Standardised assessments are attractive at graduate scale because they are cheap per candidate. What they measure is narrower than their usage implies.
What a test genuinely does
It ranks a large population quickly on a narrow ability, which is valuable when you cannot interview everyone. That is the whole of its value — it is a throughput instrument, not a quality instrument.
The coaching problem
Any widely-used assessment format in India is coached, which compresses the score distribution and weakens its discriminating power over time. Scores tell you increasingly more about preparation access than about ability.
Use it as a floor, never as a ranking
Set a threshold that removes the clearly unprepared, then discard the scores. Ranking a shortlist by test score reintroduces every bias the test carries, and it systematically disadvantages candidates without coaching access.
Where we use assessment data
On our own shortlists, assessment history is shown where a candidate has actually taken one of our tests, and marked "not attempted" where they have not. It is context alongside the profile — never a substitute for your interview.
Key takeaways
- Assessments are throughput tools that rank quickly on a narrow ability.
- Coaching compresses score distributions and erodes discriminating power over time.
- Use scores as a floor, then discard them; ranking by score imports the test's biases.
- Assessment history is context on a shortlist, never a replacement for interviewing.
Questions
What threshold should we set?+
Low enough that it only removes the clearly unprepared. A high threshold is doing ranking work under another name, with all the same problems.
Are coding tests different from aptitude tests?+
Somewhat — they are closer to the work. They are still coached heavily, so the same floor-not-ranking discipline applies.