copyright notice
link to the published version: IEEE Computer, January, 2022
2022 IEEE Computer Outstanding Column Award

accesses since January 14, 2022



Hal Berghel

Our profession has reason to take pride in the intense study of election integrity in the U.S. by computer science researchers. To a large degree it is due to these research efforts that we have come to understand the real strengths and weaknesses of modern election systems when manufacturers and software designers cloaked their products in secrecy and government oversight into the integrity of the systems is minimal. One such computer scientist is Douglas Jones.

Doug is an emeritus professor in the Computer Science at the University of Iowa. He has been involved in voting technology research since 1995 and was a principal investigator for the National Science Foundation (NSF)-funded ACCURATE project (A Center for Correct, Usable, Reliable, Auditable, and Transparent Elections; His book with co-author Barbara Simons, Broken Ballots: Will Your Vote Count? (CSLI Publications, 2012), is a seminal work in the area of current voting technology and is highly recommended to anyone who is concerned about election integrity at the ballot box. The “interview” that follows resulted from our email exchanges during September and October, 2021 . This is the third interview with Doug on this subject since 2016. (n.b., Some of Doug's published work may be found onlin e at ).

Hal Berghel: In 2016 we discussed why prominent computer scientists were critical of DRE [direct recording electronic] voting machine vendors who refused to build their equipment around robust security models. Has anything changed in the past five years?

Doug Jones: There has been a significant shift in the voting equipment marketplace. The era of direct-recording electronic voting machines is all but over. Most voters today vote on paper ballots that are tabulated by optical scanners. Hand-marked paper ballots dominate in the United States, being used universally for vote-by-mail ballots as well as in 2/3 of all polling places.

20 years ago, electronic ballot-marking devices (BMDs) emerged as a new technology. These have been widely accepted as assistive devices for voters who are unable to mark a paper ballot, but in recent years, several states have adopted ballot marking devices for all voters, and now 1/5 of all polling places operate this way. [VEVO]

Unfortunately, while voters can theoretically verify that the paper ballot marked by a BMD reflects their intent, there have been several research studies that show that most voters do not do so. While these studies do show that there are interventions that can improve the rate of verification, the rates achieved are not high enough to assure election integrity. [KORT][BERN]

A second problem is created because most of the new BMDs record the voter's selection twice on the paper ballot, once in human-readable print, and once in the form of a one or two-dimensional bar code. While voters can verify the text, the scanners for these ballots only interpret the bar code.

Aside from the use of paper ballots, the evidence I have seen about the security of the software inside voting machines and election management systems continues to suggest poor design, at least from the point of view of security. Sadly, the most recent study of this issue remains under a court-imposed seal and has not been released either to election officials or the voting system vendor.

HB: Also in 2016 you indicated that there was a new generation of optical mark scanners that could read the ballot and record the vote, and also capture the full image of the ballot for auditing purposes. This would seem to be an ideal way to return to paper ballots. What is the status of this technology and how widely was it deployed in 2020?

DJ: Essentially all of the new ballot scanners are able to retain full images of the ballots, but not all election officials opt to retain these images. The 1960 Civil Rights Act, in Section 301, requires that election officials retain “all records and papers which come into his possession” for 22 months. Clearly, any computerized device creates huge numbers of transient records as it performs its function. We cannot require the retention of a complete trace of every computation involved in an election. When a voting system offers the option of retaining or not retaining some record that has traditionally been transient, does Section 301 require retention?

Some states have conducted election audits from the digital images without reference to the original paper ballots, but this is problematic. [USEAC]

There is room for additional research on audit methodologies that both authenticate the ballot images and use them as the basis for a machine-assisted audit.

HB: Much has been made of the increased integrity of voting systems that have strong software independence (i.e., the independence of the vote tabulation and audit from any software in use). Current ballot-marking devices (BMDs), and even DRE+VVPAT ( Voter-verified paper audit trail) systems, are not strongly software independent (SSI), which means that the systems that use them are subject to misconfiguration, hacking, fraud, etc. [APPE] Travis County Texas (Austin) has experimented with a DRE+VVPAT system called STAR-Vote that uses DRE, VVPAT together with a smart ballot box that authenticates the paper ballot with respect to the vote held in limbo by the DRE. Does this achieve SSI? Are there any other alternatives that you know of?

