SLIDE 1 1
XXVIII IUSSP International Population Conference Paper ‘Childlessness’ in Colombia: Exploring the trends, associated factors, and intentionality
- f non-parenthood since the 1980s
Cristina Pérez Department of Anthropology, University College London (UK) ABSTRACT Between 1965 and 2015, Colombia’s total fertility rate (TFR) declined by 5 children per woman, reaching the below-replacement level of 2.0. Colombian women now outperform men at every level of education, and between 1985 and 2005, female labour-force participation rose faster than in any other country in the region. Against a backdrop of profound sociodemographic change, this paper presents findings from the initial quantitative analysis of a two-part mixed methods study that aims to improve our understanding of childlessness beyond infertility within Latin America. Using data from a series of Colombian Demographic and Health Surveys and censuses, this paper explores the trends in Colombian childlessness from the 1980s to the present and asks who is most likely to be childless, how this changes across the life course, and whether this is ‘voluntary’ or not. Using descriptive statistics and generalised linear modelling, it explores the factors associated with female childlessness (as compared to motherhood) around ages 30 and 40. Partnership status (never married), higher education, and higher socioeconomic status were strongly and positively associated with female childlessness, and ‘voluntary’ forms of childlessness, in particular, have increased steadily over the past 30 years. INTRODUCTION & BACKGROUND In a pattern not uncommon for Latin America, between 1965 and 2010, Colombia experienced a ‘spectacular’ fertility decline, as the total fertility rate (TFR) fell from 7 children per woman to 2.1 (Ojeda, Ordonez and Ochoa, 2011). As of 2010, approximately half (49.1%) of the Colombian population resided in regions with at or below-replacement level fertility, and the most recent Demographic and Health Survey estimated that the national TFR in 2015 was 2.0. The country is now in an advanced stage of the ‘First’ Demographic Transition (FDT), and questions around whether or not Colombia has started to
SLIDE 2 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
2
exhibit features of the so-called ‘Second’ Demographic Transition (SDT) have begun to emerge (Florez and Sanchez, 2013). The SDT in Europe is broadly characterised by: sub-replacement fertility related to postponement of childbearing and increasing childlessness; rising ages at marriage because of increased female choice and autonomy; and rising premarital cohabitation (Lesthaeghe, 2010). Many SDT theorists attribute these changes to ‘modern’ forms of social organisation and ‘post-materialist values’, as women (and men) prioritise their own self-fulfilment in a context of increasing urbanisation, education and gender equality (Arriagada, 2002). There is some evidence that this is the case in Colombia: women now outperform men at every level
- f education, and between 1985 and 2005 their labour force participation rose faster than in
any other country in the region (Amador, Bernal and Pena, 2013). The same period saw sharp moves away from marriage and towards ‘consensual unions’ (Esteve, Lesthaeghe and López- Gay, 2012), alongside some evidence of delayed childbearing (Rosero-Bixby, Castro Martin and Martin-Garcia, 2009). However, most Colombian women will have had their first child by the time they reach their mid-twenties, and little is known about the levels, patterns, and factors associated with childlessness in Colombia. Figures from the 2005 census indicate that more than 10% of Colombian women aged 45-49 were childless, which, although low compared to most European countries, is in line with estimates from Brazil (Cavenaghi, 2013). Cavenaghi and Alves (2013), who examined childlessness in a Brazilian context, note that low fertility is unevenly spread across that country according to stratifying factors such as education, socioeconomic status, and region/area of residence. This stratification owes to the large-scale socioeconomic inequalities that characterise Brazilian society, in common with most other Latin American countries including Colombia. Although some sectors of Latin American society may have
- nly recently experienced the changes associated with the FDT, other subgroups may be well
- n their way to exhibiting features of the SDT.
Recent Fertility Trends in Colombia As of 2015, almost two-thirds (65.4%) of the Colombian population resided in regions with at
- r below-replacement level fertility, an increase from just 16.7% in 2005. When looking
- nly at below-replacement level TFRs (less than 2.1), the proportions in 2015, 2010 and 2005
were 58.1%, 44.5%, and 15.3% of the country’s population, respectively. This change can be seen clearly in Figure 1 and in (Appendix) Table A1.
SLIDE 3 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
3 Figure 1: Maps of Colombian Total Fertility Rate (TFR), by Department from 2005-2015.
2015 2010 2005
Source: Author’s own elaboration using DHS data.
Figure 1 also illustrates the relatively wide regional variation, as (increasingly few) departments have TFRs above 2.5. In the 2015 DHS the highest departmental TFR (of 4.6) was registered in Vaupés, a sparsely-populated department in the south-eastern Amazonian region, or ‘National Territories’ bordering Brazil, which accounted for less than 1% of the national population. The department with the lowest overall TFR (of 1.3) was Caldas, in the ‘Central’ Andean region, which borders the larger department of Antioquia (with a TFR of 1.4), which is home to Medellin, Colombia’s second city after Bogota. Historically, the Atlantic coast, Eastern region and National Territories (Amazonia/Orinoquia) tend towards higher TFRs, while the ‘Central’ Andean region and Bogota typically have lower fertility.
Table 1: Trends in TFR by Colombian Region, DHS 1986-2015 Region 1986 1990 1995 2000 2005 2010 2015 Atlantic 4.3 3.6 3.2 2.7 2.7 2.6 2.5 Eastern 3.5 2.8 3.3 2.8 2.6 2.2 2.0 Central 2.9 2.5 2.8 2.5 2.3 1.9 1.6 Pacific 2.9 2.8 3.0 2.6 2.3 2 1.9 Bogotá 2.6 2.4 2.5 2.4 2.2 1.9 1.8 National Territories*
2.5 2.2 Total 3.2 2.8 3.0 2.6 2.4 2.1 2.0
*The first four DHS rounds, 1986-2000, excluded the National Territories, consisting mainly of the Amazonian region, on the basis of being sparsely populated and accounting for around/less than 2.5% of the total Colombian population.
SLIDE 4 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
4
As Table 1 and Figure 2 both illustrate, there has been a steady downward trend over time in both national and regional TFRs, with below-replacement levels emerging in these larger regions from 2010 onward. By 2015, four of the six regions had below-replacement level fertility, excluding only the Atlantic coast and National Territories.
