Composite indices provide a convenient means of standardising the effects of incongruous parameters, such as those encountered in the management of pavement distress. With these indices, the relative needs of one pavement can be compared with the needs of another even though they experience different distresses. This particular application often leads to a priority list for resource allocation. Composite index development is studied to learn the extent to which statistical analyses will support objectivity in the management of pavement distress. The major findings include the following: (a) The systemwide average composite index is an excellent estimator of the true value. (b) Composite indices are generally effective in sorting pavements according to their relative need. (c) On an individual, section-by-section basis, composite indices do not and cannot reliably identify the upper population percentiles generally sought. Approximately one-third of the systemwide worst 20% may fail to be identified as such. (d) Relatively coarse increments of measurement can be used in multiple parameter indices with small practical effect on index efficiency, facilitating systemwide data collection. (e) Statistical analyses will indeed help -- but must not be allowed to hide -- the extremely subjective process of composite index development.
Samenvatting