FY18 Omnibus Revives 2020 Census Funding: Question Bias and Lack of Testing Still Could Hold Back Transportation Research
Census data is critical for a myriad of research areas, including public health, education, economics, and of course infrastructure and transportation. When the U.S. government first appropriated funds for highway projects in urban areas, the U.S. Census Bureau began surveying random households for information about travel patterns to estimate travel in urban areas. Data on population/household demographics, land use and housing types, vehicle ownership, and trip characteristics such as origins, destinations, trip purposes, travel modes, and travel times are still collected today and provide a valuable time series to understand the past and plan for the future in land use and transportation. Metropolitan Planning Organizations are designated to urbanized areas with populations of 50,000 people or more, which is determined by the census, and formula funds for infrastructure projects are based on population levels.
Cross-sectional data are useful as well, and are often combined with other data sets to examine travel patterns among certain population demographics or geographic areas in individual or aggregated census tracts. A recent study out of San Francisco illustrates how researchers can look at commute mode split from the census and public transit ridership data to assess the service provision of the transit system.
Despite the importance and usefulness of these data, the 2020 census was woefully underfunded due to a failure to increase the funding allocation from the 2010 census levels. Increased population, advancements in technology, and inflation all suggest that an accurate count would require an influx of funds for the survey ten years later. A general lack of awareness of the census and it’s importance may have partially led to the low amount of outcry at the perceived defunding in 2014, but advocates harped on the issue, including focusing in on the potential inability for the Bureau to fulfill their constitutional mandate of determining populations to apportion the number of representatives for each state in the House of Representatives.
Last week, Congress got spending happy with the omnibus FY18 bill, raising funding levels for programs across the board. Luckily, the U.S. Census bureau was not left out. The bill appropriated $270 million to current surveys and programs and an additional $2.544 billion for periodic censuses and programs, allocating in total a substantial increase from the $1.47 billion total allocation for 2017.
Hopefully it’s not too late. Part of the budget needs for the 2020 census include testing of technology and trying out surveys in advance, and the survey deployment date is quickly approaching after years of an inadequate budget.
There are also more problems that may plague the 2020 data set. Non-response is a reality for every survey. Researchers tend to craft surveys to minimize the number of people who will choose to not respond in order to collect the largest sample possible, be cost-effective in their distribution, and avoid biases when certain types of people are less likely to respond. On Monday, the U.S. Department of Commerce announced the reinstatement of a question that has not been asked in a decennial census since 1950. The question asks whether or not the respondent is a citizen. The addition of this question was controversial due to concerns about non-citizens not completing the survey due to fear of people finding out about their non-citizen status. Not only could the addition of the citizenship question affect response rate and bias results, but based on analysis of American Community Survey (ACS)(a smaller scale annual survey) data, it is likely to be inaccurate to the level “that between 28 and 34 percent of the citizenship self-responses for persons that administrative records show are non-citizens were inaccurate.” Non-citizens utilize our roads and transit systems, and understanding travel patterns of all people in a region is vital to planners’ and researchers’ ability to model and study transportation.
Without good data our public funding and programming will be incorrectly allocated, districts and congressional representation will be falsely assigned, and research and practice will suffer. The additional funds for the Bureau in FY18 are a start, and as long as levels are maintained in 2019, the 2020 census has a shot at exceeding the low expectations set over the last half decade.