Skip to main content
COVID-19 Update: University offices are open but being staffed remotely. Classes will be conducted remotely for all summer sessions.

Issues in Implementing All-VBM Elections in New Jersey

by Patrick Murray

[The following was presented as testimony to the New Jersey General Assembly State and Local Government Committee on June 17, 2020.]

I am not an expert in voter access law or the mechanics of election administration. I do know something about voter motivation and behavior. So that’s what I will address today in providing some context about what we saw in New Jersey in May’s local elections and what we may see in the July primary and beyond.

In 1992, more than 90% of all votes in America were cast traditionally – that is to say at polling places on Election Day. By 2016, the Election Day vote had dropped to 60% of national turnout, with the remainder evenly split between mail ballots and in-person early voting.

It is likely that vote by mail – or VBM – will increase this year because of Covid even without a state change. Half of voters in a recent national Monmouth poll said they are very (31%) or somewhat (20%) likely to vote by mail this year. Covid is accelerating the already existing move in states to provide more convenient – and safer – access to the voting franchise.

I want to talk about two related impacts – turnout and public confidence in the system.


If you looked at the turnout in last month’s local elections here, you’ll notice a large increase for many, but not all, of those races. Looking at the six Bergen County school districts that held Board of Education elections, average turnout more than doubled in these contests compared to last year’s April elections. However, we didn’t see quite the same jump in more urban communities. Turnout was much higher in Irvington for both its municipal and school board elections, but was up only slightly in Newark’s school board race. Turnout was actually lower in municipal contests in Orange and Paterson than for the same races four years ago.  However, in both those cities, these races were less competitive than in the prior cycle – and you need to factor that in.

We need a larger sample of elections before we can draw any firm conclusions about VBM in New Jersey. However, it does seem that VBM had a bigger impact on turnout in the suburbs. There is one other side note to this. It appears that there were fewer undervotes on the public questions relative to overall turnout. So perhaps, VBM prompts more voters to fill out the entire ballot.

But July will be a statewide election, so the question is what do we know about turnout from other states that have implemented all-VBM?  Researchers are not quite agreed. It looks like there were initial bumps in Oregon and Washington when they first implemented all VBM more than a decade ago, but those gains may have been maintained inconsistently. VBM does seem to have had an impact in CO, which made the switch only recently. The impact seems to be higher in lower turnout elections.

There is no evidence that VBM has a partisan impact on turnout, even though Republicans tend to express less enthusiasm for it than Democrats. A study of recent elections in CA, UT, and WA by Stanford researchers, which was just released last week, found a negligible shift in pre-existing partisan turnout advantage related when those states moved to all VBM elections (maximum 0.7% Democrat).

Looking back at May, the overall picture is that VBM may increase turnout more in the suburbs than the cities. Irvington stands out as an exception to this limited finding. That may be in part because the municipal and school elections were combined. Holding fewer elections in a year has a measurable impact on increasing turnout by decreasing voter fatigue.

We also saw a five-fold increase in turnout for mayor and council in Montclair, which was much higher than the increased turnout in other municipal contests. However, most of the Montclair seats were uncontested in the prior cycle. And we know that the one thing guaranteed to cause a sizable increase in turnout is a competitive election. But it doesn’t look like we are going to have too many of those in July.

-Public Confidence-

The other issue I want to discuss is public confidence. Researchers have found a correlation between confidence that one’s vote will be counted accurately and the likelihood to vote. This is one area where New Jersey could have problems in July, particularly if there are close contests in high profile races.  And that’s because implementing VBM properly takes time.

A recent study by Brigham Young University political scientists found a decrease in public confidence in Utah after that state went to all-VBM. The percent of all Utah voters who said they were very confident that their ballot was counted accurately went from 69% after the 2008 election to 60% in 2018.  Nearly all the remainder were somewhat confident and only about 5% were not confident in either year. However, that dip in confidence was tempered by past experience with voting by mail. Among voters who had cast a mail ballot in two or more elections, strong confidence in the process stood at 68% – nearly identical to the 2008 number for all voters. This result was a similar 64% among those with one VBM ballot under their belts, but it was only 53% among those who voted by mail for the first time in 2018. Thus the switch to all VBM caused an initial dip in confidence that disappeared once voters experienced it.

On the other side of the coin, an examination of the voter files in one particular California county after it switched to all-VBM more than a decade ago actually found a decrease in individual voting turnout after that change. Confounding factors in that county included poor communication about the switch and large numbers of non-English speaking voters. These issues were not factors in Utah, Oregon or Washington State, but are the kind of things we could see here in New Jersey with our diverse population and sudden implementation of VBM.

It’s important to caveat this. New Jersey may face certain VBM problems that are not noticeable in July, because it is already a low turnout primary without high-profile competitive races at the top of the ticket. If we have an all- or even high-VBM election in November, though, these cracks in implementation could become more apparent.

