ABSTRACT
The last two chapters have described some techniques for investigating tables of data when the context of the investigation allows structure to be imposed and specific hypotheses to be tested. Unfortunately, there are some sets of data where this cannot be done. One may lack the information to formulate specific hypotheses and be unwilling to posit a predictor-outcome structure, yet still want to analyze the results. Obviously, starting with weaker assumptions about the data, one must draw weaker conclusions in the end. Nevertheless, there is much one can do; this chapter considers some options.
