Search results

In 1959, a cross-section survey on pol'al awareness, att's, & COMM's was conducted in the 12 provinces of Laos. The sample consisted of about 500 R's in Vientiane, 480 in the 11 provincial capitals, & 1,000 in more than 52 provincial villages. Because of the lack of a direct pop census in Laos in modern times, it was necessary to train men with no previous interviewing experience as interviewers (I's). Most of the I's used were provided by the Laos gov, & had been trained to work on public improvement programs in Lao villages. Initial I training was done over a 3-day period, with additional training provided during the pre-test & study period in Vientiane. In addition, after a group of I's had gained some experience in village interviewing, they met with other I's to relate experiences & provide further training. In the absence of sampling materials, I training & instructions were provided for each of the 4 teams of I's, including specific instructions on how to proceed in drawing the sample for specific situations, & whom to interview within sample households. A major problem was that of proper I supervision in the field. It was the job of supervisors to contact provincial governors & village headmen before interviewing, provide transportation for their teams, supervise the sampling, & review questionnaire's. It was not until some time after field work had begun that 2 of the 4 teams had satisfactory Lao supervisors. For future res in developing countries, it is recommended that training & written instructions be provided I's, so as to reduce individual judgement & deviation on their part. With previously untrained I's, gaps in specific information generally result in omission & error. (See SA 0827-A9630). Modified AA.

The retrospective question is often used to obtain R's own judgments about his changes in attitude over a specific period of time. A test was done among a panel of 918 Coll S's of the question, 'In what way has your evaluation of religion changed, if at all, since you came to Coll?' The responses in 1952 were as follows: I personally value my religion more - 39%; I have not changed my evaluation - 41%; I personally value religion less - 20%. It was thus possible to relate stability or change on the retrospective question with stability or change on several panel questions asked in 1950 & 1952. The following findings emerged: (1) there was a distinct tendency for R'; whose att's did not change on the test items to report that their religious att's had changed, & (2) among R's whose att's changed, there was a good likelihood that their retrospective answers on the direction of their change would correspond with their actual direction of change. Several possible explanations of these findings are offered. AA.

AbstractAn algorithm for calculating the probabilities of a summed multinomial density function which is recursive with n (the number of trials) is presented. Having application in inspector error models for auditing and quality control problems with Cartesian product structures, the algorithm is discussed in the context of computing optimal economic sampling plans. Computational experience with the algorithm is presented.