Attitudes towards women's rights to inheritance in District Lakki Marwat, Pakistan.

Ahmad, Eatzaz ; Bibi, Anbereen ; Mahmood, Tahir 等

This paper analyses the effects of various economic, demographic and social factors like marriage practices, education and awareness on the beliefs and attitudes regarding women's rights to inheritance, ownership and management of property in Pakistan. Based on a random sample of 507 families from district Lakki Marwat, a backward northern district of Pakistan, the study finds that education level of respondents, including religious education, awareness of law and formal sources of information regarding women's rights are the main factors shaping respondents' perceptions about women's right to inheritance, ownership and management of properly. The other factors like economic status, family demography and marriage practices appear relatively less important.

1. INTRODUCTION

Transfer of wealth from old to young generations through inheritance provides an important source of wealth to the young and the middle-aged. These transfers often constitute a major source of earning and old-age security in the form of home ownership in poor countries where formal social security networks at national level are almost nonexistent and capital markets are imperfect. Inheritance also provides a potential source of investment in education for children who become orphans at an early age. Even if it is assumed that parents accumulate wealth only for their own old-age consumption, as the theory of life cycle claims, and have no particular desire to leave something to their children, the latter will probably still receive an inheritance.

Different societies have evolved their own social norms and practices in transferring wealth from old to young generations. In certain societies inheritance is a simple matter of transferring legal possessions of deceased persons to their descendants. According to traditional tribal customs in Pakistan, sons tend to enjoy the right of receiving almost all the assets left by their parents, while women generally do not receive or are obliged to surrender their legal share in inheritance. Being daughters, women are often also expected to forego their rights to inheritance in favour of their brothers. Their claim to get their inherited share in property may result in their desertion by their parental side. All this is mostly done and justified in the name of preserving inherited land, an important source of income and the symbol of power. Dowry is often treated as the daughters' share in inheritance which obviously, in most cases, is not equivalent to legal inheritance, but customary practices are binding on women to acquiesce. Moreover, women are told that their brothers are going to take care of them and the gifts given to them on different occasions are considered their share in property.

All these practices are quite contrary to the prevalent state law and clear guidelines given in Islam that prescribe well-defined shares for the male and female descendants of a deceased person. The Quran clearly states: "Men shall have a share in what parents and kinsfolk leave behind, and women shall have a share in what parents and kinsfolk leave behind." (Quran, 4:7). Women in Pakistan do have the legal right to inherit family's wealth, yet they rarely exercise their rights. Although Islamic law (Shariah) and Pakistan's state law both entitle women to inherit immovable and movable property, the practice has been to deny women their share in inheritance, particularly if it is land in which case their entire claim would often be denied. [Mehdi (2002)]. This is especially true in rural Pakistan, where the tribal nature of social organisation undermines inheritance rights.

Based on Islamic law, state law stipulates the share of women's inheritance to be one-half of that of men in similar relationships to them (e.g., a daughter would inherit one share for every two shares that a son inherits) due to the man's greater responsibility for supporting the family. Inheritance is governed by Islamic Shariah as codified in the Family Laws Ordinance 1961 [Mumtaz (2006)]. Due to the powerful force of customary practice in the country, however, the inheritance rules most often followed are those based on custom. Women in rural areas in particular rarely receive their shares in immovable property. When women do inherit property, jt is typically controlled by male heirs [Mehdi (2002)].

The present study is carried out to analyse a number of potential factors that can be associated with the attitudes towards inheritance, property ownership and property management by women in Lakki Marwat, one of the less developed districts of Pakistan located in the province of Khyber Pakhtimkhwa (KPK). The study is based on a field survey containing both quantitative and qualitative data, which can explain the relationship of women's rights to inheritance and control over property with various socioeconomic and demographic variables.

The study considers four variables representing society's attitude towards recognising and accepting women's rights to inheritance, ownership and management of their properties. These are based on the respondent's intentions to give daughters their legal shares in inheritance and their views on whether women should be given their legal shares in inheritance; whether women should keep their property in their own names and whether women should have authority to manage their property.

The potential correlates of the attitudes towards women's rights to inheritance are the family's economic status, size and gender composition, registration of births, marriage type, pre-marriage relationship with spouse, education, religious orientation, awareness about women's rights to inheritance and the source of information about women's rights. These correlates are expected to have direct or indirect influence over the society's attitudes towards women's right to inheritance.

The study is organised as follows: Section 2 provides a brief review of the status of women regarding ownership and management of property. Section 3 provides the analytical framework and construction of variables. Data and brief descriptive analysis are presented in Sections 4. Econometric analysis of the relationship of each of the four variables describing society's attitude towards recognising and accepting women's rights to inheritance, ownership and management of property is carried out in Section 5. Finally, Section 6 concludes the study.

2. STATUS OF WOMEN IN PROPERTY OWNERSHIP AND MANAGEMENT

In Pakistan, especially in rural areas, land as an asset is considered as the basic physical resource to provide food, employment, living space, economic security and social status. Less than half of rural households own agricultural land and 40 percent of the land is owned by 2.5 percent of households. In rural Sindh, landlessness is most acute with two-thirds of rural households not owning any land and just 0.4 percent of households accounting for nearly 24 percent of the total land area [Mumtaz (2006)].

According to Kamal (1999), male relatives tend to have actual control of property even where women own property. Often women can dispose of their property with household consent only and all buying and selling is done by male members of households. Statutary and Islamic laws, however, provide that a woman has the right to acquire, hold and dispose of property. The study mentions that another widespread practice is of forfeiting of inheritance share by women in favour of their brothers or sons, often through force or social pressures. In the absence of protective measures, women generally rely on the parental home and brothers in times of need (bad marriage, illness, economic pressure, etc.) and, therefore, forgo their share in property as insurance for the future.

Mehdi (2002) reports several cases in which women's names are registered in the property transfer papers on inheritance, but in practice, they do not get it. The understanding is that brothers would take the land, and the sisters would receive shares of the harvest or gifts. In rural Punjab, it is also not uncommon for brothers to enter into an understanding with the patwari (land revenue official) not to include their sisters' names among owners under the pretext that they, the sisters, have surrendered their shares. Otherwise, as soon as a woman's name is entered in the transfer papers, a gift is made in favour of the brothers. The study also mentions the extreme practice of haq-bakhshwana (explicit giving up of rights) whereby girls are either never married, or are 'married to the Quran' as in southern parts of Punjab (Multan and Bahawalpur) and Sindh in order to prevent property going out of the family. Similarly, cousin marriages and exchange marriages whereby one set of brother and sister are married to another set are designed to prevent break up of property as the size of land and property is associated with power and status.