DJ: STAR-Vote tries hard to find such a solution, and indeed, it may be largely software independent under the assumption that the human-readable content of the paper ballots is subject to routine audits comparing it to the electronic record of the election. The weakness of STAR-Vote lies in the GUI that collects the voter's selections, before the cryptographic commitment step. It shares this weakness with all DRE+VVPAT and BMD voting systems.

The fundamental problem with these systems is that they rely on voters to proofread the paper ballot at the end of the voting session. People are notoriously bad at proofreading. With hand-marked paper ballots, the voter verification step that matters is the view of the ink flowing from the pen as the voter marks the ballot. That is immediate feedback, not something delayed.

In addition, many studies have found that when voters do find problems while proofreading their ballots, they frequently blame the problems on their own carelessness instead of on the voting machine. If a voting machine cheats on, for example, 1% of the votes, and half of the victims do not notice this, while those that do assume that they made mistakes entering their votes, the machine will have successfully added a ½% edge to its favored candidate without creating any suspicion of fraud. Even an audit of the spoiled ballots might not notice a problem against the background of ballots spoiled because of genuine voter errors.

HB: When one compares the 2018 the Federal Election Commission's list of states that do not require post-election audits [LOVA] with its list of 2020 presidential general election results ( ), one notices that all of these states that do not require post-election auditing are red states. Coincidentally, it appears that none of these 13 states have any statutory requirement for voting machines to have a paper trail. [NCSL] What accounts for these 13 states' aversion to election verification?

DJ: The push for voter-verified paper trails largely grew out of controversies arising from the election of 2000. It is an unfortunate fact that the winner in any election is reluctant to encourage any retrospective analysis of that election or to support any reform proposals resulting from that analysis. To do so could be seen as acknowledging questions about the legitimacy of that election.

In the case of the election of 2000, the Republican party took control of the White House, and the resistance to post-election reforms was strongest in Republican controlled states. After the 2020 election, the tables have been turned, and it is the Republicans who are asking questions about election integrity.

After the election of 2000, it took several years for a consensus to emerge about appropriate changes. The initial reaction, among many, was to ask for technologies that we now know were quite flawed, pure DRE machines, and then DRE+VVPAT and BMD technologies. It took most of a decade for a clearly articulated push for software independence and post-election auditing to emerge.

After the election of 2020, we are seeing some Republican critics of the election demanding major changes. I suspect that some of these changes are as poorly thought out as some of the initial reactions to the election of 2000, but I also see some demands that make sense. While forensic audits of the type conducted by Cyber Ninjas in Arizona may be seriously flawed, the proponents of these audits are asking for universal use of paper ballots subject to auditing in states that had previously resisted these changes. Perhaps, in time, this will lead to a national consensus on minimum standards that will satisfy both Democrats and Republicans.

HB: During this past election cycle, the initial automated count, hand recount, automated recount, risk-limiting audit, and signature audit of absentee-by-mail ballot oath envelopes all confirmed the initial presidential election results and collectively produced no evidence of voter fraud in Georgia during the 2020 Presidential election. [GASOS][ABSV] This may have been the most thoroughly analyzed election result in U.S. history, and yet undocumented claims of widespread voter fraud persist. [FUNKE] Although Georgia has had a spotty history regarding election integrity, [OTTO] [ROSEN] the 2020 election results would appear to be as uncontestable as might be expected in Georgia at least. Would you care to speculate on why the Republican state leadership incurred so much wrath from the Trump legal team over this election? Do you think this will become a trend in 2022 and 2024?

DJ: On October 20, 2016, then candidate Donald Trump declared that he would accept the election results, if he won. Throughout the 2020 campaign, Trump repeated the claim that the only way he could lose was because of fraud. The net result, in states like Georgia and Arizona, is that a sizable fraction of the Trump base was primed to accept claims of fraud after the election.

