Performing Marriage?

Gender Scripts and the Marital Timing in India

Lester Andrist Manjistha Banerji Sonalde Desai

India Human Development Survey Working Paper No. 9

Performing Marriage? Gender Scripts and the Marital Timing in India

Lester Andrist University of Maryland College Park landrist@umd.edu Manjistha Banerji University of Maryland College Park mbanerji@socy.umd.edu Sonalde Desai University of Maryland College Park sdesai@socy.umd.edu

Version: August 2008

India Human Development Survey Working Paper No. 9

Views presented in this paper are authors personal views and do not reflect institutional opinions. The results reported in this paper are based primarily on India Human Development Survey, 2005. This survey was jointly organized by researchers at University of Maryland and the National Council of Applied Economic Research. The data collection was funded by grants R01HD041455 and R01HD046166 from the National Institutes of Health to University of Maryland. Part of the sample represents a resurvey of households initially surveyed by NCAER in 1993-94.

Introduction:
While it is well recognized that marriage in India is more or less universal for both males and females and tends to occur at an early age (Oberoi 1998), it is often overlooked that women in India (and neighboring Bangladesh) typically enter into marriage at an earlier age when compared to most other regions of the world (See Tables 1 and 2). However, while age at marriage in India has been increasing for both men and women, increases have been fairly slow with much of the change coming through elimination of marriage for girls under 14 years of age rather than actual delays in marriage for older teenagers or women in their twenties. Interestingly, in a 30 year period spanning from 1961 to 1991, women on average delayed their marriages only by about 3 years, from a mean age of 16 to about 19 years of age (International Institute for Population and Macro 2000). In contrast, average age at marriage in Bangladesh was delayed by a full year more during the same three decades (Islam and Ahmed 1998).

<Tables 1 and 2 about here>

It is important to think about this slow change in marriage in the context of rapid changes observed in other aspects of Indian society and economy. The economy has grown annually at a rate of 7-9 percent over the past 15 years and education has expanded rapidly for all segments of the society (Desai and Kulkarni 2008). Not even families have remained immune to these changes, evidenced by rapid declines in the Total Fertility Rate from 4.8 in the 1960s to 2.7 in recent years (International Institute for Population Sciences and Macro International 2007). Thus this relatively slow change in age at marriage remains a puzzle, and one worth examination,

-1-

particularly since marriage forms the cornerstone of other aspects of Indian life such as caste and gender relations. We argue that a structural or modernization perspective may not be enough to understand the marriage process in India. In other contexts, when structural models are deemed less than satisfactory, focus often tends to turn to culture to emphasize changing roles and norms as the driving source of observed changes in behavior (Retherford, Ogawa, and Matsukura 2001). However, this residual focus on culture as a structure which is seemingly indifferent to the inputs of actors runs the danger of seeing culture as a stagnant or fossilized backdrop (Hammel 1990). Research on marriage patterns in modern India is faced with a similar dilemma. Much of the classical research on marriage in India begins with kinship norms and patterns with emphasis on suitable marriage partners, dowry and other forms of exchange. In this work, marriage often forms the key through which social hierarchies are built and articulated (Dyson and Moore 1983; Dube 2001; Oberoi 1998; Karve 1965; Dube 1996; Srinivas 1977). While these classical studies of kinship originating out of a long tradition of village studies, have contributed enormously to our understanding of intricate web of social relationships and hierarchies, they fail to illuminate marriage patterns in a changing India. Developing nuanced theoretical models for studying social change particularly in such intimate domains as marriage behaviour remain challenging. While structural forces inform individual behaviour, the behaviours of similarly located individuals aggregate as patterns capable of continually modifying that structure. In a sense, then, agency collapses into structure and nowhere is this more apparent than when looking at the corporate family in Indian society and the role it plays marriage patterns. Typically, the family is treated along side social

-2-

structures, such as labour markets, educational systems, kinship, and caste structures. For our purposes, it is important to also recognize the corporate family as an embodied agent, which is able to negotiate competing demands placed on it. In deciding on an appropriate time to arrange marriage for their sons and daughters particularly daughters parents are faced with many difficult decisions. Some of these forces push toward an early age at marriage, others toward a later age. Finding a good match depends on the availability of eligible mates and this pool rapidly dwindles once a girl reaches an age threshold where most of her peers are married. Social pressure to arrange a match begins to mount as a girl gets older, affecting parents of the bride, the bride herself and the way in which the potential groom and his family view the bride. In some ways this is not very different from the kind of pressures American women face as they approach age markers such as 35 or 40 without having a child. At the same time, the pressures against marriage at very young ages are also strong. With strong media propaganda and laws against child marriage defined as under age 18 for women and 21 for men the Indian state has set out clear expectations which are reinforced by a changing social discourse regarding how one lays claim to modernity. Moreover, most parents wish their daughters to be happy and are concerned about thrusting the responsibilities of married life onto them at too early an age. Thus, the central problematic facing research on marriage timing in India relates to two issues: (1) How is optimal timing for marriage established? (2) How do families negotiate competing demands regarding appropriate age at marriage?

