-41- 



height double sampling option was used, SEL must reflect both SE— and SE„, 

 the standard error due to height estimation. In this case, SE^ is 

 calculated as follows : 



^^voL = ^/ 4 ^ -H^ ^'4 ) 



Combined Strata Statistics 

 We have seen how INVENT calculates the mean volume/acre with its standard 

 error for each stratum in the sample. The final step in the statistical 

 analysis is to use the individual stratum statistics to calculate the total 

 volume across all strata along with the standard error. 

 Let L = number of strata sampled 



AC. = size of stratum i in acres 



N = number of sample units in stratum i 



/ 



n = number of units sampled in stratum i 



Y = mean volume per acre of stratum i 



VOL = total volume of stratum i = AC " Y 



^% 



OL = volume standard error of stratum i as a percent 



2 



2 2 2 

 VOL = (SE^Q, • VOL ) / 100 = variance of total volume 



1 

 then we calculate the stratified total statistics as follows: 



AC = Z AC = total acreage sampled 

 i=l ^ 



L 



VOL = Z VOL = stratified total volume 

 i=l ^ 



J I 'K 



»^0L "°« 



Having calculated VOL and SE^ , all that remains in calculating the 

 confidence interval is determining the appropriate percentage point of the 



Student's t distribution. In order to do this we must calculate the 



2 

 degrees of freedom associated with S . Satterthwaite (1946) states that 



2 

 the exact distribution of the complex estimate S is too involved for 



VULi 



