Hi everyone
My apologies in advance for what will be a basic question for this community. When SPSS tests for significance differences between percentages in columns in custom tables, does it use the weighted base size, effective base size or unweighted base size? And, following on from this, do I need to conduct sig testing on unweighted tables, given that we have boosted/over-sampled some key groups and then down-weighted them back to the population profile?
Thanks!
Claire
------------------------------
Claire Lister
------------------------------