Regresar

Race and Ethnicity in Research Methods

Abstract:

CXi 1 b n EMSA= bno? A+ na2AB+ a 2e MSAIMSA x B I3j a 1 n EMSB= ana2B+ na2AB+ a2e MSBIMSA x B cxl3ij 1 1 n EMSAx B= na2AB+ a2e MSA x BIMSS (AB) ek (ij) 1 1 1 EMSs (AB)= a2e aFrepresents a" fixed factor"; R represents a" random factor." fects. Rather, lower-order terms in general are affected (eg, the error term required for testing a three-way interaction may differ in a factorial with four or more factors). Note also in Tables 2 and 3 that not only do sources of" error" change, but the corresponding degrees of freedom for the denominator will also be affected. A test of MSA against MSS (AB) will yield (a-1) and ab (n-1) df, but a test of MSA against MSA x B is tested against critical F's with (a-1) and (a-l)(b-1) df, the latter of which is often far fewer than ab (n-1), which makes the test less powerful in the presence of random factors.(Note that the power of the test of factor B and the interaction in Table 2 are …