prilagođeno pretraživanje po punom tekstu

ŠUMARSKI LIST 5-6/2019 str. 12     <-- 12 -->        PDF

the volume of the antlers (K-S: d=0.17541; p>0.2; S-W: W=0.95450; p=0.10834); the difference in the mass of the antlers after the shallow cut (K-S: d=0.08928; p>0.2; S-W: W=0.97409; p=0.480); and the difference in the mass of the antlers after the regular cut (K-S: d=0.06925; p>0.2; S-W: W=0.98097; p=0.725) have normal distribution and the data may be equalized by a linear function. Therefore, simple linear regression was performed. If the data did not show normal distribution they were transformed using the Box-Cox transformation method (Sakia 1992).
The differences between trends in the trophy parameters were tested using analysis of covariance (ANCOVA). In the analysis of covariance it is desirable for the lines of the researched groups not to show interaction, that is, they should not intersect (Enqvist 2005). If they do (in the case of interaction) it is more difficult to define any differences between the groups. If a significant difference in values is found between the groups then for one group within the determined range of the continuous variable (e.g. volume) the parameter in question shows higher values than in another group, and in another range of the continuous variable the situation is reversed. According to that, this results in the impossibility of drawing a general conclusion (for the entire range of the continuous variable) but the rule only applies within the specific range (Fraas and Newman 1997). Since no statistically significant differences were found in any of the tests between the slope of the lines, the tests were undertaken using the classical analysis of covariance. The data were equalized with the lines (the method of simple linear regression), square function or potency function (the correlation of the gross mass of the antlers and the sawn-off part of the skull). In equalizing the potency function, the correlation between the gross mass and the waste from sawing the skull was found by the Gauss Newton minimization procedure. The data were analysed using the Statsoft 13 Program (TIBCO Software Inc. 2017).
Depending on the parameters measured, variability differs considerably (Table 2). The smallest variability was shown by the waste from the part of the skull after a shallow cut