547898
research-article2014
SOC0010.1177/0038038514547898SociologyYonay et al.
Article
Religious Heterogeneity and
Cultural Diffusion: The Impact
of Christian Neighbors on
Muslim and Druze Women’s
Participation in the Labor
Force in Israel
Sociology
1­–19
© The Author(s) 2014
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0038038514547898
soc.sagepub.com
Yuval P Yonay
University of Haifa, Israel
Meir Yaish
University of Haifa, Israel
Vered Kraus
University of Haifa, Israel
Abstract
This study exploits the unique demographic structure of the Arab-Palestinian minority in Israel and
their geographical immobility in order to help resolve the riddle why women in the Middle East and
North Africa are less likely to participate in the labor force than women elsewhere in the world.
We show that, controlling for economic variables, Muslim and Druze Arab women are more likely
to enter the labor force if they live in a locality where Christian Arabs live as well. A possible
explanation of this finding is the impact of social interaction among people who have different cultural
schemas. Female labor force participation is rising throughout the Middle East, including among ArabPalestinians in Israel, but the tempo of this transformation depends on various local variables, and
in this article we identify one such factor, namely, the ethno-religious composition of a community.
Keywords
Arab women, cultural diffusion, Druze women, labor market participation, MENA region,
Muslim women, Palestinian women, residential segregation, social inequality
Corresponding author:
Meir Yaish, Department of Sociology and Anthropology, University of Haifa, Haifa 31905, Israel.
Email: [email protected]
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
2Sociology 
Introduction
Despite the high level of labor force participation (LFP) of women in Israel, only about
one in four Muslim and Druze women in Israel is engaged in economic activity in the
formal economy. This relatively low rate of engagement in the labor force is similar to
the overall participation rate of women in the Middle East and North Africa (MENA).
Such a low level of LFP is commonly attributed to religion and to other cultural factors. Thus, for example, it is argued that Muslim and Arab societies place more importance on gender segregation and discourage women’s presence in the public sphere,
thus discouraging women’s employment outside their homes (Jacobsen, 2007; Minces,
1982; Salman, 1987).
Growing literature on women in the MENA region challenges the focus on the cultural explanation of women’s low rates of employment and points to the diversity of
female labor force participation (F-LFP) across MENA countries. A variety of economic
variables are suggested to account for this variability, yet such studies have not been
enough to discredit the claim that the ostensible objection to women’s equal rights in
many MENA countries is a crucial factor in curbing F-LFP, and attempts are being made
to reconcile the cultural and the economic paradigms.
In this article, we use the unique demographic structure of Israel and the lack of selfselection to place of residence due to the tendency of Arab-Palestinians to stay in the
communities where they grew up to shed light on the determinants of the low levels of
Muslim women’s LFP in Israel. We show that for Arab-Palestinians, Christians living in
a community that is also composed of Muslims and Druze positively affects the F-LFP
of the latter two. A possible explanation for this is that the exposure to ‘modern’ cultural
patterns adopted earlier and faster by Christian Palestinians facilitates similar trends
among non-Christian Palestinians. This effect illustrates the importance of cultural processes without assuming that cultural concepts and norms are static and resist adaptation
to new conditions.1
Theoretical Background
Read and Cohen (2007) claim that explaining the LFP of women from ethnic and racial
minorities in the US requires expanding the models that have been constructed on the
basis of research on majority white women. The main direction for theoretical expansion they suggest is what they call the ‘cultural gender norms’ which are ‘clearly significant for women’s employment, especially in non-Western, non-industrialized nations’
(2007: 1730). Read’s own study of American-Arab women demonstrates this point.
Many Arab women, she argues, regard education as a resource for raising children and
not as human capital (Read, 2003; Read and Oselin, 2008). Read and Cohen’s (2007)
comparison of the determinants of F-LFP among different ethnic groups in the US is
clearly relevant for a similar comparison between different societies, and especially for
the question whether ‘culture’ should be used as a variable explaining differences in
women’s paid employment in various societies. This view fits in with the revived ‘interest in culture as an explanatory variable’ (Harrison and Huntington 2000; Huntington,
2000: xiv). Many scholars, however, do not look favorably on this development and still
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
3
Yonay et al.
look suspiciously at cultural explanations of international and inter-ethnic differences
(Lamont and Small, 2008).
This controversy over the role of ‘culture’ in explaining social variance is especially
important in the Middle East and North Africa, the region that lags behind all other
regions of the world in terms of women’s employment (Moghadam, 2004; Spierings
et al., 2010). In accounting for this persistent and widespread deficiency, many scholars
have indeed resorted – too often unreflectively – to the Arab cultural heritage and to the
religious characteristics of Islam as key factors. While the emphasis on women’s chastity
and modesty have been features of many patriarchic societies throughout the world, it is
widely believed that they have played a much bigger role in Arab culture and in the
Muslim faith, restricting women’s movements and activity more than anywhere else
(e.g., Jacobsen, 2007; Minces, 1982; Salman, 1987). In many Muslim societies, women’s
work, clothes, and movement are still legally regulated by the state, and even where they
are not subject to such legislation, strong Islamist movements exert pressures on women
to follow traditional proscriptions (Ahmed, 1992; Anwar, 2009; Haddad and Esposito,
1998; Joseph and Slyomovics, 2001; Kandiyoti, 1991; Mernissi, 1993). Feminist and
post-colonial critiques challenge the essentialist assumptions behind the view that
Muslim and Arab cultures are inherently more inimical to gender equity than other societies but they do not deny the harsh conditions Muslim and Arab women face (AbuLughod, 2002; Ahmed, 1992; Haddad and Stowasser, 2004; Sonbol, 2010).
