Guide on Gender Analysis of Census Data Full Draft of 6 December 2012 Contents


Table 17: Malawi (2008) – SMAM by educational attainment



Download 3.92 Mb.
Page13/32
Date09.08.2018
Size3.92 Mb.
1   ...   9   10   11   12   13   14   15   16   ...   32

Table 17: Malawi (2008) – SMAM by educational attainment


Characteristics

SMAM

Male

Female

No Education

23.0

18.2

Primary

23.0

19.5

Secondary

25.2

22.1

Post-Secondary

28.5

26.2

236. The 2008 census of the DPR of Korea is among the ones that collected data on the age at first marriage. This age is actually quite high in the country, with a mean of 28.4 years for men and 25.5 years for women. Nevertheless, there is some association between the age at first marriage and the level of educational attainment. Only 15.7 per cent of women currently aged 30-39 who married at age 20 or below had post-secondary or higher education. For those who married between age 21 and 24, the percentage was 18.3 per cent; for those married between ages 25 and 29, it was 20.9 per cent; for those married between ages 30 and 34, it was 21.5 per cent and for those married after age 35 it was 22.4 per cent. The corresponding percentages for men currently aged 30-39 were 23.2 per cent, 25.1 per cent, 25.2 per cent, 27.1 per cent and 30.4 per cent, respectively. As was observed earlier, these data do not allow a strictly causal interpretation. It is difficult to imagine, however, that marriage between ages 25 and 29 would have had any direct effect on women’s (or men’s) ability to complete their post-secondary or higher level studies. Therefore, the higher post-secondary or higher education completion rates of those who married later are more likely to have a different explanation, like the lower propensity to marry of those who are still investing in their studies or early post-university careers, perhaps due to the difficulty of saving for marriage while still in the university. The other conclusion that emerges from these data, of course, is that, regardless of the age at first marriage, there was still a gap of about 8 per cent between male and female completion rates of post-secondary or higher education.


IT WOULD BE GOOD TO REPEAT THIS ANALYSIS FOR A COUNTRY WITH MUCH LOWER AGES AT MARRIAGE.
237. Child marriage may result in difficulties entering the labour market, especially where child brides are taking care of their families. A first step to understanding this issue could be to study whether among women of a given age group, for example 25-29 years old who married before 15, or before 20, are economically active (or employed) in lower numbers than those from their cohort who married later. Obviously, this is only possible if the age at first marriage is declared in the census. It is important to conduct the analysis by age group, because behaviours change from one generation to the next due to changes in social norms and beliefs over time. Adding the number of children, presence of children, or educational attainment may also further explain labour market participation. The unemployment rate across women who married as children compared to women who did not marry as children, with similar background characteristics, can then serve as a measure of women’s status in society that can be monitored over time with future census data. In addition to the problems associated with the age at first marriages, one limitation to keep in mind with this analysis is the possible under-reporting of unemployment for young women, as they are more likely than young men to be classified as not economically active.

