nonindustrial private ownerships in Vermont were analyzed. 



Probit analysis was used to estimate the relationship 

 between a dichotomous dependent variable, coded "1" if 

 the woodland was used for recreation and "0" otherwise, 

 and selected explanatory variables. Probit procedures can 

 be used to estimate the strength of correlations between 

 recreational use of forest land and selected characteristics, 

 as well as the probability that parcels with a given set of 

 characteristics will be used for recreation. Judge and others 

 (1982) and Aldrich and Nelson (1984) provide detailed 

 discussions of probit models. 



Results 



Much privately owned woodland is used for recreation. 

 Approximately 77 percent of the 258 sample plots were 

 within ownerships that were used for recreation. The estimated 

 probability that woodland was used for recreation was 0.57, 

 evaluated at the mean values of the explanatory variables. 



Table 1 provides a brief description of the variables, and 

 Table 2 shows the probit results and the estimated elasticities. 



Table 1. Definition of variables 



Variable Definition 



REC Dependent variable, coded "1" if woodland is 

 used for recreational purposes and "0" 

 otherwise 



PINE Proportion of eastern white pine in stand 



ED Years of formal education 



AGE Age of landowner (years) 



PRO Variable, coded "1" if owner is employed in a 



white collar or professional occupation and 



"0" otherwise 



CITY Variable, coded "1" if landowner spent the first 

 12 years of his or her life in a large city 

 (population > 100,000) and "0" otherwise 



The signs of the coefficients indicate the direction of change 

 resulting from an increase in an explanatory variable, but 

 since the model is nonlinear, the magnitude of the change 

 is influenced by the values for all the variables and 

 coefficients. Elasticities measure the percentage change in 

 the probability that woodland is used for recreational 

 purposes resulting from a 1 percent increase in an 

 explanatory variable. The elasticities shown in Table 2 were 

 calculated at the mean values of the explanatory variables. 



The only forest characteristic statistically correlated 

 (10-percent level) with recreational use was the proportion 

 of eastern white pine occurring in the stand. There is no 

 clear intuitive explanation for this correlation other than 

 preference for the aesthetic appeal of white pine or that 

 white pine is more likely to occur on better drained sites. 

 Other parcel characteristics examined but not statistically 

 discernible included: size of ownership, proximity to a 

 maintained road, per-acre timber volume, and several 

 variables that measured species composition other than 

 white pine. Since forest measurements were taken on only 

 one sample area for each parcel (see Frieswyk and Malley 

 1985) and may not portray average characteristics 

 accurately, results with respect to forest characteristics 

 should be used with caution. 



Landowner characteristics had much greater significance in 

 explaining differences in recreational use among wooded 

 parcels. Woodlands owned by more highly educated 

 persons had a greater likelihood of being used for 

 recreation. A strong positive correlation (significant at the 

 1 -percent level) was estimated between recreational use 

 and the landowner's level of formal education. 



Parcels held by owners reared in large cities also were 

 more likely to be used for recreation. Recreation may be a 

 more important reason for owning forest land for these 

 owners than for those with a more rural background. 

 Preliminary regressions provided weak evidence, significant 

 only at the 20-percent level, that farmers were less likely to 

 use their woodlots for recreation. 



Table 2. Probit results and estimated elasticities 



Mean 



Explanatory 





Standard 



Recreational 



No recreational 





variable 



Coefficient 



error 



use 



use 



Elasticity 



Constant 



0.511 



0.613 



1.00 



1.00 





PINE 



0.828* 



0.432 



0.13 



0.06 



0.05 



ED 



0.083*** 



0.028 



14.90 



13.00 



0.60 



AGE 



-0.017** 



0.008 



55.58 



60.51 



-0.47 



PRO 



-0.308 



0.204 



0.42 



0.41 



-0.06 



CITY 



0.506* 



0.302 



0.19 



0.07 



0.04 



N = 258 



-2 LOG (Likelihood Ratio) = 26.63 



** 'Significant at 1-percent level 

 * 'Significant at 5-percent level 

 'Significant at 10-percent level 



2 



