ABSTRACT
The construction quality of a lot is often measured by collecting and testing a few samples. The Acceptance Quality Characteristic (AQC) statistics, based on samples, are used to make an inference on its overall construction quality. The commonly used Quality Measures (QM) include average, and a measure of variability (i.e., standard deviation or range) of an AQC from the samples. While a contractor is paid an incentive if the measured quality is better, he may be penalized if the quality is worse than specified. Average AQC, if used as a quality measure, can be manipulated by the contractor to meet the specified quality; thus increasing the construction variability. Variability in AQC alone as a QM can be misleading. For example, the construction quality of a lot could be bad overall but have low variability. Therefore, combining the AQC average and variability into a single QM may better represents the true lot quality. This QM is called the Percentage Within Limits (PWL). The PWL represents the percent of a lot constructed within the specified limits of an AQC. Any incentives or disincentives paid to the contractor can be based on the PWL.
The work presented herein shows why PWL is a better QM than just the sample average. Air void contents after construction were used as the AQC with two-sided specification limits. Different sampling methods used in the paper include: (a) random sampling with replacement, (b) stratified sampling, and (c) random sampling without replacement. For each sampling method, samples of sizes 3, 5, 10, and 20 were used to compute the average and PWL for air voids contents. The PWL of a lot after construction was determined based on the commonly used air void content specifications limits (i.e., lower and upper limits of 2% and 8%, respectively). The results of the analyses show that the PWL represents the lot quality better than the sample average i.e., two lots that have the same average AQC could have different PWL values. In addition, it was found that for different construction qualities and sample sizes, the percentage of times a contractor gets full pay is more if average is used as a QM instead of the PWL.
