Problem: Superficially, it seems preferable to use objective performance data (such as productivity figures) when available, rather than subjective supervisory ratings to assess employees.
When might objective data be less effective as performance measures than are subjective ratings?