Universe and Sample. Page 26. Universe. Population Table 1 Sub-populations excluded

Transcription

1 Universe and Sample Universe The universe from which the SAARF AMPS 2008 (and previous years) sample was drawn, comprised adults aged 16 years or older resident in private households, or hostels, residential hotels and similar accommodation in the Republic of South Africa. From 2009, the universe was expanded to include 15 year olds. The Dec 2009 release incorporated this new 15 year old universe for the first time. This added to the adult population. Excluded from the universe are residents of such institutions as prisons or hospitals, military personnel on active service and minority sub-populations in certain geographical areas, as shown in Table 1. Table 1 Sub-populations excluded Province Adults Population excluded % Western Cape Northern Cape Eastern Cape KwaZulu-Natal Free State Gauteng Mpumalanga Limpopo North West TOTAL Population 2012 For SAARF AMPS Jan 12-Dec 12, the same population figures were used as for AMPS Jul 11-Jun 12 and AMPS Jan 11-Dec 11, taken from the IHS Global Insight Regional explorer suite of models. (See Section 2 Special Notes for SAARF announcement on 2013 Population Updates). The Regional explorer model is a living model of the economic and socio-economic environment of South Africa. The ReX is a spatial model which aims to produce an internally consistent set of data for a range of variables across every region in the country. The Regional explorer suite of models are updated on a continuous basis, constantly taking into account new data sources, and new releases of existing data sources that span the full breadth of the South African socio-economic landscape. The methodology employed in the demographic model is different to the one previously used in a number of technical areas. Firstly, on a national level, the new cohort component model handles mortality in a fundamentally different way, returning to the technique employed in earlier BMR (Bureau of Market Research) publications. Secondly, on a regional level, a slightly different set of techniques are employed which will change the spatial distribution of the population to some extent. Page 26

2 2011 Population Estimates The following is a summary of the methodology. Methodology The IHS Global Insight population numbers are estimated in three distinct phases. The first and second phase both make use of Cohort-Component Population Projection methods, whereas the third phase makes use of the ratio method to regionally distribute the results determined by the first two phases. An interpolated ratio calculated from all available census data for all regions, age, gender and population groups are used in the ratio method. Cohort-Component Population Projection models are a class of models that are known for their ability to accurately consider the structure of a population as it grows over time. These models make use of population fundamentals (births, deaths, migration, etc.) to project a population as it experiences change. All of the demographic model outputs form part of a larger cross-regional model that aims to be entirely internally consistent with the economic, labour, income and development factors of every region in South Africa. Thus, as new data becomes available on any regional or national indicators, the Cohort Component Demographic model is updated and checked for internal consistency. The Cohort Component aspects of the demographic model require a number of inputs on various spatial and demographic levels. In total, a unique cohort component model is built for each population group and province, resulting in 45 distinct models each requiring their own set of inputs plus four national level models. Each model requires various input assumptions, of which the most important are listed below. The typical approach to estimating any inputs follows a three stage process: (1) Literature review of published demography work. (2) Additional calculations based on Censuses (and other surveys where available). (3) Calibrating and benchmarking the model against empirical outputs. Fertility Using the literature review fertility rates as a starting point, and combining those with additional calculations from various census and survey results, a set of final input TFRs is derived. Final national TFRs were estimated by calibrating the model such that the population estimate started at a given population in 1970 and passed through each of the population figures from the 1985 census up until the Community Survey of 2007, within an adjustment factor that recognised the quality of each individual dataset. Mortality Determining accurate mortality rates is complicated by a number of factors. However, the effect of HIV and AIDS on the mortality rates across the various population groups is the most contentious. Various techniques are available to overcome this problem, but regardless of the approach, the methodology employed should aim to correctly maintain the age-distribution of output deaths. The IHS Global Insight Demographic model makes use of the AIDS excluded model life tables, and accounts for the impact of HIV / AIDS ex post. The precise method used to adjust for HIV / AIDS is to use the AIM model. This technique is chosen in favour of the AIDS included model life tables approach on the basis that it is more accurate under scenarios where the AIDS progression rates are unstable and for various other technical reasons related to the actual construction of the model life tables themselves. Net migration Emigration is measured on an annual basis from the receiving countries side. This follows an original 2002 study by Statistics South Africa which IHS Global Insight has updated annually to reflect the change in the population s propensity to migrate. Immigration is estimated using the foreign-born population as measured in the various Statistics South Africa Censuses and surveys, and confirmed by various other sources where available. HIV / AIDS Estimates Various HIV and AIDS estimates are required to accurately account for the effect of the disease on the size of the population. The most contentious is the HIV prevalence rate. The IHS Global Insight demographic model derived prevalence rates largely from ASSA 2008, with some adjustments to account for the different model life tables by ASSA. These prevalence rates are built on significant work by ASSA on various primary data sets particularly the Ante-Natal prevalence surveys conducted by the Department of Health and, to some extent, the HSRC household surveys on HIV / AIDS. Page 27

