Delayed marriages in Pakistan.

Sathar, Zeba A. ; Kiani, M. Framurz K.

INTRODUCTION

Delayed marriages played a very important role in slowing down population growth during the European Demographic Transition. Similarly, some developing countries have recently undergone even more rapid changes in marriage patterns, leading to declining levels of fertility. Curtailing marriage or entry into sexual unions is one of the "positive" checks posited by Malthusian theory and is worthy of some renewed attention because of the lack of decline in marital fertility in Pakistan.

Several researchers have identified changes in nuptiality behaviour in Pakistan, in terms of a rise in both the average age at marriage [8; 11; 12] and changes in cohort nuptiality [7]. One researcher observed a slight decline in fertility and attributed it to a rise in the age at marriage in the late Seventies [1], but his observation was found to be an artefact of data and was, therefore, refuted [18]. Thus, nuptiality behaviour has been noted to have changed in Pakistan since the Fifties with no notable accompanying changes in marital fertility. This paper's primary objective is to explore the impact of modernization, particularly of expansion of education and modern sector employment, urbanization and migration, on proportions never married in various age groups.

Data

The primary source of data for this study is the Migration module of the Population, Labour Force and Migration (PLM) Survey of 1979. Supplementary data have also been taken from Pakistan's population censuses of 1961, 1972 and 1981. The main advantage of PLM data is that they provide household characteristics for every member of the more than 10,000 households sampled in Pakistan. The households were chosen on a nationally based sample and, unlike the Census, the PLM Survey contains more details about personal characteristics which can be cross-tabulated and give the analysis greater flexibility than is possible from published census tables.

Whereas previous studies based on the PLM Survey have mainly utilized data from samples of ever-married females, we in this study are using household data. This avoids the selection bias which occurs when using a sample of ever-married females and allows us to identify characteristics of never-married men and women. However, the PLM Survey data may contain biases owing to misreporting of ages which may affect the accuracy of the estimates derived in the paper. In the absence of a thorough evaluation of the quality of PLM Survey data, the similarity of the age--sex distributions in the PLM Survey with the corresponding distributions in the 1981 census and the Pakistan Fertility Survey is somewhat reassuring [19].

CHANGES IN MARRIAGE PATTERNS (1961-1981)

Although marriage still remains universal, there seems to have been, as is evident from Table 1, a noticeable and substantial increase in proportions never married between 1961 and 1972 and subsequently till 1981. The increase in proportions never married is more pronounced for young females aged 15-29 years than for males in the same age-group. The figures for 1972 and 1981 are not very different, indicating that increases in the proportions never married occurred more in the Sixties. The singulate mean age at marriage for females was computed to be 18.1 years in 1961, 19.8 years in 1972 and 20.7 years in 1981.1 Marital postponement for males in the period from 1961 to 1981 was much less pronounced. In the age groups above age 30, the proportions never married were lower in 1981 than in 1961, for both males and females. This may be due to the fact that previously persons who passed the "preferred" age of marriage had less chances of getting married, but with the changes in attitude due to impact of gradual modernization there seems to be a tolerance of late marriage. Thus, with the pattern of later marriages, celibacy levels actually seem to be declining.

It is assumed that whatever the roots of changes in marriage behaviour, they are more likely to occur in metropolitan cities than in smaller cities and towns and least likely to occur in the tradition-bound rural sector. Urban areas have, for this reason, been subdivided into Lahore, Karachi and other urban areas, to study the gradation of changes in nuptiality behaviour. Table 2 shows that the increases in proportion never married, both for males and for females are more pronounced in Lahore and Karachi.

Overall increases in the proportions never married are not so marked in the case of Pakistani males. This may be attributed to the fact that as far back as 1961 male age at marriage was already considerably higher than female age at marriage (23.6 years) and rose to 25 years in 1981. Furthermore, although early marriage is as desirable for sons as for girls, boys, unlike girls, are required to be economically active before they get married--a precondition that has not changed much over time. The slighter changes in proportions married for males over the 1961-81 period, as compared with those for females, have led to a narrowing of the historically large age difference at marriage between the two sexes in Pakistan.

The difference has declined much more in metropolitan cities and to a somewhat lesser extent in other urban and rural areas. This suggests that there is a greater trend for selecting marriage partners closer in age than was the case previously. The possible cause of this may be that girls are more likely now to be able to postpone marrying, or that they can attempt to resist unwanted marriage proposals because of lesser social pressures and greater chances of schooling and employment opportunities. In metropolitan areas, this change is likely to be more pronounced, as there may be increased chances of marriage outside the kin group or biradari. Although, by and large, marriages are still arranged by parents (which reflects the family influence on nuptiality behaviour), the evidence from other studies suggests that some degree of personal involvement in mate choice has begun to occur, particularly in urban areas [3]. Another important factor which indicates a change in marriage behaviour is the decline in celibacy during the 1961-81 period. Although increases in proportion never married usually lead to higher levels of celibacy, both males and females demonstrated, during the 1961-81 period, a decline in proportions who never married at older ages.

Geographic Patterns of Proportion Never Married

Pakistan consists of four provinces, viz. Punjab, Sind, NWFP and Baluchistan, which account for roughly 58 percent, 21 percent, 17 percent and 4 percent of the country's total population, respectively. Patterns of marriage behaviour are expected to vary in these four units because of differences in cultural patterns, levels of development, and urbanization, and are likely to vary substantially from one province to another. Urban-rural variations in nuptiality within each province are thus also of interest.

