Raudenbush, S., 2002. Hierarchical linear models. Thousand Oaks, CA: Sage.
Buijs, F. and Rath, J., 2002. Muslims in Europe. New York: Russell Sage Foundation.
Burn, S.M. and Busso, J., 2005. Ambivalent sexism, scriptural literalism, and religiosity.
Psychology of Women Quarterly, 29, 412418.
Butler, D. and Stokes, D.E., 1974. Political change in Britain: the evolution of electoral choice.
London: Macmillan.
Fetzer, J. and Soper, C., 2005. Muslims and the State in Britain, France and Germany.
Cambridge: Cambridge University Press.
Fish, M.S., 2002. Islam and authoritarianism. World Politics, 55, 437.
Ghoussoub, M., 1987. Feminism- or the eternal masculine- in the Arab world. New Left
Review, 161, 313.
Harkness, G., 1972. Women in church and society. Abingdon Press
Hewer, C.T.R., 2006. Understanding Islam. London: SCM Press.
Heine, P. and Spielhaus, R., 2009. Was glauben Muslime? [What do Muslims believe?] In
Woran glaubt die Welt? Analysen und Kommentare zum Religionsmonitor 2008 [What does
the world believe? Analyses and comments to the Religion Monitor 2008], Gutersloh:
Bertelsmann Stiftung, 609632.
Hood, R.W., Hill, P.C., and Williamson, W.P., 2005. The Psychology of Religious
Fundamentalism. New York: Praeger.
Inglehart, R. and Norris, P., 2003a. The true clash of civilizations. Foreign Policy, March/
April, 6270.
Inglehart, R. and Norris, P., 2003b. Rising tide. New York: Cambridge University Press.
Joanes, D.N. and Gill, C.A., 1998. Comparing Measures of Sample Skewness and Kurtosis.
The Statistician, 47 (1), 183189.
Karam, A., 1998. Women, Islamisms and the State. New York: St. Martins Press.
660
665
670
675
680
685
690
695
700
705
International Review of Sociology Revue Internationale de Sociologie 273
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:43
UNCORRECTED PROOF
Koopmans, R. and Statham, P., 1999. Challenging the liberal nation-state? American Journal
of Sociology, 105, 652696.
Kosmin, B. and Mayer, E., 2001. Prole of the US Muslim population. New York: The
Graduate Center, City University of New York.
Laurence, J., 2007. Integrating Islam. Migration Policy Institute. http://www.migrationpolicy.
org/pubs/LaurenceIslamicDialogue100407.pdf.
Laurence, J. and Vaisse, J., 2006. Integrating Islam: political and religious challenges in
contemporary France. Washington DC: Brookings.
Lesthaege, R., 2004. Value orientations and the second demographic transition in northern,
western, and southern Europe: an update. Demographic Research, 3, 14351471.
Lewis, B., 2004. The crisis of Islam: holy war and unholy terror. New York: Random House.
McCallum, R.C. and Browne, M.W., 1993. The Use of Causal Indicators in Covariance
Structural Models. Psychological Bulletin, 114 (3), 533541.
Minces, J., 1982. The house of obedience. London: Zed Books.
Moghissi, H., 1999. Feminism and Islamic fundamentalism. London: Zed Books.
Morin, R. and Horowitz, J., 2006. Europeans debate the scarf and veil. Pew Research Center
Report, Nov. 20. Pew Research Foundation.
Okin, S., 1999. Is multiculturalism bad for women? In: J. Cohen, M. Howard and
M. Nussbaum, eds. Is multiculturalism bad for women? Princeton, NJ: Princeton University
Press, 726.
O
¨
gelman, N., 2003. Documenting and explaining the persistence of homeland politics among
Germanys Turks. International Migration Review, 37, 163192.
Paxton, P. and Hughes, M., 2007. Women, politics and power. Los Angeles, CA: Pine Forge
Press.
Peek, C., Lowe, G., and Williams, L.S., 1991. Gender and Gods word. Social Forces, 69,
12051221.
Pfaff, S., 2007. Integrating Muslims in Europe and the United States. Paper prepared for
Multiculturalism and its Discontents Conference. Boulder, CO.
Pipes, D., 2002. The challenge of Muslim immigration. Policy Review, August-September.
Rizzo, H., Abdel-Latif, A., and Meyer, K., 2007. The relationship between gender equality
and democracy. Sociology, 41, 1151.
Ross, M., 2008. Oil, Islam and women. American Political Science Review, 102, 107123.
