d) Cohort badges with readings that are markedly different from all other cohorts and whose badge wearerappears to be a poor exemplar for the cohort composition. For example, for the period 1 through 8 May on COCOPA,the badge wearer for a cohort of twelve enlisted men was a Hospital Corpsman First Class. He had a recorded reading of 3150 mrem. The cohort consisted of ships cooks, storekeepers, stewardsmen, and one seaman, most of whom werein the same cohort for three other badging periods, with readings of 190, 0, and 175 mrem (all below the overall averages for those periods). It is doubtful that a hospital corpsman could have received such a dose. Stipulating that he did, it is very unlikely that the other members of the cohort had similar exposures. These and other similar examples, such as obvious alphabetical cohorts with disparate rating groups, generated a need to developa set of rules for interpretation and evaluation of cohort badging data. The approach adoptedis illustrated in tables 6.1 and 6.2. As indicated by the wording of the entries in the tables, the resultant two-step screening processis qualitative and requires experienced judgmentin application. As applied in this evaluation,the processis a useful tool. The first step, indicated in table 6.1, consists of a general evaluation of the apparent statistical validity of the results of cohort badging of a given unit for a given period. Theresults are then compared with the reconstructed dose for the period. If it is found that the average reading of the cohort badgingfor the periodis significantly higher than the reconstructed dose, but the overall quality of the badging procedure is evaluated as low in all or nearlyall of the criteria in the table, the reconstructed dose should be assigned. In all other cases, it may be advisable to assign the higher of the two values. Table 6.2 summarizes the results of the cohort dosimerry analysis. In units with more than one cohort badging period,there are significant variations in the memberships of cohorts. Therefore, the table is applied to each badging period and in the context of the preceding evaluation in table 6.1. Where a cohort badge readingis significantly higher than the averageofall the cohort badges for the period, but the validity of assignment of the indicated dose to an unbadged individual in the cohort is generally low, the calculated dose is more credible. Figures 6.1 and 6.2 summarize the cohort dosimetry data available for RECLAIMER and SHEA, respectively. These two ships have similar exposure scenarios (both provided support for Project 3.4 during the same time-frame), and the radiation environments in which they operated 91