Skip to Content

AFBAmerican Foundation®
for the Blind

Expanding possibilities for people with vision loss

Counting Down to 2020: The Census Is Coming; Are We Ready?

Published January 17, 2018
a laptop computer with news headlines on the monitor

Welcome to part two of our series looking at the census and ways people who are blind or visually impaired are included in data and research. Last month, in "O Come and Be Counted!" we revealed that some of the earliest organized research in history occurred through censuses, and that surveys and other forms of research are now commonly used to collect information from samples of individuals in order to better understand a larger population (see Censuses, Surveys, Research - What's the difference?). We considered how census takers and other researchers protect people's rights and personal information, and we discussed the reasons why it is mandatory for individuals to participate in the census and the American Community Survey (see Your Rights and Responsibilities as a Census-responder and Research Participant).

In this edition of the Research Navigator, we'll focus on issues, challenges, and concerns related to the upcoming decennial U.S. Census, including:

About This Series

The Research Navigator is a quarterly series - accompanying AFB's DirectConnect newsletter - from the AFB Public Policy Center . The purpose of this series is to keep you informed of user-friendly facts and figures and the latest research pertaining to people with vision loss. The series will also include the necessary background information so you may use the information most accurately. Have an idea for a Research Navigator topic? Want to know more about a statistic or line of research? Send your thoughts to AFB's Senior Policy Researcher, Rebecca Sheffield. Readers are also encouraged to check out AFB's Statistical Snapshots . This webpage is regularly updated with a wide variety of information and tools that address commonly asked questions about people with vision loss.

The Problem with Bad Data - When Censuses Go Wrong

Overseeing a nation's data is an enormous responsibility requiring well-supported staff, ongoing and detailed oversight, and careful adjustments to change. The following historical examples demonstrate unintended consequences that can follow from deviations in the way a census or survey is conducted.

Canadian Census, 2011

In Canada, a national census has been undertaken every five years since 1971. Typically, in addition to every household being required to respond to the census, one in five Canadian households also receives and is required to respond to a longer set of more detailed questions (similar to the American Community Survey) (Cecco, 2015). In 2006, the response rate for the mandatory long-form census was over 93% (Marshall, 2015); that's a bit lower than the 98% of Americans who responded to the 2010 U.S. Census (Long, 2017) and the 95% response rate for the American Community Survey (ACS) (U.S. Census Bureau, 2017a), though - as we will discuss later in this edition of the Navigator- the U.S. Census Bureau puts a lot of work into tracking down non-responders.

In 2010, the Canadian government decreed that the long-form census (which they renamed the National Household Survey [NHS]) would become a voluntary survey, sent to a randomly selected 30% of households. In 2011, these changes were implemented, and only 68% of the Canadians who received the NHS responded to it (Marshall, 2015). Since the NHS was designed to function as a research survey (based on a sample) rather than as a census (based on a count of every member of the population), the NHS needed a random sample to produce accurate estimates about the entire population. With such a low response rate, and with no authority to require Canadians to respond to the NHS, Statistics Canada (the Canadian equivalent of the U.S. Census Bureau) was faced with major gaps in their data and their access to important demographic information. People who responded to the NHS were more likely to be wealthy, white, and well educated than those who did not respond; thus, any estimates based on the NHS provided a skewed perspective of the Canadian population. Canada's auditor General estimated that 25% of geographic areas in Canada could not make reliable use of the NHS data for demographic, health, and other purposes. The head of Statistics Canada resigned, saying "the question of whether a voluntary survey can become a substitute for a mandatory census… It cannot" (Marshall, 2015).

In 2015, with a transition in political power in Canada, the mandatory long-form census was reinstated, and in 2016, data was once again collected with much higher response rates. Because the 2011 NHS data was so problematic, it is not used, and the 2016 long-form census data has to be compared to the 2006 data to determine demographic changes over the past ten years. The missing 2011 data will remain a difficult gap in Canada's national statistics, which otherwise have included updated data points every five years. For the foreseeable future, the 2011 gap will impact researchers' and statisticians' abilities to understand past trends and make future predictions about the Canada's population.