One explanation for such discriminatory practices often cited is the adherence to age-old traditions and cultures. Women are said to be lacking in information about legal, economic or political rights. They are considered vulnerable to violence due to absence or lack of access to protection and justice. They are also considered to have restricted mobility. An important barrier for women is their lack of knowledge about their property rights and limited understanding of land registration systems, transaction procedures and other legal matters involved in possession of land.

Tirmazi (1999) regards women's lack of mobility as a barrier to their access to property and freedom to manage the property if they happen to own it. The excuse often made, especially in South Punjab, is that women are not supposed to leave the home without purdah (seclusion) to physically own and manage property. This is despite the observation that women are expected to leave the home to fetch water from far flung areas for home consumption. Tirmazi (1999) asserts that women are relegated to work as service providers, but they are not considered responsible enough to own and control property and make important decisions related to it. This behaviour is not confined to Pakistan. Bennett (1981) has found that only less than two percent of all titled land around the world is owned by women, which shows worldwide gender-based inequity practices that reflects on the women's state of economic wellbeing. In its extreme form this type of discrimination can restrict the women's ability to acquire even the minimum means of survival.

Owning property provides a hedge against risk for women. According to Agarwal (1994) in South Asia the bargaining position of women is significantly affected by the amount of assets they own. This means that in seeking gender equality and justice the right to inheritance can be used as an effective tool. Agarwal (1998) finds that in rural India women tend to be economically and socially vulnerable and have limited range of choices for livelihood because of their limited access to own and manage property. Deere and Leon (2001) observe that property ownership lends women numerous advantages, especially in late years of life.

Some cultures consider dowry as a substitute for inheritance, which is not necessarily a fair practice. Dowry is given as a favour, while inheritance is received as a right. According to McCreery (1976), for example, women given dowry in China do not enjoy the same rights and privileges that are associated with inheritance. Similarly, Brown and Chowdhury (2002) find that only few women in rural West Bengal own agricultural land and even the house where they live, while dowry is a commonly practised alternative to inheritance.

Analysing intergenerational land transfer practices in West Ghana, Quisumbing, et al. (2004) find that the growing demand for female labour due to rising intensity of land utilisation had a favourable effect on ownership and management of land by women. In contrast, Matashane and Marite (2005) find that in Lesotho despite various rural development measures especially targeting women's empowerment, women remain subjected to discrimination when it comes to inheritance of agricultural land.

Deere and Leon (2003) find that in Latin America the inequality in land holdings between men and women remains significant. The study attributes this inequality to preferential treatment of men in society, marriage traditions that favour men and men-biased government's land distribution schemes. The study also finds that gender inequality in land holding is followed through generations because of the prevalence of parallel inheritance channels by which men tend to bequeath to sons while women bequeath to daughters. This is despite the observation that women land-owners play greater role in decisions regarding household consumption and farm production. A different form of gender-based intergenerational transfer of wealth has also been observed in the Philippines by Estudillo, et al. (2001). The study finds that educated fathers tend to transfer wealth to sons both through investment in their education and ownership of land, while mothers prefer direct transfer of land to daughters. The study, however, observes that this form of gender discrimination is declining among younger generations.

3. THE MODEL

The above review of previous literature shows that there are various aspects of gender discrimination in the way intergenerational transfers are carried out. In order to capture the multidimensional nature of the issue, we consider four specific variables that indicate gender-biased attitudes regarding intergenerational transfer. These variables are: (a) intention to give legal share in inheritance to daughters, (b) opinion regarding whether women should be given their legal share in inheritance, (c) opinion regarding whether women should own property in their name, and (d) opinion regarding whether women should be given authority to manage their property. Each of these questions has binary answers of yes or no. Thus, we construct binary variables assigning the value of one to 'yes' answers and zero to 'no' answers.

The first variable indicates whether the respondents intend to give their daughters their legal shares in property. Since respondents are likely to avoid a direct question, they are asked to pick one of the four responses, which are: (a) intend to give legal share to daughters in property; (b) intend to give dowry to the daughters rather than their legal share; (c) intend to give gifts to daughters rather than their legal share in inheritance; and (d) intend to compensate daughters in other forms in place of their legal share in property. Only the first option is considered as yes response, while the other three options are regarded as no response. The second variable is aimed to assess the respondents' belief rather than their intention. The analysis of these two variables, which represent intentions and beliefs, would indicate whether intensions differ from beliefs due to social norms and pressures. The third and fourth variables assess the respondents' general opinion on whether or not women could have the right to formally own and manage property. Answers to these questions would indicate whether women are denied their legal inheritance rights only or are they also discouraged from owning and managing property in general.

The study considers five categories of independent variables, which are likely to influence their attitudes regarding women's rights to inheritance, ownership and management of property. These are: (a) economic variables, (b) demographic variables, (c) variables concerning marriage practices, (d) education related variables, and (e) awareness variables.

The economic status of respondents is measured by income and assets. Income includes wages and receipts of interests, rents and other forms of net transfers. Assets are measured in terms of financial assets plus the current market value of all physical durables net of liabilities. On prior basis we do not pose any hypothesis regarding the nature/direction of relationship of the dependent variables with income or assets. Richer respondents may be more generous to daughters if they consider the son's wellbeing a necessity and daughter's wellbeing a luxury. Or in simple terms, they may value the daughters' wellbeing after ensuring a certain minimum level of wellbeing for the sons. On thcother hand, the relationship could be reversed if respondents considered the son's wellbeing a 'normal good' and the daughter's wellbeing an 'inferior good'.

The first demographic variable i.e., family size captures various aspects. It is expected that in smaller families the generation gap is narrower and, hence, parents would be more inclined to give daughters their due shares in property. Furthermore, for the given amount of assets, smaller families would have better economic status. Thus, family size will matter if the respondents' attitudes towards daughters' rights are related to their economic status.

The number of sons and daughters can also affect parents' attitudes. In families with more sons parents can favour them at the cost of daughters to provide them with better economic security assuming that economic security of the daughters is the responsibility of their husbands. Similar considerations may apply where there are more daughters as giving legal share to the daughters would mean smaller share left for the sons. During the survey for this study, it was observed that a few parents having just one daughter intended to give her legal share in inheritance because it would not cost much to the sons and parents tend to love a lone daughter amongst several sons. Another observation was that daughters are often refused their share in property because it will divide the family's property and reduce its economic status. If this be the case, then the same argument can be applied to division of property among sons. Therefore, on this particular basis, gender discrimination can be explained only when there is just one son and therefore non-division of property would imply giving all the property to him. It follows from this discussion that it is the absolute number of sons and daughters, rather than their relative numbers, that matters.