So, let me mention some key factors that should be considered in implementing and examining VBM elections:

First, if you look at states that have successively implemented VBM, “vote by mail” is actually an inaccurate term. Yes, all voters receive their ballot by mail. But the vast majority actually deposit it at one of many offices and drop-box locations near them. Only about one-third of voters in these states actually return their vote by mail.  In other words, they do not rely on the U.S. Postal Service to carry the full load of delivering ballots. We already saw some problems in May with relying almost exclusively on the U.S. Postal Service to handle returned ballots. Some were missing the requisite postmark and others were not delivered to the county clerks in time, or at all. A successful VBM process requires multiple drop off locations that voters can easily access.

Second, we need to give voters the opportunity to rectify rejected ballots. According to the National Conference of State Legislatures, at least 16 states require clerks to notify voters if their ballot signature does not match the voter file. Election officials are not handwriting experts and should not have the final say based only on their own visual examination.

Third, allowing additional time for ballots to be received and problems to be addressed is also important for building public confidence. Some states only rectify problems for ballots received before Election Day, but others allow for a longer period to ensure full voter access. Washington, for example, gives voters from 10 to 21 days after the election to rectify signature discrepancies.

Fourth, another feature of a high-confidence VBM process is providing an easy way for a voter to verify that their ballot has in fact been accepted – via either automated phone interface or online.

Finally, VBM actually addresses a glaring security problem in New Jersey’s current voting system – the lack of a paper trail. While there are costs associated with moving to all-VBM elections, they more than offset the capital investment required to address the urgent need to replace all of the state’s existing voting machines.

These things, along with an extensive voter communication campaign, are some of the things that have contributed to the successful transition to all-VBM elections in other states. We lack most of these factors here in New Jersey. Even so, we may have a seemingly “successful” primary election here next month. However, that experience may not tell the whole story of how VBM might perform if we administer November’s high turnout general election in a similar way.

Trump Job Rating “Bump” in Context

By Patrick Murray

Fact 1: Donald Trump’s job rating is at an all-time high.

Fact 2: Donald Trump has not received the same approval “bump” as past presidents in a crisis.

Recent shifts in the president’s job approval have been met with “either alarms or fist pumps,” as one reporter put it to me.  But we really have to keep this in context.  We have become so accustomed to the fact that Trump’s numbers never move all that much, that we accept that as the norm. The current crisis is just an exceptionally stark example of that.

To put this in perspective, if this were any other president, we would expect job ratings to have swung almost instantaneously by at least 10 points.  George W. Bush got a nearly 30 point bump after 9/11.  John F. Kennedy saw a double-digit hike in his already high ratings during and after the Cuban Missile Crisis. Even Jimmy Carter got a 25 point bump in 1979 when Americans were taken hostage in Iran.

In Monmouth’s polling, Trump’s approval rating is only 2 points higher than where it was one month ago, before the pandemic really spread in the country. And for context, his current rating is 3 points higher than two months ago in the midst of his impeachment trial, and 5 points higher than six months ago, when the impeachment process was just getting underway.  Monmouth’s numbers track consistently with the average of all polls.

The scale of these shifts means that we end up trying to discern significance from infinitesimal amounts of evidence.  I am not saying that these small movements cannot be consequential. When the country is as evenly divided as it is now, they most certainly can be the tipping point for political change. What I am really trying to say is that is it very difficult to explain the reasons for these shifts at the microscopic level of detail many observers want. That’s because standard public opinion polls are not the right tool for the job. They are more like magnifying glasses than microscopes.

Let’s take the recent shift in Monmouth’s poll numbers as an example. The one major change we saw in Trump’s job rating was that approval among Democrats went from 6% last month to 11% now. The numbers for Republicans (91%) and independents (44%) stayed exactly the same. Now, the fact that the latter groups were exactly the same in the poll does not mean they are exactly the same in reality, because of the potential margin of error in the poll sample. It’s just that we know they did not move as much (if at all) as the Democrats.

A five percentage point movement among a group that makes up about a third of the population is microscopic in polling terms. Absent a sample size in the tens of thousands, we just don’t have the ability to examine this group with any level of precision. In real terms, this shift may represent about 3 or 4 million adults across the country. In polling terms, this equates to approximately 15 respondents in a sample of 850).

It is likely that this group had a range of reasons for changing their opinion. For some it was probably movement from soft disapproval to soft approval for a specific thing Trump had done. For others it may be aspirational.

There’s a body of literature about the psychological need to rally around a leader in times of crisis, which is why the bigger research question for a student of public opinion is why that effect isn’t bigger right now rather than finding explanations for the few people who have become more positive toward the president.

Part of the explanation is certainly down to Trump’s inability to project a more inclusive, non-partisan persona as well as a steady hand on how his administration is tackling this situation. Part of the explanation is the failure of opposition leaders to signal to their followers that they should get behind the president (which admittedly is difficult for them to do as Trump’s rhetoric continues to lambast those who don’t show due deference to him).

Basically, the current times are blowing away a lot of the political theories about what typically happens in a time of crisis.  And that, to me, is the more important public opinion story right now.

