Does a More Precise Score Make a Judgement More Objective?

A selection panel scores two candidates at 82.4 and 82.1. The result looks more objective than calling one “better” and the other “slightly weaker”. Yet extra decimal places primarily add precision: they express a difference more finely, without showing that the difference is real, important or correctly measured.

Before a score exists, people have already decided what to measure, how much each element should count, and how experience, communication and work samples will become numbers. If “communication” is scored mainly through interview fluency, a rule applied consistently to everyone may still miss the written collaboration the job actually requires. A reproducible procedure can reduce an assessor’s moment-to-moment discretion while reliably repeating a biased standard.

Philosophical discussion of scientific objectivity no longer rests only on Thomas Nagel’s ideal of a “view from nowhere”. Work by Kuhn and others emphasises that observation and evaluation occur within concepts and shared standards. Later accounts therefore also locate objectivity in practices of public criticism, calibration, replication and correction. Numbers can assist these practices by making standards and differences easier to inspect, but they cannot justify the standards themselves.

My view is that objectivity does not remove judgement from human hands; it makes judgement answerable to shared tests. Faced with 82.4 and 82.1, ask whether the scoring error exceeds 0.3, why the weights were chosen, whether another qualified assessor would reach a similar result, and whether a candidate can identify and appeal a factual mistake. Without such conditions, a precise number merely lends unexamined choices an air of certainty.

https://plato.stanford.edu/entries/scientific-objectivity/