Developing model schedules of female proportions single for India

The Indian Census provides marital status data cross classified by sex and age. There are, however, several limitations in the general state of information and the quality of data on marital status. For instance, the ratios proportions single in corresponding age groups in 2 consecutive censuses have been found to exceed unity in the absence of any significant migration. In the present study, the authors adjust the the schedules of proportions single for 1901-1971 with a logit linear model. In spite of the serious nature of errors in the data, the model appears to work well. In assessing its quality, the technique is applied to more accurate Japanese data, and found to have an extremely good fit. A comparison with the Coale-McNeil nuptiality model reveals that for the Indian situation the logit model may be more appropriate.