-3-

Gender Performance and Marriage Timing:

This paper argues that recent literature on the performative aspect of gender offers an interesting avenue for studying marriage patterns in the Indian context. Goffman (1976) first argued that men and women engage in a visible display of gender where a stylized mode of interaction may indicate deference or dominance. This concept was further elaborated in a provocative paper by West and Zimmerman (1987), titled Doing Gender, where the authors argue that gender is a powerful ideological device which shapes choices and limits actions based on the actors sex and leads individuals to consistently act in a way that produces gendered behaviors in day-to-day interactions between individuals. In anthropology these ideas have been carried forward in the form of performance theory where gender is constituted through symbolic enactment in a highly visible manner (Morris 1995). Steve Dern (1994) in his qualitative work in Banaras (Varanasi) in north India, similarly finds that in every interaction in which a husband gives his wife permission to go outside the home, he reconstitutes the normal state of affairs in which restrictions on women are necessary (p. 210). The synergy between this performative approach and classical social anthropology of M.N. Srinivas and colleagues is striking. Srinivas first identified the role of women as custodians of family status and caste purity (Srinivas 1977). But while focusing on the notion of sanskritization, the process through which castes manipulated their ritual status by embracing gendered practices such as prohibition of widow remarriage, he also acknowledged that this might conflict with other forces such as modernization and Westernization. Srinivass work has been highly influential for several generations of scholars working within this framework.

-4-

However, it seems that the original insight on the performative aspect of culture has been obfuscated by village studies of caste structure. The concept of gender performance has many potential uses in social science research that have not been fully exploited given the roots of performance theory in day-to-day interactions and enactments in an interpersonal setting away from a study of structural relationships. Some exceptions are interesting, however. In a highly controversial essay titled Doing Difference, West and Fenstermaker (1995) expand this notion to incorporate performance of other forms of difference, particularly race, in this framework. In a paper titled Performing Modernity, Schein (1999) approaches the issue of constitution of modernity through its enactment in China. For our purposes, seeing performance theory as a subset of the new semiotic school of sociology of culture is fruitful because it allows us to focus on the way in which social actors use culture to fabricate meaning in and of their own lives (Kaufman 2004). We argue that a notion of scripts that frame actors day-to-day behavior and yet are constantly modified as actors face competing demands provides an interesting framework for a study of marriage in India. Travels across India document wide diversity in the way gender is performed. Purdah or ghunghat is probably the most visible marker or public performance of gender and it varies from a sari pulled over the face to make women virtually invisible to prying eyes in north-central India to a polite nod at segregation when an older relative is present in Gujarat to a total absence of purdah in southern India. While ghunghat or purdah is the most visible marker to an outsider, there are many other more subtle markers of gender segregation. In some parts of India, it is common for men and women to eat together, in others segregation within the family would make it unthinkable for a young daughter-in-law to eat with her fatherin-law. Restrictions on womens physical mobility is yet another marker of gender segregation

-5-

where women must seek permission from family elders before venturing outside the home to visit health centers, friends or bazaar. This gender segregation is not necessarily a marker of gender inequality in the household. Secluded women may retain substantial power in the household and women with considerable freedom of movement may not find this freedom translating into control over economic resources. This observation is consistent with a host of demographic studies of gender which have remarked upon the multidimensionality of gender inequality (Kishor 2000; Mason 1986). For our work, the focus on value placed on performance as measured through gender segregation is particularly important. Linking of gender scripts to age at marriage must also be viewed in the historical context of the late 19th and early 20th Century conflicts between the colonial state which set itself up as the protector of Indian women and the nationalist movement which needed to find an alternative construction of Indian women in order to deflect the colonial discourse. Opposition to the Age of Consent Act of 1891 proved to be one such turning point (Heimsath 1962). This act set a minimum age for a consenting bride to be 12. Nationalist Indians saw this as an attack on Indian religious autonomy and a vigorous protest emerged, led by a charismatic Indian politician, Lokmanya Balgangadhar Tilak. A subsequent increase of minimum age at marriage to 14 in 1929 in an act that came to be known as the Sarda Act also led to significant protests. Partha Chatterjee has written persuasively about the process through which the nationalist movement of early 20th Century created a vision of modern Indian womanhood that was at once modest, decorous, spiritual and refined (Chatterjee 1989, 1993). This positioning of Indian women of refinement against their Western counterparts emerged as a response to the colonial state and Western discourse which continually saw Indian women as dispossessed and subjugated.

-6-

While the resistance to colonial construction of early marriage is probably less relevant after sixty years of independence, other practical concerns persist. One of the greatest concerns for most parents is to arrange a marriage for their daughter in a good family where she would thrive. While the definition of good marriage may vary across families, there is a universal concern that nothing should damage the value of a daughter in the marriage market. Popular literature, films and social science literature all emphasize a fear of womens sexuality, particularly among upper class, upper caste families, and argue that even a possibility that the bride may not be a virgin reduces her desirability to her prospective parents-in-law. In practice, a girl does not even have to be sexual active to be labeled promiscuous. Simple contact and platonic friendship with the opposite sex can be enough to damage her reputation (Caldwell et al. 1998; Caldwell, P. H. Reddy, and Caldwell 1983; Lindenbaum 1981). Thus a long gap between puberty and marriage is seen as a risky period by parents who seek to minimize this risk by arranging an early marriage. Based on fieldwork in Hyderabad, Leonard argues that while all castes of Kayastha in her study had preferences regarding normative age at marriage, deviation from this normative age was permitted for men far more readily than for women (Leonard 1976). However, this concern with womens sexual purity is neither universal nor predominant across class and geographic boundaries (Mendelbaum 1988; Papanek 1973). Reification of womens modesty is the privilege of upper social classes, and higher caste status is often demonstrated through such reification (Sharma 1980; Dube 2001). Lower class and lower caste women rarely have the privilege of secluding themselves. Similarly, casual contact with men is viewed with much greater fear in certain areas of the country than others. We seek to better understand the role this fear of womens sexuality and immodesty plays in shaping marriage patterns via an examination of these differences across different cultural contexts. Fortunately for

