Category: Government / Topics: Culture Education Government Policy Research Methodology Statistics

Will I be Counted?

by Stu Johnson

Posted: August 21, 2020

Don't be left out of the 2020 Election or Census!…

Election 2020 Update

As Election 2020 in the United States heats up with the Democratic convention this week and the Republicans next week—both resorting to virtual formats because of the COVID-19 pandemic—controversy over mail-in voting has also boiled over.

When I wrote "How Should I Vote?" on August 8, there were assurances from the United States Postal Service (USPS) that mail-in ballots could be handled in a timely manner. The credibility of that claim came into question since then as increased attacks by President Trump on mail-in voting and service cutbacks by Postmaster Louis DeJoy combined to produce charges of sabotage by Trump critics. This was heightened by a letter sent by the Postal Service to a number of large states warning of potential problems in being able to deliver mail-in ballots on time (requirements vary by state). Then, earlier this week, DeJoy announced that changes (ostensibly to combat perennial financial problems) would be delayed until after the election.

Much of the controversy is political warfare. However, in this atmosphere, many news reports have been quick to point out Trump's unsubstantiated (their word, or similar) charges of fraud and DeJoy's role as a Trump donor (with no mention that he was a former shipping executive, except by columnist Byron York). Most postmasters have come from government service or executive positions in the private sector, and several had tenures that spanned presidents of both parties, so emphasizing that he is a Trump donor does make DeJoy more vulnerable to attacks of cronyism.

While too many reporters simplistically dismiss as ludicrous Trump's claims of fraud, I have been pleased to see more stories recently taking a serious look at potential problems of any kind with our election system(s). On the other hand, I have yet to hear much, if anything, related to the potential for trouble I pointed out in the previous article: even if ballots arrive on time, when will they be counted?

If there is very high mail-in voting in states new to that practice and counting is not done until after November 3, the results of the election all the way down the ballot could be in question through mid-November and beyond.

Therefore, let me reiterate and update a few points from the previous article:

  • Be an informed voter. Federal law sets the date and protections of voting rights, but the way the election is conducted is a matter of state and local authority. In most states, voting is administered by a county election board, overseen by a state-wide election authority. Check the website of your local election authority, or contact the county clerk's office. Go there first, before relying on information from any of the myriad of websites and organizations offering help, which can vary widely in accuracy and trustworthiness.
  • If you vote in person at your polling place on Election Day, or at a designated Early Voting site, be prepared to follow COVID protocols (face mask and social distancing). Your ballot should be included in the official Election Day count. In some cases (same day registration, change of address, etc.), your ballot may be set aside as "provisional," requiring further verification before being added to the final tally. Again, this can vary by state.
  • If you are considering vote-by-mail find out how it works where you live. If your election authority will not count mail-in ballots until after Election Day, seriously consider voting in person at your precinct polling place on Election Day or an Early Voting site if available. Otherwise, your vote will not be included in the Election Day totals, but in the final count up to two weeks later.
  • If you vote-by-mail.
    • Make application early. If your state offers mail-in voting and you have not received an application from your election authority, check its website or call. Be wary of anything sent in any manner by anyone else. Most states are already accepting applications, though ballots will usually not be sent out until September or early October. If there is a tracking option, use it.
    • Don't procrastinate! Return the ballot early. It may not be counted until after Election Day, but it will have arrived early enough to be counted. If you are uneasy about sending it in the mail, many areas are providing secure drop boxes. Check with your election authority for locations and availability.
    • Need assistance? If you need help at any stage, use people you can trust. Be extremely wary of individuals or groups offering to help you obtain, fill out, and return your ballot.

2020 Census

When it was announced that field interviews, used to contact people who have not responded, would end September 30, instead of October 31, there was an outcry from some corners not unlike the current brouhaha surrounding mail-in ballots. Was this, in fact, an attempt to under count certain people?

That sounds plausible, given our current political climate. However, over the years, I have worked with Census Bureau data and can attest to to the Bureau's professionalism and desire for accuracy. I have used reports from the Census Bureau for numerous articles. (See a list). So, what is the story with Census 2020? Basically, COVID-19 put much of the census operation on hold just as it was getting startied.

In addition, the focus on racism and inequality that began with the murder of George Floyd by police in Minneapolis continues into the presidential election. This points out the need for everyone, citizen and non-citizen alike, to respond to Census 2020, which will impact representation in Congress.

