NBC6 Investigates

Lawsuit to invalidate abortion rights amendment relies on flawed state report

NBC6 Investigates finds that the Florida Department of State report on alleged Amendment 4 petition fraud contains misstatements, questionable methodology and inaccurate numbers. Days after their release, those same numbers appeared in a lawsuit seeking to invalidate the abortion rights amendment. 

NBC Universal, Inc.

Twenty-five days before Florida voters would be asked to establish a woman’s right to abortion in the state constitution, the Department of State issued a report that appears to contain inaccurate calculations used to estimate what it claimed was fraud in the petition collection process, according to an NBC6 Investigates review.

The exact scope of that fraud is critical to a lawsuit seeking to have the amendment stricken from the state constitution if voters approve it on Nov. 5. It relies heavily on the report made public by the Department of State’s Office of Election Crimes and Security (OECS) five days before the lawsuit was filed.

Watch NBC6 free wherever you are

>
Watch button  WATCH HERE

But a review of the report by NBC6 Investigates found it includes apparent factual misstatements, questionable methodology and inaccurate numbers. After requesting more time last Thursday to do “research and find answers to” questions posed by NBC6 about its report, the Department of State has declined to respond with any of the promised answers.

Amendment 4 -- “the amendment to limit government interference in abortion” – would allow abortions up to the point of fetal viability (usually around 24 weeks) or when necessary to protect the patient’s health, gutting Gov. Ron DeSantis’ and the state legislature’s six-week abortion ban that took effect in May.

Get local news you need to know to start your day with NBC 6's News Headlines newsletter.

>
Newsletter button  SIGN UP

Even if the amendment gets the required support of more than 60% of the voters, the lawsuit filed in Orange County by Amendment 4 opponents on Oct. 16 seeks to have it declared “null and void.”

It was filed by C. Alan Lawson, a former Florida Supreme Court justice whose firm has been paid more than $838,000 by the DeSantis administration since he retired from the court in 2022, according to a state database. He has not responded to an email seeking comment from him or his clients.

In their lawsuit against Floridians Protecting Freedom, Inc., the group behind Amendment 4, and state and county elections officials and offices, four voters in St. Lucie and Taylor counties are challenging whether supporters of Amendment 4 in fact obtained enough signatures to make it onto the ballot.

But law professor Bob Jarvis of Nova Southeastern University said to succeed, the lawsuit needs more than the statistical speculation produced by the OECS.

“Ultimately it'll come down to whether or not the people pushing this lawsuit can prove fraud,” Jarvis said. “There's no reason to think that the supervisors of elections didn't do their job properly. So, if I had to make a prediction, I would predict that ultimately the lawsuit will fail. I'm not sure, though, that this lawsuit is really about winning or losing.”

Instead, he said, “I think what this lawsuit really is about is throwing a monkey wrench into the works a couple of weeks before Election Day in the hope that it will get some people either not to vote at all or to not vote for Amendment 4 who were planning to vote for Amendment 4 because they've heard that there's a problem.”

THE MAGIC NUMBER: 891,523

For this election, 891,523 valid petition signatures statewide were required, including a sufficient number in at least 14 of the state’s 28 congressional districts.

After supervisors of elections reviewed petitions and rejected about 400,000 of them as invalid, 997,035 verified petitions remained -- 11.8% more signatures than required statewide. They also exceeded the number required in 17 of the state’s 28 congressional districts – three more than necessary.

Relying on the numbers in the Department of State report, the lawsuit claims the “actual number of valid petition forms” submitted is 16.4% lower than the number verified by county supervisors of elections, 

If that were accurate statewide, Amendment 4 would not have had enough valid signatures to get on the ballot.

To supporters of abortion rights, that’s a big “if.”

“I question the entire statistical analysis,” said Lauren Brenzel, director of the Yes on 4 campaign. “This lawsuit is frivolous and salacious and it's distracting from the reality that Florida has a near total abortion ban without workable exemptions that is harming Florida women and putting their lives at risk.” 

In addition to apparent factual misstatements in the report, NBC6 Investigates also found incorrect calculations and methodology to produce a key figure: the 16.4% “average total invalidity rate” relied on in the lawsuit to estimate more than 163,000 verified petitions were actually invalid.

FACTUAL MISSTATEMENTS

For its report, the Office of Election Crimes and Security (OECS) reviewed petitions in only three of Florida’s 67 counties: Palm Beach, Orange, and Osceola.