Table 3 supports the above expectations, showing variation in the proportions never married in four provinces by both urban and rural areas. Once more, proportions never married of males reflect relatively slight differences in all the provinces by urban-rural residence. Punjab, being the most developed province, contains the bulk of the proportion never married in the 15-19 age group, both in urban and in rural areas. Thus Punjabis, as compared with other ethnic groups, seem to be leading the delayed-marriage pattern.

The provincial differential in Proportions never married is much greater for females than for males (Table 3). Punjab has the highest proportions of never married females, followed by the NWFP and Sind, in both urban and rural areas. Earlier studies have shown that female nuptiatity patterns appear to have shown the most marked change in Sind [8], while another study using the 1979 PLM Survey data also showed that after the Punjab, Sind showed the greatest changes in nuptiality patterns [12, p. 20]. Whereas the present study is mainly based on household characteristics of the sample, the previous studies were based on samples of ever-married females and it is interesting, therefore, that our study showed greatest delays in marriage in the Punjab, followed by the NWFP, not by Sind. The 1981 Census also supports the differentials derived from the PLM Survey. A similar exercise done on the 1972 Census supports the above findings in which the NWFP, as compared with other provinces, reflects higher proportions never married.

A comparison of the male and female singulate mean ages at marriage suggests that urban Punjab has the least differences in average ages at marriage between men and women (4.6 years), while in other provinces this difference is more than 5 years. In rural areas of the Punjab, once more, the age difference is the smallest, while in Sind this difference is the greatest.

IMPACTS OF EDUCATION AND URBANIZATION ON DELAYS IN MARRIAGE

The positive association between female education and age at marriage both across and within countries experiencing transition has been strongly supported [4; 6]. Even little schooling has been found to have the effect of delaying marriages well beyond the number of years spent in school and is thought to represent changes in attitudes and values leading to marital postponement. In this section we measure the impact of acquisition of education on postponement of marriage, particularly in the younger age groups, in a largely illiterate society like Pakistan. We also investigate the effect of interaction between education and rural-urban residence on postponement of marriage.

Table 4 demonstrates a clear pattern of higher proportions never married by each level of education for both males and females in urban and rural areas. Tile extent of postponement in marriage is more pronounced among younger educated females in urban areas than among rural females. Although the numbers of persons sampled with 9 or more years of schooling are on the low side in rural areas, there is a distinct relationship between acquisition of education and marriage postponement there as well.

Male nuptiality behaviour also varied by educational levels, though to a lesser extent. Similar differences have been reported by earlier studies for both males and females in Pakistan [10; 12; 13; 20]. Undoubtedly, the impact of each successive level of education is much greater on females than on males. In fact, for men the acquisition of a few years of schooling (1-8 years) has practically no impact on their singulate mean age at marriage.

Another point of interest is that differences in SMAM between men and women of equal educational attainment are greater for the uneducated (5.6 years in urban and 5 years in rural areas) than for the most educated (3.6 years in urban and 1.5 years in rural areas). Those with only 1-8 years of schooling fall in between. Thus there seems to be an inverse association between the difference in SMAM of the two genders by educational level. Since there are more educated men than women in Pakistan, women are more likely to marry men belonging to higher educational group than their own, but, generally speaking, the pattern found seems to be that more educated persons are likely to marry spouses closer in age to themselves. The rather large difference between ages at marriage of men and women may be declining over time as a result of the increasing preference for partners closer in age.

Since most educational facilities are concentrated in large urban centres, it is expected that the impact of education would be compounded by urban residence to create even greater postponements in marriage. However, such a compounding effect is not found (Table 5) and, in many instances, women of equivalent levels of education showed greater postponement of marriage in "other urban areas" than in Karachi and Lahore. For men, the effect of the same level of education manifested itself in generally greater proportions never married in Karachi than in other urban areas. In the case of women, the singulate mean ages at marriage were highest in Karachi for all educational levels except for the category of 1-8 years schooling. This finding is curious, since a much earlier study found highest ages at marriage in Lahore and had attributed it to the city's being a cultural and educational centre [17]. In the case of men, however, Lahore does have generally the highest SMAM across educational groups. We will further explore the independent effects of metropolitan residence and successive levels of education in a later section.

IMPACT OF EMPLOYMENT AND EDUCATIONAL ACTIVITY ON MARRIAGE DELAYS

When males and females are classified by their activity and marital status, we can discern the likelihood of a person's being never married as a function of whether the person is employed, unemployed, involved in housekeeping or attending some educational institution. In Table 6 we have grouped together some categories of activities which contained only very few people (such as disabled persons, those living on rent, etc.) and used the categories of (a) employed, (b) unemployed and doing nothing or living on rent and (c) student, for males; for females the employed and unemployed form one category, while housekeeping and students are the two other major categories.

The findings in this area are most interesting: while in the case of females, particularly in the urban areas, employment is associated with delayed marriage, for employed men, proportions never married are much lower than for other categories up to ages 30-34. Those young men and women who are still involved in educational activity are most likely to be never married. Thus, the impact of current involvement in some educational activity (and at ages beyond 19 this would be synonymous with fairly high levels of education), not surprisingly, is to delay marriages of men and women considerably.