Ruether, R.R., ed., 1974. Religion and sexism. Simon and Schuster: New York.
Tabari, A., 1982. Islam and the Struggle for the emancipation of Iranian women. In Tabari
and Yeganeh, eds. In the shadow of Islam. London: Zed Books, 525.
Walby, S., 1989. Theorizing patriarchy. Sociology, 23 (2), 213234.
Appendix
Uniform scaling
Like the patriarchal values index, all variables in the analyses are normalized into a uniform
format from minimum 0 to maximum 1.0, which means that multi-point indices take any
fractional value between 0 and 1.0, such as .25, .33, .50, .66 or .75. We use this standard format to
allow for a uniform interpretation of regression coefcients across all variables: coefcients will
appear as fractions between 0 and 1.0 and their value will tell us how many scale points we have to
add to or subtract from the overall mean in patriarchal values when the respective independent
variable is at its maximum 1.0. A coefcient of 12 for formal education, for instance, tells us
that when education is at its maximum value of 1.0 (for respondents with a university degree), this
reduces patriarchal values by .12 scale points. If the score in education is lower than 1.0, this
reduces patriarchal values by a .12 fraction of whatever that score in education is.
Recoding the patriarchal value items
Thus, variable V44 was coded 0 for disagree,.5forneither agree, nor disagree, and 1.0 for
agree. We interpret the recoded variable as an indication of patriarchal values in the domain
710
715
720
725
730
735
740
745
750
755
760
274 C.Welzel and A. Alexander
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:43
UNCORRECTED PROOF
of labor-market participation. For both items, the most patriarchal position is the strongly
agreeoption and the least patriarchal one is the strongly disagreeoption. Thus, we code 0 for
strongly disagree, .33 for disagree, .66 for agree and 1.0 for strongly agree.
List of countries in the sample with proportions of Muslims above 10%
India (11%), Bulgaria (12%), Ghana (15%), Singapore (29%), Macedonia (30%), Cyprus
(33%), Tanzania (39%), Nigeria (40%), Burkina Faso (49%), Malaysia (56%), Kyrgyzstan
(88%), Bangladesh (89%), Indonesia (90%), Egypt (92%), Mali (93%), Jordan (97%), Pakistan
(98%), Saudi Arabia (98%), Iran (99%), Iraq (99%), Morocco (99%), Turkey (99%).
Taking logs of the percentage of Muslims
The distribution on the percentage index of Muslims deviates from a normal distribution.
The skewness shows a positive value of 1.76, indicating a right-skewed distribution. This
reects an imbalance between many societies with few Muslims and few societies with many
Muslims. In addition, the kurtosis yields a positive value of 1.47, indicating a leptokurtic
distribution. This reects the fact that a large proportion of our societies have either pretty
low or pretty high percentages of Muslims while fewer societies have medium-level
percentages. To normalize the distribution, we take logs of the percentage of Muslims,
which stretches differences at the lower end of the percentage scale and condenses differences
at the upper end. The result is a closer-to-normal distribution: the skew of the logged
percentages is .34 and the kurtosis is 1.21. Thus, with logged percentages, fewer societies are
found at extreme values and the distribution is less left-leaning. To put the logged percentages
into the 0-to-1.0 standard format of all other variables, we set the highest logged percentage
(4.6ln(99)) for Saudi Arabia at maximum 1.0 and the lowest logged percentage
(2.3ln(0.1)) for Japan at 0.
Coding of the cohort variable
The coding is: 0, born before 1921; .125, born between 1921 and 1930; .250, born between
1931 and 1940; .375, born between 1941 and 1950; .500, born between 1951 and 1960; .625,
born between 1961 and 1970; .750, born between 1971 and 1980; 1.0, born after 1980.
Coding of The Education Variable
The coding is: 0 for no formal education, .125 for incomplete primary school, .250 for
complete primary school, .375 for incomplete secondary school of the technical/vocational
type, .500 for complete secondary school of the technical/vocational type, .625 for
incomplete secondary: university-preparatory type, .750 for complete secondary: univer-
sity-preparatory type, .875 for some university-level education, without degree and 1.0 for
university-level education, with degree.
Coding of the religiosity variable
The recoding is as follows: 1 into 0, 2 into .12, 3 into .23, 4 into .34, 5 into .45, 6 into .56, 7 into
.67, 8 into .78, 9 into .89, 10 into 1.0.
Coding of the service attendance variable
The coding is: 0 never, practically never, .17 less than once a year, .33 once a year, .50 only
on special holy days, .67 once a month, .83 once a week, 1.0 more than once a week.
