The purchase and preparation of data from the March 1973 Current Population Survey for New York City has been in the nature of an experiment. The U.S. Bureau of the Census in conducting its Current Population Surveys (CPS) includes a sample of approximately 2,100 households in New York City, chosen in such a way as to be representative of the city as a whole though not necessarily of the individual boroughs. It covers a wide range of data including age, sex, ethnicity, family size, income, earnings, occupation, education, etc., and these can be cross-tabulated in a variety of ways. A sample of 2,100 households is a respectably sized sample. Nevertheless, it has its limitations, especially when it is more and more finely broken down, for example, first by income groups, then by ethnic groups, and then by family size. The degrees of variance of the standard errors are shown in the appendix. We have done a great deal of "collapsing" of the data as they were obtained from the Census Bureau printouts. In preparing the tables, we tried to strike a balance between reasonable reliability, on the one hand, and flexibility for the users of the data on the other hand. Whenever the Census Bureau presented medians, eg., income, years of schooling, for categories that did not require additional calculation by us, our tables show the CPS figures. (Author/JM)