Methodology

Identifying what does and doesn’t make a state’s elections susceptible to political interference turned out to be more complicated than it first appeared. Because every state runs its own elections, there are 50 separate election systems to assess, each a bit different from the others. Fortunately, election law is a well-studied field with a lot of expertise and data available to help guide the way.

To assess the state of the US election system, Bloomberg reporters around the US and others who cover national politics began with the 5,661 election-related bills proposed in state legislatures since 2020. We talked with lawmakers and state elections officials from both political parties, studied the bill texts, read signing statements and veto messages from governors, and looked at the existing statutes that would be changed if new ones took their place.

We also looked at the public statements of hundreds of elected officials to see how they responded to Donald Trump’s baseless claims that Joe Biden stole the 2020 election, and his attempts to overturn the results.

We then looked at the 396 bills that were ultimately signed into law to identify changes that would affect the act of casting a ballot, and the official counting of those ballots. We searched for trends and commonalities across states.

Next, we talked with local and state elections administrators; a variety of election security experts; political scientists with the US Elections Project and the MIT Election Data and Science Lab; researchers at the Bipartisan Policy Center; and advocacy groups including the National Vote at Home Institute and the Brennan Center for Justice.

This yielded many different perspectives on what does and doesn’t affect the fairness of elections, and the integrity of the results. But there was also broader agreement among experts on several points, which we narrowed to seven benchmarks for ease of voting and eight benchmarks for ballot security. We simplified each benchmark into a simple yes-or-no question so that we could better compare states and set values for each of the benchmarks.

Here are the benchmarks we settled on for ease of voting:

• Does a state allow voters to register at their polling places on Election Day?

• Can they register to vote online?

• Does a state automatically register an eligible voter when they interact with a government agency other than the motor vehicles bureau?

• Can anyone vote by mail or early in person?

• Can voters sign up to continue receiving mail ballots in the future?

• Can they check whether their mail ballot was accepted online?

• Can they go to a larger voting supercenter that handles all types of ballots instead of just the one in their neighborhood?

And here’s what we chose for ballot security:

• Does the state compare voter rolls with other states to find voters who’ve moved?

• Does it remove dead voters from the rolls regularly?

• Does it allow elections workers to begin processing mail ballots before Election Day?

• Does it have a reasonable deadline for mail ballots to be received?

• Does it require voter ID with reasonable accommodations for those who forget it?

• Does it count ballots using machines, which are faster, more accurate and cheaper than hand-counting?

• Is there a reasonable standard for automatic recounts?

• Does the state regularly audit election results?

Our goal was to find specific measures that were likely to have the most impact. Some were straightforward, such as whether a state allows citizens to register to vote on Election Day, one of the few measures that has consistently been shown to improve voter turnout.

The impact of others was less clear.

Take voter ID requirements, a source of controversy for nearly two decades. Democrats have long decried restrictive ID requirements for making it harder to vote, while Republicans have said they are necessary to fight fraud and instill voter confidence. But there’s little evidence for either assertion.

Research is mixed on whether voter ID laws hurt turnout in general or among minorities, with some studies finding it can reduce turnout and others finding no effect. Other studies suggest that restrictive ID laws may be counteracted by more campaign outreach to help voters get proper identification.

But there’s also little evidence that voter ID reduces fraud, already extremely rare in the US, or that it boosts voters’ confidence in the results.

Voter ID laws also vary state to state, which makes them hard to compare. In the end, we chose to ask what states do when someone forgets their ID; specifically, whether voters are required to take extra steps to get their ballot counted or if they were allowed to provide an alternate confirmation of their identity. The answer to that question yielded information about the possible intention of a state’s ID law—whether it was meant to prevent fraud or merely to make it harder to vote.

In other cases, we excluded something because good data wasn’t available. One was whether long lines at polling places suppresses turnout. Although no one likes waiting in line, there is no reliable source of data on its impact on turnout and how it compares state to state. Most studies we found have simply asked voters how long they waited, which isn’t a particularly useful measure.

Instead, we focused on the potential impact of well-studied rules that would reduce the likelihood of long lines, such as allowing any voter to cast a ballot early or by mail and creating supercenters where anyone from an area can vote.

For same-day voter registration, where a person can register and cast their ballot in one stop, we did not count Connecticut and Michigan as having effective same-day registration, even though it’s technically available. In both those states, some voters have to register in one location then go somewhere else to vote while others have to cast provisional ballots. Those policies made same-day registration more onerous than others that have a one-stop system.