In my 25 years of work on election technology, I have had several encounters with lawyers in both parties who clearly stated that their job was to see that their candidate won, by any means. One was working on Trump's challenges to the 2020 election. In light of this personal experience, I don't find the continuing attack on the election results in Georgia or other states to be surprising.

Unfortunately, I expect this strategy of contesting election results to continue because it has been effective at keeping key parts of the Trump base united and engaged.

HB: There has been considerable support for risk-limiting audits (RLA) in the past decade or so. [STARK]. What varieties of risk-limiting audits are currently in use? Since RLAs are used for detection of anomalies and not used for counting ballots, why aren't they more widely deployed? Also, how do you account for the confusion over what an RLA is (e.g., vs. canvass).

DJ: Risk limiting audits typically involve drawing randomly selected ballots from among all ballots cast. Randomly selecting one particular ballot from among all ballots in such a way that each ballot is equally likely to be selected is easiest if the set of all ballots is sequentially numbered and stored in sequence. This, in turn, is easy in states like Colorado where the majority of ballots are cast by mail and all ballots are counted on centralized scanners. In such a case, the scanners themselves can be equipped to print sequence numbers of the ballots as they are scanned and the scanned ballots can be stored in order.

After a sample of ballots is drawn, the next question is, do you count the votes in the sample and ask if the sample confirms the election result, or do you compare the individual ballots to the result of the first scan. If the ballots are serial numbered in a way that allows each ballot to be compared with the record of scanning that particular ballot, you can do ballot comparison.

With precinct-count voting technology, ballot comparison is difficult because retaining the order the votes were cast threatens ballot privacy. Voting is a public act, so observers can directly observe who votes in what order, and in many cases, the check-in process at the polling place creates a record of the order votes were cast. So long as voter coercion or intimidation remain significant threats, both ballot images and the ballots themselves really need to be shuffled before an audit.

As a result, jurisdictions that use precinct-count voting equipment cannot do ballot-comparison audits without first re-scanning and serial numbering all of the ballots. This amounts to a complete machine recount as part of the audit. The alternative is a ballot tabulation audit that focuses on counting the votes in the sample.

The next great question is how does the audit respond to a discrepancy. Typical risk limiting audits respond to discrepancies by escalation of the sample size. The initial sample size is chosen to offer a specified level of assurance that the announced election result is correct, assuming that no discrepancies are found. If a ballot comparison audit finds a particular number of discrepancies, the sample needs to be enlarged under the assumption that that rate of discrepancies is typical. If a ballot tabulation audit finds a different winner than the winner that was announced, the sample size needs to be increased to see if that was a statistical fluke, an artifact of taking a small random sample, or actual evidence of a problem. The judge found these claims vacuous and dismissed them with prejudice. Was there any substance to these claims?

  1. “‘[T]he absentee voting counts in some counties in Michigan have likely been manipulated by a computer algorithm,' and at some time after the 2016 election, software was installed that programmed tabulating machines to ‘shift a percentage of absentee ballot votes from Trump to Biden.'”
  2. “Smartmatic and Dominion were founded by foreign oligarchs and dictators to ensure computerized ballot-stuffing and vote manipulation to whatever level was needed to make certain Venezuelan dictator Hugo Chavez never lost another election.”
  3. “The several spikes cast solely for Biden could easily be produced in the Dominion system by reloading batches of blank ballots in files such as Write-Ins, then casting them all for Biden using the Override Procedure (to cast Write-In ballots) that is available to the operator of the system.”

DJ: I find these claims to be bizarre.

  1. By midsummer of 2020, it was widely understood that Democrats were far more likely to vote absentee than Republicans, largely because of highly partisan COVID denial. Actual election results bore this out in several ways.

    In-person versus absentee ratios were distinctly different for the two parties in 2020 compared to 2016. This difference does show that the election of 2020 was different from past elections, but the confusing graphs of this data presented by the Trump legal team as evidence of fraud show nothing more than the expected result of the partisan split over absentee voting.