Figure 2: Regional Trends in the Total Fertility Rate (TFR), DHS 1986-2015
Against this backdrop of pronounced social and demographic changes, this paper will present the initial, quantitative findings from a multi-stage mixed methods research project focusing
- n childlessness in Colombia. It will ask:
1. How has the overall level of childlessness amongst Colombian women changed since the mid-1980s, with a particular focus on women around ages 30 (25-34 year olds, to explore delayed entrance into motherhood) and 40 (35-44 year olds, representing childlessness close to the end of the reproductive life)? 2. What factors are associated with childlessness in Colombia around ages 30 and 40? 3. To what extent can ‘voluntary’ childlessness be distinguished from other forms using DHS data? How has ‘voluntary’ childlessness changed since the 1980s? DATA & METHODS Data Sources The quantitative analysis presented here is based on a series of seven Demographic and Health Surveys (DHS) for Colombia, from 1986, 1990, 1995, 2000, 2005, 2010, and 2015, as
1.1 1.6 2.1 2.6 3.1 3.6 4.1 4.6 1986 1990 1995 2000 2005 2010 2015 TFR Atlantic Eastern Central Pacific Bogotá National Territories National TFR
SLIDE 5 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
5
well as Colombian census data extracts from 1985, 1993, and 2005. The census extracts are from the Integrated Public Use Microdata Series, International, or IPUMS-I (Minnesota Population Center, 2017). Methods All of the analysis presented in this paper applies only to female respondents, as available male data (from the census and 2015 DHS only) were excluded from the present analysis. Male perspectives will be explored in future work. Census data were only used for basic descriptive statistics, in order to cross-check DHS-based estimates of overall levels of childlessness and changes over time. Using the number of children ever born (CEB), I calculated the proportion of women aged 15-49 (in 5-year age groups) who reported being
- childless. The DHS does not give the option of an ‘unknown’ number of CEB; however, the
census is subject to relatively substantial amounts of missing data, particularly in the earlier 1985 and 1993 rounds. Therefore, because more than 2% of all census CEB data was missing, I followed the methods outlined by Moultrie, et al. (2013), to assess each round and apply the El-Badry correction. Despite this, I will present a series of three estimates for the proportions childless: the ‘crude’ proportion (including only those women with zero CEB); the ‘aggregated’ proportion (adding all women with an ‘unknown’ number of CEB to those with zero CEB); and finally, the (El-Badry) ‘corrected’ estimates. Using the series of DHS, I first explored trends in overall childlessness (for all women aged 15-49), before focusing on age-specific childlessness in two groups: ‘younger’ women, around age 30 (25-34 years old) and ‘older’ women, around age 40 (35-44 years old), in order to capture whether trends in and factors associated with postponement (childlessness around age 30) differed from those for female childlessness towards the end of the fertile years (around age 40). Ten year age groups were used to increase robustness, as the DHS are sample surveys. ‘Childlessness’ was defined as a binary outcome (yes/no), with women who were not pregnant at the time of survey and who reported zero children ever born (CEB) classed as ‘childless’, irrespective of partnership status. Then, focusing only on two age groups: women around age 30 and around age 40, I explored how the proportions childless varied over time, according to six hypothesised ‘explanatory’ variables: partnership status (‘marriage’ includes cohabitation); level of education; socioeconomic status (wealth index quintiles); occupational status; area of residence; and
SLIDE 6 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
6
- region. Focusing on the 2010 DHS only, I then used bivariate logistic regression to test the
statistical significance of all independent associations and calculate ‘crude’ odds ratios (ORs), before formulating multivariate logistic regression models to probe the relative importance of each explanatory variable, adjusted for the effects of the other variables in the model. Continuous age was included as an a priori confounder, since childlessness tends to decrease with increasing age. In order to arrive at a ‘final’ explanatory model, I iteratively compared a series of different multivariate models using the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) values. AIC and BIC are used for judging relative goodness of fit (comparing models, which do not have to be nested), and both measures include a penalty for overfitting. Lower AIC and BIC values indicate better model fit and are therefore preferred. The BIC has a higher penalty for superfluous information than the AIC does, meaning that it tends to favour more parsimonious models. Finally, I broke down the ‘childless’ category further, exploring proportions of so-called ‘voluntary’ and ‘involuntary’ (or temporary) childlessness in each DHS over time. I did this using DHS variables derived from questions regarding desired and ideal fertility. For more details regarding the phrasing of these DHS questions, and my categorisations of ‘voluntary’ and ‘involuntary/temporary’ childlessness, see Appendix A. As in the logistic regression modelling described above, I chose to focus my analysis of intentionality on the 2010 DHS, as this year had the largest sample, including nearly 50,000 women aged 15-49, which was
- ver 10,000 more than the most recent 2015 DHS. IBM SPSS version 24 and R version 3.3.1
were used for data management and analysis. The analysis presented here is based on survey- weighted data analysed using the R Survey package (Lumley, 2004, 2017). RESULTS & DISCUSSION Background Characteristics Between 1986 and 2010, the proportion of all women in the DHS sample who worked outside the home more than doubled from 24% to 53%. This pattern was reflected at an even higher level amongst women aged 35-44, with an increase from 28% to 67% working in the same
- period. Women’s educational attainment rose at a similarly dynamic rate: whereas in 1986,
less than one in 10 women aged 25-34 had a ‘higher’ level of education (just 8%), by 2015 this had more than quintupled to nearly one in two (45.1%), as well as more than one-third (33.6%) of 35-44 year olds. Also amongst women aged 35-44, the proportion who never
SLIDE 7 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
7
married or cohabited remained stable and relatively low (between 8-9%). The changing characteristics of Colombian DHS respondents in this 30 year period reflect the broader socio-demographic transitions in Colombian society in the same time. Overall Childlessness Trends Since the 1980s: Tracking Childlessness Across Seven Rounds of DHS
Figure 3: Proportions Childless Amongst All Colombian Women by Age Group, DHS 1986- 2015
Based on analysis of the DHS, overall levels of childlessness do not appear to be increasing
- ver time in the Colombian context, despite continuously declining fertility and a national
TFR that is now below replacement level, as well as substantial increases in women’s education and work outside the home. With respect to childlessness, the proportion of all women aged 15-49 with no children actually decreased slightly over time: from 36% in 1986 to 30% in 2010. This reflects decreasing childlessness at younger ages, likely related to rising adolescent motherhood (Florez and Soto, 2006). In contrast, childlessness amongst ‘younger’ women aged 25-34 has increased slightly (from a low of 17.5% in 1986 to a high of 19.5% in 2015), although in the intervening DHS rounds, it varied from 19.3% in 1990 to 17.6% in 1995 and then remained fairly steady – not exhibiting a clear upward or downward year-on- year trend. Childlessness amongst ‘older’ women aged 35-44 has also remained fairly steady (7.5% in 1986 versus 6.8% in 2015), peaking in 1990 at 8.7%, before declining slightly for the following two DHS rounds, and then reaching 8.6% again in 2005. Since then, it has slid back down to 6.8% in 2015. Over the past 30 or so years, from 1986 to 2015, childlessness
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 15-19 20-24 25-29 30-34 35-39 40-44 45-49 % Childless 1986 1990 1995 2000 2005 2010 2015
SLIDE 8 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
8
around ages 30 and 40 has varied within a relatively narrow range of around 2%, and without displaying a consistently increasing or decreasing pattern over time, instead fluctuating slightly with each DHS round. Trends Since the 1980s: Childlessness in Three Rounds of the National Census Unlike the DHS, the Colombian census is subject to missing/unknown data regarding number
- f ‘children ever born’ (CEB). This missing data ranges from relatively low values of around
3% across all age groups in 2005 to highs of around 32% for 15-19 year olds (declining to around 5 or 6% in the older groups) in the 1993 census. As El-Badry (1961) noted, when census data include large numbers of women (over 2%) with ‘unknown’ or missing information regarding CEB, we can assume that at least some of them are actually childless.