It states Osceola was chosen because it “would be helpful” to compare the larger Palm Beach and Orange counties “with a smaller county where petitions submitted by known and suspected fraudsters accounted for a smaller percentage of validated forms.”

But that is incorrect, based on the Department of State’s own data.

The 2,216 suspected fraudulent petitions in Orange County comprise 3.1% of the 70,886 verified in that county and the 41 in Palm Beach a minuscule share of those verified.

But Osceola’s 1,378 suspected forms make up 8.3% of the 16,700 verified forms in that county – a much larger percentage than in the larger counties and in direct contradiction of one of the investigators’ stated rationales for choosing Osceola as part of its audit.

And the apparent mistakes in the state’s report continue in that same paragraph.

The OECS claims that 19.6% “of the 11,306 forms validated in the portion of Orange County within the Ninth (Congressional) District” were submitted by known or suspected fraudsters. That’s 2,216 validated suspect forms in just that district, the report states.

But according to a footnote and table on the previous page, those 2,216 suspect petitions were drawn from all over Orange County, which has all or part of four congressional districts. Orange County voters in the Ninth District produced just 16% of the county’s 70,886 validated Amendment 4 signatures, state records show.

So that 19.6% figure cited by the OECS is vastly inflated from what may have been less than 4%, based on the Ninth District’s contribution to the county total. 

Aside from those apparent factual misstatements – which, again, the Department of State has failed for eight days to refute, explain, or clarify – the report also uses what a national authority on statistics confirms to NBC6 was a flawed statistical analysis and what appears to be an outright mathematical error.

FLAWED METHODOLOGY

In creating a key statistic relied on heavily in the lawsuit seeking to invalidate the amendment, the state committed what an NBC News data analyst described as a “math trap” of taking an average of averages.

To see how the state audit fell into that trap -- in producing what it called an “average known suspected fraud invalidity rate” -- consider the number of petitions from known and suspected fraudsters (KSF) reviewed and their share of all validated forms in each county.

Despite reviewing just 0.05% of the petitions in its “audit” of Palm Beach, versus 3.1% and 8.3% in Orange and Osceola, respectively, the report gives Palm Beach equal weight to the others in computing what it calls its “average known suspected fraud invalidity rate” of 25.4%.

It simply added the three counties’ rates of 36.6%, 32.3% and 7.4% and divided by three – producing an average of averages.

But Ronald Wasserstein, executive director of the American Statistical Association, said he “can’t think of any good reason to compute the average the way the agency did.”

Instead, Wasserstein said, it “should be calculated the way” NBC6 Investigates suggested: add the number of invalid ballots in all three counties (832), divide it by the number reviewed (3,635) and produce a known/suspected fraud invalidity rate of 22.9% – down from the 25.4% the state calculated.

That may not seem significant, but -- combined with an outright mathematical error -- it would have a cascading effect on a key figure relied on by the lawsuit seeking to invalidate Amendment 4: the 16.4% “average total invalidity rate” the lawsuit argues proves fraud carried Amendment 4 over the finish line.

“It's very clear that they cherry-picked, that it's not statistically significant and that they were out to get a result and they were going to get that result no matter what it took.” Jarvis said. “So there was no reason to believe that the report and its findings have any validity.”

CHECKING THE MATH

Then there’s what appears to be a simple mathematical mistake.

This table from the report claims a review of 9,810 verified petitions in the portion of Orange County in Congressional District 9 revealed 20.9% were invalid. 

Total Verified Petitions Reviewed9,810
Valid6,840
Invalid2,017
First review indeterminate815
Missing341
Invalidity rate20.9% or more
2017/9810=20.6%

But check the math: 2,017 invalid petitions divided by 9,810 reviewed actually equals a 20.6% “invalidity rate.”

Running that same calculation in three other similar tables in the report confirms the figures the state says were the “invalidity rates” in those tables.

Table 1: 15/41=36.6% - Table 2:  715/2216=32.3% - Table 3:  102/1378=7.4% - Table 4: 2017/9810=20.6%
Table 1: 15/41=36.6% - Table 2: 715/2216=32.3% - Table 3: 102/1378=7.4% - Table 4: 2017/9810=20.6%

But not so in the Orange County Congressional District 9 table.

That 20.9% figure was used in other calculations to support the state’s conclusions, including ultimately the 16.4% “average total invalidity rate” cited in the lawsuit.

