Measuring the performance of engineering: the data fusion problem

People need assistance when assessing performance based on multiple measures. It is simply not a part of people's intuitive judgment to be able to weigh up more than a few factors simultaneously. Since performance factors are manifold, setting goals and providing feedback to engineers involves data fusion: assembling multiple measures that tell people what, on balance, is good performance and what is bad. The author discusses explicit ways of doing data fusion, data envelopment analysis, and the relationship with prices and weights.