But the impact of pre-nuptial employment in terms of delay in marriage (as seen by the SMAM) is greater than that of educational activity amongst women. The SMAM for females who are employed in urban areas (25.4) is six years higher than that of women involved in housekeeping (20.3). Differences are less pronounced in rural areas. Men who are employed marry, on average, about 3.7 years earlier than those who continue with higher education in urban areas. It is also interesting to note that unemployed men have a higher age at marriage than those already employed. Although it is difficult to deduce cause from effect, it is widely acceptable that men are considered "marriagable", in the Pakistani context, only when they are employed. It is quite possible that if a man wants to be married he is under greater pressure to accept any job, while those men who are not in a hurry to be married can further "afford" to remain unemployed and wait for an adequate job.

Since female labour force participation rates are extremely low in Pakistan, and are even lower in urban (5.2 percent) as compared with rural areas 16 percent [18], it is quite significant that the small minority who do work delay their marriages considerably. It has been well established elsewhere that there is a positive association between economic activity, socio-economic status and delay in marriages [5]. It is quite likely that a large proportion of females who work may be quite well educated [2] and belong to families where they take on work out of choice rather than out of necessity. But, in many cases, never married women who work may be supporting families and siblings and may have to delay marriage, as the family cannot afford to forgo their contribution to household income.

From the fertility module of the PLM Survey, which comprised only ever-married females, it was found that women in professional and clerical occupations married, on average, at 19.6 years, i.e. three-and-a-half years later than those women who worked as agricultural, skilled and unskilled workers (at 16.1 years). Thus, there may be significant differentials in the impact of employment on delays in marriage, depending on whether women take up high-status or low-status occupations. However, at the moment, the numbers of women employed in the modern sector are extremely small.

IMPACT OF INTERNAL MIGRATION AND EMIGRATION ON MARRIAGE DELAYS

The impact of internal migration on marriage delays has also been analysed. The likely impact of internal and external migration on marriage behaviour should be to delay marriages, particularly if male-selective migration causes a shortage of suitable men. However, only very negligible differences between proportions never married of migrants and non-migrants were found and are consequently not presented here.

More detailed and specifically collected data are a prerequisite for analysing the impact over time of migration on delayed marriages. As such, the PLM Survey sample was unable to capture many migrants, and cases in which an out-migrant was the likely head of household were excluded. Thus, a more purposively designed questionnaire and sample than those of the PLM Survey could bring out the very likely impact of migration on nuptiality.

MULTIPLE CLASSIFICATION ANALYSIS

The preceding discussion is mainly a description of patterns of marriage differentials in Pakistan (as measured by proportions never married in various 5-year age groups) by geographic areas, education, migration and activity status. We now turn to discuss whether differentials caused by the above-mentioned characteristics portray independent and significant effects on marriage behaviour. Since changes in marriage behaviour are mainly occurring amongst the younger females aged under 30, we will only consider differentials in the proportions married in these age groups.

Proportions married, the exact converse of proportions never married, by each category of the variables, is the dependent variable in the multiple classification analysis. Thus far we have been discussing some important differentials leading to delayed marriage within certain subgroups, but we now proceed to control the effect of all the other variables and single years of age in each five-year category to see if any statistically significant differentials persist.

Table 7, which pertains to females, clearly shows that educational attainment is the most important factor which is statistically significant in five out of six cases. Large differences in proportions married persist in urban and, to a less extent, also in rural areas by each level of education. Employment is important in the urban areas but not at all in the rural population. This confirms that it is pre-nuptial employment in the modern sector which seems to be associated with deferment of marriage. In rural areas, women work mainly on family farms and this has no significant impact on marriage postponement. Residence in the cities of Lahore and Karachi, particularly in the former, is associated with later marriages, even after the impact of education and employment differences in the urban sector has been controlled. Also, provincial differentials were found to be significant for the two younger age groups in both urban and rural areas. This fact presumably reflects genuine difference in the tempo of first marriages across provinces and reflects that by ages 25-29 the proportions who remain never married are not significantly different. Finally, migrant status, at least as classified in this study, is only significant in the 20-24 age group in urban areas.

CONCLUSIONS

This study mainly underscores the importance attached to education, particularly of females, as an important policy tool for bringing about significant delays in marriage and subsequently in lowering fertility. Pre-nuptial employment of women, particularly in the modern sector, though of small magnitude, is of critical importance in changing attitudes not only towards marriage but, subsequently, towards child-bearing and contraception also.

Education for women and their involvement in paid jobs is also likely to lead to enhancement of women's status and to help in bringing about profound changes in their stereotyped role as wives and mothers. Apart from making efforts for expanding educational and employment opportunities for women, the Government can adopt measures to encourage a higher age at marriage in Pakistan. Such measures would reduce the period of exposure to child-bearing for many women who currently marry much earlier. Also, since a delay in marriage allows some degree of maturity, which is a prerequisite for conscious efforts at birth-spacing and birth control within marriage, it would be supportive of population programme activities.

Efforts such as those made in China (where the average age at marriage has been raised from 18 years for women and 20 years for men in 1959 to 20 years and 22 years by the Eighties [21], are not feasible in Pakistan as the former has a more totalitarian system. However, there is the example of a Muslim country like Tunisia where government-sponsored efforts have raised the age at marriage. Neighbouring Sri Lanka also presents an alternative approach, where no conscious policy has been followed to raise age at marriage, but where women's educational and labour force participation rates are notably high and changes in nuptiality have brought about a sharp decline in fertility.

