SYSTEMATIC SAMPLING FOR NON-LINEAR TREND IN MILK YIELD DATA
The present paper utilizes systematic sampling procedures for milk yield data exhibiting some non-linear trends. The best fitted mathematical forms of non-linear trend present in the milk yield data are obtained and the expressions of average variances of the estimators of population mean under simple random, usual systematic and modified systematic sampling procedures have been derived for populations showing non-linear trend. A comparative study is made among the three sampling procedures for five data sets by calculating average variances using best fitted trend equations. Usual systematic sampling is found more precise than simple random and modified systematic sampling for four data sets whereas modified systematic sampling is found better than the other two procedures for one data set. Estimates of parameters for these trend functions are obtained with the help of SAS software.