The South African Audience Research Foundation (SAARF) conducts the All Media and Products Survey (AMPS) to collect data on media usage of South Africans, as well as their ownership or usage of certain products and services.
Kind of Data
Sample survey data
Unit of Analysis
Households and individuals
v1: Edited, anonymised data for licensed distribution
Version 1 of the survey was orignially acquired by DataFirst from SAARF in column-binary (mulitipunch) format.
Version 1.1 of the survey is the data converted to statistical software formats for research usage
Version 1.2 of the survey is a result of changes made to version 1.1. Changes made are correction of the age variable (ca56co36) in the "personal" data file.
Version 1.3 of the survey includes the corrected amps-2012-personal file with variables from ca57co09 to ca57co55 that were missing in version 1.2 of this data file.
The survey collects demographic data on the surveyed households, including data on race, sex, age, income, education level and home language. Data is also collected on media used by households, including newspapers and magazines, television, and radio, as well as cinema attendance. The survey also collects data on ownership and usage of products and services.
The survey covers major rural and urban areas in South Africa. This includes major metropolitan areas in Gauteng, Western Cape, Eastern Cape, KwaZulu-Natal, Free State and Northern Cape.
The lowest level of geographic aggregation covered by the data is metropolitan area.
The universe which the AMPS sample is drawn comprises virtually all South African adults aged 15 years or older throughout urban and rural areas of major metropolitans in South Africa. The survey covers residents in private households, or hotels, residential hotels and similar accommodation in the Republic of South Africa.
Producers and sponsors
South African Audience Research Foundation (SAARF)
The universe which the AMPS sample is drawn comprises adults aged 15 years or older in South Africa. A multistage, area stratified, quasi-probability design was employed on a full annual sample. Using population estimates, the sample is then grossed up to the total population so that findings can be looked at in terms of both percentage and thousands of people. The data were collected by personal, in-home interviews. In the case of each racial group, certain areas were excluded from consideration, as containing no persons or a negligible number of persons in a given group.
The AMPS dataset has the following weight variables:
1) Adult Population Weights – “pwgt”.
2) Household Weights (not to be used with personal demographics) – “hhwgt”.
3) Household Decision Maker Weights - to be used with personal demographics. These must be used for personal analyses on household items – only on a filtered of male heads of household and female housewives – “hhdwgt”.
4) Purchaser Weights – These must be used for the household products only on a filtered base of those wholly or partly responsible for household purchases – “hhpwgt”.
South African Audience Research Foundation. All Media and Products Survey 2012 [dataset]. Version 1.3. Johannesburg: South African Audience Research Foundation (SAARF) [producer], 2013. Cape Town: DataFirst [distributor], 2015. DOI: https://doi.org/10.25828/7756-k204