Figure 4: ‘Crude’ Proportion of Colombian Women Aged 15-49 Who were Childless (Zero CEB Only), 1985-2005 Census Rounds Figure 5: ‘Aggregated’ Proportion of Colombian Women Aged 15-49 Who were Childless (Zero or Unknown CEB), 1985-2005 Census Rounds Source: Colombian census data (10% sample) from IPUMS-I. Author’s own calculations.
Because of this, I have presented ‘childlessness’ in three different ways: (1) the ‘crude’ proportion of women who are childless (Figure 4), based only on women with zero CEB and excluding all missing data; (2) the ‘aggregated’ proportion of women who are childless (Figure 5), which assumes that all women with missing CEB data in fact have zero CEB; and (3) the ‘corrected’ proportion of women who are childless (Figure 6), using the El-Badry
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 2005 1993 1985 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 2005 1993 1985
SLIDE 9 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
9
correction to determine what proportion of women with missing CEB data are likely childless, and adding only these women to the childless category. The ‘crude’ level of childlessness is
- bviously flawed, as it is very unlikely
that only 54% of 15-19 year olds were childless in 1993 (see Figure 4). In contrast, the ‘aggregated’ level of childlessness (which added all women with an ‘unknown’ CEB to those with zero CEB) most likely over-estimates childlessness in the earliest rounds (1985 and 1993), as they had more missing data. In the youngest age groups, there appears to be very little change in the proportions childless over time, and even in the older groups (35+), the change seems negligible, as all estimates are tightly clustered between approximately 11 and 15% childless (see Figure 5). I should note that this overall level of ‘aggregated’ childlessness is substantially higher than the ‘crude’ estimates. Finally, turning to the trends in childlessness over time in the ‘corrected’ estimates (see Figure 6), these seem to display some divergence between the earlier 1985 and 1993 rounds and the most recent census in 2005, especially in the oldest groups (35+), where it appears that childlessness has risen from around 7-8% in the 1980s and 1990s to 10-12% in the early 2000s. These estimates of childlessness differ from the DHS in two important ways: firstly, in terms
- f overall levels and secondly in terms of trends over time. With respect to levels, for all
women over 35 in the census, regardless of the measure used, childlessness is over 10% in the most recent data from 2005. In contrast, none of the DHS rounds, especially not those from 2005, 2010, or 2015 exhibit such high levels of childlessness (all staying under 10%). It is unclear why this discrepancy exists, and whether the census or the DHS are more accurate in estimating the proportion of older women (35+) who are childless. With respect to trends, the census and DHS exhibit different patterns of change over time: whereas childlessness appears to be low and (relatively) stable in the DHS, in both the ‘crude’ and ‘corrected’
Figure 6: El-Badry ‘Corrected’ Estimates of Age-specific Proportions of Colombian Women who are Childless, 1985-2005 Census Rounds Source: Colombian census data (10% sample) from IPUMS-I. Author’s own calculations.
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 2005 1993 1985
SLIDE 10 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
10
census estimates childlessness appears to have increased between 1993 and 2005. As this increasing pattern relies only on change between two data points, it remains to be seen whether this trend will continue in future census rounds. Due to the greater detail provided by the DHS and the availability of more recent surveys, the remainder of this paper will focus exclusively on DHS. We should keep in mind, however, that the DHS estimates of overall childlessness are consistently lower than those for the equivalent census years. Factors Associated with Childlessness Trends Over Time in Proportions Childless Around Ages 30 and 40 (DHS 1986-2015) In order to understand who is most likely to be childless in Colombia, it is important to explore childless as distributed according to personal characteristics (the six ‘explanatory variables’ described above in the Methods section), as it is likely that there are substantial subgroup differences. In fact, the most striking difference in proportions childless occurs as a result of partnership status. Figure 7 and Figure 8 show that whereas women around age 30 and 40 who have ever been in a union are unlikely to be childless (generally less than 10% of the younger group and 5% of the older group), amongst women who have never been married
- r cohabited, these figures are substantially higher: in 2015, around 60% of the younger
women and 35% of older women. Based on the literature from other countries, we would also expect that larger proportions of women with a higher level of education and higher socioeconomic status will be childless (Fieder, Huber and Bookstein, 2011; Cavenaghi and Alves, 2013). Childlessness in Colombia does indeed vary according to educational outcomes, in the expected direction, whereby the highest proportions childless are found amongst women with a ‘higher’ level of education in both around age 30 and 40, as can be seen in Figure 9 and Figure 10. Secondary education appears to have a relatively negligible (and declining) impact on childlessness. Figure 11 and Figure 12 demonstrate that wealth quintiles have a similar, although less pronounced effect than education on childlessness, particularly in the younger group, where the poorest women have the lowest levels of childlessness, and this increases relatively consistently across wealth quintiles. Although the richer/ richest older women also have higher levels of childlessness than the other groups, the patterns over time and between groups are less clearly visible.
SLIDE 11
Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
11
25-34 year old women only: 35-44 year old women only:
Figure 7: % Childless by Partnership Status Figure 8: % Childless by Partnership Status Figure 9: : % Childless by Educational Attainment Figure 10: % Childless by Educational Attainment Figure 11: % Childless by Wealth Quintile* Figure 12: % Childless by Wealth Quintile* *The Colombia DHS in 1986 did not include information regarding wealth quintiles.