-7-

our purposes, India provides a fascinating laboratory of different gender scripts, allowing us to test our hypothesis that early marriage is a part and parcel of gender scripts in which public performance of womens modesty is valued. Our focus on gender scripts emphasizes a concern with public performance of modesty and implied control over womens sexuality but is quite distinct from other measures of womens empowerment such as their control over resources or general power in household decision making (Mason 1986; Mukhopadhyay and Higgins 1988). Thus, we argue that age at marriage will be lower in areas and in communities where there is a greater concern with womens sexuality indicated by greater segregation of men and women in separate spheres but other dimensions of gender relations will not have an impact on age at marriage. It is important to also note that a deference to gender scripts embodied in early marriage often collides head on with parental desire to let their daughters mature before facing the pressures of the married life and increasing public consensus about the undesirability of child marriages. Most importantly, early marriage is often associated with curtailment of education. Since education is one of the most important claims to modernity in India, early marriage is not something parents enter into lightly. One of the interesting ways these two conflicting demands may be combined is by arranging marriage early but then delaying consummation. The Indian marriage system is characterized by a disjunction between formal marriage and cohabitation and initiation of sexual activity. Historically, marriage was quite different from gauna or effective marriage where the bride was sent to her husbands home to begin a married life. A gap between marriage and gauna was common for child marriages. But we suggest that above and beyond, ensuring a mature age at sexual initiation, this tradition may also be used by parents with claims to modernity to ensure their daughters do not curtail their

-8-

education. The grooms family must also acquiesce in this for the process to work, but frequently a desire for this obvious marker of modernity higher education is shared by both parties.

Data:
Results presented in this paper come from India Human Development Survey 2005, spanning 41,554 households over all 25 states and union territories of India (with the exception of Andaman/Nicobar and Lakshadweep). The survey was conducted by researchers from University of Maryland and the National Council of Applied Economic research and was funded by the U.S. National Institute of Health. It was specifically designed to study various dimensions of gender relations and since the data are collected in structured interviews, considerable attention was directed to framing questions which would provide information that would meaningfully tap into womens experiences within the Indian context. For this analysis, we restricted our sample to 27,932 ever married women age 25-49 for whom complete data was available. Results from 2001 Indian census indicate that nearly 95 percent women are married by age 25 and restricting our sample ever married women aged 25 and above allows us to minimize the selection bias due to the omission of women who marry late. These women were interviewed in their homes by female interviewers in local language.

Marriage Patterns in Modern India:

Given a lack of national information on marriage patterns in India we start with descriptive statistics from IHDS. Table 3 shows that average age at marriage varies considerably across demographic characteristics among ever married women aged 25 years and above in our sample. Regional differences in age at marriage are striking, with average age being 15-17 years

-9-

in central states like Bihar and Madhya Pradesh and higher age at marriage in Punjab and Himachal Pradesh as well as in the southern states. Women in poor and less educated households often marry around 16 years of age while women from better off and more educated households get married around age 19-20 years. Average age at marriage is 19.2 years in metro cities and is considerably lower in less developed villages.

<Table 3 about here>

Not surprisingly many of the young brides were physically immature and had not attained puberty at the time of marriage. For instance, in Bihar and Rajasthan, states with earliest age at marriage, around 25 percent of girls had not attained puberty at the time of their marriage. At the same time, a focus on formal age at marriage may well be mistaken in a context where early marriage is not synonymous with early age at entry into a sexual union. As documented by many anthropologists, early marriage is often associated with a delay in consummation and the bride remains with her parents until a formal gauna or bidai ceremony occurs. States with very early age at formal marriage also follow the custom of a gap of a year or more between gauna and marriage. Table 3 indicates proportions waiting at least six months following the wedding before cohabitation. In Bihar about 65 percent of women waited for six months or more to begin living with their husbands as did about 43 percent of women in Rajasthan. As Figure 1 shows, this waiting period is often associated with the relative youth and immaturity of the bride and tends to decline as the age at marriage increases. But it is important to note that regardless of the age at which formal marriage occurs, average age at which cohabitation or effective marriage begins is barely about 18-19 years in many states and even younger in others.

- 10 -

<Figure 1 about here>

Most marriages are arranged. In spite of the Valentines Day articles in English newspapers emphasizing the importance of love in marriage among urban elites, in our sample less than 5 percent of the women said they chose their husbands independent of their parents (see Table 4). The rest reported a variety of arrangements through which their families made marriage decisions. Most reported a very limited contact with their husbands before marriage; 68 percent met their husbands on the day of the wedding or shortly before; an additional 9 percent knew their husbands for a month before the wedding. Only 23 percent knew their husbands for more than a month when they got married. While educated women are more likely to have a longer acquaintance with their husbands, as Figure 2 indicates, even among women with college education, a long period of acquaintance is not normative. It is important to note that since our data were collected from women only, much of this discussion has focused on womens choices and lack thereof. However, much of this discussion also applies to males who have little opportunity to get to know their wives.