First, some background

The purpose of the decennial (ten-year) census is to enumerate (count) all residents by state (citizens and non-citizen, regardless of age) in order to apportion members of the U.S. House of Representatives among the 50 states.

Interestingly, the birth and development of the United States government is very much tied to the census. With a bicameral Legislative branch, the Constitution gave each state two members in the Senate, but apportioned members of the House of Representatives by population. Naturally, you need to know how many people there are. Thus, the Constitution established the decennial census.

You may think, so what? Census taking was common, wasn't it? In fact, that was a major innovation in governance. At the time of the creation and adoption of the Constitution (1780s), only Sweden had conducted a regular nationwide census in modern times. We were already conducting our second census by the time Britain and France started the practice. It was part of the Enlightenment embrace of reason and empirical observation.

Because there was no Census Bureau yet, the first census was conducted in 1790 using U.S. Marshals to supervise hired enumerators who did the actual counting.

Census Day remains April 1. The Census is intended to be a snapshot of where everyone considered home on that day, even though the actual collection of data goes on for months.

The amount of demographic data collected has changed over the years. The population had grown so much and the questions so numerous that the 1880 census ran out of money and the detailed tabulation could not be completed before it was time to prepare for another census. Much of that was driven by the rapid growth in the 19th century of "political arithmetic" and "social mathematics"—which coalesced into the what we today call statistics.

That crisis, in turn, led to the development of tabulating machines using punch cards representing census responses, precursors of today's computers. The central, Constitutionally-mandated purpose remains the same: to determine the number of members allocated to each state in the House of Representatives.

Because of its continual efforts to collect detailed demographic data, the Census of 2010 began using only the short formquestionnaire required for apportionment and redistricting. Before that, a select number of people (1 in 6 in 2000) received a long form, with considerably more detail, often including such things as whether you had indoor plumbing, electricity, a radio, a car, and more, reflecting the nature of American life at the time the census was being taken

There are two steps in reporting the results:

Apportionment is the process of dividing the 435 memberships, or seats, in the U.S. House of Representatives among the 50 states based on the apportionment population counts from the decennial census. By law, the apportionment results must be submitted to the President by December 31 of the census year.

Redistricting is the process of revising the geographic boundaries of areas from which people elect representatives to the U.S. House of Representatives, a state legislature, a county or city council, a school board, etc. By law, redistricting data must be submitted to the states by April 1 of the year after the census.

For more detail see the Census Bureau's FAQ on Congressional Apportionment.

It seems redundant now, but the process historically—in the pre-digital age— involved three steps, still contained in the U.S. Code. The President received the apportionment report (state populations and calculated number of representatives) by December 31, then passed it on to the Clerk of the U.S. House of Representatives within one week of the opening of the next session of Congress in the new year. Within 15 days of receiving the report the Clerk had to inform each of the state governors of the number of Representatives apportioned.

The report to the governors has evoked into the Redistricting Data due April 1, which provides more detailed information (i.e., breaking down population by county or smaller units) that may be used by states when they redistrict. If there is a change in the number of representatives, the Congressional district map would have to be redrawn. Even with no change in apportionment, redistricting will adjust for changes in the distribution of population within each state since the last census.

Gerrymandering—oddly contorted districts favoring one party—occurs in states where the majority party in the legislature controls redrawing of the map. The map can change radically with population shifts and more so if the majority party changes hands. (Illinois, where I live, is like this. Attempts to get a "fair map" referendum on the ballot have so far been unsuccessful. A fair map would be created by an independent or bipartisan group, with an emphasis on districts that are compact and contiguous.)

The U.S. Constitution mandates apportionment—setting the number of representatives— but is silent on redistricting—how maps are redrawn and what data is used. According to the National Conference of State Legislatures:

  • 21 states explicitly require the use of census data for legislative, congressional or both redistricting
  • 17 states have an implied basis or in-practice reliance on using the census for legislative, congressional or both redistricting
  • 6 states permit the use of the census or other datasets for their redistricting, depending on circumstances.
  • 6 states do not fit the categories above.

Thanks for allowing me to take this side trip through history. If you are interested in digging deeper, check out The Sum of the People. by Andrew Whitby, a new book that describes "How the Census Has Shaped Nations, from the Ancient World to the Modern Age."

Much of the following information comes from an August 17 Census Bureau presentation, "Review of 2020 Operational Plan Schedule."

2020, before COVID-19

The work of conducting the census is coordinated across 248 Area Census Offices (ACOs). By early May, materials were to be distributed to known residential addresses or field workers were to make contact with other subsets of the population, some to leave materials, others to collect data. This process is aimed to catch everyone, including the homeless, people in transitory locations, group quarters, and others unreachable by mail.