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 1986 1990 1995 2000 2005 2010 2015 Never in union Currently in union Formerly in union 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 1986 1990 1995 2000 2005 2010 2015 Never in union Currently in union Formerly in union 0.0 10.0 20.0 30.0 40.0 50.0 60.0 No education Primary Secondary Higher 0.0 5.0 10.0 15.0 20.0 25.0 No education Primary Secondary Higher 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 1990 1995 2000 2005 2010 2015 Poorest Poorer Middle Richer Richest 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 1990 1995 2000 2005 2010 2015 Poorest Poorer Middle Richer Richest
SLIDE 12
Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
12
25-34 year old women only: 35-44 year old women only:
Figure 13: % Childless by Occupational Status Figure 14: % Childless by Occupational Status Figure 15: % Childless by Area of Residence Figure 16: % Childless by Area of Residence Figure 17: % Childless by Region (excluding National Territories) Figure 18: % Childless by Region (excluding National Territories)
The proportions childless, stratified by occupational status (currently working outside the home or not), display different patterns over time in the younger and older age groups. Figure
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 1986 1990 1995 2000 2005 2010 2015 Not working Currently working 0.0 2.0 4.0 6.0 8.0 10.0 12.0 1986199019952000200520102015 Not working Currently Working 0.0 5.0 10.0 15.0 20.0 25.0 1986 1990 1995 2000 2005 2010 2015 Urban Rural 0.0 2.0 4.0 6.0 8.0 10.0 12.0 1986 1990 1995 2000 2005 2010 2015 Urban Rural 0.0 5.0 10.0 15.0 20.0 25.0 30.0 Atlantic Eastern Central Pacific Bogotá 0.0 2.0 4.0 6.0 8.0 10.0 12.0 Atlantic Eastern Central Pacific Bogotá
SLIDE 13 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
13
13 shows that the proportion of working younger women who are childless is decreasing over time, whereas amongst those who do not work, it has remained steady. However, even in 2015, there was a substantial gap between the approximately 10% of non-working women who were childless versus almost 25% of those who were working. Amongst older women in 2015, the gap had narrowed to such an extent that around 7% of both working and non- working women were childless, as can be seen in Figure 14. This declining proportion of working women who are childless could perhaps be due to more mothers joining the workforce over the 30 year reference period (Amador, Bernal and Pena, 2013), meaning that the share of working women who are childless has decreased modestly over time. Finally, and turning our attention from individual to geographic characteristics, the patterns displayed according to urban or rural residence in Figure 15 and Figure 16 also display a difference in the expected direction, whereby urban women have consistently higher rates of childlessness than their rural counterparts. In the younger age group, urban women have a fairly steady level of childlessness (around 20%, although very modestly increasing over time), whereas amongst the younger rural women around 10% are childless, with a modestly decreasing trend over time, indicating a small increase in rural-urban polarisation according to motherhood/non-motherhood. Finally, Figure 17 and Figure 18 illustrate the regional patterns of childlessness (excluding the ‘National Territories’ because the DHS collected no data in this region until 2005). Although the regional trends are not entirely clear (particularly in the older age group), amongst younger women, Bogota and the Central region (home to Medellin, Colombia’s second-largest city), have consistently had the highest proportions childless since the 1980s, and particularly Bogota appears to exhibit a modest upward trend
- ver time, reaching 25% by 2015.
Multivariate Analysis of the 2010 DHS Table 2 presents the crude and adjusted odds ratios (ORs) from bi- and multivariate logistic regression models exploring the strength of the associations between each explanatory variable and childlessness around ages 30 and 40. The ‘full’ model includes all explanatory variables, while the ‘final’ model presents the model that provides the best explanation of childlessness in each of the two age groups, by eliminating superfluous factors through iterative model-building and the comparison of model AIC and BIC values (lower values indicate a better model).
SLIDE 14 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
14 Table 2: Associations between explanatory variables and childlessness around ages 30 and 40, using 2010 Colombia DHS Hypothesised Explanatory Variables Women Aged 25-34 Only (N=14,114) Women Aged 35-44 Only (N=12,773) Crude Odds Ratio (OR) Full Model: Adjusted OR Final Model: Adjusted OR Crude OR Full Model: Adjusted OR Final Model: Adjusted OR Age (Continuous) 0.86 *** 0.90 *** 0.90 *** 0.96 * 0.99
status: Ever married 1.0 1.0 1.0 1.0 1.0 1.0 Never married 21.33 *** 15.65 *** 15.16 *** 27.35 *** 27.22 *** 26.73 *** Education: None/Primary
1.0 1.0 1.0 1.0 1.0 1.0 Secondary 2.24 *** 1.33 * 1.38 * 1.59 *** 1.13 1.16 Higher 10.49 *** 4.02 *** 4.26 *** 5.26 *** 3.42 *** 3.60 *** Wealth index: Poorest (ref) 1.0 1.0 1.0 1.0 1.0 1.0 Poorer 1.80 *** 1.44 * 1.33 . 1.41 . 1.02 1.08 Middle 2.31 *** 1.96 *** 1.66 ** 2.13 *** 1.27 1.37 Richer 3.42 *** 1.99 *** 1.59 ** 2.61 *** 1.35 1.43 . Richest 6.82 *** 3.05 *** 2.32 *** 4.15 *** 1.80 * 1.89 *** Occupational status: Not working 1.0 1.0 1.0 1.0 1.0
2.29 *** 1.34 *** 1.29 ** 1.71 *** 1.13
residence: Urban (ref) 1.0 1.0 1.0 1.0 1.0
0.42 *** 1.53 ** 1.40 * 0.44 *** 0.93
Bogotá (ref) 1.0 1.0
1.0
0.68 *** 1.81 ***
* 1.34
0.76 * 1.37 *
1.19
0.54 *** 1.01
** 0.77
0.76 * 1.47 **
1.04
Territories 0.44 *** 1.44 *
*** 0.70
** 0.44 .
*** 0.01 *** AIC
8714.1
4413.6 BIC
8761.5
4467.7 Level of statistical significance: p≤0.001 ‘***’ p≤0.01 ‘**’ p≤0.05 ‘*’ p≤0.10 ‘.’
Echoing the crude results presented in the previous section regarding trends over time stratified by each ‘explanatory’ variable, several factors stand out as highly important to childlessness in both age groups, namely: partnership status, education, and socioeconomic
- status. Partnership status exhibited by far the strongest association with childlessness.
Women who had never married or cohabited were significantly more likely to be childless at both ages 30 and 40. Even when adjusted for other factors, like education and socioeconomic
SLIDE 15 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
15
status, never partnered women in the younger group had odds of childlessness 15 times those
- f ever partnered women, while the equivalent figure in the older group was nearly 27,
indicating that a large proportion of childlessness may result from circumstance. Even when adjusted for the effects of partnership status and other important factors, having a higher level of education was another strongly significant factor increasing the odds of childlessness for women around ages 30 and 40 (OR=4.26 in the younger group and 3.6 in the older group), compared to those with no education or primary only. Although women with secondary education were also more likely to be childless than women with lower levels
- f education, the effect was much smaller (and not statistically significant in the older group).