<Figure 2 about here>

Yet, in spite of the popular stereotype of women getting coerced into arranged marriages, about 65 percent felt that their wishes were considered in selection of their partners. Perhaps the most striking change in Indian marriage patterns is the extent to which womens consent is sought in marriage arrangements. A glance at Table 4 indicates that women between the ages of

- 11 -

25 and 29 reported that more of their marriages were self-arranged than any earlier birth cohorts. Indeed about 6.3 percent arranged their own marriages, as opposed to 4.5 percent of women between the ages of 45 and 49. Similarly, when compared with older cohorts, fewer 25 to 29 year olds reported that their marriages were arranged without their consent.

<Table 4 about here>

While women appear to be more inclined than older cohorts to emphasize for themselves their choice and efficacy in determining their marriage partner, one can still argue that entrenched marriage patterns make some choices more probable than others. In parts of India, especially in the north, the practice of exogamy prevails. As demonstrated in Table 5, in the northern states of Punjab, Haryana and Uttar Pradesh, women who married in the same village or town numbered barely 5 percent. An even smaller percentage of women from these states reported marrying cousins or close relatives. In contrast to the north, women in the south may not only be encouraged to marry within the natal village, custom may prescribe that marriage to a close cousin or uncle is preferred. In Tamil Nadu, for instance, nearly 27 percent of women married within the same village or town, and about 26 percent of women reported marrying a close relative.

<Table 5 about here>

- 12 -

Gender Performance and Marriage:

If age at marriage is a component of a gender script that views early marriage as a marker of decorum and propriety, we would expect it coincide with other markers of gender performance. Specifically, we highlight the relationship between age at first marriage and the practice of purdah, an eating order during meal times, and restricted mobility. Purdah or ghunghat is probably the most visible marker or public performance of gender. In the IHDS, women responded yes or no to the question, Do you practice ghungat/purdah/pallu? Only about half of women between the ages of 25 and 49 reported they practiced purdah; however, Table 6 demonstrates marked regional variation. Dividing India into northern and southern halves by the Satpura hill range, purdah practice is nearly ubiquitous in the northern states of Rajasthan, Madhya Pradesh and Bihar. Indeed 93 percent of women claimed to practice purdah in Rajasthan and in the northeastern state of Bihar nearly 87 percent of women claimed to practice. In contrast, women in south India practiced purdah far less. In Tamil Nadu only 10 percent of women claimed to practice purdah.

<Table 6 about here>

Slightly less visible to public scrutiny are the behaviors of households associated with meal time; and in some parts of India, a gendered eating order is followed. The IHDS asked women, When your family takes the main meal, do women usually eat with the men? Do women eat first by themselves? Or do men eat first? The options eating together and varies were coded together, while the options women first and men first were coded together. In the northern state of Gujarat only about 4 percent of women reported that their household

- 13 -

practiced an eating order. Thus while there is a less discernable north-south pattern in eating order, the percentage of families who ate separately during meal time was highest at 91 percent in the northern state of Bihar. Neighboring Uttar Pradesh followed with nearly 70 percent of women reporting they practiced an eating order with their families. Finally, restrictions on womens physical mobility is yet another marker of gender segregation where women must seek permission from family elders before leaving the home alone to visit health centers, friends or the local bazaar. For each of these three destinations interviewers asked women, Can you go alone? (yes or no)1. 71 percent of respondents could travel unescorted to the local health center, while 73 percent of respondents could go to a friends home alone. At 78 percent, most women reported being able to travel alone to the local market or kirana shop. Table 6 shows a dichotomous mobility variable, where women were counted as mobile if they could travel alone to all three destinations. 63 percent of all women fit these criteria. The majority of women, 64 percent, reported being able to travel alone to all three destinations, and 14 percent reported they could not travel alone to any of the three destinations. In Bihar, about 30 percent of women reported being able to travel unescorted to all three destinations, and in contrast, 88 percent of women in Northeast reported being similarly mobile. Figures 3, 4 and 5 graph state specific markers of gender performance by age at marriage. As the trend line indicates, the states with greater emphasis on gender performance are also states with lower age at marriage. In results not reported here, we have undertaken multivariate analysis using hierarchical linear models which control for womens age, education, household economic status and place of residence (Desai and Andrist 2008). Even after controlling for these
1

A preceding question asked the respondent whether she needed to acquire permission to travel outside the home. In cases where women reported they did not need to acquire permission, interviewers often failed to ask whether she could go alone to a particular destination. Because it is impossible to know whether women who did not need permission to go out were allowed to travel alone, we have opted to drop these records from the analysis. In total, there were 2,256 such cases.

- 14 -

factors, the district level gender performance indicators seem to be significantly associated with age at marriage. These results suggest that in regions where gender segregation is more prevalent, early marriage is also preferred. This bolsters our argument that for women, early marriage is part of a pattern in which seclusion, segregation and modesty mark claims to refinement and status.