  2. Smartmatic was founded in 2000 by Venezuelan expats, and it did develop an election system sold for use in Venezuela before buying Sequoia Voting Systems in 2005 from its previous British owners. Smartmatic sold Sequoia in 2008 amid concerns about foreign ownership of US voting companies, and Dominion (originally a Canadian company founded in 2002) acquired all of the assets of Sequoia in 2010, the same year it also acquired the assets of Premier Election Systems (formerly part of Diebold).

    In addition to real and serious questions about the security of Diebold's voting systems and about entanglements between Smartmatic and Venezuela, numerous conspiracy theories have swirled around both companies. Unfortunately for Dominion, it has inherited these along with the assets of the companies it acquired.

  3. The time-series data for the publicly released partial vote totals do indeed show spikes as totals from different precincts and absentee ballot batches are added to the running totals. While it is true that the fraud mechanisms suggested would produce such spikes, it is also the case that absentee ballots are counted in batches, and as each batch count is added to the total, there is a spike. Furthermore, in jurisdictions where there are large numbers of absentee ballots, absentee ballot processing is expected to continue after election day. When party affiliation is a factor in the decision to vote absentee, this leads to vote totals that swing after election day. All of this was well known before the election. [RAKI]

HB: One of Trump's attorneys, Sidney Powell, claimed among other things that Dominion Voting Systems rigged their voting machines to allow Biden to win the popular vote. This led Dominion to bring a defamation lawsuit against her, seeking $1.6 billion in damages for defamation. [FEUER] [121-1] Powell has undertaken what appears to me to be an extraordinary and tenuous defense. She claims that she can't have defamed Dominion because no ‘reasonable' person would have taken her accusations seriously in the first place. [121-222] Is there any evidence that supports her original claims about Dominion and Smartmatics? Is there any reason to believe that her defense strategy might prevail in federal court?

DJ: I find it difficult to find a coherent story in the “evidence” presented by Powell's lawsuits. Smartmatic controlled Sequoia for 3 years prior to its purchase by Dominion a decade ago. How this could allow Smartmatic to reach into Dominion's product line acquired from Premier (formerly Diebold) baffles me. Some of the allegations seem to require unwieldy conspiracies between election officials, vendors and foreign powers, necessarily involving hundreds or even thousands of co-conspirators.

I'm not a lawyer, but I certainly find Powell's “no reasonable person” defense to be bizarre. Her lawsuits have certainly asked the courts to take her allegations seriously, so is she implying that the courts are not reasonable?

HB: William Barr's response to the Trump campaign's claim of election irregularity was: “To date, we have not seen fraud on a scale that could have effected a different outcome in the election.” [KARL] It is noteworthy that he put a negative spin on this. What he did not say is that the available evidence confirms election integrity. Wasn't the voter-verified paper audit trail (VVPAT) environment created so that public could be assured of election integrity? It appears that either Barr didn't even find VVPAT results fully compelling, or that his response was politically motivated. What are your thoughts?

DJ: A general election in the United States involves over 5000 different local election jurisdictions, over 170,000 precincts, and over 2,000,000 election workers, including precinct officials and election office employees. Furthermore, most of those election workers are temps, not professionals. In any enterprise this large, no reasonable person should expect perfection. There will be clerical errors and some of those workers will be less than honest. Barr's wording carefully avoids the implication of a perfect election.

Paper ballots introduce well-known problems maintaining a secure chain of custody. Under the pressure of election day, election workers sometimes cut corners on properly documenting chain of custody. The so-called “forensic audits” at the behest of Trump supporters after the election have found instances of this, at a low and not shocking level. I have seen evidence that similar problems have been covered up in the Georgia audit and recount. I think that Barr's wording correctly anticipated that such discrepancies might be found. [FULT][MARI]

Much of the press for election auditing has taken for granted that chain-of-custody issues would be dealt with. As these “forensic audits” have shown, current election procedures are not always as good as we would like. Simple measures such as check lists that have been used in safety critical systems since World War II are not being widely used in election set-up or take-down.