3 Community Size Changes As a result of the 2011 population updates, a few community size changes occurred in SAARF AMPS Jan 11-Dec 11. There were no further changes for AMPS Jul 11-Jun 12 and AMPS Jan 12-Dec 12. Trends 1997 to 2000 In view of the release of lower than anticipated preliminary population figures from the 1996 Census and the paucity of detailed age and area data available, the decision not to adjust the SAARF AMPS 97, 98 and 99 population figures was made. Population figures were again updated from SAARF AMPS Tables The estimates of the total population of the RSA, including children under the age of 15, who do not form part of the adult SAARF AMPS sample - are shown in Table 2A for 2012 and earlier years, by sex. Figures for years prior to 1996 contained estimates for the previous TBVC and self-governing states. The estimates for 1988 to 1999 are as at January, while for previous years and from the year 2000 they are mid-year estimates. As a comparison, Table 2B shows the SAARF AMPS universe, excluding children below the age of 16 for years 1984 to 2008, and excluding children under 15 years for 2009 to date. Also excluded are the sub-populations which do not form part of the AMPS universe (see Table 1 for details). The breakdown of the adult population, by province, community, metropolitan area, sex, age group, home language, household income and population group forms part of Table 4, which appears at the end of the sampling section. Page 28

4 Total Population of South Africa Table 2A TOTAL POPULATION OF SOUTH AFRICA SEX TOTAL MEN WOMEN 1970(JUN) 1971(JUN) 1972(JUN) (JUN) 1974(JUN) 1975(JUN) (JUN) 1977(JUN) 1978(JUN) (JUN) 1980(JUN) 1981(JUN) (JUN) 1983(JUN) 1984(JUN) (JUN) 1986(JUN) 1988(JAN) (JAN) 1990(JAN) 1991(JAN) (JAN) 1993(JAN) 1994(JAN) (JAN) 1996(JAN) 1997(JAN) (JAN) 1999(JAN) 2000(JUN) (JUN) 2002(JUN) 2003(JUN) (JUN) 2005(JUN) 2006(JUN) (JUN) 2008(JUN) 2009(JUN) (JUN) 2011(JUN) 2012(JUN) Page 29

5 ADULT SAARF AMPS UNIVERSE* Table 2B ADULT SAARF AMPS UNIVERSE* SEX TOTAL MEN WOMEN 1984(JUN) 1985(JUN) 1986(JUN) (JAN) 1989(JAN) 1990(JAN) (JAN) 1992(JAN) 1993(JAN) (JAN) 1995(JAN) 1996(JAN) (JAN) 1998(JAN) 1999(JAN) 1999(JUN) 2000(JUN) 2001(JUN) (JUN) 2003(JUN) 2004(JUN) (JUN) 2006(JUN) (JUN) 2008(JUN) (JUN) 2010(JUN) 2011(JUN) (JUN) * See Table 1 for exclusions NOTE: From 2009 (Jun), the SAARF AMPS Universe includes 15 year olds (previous years include 16+) Page 30