765
770
775
780
785
790
795
800
International Review of Sociology Revue Internationale de Sociologie 275
{CIRS}articles/CIRS581801/CIRS_A_581801_O.3d[x] Friday, 3rd June 2011 20:15:43
... Such reasoning is found, for example, when gender gaps at higher systemic levels are held responsible as causes of emergence of gender gaps at lower systemic levels. A well-known example is the assumption that gender gaps on the individual level can be attributed to patriarchal social structures that favor men [27,28]. However, the reverse line of reasoning, that gender gaps at lower systemic levels cause gender gaps at higher systemic levels is equally possible [29]. ...
... We chose the KSA as the setting for our study. Its prevailing patriarch structures favoring men [27,28] and the reversed gender achievement gap that demonstrated for the KSA in many studies [51] make it particularly suitable for ex the filter-empowerment heuristic (in contrast to the correspondence heuristic going into more detail about our study, we will briefly describe the education s the KSA. ...
... We chose the KSA as the setting for our study. Its prevailing patriarchal social structures favoring men [27,28] and the reversed gender achievement gap that has been demonstrated for the KSA in many studies [51] make it particularly suitable for examining the filter-empowerment heuristic (in contrast to the correspondence heuristic). Before going into more detail about our study, we will briefly describe the education system in the KSA. ...
Article
Full-text available
The reversed gender achievement gap in the Kingdom of Saudi Arabia (KSA) in favor of girls developed in a societal environment that still predominantly favors men. The finding illustrates how equity gaps at higher systemic levels may not translate seamlessly to lower systemic levels. We propose that this lack of correspondence between systemic levels regarding equity gaps can be explained by the fact that more exogenous learning resources (educational capital) selectively reach girls’ actiotopes, enabling them to build more effective endogenous learning resources (learning capital) and, in particular, effective action repertoires. To investigate this assumption, we introduce a filter-empowerment heuristic and contrast it with a simpler correspondence heuristic. We employ the filter-empowerment heuristic and the education and learning capital approach to investigate the reversed gender achievement gap in a sample of secondary school students in the KSA. We test four hypotheses: (a) Girls have higher academic achievement than boys. (b) Girls have more educational capital and (c) more learning capital than boys. (d) Educational and learning capital mediate the relationship between gender and academic achievement. The hypotheses were tested with a sample of 2541 ninth-grade students from 55 KSA schools. The four hypotheses were confirmed and support a filter-empowerment heuristic rather than a correspondence heuristic.
... Prior research has shown that MENA countries remain among the least secularized in the world (e.g., Eller, 2010;Kasselstrand et al., 2023). Religious people tend to hold more traditional or patriarchal views on gender issues (Zuckerman, 2009;Schnabel, 2016), a pattern that remains particularly strong among Muslims (Alexander and Welzel, 2011). Being raised religious and within a MENA context that overall tends to be characterized by strong religious authority thus exposes people to more traditional gender roles. ...
... The article focuses on a region that struggles with the stigma of fostering and protecting patriarchal structures, which is often ascribed to prevalent conservative religious beliefs (Inglehart et al., 2003;Alexander and Welzel, 2011;Tausch and Heshmati, 2016). By analyzing data collected by the Arab Barometer in 2018-19, we were able to simultaneously test which mechanisms are at play for the wider region while controlling for country clustering in the data. ...
... However, there is still a leeway for improvement, as we observe substantive cross-country variation, in line with the findings of previous studies (e.g., Coffé and Dilli, 2015). Exposure to diverse worldviews especially via the Internet and social media and by separating religious traditions and potential resulting constrains from social and political lives may contribute to more liberalization (see also Zuckerman, 2009;Alexander and Welzel, 2011;Schnabel, 2016;Tausch and Heshmati, 2016). We also find some support for the pathway of educational attainment, following the previous literature on the impact of education (e.g., Bolzendahl and Myers, 2004;Auletto et al., 2017;Kyoore and Sulemana, 2019). ...
Article
Full-text available
This paper explores the question of what explains public opinion of women empowerment in the Middle East and North Africa. Muslim societies have often been accused of conservatism toward empowerment, stripping women of equal access to education and opportunities. However, many predominantly Muslim societies in the MENA region seem to be on the way to implement change to provide women with more rights. Prior research points to exposure to diversity as a contributor to the acceptance of a more egalitarian role of women in society. This article analyzes different mechanisms of the exposure hypothesis and whether they contribute to predicting positive public perceptions of women empowerment in the region. The empirical analyses rely on public opinion data collected by the Arab Barometer in 2018–19. The descriptive findings suggest attitudinal differences across countries, but also significant gender gaps and divergences across core explanatory factors found under the umbrella of the exposure hypothesis, such as diverse urban living, keeping religion a private matter, and connecting with the world via social media. These factors seem important to shift people's minds and to pave women's long way to liberalization.