HB: You recently attended the Mike Lindell Cyber Symposium in Sioux Falls, SD which purported to offer “proofs” that the 2020 presidential election was stolen proffered by computer hacking experts. Who were these experts and what was the specific nature of their “proofs”?

DJ: Attending that event was a truly odd experience. Lindell advertised that he would reveal 37 terabytes of Internet packet-captures (Pcaps) that proved foreign interference in the election, but only a token fraction of this was released to the technical experts at the event. By the end of the event, the consensus was that there were no Pcaps in this data. My guess is that Mike Lindell was swindled by those who provided him with the purported Pcaps. While this data was being analyzed, speakers offered other “evidence.”

One featured speaker, Douglas Frank, presented elaborate graphs and charts that I never fully understood, purporting to show that the coefficients of a higher order polynomial were the key needed by an apparently vast conspiracy that has had complete control of US elections since the passage of the Motor Voter Act of 1993. Frank's allegation implied a complex algorithm that customized the theft of votes based on local demographics.

In contrast, Seth Keshel had data that showed that the election of 2020 was unique while the elections of 2012 and 2016 were normal. Furthermore, his data pointed to a small fraction of the counties in the US where something was different. He considered those counties to be the ones where fraud was likely. It was my impression that what he had identified were the counties where exurban "Main Street Republicans" were so offended by Trump that they voted Democratic, many for the first time in their lives.

Meanwhile, whenever nothing was going on, there was a projected data showing the “fake news” (that is, official) vote totals for Trump and the “real” (that is, Lindell's) numbers. One of the speakers at the Symposium, Shiva Ayyadurai, later pointed out (in an interview with Steve Bannon), that Lindell's numbers were uniformly 4.2% above the official numbers. Shiva's talk at the symposium focused on his experience as a candidate in Massachusetts, not on the presidential race, but in his discussion with Bannon, he said of Lindell's data, “either it's a hoax or people read The Hitchhiker's Guide to the Galaxy ” (in which the number 42 plays a special role). [SHIVA]

These different positions are incompatible. A uniform 4.2% theft does not require fancy demographic data or fitting higher order polynomials the way Frank proposed, and it is not compatible with Keshel's relatively small number of counties. Frank's theory that election fraud has been universal and running for decades is equally at odds with Keshel's data.

HB: What was Steve Bannon's contribution in this symposium? It isn't obvious to me how a partisan political advisor fits into the veil of “proofs” claimed by Lindell.

DJ: In addition to producing his War Room podcast from the event, Bannon gave a stirring talk saying that “what happened in Wuhan, that virus, what happened on Nov. 3, and what happened on Jan. 6. We'll get our country back on the answer to those questions. … It's a sucker's play to try to fight the next election, we must go back and unwind the last.” This was met by cheers from an audience that included significant numbers of state legislators. This suggests that, to those agreeing with Bannon, there is little room to negotiate for election reform prior to a reversal of Trump's defeat.

HB: Over the past five years, we have had discussions of voting machines [HBDJ1] and election integrity [HBDJ2] printed in Computer. This year, we add to our list the topic of election delirium. I didn't think it was possible that an election that could produce a popular vote plurality as well as an electoral college majority could be more fractious than the five elections where the president was elected by overruling the popular vote (1824, 1876, 1888, 2000 and 2016). It appears that our topics are getting more onerous over time. Are there any grounds for optimism in 2022, 2024, and beyond?

DJ: It was hard to find any cause for optimism after listening to Bannon's talk. But Shiva, in his interview with Bannon, ended on a more positive note, suggesting that the questionable numbers being offered by Lindell were a distraction from the real issues of how we conduct elections. Many in the Stop the Steal movement are reading the work done by the election integrity movement over the past 20 years, and despite the fact that radicals are twisting that work to their own ends, there are some who may be open to reasonable reforms.

HB: These interviews are becoming a regular feature with us.  Would you care to make a prediction on the topics that will occupy our attention in 2024?

