stratifying by area 



Stratification by areas, that is, dividing a sale into blocks of known 

 difference in regard to timber size or quality and sampling each block 

 independently, is good sampling procedure. Ordinarily, however, a 

 stratification by volume classes will eliminate more variation than 

 area stratification because of the mixed size classes on most sales. 

 An important advantage of area stratification is that sales can be 

 divided into operating blocks and separate volume determinations made 

 for each block. In cases of reappraisal or change of operators on a 

 given sale, block stratification might prove to be of great value. 



Table 2, — Comparison of number of trees needed in sample to meet an 

 accuracy of 2.5 percent without and with stratified sampling 



Total trees 

 in 

 sale 



(number) 



Number of trees 



needed in sample 





V/ithout 

 stratified 

 sampling 



With stratified sampling 



Total 



Large trees 



tallied 

 completely 



Sample of 

 "medium" 

 stratum 



Sample of 

 "small'' 

 stratum 





Number 



Number 



Number 



Number 



Number 



3850 



1080 



275 



113 



32 



420 



1192 



435 



104 



30 



41 



175 



In both tree measurement and sample tree measurement, it is difficult 

 to determine exactly how much volume has been cut at any given time. 

 If the sale area is divided into blocks, preferably according to timber 

 size and quality for full value of stratification, then the determination 

 of the amount cut at the end of a season, for example, is greatly facili- 

 tated. For ease in determining the amount cut periodically it might 

 prove necessary on large sales to sample by separate blocks even if 

 blocks were divided only according to a logging plan v;ith no stratifi- 

 cation of volume classes associated with them. Sampling by blocks would 

 mean, of course, a larger sample proportionately in order to get the 

 desired accuracy for a specific block, but with the use of volum.e 

 stratification within blocks, the sample size might still be kept small 

 enough for practical purposes. 



-14- 