Following these initial contacts, "Nonresponse Followup," where field workers attempt to reach everyone who has not responded, was to be conducted between May 13 and July 31. That would leave five months to tabulate and publish results to meet the statutory deadline of December 31,2020 to deliver apportionment data. That tells the states whether there will be a change in the number of members in the U.S. House of Representatives. The redistricting data was originally due April 1, 2021. It would be expected that a partial roll-out of data could begin 6 weeks prior.

2020, adjustments because of COVID-19

Just as Census 2020 was about to get started in earnest, COVID-19 went from threat to global pandemic. In the U.S., states declaring stay-at-home orders began to spread until it was clear this was a national crisis. Census 2020 field operation ground to a halt, but mailings to residential addresses continued. By the end of April, the National Self-Response was at 55.6%. The December 31 statutory deadline for Apportionment data was still in place. Not bad, but collection of data requiring people in the field was in jeopardy, including Nonresponse Followup, originally scheduled to start May 13.

Congress considered moving the Appropriation Data deadline to April 30, 2021 and Redistricting Data to July 31, 2021. This would allow the Nonresponse Followup to be moved from May 13-July 31 to August 11-October 31 (roughly 11 weeks each). In the end, the Appropriations Data deadline remained at the original December 31 date, with Nonresponse Followup scheduled for August 9-September 30 (8 weeks).

That is three weeks less than originally planned, but was it a conspiratorial attempt by the administration to purposely under count segments of the population? It is far more complicated than that, as indicated by material on the National Conference of State Legislature website, showing the patchwork legal requirements that states have regarding use of data and timelines. Besides, the field work for special categories was shifted on the calendar, with little impact on the actual time dedicated to each one.

The time from the end of field work to the release of Appropriation Data has been reduced from five weeks to three. For the much more detailed Redistricting Data—originally set for March 30, 2021 with a suggested extension to July 31, 2021—the August 17 presentation indicates "plan in development." Here, too, the requirements of states to begin redistricting will have an important impact on the final plan.

Perhaps it would help you understand the difference in producing the two reports fi you can visualize the punch cards used for so many years to tabulate results, with each card representing one person.

  • To produce the Appropriation Data requires the relatively simple task of sorting the cards by state and then counting the cards for each state.
  • Developing the more detailed Redistricting Data requires further sorting of each state's cards, for example, dividing by geographic region (county, MSA—Metropolitan Statistical Area, city, zip code), age, sex, race, education, income or any other responses.

    Cross tabulation, necessary for analysis and development of policy and programs, requires even more refined sorting: i.e., county by age, race by education, or city by age by income (which I had to do in my role on my city's Housing Commission to answer the question, "How many seniors in the city of Wheaton have income below the poverty level?")

    With physical punch cards, that required resorting each state's cards for each criterion and desired cross-tab. Computers still require capturing the data, but can then do all the sorting and cross tabulating electronically. The census produces a tremendous amount of detail, which will continue to roll out after what is needed for Redistricting is reported. In between decennial censuses the Bureau continually monitors available data and conducts surveys to produce estimates of current statistics.

Where do we stand? Have you responded?

As of today (August 21) the Census Bureau reports (for data collected through August 20) total response of 73.9%—64.2% by self-response, 9.7% by field enumerators for nonresponse followup. This is the first decennial census where nonresponse followup results have been included in interim reports.

For comparison, the final response rate for the 2010 census was 74.1%—66.5% by self-response, 7.6% by nonresponse followup. (This information was surprisingly hard to find, and may not be accurate. I will update it if I can find or be directed to more reliable information).

Go to the 2020 Census Response Rates page to see how the census is going, including the ability to drill down by state, county, city, congressional district or tribal area. From this page you can get other information about the census and respond, if you have not done so yet—and do it now. Join the people where you live to count toward representation, and to help Census 2020 achieve its mission of counting every person, counting them once, and counting them in the right place.

This article also appears on, which Stu edits.

Stu Johnson is owner of Stuart Johnson & Associates, a communications consultancy in Wheaton, Illinois focused on "making information make sense."

E-mail the author (moc.setaicossajs@uts*)

* For web-based email, you may need to copy and paste the address yourself.

Posted: August 21, 2020

Go to the list of most recent InfoMatters Blogs
Search InfoMatters (You can expand the search to the entire site)

Share and Comment via Social Media, or Print this Post