Studying urban patterns, I have necessarily worked with census data my entire career. I have come to know the critical importance of understanding how data have been collected and how things are defined when using any data collected by others.
At the urban research center where I worked, I often sent around updates on census changes that were important to our research and to understanding data for our region. Some of the staff got to calling me “the census nerd.” This attention to the details even led me to the writing of a journal article.
A few years after the 1990 census, the Census Bureau sent out a 4-page brief on trends in metropolitan area population. (This was back when such things were still being distributed on paper.) The first page had a chart showing the percentage of the U.S. population living in metropolitan areas from 1910 to 1990, broken down by the portions inside and outside central cities. Of course it showed the continuing large increases in the percentage in metropolitan areas. It also showed the increase in the percentage in central cities up through the middle of the century, followed by a decline as the greatest portion of metropolitan growth then occurred in the suburbs. Not news to anyone familiar with urban trends. Except that from 1980 to 1990, the chart showed a small uptick in the percentage in central cities. Nothing dramatic, but any time a trend that had been steady for a number of decades shows signs of a reversal, that may indicate something interesting is happening. The publication made no mention of this change. The chart did include a note indicating that the data reflected the definitions used at the time of each census.
I then recalled reading that the Office of Management and Budget had changed the way central cities in Metropolitan Statistical Areas (MSAs) would be identified. Before, the largest city in an MSA would be designated the central city, and up to two more cities could also be central cities if they were sufficiently large in relation to the first. So many MSAs had only a single central city, but this allowed cities like Dallas and Fort Worth to both be designated central cities in their MSA. But under the new definition, the number of central cities would not be limited and they could be much smaller than the largest city. The primary criterion required that a central city have a significant concentration of employment. (Note that what were central cities are now called “principal cities” in the current MSA definition, reducing potential confusion.)
This led me to ask a host of questions: Was the reversal I saw in that chart caused by the additional central cities? How many new central cities were there and what effect did they have on total central city populations for different MSAs? How did the new central cities affect traditional central city-suburb comparisons commonly made in urban analysis?
I collected the data and did the analysis to answer these questions. This resulted in the publication of the article “The New Central Cities” (Urban Affairs Review, 1996). This had the best beginning sentence I have written: “What do Cambridge, Massachusetts, East St. Louis, Missouri, Irving, Texas, and Palm Springs, California, have in common?”