Scholars who feel uncomfortable with culturalist accounts claim that low F-LFP in
the MENA region is, as in other countries, a consequence of their economic structure
and the specific development of their polity (Kandiyoti, 1991; Moghadam, 1998,
2003; Shukri, 1999). As a consequence of those economic and political factors, F-LFP
rates greatly differ across the Muslim world. For example, in 1991, F-LFP was only
9.8% in Iran, 11.1% in Algeria, and 13.2% in Jordan, while in Libya it was 24.4%, in
Egypt 26.9%, and in Turkey 35.3%. Furthermore, in most countries, F-LFP has continued to rise: for example, in 2010, to 23.9% in Algeria; in Jordan to 28.3%; in Iran
to 23.2%; and in Libya to 41.7%.2 Such figures demonstrate that under favorable
circumstances Muslim women do join the workplace, politics, and voluntary associations (Clark, 2003; Clark and Schwedler, 2003; Joseph and Slyomovics, 2001;
Moghadam, 2003, 2004).
Demonstrating the impact of economic and political variables on the female labor
force, the consequent variance of women’s participation rates among Arab and Muslim
countries, and the rising rates in almost all those states are enough to dispel what Lamont
and Small call ‘thin understanding of culture’ (2008: 76), according to which Arab and
Muslim cultures are powerful enough to block women’s entrance into the labor force. Yet
such evidence is not enough to disprove the claim that ‘Islam’ and ‘Arab culture’ do have
lingering influences that slow down women’s activities in the public sphere (Norris and
Inglehart, 2004). It is therefore possible that F-LFP rates in Muslim countries diverge
due to differences in the strength of religious authorities, their adjustment to industrial
societies, and their alliances with other social forces (Anwar, 2009; Charrad, 2001; Clark
and Schwedler, 2003; Clark and Young, 2008). Acknowledging both structural and cultural influences, several researchers have used quantitative methods in an attempt to
isolate and measure the impact of each factor. Yet since the ‘measuring’ of culture is very
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
4Sociology 
indirect and imprecise, evidence about relative importance of various factors is far from
conclusive (Clark et al., 1991; Karshenas, 2001; Spierings et al., 2010).
Thus, the role of culture in shaping women’s status in the Middle East continues to stir
up fierce controversies. The status of women in Muslim and Arab societies stands in the
center of current political and ideological struggles within those societies and in the
alleged ‘clash’ of civilizations between Western and Muslim societies.3 Resorting to
‘culture’ to explain different behavioral patterns runs the risk of orientalism, that is, of
assuming that other peoples, Muslims, in this case, are inherently different from ‘us,’
Westerners (Ahmed, 1992; Beck and Keddie, 1978; Said, 1978). Yet people’s beliefs,
conceptualizations, images, and values – that is, ‘culture’ – are an essential component
of any society. They channel people’s behavior everywhere, allowing them to coordinate
their expectations and work together. The question therefore is how to use ‘culture’ in our
accounts without reverting to explanations that assume culture to be an immutable force
that overrules all other forces.
Following Read and Oselin (2008), we think that Sewell’s (1992) discussion of
resources (structure) and schemas (culture) offers a way how to do so. Swell seeks ways
to explain how material resources and virtual ideas reproduce each other. One cannot
decide which side is the ‘real world,’ and thus determine the other part as a mere derivative thereof. Rather they are two essential components determining each other. But in
order to be able to explain social and cultural change, Sewell offers ‘a far more multiple,
contingent, and fractured conception of society – and of structure’ (Sewell, 1992: 16), a
conception in which different, and even contradictory cultural schemas are open to interpretation and exist in overlapping structures. This multiplicity of structures and schemas
gives human agents a leeway in choosing and interpreting which cultural schema to follow (cf. Delanty, 2011; Lamont and Small, 2008; Swidler, 1986).
Read and Cohen’s (2007) analysis of the F-LFP of various ethnic groups in the US
mentioned above may serve as an illustration. The experience of immigration suggests
that individuals are subject to both the cultural schemas of the hosting society (often
numerous, overlapping, contradictory, and flexible) and to those carried with them from
their place of origin (again, potentially diverse and inconsistent); they have more ‘tool
kits’ (Swidler, 1986) to choose from. Indigenous minority groups have a similar leeway.
As citizens of a modern state they are subject to the norms distributed and sanctioned by
state agencies, they are influenced by cultural schemas of the majority, and they are also
subject to ideas and conceptualizations that originated in their own tradition. This article
posits that social heterogeneity of some Arab-Palestinian communities fosters such a
plurality of cultural images, thus changing the likelihood of women’s employment from
that found in purely Muslim or purely Druze communities. To better understand the postulated impact of social heterogeneity, we elaborate in the next section on some of the
less-known features of Arab-Palestinian society in Israel.
The Setting: Arab-Palestinians in Israel
Since the Arabs conquered the Middle East and North Africa in the 7th century, Muslims,
Christians, and Jews have lived side by side in Palestine (Sabella, 1996; Tibawi, 1969).
The Druze communities were established in the Galilee and Carmel mountains in the
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
5
Yonay et al.
16th century by Druze who came from Lebanon. They either established their own villages or joined villages in which Christians were present (Firro, 1999). Today, 17.3 percent of the Israeli citizenry are Arab-Palestinians.4 Among these Palestinians, the
Muslims, all Sunni, constitute the vast majority, 82.5 percent. The Christians, who themselves are divided into more than a dozen denominations (Tsimhoni, 2001), used to be a
fifth of all Palestinians in Israel in 1949 and still in 1961, but due to lower fertility and
emigration (Sabella, 1996; Tsimhoni, 2001) are now just 8.8 percent of all Arab citizens.
The Druze account for about a tenth (9.9%) of all Arab-Palestinians.
