Laserfiche WebLink
<br />. <br />. <br />. <br />. <br />. <br />. <br />. <br />. <br />~ <br />~ <br />~ <br />~ <br />I <br />I <br />. <br />. <br />. <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br />I <br /> <br />Discussion of Results <br /> <br />Because the correlation coefficients for the targeUcontrol relationships are not <br />perfect. i.e., 1.00, the results obtained through use of these equations should be <br />considered as approximate indicators. The degree of confidence that can be placed in <br />the indications is affected by a) the statistical correlation between the target area sites <br />and the control area sites, b) the strength of the seeding effect and c) the number of <br />seasons included in the analysis, It is conceivable that a single season analysis could <br />yield a negative mathematical indication of seeding effect, although it is not likely that <br />properly conducted seeding operations would cause true decreases. Any such <br />indication of a decrease would most likely be due to the inability of the chosen <br />regression equation to eliminate all of the natural variance between the control and <br />target sites. Statistical tests can be conducted to estimate the probabilities that the <br />single season and combined-season indications are not due to chance, but most are <br />designed for use in randomized data sets. <br /> <br />~I <br />