In addition to his concerns about the biases of the investigators’ methods, when Jarvis was told of the factual misstatements and basic math error he concluded, “The report has no integrity … I don't think anybody should put any stock in this report.”

JUST ENOUGH TO FLIP COUNTIES

That 16.4% becomes even more important when analyzing its effect on each of the state’s 28 congressional districts.

Remember: in addition to getting 891,523 signatures statewide, the campaign needed to reach certain thresholds in at least half of the state’s congressional districts.

They did so in 17 districts – three more than required.

But if one assumes that 16.4% “average total invalidity rate” can be fairly applied to all 64 unaudited counties, it would invalidate just enough petitions to flip four of those districts to the opponents’ side.

That means the amendment would not have been approved for the ballot regardless of the statewide total.

But two of those four districts would flip back if a lower invalidity rate were applied – one that corrects for the Orange County math mistake and for the flawed methodology the national statistical expert confirmed the state used to produce the “average known suspected fraud invalidity rate” in the three “audited” counties.

Making those two changes would reduce the suspected fraud invalidity rate in the unaudited counties from 16.4% to 14.6%.

That is just enough of a decrease to return two of those four districts into the Amendment 4 supporters’ column, giving them enough to qualify for the ballot in 15 congressional districts -- one more than necessary.

Even applying the lower, more fairly calculated invalidity rate, the statewide total would still fall below the 891,523 signatures threshold. Opponents would need to prove 105,513 (11.8%) of the signatures were fraudulent to get below that required statewide threshold.

So far, the OECS report says, it has found fewer than 2,900 to be invalid.

BEYOND THE STATISTICS

Statistical speculations aside, rejecting 105,513 petitions would be a high bar because about 400,000 invalid petitions were weeded out by supervisors of election before they could be included in the total count, said Broward Supervisor of Elections Joe Scott.

In Broward County, one third of the nearly 150,000 petitions submitted were found to be invalid by the supervisor of elections, leaving nearly 100,000 verified petitions added to the statewide total.

Accepting the state’s math, the lawsuit would claim Scott’s office missed another 16,400 invalid petitions – something he said could not have happened.

“I think that the number of rejections just goes to show that the ones that we accepted are good,” Scott said.

He said those rejected include some innocent mistakes or sloppiness.

“I wouldn't say that it's necessarily indicative that anybody was trying to commit fraud,” he said. “I'll leave that up to investigators, leave it up to the courts to determine if that's the case.”

INTO THE COURTS

And that may, in fact, be where this is headed, though Brenzel of Yes on 4 said she was “absolutely enraged by the idea that we're still having a conversation about petition collection. This is a done and settled issue.”

But it may not be so settled.

In claiming the courts could throw out the amendment -- even after it might pass -- the lawsuit cites a 2006 appeals court decision that appears to support its claim.

That case involved a 2004 amendment that authorized voters in Miami-Dade and Broward counties to approve slot machines at parimutuel facilities, such as horse racing tracks and jai alai frontons.

Before voters approved it, opponents filed suit claiming the supporters’ paid petition gatherers committed fraud – similar to the allegations made in the Amendment 4 lawsuit.

A trial court judge in Tallahassee dismissed the case after voters approved the amendment, but the First District Court of Appeal reversed, finding the opponents could indeed have the amendment struck from the constitution after voters approved it -- if they proved sufficient fraud at trial.

“It is clear that a favorable popular vote cannot cure deception,” the court held. “To the extent that, but for the fraudulent actions, the constitutional amendment would not have been presented to the public in the general election, there is no room in the process for fraud that places an amendment on the ballot.”

Of course, allegations are one thing; proof is another.

And while the slots case was sent back for a possible trial where evidence of fraud could have been presented, the parties wound up jointly dismissing the action without going to trial.

In addition to challenging the number of valid signatures, the current lawsuit also alleges violations of law in how petition gatherers were paid and in how records were retained and submitted.

Jarvis suggests the opponents’ motive is to “push the vote total in favor of Amendment 4 below the 60 percent threshold needed to pass and then the lawsuit, of course, becomes moot.” If it does pass, he said the lawsuit could continue, but, “my prediction is that eventually the lawsuit will get thrown out.”

It names as defendants Floridians Protecting Freedom, Inc. (Brenzel’s organization), the Department of State, Secretary of State, Elections Canvassing Commission and 21 county supervisors of election.

Brenzel’s take: “This is yet another example of a small group of extremist individuals who believe that they should speak for the majority.”

Contact Us