Sampling errors (standard errors of estimate) on the other hand do 

 not involve human errors but rather are theoretical measures of the 

 reliability of estimates based on the variability exhibited by 

 sample measurements. They generally vary inversely with the square 

 root of the number of samples and directly with the square root of 

 the unsampled proportion of the total population. Hence , they can 

 be controlled by altering either the number of samples, the size of 

 individual samples, or both. 



Analysis of sample variations indicate that the standard errors of 

 estimate for the unit as a whole are _ 1.8 percent for total forest 

 land, t 2.4 percent for commercial forest land, and i 3.9 percent 

 for noncommercial forest land. Accordingly, the probabilities are 

 2 out of 3 that the total forest land, commercial forest land and 

 noncommercial forest land are, respectively, within 1 85,000, 

 1 87,000 and 1 48,000 acres of the estimated areas if measurements 

 and computed errors introduced no bias. 



In determining timber volumes, the possible sources of error in- 

 clude in addition to those cited above (3) inaccurate measurement of 

 sample plots, tree diameters, tree heights, and cull, and (4) bias 

 resulting from improper construction, selection, and use of tree- 

 volume tables. All reasonable effort was made to eliminate errors 

 from these sources. The standard error of the cubic-foot volume 

 estimate of the primary growing stock is 1 3.3 percent. Accordingly, 

 the probabilities are 2 out of 3 that the actual volume of the 

 primary growing stock is within t 163,000 M cubic feet of the esti- 

 mated volume of 4,939,000 M cubic feet. 



The reliability of one statistic as compared with another presented 

 in the same or a related table can be judged roughly by its relative 

 magnitude. In general, the larger quantities warrant greater con- 

 fidence; the smaller quantities indicate only relative magnitude. 



-18- 