Since official registration of births is considered as the main legal proof of the identity of a child, the parents who do not register the births of their children may find it easier to discriminate against their daughters in the distribution of inheritance wealth. Thus, the percentage of births officially registered may have some relationship with parents' attitudes towards the daughters' rights to inherit property.

With the projection of progressive and liberal thought through the global media, it is expected that in recent years people should have become more aware about the importance of giving women their rights and independence. This means that in a cross section of respondents, the younger ones are more likely to recognise the rights of women with regard to their due share in inherited property and authority in making their own economic decisions. To capture this factor we include the age of the respondent as a potential independent variable.

The status of marriage is another factor considered that can affect attitudes towards women's rights to inheritance. Denial of women's right to inheritance is often practised in the form of implicit contract in exchange marriages (A brother and sister marry another set of siblings) whereby women of the two families are denied their inheritance rights to yield zero inter-family transfers of assets. Thus, women in exchange marriages are less likely to get their rightful shares in inheritance. A somewhat weaker form of this practice is marrying within families. Here we consider three categories; cousin marriage, marriage with other relatives and marriage with non-relative.

The education level of respondents can also influence their attitudes towards women's rights. We classify education into five levels from illiterate to higher education. In addition, we also consider religious education as another factor that can potentially affect attitudes towards women's property rights. The respondents with higher levels of formal and religious education are more likely to be fair in recognising and accepting women's rights to inheritance and ownership of property.

The last set of factors pertains to the respondents' awareness in respect of women's rights. In this regard we consider various variables representing the state of the respondents' awareness and the sources of their information about women's rights. The inheritance laws in Pakistan have been prepared in accordance with the detailed instructions in the Holy Quran about the inheritance rights of wives, sons, daughters, brothers, sisters, parents and other relatives of a deceased person under various circumstances. These instructions are very elaborate and clearly mention women's rights to inheritance. It can be expected that the respondents who understand the contents of the Quran would recognise and accept women's rights to inheritance and ownership of property. This justifies the inclusion of the first awareness variable, 'understanding of the contents of Quran'. A more specific variable in this context is the respondents' awareness about Islamic laws of inheritance. Finally, another variable considered is the respondents' awareness about the country's inheritance laws.

Although the country's laws on inheritance are based on Islamic laws, the two laws pose different type of deterrence, one in the form of punishment by man-made institutions and the other in the form of punishment by divine authority. Thus, how one reacts to the two types of perceived risks will depend on one's sense of civic responsibility and one's level of understanding of religion. It is expected that the respondents who are aware of any of these two types of inheritance laws are more likely to recognise women's rights to inheritance and ownership of property.

It is also important to analyse the role of various sources of information that make people aware of women's rights. In this connection, we consider four sources, which are formal religious sources, formal education, media and parents/elders. Awareness obtained through formal religious sources, education and media is likely to be relatively more effective in shaping the respondents' attitudes towards women's rights to inheritance and ownership of property. As to the relative effectiveness of the three sources, we do not pose any prior hypothesis and leave the matter to empirical results. Table 1 summarises the above descriptions of variables.

As should be clear from the nature of research problem at hand, our analytical framework involves binary options. In one case the binary options are regarding the choice between one of the two intended actions, that is, to give or not to give daughters their legal shares in inheritance. In the other three cases the binary options involve one of the two opinions, one in favour and the other against daughters regarding their right to inheritance, ownership of property and management of property. In either case the dependent variable can take only two binary values: one or zero. Obviously, in this situation linear regression models are unsuitable. Therefore we consider two non-linear models, which are logistic and Probit models.

Denoting observation i on the binary dependent variable by [Y.sub.i], the row vector of observation i on explanatory variables by Xj and the column vector of the regression parameters by P, the two models for the determination of the binary variables considered in Table 1 can be represented as follows:

Logistic Model: [Y.sub.i] = [[e.sup.X,[beta]]/1 + [e.sup.X,[beta]]] + [U.sub.i] (1)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

For X variables we use the independent variables listed in Table 1. Both the models are estimated by standard methods of estimation for non-linear equations following the ML principle. Once the models are estimated, the probability derivatives or marginal effects with respect to various -V-variables are obtained as follows, where bar on X indicates the vector of sample means and hat on p indicates the vector of estimated parameters.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

In the first round we will estimate the models with all the independent variables listed in Table 1. However, it is understood that not all the explanatory variables will turn out to be statistically significant. Therefore, we will apply the Wald tests for the joint significance of various sets of variables such as economic variables, demographic variables, etc. In addition, we will also follow general-to-specific (step-wise backward elimination) procedure to arrive at the final regression estimates.

4. DATA AND DISCRIPTIVE ANALYSIS

For empirical analysis a sample of respondents is collected from district Lakki Marwat, Khyber Pakhtunkhwa (formerly North West Frontier Province), Pakistan. The district is characterised by agrarian economy and very low level of literacy among women. According to the 1998 census [Pakistan (1998, 2000)], the district has a population of 490025 residents, 90.43 percent of whom live in rural areas. The district has an area of 3,164 square kilometers. The literacy rate is 50.3 percent among male and only 8.6 percent among female population. The average family size is 9.2 persons per unit and the total units are about 53302.

Most of the people in this region are Sunni Muslims and most young and old men go to mosques for prayers five times a day. The majority of people are farmers, while a considerable number of young people are engaged in government and private services (e.g. banking, education, medical, shop keeping and other small businesses). Agriculture plays an important role in sustaining the economy of the district.

The custom in the arrangement of marriage is to pay money to the guardian of the girls. Educated and upper class people, however, do not insist on monetary payment. There is always some secrecy about the amount settled to be paid and an element of shame attached to the contracts, should knowledge about a transaction becomes public. Boys and girls are not betrothed until they attain puberty. Girls in towns as well as in the villages start wearing veil when they are 10-12 years of age.

Lakki Marwat district consists of two tehsils (administrative units), namely Lakki Marwat and Sarai Naurang and 36 union councils (administrative sub-units). Using random numbers, these union councils were sorted in arbitrary order and the first 10 union councils were selected. From each union council one village/suburb was selected randomly and from each village approximately 50 families were selected using a mix of convenience, random and judgment sampling. This resulted in an overall sample size of 507 families. For the obvious reason only such potential respondents were selected who had both daughters and sons. The distribution of the sample by regions is shown in Table 2.

The head of each selected family was requested to provide oral information according to a pre-specified questionnaire. To help interpret the statistical results of the study, group discussions were carried out in five villages where the respondents were willing to spare time.

The respondents' opinions about recognising women's rights to inheritance are recorded and analysed both qualitatively and quantitatively. The data and group discussions reveal that women's rights regarding access to and control over land, housing, and property are important in determining their overall living conditions, economic security and even physical safety. The recognition of women's rights to inheritance is also accepted as a serious social consideration. Table 3 provides the most basic statistics.

The data in the table--specifically the two middle rows--, along with discussions with respondents show that the majority of respondents recognise women's rights to inheritance and almost half of them also acknowledge women's possession of property as their right but when it comes to management of the property, a vast majority believes that the women should not be given the right to manage their own property. The majority also do not intend to give their daughters their legal shares in inheritance. The apparently contradictory responses arise partly due to cultural constraints and lack of confidence in women's ability to manage property. The majority of respondents consider ownership of property as a source of protection and security to women in times of emergency. However, it is generally accepted that women are incapable of managing and dealing with matters relating to property and legal issues as they lack sufficient understanding of the law regarding property. Based on this general perception related to land management by women, respondents seem to be reluctant to give women their rights to inheritance on customary grounds. Thus, an overwhelming majority of the respondents believe that it is difficult for women to manage property on their own and, hence, do not intend to act on their positive attitudes towards daughters.

Extensive discussions with the respondents reveal that many of them do not intend to act on their beliefs due to social pressure or other considerations. In particular, quite a few respondents believe that women should be given their due share in property but they do not intend to act on this belief due to perceived difficulties in management of the property that women are expected to face in the so-called 'men dominated society' and pressure of other family members, especially elders of the family and wives. However, a common observation is that men are reluctant to put their economic power at stake by giving women their due right to inheritance. Considering the views about the reasons for not giving the due right to women in inheritance, it is found that 58 percent of the respondents believe that if women are given property in their name, it would lead to disputes in the family. Another 30 percent argue that women do not need property in their name and about 12 percent respondents blame women's mental capability in handling property.

It is also observed that a majority of the respondents recognise women's rights to inheritance and almost half of the respondents also agree that women could be given legal possession of the inherited property, but only 7 percent believe that women should be given authority to manage the property. The majority of the respondents argue that it is the sociocultural value system prevailing in society that makes it difficult for women to approach and interact with male officials for settling matters of their inherited property and consider this difficulty a hurdle for the women to manage their property. Here it can be concluded that the data endorse the recognition of women's right to inheritance and possession of property to some extent but do not endorse their right to exercise control over the property.

We now turn our attention to the estimation of quantitative relationships between various indicators of attitudes toward women's rights to inheritance and the selected independent variables on the basis of the observed data.

5. RESULTS AND DISCUSSION

Since the dependent variables are binary variables, we have estimated the probability function using the Logistic and Probit regression models. In order to avoid repetition, only the results of Logistic regression models are presented here, while the results of Probit models are placed in an appendix. In the regression results presented here we report the probability derivatives (marginal effects) estimated at the sample means and the corresponding t-statistics.

Following the framework given in Section 3, we estimate four different Logistic regression equations corresponding to the four binary variables (Equation 1). For the obvious reason for each categorical independent variable, one of the categories is excluded from the regression equation to serve as the reference category for comparison with the included categories. In the first round of estimation, quite a few independent variables turned out to be statistically insignificant. Therefore, in order to determine the strength of the various sets of variables in explaining the models, we applied the Wald tests for joint significance of the corresponding parameters. In addition, to improve the quality of the estimates we also followed general to specific (stepwise backward elimination) procedure. In the first step, the original equations with all the variables specified in Table 1 are estimated. Then in each estimated equation the variable with the smallest t-statistic (most insignificant) of its regression coefficient is dropped and the restricted equation is re-estimated. Again the variable with the smallest t-statistic of its regression coefficient is dropped and the process is continued till all t-statistics are greater than one in absolute terms. In order to establish the robustness of the procedure, the variables dropped at the initial stages are included again in the equation at later stages. As it turns out, in no case could a variable dropped at the initial stage be included at a later stage, which confirms that the process of model specification has been robust.

The Wald test results are reported in Table 4. It is obvious from the table that not all sets of variables are jointly significant in all the four equations. The first set of variables, that is, economic variables namely income and assets appear to be statistically significant in determining whether respondents intend to give their daughter their due share in property and their opinion about whether women should be given their legal share in inheritance. On the other hand, the economic variable turns out to be insignificant in determining the respondents' opinion regarding whether women should own property in their names and whether they should be allowed to manage their property.

The demographic variables are significant only in determining whether the respondents intend to give daughters their legal shares in property and the respondents' opinion on whether women should have property in their names. The narrower set of variables consisting of the family size, number of sons and number of daughters is significant only in determining the respondents' opinion on whether women should have property in their names. Thus, the demographic variables do not appear to be much important in determining the respondents' attitude towards women's property rights.

Marriage related variables are jointly significant in determining the respondents' opinions on whether daughters should be given their legal shares in property and whether women should own property in their names. But the practices of exchange marriage and marriage with relatives do not affect significantly the actual intention of respondents for giving daughters their legal shares in inheritance and their opinion on whether women should be allowed to manage their property.

The results show that the education level of respondents including religious education is the most significant factor in determining their attitudes towards women's property rights. Education is found to affect all aspects of attitudes towards women's rights to inheritance, and the ownership and management of property.

Coming to variables representing the respondents' awareness about women's rights, we have tested several sets of variables. First, an important result is that respondents' understanding of Holy Quran and their awareness about Islamic and country's laws of inheritance do not affect significantly their intentions about giving the daughters their legal shares in inheritance, even though these awareness variables do matter in forming their opinions on women's rights to own property. On the other hand, respondents' attitudes seem to be affected by the source of their information about women's rights.

All in all, we conclude that none of the five categories of potential explanatory variables, that is, family's economic status and demography, state of respondents' marriage, respondent's education and respondents' awareness about women's rights, can be rejected altogether in the overall framework of analysis. However, the respondents' opinion on whether women should have authority to manage property is not much affected by any set of variables except the education level of the respondents.

Next, following the general to specific (stepwise backward elimination) procedure as outlined earlier, we obtained the estimates for the final selected Logistic and Probit regression models. The results of the Logistic model are presented in Table 5, while the results of Probit model are given in the Appendix Table A5. A detailed discussion of the results of the Logistic model proceeds as follows:

The overall performance of the estimated equations is satisfactory as indicated by the values of McFadden R-squared and Likelihood statistics. However, relatively low values of the McFadden R-squared in all the equations and the presence of significant intercepts in three equations indicate that some variables that could explain the variations in the dependent variables are missing from the specified model. But this is quite a common observation for behavioural models estimated on the basis of cross-section data.

Coming now to the role of individual variables, we observe that the economic status of the families, as indicated by the family's current income, serves as an important factor contributing to the respondents' favourable opinions regarding women's right to ownership and management of property.