Comments on "Delayed Marriages in Pakistan"

This study has been skilfully designed and executed both quantitatively and qualitatively. Almost all the data used in this study have been taken from the Population, Labour Force and Migration (PLM) Survey of Pakistan. As a rule, the authors should have devoted some space to the evaluation of the quality of PLM Survey data in this paper.

Since the data utilized for the quantitative measurement belong to the variables of current age and marital status of either sex in the sample population, it is highly likely that they suffer from both age misstatement and reporting errors which tend to vary differently across the regions and are thus bound to affect the derived measures differently.

The sample figures and not the weighted figures of PLM Survey data have been utilized for studying the age pattern of nuptiality, both over time and at a given point in time. Only in one case have the census data been utilized to support the trend in nuptiality. Because of the small number of observations in age groups, especially in the terminal age groups, the levels of singulate mean age at marriage calculated from these observations do not seem to portray the true picture of nuptiality in Pakistan. This phenomenon seems to be prevalent at all levels of the data including the two sexes, rural-urban categories, provincial-level categories, the city-level categories, etc. Moreover, the same phenomenon seems to be prevalent across the two main regions of the country, for all the educational categories by sexes, socio-economic activities by sexes and migration status by sexes. The main drawback of Hajnal technique for deriving the singulate mean age at marriage is that the results are mostly affected by the variation of data given in the terminal age groups. In this study, the starting age group for both males and females has been taken to be the 15-19-year age group. It is a fact that age at marriage in Pakistan is low for females but that for males lies somewhere near the end of the 15-19-year age group. The result is a very high proportion of single males in the 15-19 age group, which prevails in all the regions of the country as well as in all the socio-economic groups of the society. Because of a very high proportion of single males in this age group, the singulate mean age at marriage derived from the relevant age structure shows an upward bias. Under such circumstances, the starting age group for males should be the 20-24 age group whereas for females it should remain the 15-19 age group.

The quantitative outcome of the present study seems to be in complete agreement with the usual phenomenon of nuptiality pattern observed in other societies as well. However, it is necessary to make desirable changes in the quantitative part of the study so that the results from both parts of the study tend to agree completely. There is no denying the fact that the authors have written a paper which is the first of its kind in Pakistan and is also the best in that its approach is unique in the field of social research. The authors deserve congratulations on their remarkable achievement.

M.N. I. Farooqui

National Institute of Population Studies, Islamabad

REFERENCES

[1.] Alam, Iqbal. "Fertility Levels and Trends". In Iqbal Alam and Betzy Dinesen (eds.), Fertility in Pakistan: A Review of Findings from the Pakistan Fertility Survey. Voorburg, Netherlands: International Statistical Institute. 1984.

[2.] Azhar, Mohammad. "Differentials in Cumulative Fertility by Education of Mothers". Islamabad: Pakistan Institute of Development Economics. 1980. (Research Report Series, No. 112)

[3.] Cheung, Paul et al. "Cultural Variations in the Transition to Marriage in Four Asian Societies". IUSSP International Population Conference. Florence. Vol. 3. 1985.

[4.] Cochrane, Susan H. Education and Fertility: What Do We Really Know? The Johns Hopkins University Press. 1979.

[5.] Dixon, R.B. "Explaining Cross Cultural Variations in Age of Marriage and Proportions Never Marrying". Population Studies. Vol. 25, No. 2. 1971.

[6.] Dixon, R. B. "The Social and Demographic Determinants of Marital Postponement and Celibacy: A Comparative Study". Ph.D. Dissertations, University of California, Berkeley (USA). 1970. (Unpublished)

[7.] Farid, Samir M. "Recent Transitions in Cohort Nuptiality". In Iqbal Alam and Betzy Dinesen (eds.), Fertility in Pakistan: A Review of Findings from the Pakistan Fertility Survey. Voorburg, Netherlands: International Statistical Institute. 1984.

[8.] Karim, S. Mehtab. "Differentials in Age at First Marriage". In Iqbal Alam and Betzy Dinesen (eds.), Fertility in Pakistan: A Review of Findings from Pakistan Fertility Survey. Voorburg, Netherlands: International Statistical Institute. 1984.

[9.] Karim, S. Mehtab. "socio-Economic and Cultural Aspects of Marriage and Fertility in Urban Pakistan". Honolulu, Hawaii: East-West Center. December 1979. (Papers of East-West Population Institute No. 64)

[10.] Kiani, Mohammad Framurz, K. "Patterns of Female Nuptiality and their Implications for Fertility in Pakistan". M. Phil. Thesis, Cairo Demographic Centre, Cairo. 1984. (Unpublished)

[11.] Korson, J. Henry, and M. A. Sabzwari. "Age and Social State at Marriage, Karachi, Pakistan 1961-64 and 1980: A Comparative Study". Journal of Comparative Studies. Vol. 13, No. 2. 1984.

[12.] Mahmood, N., and S. Mubashir Ali. "Nuptiality Patterns in Pakistan". Islamabad: Pakistan Institute of Development Economics. 1983. (Studies in Population, Labour Force and Migration Project Report No. 2)

[13.] Mahmood, N., and Zubeda Khan. "Literacy Transition and Female Nuptiality: Implications for Fertility". Pakistan Development Review. Vol. XXIII, Nos. 3&4. Autumn-Winter 1985.