5. Indicators
238. Proportion of ever-married (and never-married) men and women. Differences in women's and men's behaviour towards marriage can be analysed by looking at the proportion of ever-married women and men and its complement, the proportion of never married women and men. The proportion of ever married and never married women and men, and their trend, can reveal important gender differences. In the situation referred to in Country Example 2, for example, where the sex ratio at birth is considerably imbalanced, the proportion of women and men ever married and never married below a certain age is skewed since men have fewer opportunities to find a spouse as women are less numerous. Polygamy can have this same effect, especially among men of lesser means. Note that, in some countries, it is necessary to take into account consensual unions and same-sex partnerships to have a complete picture.
239. Proportion of men and women in polygamous unions (prevalence and evolution). The proportion of men/women (global or by age group) living in polygamous unions should be interpreted with caution as there is important underreporting of polygamy for women. Often, analysts therefore chose to look at the marital status of men only to determine the prevalence and trend of polygamy. The evolution of polygamy can be studied on the basis of successive censuses, as long as the definitions and questionnaire are comparable. If not, it is possible to estimate the trend by comparing the behaviour of successive generations, bearing in mind the limitations of this approach that were commented on in the Box on Visualization of Spatial Data.
240. Age at first marriage. The Principles and Recommendations (United Nations, 2008 a) do not recommend any tabulations specifically to measure age at first marriage. Where age at first marriage is included as a census question (e.g. Algeria, Azerbaijan, Bermuda, China, Democratic People’s Republic of Korea, Guinea-Bissau, India, Israel, Kazakhstan, Lesotho, Maldives, Malta, Occupied Palestinian Territories, Republic of Korea, St. Lucia, Sudan, Swaziland; the 2000 census of Switzerland asked how long people had been in their current union), misreporting may be widespread, to conceal illegal early marriage. Household surveys are generally believed to be better suited to analysing child marriage, but face the same underreporting limitation. Censuses can provide complementary information on girls and boys, who are under the age of 18 and reported as currently married, at the time of the census. However, in some countries marital status is not collected for household members under the minimum legal marriageable age. The Minimum Gender Indicator Set approved by the UN Statistical Commission in February of 2012 contains one marriage indicator, which can be computed from census data if the relevant question was asked, namely the percentage of women aged 20-24 years old who were married at or in a union before age 18.
241. Singulate Mean Age at Marriage. As noted above, when age at first marriage is not collected in a census, it is advisable to calculate Singulate Mean Age at Marriage as a proxy. The SMAM is the average length of single life, expressed in years, among those who marry before age 50. It is calculated from the proportion of single persons (not including persons separated, divorced or widowed) by age. The main disadvantage of the SMAM in comparison to individual data on age at marriage is that it is an aggregate indicator. It can be broken down by major population groups, but it cannot be related to individual characteristics. The other major limitation of the SMAM is that it does not function well in circumstances where there are a lot of informal unions and where those leaving such unions tend to declare themselves as “single”, rather than “separated”, “divorced” or “widowed”.
Methodology Box 5: Calculating the SMAM
The steps for calculating the Singulate Mean Age at Marriage are the following
The following illustrates the computational steps with the proportion of never-married women by age group, in the 2008 census of Malawi:

15-19 70.6 per cent

20-24 17.4

25-29 6.7

30-34 3.0

35-39 1.9

40-44 1.5

45-49 1.2


Step 1. Calculation of the person years lived in a single state:

15*100+5*70.6+5*17.4+5*6.7+5+5*3.0+5*1.9+5*1.5+5*1.2 = 2004.5 (A).



Step 2. Estimation of the proportion remaining single at age 50: 0.9 per cent.

Step 3. Estimation of the proportion ever marrying by age 50: 99.1 per cent (C).

Step 4. Calculation of the number of person-years lived by the proportion not marrying:

50*0.9=45 (D).

Step 5. Calculation of Singulate Mean Age at Marriage (SMAM):

SMAM = (A - D)/C = (2004.5 - 45)/99.1 = 19.77.

242. In some countries, mostly in Africa, the age at first marriage for women is below 20, such as in Niger, where the SMAM was 17.6 in 2006. At the other end of the spectrum, Northern European countries have the highest age at first marriage for women, as in Sweden where the SMAM was 32.2 in 2006. The analysis of the trend is also necessary to understand the dynamic: in Niger, the age at first marriage for women has increased from 16.2 in 1977 to 17.6 in 2006 (United Nations, 2009 a).
243. Proportion of women married below the legal age, by age. To answer the question of whether child marriage is decreasing in a country or not, a graph can be presented showing the proportion of women married below the legal age, by their current age. If the curve is increasing with age, it means that younger women get married less early than their elders. If only marital status is available in the census questionnaire, it is necessary to combine successive censuses to analyse the trend. Attention should be paid to the comparability of these censuses, in terms of age, marital status reporting (definition and methodology), and coverage.

244. Age at first marriage of women and men, by age group. Comparing the age at first marriage of men and women, by age group or cohort, is important for understanding the scope of early marriage.



_________________________________________________________________________Country_Example_9:_Using_the_Singulate_Mean_Age_at_Marriage_to_Examine_Early_Marriage_in_Malawi'>________________________________________________________________________
Country Example 9: Using the Singulate Mean Age at Marriage to Examine Early Marriage in Malawi
Table 18 shows how the general proportionate SMAM increased over time for all age categories and for both sexes in Malawi. For women, the SMAM increased from 17.8 in 1977 to 19.8 in 2008, while that for males increased from 22.9 in 1977 to 23.9 in 2008, indicating that generally, more women still marry younger than their male counterparts. Consequently, the proportion of the population staying single increased, with that of males aged 15-19 increasing from 93.8 per cent in 1977 to 95.2 per cent in 2008, while that of females aged 15-19 years increased from 48.9 per cent in 1977 to 70.6 per cent in 2008.
Nonetheless, the proportion of females getting married early is still much higher than that of their male counterparts and requires addressing. For example the proportion of females aged 15-19 getting married was 29.4 per cent while that of their male counterparts was 4.8 per cent. Similarly, the proportion of females aged 20-24 years and 25-29 years (17.4 per cent and 6.7 per cent respectively) staying single is still much lower than that of their male counterparts (54 per cent and 21.2 per cent respectively). These trends suggest that girls are still marrying young and strategies to prevent this, such as retaining girls in school, especially secondary school, would go a long way in keeping girls in school longer, increasing the age at which they marry and reducing their fertility rate.
Table 18: Malawi - Proportion single and Singulate Mean Age at Marriage: 1977, 1987, 1998 and 2008