About 92 percent of the Arab-Palestinians in Israel live in exclusively Arab localities,
while the remaining reside in mixed Arab-Palestinian and Jewish cities. The bulk of
Arab-Palestinians in Israel, 57 percent, live in the Galilee (northern Israel) and most of
the rest are equally distributed among what is locally known as ‘The Triangle’ (along the
border between Israel and the northern part of the West Bank) and those who live in the
southern part of the country (Alnaqeb/the Negev) where the native Arab population is
Bedouin, a traditionally nomadic population that has undergone a long process of sedentarization in seven new towns, partly voluntarily and partly under pressure by the government (Falah, 1989).5
The Arab communities in Alnaqeb/Negev and the Triangle areas are homogeneously
Muslim, whereas in the Galilee many villages are religiously heterogeneous. The
Christians have different demographic patterns than the Muslims and Druze. They marry
at an older age, and they have fewer children (Lewin, 2012). The average education of
Christians is higher than that of Muslim and Druze, and consequently Christian women
are much more likely to be employed than other Palestinian women. According to the
1961 census, Christians already enjoyed a higher status and larger income than Muslims
at that time, and higher rates of women were employed; those findings have been corroborated by subsequent studies (Lewin-Epstein and Semyonov, 1992; Sa’ar, 1998;
Khattab, 2002; Kraus and Yonay, 2000). The Palestinian groups interact with each other
in all spheres of life but have remained separate, preserving their unique institutions and
organization. Since each group has been exposed to distinct economic and political conditions, their cultures have evolved along somewhat different trajectories. These cultural
differences are not necessarily the result of differences in religious doctrines and precepts. Hence, it should be noted that when we talk about Muslims, Christians, and Druze,
we refer to distinct cultural patterns and not to systems of religious faith.6
Traditionally, in the Middle East, Christians and Jews were under the influence of
Muslim societies and adopted similar cultural codes (Keddie, 1991; Lamdan, 2000;
Read, 2003; Read and Oselin, 2008: 310–1). In Palestine, however, cultural divergence
of Muslims and Christians might be traced back to differential opportunity structures due
to European intrusion into Ottoman Palestine in the 19th century. During that century,
new commercial cities developed along the coast following expanding trade with Europe
that attracted many Christian and Jewish merchants (Sabella, 1996; Schölch, 1982). The
activity of Christian churches in Palestine and the influx of Christian pilgrims also benefited Christian Palestinians. Church schools, still prominent within the Palestinian society to this day, were first built during the 19th century, and while serving the sons of
some of the Muslim notables as well, their location in Christian villages hampered the
attendance of Muslims from distant villages (Schölch, 1982; Tsimhoni, 2001).
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
6Sociology 
‘During the 30-year British mandate,’ the Christian Palestinians ‘experienced unprecedented security and prosperity’ (Tsimhoni, 2001: 3). ‘By 1920,’ Tsimhoni continues,
their majority ‘had become urbanized, … Christians formed a salient portion of the
Palestinian urban middle class; they had higher and westernized education, smaller
households, and a lower birthrate than the Muslims’ (Tsimhoni, 2001: 3). The socioeconomic advantage of Christian Palestinians over Muslim Palestinians following the
establishment of Israel might have been supported by two new factors. First, the
Christians were able to restore their community organizations faster due to the support of
European churches. This advantage facilitated Christians’ ‘ascendancy within Palestinian
politics in Israel’ (Pappe, 2006: 155–6). Second, the Israeli state favored Christians over
Muslims, presenting and treating the former ‘as more loyal to the Jewish State’ (Pappe,
2006: 155; Cohen, 2010).
The Druze religion is very different from both Islam and Christianity, but their ways
of life are closer to those of Muslims than to those of Christians. Druze and Muslims
share similar socio-economic characteristics, including family size, education, income,
and female labor force participation. Both groups also monitor the chastity of their
women and limit their freedom of movement much more strictly than Christians. We
therefore subsumed these two populations under the same grouping for the purpose of
this study. This is also a methodological necessity as explained below, but the similarities
in life experiences of these two groups makes this merging useful for our analytical
endeavor.
Since the establishment of Israel the state separated the Druze from the rest of
Arab-Palestinians whereby many of them reject the label ‘Palestinian’ (Hajjar, 2001).
The most noticeable feature that separates Druze from other Palestinians is the conscription of the former to and the exemption of the latter from military service (Firro,
1999). The intensified social contact between Druze and Jews due to military service
might have facilitated the penetration of Western culture into Druze communities. Yet
the wish to maintain group boundaries enables religious leaders to present traditional
codes of behavior as part of the group’s uniqueness, and heavy pressure is put on
members of the group to adhere to those codes, including women’s domesticity and
modesty. The fact that many Druze men work in the Israeli armed forces (army, police,
and prisons) discourages employment of their spouses due to the stability of their own
employment and its generous accompanying social benefits. As a result of all the
above factors, Druze women’s rates of employment as calculated by us from Labor
Force Surveys (not reported here) were slightly lower than those of Muslim women
until about 2000 and slightly higher since then. The statistical similarity that reflects
cultural resemblance enables us to treat the Muslims and the Druze as a unified category in our analysis.
This article exploits the divergent social patterns of Palestinian Christian, Muslim,
and Druze citizens of Israel to shed more light on the question of Muslim and Arab
‘culture’ and women’s participation in paid labor. It also exploits two opportune facts
that create almost an experimental design. First, many Christian Palestinians in Israel
reside in the same locality alongside Muslim and Druze Palestinians. In such mixed
localities, the members of these groups have lived side by side for hundreds of years,
and although maintaining the clear-cut boundaries that have kept them apart, they do
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
7
Yonay et al.
have long-standing extensive social ties. For our goal here the key point is that in some
localities there is no Christian presence, and the question is whether we will find a difference in Muslim and Druze F-LFP between those purely Muslim or Druze localities
and the heterogeneous ones.7 Second, Arab-Palestinians in Israel tend to maintain the
traditional pattern of patrilocality in their residence. This means that men build their
homes near their parents’ houses. Women who marry move to the houses built by their
husbands and thus live in proximity to the husbands’ parents, brothers and other relatives (Abou-Tabickh, 2010).8 For our purpose this pattern is crucial, because it means
that there is little self-selection in place of residence among Arab-Palestinians in Israel,
thus excluding the possibility that differences among various localities would be the
result of the attraction of certain types of people by certain types of communities.9 Our
hypothesis is thus as follows:
H: Other things being equal, the employment rate of Muslims and Druze women in
localities where Muslim and Druze Palestinians live side by side with Christian
Palestinians is higher than in communities where Christians are absent.
