S Shujahat Ali
While working as a consultant, I was once told about an educationist that “You get what you test for.” But in the case of CSS, a more troubling question emerges: do we even know what is being tested? Memory, perhaps?
At present, the system operates with a level of opacity that would be unacceptable in most modern evaluation systems. Yes, examiner reports are published, but these are generic, post-hoc summaries that neither capture the diversity of 10,000 candidates nor provide actionable insight to an individual who has failed. A candidate is left guessing: was it weak argumentation, poor structure, factual gaps, or simply harsh marking? You never get to know.
The opacity deepens further. MCQs and their correct answers are not consistently released. Full solved papers by examiners are unavailable. Rechecking mechanisms are limited to superficial recounting rather than substantive review. In effect, candidates are evaluated by a system that reveals neither its rubric nor its reasoning.
Contrast this with global best practices. The UK’s Civil Service Fast Stream provides detailed feedback to candidates, including performance across competencies such as analysis, communication, and decision-making. Similarly, the US’s Educational Testing Service, which administers exams like the GRE, publishes percentile distributions, scoring guides, and even sample high scoring responses. Candidates are not left in the dark; they are guided throughout all the process.
Even within Pakistan, transparency is not an alien concept. Many Boards of Intermediate and Secondary Education such as BISE Lahore routinely publish answer sheets of topper. This practice demystifies excellence, allowing students to calibrate their preparation against real benchmarks rather than hearsay.
CSS, by contrast, withholds precisely the information that would make preparation more rational and equitable. And, would cut out academies.
A modern examination system should publish, at minimum, core statistical metrics: mean scores, medians, and standard deviations for each paper. These are not academic luxuries; they are essential signals. A paper with an average score of 35 tells a very different story from one with an average of 65. Without such context, candidates cannot distinguish between personal failure and systemic harshness.
Equally important is the publication of scoring patterns. Are certain questions consistently underperformed? Do some subjects exhibit extreme variance year to year? Are evaluators aligned in their marking standards? These are questions that data can answer but only if the data is made public.
Reforms here are straightforward but powerful. First, release official solved papers written by examiners to establish a gold standard. Second, publish anonymized top-performing scripts across subjects. Third, introduce a robust rechecking system that allows candidates to understand, not just recount, their marks. Finally, institutionalize the release of statistical summaries for every paper.
Transparency is not merely about fairness; it is also about efficiency. When candidates understand the rules of the game, they prepare better. When evaluators are held to visible standards, they grade more consistently. And when the system communicates clearly, it builds trust.
Opacity, by contrast, breeds speculation, coaching myths, and wasted effort.
If CSS is to select the best minds for public service, it must first demonstrate a basic principle of governance: that power, whether administrative or evaluative, must be accountable.
Because in the end, a system that cannot explain its decisions cannot claim legitimacy for them.
S Shujahat Ali
Fulbright Scholar and Civil Servant