Age Group

Percentage Single

Male

Female

1977

1987

1998

2008

1977

1987

1998

2008

15-19

93.8

91.1

91.7

95.2

48.9

55.1

61.8

70.6

20-24

49.3

51.4

53.0

54.0

7.4

11.5

14.6

17.4

25-29

13.3

17.4

18.0

21.2

2.2

3.5

4.8

6.7

30-34

4.9

6.3

6.0

7.7

1.3

1.6

2.1

3.0

35-39

2.9

3.4

3.4

4.1

1.0

0.9

1.3

1.9

40-44

2.3

2.3

2.6

2.7

1.0

0.8

1.1

1.5

45-49

1.8

1.7

1.7

2.1

0.9

0.7

1.0

1.2

SMAM

22.9

23.2

23.4

23.9

17.8

18.4

19.0

19.8

Source: Malawi. Gender in Malawi. Analytical Report 3 of the 2008 Census: Table 4.5


Table 18 confirms the earlier statement that females generally entered marriage at a younger age than their male counterparts, regardless of residence and educational attainment. It shows that rural women entered marriage 2.1 years earlier than urban women, while women with no education entered marriage 8 years earlier than those who had post secondary education. This suggests that being rural and being uneducated or less educated renders a young woman more vulnerable to early marriage. It confirms previous assertions that education and residence have an impact on a women’s entry into marriage and consequently their fertility.
Table 19: Malawi - Singulate Mean Age at Marriage by residence and educational attainment


Characteristics

SMAM

Male

Female

Residence

Urban

25.9

21.9

Rural

23.4

19.8

Educational Attainment

No Education

23.0

18.2

Primary

23.0

19.5

Secondary

25.2

22.1

Post-Secondary

28.5

26.2

Source: Malawi. Gender in Malawi. Analytical Report 3 of the 2008 Census: Table 4.6



_________________________________________________________________________
245. Analogously to the SMAM, other measures can be defined for the timing of events. It has been suggested, for example, to define a Mean or Median Age at Widowhood and to compare this measure between men and women. Although this measure can be constructed from most census data, it has the following limitations:

1. Women may remarry, so that the mean/median age will be over-stated. In Ireland (2006), for example, the mean age of widowhood for ever-married women was 55.9 years if remarriage is taken into account and 56.2 if not.

2. Most women marry at some point, but a lot of women (and particularly men) never become widow(er)s. In the Ethiopian census of 2007, for example, widowhood in the highest age group (75+) was 62.1 per cent for women and only 11.2 per cent for men.

3. Because widowhood is most prevalent in the very highest age groups, the results will be sensitive to where the cut-off point for the last age group is placed.

4. The concept of widowhood is problematic in contexts where a high proportion of unions are informal.

5. If there is differential mortality of widows, the results will be distorted, especially in the highest age groups.


246. Mean difference in the age at first marriage of the spouses. An indicator that can be derived from the mean age at first marriage (measured directly through the relevant census question or indirectly through the SMAM) is the mean difference in age at first marriage of the spouses. This is relevant from a gender perspective because women who are much younger than their husbands generally have less autonomy and authority in the marital relationship. By and large, differences in age at first marriage between the spouses have been diminishing, but they remain large in some countries in West Africa, such as Mauritania (7.6 years in 2001) and Sierra Leone (6.8 years in 2004) (United Nations, 2009a). The indicator is less adequate in societies in which remarriage is frequent or where polygamy is widespread because it does not measure the age differences in these later unions. In second and third unions or marriages, age differences between the spouses tend to increase as men often remarry with substantially younger wives. Consequently, the gap between the mean ages of husbands and wives tends to widen as they grow older, which increases the probability of widowhood and its economic and social consequences for women, as discussed earlier.
247. A simpler measure to compute is the difference between the mean age of married men and married women. This measure can be compared to the SMAM or to the average age at first marriage if these data are gathered. In the 2008 census of Mozambique, for instance, the SMAM was 18.1 years for women and 22.4 years for men, a difference of 4.3 years. But the average age of women that were married or living in consensual unions was 33.4 years, compared to age 40 for men, which means a difference of 6.6 years (NSO Mozambique website, accessed 8 April 2011). It should be pointed out, however, that these two indicators measure different things. The SMAM only refers to first marriages, but the average age difference of married persons mixes first marriages with remarriages.