<Figures 3, 4 and 5 about here>

Competing Claims of Gender Performance and Modernity:

In our theoretical discussion we noted that families are faced by competing demands of gender performance and modernity. Whereas status claims based on sanskritization emphasizes behaviors in which modesty and decorum on womens part mark the status of the family; status claims based on modernity dictate emphasis on education and increased protection of childhood. How families resolve this contradiction remains an interesting empirical question. We suggest that a lengthy gap between marriage and cohabitation may be one avenue through which these competing claims may be resolved. Thus one might expect variation in the gap between marriage and gauna to be associated with the level of education a woman reports to have. That is, we are arguing that women who have spent a substantial amount of time in and around educational institutions are likely to extend the gap between marriage and gauna as a means of reconciling the demands of these competing scripts. In what follows we use ordinary least squares regression to analyze the gap between marriage and cohabitation, and how it might vary by the degree to which women pursue an

- 15 -

education2. Thus our principle explanatory variable is the eligible womens education level, broken down into five discreet categories: "Illiterate, 1-6 years, 7-9 years, 10-11 years, and 12 or more years. The category Illiterate is dropped from the model and used as the reference category. We include controls for caste, tribe and religion to mitigate against the conflating effects of differential marriage and gauna patterns associated with these groups. As women living in an urban setting may be more inclined to seize upon and enact the prescriptions of scripts associated with modernity, we include a dichotomous variable indicating whether the eligible woman lived in an urban setting. As we have demonstrated throughout this paper, regional diversity in India is substantial, and we attempt to control for the conflating effects of that diversity by adding state dummies to the model. Finally, because we know age at first marriage is associated with the gap between marriage and gauna, we add age at first marriage as a control variable to the model. After adding controls, the results indicate that higher education is associated with longer gaps between marriage and cohabitation. Note that this is not simply a wealth effect. Other variables measuring socioeconomic status and urban residence do not appear to be correlated with the gap between marriage and gauna once controls are added. It is only higher education that lengthens this gap.

Discussion:
The age in which a daughter or son marries is a pivotal moment of a process and is undoubtedly the result of careful thought and planning. Perhaps the timing of a daughters first marriage is conceived to hinge largely on what is perceived to be the depth of the pool of eligible
The IHDS obtained information about first and second marriages. This analysis is restricted to age at first marriage.
2

- 16 -

bachelors. Perhaps too, as some have argued, the size of the dowry a daughter is able to provide at age 18 as opposed to, say, age 25 is a salient consideration. However, at base, these explanations regard women and their persuasive families as calculative agents, but they are agents seemingly devoid of culture. In this paper, we argue for a different understanding of agency, one which can not be reduced to actors primarily incited by opportunities to maximize profit or hedge against risk. Instead, they are embodied actors, and while they are capable of taking account of the consequences of choosing particular courses of action, they are also subject to the demands of dominant scripts. What happens when the demands associated with one script seem to preclude ones ability to meet the demands of another? If an early marriage reduces a womans ability to complete ever higher levels of education, what then happens when an actor is encouraged to retain respectability through an early marriage, while at the same time, achieve a greater measure of respect through additional investments in education? We have argued that families in India are utilizing and even expanding the space between marriage and gauna as a means of negotiating these competing demands. However, underlying the principle finding that the gap between marriage and gauna is lengthened as a strategy to address the competing claims of scripts, is the notion that families need to be taken seriously as embodied actors but they are also sites where discourse on modernity, gender and even sexuality coalesce and manifest. Such a conception allows one to move away from positing marital timing as strictly the product of simply structural conditions which exert influence over household and daughter, social actors and the family in particular are seen here as both subjects and authors of the social process in which they find themselves immured.

- 17 -

Data presented in this paper pose an interesting paradox. On one hand, they emphasize the continuing relevance of traditional scripts in which gender performance continues to be a significant marker of claims to status and culture. On the other hand they demonstrate that competing claims of modernity lead to unexpected behaviors through which families seek to deal with the onslaught of globalization. We have focused on one aspect of this novel behavior, long gap between marriage and cohabitation as a way of increasing educational attainment.3 Other studies focus on changing marriage arrangements in which arranged marriages coexist with increasing input and participation of brides (and presumably grooms) (Banerji, Martin, and Desai 2008) and increasing cross-region marriages in context of bride-shortages associated with declining sex-ratios (Ravinder Kaur 2004). These observations suggest that marriage patterns in India may well be changing, however, this change may not necessarily involve a movement towards a Western pattern of delayed marriage with dissolution of arranged marriage system. Instead, new forms of marital arrangements may evolve and may deserve attention in future research.

We do not argue that the gap between marriage and cohabitation is novel in itself, just the use of this gap to increase educational attainment. Interestingly, this has also been noted by some activist groups and delayed cohabitation is seen as one of objectives of programs such as Doosra Dashak in Rajasthan.