There were some surprises. While Democrats typically emphasize voting access, some Democratic-led states, such as New York, did poorly in that category, while some Republican-led states, such as Mississippi, where officials have stressed the need for ballot security, fared poorly on those measures.

On ballot security, we found that many of the proposals considered since 2020 are what might be called “election security theater”—designed to address concerns about voter fraud that stem from conspiracy theories and baseless claims. Some state laws barred voting machines from being connected to the internet, even though most already typically are not internet-connected, amid false claims from Trump and his supporters about Italian satellites and smart thermostats being used to hack into the systems.

Instead, we looked at measures such as how accurately the state updates voter rolls and whether it regularly audits election results. As with our benchmarks on voting access, the point was not that any one measure makes elections secure, but that the more a state is doing, the less likely it is that problems could arise.

We did not attempt to rank the states or give them a numerical score based on the answers to the benchmark questions. That would imply a false precision, when many of our decisions about what to include or exclude in this assessment were somewhat subjective. We also didn’t weight one measure as being more important than another, since the data again suggests it’s the overall picture that matters. Our goal was to give a broad gauge of a state’s overall ballot security efforts.

Even without a numerical score, the benchmarks make it possible to see each state’s individual election profile as well as where a state sits generally in relation to others. A state with six or seven “yes” answers to the benchmarks on voting access and ballot security, for example, provides a much different voting experience for citizens than a state only meeting two or three.

To find the answers to those questions, we started with the 2021 Election Administration and Voting Survey report from the US Election Assistance Commission, a federal agency that collects and shares information on the state of American elections created in the aftermath of the contested election in 2000 between Al Gore and George W. Bush.

We then interviewed dozens of state elections officials to check the data for accuracy and fill in incomplete information from the survey, which asks dozens of questions about the elections process. We added any changes made by legislation enacted in 2021 and 2022, using a database maintained by the National Conference of State Legislatures.

We also referenced information from nonprofit organizations such as the Electronic Registration Information Center, the Verified Voting Foundation and Ballotpedia, and Bloomberg Government’s state legislation and state regulation trackers.

For the statements of office-holders, we checked press releases, social media and contemporaneous news reports for comments from governors, attorneys general, secretaries of state and members of Congress on their positions on the 2020 election between Nov. 3 and Jan. 6, the day the Congress voted to certify the electoral college for Biden.

We examined whether they backed up their words with actions, such as supporting lawsuits by Trump-friendly states to challenge the results, and voted in the House to certify the election after the attack on the US Capitol. We checked more recent statements to see whether they say they still don’t believe that Biden is the legitimately elected president. And we looked at what they pledged to do about elections in the future.

We also consulted an archive of older and deleted social media posts relating to the election created by US Representative Zoe Lofgren of California, a Democratic member of the committee investigating the Capitol assault that occurred the day the vote was certified; amicus briefs and public letters they signed or submitted; op-eds they wrote for local news outlets; and speeches they made in Congress.

Finally, we looked at campaign websites, interviews, convention speeches, political ads and social media posts for candidates for key posts such as governor, secretary of state, attorney general, US senator and US representative for all 50 states.

There were some judgment calls. For each state, we tallied major elected officials and nominees who have explicitly denied that Biden won, or said the 2020 election was not free and fair. We also included those officials who only raised procedural concerns about how the election was conducted as long as they also took some action, such as publicly urging the Department of Justice to investigate, seeking to have the Supreme Court intervene or objecting to certification on Jan. 6. We included detail on what they have said and done so that readers can decide for themselves.

We will periodically update this database to note when new laws have passed or election deniers win primaries or general elections. We will also be writing further stories based on what we’ve learned while reporting this story.


Story by: Ryan Teague Beckwith and Bill Allison
Graphics by: Paul Murray, Allison McCartney and Mira Rojanasakul
With assistance by: Rachael Dottle, Marie Patino, Jenny Zhang, Gregory Korte, Romy Varghese, Vincent Del Giudice, Nathan Crooks, Margaret Newkirk, Shruti Date Singh, David Welch, Elise Young, Dina Bass, Brendan Walsh, Carey Goldberg and Maria Wood
Editors: Wendy Benjaminson, Wes Kosova, Alex Tribou and Yue Qiu
Photo editors: Eugene Reznik, Marisa Gertz and Maria Wood
Photo credits: Getty Images, Bloomberg and AP Photo