Data, Variables and Analytical Technique
Data and Methodology
Our aim was to compare the LFP of Muslim and Druze women in homogeneous communities with those living in heterogeneous communities. For this end we estimated the
net effect of residing in heterogeneous localities on Muslim and Druze F-LFP, adopting
a multilevel modeling technique (Raudenbush and Bryk, 2002). This technique enabled
us to estimate the effect of heterogeneous localities, net of other locality level characteristics as well as those of individual level characteristics. Specifically, the study is based
on two levels of analysis: Muslim and Druze women embedded in localities. Our data is
taken from Israeli Labor Force Surveys conducted annually by the Israeli Central Bureau
of Statistics (CBS). We pooled together the data from 2001 to 2006 to ensure a solid
number of observations even in small localities.
Our target population is comprised of Muslim and Druze women at the main employment age (18 to 45 years). In the pooled survey, this target population included 20,886
female respondents, but for several methodological and substantive reasons a few groups
were excluded. First, we excluded Palestinian women who lived in mostly Jewish cities
where both employment opportunities and cultural atmosphere are very different from
those common in purely Palestinian localities. There were only eight such cities and we
thought that their inclusion would obscure the impact of the factor that we are studying
in this project. Second, Bedouin women residing in ‘unrecognized villages’ are not
included in the labor force surveys, and are thus left out of our study. Third, women from
the Israeli-occupied Golan Heights, almost all of whom are Druze, were also omitted
because the Golan residents consider themselves as Syrians and their ties with Israeli
institutions and Israeli culture are much weaker than those of the Israeli Druze within the
pre-1967 borders. Finally, respondents residing in localities with a pooled sample size of
fewer than 150 women were excluded, as well as the very few women who lived in
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
8Sociology 
institutions. We thus ended up with 16,046 Muslim or Druze women, aged 18–44 years,
residing in 59 localities. As Appendix 1 indicates, the study population represents 76
percent of our target population, and is fairly similar in key characteristics.
Variables. The dependent variable, labor force participation (LFP), is measured by the
standard question on employment activity in the week preceding the survey. Women
employed in that week, those on short leave or illness, as well as those seeking employment were considered part of the labor force and coded ‘1’ (otherwise ‘0’).
Independent variables at the individual level include: age (in years), age square;
post-secondary degree (coded ‘1’ if one holds a degree and ‘0’otherwise), marital status
(coded ‘1’ if married and ‘0’ otherwise); young children (coded ‘1’ if a woman has at
least one child below age ten and ‘0’ otherwise); household density (number of rooms
in the household divided by household size); and a set of dummies to control for year of
survey. With regard to education, we refer only to post-secondary education because
this level was found crucial in determining employment prospects. Yonay and Kraus
(2009) for example, found that F-LFP jumped from 17 percent among Muslim women
with 9–12 years of education to 66 percent among those with at least 13 years (2009:
246). We refer to children below 10 based on our calculations (not presented here)
which show that having young children is what limits F-LFP regardless of their
number.
At the locality level, we employed three variables to characterize the social, demographic and economic makeup of each community. The variables are calculated by
aggregating the data on individuals within a locality. The first variable is our key variable
in this study, namely, the religious heterogeneity of a given locality. Since the majority
of the 59 communities in our data are entirely Muslim (40 localities), treating this variable as a continuum would be misleading. Hence we used a dummy variable coded ‘1’ if
the community was heterogeneous and ‘0’ otherwise.10
Since many studies have already documented the positive relations between economic
opportunities and F-LFP (e.g. Goldin, 1990; Huber, 1991), we tested several indicators
of economic activity. We first took the size of the local labor force as a measure of total
economic activity, and then checked also the relative size of public sector employment in
each locality where women are much more likely to be employed (we tried both total
employment in the public sector and female employment).11 Since these variables had no
impact we used two other locality-level variables as control variables. The first is a
dummy variable for the community size, which serves as an indicator of the opportunity
structure of the community. The variable was coded ‘1’ if a community had more than
10,000 inhabitants (at the middle of the pooled period) and ‘0’ otherwise. The other control variable is the locality’s economic condition. We wanted to use locality’s socioeconomic index (SEI) reported by the CBS (2006), but such an index is available only
for communities with more than 10,000 inhabitants. We therefore replicated this index
ourselves for all the 59 localities in our study. Comparing our measure with that of the
CBS for the big localities (those above 10,000 inhabitants) shows that the indices are
quite similar (R2 = 0.92). The CBS index that we replicated is based on three variables:
household density (average number of rooms per person), dependency rate (number of
individuals below the age of 18 or above 64 relative to the number of those 18–64 years
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
9
Yonay et al.
_
Figure 1. Muslim and Druze Women LFP rates by locality (X = 0.24).
old), and educational level (the proportion of individuals in the locality aged 21–65 who
hold an academic certificate). We then used the parameter estimate from the regression
model predicting the SEI index for these large localities in order to calculate the predicted values of the SEI index in our full set of 59 localities.12 Appendix 2 presents the
descriptive statistics of these variables.
Findings
We begin with a description of cross-community variation in Muslim and Druze F-LFP.
As can be seen in Figure 1, substantial cross-community variation exists, ranging from
11 to 44 percent. The aim of this analysis is to explain this variation. Table 1 reveals that
this cross-community variation is related to locality size, to locality SEI, and also to the
locality religion composition, as indicated by the statistically significant associations
between Muslim and Druze F-LFP and each of the community level characteristics.
In accordance with previous studies, Muslim and Druze women tend to work more
outside their homes in more socio-economically advanced localities. A standard deviation change in the SEI is translated into half a standard deviation change in the F-LFP
rate. Put differently, as Table 1 reveals, such an increase in the locality’s SEI would raise
the Muslim and Druze F-LFP by almost 5 percentage points. Figure 2 illustrates this
relationship. Surprisingly, Table 1 indicates that Muslim and Druze women in smaller
rather than larger localities are more likely to engage in paid employment. Whereas only
about a quarter of the Muslim and Druze women in small localities engage in paid
employment, this number is even smaller in large localities (22%). This relation is related
to in the multivariate analysis.
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
10Sociology 
Table 1. Bivariate relationships between locality LFP rates and the three locality-level
variables.
