Problems in the Analysis of Survey Data, and a Proposal

摘要: Abstract Most of the problems analyzing survey data have been reasonably well handled, except those revolving around existence interaction effects. Indeed, increased efficiency in handling multivariate analyses even with non-numerical variables, has achieved largely by assuming additivity. An approach to is proposed which imposes no restrictions on effects, focuses Importance reducing predictive error, operates sequentially, and independent extent linearity classifications or order explanatory factors are introduced.