654 



TIMBER RESOURCES FOR AMERICA'S FUTURE 



MULTIPLYING FACTOR FOR STANDARD ERROR OF TOTAL 









































































































































































































N, 



























X 

































V 



V. 



































^Ns 



^ 





























^ 



V. 































^s, 



\ 



v^ 































^s 



v 



\ 



\ 





1 3 4 5 6 8 10 20 30 40 50 60 80 100 

 i PERCENT OF AREA OR VOLUME IN A BREAKDOWN 



Figure 1. — Ratio of standard error of an area or volume breakdown to standard error of an area or volume total. 



percent of the total sawtimber volume in the 

 Pacific Northwest Region. The multiplying factor 

 for 49 percent is read as 1.4. From table 83, the 

 sampling error of the estimate of total sawtimber 

 volume in the region is 1.3 percent. Therefore, 

 the sampling error of the Douglas-fir volume figure 

 is 1.3 X 1.4, or 1.8 percent. 



The procedure for estimating the sampling error 

 of breakdowns of growth is parallel to that illus- 

 trated for area and timber volume. 



NONSAMPLING ERRORS CANNOT BE 



Measured 



In using Forest Survey procedures, sampling is 

 probably the major source of error, but mistakes 

 and errors in judgment are also possible. The 

 magnitude of these other errors cannot be meas- 

 ured. However, mistakes are kept to a minimum 

 by careful checking of photo interpretation, field 

 work, and office compilations, and by extensive 

 review of the resulting tabulations. Professional 



judgment is probably less of a factor in the Forest 

 Survey estimates than elsewhere in the Review 

 because Survey procedures consist largely of 

 routine measuring with relativeh' few opportu- 

 nities for the exercise of judgment. 



The area data may suffer from misinterpretation 

 of aerial photography, from mistakes in classifying 

 land uses, or from failure to apply correctly the 

 proper definitions of forest types, stand-size class, or 

 stocking. The volume data may contain mistakes 

 in tree measurement, misapplication of volume 

 tables or log grades, or incorrect converting factors 

 and cull percentages. The growth data are 

 affected by possible mistakes in counting annual 

 growth rings, in applying tables of average height 

 by species, and in estimating ingrowth and 

 mortality. Some of these errors are undoubtedly 

 compensating but there is no way of telling how 

 much. 



One other possible source of error in the area, 

 volume, and growth data lies in tlie procedures 

 used for adjusting the data to a common date. 