LFP
LFP
(r)
(etha2)
Christian presence
Large (over 10,000)
Socio-economic index
0.128*
−0.186*
0.492*
0.016
0.035
–.–
Size
Religious composition
Small
Large
W/ Christians
No Christians
LFP Rate
0.248
(0.001)
0.219**
(0.001)
0.241
(0.001)
0.222**
(0.001)
Note: * statistically significant correlations; ** statistically significant differences.
0.5
R² = 0.242
0.45
0.4
L 0.35
F
P 0.3
0.25
R
a 0.2
t
0.15
e
0.1
0.05
0
-2.5
-2
-1
-1.5
-0.5
0
Locality SEI
Figure 2. The relationship between locality SEI and Muslim and Druze women LFP rates.
Concerning our hypothesis, as expected, Muslim and Druze women are more likely to
engage in paid employment when they live in communities composed of Christians as
well (see Table 1). The focus of the multivariate analysis is whether or not this pattern
holds even after individual compositional differences as well as the locality characteristics are controlled for.
Multivariate analysis. This stage of the analysis focuses on the effect of locality characteristics on Muslim and Druze F-LFP rates net of individual level characteristics. Table 2
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
11
Yonay et al.
Table 2. Estimated coefficients from multi-level logistic regression models predicting Muslim
and Druze women LFP.
Model 1
Age
Age2
Density
Children <10 at home
Marital status (married)
Education (post-sec. or
academic degree) Locality SEI Large locality (over 10k)
Religion mixed locality
Mix × large locality
Constant
Random effect
Constant
N
AIC
Model 2
0.476***
0.475***
(0.029)
−0.007***
(0.000)
−0.256***
(0.036)
−0.406***
(0.077)
−1.152***
(0.081)
2.822***
(0.061)
(0.029)
−0.007***
(0.000)
−0.256***
(0.036)
−0.405***
(0.077)
−1.151***
(0.081)
2.825***
(0.061)
−0.032
(0.166)
−0.223
(0.117)
0.259*
(0.126)
−8.141***
(0.439)
−8.123***
(0.475)
0.183***
(0.041)
16,046
13,309
0.159***
(0.037)
16,046
13,309
Model 3
0.475***
(0.029)
−0.007***
(0.000)
−0.256***
(0.036)
−0.403***
(0.077)
−1.153***
(0.081)
2.824***
(0.061)
−0.074
(0.163) −0.063
(0.138)
0.542**
(0.186)
−0.495*
(0.245)
−8.239***
(0.477)
0.148***
(0.034)
16,046
13,307
Standard errors in parentheses.
*p < .05; ** p < .01; *** p < .001.
Source: LFS 2001–2006.
presents three models predicting F-LFP: the first model controls only for individual
level characteristics, while the intercept (β0j) is a random effect with error terms (i.e.,
without predictors). This model serves as a benchmark against which we can assess the
amount of variance explained by the inclusion of macro-level characteristics in subsequent models. In the second model we add the three locality-level variables to explain
the variation in this random effect. The third model represents our preferred model in
predicting Muslim and Druze F-LFP after all possible interactions between the locality
characteristics are examined (see Appendix 3 for more formal tests of alternative
models).
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
12Sociology 
Figure 3. Predicted probability of Muslim and Druze women LFP rates, by locality size, and
composition of religions as predicted from Model 3 in Table 2.
Model 1 indicates, as expected, that the effects of individual level characteristics are
in line with findings in prior studies on F-LFP (Read and Cohen, 2007). For example, the
presence of young children at home reduces women’s likelihood to engage in paid
employment. Likewise, education is positively associated with F-LFP. In this model,
however, our main interest is the random effect for the intercept, presented at the foot of
the table. This figure simply indicates that the level of F-LFP varies significantly statistically across the 59 localities in the study. Our aim in Models 2 and 3 is to relate this variation to locality characteristics.
In Model 2, then, we add the three contextual characteristics and find that contrary to
the bivariate analysis, the locality SEI does not have a significant effect on F-LFP, while
the impact of the composition of the religions in the locality remains statistically significant. That is, our hypothesis, that in communities heterogeneous with respect to religion
Muslim and Druze women are more likely to be involved in the formal labor force, is
supported.13 Exploring potential interaction effects in the data between the contextual
effects, Model 3 indicates that the effect of Christian presence on Muslim and Druze
women’s employment is much more pronounced in small rather than large communities.
This is illustrated in Figure 3 and can be shown by examining the parameter estimates in
Model 3. Such an examination reveals that most of the advantage gained by the presence
of Christian residents in the locality (γ = 0.542) in fact vanishes in large localities (γ =
−0.495). A possible explanation for this result is that in large communities, the group
within each religion is more self-sufficient and most people maintain ties with people of
the same faith, whereas in small communities people need to go out of their faith to provide their needs.14
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
13
Yonay et al.
Discussion and Conclusions
Our findings corroborate earlier studies that show that education increases Muslim and
Druze F-LFP while having young children lowers those rates. This finding is in line with
Read and Cohen’s claim (2007) that all women are susceptible to the same family and
human capital factors, a conclusion that negates the view that some societies are completely different from others. Yet, the fact that the same factors affect all groups of women
does not mean that there are no substantial gaps between them (Read and Cohen, 2007);
communities do differ from each other in their ‘base lines.’ In multi-religion Palestinian
villages and towns that also include Christians there are higher rates of participation in
the formal labor market among Muslim and Druze women.
What does this finding mean? What can we learn from it regarding the more general
question of cultural change, and of gender relations in particular? Since we control as
much as possible for personal characteristics and the economic conditions of the community, our results suggest that Muslims and Druze behave differently in communities
composed of Christians as well. Why? It is impossible to give a definite answer but we
think that a reasonable explanation is the diffusion of cultural norms and ideas from one
group to another. This view is in line with more recent sociological perspectives that
acknowledge the importance of cultural schemas as the necessary link between social
structures and the behavior of individuals (Sewell, 1992; see also Hechter and Kanazawa,
1997; Lamont and Small, 2008). Ideas about gender relations in the entire MENA region
have changed considerably with new economic and political circumstances (Beck and
Keddie, 1978; Joseph and Slyomovich, 2001; Read and Oselin, 2008; Spierings et al.,
2010). The question then is not to adjudicate whether the causes for the relatively low
F-LFP among Muslims and Arabs in the MENA region are ‘cultural,’ ‘economic,’ or
‘political,’ but to identify the factors that affect the tempo of the change of the whole
social complex.