The opinions in favour of women's rights to inheritance and management of property are positively and significantly related to the level of family income. Although the respondents with higher current family income appear to be less likely to give their daughters their due share in property, the relationship is statistically insignificant. The improvement in economic status, as indicated by assets is positively (and somewhat significantly) related to the respondents' favourable attitudes towards women. For example, on average an increase of one million rupees in net assets would bring about 4.8 percentage points increase in the likelihood that a respondent would intend to give his daughters their legal share in property.

Turning now to demographic variables, the table shows that family size is an important determinant of respondents' attitude towards women's rights. The respondents with larger family sizes are less likely to believe that daughters should be given their shares in property and women may have property in their names. A possible interpretation is that, given the economic status of a family, the respondents belonging to larger families tend to believe that giving daughters their share in property would make all their offspring poorer. The belief is that since the daughters are supposed to be supported by husbands, it would be fair to ensure economic security for the sons, hence justifying reservation of inheritance wealth for sons only. It is important to note here that the family size does not affect the intention of giving daughters their due shares in property. Thus, even though the respondents with larger families tend to believe that daughters should not be given their shares in property, when it comes to intentions, family size appears to be a redundant variable. Discussions with the respondents indicate that this result does not mean that larger families with unfavourable opinions towards daughters do not intend to act on this belief. A more realistic interpretation is that even the smaller families having favourable opinion towards daughters do not intend to act on their beliefs. Social norms and prevalent practices come in their way to act on their beliefs, which are often not backed by strong convictions.

Against theoretical expectations, the number of sons and the number of daughters turn out to be 'redundant' except in only one equation each and in these equations the relationship is statistically insignificant. Thus, as already noted with reference to the results of the Wald test on joint significance of demographic variables, the number of sons and the number of daughters are not much important in determining the respondents' attitudes towards women's rights.

The next variable under consideration is the percentage of births registered with the local births and registration office. The results show that families that get the births of their children officially registered are more likely to give daughters their legal share in property even though they are less likely to have favourable opinions regarding women's rights to inheritance and ownership of property. This obviously means that the official registration of births, being the main legal proof of the identity of a child, makes it difficult for the parents to discriminate against daughters in the distribution of inheritance wealth.

The age of the respondent has negative regression coefficients in the three equations in which it appears and the regression coefficient is highly significant in the first equation. This shows that the younger respondents have more favourable attitude towards women. The probability derivative estimated at the sample mean shows that a 10 years younger respondent is five percentage points more likely to intend to give his daughters their legal share in property and 4 percentage points more likely to view that women may keep property in their names.

All the three marriage related variables turn out to be redundant in, and hence dropped from, the first equation. The variable, however, has significant correlation with the three dependent variables describing the respondents' views on women's right to inheritance, ownership and management of their properties. The result shows that exchange marriages and marriages with cousins or other relatives tend to promote negative attitudes towards women's economic rights. For example, respondents having cousin marriage are about 12 percentage points more likely to view that daughters should not be given their legal shares in property and about 22 percentage points more likely to view that women should not keep property in their names. However, a vast majority of respondents do not intend to give daughters their legal shares in inheritance irrespective of whether they had exchange marriage or married within the family.

The level of education of the respondents appears to be the only variable which plays a significant role in all the four aspects of attitudes towards women's rights to inheritance, ownership and management of property. The most reassuring result is that the respondents with higher education attainments are more likely to give their daughters their legal share in property. For example a respondent with 16 or more years of schooling is 31 percentage points more likely to give his daughters their legal shares in property as compared to illiterate respondents. In addition, in all the cases reported in the table, the education level is also affirmatively correlated with the respondents' positive views towards recognising women's rights to inheritance, ownership and management of property. For example, as compared to illiterate respondents, those with 16 or more years of schooling are 62 percentage points more likely to view that a woman may keep property in her own name and 8.5 percentage points more likely to view that a woman may have authority to manage her property in her own name.

Education up to the middle or higher levels seems to be a critical variable in positive attitudes as seen in the table. Since about half of the sampled respondents (48 percent) have obtained middle or higher level of education, while only 3.75 percent have obtained 'higher' education, the overall position for women's rights to inheritance in the region does not appear much promising. This is particularly the case because the level of literacy is expected to be even lower in the region than revealed in the sample as quite a large number of potential respondents with little or no education could not understand or respond to many questions and were replaced by others.

Quite interestingly, religious education appears to be quite a significant factor in determining the respondents' attitudes towards women's economic rights. Specifically, the respondents with religious education are 22 percentage points more likely to give their daughters their due shares in property, 51 percentage points more likely to view that women should be given their legal shares in property and 69 percentage points more likely to view that women may keep property in their own names.

The final set of variables that we consider now is about the respondents' awareness about inheritance rights of women. The results show a surprising trend whereby the three awareness variables, namely understanding of the contents of Holy Quran and awareness of Islamic laws of inheritance and the country's basic laws of inheritance have no significant effect on the respondents' intentions to give their daughters their lawful shares in property. As regards the relationship of awareness with respondents' opinions about women's rights, the results show mixed patterns. While understanding of Quran and awareness of Islamic laws of inheritance seem to have adverse effects on the respondents' opinions regarding women's rights, the respondents' awareness about the country's basic laws of inheritance contributes favourably to respondents' opinions about women's rights. Thus, the respondents seem to be influenced positively more by their awareness about the country's basic laws than by their awareness about religion. A possible reason that could explain this behaviour may be that the respondents who are aware of the Islamic laws of inheritance have tendency to over-value rights of men against the rights of women because according to Islam women do not have equal status in inheriting property or giving witness in a court of law. Thus, while Islam defines, recognises and emphasises on adhering to women's rights, men are likely to misinterpret its teachings to economically discriminate against women. However, a reassuring result is that the respondents having acquired formal religious education tend to have highly favourable attitudes towards women. This means that mere awareness is not sufficient; one also has to have deeper understanding of religion to recognise the place of women in social order.

We now consider the last three variables, which relate to the sources of respondents' information about women's inheritance rights. The results show that all the formal sources of awareness, that is, religion, education and media, as compared to informal source of awareness through parents and elders, play an important role in the recognition and acceptance of women's rights to inheritance, ownership and management of their property. For example, the respondents who have become aware of women' inheritance rights through media are 11 percentage points more likely to intend giving their daughters their due shares in inheritance, 29 percentage points more likely to view that women should be given their lawful shares in property and 33 percentage points more likely to view that women may keep their property in their own names.

