A combination of a more punitive policy situation and the greater range in size and financial performance in farms, especially across regions, has led to policy-makers focusing more on how new policies or policy proposals affect regions or types of farms, not just the national aggregates or average. While some new research is focusing on the implications of policy changes on farms in different areas and of different sizes, less has been done to improve how one goes about identifying and defining a "representative" farm. The basic objective of this study is to explore the relationship between structural, technical efficiency, and financial performance data for a large sample of New York dairy farms. To what extent are financial or technical performance related to farm size or some other structural characteristic? If we conclude that most highly profitable farms are large, for example, can we also infer that most large farms are highly profitable? Ultimately, we want to be able to improve how we go about defining and talking about "representative farms." This research provides an improved understanding of the differences between top performing and low performing dairy farms. Performance is measured by five financial criteria; four profitability and one cash flow measure. The distinction between high and low performance is specified as being in the top 25 percent versus the low 25 percent by financial performance criteria. As judged by the four profitability measures, top performers from the 387 specialized, large breed farms participating in the 1989 New York Dairy Farm Business Summary are significantly higher, in herd size, crop yields, net milk prices and milk sold per cow. Top performing farms are significantly lower in capital investment per cow, feed and crop expense per cwt. of milk, purchased feed per cwt. of milk, and labor and machinery costs per cow. The cash flow criterion provides fewer significant differences between low and high performing farms. The high performing farms average 170 cows and the low performing farms 70 cows. Of course, this does not mean that farms of 170 cows or so are necessarily more profitable than smaller farms. An analysis of 10 percent of the farms clustered around each herd size is conducted to determine if very large farms or very small farms bias the analysis. The comparison reveals significant differences exist between clusters for all four profitability measures, but not cash flow. However, milk sold per cow, feed costs, and non-feed costs are not significantly different. Within clusters, farms are ranked by return on investment, and differences between upper and lower ranking halves are analyzed. Significant differences are found in all four measures of profitability, but again, not in cash flow. Significant differences in performance exist in capital efficiency and machinery expense per cow for the 70 cow group and only in debt to asset ratio and com silage yield for the 170 cow group.