- 18 -

References:
Banerji, Manjistha; , Steven P.; Martin, and Sonalde Desai. 2008. Is Education associated with a Transition towards Autonomy in Partner Choice? A Case Study of India In IUSSP Seminar on Changing Transitions to Marriage: Gender Implications for the Next Generation. New Delhi. Caldwell, J. C., P.H. Caldwell, B. K. Caldwell, and I. Pieris. 1998. The Construction of Adolescence in a Changing World: Implications for Sexuality, Reproduction, and Marriage. Studies in Family Planning 29 (2, Adolescent Reproductive Behavior in the Developing World):137-53. Caldwell, J. C., P. H. Reddy, and Pat Caldwell. 1983. The Causes of Marriage Change in South India. Population Studies 37 (3):343-361. Chatterjee, Partha. 1989. Colonialism, Nationalism, and Colonialized Women: The Contest in India. American Ethnologist 16 (4):622-33. . 1993. The Nation and Its Fragments: Colonial and Postcolonial Histories. Princeton: Princeton University Press. Derne, Steve. 1994. Hindu Men Talk About Controlling Women: Cultural Ideas as a Tool of the Powerful. Sociological Perspectives 37 (2):203-227. Desai, Sonalde, and Lester Andrist. 2008. Gender Scripts and Age at Marriage in India. In Under Revision Demography. Desai, Sonalde, and Veena Kulkarni. 2008. Changing educational inequalities in India: In the context of affirmative action. Demography 45 (2):245-270. Dube, Leela. 1996. Caste and Women. In Caste: Its Twentieth Century Avatar, edited by M. N. Srinivas. New Delhi: Penguin Publishers. . 2001. Anthropological Explorations in Gender: Intersecting Fields. New Delhi:: Sage Publications. Dyson, Tim, and Mick Moore. 1983. On Kinship Structure, Female Autonomy, and Demographic Behavior in India. Population and Development Review 9 (1):35-60. Goffman, Erving. 1976. Gender Display. Studies in Anthropology of Visual Communications 3 (69-77). Hammel, E. A. 1990. A Theory of Culture for Demography. Population and Development Review 16 (3):455-485. Heimsath, C. H. 1962. The Origin and Enactment of the Indian Age of Consent Bill, 1891. The Journal of Asian Studies 21 (4):491-504. International Institute for Population Sciences, and Macro International. 2007. National Family Health Survey (NFHS-3) 2005-2006, India: Volume 1. Mumbai: IIPS. International Institute for Population, Sciences, and O. R. C. Macro. 2000. National Family and Health Survey (NFHS-2), 1998-99: India. IIPS, Mumbai, India. Islam, M. Nurul, and Ashraf U Ahmed. 1998. Age at First Marriage and its Determinants in Bangladesh (Demographers' Notebook). Asia-Pacific Population Journal 13 (2). Karve, Irawati Karmarkar. 1965. Kinship Organisation in India. Bombay: Asia Publishing House. Kaufman, J. 2004. Endogenous Explanation in the Sociology of Culture. Annual Review of Sociology 30:335-57.

- 19 -

Kishor, Sunita. 2000. Empowerment of Women in Egypt and Links to the Survival and Health of their Children. In Women's Empowerment and Demographic Processes, edited by H. Presser and G. Sen. Oxford: Oxford University Press. Leonard, K. 1976. Women and Social Change in Modern India. Feminist Studies 3 (3-4):117-30. Lindenbaum, S. 1981. Implications for Women of Changing Marriage Transactions in Bangladesh. Studies in Family Planning 12 (11):394-401. Mason, K. O. 1986. The Status of Women: Conceptual and Methodological Issues in Demographic Studies. Sociological Forum 1 (2):284-300. Mendelbaum, D. 1988. Women's Seclusion and Men's Honor: Sex Roles in India. Bangladesh and Pakistan: Tuscon, Arizona: University of Arizona Press. Mensch, Barbara S., Susheela Singh, and John B. Casterline. 2005. Trends in the Timing of First Marriage among Men and Women in the Developing World. The Population Council Working Paper No. 202. for other developing countries. Morris, Rosalind C. 1995. All Made Up: Performance Theory and the New Anthropology of Sex and Gender. Annual Review of Anthropology 24 (1):567-592. Mukhopadhyay, C. C., and P. J. Higgins. 1988. Anthropological Studies of Women's Status Revisited: 1977-1987. Annual Review of Anthropology 17:461-95. Oberoi, Patricia. 1998. Family, Kinship and Marriage in India. Delhi: Oxford India Paperbacks. Papanek, H. 1973. Purdah: Separate Worlds and Symbolic Shelter. Comparative Studies in Society and History 15 (3):289-325. Ravinder Kaur. 2004. Across-Region Marriages: Poverty, Female Migration and the Sex Ratio. Economic and Political Weekly 39 (25):2595. Retherford, Ogawa, and Matsukura. 2001. Late Marriage and Less Marriage in Japan. Population and Development Review 27 (1):65-102. Schein, L. 1999. Performing Modernity. Cultural Anthropology 14 ( 3):361-95. Sharma, Ursula. 1980. Women, Work and Property in North-West India. New York: Tavistock. Srinivas, M. N. 1977. The Changing Position of Indian Women. Man 12 (2):221-238. West, C., and S. Fenstermaker. 1995. Doing Difference. Gender & Society 9 (1):8-37. West, Candace, and Don H. Zimmerman. 1987. Doing Gender. Gender & Society 1 (2):125-151.