Muslim and Druze women even in the remotest villages in Israel might be influenced
by ideas and values developed in modern state agencies, schools and colleges, women
organizations, political parties, Israeli and foreign mass media, and places of work. These
institutions may pull in different directions, but actual behavior depends on specific configurations of social networks and the resources of relevant actors. Christian communities in the same village or town might be perceived from the viewpoint of Muslim and
Druze women as part of this network that might have some impact on the final valuation
of female employment (cf. Lewin-Epstein and Semyonov, 1992: 49–50).
For various historical reasons, Christian Arab-Palestinians have developed differently
than Muslims and Druze, and among other things, have legitimized women employment
earlier and faster (see discussion above). Although maintaining the clear-cut boundaries
of religion that keep the groups apart, Muslims, Druze, and Christians do have longstanding extensive ties including social, political, and economic cooperation. Our study
is based on statistical data, and is not based on direct observation. Hence it is only an
interpretation based on our findings and on the negation of other plausible explanations.
It is also based on our knowledge of Arab-Palestinian communities from the literature.
Many Muslim, Druze, and Christian children go to the same schools, including church
schools where most Christian parents prefer their children to go, but have become quite
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
14Sociology 
popular among Muslim and Druze parents as well (Badran, 2012). Such intense social
interaction creates the potential for mutual influences that may explain the fact that
Muslims and Druze living with Christians in the same locality behave more similarly to
Christians than their brethren living in purely Muslim or Druze communities. The fact
that this effect is found only in small communities supports this interpretation, because
in such communities the spatial proximity between the distinct religion groups and the
lack of services provided by co-religion groups increases the likelihood of people interacting with those belonging to other religions.
We started our theoretical review with Read and Cohen’s question (2007), why people
from various ethnic origins respond differently to the same circumstances. They suggested that we introduce ‘culture’ to models of women’s employment, but the question is
how to do so in a way that ‘moves beyond the assumption that racial groups have inherent cultural traits’ (Lamont and Small, 2008: 76–7). In our study we compared two identical groups that differ only by the presence of another social group in the vicinity. The
two are part of the same Muslim Palestinian society in Israel: they have the same cultural, political, and social institutions, and yet there is a statistically significant difference
between them even after controlling for individual and community characteristics.
Of course, the experience of Palestinians in Israel is quite unique, and there are very
few communities in the Middle East where people from different religions live side by
side in small communities. Yet our case may inform much more common situations in
which distinct social groups with divergent attitudes on gender relations interact on a
daily basis. Our story is not about the triumph of ‘modernity’ over ‘tradition’ – an interpretation that assumes two polar stationary cultures (Gusfield, 1967) – but rather on the
outcomes of interaction among distinct groups, interaction that changes the balance of
conflicting values and concepts. It is a gradual change as part of adaptations that are
omnipresent in all societies and not a ‘clash of civilization.’
Funding
This study was supported by Israel Science Foundation, grant no. 637/06 to the first and third
authors.
Notes
1. As will be further elaborated below, the religious groups of Palestinians in Israel constitute
separate social groups that maintain distinct social patterns that are not reducible to strictly
religious beliefs and habits. See section: The Setting: Arab-Palestinians in Israel.
2. The data is calculated online at ILO’s Key Indicators of the Labor Market, http://kilm.ilo.org/
kilmnet/, for 25–54-year-old women.
3. Huntington’s (1996) clash-of-civilization thesis presumes irreconcilable differences between
Western and Muslim cultures. This thesis has been harshly criticized by numerous scholars
(e.g. Halliday, 2003; Rashid, 2003).
4. The numbers in this paragraph are based on our calculations of data from Tables 2.6 and 2.11
of the 2010 Statistical Abstract of Israel published online by the Israeli Central Bureau of
Statistics (http://www.cbs.gov.il/reader/shnatonhnew_site.htm). We took the number of all
Jews, including settlers in the Occupied Palestinian Territories, but excluded Palestinians who
reside there.
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
15
Yonay et al.
5. The Bedouins are all Muslim, but socio-historical distinctions keep them and other Muslims
apart (see Falah, 1989).
6. Within each group there are deep cultural cleavages based, inter alia, on region, type of locality, and class. Here, however, we focus on the average behavior of each group after controlling for individual traits.
7. A hint of this thesis was offered more than two decades ago by Lewin-Epstein and Semyonov
(1992).
8. The endurance of this pattern in modern Israel is exceptional in the Middle East, where widespread migration to big urban centers has been the rule (Joseph and Slyomovics, 2001: 3). It
might be attributed to the small size of Israel and to pervasive discrimination against Arab
tenants.
9. Some Arab-Palestinians, probably those who care more about personal freedom, move from
rural communities to big mixed cities, yet the small proportion of Arab-Palestinians who live
in those cities means that such migration is marginal.
10. Due to the small number of heterogeneous communities we treat localities with modest
Christian presence and those in which Christians constitute a major bulk of the population
equally.
11. Since most men in Arab-Palestinian communities work outside the locality, usually in big
urban areas, male LFP rates do not measure economic activity in the locality itself. We also
included in our models locality measures of voting and religiosity behaviors. None of these
community characteristics had a statistically significant effect. Results can be provided on
request.
12. Locality SEI = 3.845–0.027 (dependency rate) − 0.935 (household density) + 3.364 (p with
degree).
13. The non-significant contextual effects in Model 2 are not the results of multicolliniarity
between these variables, as Appendix 4 indicates.
14. An example is the large town of Shefa-’Amr (Shefar’am) where there are three large groups
of Muslims, Christians, and Druze. The size of the city, over 35,000, has necessitated several
public schools and attracted several church schools. Since neighborhoods are quite segregated
by faith, schools have become more homogeneous, and less cultural diffusion takes place
(Badran, 2012).
References
Abu-Lughod L (2002) Do Muslim women really need saving? Anthropological reflections on
cultural relativism and its others. American Anthropologist 104(3): 783–90.