American Community Survey, January 2004 and October 2013

In January of 2004, due to a reduction in funding, the U.S. Census Bureau was not able to conduct telephone interviews or personal-visit follow-ups to reach non-responders to the ACS. This reduced the response rate for the 2004 ACS, leading to greater difficulty producing estimates for smaller geographies or subsets of the population. In October of 2013, a government shutdown prevented a second mailing, telephone follow-ups, and/or in-person follow-ups with ACS non-responders, also leading to a reduced response rate and challenges with the ensuring the accuracy and reliability of estimates based upon the data.

Although these may seem like small "blips" in the overall scheme of a massive survey project like the ACS, these two months mark irretrievable gaps in an otherwise incredibly comprehensive, robust, and long-running statistical history of the United States. Projects like the ACS must operate with incredible efficiency; therefore, the Census Bureau only randomly selects and samples the smallest number of people required to produce high-quality population estimates. Their pre-defined methodologies for following up with non-responders help to ensure that those people who are most likely to not respond (people of lower incomes, people with disabilities, etc.) are given every opportunity to participate, avoiding under-representing any demographic group. When the Census Bureau is not able to follow its established methods, then the value of the entire project is jeopardized, not only for that year's estimates, but also for future research which attempts to understand changes over time.

United States Capitol in Washington, DC.

Preparations and Concerns for the 2020 Census

As you have just read in the previous section, and as we discussed in part one of this series:

  • censuses must have maximum participation to truly count everybody, and
  • national survey efforts like the ACS must have very high participation rates to ensure their sampling methods are as random and representative of the actual population as possible.

Without these assurances, all the time and money invested in planning and executing census and survey projects could be worthless because the data will not provide dependable results. Potentially even more serious, outside forces could unduly influence the outcomes of the 2020 Census or the ACS, leading to a host of potential problems, from inaccurate reapportioning of congressional districts to underfunding of education and health services in various parts of the country. In most every case, the people who are the most difficult to contact (people who are older, people living in poverty, people with disabilities, etc.) are the most likely to go uncounted and to be underrepresented.

Preparations for the 2020 Census have been underway practically since 2010, including test-runs annually since 2013, release of an operational plan in 2015, and an address canvassing test in 2017. Field infrastructure has been ramping up since 2017, and an end-to-end test of census systems is set for 2018. If all goes as planned, the Census Bureau estimates that efficiencies in the 2020 Census could save up to $5.2 billion by avoiding unnecessary costs and producing higher quality data (U.S. Census Bureau, 2017b).

However, all does not seem to be going to plan. Politicians and advocates have expressed concerns that the 2020 Census is in trouble. From 2012 to 2017, the Congressional Budget for the census has only provided 90% of the funding which the Bureau requested, and costs estimates for the census are increasing due to concerns about non-responders (Scherer & Bahrampour, 2017). Former congressman Tony Coelho wrote, "By underfunding preparations for the 2020 census, including not running all the national field tests that should take place in 2018, it appears …leadership is pursuing a failed census" (Coelho, 2017). In 2017, John Thompson resigned from his post as Director of the Census Bureau; interim staff are currently filling this important vacancy, leading to further apprehension about the upcoming decennial census. A non-partisan, active Director is crucial to seeing the Bureau through the preparations for 2020, ensuring that the census will be conducted with accuracy and its data will be trustworthy (Sweetland Edwards, 2017).

The federal budget has not been established for 2018 (the government is currently operating on a continuing resolution through January 19, which mostly maintains the funding allocations from the previous year). Proposals of funding for the Census Bureau from the House, the Senate, and the Administration all fall short of what was requested by the Bureau; if enacted, these funding cuts will force the Bureau to further reduce its preparation and testing, not only for the 2020 Census but also for the ongoing ACS and other important projects of the Bureau. A letter from the Census Project (an advocacy organization) and other stakeholders expressed concern that "no funds are included in the revised FY 2018 Census Bureau budget for timely development of the full advertising campaign, launch of the Partnership Program, restoring cancelled field tests in rural areas, or to adapt operations to remedy the impact of disasters in Texas, Florida, Puerto Rico and California that increase the risk of an incomplete census count in those communities" ("Critical Census budget action needed," 2017). Meanwhile, if disagreements about the budget lead to a federal government shutdown, all Census Bureau operations could grind to a halt, leading to a repeat of problems for the ACS and further reduction in preparations for 2020.

