Part
V: Analysis of Variance (ANOVA) |
| | Keppel's Basic Ratios | SRD Structural Model | SRD Source Table | |
| Keppel outlines 3 values
(using what he calls basic ratios) which can be
calculated to derive SSbetween, SSwithin, and SStotal.
They are:
From here,
|
| A structural model is
simply a way to represent possible sources of variance
which may contribute to an individual's score on the DV.
For example, take the following equation:
This states that the score of a person in group j is a function of the overall grand mean in the population plus the variation attributed to treatment effect plus the individual's random error. Under Ho, (where we assume no significant differences among treatment groups) the treatment value is equal to 0. In fact, the Ho that the groups are equal is equivalent to the null that there is no treatment effect. |
Once we compute all of
the necessary values for a SRD, we need to represent the
values in a fairly standardized manner. This
representation is called a source table; the values are
organized and determined as follows (a = # of levels for
the between subjects factor):
|
||||||||||||||||||||
