Dear colleagues,
I am thankful for all your suggestions and help about the effects of collapsing
tables and statistical power in log linear analysis. However, to fully
understand the problem, I would like your advice on where to find information
about the relationship between degrees of freedom and power in chi-square
analyses and derived multivariate techniques. As I understand so far, reducing
degrees of freedom in contingency tables (i.e. collapsing a table) should
increase power, although this usually depends also on the hypotheses being
tested and the implicit effects of each categorical variable.
Pablo
Pablo Jarrin
Department of Biology
Boston University
|