[14.] Pakistan. Population Census Organization. "Population Census of Pakistan 1961: Age-Sex and Marital Status". Census Bulletin. No. 3. Karachi. 1963.

[15.] Pakistan. Population Census Organization. Statistical Report of Punjab, Population Census of Pakistan 1972. Islamabad. n.d.

[16.] Pakistan. Population Census Organization. 1981 Census Report of Pakistan. Islamabad. 1984.

[17.] Sadiq, Naseem M. "Estimation of Nuptiality and Its Analysis from Census Data of Pakistan". Pakistan Development Review. Vol. II, No. 4. 1965.

[18.] Sathar, Zeba A., and Mohammad Irfan. "Reproductive Behaviour in Pakistan: Insights from Population Labour Force and Migration Survey 1979-80". Pakistan Development Review. Vol. XXIII, Nos. 2&3. Summer-Autumn 1984.

[19.] Sathar, Zeba A., S. Mubashir Ali and G. M. Zahid. "Socio-economic Characteristics of the Population in Pakistan". Islamabad: Pakistan Institute of Development Economics. 1984. (Studies in Population, Labour Force and Migration Survey Project Report No. 8)

[20.] Smith, Peter C. "Tradition and Transition in South, South-East and West Asian Nuptiality". Paper prepared for Population Association of America, East-West Population Institute, Hawaii. 1977.

[21.] Wei, Jin. "Dynamics of Nuptiality in a Chinese Community: A Study in Suburban Beijing". M. Phil. Thesis, Cairo Demographic Centre, Cairo. 1982. (Unpublished)

(1) Hajnal's Singulate Mean Age at Marriage (SMAM) utilizes percent Never Married, using a single decrement procedure, the implied average number of years of never married life for a hypothetical cohort. It yields mean ages which typically differ from the means obtained from retrospective survey responses, which reflect the age compositions of the populations at risk. In a situation in which age at marriage is increasing, the SMAM will be greater than means based on direct responses.

ZEBA A. SATHAR and M. FRAMURZ K. KIANI *

* The authors are, respectively, Senior Research Demographer and Research Demographer at the Pakistan Institute of Development Economics, Islamabad. They gratefully acknowledge useful suggestions made on an earlier version of the paper by Dr Sultan S. Hashmi and Mr M. N. I. Farooqui. They are also grateful to Mr Niaz Hussain for computing assistance and to Mr Naseer and Mr Shamsi for typing the paper.

Table 1
Proportions Never Married by Sex and Current
Age for 1961, 1972 and 1981 Censuses

 Current Age Groups
 Census
 Sex Year 15-19 20-24 25-29 30-34

Females
 1961 46.6 12.0 5.1 3.0
 1972 65.6 21.3 7.2 3.6
 1981 68.9 24.4 7.9 3.7
Males
 1961 83.6 52.9 27.9 14.1
 1972 92.6 67.8 36.0 17.4
 1981 92.5 64.0 32.8 15.8

 Current Age Groups Singulate
 Census Mean Age at
 Sex Year 35-39 40-44 45-49 Marriage

Females
 1961 2.7 2.0 1.9 18.1
 1972 2.1 1.5 0.9 19.8
 1981 1.8 1.7 1.1 20.7
Males
 1961 8.8 6.4 5.2 23.6
 1972 0.9 0.6 4.3 24.9
 1981 8.1 6.3 3.9 25.0

Sources: For 1961, [14].
For 1972, [15].
For 1981, [16].

Note: SMAM calculated assuming 100 percent
proportions never married before age 15.

Table 2
Proportions Never Married of Females and Males by Current Age for
Lahore, Karachi, other Urban and Rural Areas from the 1961
and 1981 Censuses

 Current Age Groups
Area/
Year 15-19 20-24 25-29 30-34 35-39

FEMALES

Lahore
 1961 63.0 17.9 4.7 2.4 1.9
 1981 84.7 38.2 11.0 3.7 1.6
 Difference 21.7 20.3 6.3 1.3 -0.3
Karachi
 1961 50.4 13.9 5.0 2.8 2.5
 1981 81.1 36.7 13.1 5.3 1.9
 Difference 30.7 22.8 8.1 2.5 -0.6
Other Urban
 1961 51.0 14.7 6.4 3.6 3.2
 1981 74.9 30.7 9.9 4.2 1.8
 Difference 23.9 16.0 3.5 0.6 1.4
Rural
 1961 44.5 10.9 4.9 2.9 2.6
 1981 67.1 23.2 7.8 3.7 1.7
 Difference 22.6 12.3 2.9 0.8 -0.9

MALES

Lahore
 1961 92.9 63.8 29.8 13.2 7.5
 1981 97.2 75.8 37.2 13.3 5.4
 Difference 4.3 12.0 7.4 0.1 -2.1
Karachi
 1961 92.1 62.1 35.6 17.4 11.2
 1981 96.7 75.6 39.1 15.7 7.0
 Difference 4.6 13.5 3.5 1.7 -4.2
Other Urban
 1961 85.3 55.5 27.6 13.2 9.1
 1981 94.0 66.9 31.6 13.1 6.1
 Difference 8.7 11.4 4.0 0.1 -3.0
Rural
 1961 82.1 50.4 27.1 14.2 8.7
 1981 91.2 61.5 29.9 13.9 6.2
 Difference 9.1 11.1 2.8 -0.3 -2.5

