131 

 Statistical Methods 



Nonparametric tests were used to analyze census data due to small sample sizes and the high 

 likelihood that the data were not normally distributed. A potential bias due to animal observability 

 differences between forest successional stages was tested by comparing the mean sighting distances 

 between the spotter and the animal when first noted. No differences were noted (Appendix J) so it was 

 considered appropriate to compare relative abundances and population density estimates between forest 

 successional stages. The Kruskal-Wallis test (x^ approximation) was used for the analyses. The 

 following statistical terms were used; £ = mean, SE = standard error, n = sample size, d.f. = 

 degrees of freedom, x^ approximation = chi-square approximation, and P = probability value. All 

 statistical analyses were conducted using SAS (SAS Institute Inc., 1988). 



Limitations of die Data Sets 



There were at least five main factors that must be considered when reviewing these results. 

 One, these results were based on a small number of sightings over a short period of time at a single 

 site. As a consequence, a small number of sightings could have a substantial impact on the relative 

 abundance values and population density estimates for the wildlife taxa. Two, there were differences in 

 the composition and structure of the three forest successional stages. These differences likely made it 

 easier to observe animals in Late Secondary Forests than in Combined/Early Secondary Forests because 

 Late Secondary Forests were more open at ground level (but see the section on mean sighting distances, 

 below; also see Chapter 2 for additional information about the structure and composition of the forest). 

 Three, there were differences in the nature and extent of human activity in the three forest successional 

 stages. One would expect that game species intolerant of disturbances associated with human activity 

 would be less abundant in areas with extensive human activities than in areas with limited activities, but 

 this may be offset-in part--by wildlife habituation to humans. The degree of habituation by wildlife 

 species on the study area is unknown. Four, differences in animal size, behavior, and ecology may 

 have resulted in the populations of some species being overestimated, while others may have been 



