13 14 Survey sampling with model-assisted estimation has gained popularity in forest inventory 15 recently. Another option for utilizing the auxiliary information is to use post-stratification, which 16 is a special case of model-assisted estimation with class variables as explanatory variables. In 17 this study, we compared the efficiency of post-stratification with increasing number of strata to 18 model-assisted estimation. We carried out a study based on a simulated population. We 19

more »

... red four different types of post-stratifications, namely (i) stratification based on 20 predictions of a linear model, (ii) stratification based on a regression tree model, (iii) 21 stratification based on the first principal component of the explanatory variables, and (iv) 22 stratification based on the regression tree model with the first principal component as the only 23 explanatory variable. Furthermore, we examined both the traditional post-stratification mean and 24 variance estimators and the difference estimator and its variance estimator for post-stratification. 25 Within the recommended range of number of strata, the model-assisted approach was more 26 efficient than post-stratification. With a large number of strata, post-stratification produced 27 smaller standard error of estimates, but problems such as empty strata were encountered with 28 small sample sizes. Using the first principal component directly for stratification or as an 29 explanatory variable was the most efficient approach. 30 31