120 METHODS i\D TFCHNIQUES 



port it was especially useful in untangling the effects of years, seasons, sex and age. For ex- 

 ample, a series of weights of grouse collected over several years from all seasons would mean 

 little until analysed by this method. When the variability due to each of these influences 

 was separated from that of the whole it became possible to judge the significance of each. 



Regression and Correlation 



In many wildlife research problems knowledge of the relationshij) or association between 

 two or more factors is important. Regression and correlation are methods of analysis which 

 provide means of expressing the degree of relationship between such factors. They summar- 

 ize, in mathematical terms, information usually presented by graphs and visually appraised. 

 Mathematical evaluation has the advantage of leading to tests of significance. 



One of the most important uses made of regression analysis by the Investigation was in the 

 statistical adjustment of data. For example, variations in the degree of overwinter mortality 

 recorded among grouse on the study areas over a series of years required such adjustment 

 with respect to fluctuatiotis in the preceding fall population level to obtain comparable yearly 

 trends. Similarly, records of the number of fox tracks observed were corrected for differ- 

 ences in the amount of effort spent in the field from year to year. 



Simple regression, when extended bv measuring more than two factors coincidentally. gives 

 rise to a form of analysis termed "nuiUiple regression". When several groups of such data 

 are handled together the analysis of variance is combined with either simple or multiple re- 

 gression in what is known as ''analysis of covariance". This method was. for instance, used 

 to correlate records of wind. rain, man-hours and grouse flushes on the xarious compartments 

 of the Connecticut Hill area for a series of years*. 



* See diwcussiiiii of Fstiinuting Cnnisf Population!), p. 714. 



