A Statistical Analysis of Rural Community Cards

Research Project

Project/Area Number	61450022
Research Category	Grant-in-Aid for General Scientific Research (B)
Allocation Type	Single-year Grants
Research Field	社会学(含社会福祉関係)
Research Institution	TOHOKU UNIVERSITY
Principal Investigator	NISHIDA Haruhiko Tohoku University, Faculty of Arts and Letters., 文学部, 教授 (20027949)
Co-Investigator(Kenkyū-buntansha)	HASEGAWA Keiji Tohoku University, Faculty of Arts and Letters., 文学部, 助手 (00198714) 片瀬一男東北大学, 文学部, 助手 (30161061)
Project Period (FY)	1986 – 1987
Project Status	Completed (Fiscal Year 1987)
Budget Amount *help	¥3,300,000 (Direct Cost: ¥3,300,000) Fiscal Year 1987: ¥800,000 (Direct Cost: ¥800,000) Fiscal Year 1986: ¥2,500,000 (Direct Cost: ¥2,500,000)
Keywords	Quantitative Sociology / Rural Sociology / World Census of Agriculture and Forestry / Rural Community Cards / Principal Component Analysis / Latent Profile Analysis / Descriminant Analysis / 複合標本方式
Research Abstract	1. We applied principal component analysis in terms of 7 items to the 3% random sample of rural community cards 1980. We obtained two components, "agricultural activity." and "de-agriculturalization." with the cummulative proportion of 0.73. 2. On the basis of means and standard deviations of the items and coefficient vectors of the principal components. we can calculate principal component scores of any rural community and show its position in the Japanese rural communities. We took as examples 29 communities in Kashamadai. Miyagi Prefecture. 3. We obtained a solution of latent profile analysis with 5 items. We classified rural communities into therr latent classes: 1) rural communities with high agricultural activity, 2) communities with moderate agricultural activity, 3) communities where agriculture is inactive. We utilized discriminant analysis in order to determine to which class any sample community belongs. 4. We proposed a method for detemining to which class any community belongs: firstly, calculate discriminant component scores of the community in terms of coefficients of the discriminant functions; second, measure the Euclidean distances between the scores of the community and that of the centroids of the classes; finally, find the class which minimizes the distance. 5. We investigated the "rates of farming families having a male as full-time farming sucessor" in the latent classes. The proportions of communities at the rate of 0% significantly differs among the three classes. 6. Remeining problems are as follews: 1) exploration of factors affecting the "rates of farming families having a male as full-time farming successor," and 2) comparative analysis of rural community cards 1985 with those in 1980.