It is important to note that the respondents who claim that they have obtained information regarding women's economic rights through former religious sources are more likely to give their daughters their due shares in inheritance and recognise women's rights to inheritance and ownership of property and the difference is highly significant in all the three cases. This result combined with the observed significant role of religious education and the insignificant or adverse effects of the understanding of Quran and awareness of Islamic laws of inheritance means that religious teachings can shape the respondents' attitudes towards women's right provided such teachings are acquired directly from the authentic sources such as the Holy Quran. It is, therefore, important to emphasise here that awareness is a subjective concept and those who think they are aware, may not in reality be aware properly unless the source of information is undisputed like the direct reading of the original source material.

The respondents' knowledge of Islamic teachings on women's rights is not effective in shaping the respondents' attitudes unless the knowledge is acquired directly from formal religious sources and backed by formal religious education.

6. SUMMARY AND CONCLUSION

This study has been carried out to estimate and analyse the contribution of various economic, demographic, and social factors that can potentially affect the beliefs and attitudes prevailing in society regarding women's rights to inheritance, ownership and management of their property. The study is based on a sample of 507 families collected through a field survey in District Lakki Marwat, a backward district dominated by rural population, agrarian economy and low literacy of women. The district is located in the province of Khyber Pakhtunkhwa (KPK), Pakistan. The sample is selected through a stratified random sampling procedure covering various locations as well as rural urban divide of the district. The study uses both qualitative as well as quantitative data, which explain the relationship of the respondents' attitudes towards women's rights to inheritance and their control over property with various socioeconomic and demographic variables. For a formal analysis the study uses Logistic and Probit regression models.

The study arrives at several interesting conclusions. The economic status of families appears to be quite relevant in the given context. The results show that in most cases improvement in economic status of families is positively and significantly correlated with respondents' favourable attitudes towards women. The respondents with larger family sizes and/or lower level of wealth are less likely to believe that daughters should be given their shares in property and women may have property in their names. A possible reason is that other things held constant, the respondents belonging to larger families or having lower stock of wealth tend to believe that giving daughters their share in property would make all their offspring poorer. The belief is that since the daughters are supposed to be supported by husbands, it would be fair to ensure economic security for the sons, justifying sole inheritance rights of the son. It is important to note here that family size does not affect the intention of giving daughters their due shares in property. Thus even though the respondents with smaller families tend to believe that daughters should be given their shares in property, they still do not intend to act on this belief, indicating that social norms and prevalent practices come in their way when they are to act on their beliefs, which are often not backed by strong convictions.

The study finds that persuasion of families for regular official registration of births can be instrumental in letting women obtain their legal shares in inheritance. Another expected result is that the respondents belonging to younger generation have relatively more favourable attitude towards women in recognising their rights to inheritance, ownership and management of property.

A useful result is that the respondents with higher education attainments are more likely to give their daughters their legal shares in property. In addition, education level is also positively correlated with the respondents' positive views towards recognising women's rights to inheritance, ownership and management of property. Religious education also comes out to be instrumental in promoting positive attitudes towards women's rights to inheritance.

An unexpected result found in the study is that the awareness of the respondents in terms of understanding the Holy Quran and the knowledge of Islamic laws of inheritance have no significant effect on their intentions to treat their daughters fairly in the distribution of inheritance wealth. On the other hand, the respondents' awareness about the country's basic laws of inheritance contributes favourably to respondents' opinions about women's rights. A possible reason could be that people with better awareness of the Islamic laws of inheritance have tendency to over-value the rights of men against the rights of women because according to Islam women do not have equal status in inheriting property. The results, however, also suggest that the respondents who claim to have obtained information regarding women's economic rights directly from formal religious sources and/or have acquired formal religious education are relatively more likely to give their daughters their rightful shares in inheritance and form positive attitudes towards women's rights to inheritance and ownership of property.

An obvious implication of the results is that the attitudes towards women can be significantly improved through education. This not only includes formal education of individuals through schooling but also the general education of society through electronic and print media. Although various TV channels present programmes on Islamic teachings but most of them are concerned with technical aspects of Islamic practices, while those focusing on women's rights are rare.

Furthermore, the contents of education also need to be revised to educate the society about the rights of women as prescribed by law and religion. There is also a need to reform religious education system. The syllabi of Islamic studies should also include references to such Islamic teachings that clearly state the importance of women's rights. For example, translation and elaboration of the sections of Sura Nisa (a chapter in Holy Quran) that gives a very detailed guideline about the distribution of inheritance wealth can be made part of the syllabi at least at the high school level, if not earlier.

There is also a need to educate the society about the negative side of certain marriage practices prevalent in society such as exchange marriage and marriage with relative, especially with cousins. In this respect also the syllabi of the formal education system need to be revised and various sources of media may be encouraged to come forward and play their roles.

APPENDIX A Table A5 Results of Probit Regression Models Dependent Variable Intention_ Opinion_ Independent Variables Share Share Intercept -1.390 0.442 (-1.76 **) (1.30) Income -0.028 0.026 (-1.62 (2.45 *) Assets 0.473 (3.39 *) N -0.080 (-2.51 *) N_Sons N_Daughters -0.185 (-1.77 **) Birth_registration 0.013 -0.006 (2.86 *) (-261 *) Age -0.049 (-3.30 *) Marriage_exchange Marriage_cousin -0.311 (-2.42 *) Marriage_relative Edu_Primary 0.651 0.413 (1.46 (1.46) Edu_Middle 0.861 0.459 (2.59 *) (2.16 *) Edu_College 0.517 0.408 (162) (2.50 *) Edu_Higher 2.898 (3.08 *) Edu_religious 2.039 1.199 (2.08 *) (1.92 **) Understanding_Quran -1.091 -0.352 (-132) (-1.10) Aware_Islamic_law 0.401 -0.272 (1.22) (-1.70 **) Aware_Country_law 0.402 (1.93 **) Inf_religion 1.226 0.887 (3.10 *) (5.62 *) Inf_education 0.856 0.874 (1.47) (2.71 *) Inf_media 0.904 0.768 (1.22) (2.37 *) McFadden R-squared 0.321 0.190 LR statistic 83.9 * 128.4 * Dependent Variable Opinion_ Opinion_ Independent Variables Name Manage Intercept 1.051 -1.969 (2.58 *) (-3.40 *) Income 0.0298 (2.16 *) Assets 0.095 (113) N -0.084 (-2.62 * N_Sons 0.075 (1.00) N_Daughters Birth_registration -0.005 (-2.22 *) Age -0.009 -0.015 (-1.49) (-1.22) Marriage_exchange -0.416 (-2.05 *) Marriage_cousin -0.529 -0.363 (-3.26 *) (-1.56) Marriage_relative -0.326 (-2.04 *) Edu_Primary Edu_Middle 0.442 (2.22 *) Edu_College 0.553 0.811 (3.31 *) (3.07 *) Edu_Higher 1.528 1.906 (2.90 *) (4.13 *) Edu_religious 1.672 0.441 (2.96 *) (0.96) Understanding_Quran -0.828 (-2.21 *) Aware_Islamic_law -0.263 (-1.68 **) Aware_Country_law 0.248 0.406 (1.28) (164) Inf_religion 0.479 (3.03 *) Inf_education 0.573 (1.84 **) Inf_media 0.788 0.582 (2.45 *) (1.30) McFadden R-squared 0.149 0.331 LR statistic 104.3 * 82.6 * Note: The results show estimated regression coefficients and t-statistics. The statistics significant at 3 percent and 10 percent levels are marked by * and ** respectively.