... Based on the World Values Survey, which is one of the largest and most comprehensive surveys in the social sciences, Inglehart and Norris (2003a, b) substantiate an 'elective affinity between Islam and patriarchal values'. Rejection of equal status for women in life domains such as education, economics, and politics is found to be most prevalent in Muslim-majority societies and among Muslims (see also Alexander and Welzel 2011;Lussier and Fish 2016). Studies on 'Islamic feminism' (e.g., Glas et al. 2018;Glas and Alexander 2020;Glas and Spierings 2019;Masoud et al. 2016) criticize the narratives and framings of the aforementioned inquiries. ...
... Several follow-up studies replicated these findings using more sophisticated multilevel analyses. The results are straightforward and confirm that patriarchal values are more prevalent among Muslims at the individual level and in societies in which Islam is the predominant religion (Alexander and Welzel 2011;Lussier and Fish 2016;Norris 2014). In all cited studies, those empirical patterns are attributed to religious socialization effects and some unique aspects of Islam. ...
... As a consequence, patriarchal ideas that emerged centuries ago continue to shape the thinking of Islamic ulama today, a situation that in turn impacts the values of ordinary Muslims as they are usually exposed to conservative spiritual leaders upon whom they rely for the interpretation of their religion (Lussier and Fish 2016, p. 32-33). 4 Alexander and Welzel (2011) point out that mosques play the role of an important socializing institution in this context. The transmission of patriarchal values is favored by par-4 If one takes the conditions on the Arabian Peninsula during the advent of Islam as a yardstick, then Islam has contributed to an improvement rather than a deterioration of the legal situation of women-yet there can be no talk of equality that meets modern standards (Koopmans 2020, p. 101). ...
Article
Full-text available
Is Islam a religion that promotes patriarchy? In the academic debate, there are different assessments. On the one hand, there is the thesis of an elective affinity between Islam and patriarchal values. In Muslim-majority countries and among Muslims, support for patriarchal values is most pronounced. On the other hand, there is the antithesis of Islamic feminism, which shows that a significant proportion of devout Muslims support gender equality. It is therefore wrong to describe Islam as a misogynistic religion. What matters is whether the religion is interpreted in an emancipatory manner. This contribution offers a synthesis and argues that religious fundamentalism provides a more valid explanation for patriarchal values than simplistic references to Islam. The 6th and 7th waves of the World Values Survey were analyzed to test this research-guiding hypothesis. Multilevel analyses show that value differences between Muslims and non-Muslims and between Muslim-majority societies and societies with another majority religion turn out to be small or even insignificant when controlling for religious fundamentalism. Fundamentalism is the central driver of patriarchal values and generates uniform effects. At the individual-level, fundamentalism makes both Muslims and non-Muslims more susceptible to patriarchal values. Moreover, Muslims and non-Muslims adapt to the conformity pressures of their societies, resulting in egalitarian as well as patriarchal values, depending on the prevalence of fundamentalism. The high support for patriarchal values in Muslim-majority countries has a simple reason: Religious fundamentalism is by no means a marginal phenomenon in these societies, but rather the norm.
... The approach to rehearsing Islam has hence prompted isolation in instruction in Saudi Arabia, and thusly has made isolation in political, financial, and workforce conditions. With the ongoing battle of normal practices and regulations, ladies have taken extraordinary steps to acquire instruction in Saudi Arabia (Alexander & Welzel 2011;Bacha et al, 2021). ...
... Instead, it has been exploited as a pawn in political games between males arguing about unrelated issues (Charrad, 2011). Alexander and Welzel (2011), macro sociologists from the universities of Gothenburg and Leufana, consider several possible reasons for the greater patriarchy of Muslim societies in a comparative perspective. Ladies in Saudi Arabia keep on being minimized nearly to the place of absolute avoidance from the Saudi labor force. ...