Finally, in the past two weeks, another issue has developed with potential to impact the accuracy of the census. The Justice Department has asked the Census Bureau to consider adding a question about citizenship to the 2020 Census, explaining that such a question would lead to a more accurate counting of the citizen, voting-age population in each state and congressional district. The census has not required all respondents to report about their citizenship since 1960, and many advocates are alarmed at potential impacts of acting such a question. A group of Democratic senators wrote to the Commerce Department (home to the Census Bureau), stating that a citizen question "would likely depress participation in the 2020 Census from immigrants who fear the government could use the information to target them. It could also decrease response rates from U.S. citizens who live in mixed-status households, and who might fear putting immigrant family members at risk through providing information to the government" (Feinstein, Carper, Schatz, Cortez Maston, & Harris, 2018). The letter describes existing and increasing mistrust of the government (including the census) as well as other problems like funding which are putting the 2020 Census at risk. As of the writing of this issue of the Navigator, the Census Bureau was still reviewing the Justice Department's request and has not made a decision about whether to test and/or include a citizenship question in 2020.

Censuses, Surveys and People with Vision Loss

Redistricting and representation

Successful advocacy by and for people with vision loss requires that people with vision loss and their families are counted and heard not just in the census, but also at the ballot box and in Congress. One of the most well-known uses of decennial census data is in the proportional drawing of congressional districts (the districts which are each represented by a member of the House of Representatives). The number of seats in the House of Representatives has been set at 435 since 1941; therefore, census figures are needed not just to redraw districts within states but also to divide these 435 seats as evenly as possible among the 50 states based upon state populations. Furthermore, each state's number of delegates in the Electoral College (through which the President of the United States is elected) is based upon that state's number of members in the House and Senate; so, the census has a direct impact on the voice of each state's voters.

Button that says Your Vote Counts

State populations change over time due to a range of factors, including relocation of retirees from the massive Baby Boom generation, displacement of victims of natural disasters, ongoing immigration, and fluctuating birth and death rates. December 2017 projections about the 2020 Census, drawn from existing ACS data, suggest state population changes could result in Texas gaining three congressional districts, Florida gaining two, and Arizona, Colorado, Montana, North Carolina, and Oregon each gaining one House seat. The same data suggests single-seat losses in Alabama, California, Illinois, Michigan, Minnesota, New York, Ohio, Pennsylvania, Rhode Island, and West Virginia (Brace, 2017).

The field of blindness sand visual impairment is no stranger to all the factors influencing population change. The "Silver Tsunami" of the Baby Boom generation includes an increasing number of older adults with vision loss. The racial and ethnic minority groups closely associated with immigration (particularly people of Hispanic ethnicity) are experiencing increasing rates of vision loss; researchers project that "by 2050, the highest prevalence of [visual impairment] among minorities will shift from African American individuals (15.2% in 2015 to 16.3% in 2050) to Hispanic individuals (9.9% in 2015 to 20.3% in 2050)" (Varma et al., 2016). Thus, accurately counting immigrants and seniors is critically important to accurately counting people who are blind/visually impaired and ensuring that these populations are represented in government.

Distributing Funding and Services

Another very important use of census data is for the distribution of federal funds meant to support various programs and populations. According to the Census Bureau, "Federal funds, grants and support to states, counties and communities are based on population totals and breakdowns by sex, age, race, and other factors. Your community benefits the most when the census counts everyone. When you respond to the census, you help your community get its fair share of the more than $675 billion per year in federal funds spent on schools, hospitals, roads, public works, and other vital programs" (2017c).

A horizontal bar graph titled “Prevalence of Vision Difficulty By Age” x axis measures prevalence ranging from 0% to 12% 5 bars: Population under 18 years = 0.73% Population 18-34 years = 1.05% Population 35 to 65 years = 2.34% Population 65 to 74 years = 4.19% Population 75 years and over = 9.94%

AFB uses statistics gathered by the Census Bureau to support our advocacy for appropriate services. This chart was prepared for use in a letter to the United States Senate Special Committee on Aging

Independent Living Services for Older Individuals who are Blind (often known as "the OIB program" or "the Older Blind Program") is a clear example of a federal program where the allocation of funding is determined by data from the Census Bureau. Although the current funding for the Older Blind Program is just $33.4 million, and it serves fewer than 60,000 individuals, this is the only source of federal dollars dedicated to supporting older adults with vision loss to live independently. When matched by state dollars, the Older Blind Program funding makes possible essential services and resources that would otherwise not be available. The majority states receive an annual OIB funding allotment determined by a formula based upon the proportion of the U.S. population over age 55 which lives in that state; states with smaller populations receive a "minimum allotment" of $225,000. (Visit AFB's webpage for the 21st Century Agenda on Aging and Vision Loss to download an excellent summary and advocacy document about the OIB Program from AFB's Public Policy team and Paul Saner, Chair of the National Council of State Agencies for the Blind). Each year, the Rehabilitation Services Administration relies upon data about seniors from the Census Bureau to allocate OIB funding; inaccurate data could result in disproportional funding and reduced services in states with the greatest needs.

