With the move towards performance-based criteria in many areas of the construction industry, it has become increasingly important to ensure that specifications are measurable and statistically robust. Measurements that are not repeatable and/or reproducible can lead to significant contractual disputes and litigation. Those developing performance criteria often assume that measurement systems are always correct, "a dangerous assumption at best," even if the data is generated logically, using approved procedures, and calibrated instruments/equipment. This paper seeks to highlight the sources and implications of inaccurate measurement systems, with particular reference to measurement 'bias' and high measurement variation. It draws on the findings of recent detailed studies undertaken by VicRoads and PQ Systems in conjunction with five industry laboratories regarding the measurement of specification criteria for crushed rock and asphalt products. This paper concludes that: specifications are more than a band of acceptability, the way they are applied and measured are as important as the specifications themselves. One of the dangers inherent in poor measurement systems is the propensity to manufacture defective product. This and some common misunderstandings that can result from inadequate measurement systems are also addressed in this paper. (a)
Abstract