Why Examinations May Be Testing the Wrong Things
For generations, the written examination has been the cornerstone of educational assessment. Students revise, memorise, and perform under timed conditions, and their results are taken as a reliable indicator of their intelligence and potential. Yet a growing body of research suggests that this model may be measuring the wrong qualities altogether — rewarding recall over reasoning, and compliance over creativity.
The most obvious criticism of traditional exams is their emphasis on memorisation. In an era when virtually any fact can be retrieved within seconds from a smartphone, the ability to store and reproduce information on demand seems increasingly redundant. What employers and universities consistently say they want from graduates — critical thinking, adaptability, collaborative problem-solving — are precisely the skills that sit-down examinations are poorly equipped to assess. A student who performs brilliantly in a three-hour paper may struggle to work within a team or navigate an unfamiliar problem without a rigid framework to follow.
Proponents of the current system argue that examinations serve an important gatekeeping function, providing standardised, comparable data across thousands of candidates in a way that portfolio assessments or project-based evaluations cannot easily replicate. There is also the question of academic integrity: when students submit coursework completed over weeks or months, the role of parental assistance, private tutors, or — increasingly — artificial intelligence becomes difficult to verify. A supervised exam, whatever its limitations, at least ensures the work is genuinely the student's own.
The most persuasive case, however, is for a blended approach. Rather than abandoning examinations entirely, educators could treat them as one component within a broader, more diverse assessment framework. Regular presentations, collaborative projects, and reflective essays can capture dimensions of learning that no timed test can reach. Assessment, after all, should serve the learner — not merely sort them. Until educational institutions are willing to question the metrics they have inherited, they risk producing graduates who are exceptionally good at passing exams and rather less skilled at everything else.