Abou-Tabickh L (2010) Women’s masked migration: Palestinian women explain their move upon
marriage. In: Kanaaneh RA and Nusir I (eds) Palestinians in Israel Revisited. Albany: SUNY
Press, 189–205.
Ahmed L (1992) Women and Gender in Islam: Historical Roots of a Modern Debate. New Haven,
CT: Yale University Press.
Anwar Z (ed) (2009) Wanted: Equality and Justice in the Muslim Family. Selangor: Musawah,
Sisters in Islam.
Badran A (2012) Elite track passes through the barrier of religion. Unpublished MA thesis,
University of Haifa, Haifa.
Beck L and Keddie NR (eds) (1978) Women in the Muslim World. Cambridge, MA: Harvard
University Press.
CBS (2006) Characterization and Classification of Local Authorities by the Socio-Economic Level
of the Population, 2006. Jerusalem: Central Bureau of Statistics (Publication No. 1401).
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
16Sociology 
Charrad MM (2001) States and Women’s Rights: The Making of Postcolonial Tunisia, Algeria,
and Morocco. Berkeley, CA: University of California Press.
Clark JA (2003) Islam, Charity, and Activism: Middle-Class Networks and Social Welfare in
Egypt, Jordan, and Yemen. Bloomington, IN: Indiana University Press.
Clark JA and Schwedler J (2003) Who opened the window? Women’s activism in Islamist parties.
Comparatie Politics 35(3): 293–312.
Clark JA and Young AE (2008) Islamism and family law reform in Morocco and Jordan.
Mediterranean Politics 13(3): 333–52.
Clark R, Ramsey TW and Adler ES (1991) Culture, gender, and labor force participation: A cross
national study. Gender and Society 5(1): 47–66.
Cohen H (2010) Good Arabs: The Israeli Security Agencies and the Israeli Arabs, 1948–1967.
Translated by Haim Watzman. Berkeley, CA: University of California Press.
Delanty G (2011) Cultural diversity, democracy and the prospects of cosmopolitanism: A theory
of cultural encounters. British Journal of Sociology 64(2): 633–56.
Falah G (1989) Israeli state policy toward Bedouin sedentarization in the Negev. Journal of
Palestine Studies 18(2): 71–91.
Firro KM (1999) The Druzes in the Jewish State: A Brief History. Leiden: Brill.
Goldin C (1990) Understanding the Gender Gap: An Economic History of American Women.
Oxford: Oxford University Press.
Gusfield JR (1967) Tradition and modernity: Misplaced polarities in the study of social change.
American Journal of Sociology 72(4): 351–62.
Haddad YY and Esposito JL (eds) (1998) Islam, Gender and Social Change. New York: Oxford
University Press.
Haddad YY and Stowasser BF (eds) (2004) Islamic Law and the Challenges of Modernity. Walnut
Creek, CA: Altamira.
Hajjar L (2001) Speaking the conflict, or how the Druze became bilingual: A study of Druze
translators in the Israeli military courts in the West Bank and Gaza. Ethnic and Racial Studies
23(2): 299–328.
Halliday F (2003) Islam and the Myth of Confrontation: Religion and Politics in the Middle East.
London: I. B. Tauris.
Harrison LE and Huntington SP (eds) (2000) Culture Matters: How Values Shape Human
Progress. New York: Basic Books.
Hechter M and Kanazawa S (1997) Sociological rational choice theory. Annual Review of
Sociology 23: 191–214.
Huber J (1991) A theory of family, economy, and gender. In: Blumberg RL (ed) Gender, Family
and Economy. Newbury Park, CA: Sage, 35–51.
Huntington S (1996) Clash of Civilization and the Remaking of World Order. New York: Simon
and Schuster.
Huntington S (2000) Foreword. In: Harrison LE and Huntington SP (eds) Culture Matters: How
Values Shape Human Progress. New York: Basic Books, xiii–xvi.
Jacobsen J (2007) The Economics of Gender. 3rd ed. Malden, MA: Blackwell.
Joseph S and Slyomovics S (eds) (2001) Women and Power in the Middle East. Philadelphia, PA:
University of Pennsylvania Press.
Kandiyoti D (1991) Introduction. In: Kandiyoti D (ed) Women, Islam and the State. Houndmills:
Macmillan, 1–21.
Karshenas M (2001) Economic liberalization, competitiveness, and women’s employment in the
Middle East and North Africa. In: Salehi-Ishfahani D (ed) Labor and Human Capital in the
Middle East. Reading: Ithaca, 147–92.
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
17
Yonay et al.
Keddie NR (1991) Introduction: Deciphering Middle Eastern women’s history. In: Keddie NR and
Baron B (eds) Women in Middle Eastern History: Shifting Boundaries in Sex and Gender.
New Haven, CT: Yale University Press, 1–22.
Khattab N (2002) Ethnicity and female labour market participation: A new look at the Palestinian
enclave in Israel. Work, Employment and Society 16(1): 91–110.
Kraus V and Yonay YP (2000) The power and limits of ethnocentrism: Palestinians and Eastern
Jews in Israel, 1974–1991. British Journal of Sociology 51(3): 525–51.
Lamdan R (2000) A Separate People: Jewish Women in Palestine, Syria, and Egypt in the Sixteenth
Century. Boston, MA: Brill.
Lamont M and Small ML (2008) How culture matters: Enriching our understanding of poverty.
In: Lin AC and Harris DR (eds) The Colors of Poverty: Why Racial and Ethnic Disparities
Persist. New York: Russell Sage, 76–102.
Lewin AC (2012) Marriage patterns among Palestinians in Israel. European Journal of Population
28(3): 359–80.
Lewin-Epstein N and Semyonov M (1992) Modernization and subordination: Arab women in the
Israeli labor force. European Sociological Review 8(1): 38–51.
Mernissi F (1993) Islam and Democracy: Fear of the Modern World. Reading, MA: Addison-Wesley.
Minces J (1982) The House of Obedience: Women in the Arab Society. London: Zed.
Moghadam VM (1998) Women, Work, and Economic Reform in the Middle East and North Africa.
Boulder, CO: Lynne Rienner.