Should We Reform the Presidential Nomination Process?

by Vincent Grassi, Monmouth University Polling Institute Intern

The process by which the Democratic Party chooses its presidential nominee has faced increased scrutiny this election cycle, especially after a flop at the Iowa caucuses. One argument has been that the first two states that hold a caucus or primary, Iowa and New Hampshire, do not represent the demographics of the Democratic Party and hold an outsized influence on choosing the party’s nominee.  A majority of Democratic voters (56%) believe that Iowa’s caucuses and New Hampshire’s primary have too much influence over who wins the party’s presidential nomination, according to the Monmouth University Poll. One in four say having these states go first hinders the party’s ability to nominate the best candidate. It may be time to shake up the primary calendar and try something different.

To begin, what are some of the advantages of having Iowa and New Hampshire go first? Candidates with limited funds can stay competitive in small states with less expensive media markets but would not be able to compete against more well-funded candidates in bigger states. Since Iowa and New Hampshire are small states, candidates who are less well known or have fewer supporters are given the opportunity to gain traction and secure a win, something that would not be possible if larger states like California went first.

Candidates turn to different methods of campaigning in these states to establish closeness with the electorate. For example, in small states like New Hampshire, candidates hold small, intimate gatherings with voters like town halls. In bigger states there wouldn’t be an incentive to hold these types of events. Candidates would be more focused on holding large rallies and giving interviews on television. Also, candidates would have little motivation to campaign in a smaller state like New Hampshire if a larger state like Florida (with more delegates to be won) were to vote first.

However, there are some disadvantages to the current system as well. The first two states, Iowa and New Hampshire, provide poor representations of the demographics of the Democratic Party. For example, having New Hampshire as the first primary has been criticized because the state’s racial demographics are 90% white, with African Americans only making up 1.7% of the population, Asian Americans 3%, and Hispanic Americans 3.9% according to the US Census Bureau.

The addition of Nevada and South Carolina as early states somewhat offsets the problem regarding the demographics of Iowa and New Hampshire. With this, some claim the first four states in the nomination process are somewhat representative of the country. Hispanic Americans account for 29% of Nevada’s population. In South Carolina, African Americans account for 27% of the population and make up an even larger share of the Democratic electorate. The media uses these states to gauge candidates’ support among minority groups. 

 However, if a candidate does not perform well in Iowa and New Hampshire, they are usually compelled to drop out of the race and never have the opportunity to compete in more diverse states. The candidates that perform well in Iowa and New Hampshire get increased media coverage that propels their campaigns. The process is all about gaining momentum leading up to Super Tuesday.

Iowa’s caucus system has been criticized for being noninclusive. It disenfranchises large parts of the state’s population due to its design and procedures. According to the ACLU of Iowa, the physical demands of caucusing in the state, “… makes participation difficult for people who can’t get or afford child care, people living with disabilities or mobility issues, people who lack transportation, and people who work evenings.” According to the Brookings Institute, the current way, “caters to older voters and those for whom politics is a passion.” Voter turnout in Iowa was also lower than expected this year.

What are some possible solutions or reforms that can be made to better the process? There are multiple different options and proposed reforms that can replace or complement what we have now. The Monmouth University Poll shows a majority of Democratic voters (58%) favor a national primary day. A national primary day would make it so that all states hold a presidential primary election on the same day. Proponents of this reform claim that this system would be more efficient, however, opponents argue that it would be very expensive for candidates to compete on a national scale.

Another reform that garnered 15% support in the poll is having a few other states hold their contests on the same days as the Iowa caucuses and New Hampshire primary. This would keep the drawn out process of the current system and calm the criticism surrounding Iowa and New Hampshire going first.

According to the Monmouth Poll, one in ten support replacing the current system with grouped primaries. Rotating regional primaries is one reform that The National Association of Secretaries of State has supported. This would group states by their geographical location (making segments) and create the primary schedule according to region. The order of these primary elections according to region would be rotated each election cycle.

Another possible reform is the elimination of caucuses and establishing primaries with ranked-choice voting. Not only does this address the concern over the accessibility of caucuses, but it may also answer the concerns expressed by Democrats over how much influence early states like Iowa and New Hampshire have over the candidate pool. Ranked-choice voting would enable fringe candidates to stay in the race even after Iowa and New Hampshire by allowing people to rank their preferences in a primary vote.

Ranked-choice voting could ensure that nominees have broader support among the electorate. According to FairVote, “The system incentivizes candidates to work together rather than attack one another in the hopes of earning backup choices, and to campaign to a broad swath of voters rather than just their own base.” Nevada used ranked-choice voting in their early voting period which helped increase voter turnout significantly according to FairVote

It might be time to give more consideration to reforming the current presidential nomination system. Criticism over Iowa’s and New Hampshire’s status as the first states to vote on presidential candidates has caused many to reexamine the process. Within our country’s history, the way political parties nominate their presidential nominees has evolved and changed to suit cultural shifts. By weighing the advantages and disadvantages of the current system, we can propose the right reforms.