REFERENCES

Agarwal, B. (1994) A Field of One's Own: Gender and Land Rights in South Asia. Cambridge: Cambridge University Press.

Agarwal, B. (1998) Widows versus Daughters or Widows as Daughters? Property, Land, and Economic Security in Rural India. Modern Asian Studies 32:1.

Bennett, L. (1981) Women, Law, and Property in the Developing World: An Overview. Human Rights Quarterly 3:2.

Brown, J. and S. D. Chowdhury (2002) Women's Land Rights in West Bengal: A Field Study. (RDI Reports on Foreign Aid and Development, No. 116).

Deere, C. D. and M. Leon (2003) The Gender Asset Gap: Land in Latin America. World Development 31:6.

Deere, C. D. and M. Leon (2001) Empowering Women: Land and Property Rights in Latin America. Pittsburgh: University of Pittsburgh Press.

Estudillo, J. P., et al. (2001) Gender Differences in Land Inheritance and Schooling Investments in the Rural Philippines. Land Economics 77:1.

Kamal, S. (1999) Effects of the Interplay of Formal and Customary Laws on Women in Tribal Cultures. Paper Presented at The 7th Interdisciplinary Congress on Women, Women's Worlds, No. 99.

Matashane-Marite, K. (2005) Women's Rights and Participation-Including Women's access to Land and Inheritance, and the Role of Lobbying and Grassroots Organisations in Lesotho. Bangkok, Thailand: United Nations, Division for the Advancement of Women.

McCreery, J. (1976) Women's Property Rights and Dowry in China and South Asia. Ethnology 15:2.

Mehdi, R. (2002) Gender and Property Law in Pakistan, Resources and Discourses. Lahore: Vanguard Books.

Mumtaz, K. (2006) Reducing the Gender Gap/Engendering PRSP II. (Draft paper for UNDP).

Pakistan, Government of (1998) Census 1998, Lukki Marwat, Basic Population and Housing Data by Union Councils. Islamabad: Population Census Organisation, Statistics Division, Government of Pakistan.

Pakistan, Government of (2000) District Census Report of Lakki Marwat. Islamabad: Population Census Organisation, Statistics Division, Government of Pakistan.

Quisumbing, A. R., et al. (2004) Are Wealth Transfers Biased Against Girls? Gender Differences in Land Inheritance and Schooling Investment in Ghana's Western Region. (FCND Discussion Paper 186).

Tirmazi, M. K. (1999) Land, Water and Gendered Nobilities in Pakistan: Exploring the Contradictions. International Conference Organised by the Chinese University of Hong Kong.