Article
Saudi Arabia is an extraordinary intermingling between religion and culture. Based on their religion and culture Saudi Arabia women’s met challenges for the public authority over education system. In Arabic countries male dominance and feminisms plays equal roles in order to overcome it the women’s should come front to stand their rights. The place of ladies in this general public is confounded and they need to confront a great deal of obstructions to get instruction as they live under male power constantly. Along these lines, this paper will assess the writing on ladies and society in Saudi Arabia and its connection to their instructive accomplishment, considering a few unique viewpoints including: Feminist hypothesis, Islamic women's liberation in Saudi society, the situation with ladies in Saudi society and the Holy Book of the Muslims as well as their correlation in both. Arab ladies, long consigned to the confidential circle by regulation and social custom, are acquiring new admittance to public life. All nations of the Arab presently have labor force "nationalization arrangements" that expect to decrease reliance on traveler work by getting more ladies into the labor force. Saudi Arabia put forth an objective of 30% female work interest. In Kuwait, female residents dwarf male residents in the labor force. Furthermore, across the Gulf, ladies dwarf men in advanced education enlistment. Ladies are making political advances in the district, as well. In Qatar some ladies have been delegated to ecclesiastical situations beginning and some ladies have stood firm on bureau footholds in Kuwait starting around 2005, including wellbeing pastor, transportation clergyman and money serve. Development going to maintain control feminist part women come in front and restriction for women.
... In a society without a free expression of speech, public opinion will be hampered in this change, resulting in, on average, more conservative social values. Third, we include percentage of Muslims in the origin population, following the research that emphasises different cultural traditions and the divide in social values between Western and Muslim countries, with people in the latter being more conservative and the importance of origin-country traditions on immigrants' social values [35,43,44]. Including percentage of Muslims is additionally in accordance with literature emphasising different barriers to diffusion of information and norms across populations [45]. ...
Article
Full-text available
Concerns have been raised that immigrants coming to Europe bring fundamentally different social values, affecting the more liberal receiving societies negatively. However, the topic of immigrants’ social values is understudied, and much research studies only one issue at a time, lacking a systematic approach to compare immigrants and native-born across issues. We study the social values of immigrants in Sweden using a large sample of newly arrived immigrants and their opinions on 35 different moral issues. Our results indicate a large heterogeneity across different issues, with, on average, a general tendency towards liberal social values among immigrants. We find that individual characteristics are more important than characteristics of the country of origin in explaining variation of social values between immigrants. Religiosity has the largest effect, with more religious individuals having more conservative stances. Using external data sources, we compare immigrants with native-born regarding both average positions on different issues, and the correlation between issue positions. Compared with the native-born, immigrants have, on average, somewhat more conservative values, but the underlying values structure is the same.
... 3.75 Pearson's v 2 (9) ¼ 12.1897/Pr ¼ 0.203 ***P < 0.001, **P < 0.01, *P < 0.05. Explaining Backlash treatment was significantly related to gender and education level-and because prior research shows that respondent gender, interviewer gender (Benstead 2014), lineage, gender attitudes, rural residence, education, income, and age are related to attitudes toward gender equality in different societies (Alexander and Welzel 2011;Muriaas et al. 2019)-we control for these factors and include post-stratification weights the regressions. ...
Article
Full-text available
Governments promote gender-sensitive policies, yet little is known about why reform campaigns evoke backlash. Drawing on social position theory, we test whether marginalized (women’s organizations) or intrusive (Western donors) messengers cause resistance across public rights (quotas) and private rights (land reform). Using a framing experiment implemented among 1,704 Malawians, we find that females’ attitudes are unaffected by campaigns, while backlash occurs among patrilineal and matrilineal males. Backlash among men is more common for sensitive private rights (land reform) than public rights (quotas) and Western donors than women’s organizations, suggesting complex effects generally more consistent with the intrusiveness hypothesis.
... The problem, however, is that these studies do not acknowledge their narrow focus on particular public-sphere values. Rather, these works tend to imply that their findings hold for gender values across the board, as is also reflected by references not to specific dimensions of gender values but to general "patriarchal values" (Alexander & W elzel, 2011) or "gender role attitudes" (Ng, 2022). By generalizing their findings to "gender values", these studies, like the ones discussed above, brush over the notion that Islamic religiosity might relate differently to support for public-sphere equality than to other values. ...
Preprint
Full-text available
Compared to race and class, religion has not received its fair share of attention as a potentially crucial axis of intersectionality. This chapter introduces readers to the marginalized field of gender scholarship into religion, focusing on Islamic religiosity and gender values. W hile populist public debates argue that Islam necessarily opposes support for gender equality, findings from the public opinion literature are mixed. I argue that one reason for discrepancies is that different studies conceptualize "Islamic religiosity" and "gender values" in divergent ways that shape their conclusions. Studies that consider Islam and gender values as flat, unidimensional concepts tend to conclude that Islam only has patriarchal effects, whereas studies that opt for more nuanced, multidimensional conceptualizations also lay bare the feminist potential of Islam. However, including fine-grained disaggregations of both "gender values" and "Islamic religiosity" might be too complex, and I provide readers with two possible ways to overcome this issue.