Further national level interpretation on this issue can be found in the CEDAW Committee concluding comments for its countries, at

(http://www2.ohchr.org/english/bodies/cedaw/cedaws).

All of these indicators should be analysed by region within a country and by religious/ethnic group, if available, as the prevalence of child marriage will be higher where a culture of gender inequality prevails, as well as in regions prone to conflict or natural disaster.


6. Multivariate and further gender analyses
248. An obvious use of logistic regression is to analyse the marital status of women based on certain explanatory variables such as age, educational attainment and/or literacy of both spouses, religion/ethnicity, and place of residence (rural/urban). Where available (the SMAM will not do in this case), the age at first marriage should also be used. Widowhood, for example, is associated with early marriage, male over-mortality, and social norms regarding remarriage. Where female age at marriage increases, levels of widowhood decline (UNICEF, 2005).
249. Measuring the scope and frequency of early marriage and its trend over time is essential for developing national policies and legislation. In particular, knowing what individual-level characteristics are associated with child marriage may be useful to plan policy interventions to prevent it. In multivariate analysis, age at marriage (where it is available) could then be treated as a dependent variable in order to model the factors that affect age at marriage. Taking this line of analysis, Maitra (2004) finds that ethnicity, religion and parental education all are significantly associated with age at marriage. In a cross-country study with 50 countries, UNICEF (2005) found that the educational level of girls was significantly associated with higher ages at marriage. The spousal age gap was negatively associated with the woman's age at marriage: women more than four years younger than their partners were more likely to be married as children.
250. An excellent example of the use of these kinds of methods for the analysis of marital status comes out of the 2009 census of Viet Nam (Viet Nam, 2011), which performed a series of logistic regressions of different marital status categories. As an illustration, the following reproduces the table with regression coefficients and the comments of the report on the probability of never marriage among population aged 40–69.
251. “In this analysis, based on social norms and distribution of marital status by age in Viet Nam, delayed marriage is defined as the situation of individuals who delay marrying till after the age of 40. The term delayed marriage is used for convenience, but in fact, includes also people who will never marry. In addition, it should be noted that delaying marriage, as defined in this monograph, does not necessarily correspond to the level of SMAM in the population. According to estimates from the Census sample survey data, by the time of the 1999 Census in Viet Nam, there were more than 84,000 males and 371,000 females aged 40 and older who had never been married, accounting for 1.1% and 3.8% of males and females respectively in this age cohort. Ten years later, by the time of the 2009 Census, the corresponding numbers had increased to 210,000 males and 635,000 with the proportions at 1.7% and 4.4%, respectively. The absolute size of the never-married population increased greatly over the past ten years not only because of the increases in the size of the total population but also because of increases in the proportion never-married in the population. Particularly, from 1999 to 2009, the proportions never-married among males aged 40–49 and of both sexes aged 50–59 and 60–69 had all increased. Only the proportion never-married among females aged 40–49 had decreased (from 6.2% to 5.7%), most likely because of the recent decline in the population sex ratio. However, in general, the number and the proportion delaying marriage among females are much higher than among males, reflecting the situation of low sex ratio of the population in Viet Nam in the last several decades.
From the birth cohort perspective, the size of the never-married population has decreased during the period 1999-2009. In 1999, about 58,000 males aged 40–49 were never-married, accounting for 1.6% of the cohort. In 2009, this cohort now aged 50–59 years had only 42,000 never married males, accounting for 1.2% of this birth cohort. The numbers declined not only because of marriage, but also because of mortality and international emigration. However, if mortality and international migration rates are not much different by marital status, the decline of about one third (from 1.6% to1.2%) would be close to the proportion getting married in this cohort over the 10 years between the Censuses. For other cohorts (except the cohort 70+ because of the strong effects of mortality), the probabilities of getting married in the ages 40 and older for males (about 25% after 10 years) are higher than for females (less than 15% after 10 years).
Figure 7: Viet Nam (2009) - Maps of the proportion never married among the popula-tion aged 40 and older by province



Download 3.92 Mb.

Share with your friends:
1   ...   9   10   11   12   13   14   15   16   ...   32




The database is protected by copyright ©sckool.org 2020
send message

    Main page