DISCUSSION AND CONCLUSIONS 



The logistic function is preferable to a linear model for describing the rela- 

 tionship between a dichotomous dependent variable and a set of independent variables. 

 The algorithm discussed in this paper provides a valid method of estimating the param- 

 eters of the logistic function. 



Both the algorithm and the computer program have been modified to handle data 

 collected with arbitrary probability. This modification is of particular value in 

 cases such as estimating tree mortality when it would not be feasible to measure all 

 observations with equal probability. 



LITERATURE CITED 



Bartlett, M. S. 



1951. An inverse matrix adjustment arising in discriminant analysis. Ann. Math. 

 Stat. 22:107-111. 



Cochran, William G. 



1963. Sampling techniques. 2nd ed. 413 p. New York: John Wiley and Sons, Inc. 

 Lee, Y. 



1971. Predicting mortality for even-aged stands of lodgepole pine. For. Chron. 

 47(2) :29-32. 



Neter, John, and E. Scott Maynes 



1970. On the appropriateness of the correlation coefficient with a 0,1 dependent 

 variable. J. Am. Stat. Assoc. 65 (330) : 501-509 . 



Sterling, Theodor D. , Randall G. Binks, Shelby Haberman, and Seymour V. Pollack 



1969. Robot data screening--a ubiquitous automatic search technique. In: Milton, 

 Roy C., and John A. Nelder (ed.), Statistical Computation, p. 319-333. 

 New York: Academic Press. 462 p. 



Walker, Strother H. , and David B. Duncan 



1967. Estimation of the probability of an event as a function of several indepen- 

 dent variables. Biometrika 54:167-179. 



18 



U.S. GOVERNMENT PRINTING OFFICE: 1974 „ 784-4 9^/60 REGION NO. 8 



