Science  /  Origin Story

The Public Health Origins of Census Data Collection

How the "census tract" came into being.

Walter Laidlaw (1861–1936), was a Canadian Presbyterian minister. As a child, he was adopted by his uncle, Robert Laidlaw, founder of a prominent lumber company in Esquesing Township, southwest of Toronto. After graduating from the University of Toronto in 1881, Princeton Theological Seminary in 1884, and going on for further study at the University of Berlin and again at Princeton Seminary, he was called to be pastor of the Jermain Memorial Church in Watervliet (now West Troy), New York, a post he held from 1886 to 1892.[1]

After a year as president of the new University of Fairhaven (which later became Western Washington University, Bellingham) from 1892 to 1893, he settled in New York City, at St. Nicholas Collegiate Reformed Church, part of the Dutch Reformed tradition, from 1894–1895. But a new opportunity called: in 1895, he was appointed the first executive director of the newly formed New York Federation of Churches and Christian Workers. He held this position for almost 30 years, until 1922. Along the way, Laidlaw earned a Ph.D. from New York University in 1896; the field has not been determined, but statistics was his passion.

From his position as executive director of the Federation of Churches, and as editor of its journal, Federation, Laidlaw sought to put the work of religion on a secure scientific basis. Who were the people of New York? What social and economic needs did these people have? Where were the (Protestant, at least) houses of worship and settlement houses? Data were needed, and the census seemed a good place to start.

At this time, there were two: the Federal census, conducted in the years ending in “0,” the New York State census in the years ending in “5,” and they didn’t work the same way. As he details in a classic 1906 article in Federation, the counting principles differed between the Federal and state censuses, from one year’s census to the next, and even for different parts of the city within the same census. The Federal census of 1900 rolled up its data differently: for Brooklyn, Queens, and Staten Island, city council wards were used, and for Manhattan and the Bronx, New York State Assembly districts. And then the 1905 state census used New York State Assembly districts in all five boroughs. In Laidlaw’s words: “The ward is a fixed boundary, immobile as the orthography of a dead language,” he said; drawing out the metaphor, he continued, “the Assembly district is a changing boundary, a phonetic spelling arrangement which responds to the alien accents in the makeup of the city.” To get good data, the Federation found itself retabulating first the 1900 Federal census for 2 boroughs, and then, for the 1905 state census, for all 5 boroughs. At this point, Laidlaw called for a new system: “The scientific sociological study of Greater New York requires a ‘dead language’ boundary for tabulations. . . . Federation respectfully suggests a scheme which does away both with ward and Assembly district outlines, and which can be permanent.” [2]

The system he proposed was securely within the American tradition. It was, in fact, to use the system that was enshrined in law in the Land Ordinance of 1785: the “section” system set up to survey and sell the undeveloped lands west of the Appalachians. (An arial view of a Midwestern county, in Indiana, say, would reveal the regularity of the system!) Laidlaw’s first unit of analysis was the quarter-section: a quarter of a square mile, or 160 acres.