84 CHAPTER THREE for the survey; institutionalized persons (e.g., residents of hospitals and persons living in military installations) were not eligible. Not all eligible persons in survey sites participated in the survey, though. Figure 3.1 presents a schematic display of the sample results as various additions and deletions were made to the survey sample. The initial survey sample consisted of 105 sites, 97 rural and 8 urban. Survey enumerators visited these 105 sites and enumerated a total of 46,972 persons as residents of the households in those sites. However, during the implementation of the survey, it was discovered that in all six sites of the Far West Terai, which included 3,356 enumerated persons, the survey protocol (especially that portion concerning the turning of eyelids on all sample persons) had been incorrectly applied. Survey staff were concerned that this would cause serious error in estimates for trachoma from that important area, but since treatment for trachoma had already been given to many persons in these sites, it was not possible simply to return and collect the data correctly. Consequently, these six sites were dropped from the sample (and from the population), and a new sample of six sites was drawn from the same stratum. These six new sites were selected and visited by survey teams and a total of 3,879 persons were enumerated in those sites. Thus, for the final 105 survey sites (91 initial rural, 6 replacement rural, and 8 ur- ban), a total of 47,495 persons were enumerated. A more complete discussion of this problem is presented in Chapter 2, Section 2.2. In one of the six replacement sites, a special study had been conducted to assess the importance of observer reliability to survey findings (see Section 2.4 for a description of the study). A total of 590 persons had been enumerated for this study, and two records had been retained for each person in the data file as part of the study. Thus, the actual number or persons enumerated in the final 105 survey sites was 47,495 - 590 = 46,905. Not all of these persons chose or were able to participate in the survey examinations. A total of 4,257 persons or 9.1 percent of those enumerated were not present in the site during the entire sur- vey examination period. An absentee rate of 9.1 percent is disturbing, but with a mobile population such as exists in Nepal it is not a surprising finding. Many Nepalese can be found visiting or working in India, and many others migrate temporarily to urban areas for employment. Among the remaining 42,648 enumerated persons present in the site at the time of examination, 2,761 persons (6.5%) were not examined for various reasons. For the most part, these persons