Socioeconomic status, measured by wealth index, also had a significant effect in both age groups: the richest women were more likely to be childless than the poorest women (OR~2- 3), even after adjusting for education and other important covariates, although its ‘crude’ effects were much stronger. While these three factors were the best predictors of childlessness around age 40, the final model for the younger group also included
- ccupational status and area of residence, with working women and those living in urban
areas displaying higher levels of childlessness between ages 25-34, potentially indicating that these factors are associated with postponement, but not definitively with childlessness later
Intentionality: Fertility Ideals, Desires and Whether Childlessness is Voluntary or Not The overarching ‘childless’ category, which refers to all women with zero CEB who were not pregnant at the time of survey, was further broken down into according to different measures
- f intentionality, as described in the Appendix. Here, we will first examine fertility ‘ideals’,
before turning our attention to ‘voluntary’ and ‘involuntary or temporary’ childlessness, which combines women’s answers to questions regarding ideal family size and desired future fertility. Fertility Ideals: Trends Over Time Figure 19 shows that, between the first DHS in 1986 and the most recent round in 2015, the proportion of women 15-49 years old, who say that their ‘ideal family size’ is zero, one or two children has grown from around 55% to over 70%, with 45% and 52% expressing a two-
SLIDE 16
Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
16
child preference in 1986 and 2015, respectively. Although the two-child norm has persisted and strengthened, since the 2000 round, the proportion of women with a one-child ideal (15- 16%) has nearly caught up with three children (17-18%). Additionally, an ideal of more than three children has halved in proportion, from just over one-fifth in 1986 to 10% in 2015. A parallel change has occurred in the lower orders since 1986, when only 10% of women stated an ideal of zero or one child, doubling to 20% by 2015.
Figure 19: Distribution of Changing Ideal Family Size over Time (All Women, DHS 1986-2015)
Figure 20 compares the actual proportion of women who are childless in their 40s with the proportion of women in their 40s, and of all women aged 15-49, who declare that their ‘ideal’ is childlessness, from 1986 to 2015. ‘Actual’ childlessness exhibits an inverse U-pattern (from 5% in 1986 to a high of almost 8% in 2000 and back down to around 6% by 2015). The pattern for ideal childlessness is different, with a modest increase over time amongst all women (15-49), and for the 40-49 year old age group. In 2015, amongst women in their 40s, the proportions actually childless and those declaring that their ideal is childlessness almost converge around 5-6%. However, as the denominator for both proportions is all women in this age group (including mothers), not all those declaring a childlessness ideal are also actually childless.
2% 3% 2% 3% 4% 4% 6% 8% 9% 10% 15% 16% 15% 15% 45% 50% 51% 52% 53% 53% 52% 23% 22% 21% 18% 17% 17% 18% 12% 9% 9% 6% 6% 6% 5% 4% 3% 3% 2% 2% 2% 2% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1986 1990 1995 2000 2005 2010 2015 Proportions Stating each Numerical Ideal (100% Total) 7+ / Non- numeric 6 5 4 3 2 1
SLIDE 17 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
17 Figure 20: Comparing Actual and Ideal Proportions of all Colombian Women who are Childless (DHS 1986-2015)
Voluntary and Involuntary or Temporary Childlessness: Trends Over Time ‘Voluntary’ childlessness is a wider category than ‘ideal’ childlessness, as it includes women who have been sterilized and those who declare that they want ‘no’ children in future, alongside those with an ‘ideal’ family size of zero. Figure 21 shows that, if there is a trend
- ver time with respect to overall childlessness amongst all women aged 15-49, it is that it is
becoming less common over time (from just under 40% in 1986 to around 30% in 2015), likely due to earlier transitions to motherhood, and perhaps also to better treatments for infertility. When we examine ‘voluntary’ or ‘involuntary/temporary’ forms, it becomes clear that, as childlessness overall has gone down, there has been an increase in the proportion of all women who can be classed as ‘voluntarily’ childless, from between 1-1.5% in the earliest DHS to 5.6% in the most recent survey, alongside a concomitant decrease in the proportion
- f women who are ‘involuntarily or temporarily’ childless. The same pattern is observable in
Figure 22 and Figure 23, focusing on 15-24 and 25-34 year olds, respectively, with the largest increase in the proportion of women ‘voluntarily’ childless in the youngest age group (from less than 1 in 50 women in 1990 to 1 in 10 by 2015). Figure 22, Figure 23, and Figure 24
- nly present change between 1990 and 2015, as the unweighted numbers of childless women
(particularly ‘voluntary’) were too small to be reliable in the earliest survey from 1986, except when aggregated across all age groups.
0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 1986 1990 1995 2000 2005 2010 2015 Proportion (%) Actual Childless % (40-49 y.o.) Ideal Childless % (40-49 y.o.) Ideal Childless % (15-49 y.o.)
SLIDE 18
Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
18
Changes over Time in the Relative Distribution of Motherhood and Voluntary or Involuntary / Temporary Childlessness Among Colombian Women, DHS 1986/1990-2015
Figure 21: Women of All Ages (15-49) Figure 22: 15 to 24 Year Old Women Only Figure 23: 25 to 34 Year Old Women Only Figure 24: 35 to 49 Year Old Women Only
Shifting our frame from considering all women as our denominator to focusing only on women without children, Table 3 and Figure 25 explore changes over time in the proportions of all childless women who are ‘voluntarily’ so. Amongst all childless women, there has been a steadily increasing number who declare that their childlessness is ‘voluntary’: from just 3% in 1986 to 18.6% in 2015.