 Current Age Groups Singulate
Area/ Mean Age at
Year 40-44 45-49 Marriage

FEMALES

Lahore
 1961 1.1 0.9 19.4
 1981 1.4 0.9 21.6
 Difference 0.3 0 2.2
Karachi
 1961 1.4 1.3 18.5
 1981 1.7 1.1 21.5
 Difference 0.3 -0.2 3.0
Other Urban
 1961 3.2 2.6 18.4
 1981 1.7 1.1 20.8
 Difference -1.5 -1.5 2.4
Rural
 1961 2.0 1.9 17.9
 1981 1.6 0.9 19.8
 Difference -0.4 -1.0 1.9

MALES

Lahore
 1961 5.2 3.8 24.9
 1981 3.6 2.4 26.1
 Difference -1.6 -1.4 1.2
Karachi
 1961 6.5 4.4 25.5
 1981 4.5 2.8 26.4
 Difference -2.0 -1.6 0.9
Other Urban
 1961 6.7 5.9 23.8
 1981 4.4 2.7 25.2
 Difference -2.3 -3.2 1.4
Rural
 1961 6.4 5.1 23.5
 1981 4.5 2.4 24.8
 Difference -1.9 -2.7 1.3

Sources: For 1961, [14].

For 1981, [161.

Notes: 1. SMAM has been calculated assuming 100 percent
proportion never married before age 15.

2. Proportions never married for Lahore and Karachi
1981 have been taken from District Census Reports.

Table 3

Male and Female Proportions Never Married by Province and Current Age
for Urban and Rural Areas

 Current Age Groups

Area 15-19 20-24 25-29 30-34 35-39

FEMALE

Urban
 Punjab 82.6 38.3 8.3 2.4 1.3
 (691) (522) (108) (337) (311)
 Sind 80.2 34.4 9.1 2.8 1.5
 (449) (358) (27) (213) (198)
 NWFP 80.3 35.1 14.4 9.9 8.7
 (127) (97) (13) (71) (46)
 Baluchistan 74.6 26.2 5.2 7.0 3.6
 (59) (76) (3) (43) (28)
Rural
 Punjab 74.4 22.0 6.4 2.1 2.2
 (1075) (914) (762) (708) (603)
 Sind 44.7 12.8 4.1 2.1 0
 (235) (281) (244) (192) (145)
 NWFP 62.7 23.3 7.8 3.5 3.9
 (233) (189) (179) (142) (127)
 Baluchistan 98.1 11.1 4.9 0 0
 (52) (63) (61) (45) (45)

MALE

Urban
 Punjab 9.5 75.1 38.9 12.9 6.1
 (734) (551) (378) (326) (330)
 Sind 97.2 73.8 37.0 14.3 5.1
 (468) (413) (289) (230) (216)
 NWFP 97.4 77.6 47.1 14.3 10.3
 (151) (116) (40) (63) (58)
 Baluchistan 97.3 75.8 23.6 10.6 12.5
 (75) (62) (13) (47) (40)
Rural
 Punjab 94.7 62.2 30.3 11.1 6.7
 (1180) (813) (690) (637) (623)
 Sind 87.6 61.0 33.9 9.2 8.7
 (291) (254) (292) (228) (173)
 NWFP 94.2 69.9 27.6 10.6 6.0
 (223) (153) (123) (123) (83)
 Baluchistan 92.6 62.0 28.8 12.2 2.7
 (81) (50) (66) (49) (37)

 Current Age Groups Singulate
 Mean Age at
Area 40-44 45-49 Marriage

FEMALE

Urban
 Punjab 2.5 1.6 21.5
 (277) (258)
 Sind 0.6 0.0 21.3
 (177) (145)
 NWFP 6.4 3.3 21.6
 (47) (61)
 Baluchistan 5.6 0 19.7
 (36) (19)
Rural
 Punjab 0.4 0.6 20.2
 (570) (525)
 Sind 0.8 0 17.8
 (123) (124)
 NWFP 4.0 2.1 19.8
 (124) (94)
 Baluchistan 0 0 17.6
 (28) (22)

MALE

Urban
 Punjab 4.0 1.9 26.2
 (274) (266)
 Sind 5.5 1.1 26.4
 (182) (184)
 NWFP 3.3 2.9 27.1
 (61) (34)
 Baluchistan 2.9 5.1 25.1
 (34) (39)
Rural
 Punjab 2.7 2.3 24.9
 (548) (526)
 Sind 4.7 2.8 24.8
 (149) (141)
 NWFP 1.1 1.0 25.2
 (95) (100)
 Baluchistan 2.4 0 24.4
 (42) (37)

Source: Population Labour Force and
Migration Survey 1979-80.

Note. Figures in parentheses represent the number of
cases upon which percentages are based.

