Regression Estimation 



Estimates of use and associated standard errors were generated for the three key 

 variables using simple linear regression estimation procedures. The three basic 

 regression equations used in this analysis expressed linear functions of the 

 corresponding covariate. 



The precision of total estimated summer use was high for all three key variables, 

 but unacceptably low for all fall-use estimates (table 7). The reasons for the large 

 standard error terms for all variables in the fall cannot be determined precisely. 

 Sampling intensity should have been adequate with a 29-, 14-, and 9-percent sampling 

 rate, respectively, in strata 5, 6, and 7. A zero registration rate likely accounted 

 for the lovv precision level in stratum 7, yet registration rates of 58 and 57 percent 

 were obtained for strata 5 and 6, respectively. The high standard errors may be a 

 function of the greater inherent variability of fall use and registration rates. The 

 average length of stay was a little longer during the fall than it was during summer, 

 but it was much more variable. Such use included (a) a mixture of deer and elk hunters 

 walking the lower elevation trails for a few hours, (b) fishermen and hikers out for the 

 day, and (c) mountain goat hunters camping for days in the high country. 



Summer and fall strata were pooled to generate season-long estimates of use for 

 the key variables using simple linear regression estimation methods (table 8). The 

 advantage of pooling all strata, of course, is the large sample size obtained (n = 110 

 2-day sample units). Fairly precise season-long estimates of use were obtained for 

 all variables. Pooling all strata can, of course, yield misleading results if some of 

 the strata are substantially different populations. 



Table 7 . --Regression estimates of veeveatian use on the Mission Mountains Primitive 

 Area^ Montana^ based on self -registration — interview relationships ^ 



summer- fall, 1968 



Stratum 



. Visits 



Total 



use 



. Groups 



" Estimate 



Standard 

 : error 



■ Estimate 



Standard 

 error •'■ 



■ Estimate 



Standard 

 error-'- 





number 



Percent 



Visitor-hours 



Percent 



Number 



Percent 



Summer 



3,622 



8.9 



39,544 



13.9 



910 



10.0 



Fall 



634 



49.0 



8,046 



59.8 



220 



36.7 



Summer + Fall 



4,256 



10.4 



47,590 



17.7 



1,130 



13.7 



1 



836 



4.2 



7,976 



7.1 



208 



5.3 



2 



765 



5.5 



5,666 



12.0 



214 



6.1 



3 



739 



7.2 



5,036 



17.0 



176 



8.5 



4 



1,282 



22.3 



20,866 



21.3 



312 



23.7 



5 



199 



28.6 



2,004 



48.0 



70 



22.8 



6 



277 



68.2 



4,314 



102.5 



99 



49.5 



7 



158 



72.8 



1,728 



112.4 



51 



59.0 



Standard error, expressed as percentage of the estimate. 



28 



