## ＜学術雑誌論文＞よみ書き調査に於ける標本調査計畫の概要

その他のタイトル よみ書き調査に於ける標本調査計画の概要 作成者名 所属機関 所属機関名 統計數理研究所 The Institute of Statistical Mathematics 日本語 統計科学研究会 Research Association of Statistical Sciences 1949-12-20 3 1/2 1 10 Version of Record open access The Literacy Survey was performed under the direction of the Comittee organized for this purpose. This short report is that of sampling design in this survey, a part of works done by the statistical c...omittees. Under some restrictions, for example, time, expenses, or testers required, this design was made. (i) The form of this Survey The test materials having been constructed, the testees are to be made answer by filling in each entry of the test paper, instructions given in the same way as in a general intelligence test. In order to carry out the test as strictly as possible, the testees should be tested under the same condition. To do this, they are to be called to a certain place, where the test with its instructions is given within certain limits of time. (ii) Sampling Nest, making use of the results obtained from the pretests, we established a sampling design and a concrete plan for the survey. It is rather difficult to draw each individual with equal probability in such a large scale sampling. For this, it is very useful to divide the universe (the Japaneses) into homogeneous groups. Then we have to consider the factors affecting the abilities. The factors which are taken into consideration , for the stratification are naturally assumed and known by the pretest. These are followings. (A) Japaneses Area, (B) Urban and Rural Area. This reflects such characteristics as population size and density, economic-ecological structure and degree of urbanization. Thus the numbers of strata are 87 in urban area and 86 in rural. The sub-sampling method from these strata was adapted. The next step is to determine the number of testees to be sampled and to allocate to each test spot. In order to obtain the required precision, necessary sample size is 17,100. Including nonresponses which may not give any bias to the whole, the sample size was determined to be 21,003. Having fixed the sample size in this manner, we preceeded to allocate them. This sampling design, however, is intended for estimating the distribution of population in terms of scores. Allocation such as is optimum for a single label is considered to be improper in the survey consideration which is so complicated that the reading and writing abilities can be measured only by many ceiteria combined. Therefore, in view of the conveniences of the counting, the samples were allocated in proportion to the population in each stratum.続きを見る

### 本文情報を非表示

p001 pdf 792 KB 94

### 詳細

レコードID 12933 査読有 統計數理研究 || 3(1/2) || p1-10 Bulletin of mathematical statistics || 3(1/2) || p1-10 http://bic.math.kyushu-u.ac.jp/ 論述 学術雑誌論文 2009.04.22 2019.08.27