- 20 -

Table 1. Singulate Mean Age at Marriage of Females in India and Bangladesh, 1961-1991 age at marriage India* Bangladesh**

1961 16.1 13.9

1971 17.2 --

1981 18.4 16.6

1991 19.3 18.0

sources *International Institute for Population and Macro 2000; **Islam and Ahmed 1998

- 21 -

Table 2. Percent of Married Women Aged 20-24 in Various Developing Regions Eastern/Southern Africa Western/Middle Africa Eastern Asia Former Soivet Asia Caribbean/Central America South America Middle East/North Africa All India Rural India Urban India 66 79 46 54 56 51 55 77 83 63

Source: Mensch, Barbara S., Susheela Singh, and John B. Casterline. 2005. "Trends in the Timing of First Marriage among Men and Women in the Developing World" The Population Council Working Paper No. 202. for other developing countries. Census of India, 2001, Table C-2 for India

- 22 -

Table 3. Marriage and Cohabitation Patterns by Selected Characteristics percent not cohabitating immediately

mean age at marriage All India Woman's Age 25-30 31-35 36-40 41-45 46-49 Woman's Education Illiterate 1-6 standards 7-9 standards 10-11 standards 12 and some college Place of Residence Metro Cities Other Urban area More dev village Less dev village Income Lowest quintile 2nd quintile 3rd quintile 4th quintile Highest quintile Social Groups High Caste Hindu OBC Dalit Adivasi Muslim Other Religion

mean age at cohabitation 17.7

17.2

16.41

17.4 17.1 17.2 17.1 17.3

14.71 15.64 18.09 17.48 18.39

17.8 17.6 17.7 17.6 17.9

15.9 17.1 18.2 19.4 21.6

23.70 11.69 7.41 7.40 3.84

16.6 17.4 18.4 19.6 21.7

19.2 18.4 17.0 16.3

2.94 7.94 16.67 24.38

19.3 18.6 17.5 17.0

16.3 16.6 16.8 17.4 18.8

22.02 19.02 17.95 13.64 9.20

17.0 17.1 17.3 17.9 19.1

18.3 16.9 16.3 16.9 17.1 20.8

10.37 22.35 19.13 13.44 9.13 2.46

18.6 17.6 16.9 17.3 17.3 20.8

- 23 -

Table 3 (contd). Marriage and Cohabitation Patterns by Selected Characteristics States Jammu & Kashmir Himachal Pradesh Uttarakhand Punjab Haryana Delhi Uttar Pradesh Bihar Jharkhand Rajasthan Chhattisgarh Madhya Pradesh Northeast Assam West Bengal Orissa Gujarat Maharashtra, Goa Andhra Pradesh Karnataka Kerala Tamil Nadu

18.8 18.5 17.4 19.6 17.3 19.1 15.7 15.0 17.2 15.2 15.7 15.8 20.4 19.2 17.5 18.0 18.0 18.0 15.9 17.4 20.7 18.6

1.41 6.11 5.19 2.08 16.40 4.11 38.91 64.83 8.54 43.02 31.45 33.33 2.14 1.10 3.31 1.23 9.64 2.79 8.76 5.61 1.29 1.18

18.8 18.7 17.5 19.7 17.9 19.2 17.0 16.4 17.5 16.9 16.5 16.7 20.5 19.3 17.5 18.1 18.4 18.1 16.2 17.7 20.7 18.7

- 24 -

Table 4. Distribution of Marriage Types for Women (25-49) Entering First Marriage at Ages 15-24, by Selected Characteristics Type of Marriage parent- arranged with consent from the respondent 23.18 parent-arranged with no consent from the respondent 35.36

N full sample women's age 25-30 31-35 36-40 41-45 46-49 level of education Illiterate Primary Upper primary Secondary Senior secondary College age at (current) marriage 15-16 years 17-18 years 19-20 years 21-22 years 23-24 years current residence Rural Urban 9,648 3,704 3,002 2,988 1,053 996 5,008 5,047 5,148 3,790 2,621 21,614

selfarranged 4.94

jointlyarranged 36.52

6.25 4.67 4.40 4.64 4.48

35.53 36.08 37.27 36.57 37.74

25.00 25.01 22.23 22.24 19.39

33.23 34.25 36.09 36.55 38.39

4.10 4.79 4.45 6.57 7.24 7.45

33.31 36.54 38.24 39.68 43.31 44.32

15.68 24.33 28.24 34.05 31.74 37.41

46.91 34.34 29.07 19.69 17.71 10.83

7207 7063 4261 1955 1127

4.36 4.04 5.70 6.60 8.59

29.60 38.61 41.40 39.35 44.36

19.64 21.81 25.56 30.46 32.71

46.40 35.54 27.34 23.59 14.34

14,815 6,799

5.07 4.66

34.50 40.93

20.34 29.36

40.09 25.05

- 25 -

Table 4 (contd). Distribution of Marriage Types for Women (25-49) Entering First Marriage at Ages 15-24, by Selected Characteristics states Jammu and Kashmir Himachal Pradesh Uttarakhand Punjab Haryana Delhi Uttar Pradesh Bihar Jharkhand Rajasthan Chattisgarh Madhya Pradesh North-East Assam West Bengal Orissa Gujarat Maharashtra Andhra Pradesh Karnataka Kerala Tamil Nadu

245 159 416 632 419 381 2402 1117 818 825 471 919 243 582 1675 929 1283 2496 1736 1156 825 1886

3.41 6.43 1.28 0.60 2.19 1.46 2.71 2.86 6.56 0.21 1.01 0.69 44.01 7.87 8.21 5.56 9.83 3.11 4.74 4.62 7.16 6.46

16.41 8.90 6.47 36.63 56.60 30.12 21.72 15.24 17.91 15.89 59.71 43.53 19.66 52.25 28.55 19.18 79.86 33.75 30.20 63.75 54.38 52.70

