ABSTRACT
The substantively weighted analytical technique (SWAT) is a new system for analyzing data, but not a completely original system: it rests on two familiar pillars. These two pillars are: quantitative pillar and qualitative pillar. An effective combination of the qualitative and quantitative approaches can provide a powerful data-analytic tool. SWAT runs against such rigid doctrine requiring the analyst to "get messy" with the data —to look at graphical output, to consider alternative weighting schemes, and to analyze coefficients from hypothetical populations. Weighting in SWAT is completely flexible and completely at the discretion of the analyst. At one extreme is the standard linear model with uniform weighting, and at the other extreme is a completely qualitative approach of looking at one case at a time, which represents weighting one case at unity and all the others at zero. Classical statistical procedures are particularly useful in sorting through large datasets and in classifying specific cases.
