ABSTRACT
A study is reported of the judgement strategies used by engineers in estimating task durations in complex development projects. These judgements are critical to commercial survival, yet much of the prior experimental literature suggests that human predictive judgement suffers from a number of critical flaws. The work involved 20 unstructured interviews, direct observation of three engineers, and the collection of traces: 300 completed estimating forms. The transcripts and traces were analysed qualitatively and inductively. Several findings were consistent with the prior, experimental literature but several were inconsistent. Much of the analysis suggested that judgement strategies were in fact strongly situated: they were explicable more in terms of the environment in which estimators worked than the estimators5 native cognitive processes.
