Digital tools are increasingly vital in addressing complex challenges such as youth skill and opportunity gaps and high unemployment in many parts of the Global South. Catalyzed by adaptations due to the COVID-19 pandemic, the online interface is now increasingly relevant in education from remote learning and online certifications to job matching.
These technologies democratize access to educational content, benefiting millions gaining internet access each year in emerging economies. As always, along with the promise they hold, there are a plethora of pitfalls.
CHALLENGES IN EDUCATION AND OPPORTUNITY MATCHING
Numerous platforms have emerged that aim to address youth training needs, such as digital skills in IT, web development etc. They also provide key staffing for businesses and organisations struggling to identify talent in tight labour markets and to fulfill their CSR goals.
Yoma, an ecosystem of partners including but not limited to UNICEF, Atingi, UMUZI, and RLabs is one such example, connecting youth to opportunities for learning and earning in over 8 countries. The government of South Africa has also developed the SAYouth platform to address these issues and tackle the country’s unemployment crisis. Likewise, in Germany, the ReDI School of Digital Integration upskills youth with a migrant or marginalized background to accelerate integration into the digital economy.
Despite the benefits from these types of programs being globally or nationally accessible, they often struggle to localize to the needs of culturally and socially-diverse users, especially those belonging to marginalized populations.
Delivering effective and equitable matching with opportunities thus becomes a challenge, as candidates from different countries, with different backgrounds and levels of education are exposed to and compete for the same training or earning opportunities.
Additionally, not all opportunities can be available to every young person, and in many cases spots for learning or earning opportunities are limited. Selection processes, often ‘funnel’-based, lead to many youth being excluded due to mismatches with predefined criteria or aptitude tests specific to the job or training opportunity.
These selection structures pose two major challenges.
CHALLENGE #1 Due to the nature of the funnel, a large population of youth fail to receive benefits as they are filtered out. In doing so, this also reduces the candidate pool for employers. Each failure to match a candidate to training or opportunities can be seen as a lost chance to deliver impact - we run the risk of excluding the most vulnerable, the population we seek to reach.
CHALLENGE #2: Bias in the filtering process further compounds inequities, as candidates may be funneled out by the selection criteria for factors such as educational attainment, language skills, time availability and internet access. Even though this is a data driven approach to identifying the best candidate, there are underlying risks of bias in the selection process. For example:
Educational attainment achieved as a proxy for gender: in many contexts, women are less likely to reach the highest levels of schooling.
Time availability as a proxy for gender: women with children will have to spend their time on childcare. As such, women might be less likely to be able to commit to the requisite amount of time for training programs.
Language as a proxy for ability: if the selection process for training programs includes language assessments, this can filter out candidates who have a strong ability — for example in STEM subjects — but who do not get through the application process because they misinterpret test questions.
Internet access as a proxy for economic status: if a candidate only has limited access to the internet, they may be rejected from a training program, but this might well be as a result of their socio-economic background.
Whilst it is logical that training and career development programs look to identify candidates with a suitable CV for their programmes, it is also important to account for such biases during the selection process.
Indeed, these challenges are not new, as the development sector has long struggled to find a balance within ICT between innovation and equity/inclusivity. So, how can we best leverage the potential of these innovations in education, without leaving the most vulnerable behind?
RECOMMENDATIONS
Although more complex, the recommended approach offers a suitable opportunity pathway tailored to the needs of each candidate. At Outsight, we build on the range of expertise offered by our network of associates in order to deliver quality results adapted to the specific tasks at hand.
We believe it is developing these kinds of complex approaches that maximizes sustainability, effectiveness and impact.
Transitioning from Funnels to Matching
A strategic shift: Moving away from funnel-based selection to matching candidates with suitable opportunities can reduce dropout rates and bias in the narrow selection process.
Opportunity for all candidates: Rather than narrow criteria focused on a specific role or type of training, programs could evaluate candidates based on diverse skills and match them to relevant opportunities (ie. mentorship and entrepreneurship training or further skills development activities)
The result? A larger percentage of candidates access an opportunity and are matched to those where they can succeed. This approach is structurally designed to broaden access to opportunities for a wider range of candidates with diverse backgrounds and talents.
2. Proactively address bias with a data driven approach
Identify and Mitigate the Impact of Bias: A matching approach opens up more paths, considering a richer view of candidates and their context. As such, it reduces the impact of bias in underlying data and data-driven algorithms.
Using Data to Understand Bias: Instead of “weeding out” candidates, aptitude tests could help counteract bias by collecting candidate age, gender, education, location, internet access, education, refugee status etc. The resulting dataset provides an opportunity to better understand the variables and structural biases that determine whether an applicant possesses the relevant skills to pass the test.
Exploring algorithms: Using a systems approach and data analysis, programs can develop more complex and useful algorithms that prioritize equity of opportunity for youth.
Ultimately, for programs to be transformative and truly unlock new opportunities for young people, the methodologies they use must place equity at the center.
about the authors
Denise Soesilo
Denise is an expert in social innovation particularly in humanitarian and development settings.
Maria Zaharatos
Maria is a consultant specializing in research, program design, partnership development, and organizational systems, who champions co-creation and engagement with stakeholders. Her areas of focus are green education, youth empowerment, and workforce development. She has supported various education-focused programs and organizations, including UNICEF, where she helped develop key partnerships and implementation strategy for Yoma’s scaling across the East and South African region.
If you’d like to discuss working with the Outsight team, please get in touch or follow us on LinkedIn for regular updates.