From instinct to infrastructure: fourteen years building a fellowship selection system

The Client

An internationally recognised university in a major Middle Eastern capital, partnering with a UK-based philanthropic foundation to deliver two flagship graduate fellowships. One focused on business and entrepreneurial leadership, the other on public affairs and policy development. Fully funded, highly competitive, and designed to cultivate a new generation of ethical, mission-driven leaders. The moral weight of the work was not lost on anyone involved.

The Challenge

The foundation was selecting for potential in a context where potential is difficult to measure. The applicant pool was exceptional almost by definition. These were high-calibre candidates who had already demonstrated academic excellence and personal ambition. The differentiation problem at the top end of a highly selective applicant pool is one of the hardest challenges in selection design, because the standard tools that distinguish between an average and a strong candidate often produce very little signal when everyone is strong.

The cultural dimension added further complexity. What leadership orientation looks like in a Middle Eastern context, what mission alignment means to candidates from different national and professional backgrounds, how to assess ethical reasoning in scenarios that carry different weight depending on who is reading them, these were not problems a standard Western selection framework could address without substantial adaptation.

And then there was the accountability dimension. The foundation was making a promise, to the scholars, to their families, and to the national vision it was serving. The selection system had to be rigorous enough to honour that promise year after year, not just in a good intake cycle.

We needed to stop relying on instinct and start building a system we could trust year after year. It was not just about fairness. It was about accountability to the future we are trying to build.

Senior Programme Director, Foundation

What We Did

Wharton Global began with structured discovery sessions involving both the university and the foundation's UK-based leadership, which surfaced something important immediately: the two organisations, despite their shared purpose, held meaningfully different views on what they were selecting for. Getting those views aligned upfront was what made the design possible.

Years 1 to 3: Building the Evidence Base

The initial framework was deliberately comprehensive. When you do not yet know what predicts performance in a specific context, you measure more, not less, and let the data tell you what matters. Candidates were assessed across personality profiling using the Big Five model for resilience, openness, and collaborative orientation; cognitive ability covering numeracy, logic, and reasoning; Situational Judgement Tests grounded in public and private sector dilemmas drawn from the region's specific context; and structured competency interviews assessing five capabilities: leadership, ethical reasoning, mission alignment, stakeholder awareness, and problem-solving.

The SJTs were written around scenarios that reflected the actual decisions the scholars would face in their careers rather than generic situations, such as resource allocation under political pressure, ethical trade-offs in public administration, stakeholder management across institutional boundaries. The cultural nuance of these scenarios was central to the approach.

After three years, the data was clear. Structured interviews consistently predicted long-term leadership potential. Cognitive testing added almost no differentiation at this level of applicant, because everyone scored well. Personality data provided context but showed less predictive validity than the interview data. This is a finding that many organisations resist, because cognitive testing feels objective and therefore reassuring. The evidence did not support keeping it.

Years 4 to 14: Refining, Cutting, and Deepening

Cognitive testing was discontinued. Standalone personality assessments were removed. SJTs were replaced by behavioural probes embedded within the structured interview. These were not cuts made for efficiency, they were made because the data showed they were the right cuts for validity. A selection process that carries components it cannot justify is not more rigorous, it is less honest.

What remained was deepened significantly. Interview question banks were refined across multiple cycles, with new questions developed and old ones retired as they became familiar to candidates. Scoring calibration workshops were introduced so that assessors were not just trained once but recalibrated regularly. Cross-rater validation was built in to catch drift before it distorted outcomes.

By year five, the process had consolidated into a ten-point evaluation framework that has been applied consistently across every cycle since. When the pandemic arrived in year nine, the programme moved fully online without loss of quality and without interruption to the cycle. The infrastructure that had been built over a decade proved robust enough to absorb a global disruption that broke many less well-designed processes.

The Results

When you move from instinct to infrastructure, it stops being just a process change and becomes something you can rely on. The foundation can stand behind every cohort selection knowing it was made through a system that has been continuously tested, refined, and validated against outcomes.

In year fifteen, a longitudinal alumni study is under development to measure sector-specific leadership impact, social return on investment over ten to fifteen years, and individual contributions to national reform priorities. The findings will shape the next phase of the programme and inform how the fellowship positions itself as a national investment rather than an individual opportunity.