DJ: My crystal ball is broken, so I can't really make useful predictions. I expect that the stop-the-steal movement will continue to challenge elections and push for roadblocks to voting for years to come. I also expect that local election jurisdictions will continue to be seriously underfunded and I expect the shortage of technically knowledgeable election staff to continue.

Despite this dismal situation, some reforms are possible. I hope that more states will require such simple things as rudimentary inventory control and checklists to help reduce the incidence of chain-of-custody errors, double-counting of ballot batches and ignored ballot batches. I hope more states require routine post-election audits, and among the states that audit, I hope that more of them will move toward more rigorous audits such as risk-limiting audits.

I don't expect anything rapid in the world of election technology. Voting equipment is built to last a decade or more, so even if we change our voting system requirements now, most of the hardware and software now in service will still be there in 4 years. On the other hand, we must find a way to update our requirements so that they take an end-to-end perspective on voting systems instead of merely focusing on the voting machine in the precinct, which is just one link in the chain from voter to the official election result.



[KORT] Kortum, P., M.D. Byrne, J. Whitmore, Voter Verification of Ballot Marking Device Ballots Is a Two-Part Question: Can They? Mostly, They Can. Do They? Mostly, They Don't, Election Law Journal, Vol. 20, No, 3 (2021). ( )

[BERN] Bernhard, M., A. McDonald, H. Meng, J. Hwa, N. Bajaj, K. Chang, J. A. Halderman, Can Voters Detect Malicious Manipulation of Ballot Marking Devices?, 41st IEEE Symposium on Security and Privacy , May 18-20, 2020. ( )

[USEAC] Post-Election Audits, US Election Assistance Commission, Oct. 23, 2020. ( )

[APPE] Appel, A., R. DeMillo and P. Stark, Ballot-Marking Devices (BMDs) Cannot Assure the Will of the Voters, April, 2018. ( )

[LOVA] Lovato, J., Risk-Limiting Audits – Practical Application, US Election Assistance Commission, June 25, 2018. ( )

[NCSL] National Conference on State Legislatures, Voting System Paper Trail Requirements, 6/27/2019. ( )

[GASOS] Historic First Statewide audit of Paper Ballots Upholds Result of Presidential Race, Georgia Secretary of State news release[undated]. ( )

[ABSV] Georgia Secretary of State/Georgia Bureau of Investigation ABM Signature Audit Report, GASOS Investigation report SEB2020-257, 12/29/2020. ( )

[FUNKE] Funke, D., Fact check: No evidence of fraud in Georgia election results, USA Today, June 1, 2021. ( )

[OTTO] Ottoboni, K. and P. Stark, Election Integrity and Electronic Voting Machines in 2018 Georgia, USA, E-Vote-ID 2019 Proceedings, 2019. ( )

[ROSEN] Rosenfeld, S., Georgia's Hand Count of 2020 ballots was No Risk-Limiting Audit,, November 22, 2020. ( )

[STARK] Stark, P., Risk-Limiting Postelection Audits: Conservative P-Values From Common Probability Inequalities, IEE Transactions on Information Forensics and Security, 4:4, December, 2009, pp. 1005-1014. ( )

[RAKI] Rakich, N. and E. Mejia, When To Expect Election Results In Every State, FiveThirtyEight , Oct. 31, 2020. ( )

[FEUER] Feuer, A., Dominion Voting Systems files defamation lawsuit against pro-Trump attorney Sidney Powell., The New York Times, March 26, 2021. ( )

[121-1] Complaint and Demand for Jury Trial, US Dominion, Inc. v. Powell (1:21-cv-00040), Document 1, U.S. District Court for the District of Columbia, 01/08/21. ( )

[121-222] Motion to Dismiss (1:21-cv-00040-CJN Document 22-2), U.S. District Court for the District of Columbia, 03/22/21. ( )

[KARL] Karl, J., Inside William Barr's Breakup With Trump, The Atlantic, June, 2021. ( )

[FULT] Fulton County Pennsylvania Election System Analysis, WAKE Technology Services, Inc., February 19, 2021. ( )