Table 3: Trend in the Proportion of All Childless Women Ages 15-49 who are 'Voluntarily' Childless (DHS 1986-2015) 1986 1990 1995 2000 2005 2010 2015 % of All Childless Women 3.0 3.7 4.0 10.0 12.3 12.0 18.6 Total (Weighted) N, Voluntarily Childless 58 112 140 363 1411 1760 2019
This increasing trend in ‘voluntary’ childlessness, as well as the greater number of women of
63.7 64.0 68.7 68.5 70.2 70.5 70.1 1.1 1.3 1.3 3.1 3.7 3.5 5.6 35.2 34.7 30.0 28.3 26.2 25.9 24.3 0% 20% 40% 60% 80% 100% 1986 1990 1995 2000 2005 2010 2015 Involuntary / Temporary Childlessness Voluntary Childless Mother 32.0 37.1 35.7 37.6 35.8 35.2 1.8 1.1 4.1 4.8 5.4 10.2 66.2 61.8 60.2 57.6 58.8 54.7 0% 20% 40% 60% 80% 100% 1990 1995 2000 2005 2010 2015 Involuntary / Temporary Childlessness Voluntary Childless Mother 80.7 82.4 82.1 81.2 82.1 80.5 0.9 0.8 2.0 2.6 2.3 3.9 18.4 16.8 16.0 16.2 15.6 15.6 0% 20% 40% 60% 80% 100% 1990 1995 2000 2005 2010 2015 Involuntary / Temporary Childlessness Voluntary Childless Mother 92.1 91.9 92.2 91.9 92.9 93.5 1.1 1.9 3.1 3.5 2.8 2.7 6.8 6.2 4.7 4.6 4.3 3.7 0% 20% 40% 60% 80% 100% 1990 1995 2000 2005 2010 2015 Involuntary / Temporary Childlessness Voluntary Childless Mother
SLIDE 19 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
19
all ages declaring that their ‘ideal family size’ is zero could indicate that, even as childlessness overall is not becoming more common, the idea of childlessness might be more acceptable than it was 30 years ago. Perhaps the stigma of childlessness and pro-natalist pressures are slowly subsiding, making it more possible even for non-mothers to conceptualise and state an ideal of zero children and no desired children in the future. This increasing trend could indicate either that more women today are truly ‘voluntarily’ childless than in the past, or alternatively, there could always have been a substantial minority of women who were ‘voluntarily’ childless, but in the past they may have felt more reluctant to declare their true feelings in a survey like the DHS, whereas they increasingly feel freer to do
- so. It could also be a combination of these two hypothesised explanations.
Figure 25: Trends in the Proportion (%) of Childless Women of Different Ages who are ‘Voluntarily’ Childless (DHS 1990-2015)
When broken down into three rough age groups (as in Figure 25), there is still a strongly increasing trend in ‘voluntary’ childlessness over time, as well as one that also emerges across age groups. The vast majority of young women without children are not ‘voluntarily’ childless, even in 2015, with most instead indicating that they would like to have children in
- future. Only 15% of non-mothers aged 15-24 and 20% of those aged 25-34 indicated that
they do not desire children in future and/or that their ideal is zero children (and that they are therefore ‘voluntarily’ childless). However, approximately 40% of the older age group (35+ years) has fallen into the ‘voluntary’ category since at least the 2000 DHS, with small year-
- n-year fluctuations. This higher level of ‘voluntary’ forms amongst the oldest, but not the
youngest, childless women could indicate that there is some adjustment effect over the life course amongst women who realise that it is unlikely that they will have children in future,
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 15-24 25-34 35+ 1990 1995 2000 2005 2010 2015
SLIDE 20 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
20
and therefore adjust their childbearing goals or ideals downward. There is longitudinal research from the US which suggests that women’s fertility expectations decline over time (Hayford, 2009), and from Australia, suggesting that the ‘childbearing desires’ of childless adults tend to be adjusted downward with age (and conversely, upward with relationship formation) (Gray, Evans and Reimondos, 2013). At least some women (and men) will change their minds and adjust their fertility preferences and life goals with other life changes; desired fertility is not fixed from the age of 15 until 49 or the end of life. Hayford (2009, p. 767) terms this a “sequential model of fertility intentions and behavior”, which is “a model that conceptualizes fertility decision-making as a dynamic process, in which people revisit plans for childbearing in response to other changes in life experiences.” For example, women in their forties who may have wanted to have children earlier in life, but did not achieve this initial goal, may decide that they are happy as non-mothers and therefore declare that not
- nly do they not want to have a child in future, but that their ‘ideal’ family size is also zero
- children. Similarly, other women might move from the ‘voluntarily’ childless younger group
into the 60% of childless women over 35 who are ‘involuntarily’ (or temporarily) so, as their lives and personal relationships change – influenced perhaps by friends, family or romantic
- partners. Unfortunately, because the DHS is cross-sectional, it was not possible to determine
whether these hypothesised trajectories suggested by the data are borne out in reality, but the patterns discussed above raise questions for future research. CONCLUSIONS, LIMITATIONS & FUTURE WORK Conclusions Using Colombia DHS from 1986 to 2015, this paper has shown that overall levels of childlessness amongst women aged 15-49 have not increased over time, despite the constantly decreasing TFR observed over the same period. Additionally, focusing in on childlessness in two age groups of women: the first indicative of ‘postponement’ in women around age 30 (25-34 years old), and the second, addressing childlessness towards the end of the reproductive lifespan amongst women around age 40 (35-44 years old), I demonstrated that the overall level of childlessness was steady, showing little movement aside from year-
- n-year fluctuations of around 2%. In 2015, approximately 19.5% of women around age 30
and 6.8% of women around age 40 were childless. However, after cross-checking these DHS estimates against analysis of the 1985, 1993 and 2005 censuses, it became clear that the
SLIDE 21 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
21
census data display a different pattern of childlessness, consisting of higher overall levels than the DHS and an increasing (rather than stable) trend between 1993 and 2005. However, it is as yet unclear whether the data from the DHS or census provide a better reflection of Colombian reality. Then, using bi- and multivariate modelling with the 2010 DHS data only, I explored the relative importance of six hypothesised ‘explanatory’ variables for childlessness, demonstrating that for both age groups (around age 30 and 40), partnership status was the most important factor, as women who had never been married or cohabited exhibiting much higher rates of childlessness than those who had ever been partnered. This was followed by education (especially having a ‘higher’ level), and then socioeconomic status, with women in the ‘richer’ and ‘richest’ wealth quintiles having higher rates of childlessness than others. With respect to the intentionality of childlessness, I showed that, since the 1980s, a consistently increasing proportion of all women have had an ideal of no children or one child. Similarly, the share of women who are ‘voluntarily’ childlessness has increased with each DHS over the past 30 years as a proportion of all women, and when looking only at women without children. This increasing proportion of childlessness that can be viewed as ‘voluntary’ could be an initial sign that ideas and attitudes are changing, and that some features consistent with the SDT are slowly spreading in Colombia. Despite this increase, even in 2015, only 5.6% of all Colombian women aged 15-49 could be classed as ‘voluntarily’ childless. Finally, when focusing on childless women only, the ‘voluntary’ component also appears to increase positively with age, as approximately 42.1% of women