Chapter
This book re-evaluates Almond, Verba, and Pye's original ideas about the shape of a civic culture that supports democracy. Marshaling a massive amount of cross-national, longitudinal public opinion data from the World Values Survey Association, the authors demonstrate multiple manifestations of a deep shift in the mass attitudes and behaviors that undergird democracy. The chapters in this book show that in dozens of countries around the world, citizens have turned away from allegiance toward a decidedly 'assertive' posture to politics: they have become more distrustful of electoral politics, institutions, and representatives and are more ready to confront elites with demands from below. Most importantly, societies that have advanced the most in the transition from an allegiant to an assertive model of citizenship are better-performing democracies - in terms of both accountable and effective governance.
Chapter
This book re-evaluates Almond, Verba, and Pye's original ideas about the shape of a civic culture that supports democracy. Marshaling a massive amount of cross-national, longitudinal public opinion data from the World Values Survey Association, the authors demonstrate multiple manifestations of a deep shift in the mass attitudes and behaviors that undergird democracy. The chapters in this book show that in dozens of countries around the world, citizens have turned away from allegiance toward a decidedly 'assertive' posture to politics: they have become more distrustful of electoral politics, institutions, and representatives and are more ready to confront elites with demands from below. Most importantly, societies that have advanced the most in the transition from an allegiant to an assertive model of citizenship are better-performing democracies - in terms of both accountable and effective governance.
Article
This paper takes one of the first, direct approaches to understanding which factors shape which attitudes towards gender equality among political elites. We examine support for gender equality among legislators in 13 Latin American countries, using 10 new questions from the 2015–2018 wave of the Latin American elites survey (PELA). We argue that legislators’ attitudes about gender equality fall into three distinct dimensions: holding egalitarian views, recognizing that gender inequality is a problem, and supporting state action to hasten gender equality. Overall, women express the more gender-equal orientation on all three dimensions, while factors like religiosity and ideology matter differently for different dimensions. These findings demonstrate the need for more nuance in measuring and analyzing attitudes towards gender equality, in order to better understand the link between descriptive and substantive representation.
Book
Full-text available
This book demonstrates that people’s basic values and beliefs are changing, in ways that affect their political, sexual, economic, and religious behaviour. These changes are roughly predictable: to a large extent, they can be interpreted on the basis of a revised version of modernisation theory presented here. Drawing on a massive body of evidence from societies containing 85 percent of the world’s population, the authors demonstrate that modernisation is a process of human development, in which economic development gives rise to cultural changes that make individual autonomy, gender equality, and democracy increasingly likely.
Book
European governments must struggle with assimilating Muslim newcomers into their countries, with so many more now living in Western Europe. Britain, France, and Germany have dealt with the related problems differently. This book explains why their policies differ and proposes ways of ensuring the successful incorporation of practicing Muslims into liberal democracies. Resolving their issues has become all the more urgent in the aftermath of the terrorist attacks of September 11, 2001. Considers complicated questions of policy towards Muslims in Europe after 9/11. Emphasis on specifically religious aspects of Muslim mmigrants in Europe. Unique focus on church-state relations
Chapter
Religion is obviously shaped by the social context within which it exists. Hence, one might expect some legitimacy to the concept of rural religion. Furthermore, if there is even partial truth to the claims that, relative to more cosmopolitan cultures, rural cultures are less heterogeneous, less tolerant of divergencies in belief and action, and less abstract in their conceptualizations, then these claims should hold true to the same extent for rural religion as they do for rural culture in general. In this sense, religious fundamentalism is not unanticipated in rural areas. Yet, as we shall see, it is more than just a curiosity that the persistence of the faith of the rural Roman (the paganus) became an English word for unbeliever (pagan) and the religious conservatism of the English heath dwellers became another word for unbeliever, heathen (Smith & Zopf, 1970). Yet these terms conceal a bias that also characterizes the scientific study of religion in general and of religious fundamentalism in particular. The problem of bias, noted through many chapters of this text, is even more apparent when investigators confront not only a rural culture unlikely to be their own, but also a religious fundamentalism dominant in that culture and to which they are unlikely to subscribe.