22.83 49.55 35.80 24.52 6.89 31.58 8.44 3.64 12.60 7.09 7.77 8.17 14.09 35.72 41.11 14.40 5.13 34.45 45.19 23.88 37.23 29.04

57.34 35.12 56.45 38.26 34.32 36.85 67.13 78.26 62.93 76.82 31.51 47.61 22.25 4.16 22.12 60.86 5.17 28.70 19.88 7.75 1.23 11.80

source: Banerji, Martin and Desai, 2008

- 26 -

Table 5. Percentage of Marriage to a Relation and Village Endogamy Percent of women married to a close relative or uncle 9.85

Percent of women married within the same village All India States Jammu and Kashmir Himachal Pradesh Uttarakhand Punjab Haryana Delhi Uttar Pradesh Bihar Jharkhand Rajasthan Chattisgarh Madhya Pradesh North-East Assam West Bengal Orissa Gujarat Maharashtra Andhra Pradesh Karnataka Kerala Tamil Nadu coefficient of variation 13.74

23.13 10.53 7.69 4.75 2.86 18.92 5.26 6.13 8.38 10.52 6.88 10.37 41.80 27.37 20.38 17.01 8.30 12.17 16.67 11.65 27.61 27.19 67.93

17.91 0.33 0.63 0.77 1.25 1.41 4.09 5.01 4.99 1.49 0.91 3.20 1.95 0.92 2.96 7.56 2.35 25.59 27.49 22.02 2.07 25.95 138.95

- 27 -

Table 6. Mean of Gender Performativity Variables by Selected Characteristics eating order 0.344 less mobility 0.359

purdah All India Current residence Rural Urban States Jammu and Kashmir Himachal Pradesh Uttarakhand Punjab Haryana Delhi Uttar Pradesh Bihar Jharkhand Rajasthan Chattisgarh Madhya Pradesh North-East Assam West Bengal Orissa Gujarat Maharashtra Andhra Pradesh Karnataka Kerala Tamil Nadu 0.532

0.586 0.404

0.404 0.204

0.400 0.266

0.766 0.447 0.423 0.310 0.799 0.429 0.855 0.874 0.565 0.931 0.554 0.917 0.269 0.668 0.666 0.651 0.743 0.364 0.118 0.116 0.138 0.102

0.201 0.106 0.404 0.231 0.099 0.142 0.701 0.906 0.541 0.426 0.467 0.488 0.048 0.336 0.250 0.601 0.045 0.175 0.074 0.264 0.088 0.147

0.373 0.207 0.238 0.277 0.314 0.248 0.435 0.696 0.577 0.578 0.571 0.513 0.123 0.490 0.276 0.374 0.236 0.171 0.295 0.312 0.141 0.249

- 28 -

Table 7.
Coefficients from Ordinary Least Squared Model Predicting the Gap between Marriage and Cohabitation coefficients Women's Education 1-6 years 7-9 years 10-11 years 12 or more years age at marriage income High Caste Hindu OBC Dalit Adivasi Muslim Sikh, Jain Christain urban States Himachal Pradesh Uttarakhand Punjab Haryana Delhi Uttar Pradesh Bihar Jharkhand Rajasthan Chhattisgarh Madhya Pradesh Northeast Assam West Bengal Orissa Gujarat Maharashtra, Goa Andhra Pradesh Karnataka Kerala Tamil Nadu constant p<= 0.01; p<=0.05 -0.134 -0.394 -0.075 -0.020 -0.088 0.349 0.354 -0.277 0.810 -0.016 0.062 0.066 -0.011 -0.375 -0.380 -0.062 -0.342 -0.565 -0.311 -0.062 -0.337 3.892 0.116 0.091 0.087 0.090 0.094 0.075 0.078 0.082 0.080 0.085 0.080 0.100 0.084 0.076 0.082 0.079 0.076 0.076 0.079 0.083 0.077 0.092 ** ** ** ** ** ** ** ** ** ** ** -0.076 0.040 0.222 0.570 -0.199 0.008 -0.037 0.176 0.022 -0.055 -0.273 -0.015 0.366 -0.035 0.020 0.024 0.030 0.033 0.002 0.006 0.037 0.035 0.037 0.045 0.040 0.078 0.070 0.019 ** ** ** ** ** ** ** std. error

- 29 -

Figure 1.

Average gap between marriage and cohabitaiton by age at marriage

6 5 Years 4 3 2 1 0 10 12 14 16 18 20 25 26 Age at Marriage

- 30 -

Figure 2.

Women's Length of Acquaintance with Husband before Marriage by years of woman's education
100 80 Percent 60 40 20 0 0 years 1-4 std 5-9 std 10-11 std H. Sec College graduate
1-12 month < 1 month Around Wedding Since childhood 1+ years

Years of education

- 31 -

Figure 3.

Age at First Marriage and Purdah by State

percent practicing purdah 100 80 60 40 20 0 12 14 16 age 18 20 22

- 32 -

Figure 4.

Age at First Marriage and Eating Separate by State

percent eating separate 100 80 60 40 20 0 12 14 16 age 18 20 22

- 33 -

Figure 5.

Age at First Marriage and Mobility by State

percent unable to go alone to three destinations 100 80 60 40 20 0 12 14 16 age 18 20 22

- 34 -