- ver 35 could be viewed as ‘voluntarily’ childless, compared to just 15.7% in the youngest
group of women (aged 15-25). I hypothesised that this could be the result of a process of fertility goal adjustment over time, although owing to the cross-sectional nature of the DHS data, it is difficult to come to any firm conclusions with respect to the underlying causes. Limitations & Future Work This paper focused on estimating and comparing period measures of childlessness using the DHS, rather than taking a cohort approach to census data using parity progression ratios (PPRs), as exemplified by Cavenaghi and Alves’s (2013) work on Brazil. I will consider adopting a cohort approach to the analysis of the Colombian census data, as this could yield slightly different results, with the advantage that it also allows for the projection of future
SLIDE 22 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
22
levels of childlessness in younger cohorts. Future work will also seek to explore potential explanations for the observed differences between the DHS and census, as census estimates
- f childlessness are consistently higher. Based on the present analysis, regional effects did
not appear to be an important factor in explaining variations in levels of childlessness; however, one of the original ideas of this research was to probe regional differences using multilevel modelling of census data, as Colombia is known to be a heavily regional country. With respect to intentionality, of course the categories of ‘ideal’ and ‘voluntary’ childlessness derived for this paper are crude measures of fertility intention based only on the available cross-sectional survey data. Without more detailed qualitative information, it is difficult to determine whether even the ‘voluntarily’ childless women do affirmatively want to be childless and how happy they are with their current fertility state (i.e. separating the happily childfree from the reluctantly or unhappily childless and perhaps more neutral or ‘semi’- voluntary forms of childlessness). One of this study’s ultimate goals is to contribute to extending anthropological demography to Latin America, and the author is finishing approximately one year of ethnographic fieldwork in urban Colombia, which seeks to complement the quantitative analysis above by focusing on the personal experiences and meanings assigned to childlessness and parenthood, particularly in relation to shifting gender
- roles. This holistic view of ‘childlessness’ will allow for the integration of male perspectives,
as well as a fuller examination of the motivations underlying childlessness in Colombia and, as such, a critical evaluation of the usefulness of SDT theories in this particular Latin American context. In addition, it will probe whether and how childlessness relates to definitions of the Colombian ‘family’ as well as how this has changed over time and might continue to do so in future. ACKNOWLEDGEMENTS Professor Sara Randall and Dr Lucio Vinicius from the Department of Anthropology at University College London (UCL) are supervising this doctoral research project. This work is supported by a doctoral research studentship from the Economic and Social Research Council (ESRC) through UCL’s ESRC Doctoral Training Centre. The IPUMS-I census data were provided by the Minnesota Population Center, and originally produced by Colombia’s National Administrative Department of Statistics (Departamento Administrativo Nacional de Estadística (DANE)).
SLIDE 23 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
23
REFERENCES Amador, D., Bernal, R. and Pena, X. (2013) The Rise in Female Participation in Colombia: Fertility, Marital Status or Education?, Documentos CEDE. Bogota: CEDE, Facultad de Economia, Universidad de los Andes. Arriagada, I. (2002) ‘Changes and inequality in Latin American families’, CEPAL Review. Santiago, (No. 77), pp. 135–153. Cavenaghi, S. (2013) Fertility Decline and Public Policies to Address Population Rights: Perspective from Latin America. Expert Paper No. 2013/5. New York, N.Y. Cavenaghi, S. and Alves, J. (2013) ‘Childlessness in Brazil: Socioeconomic and Regional Diversity’, in XXVII IUSSP International Population Conference. Busan, South Korea. El-Badry, M. A. (1961) ‘Failure of Enumerators to Make Entries of Zero: Errors in Recording Childless Cases in Population Censuses’, Journal of the American Statistical Association, 56(296), pp. 909–924. doi: 10.2307/2282003. Esteve, A., Lesthaeghe, R. and López-Gay, A. (2012) ‘The Latin American cohabitation boom, 1970–2007’, Population and Development Review, 38(1), pp. 55–81. Fieder, M., Huber, S. and Bookstein, F. L. (2011) ‘Socioeconomic Status, Marital Status and Childlessness in Men and Women: An Analysis of Census Data from Six Countries’, Journal
- f Biosocial Science, 43(5), pp. 619–635. doi: 10.1017/S002193201100023X.
Florez, C. E. and Sanchez, L. M. (2013) Fecundidad y Familia en Colombia: hacia una segunda transicion demografica? Bogota, Colombia. Florez, C. E. and Soto, V. E. (2006) ‘Fecundidad Adolescente y Desigualdad en Colombia’, Notas de Poblacion, 83, pp. 41–74. Gray, E., Evans, A. and Reimondos, A. (2013) ‘Childbearing desires of childless men and women: when are goals adjusted?’, Advances in life course research, 18(2), pp. 141–9. doi: 10.1016/j.alcr.2012.09.003. Hayford, S. R. (2009) ‘The Evolution of Fertility Expectations Over the Life Course’,
- Demography. Population Association of America, 46(4), pp. 765–783. doi:
10.1353/dem.0.0073. Lesthaeghe, R. (2010) ‘The Unfolding Story of the Second Demographic Transition’, Population and Development Review, 36(2), pp. 211–251. Lumley, T. (2004) ‘Analysis of complex survey samples’, Journal of Statistical Software, 9(1), pp. 1–19. Lumley, T. (2017) ‘Survey: Analysis of Complex Survey Samples, R package version 3.32’. Minnesota Population Center (2017) ‘Integrated Public Use Microdata Series, International: Version 6.5.’ Minneapolis: University of Minnesota. doi: http://doi.org/10.18128/D020.V6.5. Moultrie, T. et al. (eds) (2013) Tools for Demographic Estimation. Paris: International Union for the Scientific Study of Population. Ojeda, G., Ordonez, M. and Ochoa, L. H. (2011) Encuesta Nacional de Demografia y Salud: ENDS 2010, Colombia. Bogota: Asociacion Probienestar de la Familia Colombiana Profamilia. Rosero-Bixby, L., Castro Martin, T. and Martin-Garcia, T. (2009) ‘Is Latin America starting to retreat from early and universal childbearing?’, Demographic Research, 20(Article 9), pp. 169–194. doi: 10.4054/DemRes.2009.20.9.
