This exercise and example represent, in very rough form, the technique of
creating behaviorally anchored rating scales.
As explicated by Smith and Kendall (1963), the complete process would include
generating statements like those above -- called critical incidents -- plus two other
critical stages: retranslation and scaling.
In the retranslation activity, critical incidents of performance (e.g., scheduling
people, training others, planning work for self, and all of the dimensions of performance)
are put on cards and mixed up. Then, knowledgeable individuals get together and sort all
of the critical incidents into categories (performance areas), without knowing what
categories they came from in the first place.
In the scaling activity, the critical incidents for a specific job performance
dimension are assigned scale values such as those shown above. The result is a rating
scale with behavioral examples at various levels of performance. This final scaling is
assigned through a consensus process in conjunction with some statistical analysis.