![]() It is also important in the interpretation of results. Stabilizing the variances is important for the validity of the analysis of variance. Sometimes, adding a constant to all values will make it possible to take logarithms or square roots. Of course, logarithms and square roots are nonsensical if some of the values are negative. If the data are proportions or percentages, use arcsin ( y i j ), where the y i j are the proportions. If σ 2 is proportional to the mean, take the positive square root of the y i j. If σ is proportional to the mean, use the logarithm of the y i j. These situations are easily handled with transformations. In many situations, the group dispersion tends to follow a pattern related to the group mean. ![]() Fortunately, transformations that stabilize the variance will frequently reduce skewness as well. ![]() The preferred method for dealing with skewness and unequal variances is to look for a transformation of the data that will mitigate the problem. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |