Free Trial

Safari Books Online is a digital library providing on-demand subscription access to thousands of learning resources.

Share this Page URL

8.6 Between-by-Within and Within-by-With... > 8.6.12 A Rank-Based Approach - Pg. 423

Chapter 8 Comparing Multiple Dependent Groups 423 contrasts. Note that when dealing with linear contrasts, again there are two basic goals that might be of interest. The first is to test hypotheses about linear contrasts stated in terms of the measures of location associated with the marginal distributions. Section 8.1.3 provides explicit details when dealing with trimmed means that can be used to analyze a within-by-within design. The second strategy is to use an extension of methods based on difference scores. That is, now the hypotheses of interest take the form described in Section 8.1.4. R functions specifically designed for within-by-within design are described in the next section. 8.6.11 R Functions wwtrim, wwtrimbt, wwmcppb, and wwmcpbt The R function wwtrim(J, K, x, grp = c(1:p), p = J * K, tr = 0.2) tests for main effects and interactions in a within-by-within design using a modification of the method for trimmed means described in Section 8.6.1. (The modification simply takes into account the possibility that all J K variables might be dependent.) The R function wwtrimbt(J, K, x, tr = 0.2, JKL = J * K, grp = c(1:JK), nboot = 599, SEED = T, . . .) is the same as the R function wwtrim, only a bootstrap-t method is used. The R function