Multiple regressions were also run on the data, with round trip time as the depend- 

 ent variable and various combinations of distance, number of pieces, volume per piece 

 and volume per turn as the independent variables. The equations are of the form: 



Y = b^ + biXi + b2X2 + bsXa + bi+Xi4 



The results are summarized in table 7. All regressions are significant at the 0.05 

 probability level. 



Table 7. --Multiple regression intercepts ^ coefficients, R and F-statistio 



1/ 



for skidding production— 



b : 

 o 



bi : 



b2 : b3 



: b^ : 



r2 



: F 



Intercept 



ui stance 



ilrO. rCS cO vOU 



1 U~rrl vO u 











SKTDDFR 









4.08 



0.02 



0.61 





0. 32 



54.0* 



-13.85 



.02 



.45 4.33 





.40 



49.3* 



3.82 



.02 





0. 14 



.37 



66.2* 







HORSE 









4.19 



.02 



.16 





.29 



20. 1* 



2.57 



.02 



.14 .33 





.31 



14.4* 



4.03 



.02 





.04 



.30 



21.6* 







TRACTOR 









.41 



.03 



1.2 





.70 



78.8* 



31.97 



.03 



1.2 -6.84 





.71 



54.9* 



.53 



.03 





.26 



.68 



72.8* 



— Subscripts refer to following variables: 



- intercept 



1 - distance 



2 = number of pieces per turn 



3 = volume per piece 



4 = volume per turn 



* Indicates significance at 0.05 probability level 



Using R as a measure of reliability, table 7 suggests: 



1. multiple regression does not improve the reliability of prediction equations; 



2. although the reliability of the prediction equation for the rubber-tired 

 skidder and horse is quite low, it is high for the tractor skidder. 



13 