Table 4 Male and Female Proportions Never Married by Years
of Schooling, Current Age for Urban and Rural Residence

 Current Age Groups
Residence/
Schooling 15-19 20-24 25-29 30-34

 FEMALES

Urban
 No Schooling 67.3 19.3 4.8 2.8
 (563) (522) (564) (458)

 1-8 Years of 89.1 34.5 10.9 3.5
 Schooling (376) (246) (137) (115)

 9+ Years of 93.8 67.0 22.9 7.7
 Schooling (387) (285) (153) (91)

Rural
 No Schooling 64.8 17.9 5.6 2.0
 (1362) (1292) (1148) (1028)

 1-8 Years of 82.5 28.6 10.6 6.1
 Schooling (206) (126) (85) (49)

 9+ Years of 88.9 72.4 23.1 0.0
 Schooling (27) (29) (13) (10)

 MALES

Urban
 No Schooling 95.6 62.1 34.3 13.5
 (344) (311) (248) (245)

 1-8 Years of 98.6 68.5 31.4 10.6
 Schooling (512) (302) (210) (189)

 9+ Years of 98.6 86.2 44.7 15.5
 Schooling (572) (529) (349) (232)

Rural
 No Schooling 92.5 61.9 31.3 11.2
 (902) (700) (720) (680)

 1-8 Years of 93.7 59.5 27.5 8.7
 Schooling (560) (353) (298) (252)

 9+ Years of 95.5 71.9 35.3 12.4
 Schooling (313) (217) (153) (105)

 Current Age Groups Singulate
Residence/ Mean Age at
Schooling 35-44 45+ Marriage

 FEMALES

Urban
 No Schooling 1.8 1.4 19.5
 (846) (1593)

 1-8 Years of 1.1 0.6 21.8
 Schooling (185) (167)

 9+ Years of 9.0 6.5 23.6
 Schooling (89) (46)

Rural
 No Schooling 1.4 0.7 19.8
 (1692) (2947)

 1-8 Years of 1.6 0 21.1
 Schooling (63) (40)

 9+ Years of 0.2 0 24.2
 Schooling (10) (14)

 MALES

Urban
 No Schooling 6.9 2.5 25.1
 (481) (1166)

 1-8 Years of 6.1 1.5 25.5
 Schooling (310) (538)

 9+ Years of 3.5 1.2 27.2
 Schooling (404) (402)

Rural
 No Schooling 5.5 1.7 24.8
 (1215) (2968)

 1-8 Years of 3.6 1.1 24.4
 Schooling (417) (539)

 9+ Years of 4.2 1.3 25.7
 Schooling (118) (77)

Source: Population Labour Force and Migration Survey 1979-80.

Notes: (1.) Figures in parentheses represent total number
of cases upon which percentages are based.

Table 5

Female and Male Proportions Never Married by Years of Schooling
by Current Age for Lahore, Karachi and other Urban Areas

 Current/Age Groups
Urban Area/
Schooling 15-19 20-24 25-29 30-34

 FEMALES

Lahore
 No Education 81.0 9.1 11.1 0
 (42) (44) (54) (38)
 1-8 Years of 87.3 53.1 0 0
 Schooling (63) (32) (14) (14)
 9+ Years of 94.1 70.0 26.7 0
 Schooling (51) (30) (15) (13)

Karachi
 No Education 73.0 21.1 6.5 3.6
 (100) (90) (92) (83)
 1-8 Years of 89.9 42.0 25.9 0
 Schooling (89) (50) (27) (23)
 9+ Years of 95.5 58.7 18.0 10.3
 Schooling (111) (80) (50) (29)

Other Urban
 No Education 64.4 20.1 3.6 3.0
 (421) (388) (418) (337)
 1-8 Years of 89.3 28.6 8.3 5.1
 Schooling (224) (164) (96) (78)
 9+ Years of 92.9 70.3 25.0 8.2
 Schooling (225) (175) (88) (49)

 MALES

Lahore
 No Education 100 70.0 31.6 13.8
 (26) (30) (19) (29)
 1-8 Years of 100 69.4 42.1 15.8
 Schooling (839) (36) (19) (19)
 9+ Years of 100 98.9 61.8 15.4
 Schooling (61) (57) (34) (26)

Karachi
 No Education 100 64.1 41.9 5.7
 (56) (64) (43) (35)
 1-8 Years of 99.1 80.1 40.4 6.9
 Schooling (109) (63) (52) (29)
 9+ Years of 100 87.3 47.7 24.6
 Schooling (119) (134) (86) (61)

Other Urban
 No Education 94.3 60.4 32s8 14.9
 (262) (217) (186) (181)
 1-8 Years of 98.3 64.5 26.6 10.6
 Schooling (364) (203) (139) (141)
 9+ Years of 97.7 84.6 41.0 11.7
 Schooling (392) (338) (229) (145)

 Current/Age Groups Singulate
Urban Area/ Mean Age at
Schooling 35-44 45+ Marriage

 FEMALES

Lahore
 No Education 0 0 18.8
 (84) (38)
 1-8 Years of 0 0 18.6
 Schooling (23) (9)
 9+ Years of 0 0 22.5
 Schooling (9) (2)

Karachi
 No Education 0.6 0 19.9
 (171) (69)
 1-8 Years of 0 0 20.4
 Schooling (48) (14)
 9+ Years of 7.7 0 22.8
 Schooling (26) (5)

Other Urban
 No Education 2.4 1.4 19.4
 (591) (293)
 1-8 Years of 1.7 0 21.2
 Schooling (114) (45)
 9+ Years of 11.1 14.3 22.1
 Schooling (54) (14)

 MALES

Lahore
 No Education 2.1 0 24.3
 (48) (76)
 1-8 Years of 6.9 0 25.4
 Schooling (29) (16)
 9+ Years of 5.4 3.8 28.1
 Schooling (37) (26)

Karachi
 No Education 6.4 4.0 25.2
 (109) (50)
 1-8 Years of 4.8 0 25.6
 Schooling (62) (30)
 9+ Years of 4.4 0 27.4
 Schooling (90) (41)

Other Urban
 No Education 7.7 1.9 25.5
 (324) (159)
 1-8 Years of 6.4 3.7 24.6
 Schooling (219) (107)
 9+ Years of 2.9 0 26.3
 Schooling (277) (78)

Source: Population Labour Force and Migration Survey 1979-80.
Notes: (1.) Figures in parentheses represent total number
of cases upon which percentages are based.
(2.) 45+ includes 45-85 age groups.