Eatzaz Ahmad <eatzaz@qau,edu.pk> is Professor, School of Economics, Quaid-i-Azam University, Islamabad. Anbereen Bibi is Lecturer, Department of Economics, University of Wah. Tahir Mahmood <[email protected]> is Assistant Professor, School of Economics, Quaid-i-Azam University, Islamabad. Table 1 List and Description of Variables Description Variable Type/Unit Attitudes Towards Women's Property Rights (Dependent Variables) Intention to give daughters legal Categorical: yes/no share in inheritance Should women be given legal Categorical: yes/no share in inheritance Should women own property Categorical: yes/no owned in their names Should women be allowed to Categorical: yes/no manage their property Family's Economic Status Monthly family income Thousand rupees Net family assets Million rupees Family's Demography Number of family members Count Number of sons Count Number of daughters Count Percentage of births in family Percentage registered officially Age of respondent Years Respondent's Marital Background Exchange marriage Categorical: yes/no Respondent's state of marriage: Categorical: yes/no Cousin marriage, marriage with other relative Respondent's Education Education level: 0-4 years, 5-7 Categorical years, 8-9 years, 10-15 years, 16 or more years and religious education * Respondent's Awareness on Women's Rights Understanding of the contents of Categorical: yes/no Holy Quran Awareness of Islamic laws of Categorical: yes/no inheritance Awareness of country's laws of Categorical: yes/no inheritance Source of information on women's Categorical: yes/no rights: Religious sources, education, media and parents/elders Description Notation Attitudes Towards Women's Property Rights (Dependent Variables) Intention to give daughters legal Intention_share share in inheritance Should women be given legal Opinion_share share in inheritance Should women own property Opinion_ownership owned in their names Should women be allowed to Opinion_management manage their property Family's Economic Status Monthly family income Income Net family assets Assets Family's Demography Number of family members N Number of sons N_sons Number of daughters N_daughters Percentage of births in family Birth_registration registered officially Age of respondent Age Respondent's Marital Background Exchange marriage Marriage_exchange Respondent's state of marriage: Marriage_cousin, Cousin marriage, marriage with Marriage_family other relative Respondent's Education Education level: 0-4 years, 5-7 Edu_Illiterate, Edu_Primary, years, 8-9 years, 10-15 years, 16 Edu_Middle, EdujCollege, or more years and religious Edu_Higher, Edu_Religious education * Respondent's Awareness on Women's Rights Understanding of the contents of Understanding_Quraan Holy Quran Awareness of Islamic laws of Aware_Islamic_law inheritance Awareness of country's laws of Aware_Country_law inheritance Source of information on women's Inf_religion, Inf_education, rights: Religious sources, Inf_media, Inf_parents/elders education, media and parents/elders * There are four levels of religious education. Qari is a person who can recite Holy Quran with correct pronunciation. Hafiz is a person who has memorised the complete Holy Quran. Munshi Fazal is a formal certificate of education in religion, while Alam Fazal is a higher certificate in religious education. Table 2 Region-wise Frequency Distribution of the Sample Lakki Marwat Units Urban Rural Total Union Council -- -- 7 Families 102 255 357 (29%) (71%) Sarai Naurang Units Urban Rural Total Total Union Council -- -- 3 10 Families 50 100 150 507 (33%) (67%) Table 3 Respondents' Attitudes Towards Women's Property Rights Recorded Responses Question Yes No Total Will you give your daughters their 66 441 507 legal share in property? (13.02%) (86.98%) Should women be given their legal 313 194 507 share of inheritance? (61.74%) (38.26%) Should women own property in their 239 268 507 name? (47.14%) (52.86%) Should women be given authority to 34 473 507 manage their property? (6.71%) (93.29%) Table 4 Results of Wald Test for the Joint Significance of Parameters Dependent Variable Intention_ Opinion_ Set of Independent Variables Tested share share Income, Assets 5.080 * 2.470 ** 10.160 * 4.930 ** N, N_sons, N_daughters, Birth_registration, 3.640 * 1.840 Age 18.200 * 9.200 N, N_sons, N_daughters 0.840 1.780 2.530 5.350 Marriage_exchange, Marriage_cousin, 0.110 2.490 ** Marriage_relative 0.320 7.480 ** Marriage_cousin, Marriage_relative 0.160 3.690 * 0.310 7.390 * Edu_Primary, Edu_Middle, Edu_College, 2.540 * 2.350 * Edu_Higher, Edu_Religious 12.700 * 11.740 * Edu_Primary, Edu_Middle, Edu_College, 3.140 * 1.940 Edu_Higher 12.580 * 7.770 Understanding_Quraan 1.890 2.370 1.890 2.370 Understanding_Ouraan, Aware_Islamic_law 1.720 2.610 ** 3.440 5.220 ** Aware_Islamic_law, Aware_Country_law 0.850 2.550 ** 1.710 5.110 ** Aware_Islamic_law, Aware_Country_law, 2.010 ** 6.370 * Inf_religion, Inf_education, Inf_media 10.040 ** 31.850 * Inf_religion, Inf_education, Inf_media 2.530 ** 9.930 * 7.590 ** 29.790 * Dependent Variable Opinion_ Opinian_ Set of Independent Variables Tested ownership management Income, Assets 0.720 2.230 1.440 4.470 N, N_sons, N_daughters, Birth_registration, 3.450 * 0.690 Age 17.240 * 3.440 N, N_sons, N_daughters 2.500 ** 0.520 7.490 ** 1.560 Marriage_exchange, Marriage_cousin, 4.080 * 0.900 Marriage_relative 12.230 * 2.710 Marriage_cousin, Marriage_relative 5.180 * 1.200 10.360 * 2.390 Edu_Primary, Edu_Middle, Edu_College, 3.970 * 2.100 ** Edu_Higher, Edu_Religious 19.870 * 10.480 ** Edu_Primary, Edu_Middle, Edu_College, 3.590 * 2.610 * Edu_Higher 14.360 * 10.440 * Understanding_Quraan 4.580 * . 0.170 4.580 * 0.170 Understanding_Ouraan, Aware_Islamic_law 3.830 * 0.300 7.660 * 0.610 Aware_Islamic_law, Aware_Country_law 2.280 1.560 4.550 3.120 Aware_Islamic_law, Aware_Country_law, 2.800 * 0.830 Inf_religion, Inf_education, Inf_media 13.980 * 4.160 Inf_religion, Inf_education, Inf_media 3.740 * 0.370 11.230 * 1.100 Note: The first value in a cell is the F-statistic, while the second value is Chi-square statistic. The statistics significant at 5 percent and 10 percent levels are indicated by * and ** respectively. Table 5 Results of Logistic Regression Models Dependent Variable Intention_ Opinion_ Independent Variables Share Share Intercept -2.724 0.739 (-1.85 **) (1.28) -0.163 0.169 Income -0.050 0.045 (-1.58) (2.39 *) -0.003 0.010 Assets 0.800 (3.23 *) 0.048 N -0.134 (-2.47 *) -0.031 Sons Daughters -0.304 (-164) -0.018 Birth_registration 0.024 -0.010 (2.93 *) (-2.47 *) 0.001 -0.002 Age -0.087 (-3.26 *) -0.005 Marriage_cousin -0.534 (-2.44 *) -0.122 Edu_Primary 1.158 0.702 (1.42) (1.45) 0.069 0.160 Edu_Middle 1.540 0.759 (2.52 *) (2.12 *) 0.092 0.173 Edu_College 0.903 0.687 (1.50) (2.48 *) 0.054 0.157 Edu_Higher 5.229 (3.03 *) 0.313 Edu_religious 3.697 2.218 (2-11 *) (1.85 **) 0.221 0.507 Understanding_Quran -1.970 -0.645 (-134) (-1.21) -0.118 -0.147 Aware_Islamic_law 0.850 -0.458 (1 33) (-1.70 **) 0.051 -0.105 Aware_Country_law 0.679 (1.86 **) 0.155 Inf_religion 2.247 1.454 (2.76 *) (551 *) 0.134 0.332 Inf_education 1.458 1.450 (1.26) (2.63 *) 0.087 0.331 Inf_media 1.827 1.261 (1-33) (2.33 *) 0.109 0.288 McFadden R-squared 0.317 0.190 LR statistic 82.9 * 128.2 * Dependent Variable Opinion_ Opinion_ Independent Variables Name Manage Intercept 1.799 -3.556 (2.65 *) (-2.89 *) 0.448 -0.086 Income 0.053 (2.05 *) 0.001 Assets 0.152 (108) 0.038 N -0.139 (-2.58 *) -0.035 Sons 0.162 (1.10) 0.004 Daughters Birth_registration -0.009 (-2.16 *) -0.002 Age -0.017 -0.034 (-1.59) (-1.29) -0.004 -0.001 -0.684 (-2.03 *) -0.170 Marriage_cousin -0.881 -0.679 (-3.24 *) (-147) -0.219 -0.016 -0.557 (-2.09 *) -0.139 Edu_Primary Edu_Middle 0.739 (2.28 *) 0.184 Edu_College 0.908 1.688 (3.27 *) ' (2.79 *) 0.226 0.041 Edu_Higher 2.495 3.501 (2.80 *) (3.93 *) 0.622 0.085 Edu_religious 2.778 0.855 (2.81 *) (1.05) 0.692 0.021 Understanding_Quran -1.304 (-2.14 *) -0.325 Aware_Islamic_law -0433 (-1.66 **) -0.108 Aware_Country_law 0.400 0.856 (1.26) (1.79 **) 0.100 0.021 Inf_religion 0.792 (3 00 *) 0.197 Inf_education 0.943 (1.83 **) 0.235 Inf_media 1.334 1.060 (2.47 *) (1.22) 0.332 0.026 McFadden R-squared 0.149 0.327 LR statistic 104.5 * 81.5 * Note: The results show estimated regression coefficients, t- statistics and the estimated probability derivatives at sample means. The statistics significant at 5 percent and 10 percent levels are marked by * and ** respectively.