Note: A version of this post first appeared on the Capital Broadcasting Opinion pages. Since initial publication, North Carolina has joined 16 other states in a lawsuit attempting to block the addition of the citizenship question to the census in 2020.
Once every 10 years, we count all individuals living in the United States in the census. In 2010, respondents were asked ten questions about basic characteristics, such as age, sex, race, and homeownership status. Last week, Secretary of Commerce Wilbur Ross announced that census respondents will answer one additional question in 2020: citizenship status. Introducing an additional, untested question so late in the census life cycle is concerning to demographers and social scientists, like me, who rely on the census as a key source of information about how and why our population is changing.
The concern about the newly added question is not related to its content. We have a long history of asking Americans about their citizenship status surveys conducted by the U.S. Census Bureau, including the full census (of all Americans) from 1890 to 1950. The concern is the addition of this question without the standard rigor and vetting process the Bureau typically takes for adding new census questions. Without this research, we do not understand how it will impact response rates, accuracy, and overall costs.
The census is one of the most important activities of our government. It is the foundation of our representative democracy. Every decade, the 435 seats in the U.S. House of Representatives are divided among the states based on their total population in the most recent census. The census also determines the allocation of more than half a trillion dollars of federal funds, including more than $16 billion annually to North Carolina.
Counting everyone who lives in the United States—and counting them accurately—is hard to do. One of the Census Bureau’s biggest challenges in obtaining a count of all residents is non-response. When households do not complete their census form, the Bureau sends individual enumerators door-to-door to try to obtain the required information. In 2010, it cost nearly $33 million to follow-up with the 975,000 North Carolina households that did not initially complete the federally mandated census form. Nationwide, non-response follow-up efforts cost more than $1.5 billion.
High rates of non-response also raise the risk of a net undercount. In 2010, the overall census count was highly accurate, but certain populations were undercounted, meaning they were missed in the census totals. Renters, black men, American Indians living on reservations, and Hispanics were among the groups with higher rates of undercount in 2010. But the highest rate of undercount was for young children ages 0 to 4. In North Carolina, the Bureau estimates that more than 25,000 young children were not counted in 2010, the 8th largest number of any state.
Non-response in 2020 may be even higher, for many reasons. The U.S. population is larger and harder to count than ever before. Overall response rates to statistical surveys have been declining steadily for the past few decades and Americans are increasingly distrustful of government data collection efforts. If the newly added question increases non-response among immigrant communities as some worry, this could have far-reaching impacts on North Carolina.
Nearly 800,000 immigrants were living in North Carolina in 2016, representing 8 percent of the total population, with high concentrations in both urban and rural communities. Higher non-response and a greater undercount of the immigrant population could exacerbate the undercount of children in North Carolina in 2020. Between 2000 and 2016, our state had the fourth largest numeric increase in children born to immigrant parents. Today, one in six North Carolina children—365,000—are U.S.-born citizens who live in a household with at least one immigrant parent. If these children become harder to count, the state could receive fewer funds for critical programs that help to ensure healthy growth and development for all children, such as Head Start and the School Lunch Program. At the local level, school districts may face challenges planning for the future if the 2020 data does not fully capture the population they will soon serve.
Of course, the impacts of this new question on overall response rates, accuracy, and costs could be negligible. Unfortunately, the citizenship question was officially added after the Census Bureau had already begun its 2018 End-to-End test in Providence, Rhode Island. The Bureau uses this “dress rehearsal” to identify any challenges that need to be addressed prior to the full census in 2020. Without testing properly ahead of time, we won’t know the potential consequences of this change until it may be too late to address them. When the data you are collecting influences the distribution of political power and money for the next ten years, that’s a dangerous risk to take.