SLIDE 24 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
24
APPENDIX A: Additional Methodological Details, Tables and Figures
Table A1: Population of Colombian departments with at or below replacement-level fertility (2005-2015) 2015 2010 2005 Department TFR % Nat’l. Pop. Populatio n TFR % Nat’l. Pop. Population TFR % Nat’l. Pop. Population Antioquia 1.4 13.4% 6,456,299 1.8 13.3% 6,066,003
1.8 16.3% 7,878,783 1.9 16.2% 7,363,782
1.8 2.6% 1,276,407
1.3 2.0% 987,991 1.6 2.1% 978,342 2.0 2.3% 968,586 Córdoba 2.1 3.5% 1,709,644
2.1 2.4% 1,154,777
2 2.0% 961,334
1.8 3.6% 1,744,228
1.7 1.2% 565,310 1.7 1.2% 549,662 1.7 1.2% 534,506 Risaralda 1.7 2.0% 951,953 1.7 2.0% 925,117 1.8 2.1% 897,413 Santander 1.8 4.3% 2,061,079 2.1 4.4% 2,010,393
Cauca) 1.6 9.6% 4,613,684 1.7 9.6% 4,383,277 2.0 9.7% 4,161,470 Arauca 2.1 0.5% 262,315
2.1 0.7% 356,479
0.7% 295,276 Guaviare 1.7 0.2% 111,060
1.8 0.7% 345,204
0.7% 310,132 San Andrés & Providencia 2 0.2% 76,442 2.1 0.2% 73,320
Departments)
31,512,989
22,349,896
7,167,383 Total (National) 2.0 100% 48,203,405 2.1 100% 45,509,584 2.4 100% 42,888,592 Source: Own elaboration, using TFRs from the DHS Stat Compiler (http://www.statcompiler.com/en/) and same- year population figures from DANE, via http://www.dane.gov.co/files/investigaciones/poblacion/proyepobla06_20/Municipal_area_1985-2020.xls. Last checked: 18 September 2017.
SLIDE 25 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
25
The Sub-categorisation of Childlessness Using Various Measures of Fertility Intentions (Additional Methodological Details) Questions regarding desired and ideal fertility in the DHS are phrased differently depending
- n a woman’s fertility status. The question regarding a woman’s desire for (more) children is
asked differently of pregnant and non-pregnant women, while ideal number of children is phrased slightly differently for mothers and non-mothers. In 2010, these questions were phrased as follows in Table A2.
Table A2: Phrasing of Colombia DHS 2010 Questions regarding fertility intentions and ideals Question Phrasing 1 Phrasing 2 Possible Answers Q 702: Desire for (more) children For women who are not pregnant / unsure: Now I have a few questions about the future. Would you like to have a(nother) or would you prefer not to have any more children? For pregnant women: Now I have a few questions about the future. After the child that you are waiting for, would you like to have a(nother) child, or would you prefer not to have any more children?
- Have a(nother) child
- None / No more children
- Cannot get pregnant
- Undecided / Does not know /
Unsure Q 720: Ideal number
children For women with no children: If you could choose exactly the number
- f children you would have
in your lifetime, how many would you have? For Mothers: If you could go back to a time before you had children and you could choose exactly how many you would have in your lifetime, how many children would you have?
- Number: ___
- None
- Other answer
Additional Directions: If the answer is non-numeric, probe: (Specify)__
Source: Taken from the Individual Questionnaire in Appendix E of the Colombia DHS 2010 Final Report (Ojeda, Ordonez and Ochoa, 2011) and translated from the original Spanish by the author.
The questions in Table A2 are transformed into two variables addressing (future) fertility preferences (V602) and ideal family size (V614), which I used to determine the intentionality
- f childlessness, or the degree to which it is voluntary or involuntary, described in Table A3.
Table A3: 2010 DHS derivation of main outcome measures for intentionality of childlessness Original Variables Categories
- 1. V602 (“Fertility Preference”, from Q702)
1 = Have another 2 = Undecided 3 = No more 4 = Sterilized (respondent or partner) 5 = Declared Infecund
- 2. V614 (“Ideal Number of Children (grouped)”,
from Q720) 0-5 = Numeric answers 6 = 6+ (grouped) 7 = Non-numeric answer
- 3. Childless
- Created using V210, ‘Total number of
children ever born (CEB)’, and V213, ‘Currently pregnant’ 0 = Childless (no CEB and not currently pregnant) 1 = Not Childless (1+ CEB and/or currently pregnant)
SLIDE 26 Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
26
In order to better understand the intentionality of childlessness, I first defined various categories, based on the questions in Table A2 and the variables in Table A3. My final measure of the intentionality of childlessness is ‘voluntary childlessness, wide’, described
- below. Figure A1 (below) shows the numeric breakdown of the 2010 DHS sample into these
categories, based on the questions in Table 2 and variables in Table 3. The defined categories
- f motherhood and childlessness are as follows:
1) Mother: 1+ CEB and/or pregnant at time of survey; 2) Desired childless: Childless (i.e. zero CEB and not pregnant when surveyed), as well as having a fertility preference/desire (V602) of ‘no more’ children or ‘sterilized’ / Undesired childlessness: Childless when surveyed, but declaring that they would like to ‘have another’ child, that they were ‘undecided’ or that they were ‘infecund’ in V602; 3) Ideal childlessness: Childless women whose ‘ideal family size’ (V614) is zero / Non-ideal childlessness: Childless women with an ‘ideal family size’ of 1+ children. 4) Voluntary childlessness (narrow): Childless women who fall into both the categories of desired and ideal childlessness, as defined above. All other non-mothers are involuntarily/temporarily childless. 5) Voluntary childlessness (wide): Childless women who fall into either the category of desired or ideal childlessness. All other non-mothers are involuntarily/temporarily childless.
SLIDE 27
Cristina Pérez 30 Sept. 2017 IUSSP PAPER: ‘Childlessness’ in Colombia
27
Mothers (25-34): 11,588 (82.1% of 14,144)
All Women, 15-49 y.o.: 49,818
Women 15-24 y.o.: 16,860 Women 25-34 y.o.: 14,114 Women 35+ y.o.: 18,844 Mothers (15-24): 6,037 (35.8% of 16,860) Mothers (35+): 17,510 (92.9% of 18,844) All Childless (35+): 1334 (7.1% of 18,844) 'Ideal' Childlessness: 194 (14.5% of 1,334) 'Desired' Childlessness: 498 (37.3% of 1,334) 167 (13%) 'Voluntary' Childlessness (Narrow) All Childless (25-34): 2,526 (17.9% of 14,144) 'Ideal' Childlessness: 180 (7.1% of 2,526) 'Desired' Childlessness: 314 (12.4% of 2,526) 'Voluntary' Childlessness (Narrow) 167 (7%) All Childless (15-24): 10,823 (64.2% of 16,860) 'Ideal' Childlessness: 568 (5.2% of 10,823) 'Desired' Childlessness: 887 (8.2% of 10,823) 'Voluntary' Childlessness (Narrow) 548 (5%) 'Voluntary' Childlessness (Wide): 906 (8% of 10,823) 'Voluntary' Childlessness (Wide): 326 (13% of 2,526) 'Voluntary' Childlessness (Wide): 525 (39% of 1,333)
Figure A1: Flow Chart Demonstrating Different Forms of Childlessness in the 2010 Colombia DHS