[MARI] Maricopa County Forensic Election Audit, Volume III: Result Details, Cyber Ninjas, Arizona State Senate Republican Caucus, Sept. 24, 2021. ( )

[SHIVA] Steve Bannon interview with Dr. Shiva, Bannon's War Room , Episode 1,167 (Aug. 12, 2021), minute 42 to end. ( )

[HBDJ1] Berghel, H., Douglas Jones on Today's Voting Machines, Computer, October, 2016, pp. 84-89. ( )

[HBDJ2] Berghel, H. and D. Jones, Election Integrity 2018: Are We Better Off Than We Were in 2016, interview, Computer, November, 2018, pp. 90-94.( )


Hal Berghel

My thanks to Doug Jones for participating in another interview on election integrity in the US. I will conclude with my own historical perspective.

As Doug emphasized, there is no such thing as perfect elections. We can only strive for fair ones. Bill Barr opined that the 2020 presidential election bore no evidence of significant voter fraud. Despite that, we came very close to a coup when one candidate explored extra-electoral options to nullify the results. [PERIL][LEON][BEND][BOWD] For a while in late 2020 and early 2021, the U.S. election began to resemble that of a banana republic. This shows just how fragile the voting franchise is in the U.S.

A case can be made for the claim that the Trump campaign dissatisfaction with the outcome of the 2020 election derives from the fact that the Constitution never guaranteed universal suffrage in the first place. Article II speaks of sundry congressional votes, but nothing of an inclusive popular voting franchise. It must be remembered that the principle of “one man, one vote” wasn't accepted by the Supreme Court as an entailment of the Fourteenth Amendment's Equal Protection Clause until the 1960's. The Constitution dealt with procedural issues, not equity issues.

The power elite in the late 1700s felt that the right to vote should be limited to white, male property holders of European descent. As John Jay, the first Chief Justice of the U.S. Supreme Court made clear at the time, the people who own the country ought to govern it. [JAY] No one could accuse Jay of equivocation on this point. President John Adams insisted that one cannot invest the voting franchise in property less men. Why if this were allowed, where would the franchise end? [CAMP] Republican strategist Paul Weirich embraced a similar sentiment in 1980: "I don't want everybody to vote. Elections are not won by a majority of the people. They never have been from the beginning of our country and they are not now. As a matter of fact, our [Republican] leverage in the elections quite candidly goes up as the voting populace goes down." [WEIR] The same view has been held by Mitch McConnell [BERMAN] and Donald Trump. [LEVINE] Former Supreme Court Justice Antonin Scalia correctly pointed out that “… there is no right of suffrage under Article II.” [CAMP] When Scalia opposed the pre-clearance provision of the 1965 Voting Rights act as a “perpetuation of racial entitlement” that was not grounded in the constitution [TROUTT][SORKIN], he was reaffirming the views of the founding fathers that only the rich, wellborn and able should participate in the governing of the republic. Voter disenfranchisement is baked into the Constitution of the U.S.

The popular vote trend in presidential races over the past forty years is pretty clear to everyone. One political party is losing touch with the voting public. [STAT] While the 2020 “stop the steal” vote fraud gambit was effective at animating fringe groups of partisans, it proved ineffective at distracting the majority of the public from the facts of the election outcome. I take a slightly different view of this than Doug. The “stop the steal” gambit will not go away, but it will remain indefinitely just beyond the event horizon of a cultural black hole of history along with chain letters, pyramid schemes, crop circles, vanity license plates, email scams, alien abductions, and beliefs in flat earths. What will survive, however, is the time-honored tradition of political subterfuge: vote suppression.

