Data scientist Shramik Misra from the Kantar Africa and Middle East analytics team; chief growth officer for the region Karin Du Chenne; and socio-economic measurement pioneer Neil Higgs; ably supported by analytics manager, South Africa, Kantar, Prenolan Munsamy and analytics director, South Africa, Kantar, Kent Diepraam, were presented with the award.
With the first version of the Pan-Africa Socio-Economic Solution or PA-SES created for Pamro/Esomar five years ago, the update comes as a way of better understanding and comparing socioeconomics across the continent. More countries have been included this time around, based on independent source Afrobarometer’s most recent round 7 multi-country data.
Now covering 76% of the continent, this means the refreshed sample design has increased from 32,400 respondents to 45,823 and covers a more diverse base of people, languages, cultures and geographic spread.
This is crucial as the average Gini-coefficient in Africa is 42, indicating a large difference between the richest and poorest or highest and lowest socio-economic groups. In addition, 44% of all consumption comes from the bottom end, so there needs to be a concerted effort to accurately measure not just those with the biggest bank balance.
Due to Covid-19, there’s been huge migration from face-to-face field to methods that don’t involve visiting people in their homes. In addition, according to GSMA, mobile penetration is increasing daily and is the primary means of accessing the internet across Africa. “…bringing the total connected population to more than 3.5 billion people globally. For many of these individuals, mobile is the only method of accessing the internet, so growth in mobile internet adoption also drives digital inclusion, especially in low- and middle-income countries.”
Internet penetration, therefore, means mobile penetration in Africa, which in turn required the original gold standard CAPI as best-in-class measure to be whittled down for short and sharp responses and quick mobile measurement, including household and individual-level variables for good levels of stability and granularity.
This meant iterating 38 times against a Burt matrix and correspondence map of the meaningfulness of each, until the most discerning variables remained, resulting in an improved accuracy rate that sufficiently differentiates on both ends of the socio-economic continuum. The longevity of the variables goes to long-term stability of the measure.
On measuring the changing landscape of Africa, Higgs shared that his first-ever paper, written four decades ago, was on the same topic, proving that the desire for this type of measure has been a constant in the research industry.
A household’s socio-economic status is an important construct for marketers as it measures that household’s level of wealth and access to services, and by extension, the level of participation in the formal or informal economy. This guides marketers in targeting their products in more meaningful ways to the correct group of consumers.
When it comes to socioeconomic systems, it’s more effective to measure people on a scoring system across a continuum from low to high rather than in distinct groupings, for greater longevity across tech changes. As such, the gold standard scorecard has an individual scoring system of structural and personal items, like frequency of internet use.
There were good correlation and similar respondent classification across the mobile version and the gold standard, with the final magic formula suitable for all African countries on mobile. There is also good correlation between versions 1 and 2 of the measure, which makes for easy transitioning to the update.
Those on the lower end of the SES spectrum are more likely to have smaller houses and less access to what the world has to offer, particularly transport and education. Those on the higher end of the spectrum have houses made of modern materials and better infrastructure, as well as better access to communication, transport, jobs and education.
Running the average across all countries included, the gold standard and mobile correlation show Madagascar and Niger and Malawi on the lower end of the spectrum, with South Africa, Morocco, Mauritius and Tunisia at the higher end. When mapped by SES score, input variable trace curves showed solid correlations across the measures and ability to discern differences in SES with a consistent measure, ideal for planning.
The stable, non-proprietary gold standard CAPI and shorter mobile versions are both available for all in the industry to use across Africa. All materials, questionnaire guidelines and scoring systems will be made available on the Pamro website shortly
