Abstract
Highway engineers must conduct pavement evaluations before making rehabilitation decisions. Deflection tests performed at a uniform spacing are often used for structural evaluation purposes. The spacing between adjacent tests is often determined empirically. For pavements with large variability, more tests are needed. Variability is usually measured by the coefficient of variation (CV). Several methods of determining test spacing were reviewed. A new method is proposed to calculate the test spacing based on the equivalent coefficient of variation (CVe), which considers not only the variance but also the autocorrelation among the pavement responses. The new parameter CVe is more sensitive to test spacing than is the CV.