Precision experiment: Percentage refusal density test. The panel's report

The percentage refusal density (PRD) test appeared to be suitable for assessing the compaction of dense roadbase macadam. However, before being used for specification purposes, it was necessary to determine the reproducibility of the test method. An inter-laboratory precision experiment has been carried out in which sixty core samples were removed from a motorway reconstruction contract and distributed among ten participating laboratories for testing. The results showed that the repeatability of the test method was 1.2, and the reproducibility 1.8, for adjacent single cores taken on a line parallel to the direction of rolling. Although these results were derived from only a narrow range of PRD, analysis of previous data showed that over a wider range the precision did not appear to depend upon the level of PRD. The precision estimates therefore apply to a range of PRD from 92.3 to 99.2.