Table 6

Proportions, Never Married of Females and Males by Activity
Status and Current Age for Urban and Rural Residence

 Current Age Groups
Residence/
Activity Status 15-19 20-24 25-29 30-34

 FEMALES
Urban
 Employed + 87.1 78.0 35.6 18.2
 Unemployed (31) (59) (45) (33)

 Housekeeping + 73.4 28.8 7.1 2.8
 Others (919) (93) (803) (631)

 Students 99.7 98.4 66.7 --
 (376) (64) (6)
Rural
 Employed + 78.8 32.0 11.4 0
 Unemployed (52) (25) (35) (34)

 Housekeeping + 66.1 19.5 5.9 2.3
 Others (1494) (1417) (1211) (1051)

 Students 97.9 80.0 -- --
 (47) (5)

 MALES

Urban
 Employed 96.9 69.5 36.8 12.1
 (652) (886) (769) (638)

 Unemployed + 100 81.2 25.0 26.7
 Others (28) (11) (8) (15)

 Students 99.2 95.1 1000.0 50.0
 (664) (183) (11) (2)
Rural
 Employed 91.9 60.8 29.8 10.5
 (1277) (1149) (1137) (1010)

 Unemployed + 96.1 77.4 55.5 19.2
 Others (102) (62) (27) (26)

 Students 97.2 89.5 72.7 0
 (395) (57) (11) (1)

 Current Age Groups Singulate
Residence/ Mean Age at
Activity Status 35-44 45+ Marriage

 FEMALES

Urban
 Employed + 18.7 2.5 26.4
 Unemployed (48) (75)

 Housekeeping + 1.5 1.4 20.3
 Others (1072) (1718)

 Students -- -- 24.7

Rural
 Employed + 0 3.1 20.3
 Unemployed (63) (65)

 Housekeeping + 20.0 0.7 20.5
 Others (5) (2923)

 Students -- 0
 (1)

 MALES

Urban
 Employed 5.2 1.5 25.7
 (1172) -1731

 Unemployed + 29.4 4.2 27.4
 Others (17) (357)

 Students 0 -- 29.4
 (1)
Rural
 Employed 4.8 1.3 24.6
 (1716) (2993)

 Unemployed + 14.7 2.8 27.7
 Others (34) (596)

 Students 33.3 100 21.9
 (6) (2)

Source: Population Labour Force and Migration Survey 1979-80.
Notes: (1.) Figures in parentheses represent total number
of cases upon which percentages are based.

Table 7
Female Proportions Ever Married Adjusted by Multiple Classification
Analysis for Educational Level, Activity, Migration Status, Residence
and Province with Age (in Single Years) as a Covariate

 Activity
 Overall
Ages Mean Employed Housekeeping Student

URBAN

15-19 13.7 6.4 15.1 10.8
20-24 58.4 29.5 * 62.2 * 21.1 *
25-29 88.9 65.5 * 90.6 * 52.0 *

RURAL

15-19 25.9 18.3 26.3 21.8
20-24 74.1 72.0 74.3 51.6
25-29 92.9 86.5 93.1 --

 Education (Years of Schooling)

Ages 0 1-4 5-8 9-10 11-12

URBAN

15-19 23.2 * 9.8 * 8.9 * 7.0 * 1.0 *
20-24 71.4 * 56.2 * 52.3 * 4.0 * 34.6 *
25-29 92.7 * 83.9 * 87.5 * 81.5 * 70.8 *

RURAL

15-19 27.0 24.8 19.8 11.4 (0.00)
20-24 76.7 * 82.8 * 53.0 * 10.4 * 33.7 *
25-29 93.6 * 99.5 * 78.3 * 40.3 * 100.0 *

 Residence

 Other
Ages Lahore Karachi Urban

URBAN

15-19 15.5 * 3.7 * 16.9 *
20-24 63.5 54.1 57.4
25-29 84.5 * 84.2 * 90.9 *

RURAL

15-19
20-24 Not Applicable
25-29

 Province

Ages Punjab Sind NWFP Baluchistan

URBAN

15-19 7.4 * 23.4 * 8.1 * 18.7 *
20-24 48.8 * 66.1 * 58.1 * 63.9 *
25-29 86.8 92.7 84.3 88.7

RURAL

15-19 17.7 * 51.7 * 29.5 * 47.2 *
20-24 68.4 * 86.5 * 69.0 89.2 *
25-29 91.8 96.1 91.2 95.2

 Migration

Ages Migrant Non-Migrant [R.sup.2]

URBAN

15-19 14.9 13.6 0.194
20-24 68.9 57.0 0.268
25-29 95.1 88.4 0.144

RURAL

15-19 18.9 26.3 0.243
20-24 76.2 74.0 0.119
25-29 95.7 92.7 0.047

(1.) Age is statistically significant in all the categories.

(2.) () based on less than 10 cases.

(3.) * Indicates that the variable is statistically significant
at less than .05 level in that category.