The reality is that in the absence of a popular vote mandate combined with an unwillingness to appeal to a broader segment of the population through substantive policy changes, the only way to win presidential elections is through extra-electoral means – i.e., to suppress or ignore the popular vote. As mentioned above, this has been accomplished five times in the U.S. by forcing the electoral decision to more partisan and predictable interests in the electoral college or the House of Representatives. My prediction is that we will see the strategy of vote suppression consume partisan domestic politics for the foreseeable future. Vote suppression will slowly but surely displace stop-the-steal narratives. The 2020 election results confirmed for those intent to undermine democratic elections in the US in the next decade, that the low hanging fruit in anti-democratic, partisan politics is to prevent people with opposing views from voting in the first place or to ensure that their votes don't get counted. If I am correct, you'll be able to see it in the background noise of future elections: long wait times at voting precincts, expanded use of purge lists and caging, closed primaries, primary caucuses rather than primary elections, increased dependence on rigid voter ID requirements, gerrymandering, and legislative pressure to abandon practices that encourage increased voter participation such as early registration, vote-by-mail, early voting, and absentee voting. In other words, the emphasis will continue to be on vote suppression tactics that have been unchallenged by the judiciary. In the past, the technique that seems most immune to judicial interference is vote dilution by techniques like redistricting and at-large elections. A primary goal of the so-called cracking, stacking and packing of voting districts is to force the party that is not in charge of redistricting to waste their votes based on the reality that from the perspective of electoral outcomes, all votes over the minimum required to win an election are wasted! Through gerrymandering, a minority of voters have been able to maintain control of some voting districts for centuries. Elbridge Gerry's discovery in 1800 proved that under certain conditions politicians in the U.S. are legally able to define their constituencies to circumvent the will of the majority. [BERG]

In short, anything that will favor noncompetitive elections and discourage universal suffrage will be fair game in future elections, up to and including preventing university professors from serving as expert witnesses in suits against the government on voting rights. [CHRON] The Federalist sponsors would no doubt be euphoric that the leadership of the state and university of Florida has kept the flame of their alien and sedition acts alive and well in the twenty-first century.

Given all of the limitations of today's election infrastructure that Doug mentioned in the interview – voting machines that rely on unverified proprietary code, the lack of adequate auditing, the problems of maintaining a secure chain of custody of voting records, the lack of strong software independence, and so forth – it is a wonder that our elections are as reliable as they are. This is in no small measure due to the vigorous investigation of the technology by computer science researchers. [BERG]


[PERIL] Woodward, B. and R. Costa, Peril, Simon & Schuster, New York, 2021.

[LEON] Leonnig, C. and P. Rucker, I Alone Can Fix It: Donald J. Trump's Catastrophic Final Year, Penguin, New York, 2021.

[BEND] Bender, M., Frankly, We Did Win This Election: The Inside Story of How Trump Lost, Twelve/Hachette Book Group, New York, 2021.

[BOWD] Bowden, M., and M. Teague, The Steal: The Attempt to Overturn the 2020 Election and the People Who Stopped It, Atlantic Monthly Press, New York, 2022.

[JAY] Respectfully Quoted: A Dictionary of Quotations, 1989. ( )

[CAMP] Campbell, T., Deliver the Vote: A History of Election Fraud, an American Political Tradition – 1742-2004, Carroll and Graf, New York, 2004.

[WEIR] Weirich, Paul, video, .

[BERMAN] Berman, Ari, Mitch McConnell Admits That Republicans Lose When More People Vote, Mother Jones, January 30, 2019. ( )

[LEVINE] Levine, S., Trump says Republicans would ‘never' be elected again if it was easier to vote, the Guardian, 30 Mar 2020. ( )

[TROUTT] Troutt, D., Voting Rights: Scalia V. Minority Protection, Rutgers CliME, 05 March 2013. ( )

[SORKIN] Sorkin, A., In Voting Rights, Scalia Sees a “Racial Entitlement,” The New Yorker, February 28, 2013. ( )


[STAT] O'Neill, A., Share of popular votes for the Democratic and Republican parties in presidential elections from 1860 to 2020, Statistica, Feb 18, 2021. ( )

[CHRON] Thomason, A., U. of Florida Stops 3 Professors From Taking Part in Voting-Rights Suit, Raising Cries of Censorship, The Chronicle of Higher Education, October 31, 2021. ( )

[BERG] Berghel, H., Chasing Elbridge's Ghost: The Digital Gerrymander, Computer, November, 2016, pp. 91-95.