Moghadam VM (2003) Modernizing Women: Gender and Social Change in the Middle East. 2nd
ed. Boulder, CO: Lynne Rienner.
Moghadam VM (2004) Enhancing women’s participation in the MENA region. In: Handoussa H
and Tzannatos Z (eds) Employment Creation and Social Protection in the Middle East and
North Africa. Cairo: The American University in Cairo Press, 237–80.
Norris P and Inglehart R (2004) Sacred and Secular: Religion and Politics Worldwide. Cambridge:
Cambridge University Press.
Pappe I (2006) A History of Modern Palestine: One Land, Two Peoples. New York: Cambridge
University Press.
Rashid S (ed) (2003) “The Clash of Civilizations?” Asian Responses. Dhaka: University Press.
Raudenbush SW and Bryk AS (2002) Hierarchical Linear Models in Social and Behavioral
Research: Applications and Data Analysis Methods. 2nd ed. Thousand Oaks, CA: Sage.
Read JG (2003) The sources of gender role attitudes among Christian and Muslim Arab-American
women. Sociology of Religion 64(2): 207–22.
Read JG and Cohen PN (2007) One size fits all? Explaining US-born and immigrant women’s
employment across 12 ethnic groups. Social Forces 85(4): 1713–34.
Read JG and Oselin S (2008) Gender and the education-employment paradox in ethnic and religious contexts: The case of Arab Americans. American Sociological Review 73(2): 296–313.
Sa’ar A (1998) Carefully on the margins: Christian Palestinians in Haifa between nation and state.
American Ethnologist 25(2): 215–39.
Sabella B (1996) Palestinian Christians: Challenges and hopes. Available at: www.al-bushra.org/
holyland/sabella.htm (accessed 12 August 2014).
Said EW (1978) Orientalism. New York: Pantheon.
Salman M (1987) The Arab woman. In: Salman M, Kazi N, Yuval-Davis N, Al-Hamdani L,
Botman S and Lerman D (eds) Women in the Middle East. London: Zed, 6–11.
Schölch A (1982) European penetration and the economic development of Palestine, 1856–82. In:
Owen R (ed) Studies in the Economic and Social History of Palestine in the Nineteenth and
Twentieth Century. London and Oxford: Macmillan/St. Anthony’s College, 10–87.
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
18Sociology 
Sewell WS, Jr (1992) A theory of structure: Duality, agency, and transformation. American
Journal of Sociology 98(1): 1–29.
Shukri SJA (1999) Social Changes and Women in the Middle East: State Policy, Education,
Economics and Development. Aldershot: Ashgate.
Sonbol AE (2010) A response to Muslim countries’ reservations against full implementation of
CEDAW. HAWWA( 8): 348–67.
Spierings N, Smits J and Verloo M (2010) Micro and macro-level determinants of women’s
employment in six Arab countries. Journal of Marriage and Family 72(5): 1391–407.
Swidler A (1986) Culture in action: Symbols and strategies. American Sociological Review 51(2):
273–86.
Tibawi AL (1969) A Modern History of Syria including Lebanon and Palestine. London:
Macmillan.
Tsimhoni D (2001) Israel and the territories – disappearance: Disappearing Christians of the
Middle East. The Middle East Quarterly 8(1): 31–42.
Yonay YP and Kraus V (2009) Culture or opportunity structure: Why so few Palestinian women
join paid labor? In: Azaiza F, Abu-Baker K, Hertz-Lazarowitz R and Ghanem A (eds) Arab
Women in Israel: Current Status, Future Trends. Tel-Aviv: Ramot, 237–58. (in Hebrew.)
Yuval P Yonay is a senior lecturer at the University of Haifa. He has written on the history and sociology of economic knowledge, the socio-economic status of Israeli Palestinians, and on the legal
and political standing of lesbians and gays in Israel. In the last few years, he has been involved in
two research projects on Palestinian women in the labor force and on the gay history of Israel.
Meir Yaish is an associate Professor and the Chair of the Department of Sociology and Anthropology
at the University of Haifa. His research interests lie with social stratification and mobility, sociology of education, and the puzzle of altruism. Since 2010 he serves as the Editor-in-Chief of the
journal Research in Social Stratification and Mobility.
Vered Kraus is professor of Sociology at the department of Sociology and Anthropology at the
University of Haifa. Her interests and research activity have been centered upon issues of social
inequality with a special focus on the women in the labor market.
Date submitted September 2013
Date accepted July 2014
Appendix 1: Selected characteristics of the target and study populations.
Labor force participation
Years of schooling
Percent living in localities
with less than 20,000
Household size
Target population
Study population
26.0
10.8
65.0
23.0
10.4
67.0
5.9
6.2
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015
19
Yonay et al.
Appendix 2: Descriptive statistics of selected variables.
Variables
Proportion/Mean (SD)
Individual level (N = 16,046)
Labor force participation
Age
Household density
Proportion with children under 10
Proportion married
Proportion with post-secondary Degree
Locality level (N = 59)
Labor force participation
SEI
Proportion of large localities
Proportion with Christian presence
Range
0.23 (0.42)
30.12 (7.62)
1.66 (0.79)
0.57 (0.49)
0.66 (0.47)
0.13 (0.34)
0–1
18–44
0.2–21
0–1
0–1
0–1
0.11–0.44
−0.32– −2.11
0–1
0–1
0.24 (0.08)
−0.99 (0.40)
0.49 (0.50)
0.32 (0.47)
Appendix 3: Type II Chi-square tests for various interaction terms.
Interaction terms
Chi-square
df
p value
SEI × Mix
SEI × Size
Mix × Size
0.06
0.20
3.99
1
1
1
0.80
0.65
0.05
Note: Tests are vis-à-vis Model 2 in Table 2.
Appendix 4: Association measures between locality-level variables.
Mixed
Mixed
Large (over 10,000)
Socio-economic index
0.131*
0.193*
Large
Mean
0.032*
0.33
0.70
−0.99
Note: Locality is unit of analysis. * indicates that the relevant measure of association is statistically significant
at the 0.05 level.
Downloaded from soc.sagepub.com at University of Haifa Library on March 9, 2015