The Census Bureau's data has countless additional applications to programs and services important to Americans with vision loss. The State of New Jersey's Department of Labor and Workforce Development published a list titled "50 Ways Census Data Are Used ." From this list, here are some especially relevant uses to the field of blindness/visual impairment:

Picture of older woman using white cane, getting on bus
  • Forecasting transportation needs and planning public transportation services
  • Planning health services, including deciding where health services should be located
  • Allocating funding for services for people living in poverty
  • Researching and implementing public safety projects
  • Studying and predicting the impact of natural disasters
  • Creating maps that ensure quick access to emergency services
  • Planning, organizing and developing facilities, education programs, and services to assist specific populations, such as people with disabilities, children, and/or older adults
  • Providing data to support scientific research, including data to justify and/or compete for research funding
  • Providing evidence in litigation involving land use, voting rights, and equal opportunity
  • Establishing appropriate boundaries for school districts and ensuring states and districts have necessary staff and resources to provide special education and other programs and services
  • Developing policies for fair renting and lending practices
  • Evaluating programs in different geographic areas
  • Identifying areas eligible for various types of housing assistance and loans
  • The census and ACS enable data-based decisions on all of these topics, promoting the equitable and most efficient allocation of tax dollars and other limited resources.

Are we ready?

Given that the United States of America - and specifically, Americans with vision loss - cannot afford to lose a single month of American Community Survey data, we certainly cannot risk a failed or inaccurate decennial census. As 2020 grows ever closer and we learn about new features like online submission of census forms or integration of external datasets to supplement missing data, we must ensure that people with vision loss have every opportunity to participate and be counted. Our community must stand with other census advocates to oppose funding cuts and any other threats to a continuous ACS and a thorough, accurate decennial census. With our input, vigilance, and advocacy, we can look forward to equal representation and continued and expanded use of data to make informed decisions and advocate for services and funding appropriate to the diverse and changing population of people with vision loss.


Brace, K. (2017). Some change in apportionment allocations with new 2017 Census estimates; but greater change likely by 2020 . Manassas, VA. Retrieved from

Cecco, L. (2015). What happened when Canada dumped its Census. Retrieved from

Coelho, T. (2017). Failed 2020 Census looms as funding, preparations stall. Retrieved from

Critical Census budget action needed. (2017). Retrieved from

Feinstein, D., Carper, T. R., Schatz, B., Cortez Maston, C., & Harris, K. (2018). [letter to Wilbur Ross, Secretary of Commerce]. Washington, D.C. Retrieved from

Long, H. (2017). Census 2020: How it's supposed to work (and how it might go terribly wrong). Retrieved from

Marshall, A. (2015). The tragedy of Canada's terrible census data. Retrieved from

Scherer, M., & Bahrampour, T. (2017). 2020 Census needs cash infusion, Commerce secretary will tell Congress. Chicago Tribune, October. Retrieved from

Sweetland Edwards, H. (2017, May). The head of the census resigned. It could be as serious as James Comey. Time. Retrieved from

U.S. Census Bureau. (2017a). American Community Survey: Response rates. Retrieved from

U.S. Census Bureau. (2017b). Road to the 2020 Census. Retrieved from

U.S. Census Bureau. (2017c). Why we conduct the decennial census. Retrieved from

Varma, R., Vajaranant, T. S., Burkemper, B., Wu, S., Torres, M., Hsu, C., … Mckean-Cowdin, R. (2016). Visual impairment and blindness in adults in the United States: Demographic and geographic variations from 2015 to 2050. JAMA Ophthalmol, 1-21.

services icon Directory of Services

Help Make Drug Labels Safer

Help make prescription drugs safe for people with low vision. Support our Rx Label Enable Campaign today.