New Florida Court Decision Potentially Erodes Presumption of Innocence in DNA Cases

September 5, 2025 Don Pumphrey, Jr. Criminal Defense, News & Announcements Social Share

In Florida, statistical evidence is often introduced in cases involving the presence of DNA. Typically, this comes in the form of DNA “matching,” which is often used to argue to a jury that the defendant’s DNA would not be in a particular place if they were not guilty of the charge they were accused of.

Examples of the use of DNA evidence in cases may include:

• B is accused of murder in New York, but lives in Florida. A statistician testifies that there is a 1 in 200 million chance that the DNA found at the scene would randomly match B’s. B is convicted.
• E is accused of rape. Semen recovered from the victim matches E’s DNA profile at all tested loci. The probability of a coincidental match is 1 in 10 million. 
• G is accused of burglary. A blood drop left on broken glass at the scene is tested and matches G’s DNA profile, with a 1 in 5 million random match probability. 
• L is accused of child sexual abuse. A child is born, and paternity testing shows a 99.999% probability that L is the father.

The last of these examples recently came up in a major Florida case, and involved the application of what is known as “Bayes’ Theorem.” Thomas v. State, 403 So.3d 251 (Fla. 4th DCA 2025). Bayes’ Theorem is a statistical method used to determine the odds that something is true based on calculating how likely (or unlikely) it would be to see the result randomly.

Bayes’ Theorem is a commonly relied-upon statistical method for determining the likelihood of an outcome. However, if left unchecked, the use of Bayes Theorem can lead to what is called the “prosecutor’s fallacy.” This occurs when the odds of a defendant’s guilt are overstated because of an alleged DNA match. 

As an example, imagine D is arrested for sexual battery and is a match to DNA found on the victim. D has no known ties to the victim. An expert testifies there is a 1 in 30 chance of a random match of the DNA found on the victim’s body to D. The prosecutor alleges this means there is only a 1 in 30 chance D is innocent (97 percent chance of guilt), and the jury convicts D.

But this is inaccurate. This is because 1 in 30 people, if randomly swabbed, would have also produced a match. If there were 300 people who did (or reasonably could have) interacted with the victim that day, 10 of them would return a DNA match. This would place the probability of D’s guilt at closer to 1 in 10, without additional evidence D was the assailant.

This brings us to the Thomas case. Thomas was charged with two counts of sexual battery against his minor stepdaughter. The primary evidence relied on by the State was that the baby born to the victim out of the alleged battery had 15 out of 15 DNA “loci” matching the defendant’s DNA. This is extremely improbable if the child was not Thomas’s (making him guilty of sexual battery).

However, the State went a step further. The State’s DNA expert testified there was a “probability of paternity of 99.9999996 percent” based on the results (1 in 287 million chance Thomas was not the father). This was calculated by applying Bayes’ Theorem in a controversial way that was objected to by Thomas’s defense counsel.

In determining the odds that Thomas was not the father (1 in 287 million), the State’s expert did not just focus on the percentage chance of 15 loci matches. A 1 in 287 million chance would mean that in a nation of 300 million people (America), one other person other than Thomas may match the DNA profile of the baby randomly.

Instead, the expert applied his version of Bayes’ Theorem, the formula for which requires a baseline assumption as to the percentage probability that Thomas was the father. As a defendant is presumed innocent, the typical assumed odds of guilt (fatherhood in this case) would be zero (or close). Assuming the defendant is potentially (or even likely) guilty from the start violates the presumption of innocence.

But in calculating the odds of paternity (and by extension, guilt), the State’s expert used a baseline probability of a 50 percent chance (0.5) that Thomas was the father of the baby. This led to the creation of the 1 in 287 million odds of innocence that the expert communicated to the jury.

The expert’s assumption that there was a 50 percent (0.5) chance Thomas was the father (as opposed to 10 percent (0.1), 5 percent (0.05), 1 percent (0.01), 0.1 percent (0.001), or 0) radically decreased the odds of Thomas’s innocence. 

If the odds of Thomas being the father were assumed at 10 percent, this would create a roughly 1 in 30 million chance of a random match using Bayes’ formula. If 1 percent odds were assumed, this would drop to 1 in 3 million. If 0.1 percent odds were assumed, the chance of a random match would be roughly 1 in 30,000. This means about 10,000 Americans (as opposed to just 1) could randomly match.

On appeal, Thomas’s counsel argued that although the DNA match (15 out of 15 loci) could be introduced, the use of Bayes’ Theorem in the manner relied upon by the State’s expert was inappropriate. This is because by assuming a 50 percent chance of Thomas’s guilt from the start, this violated the presumption of innocence.

However, the court (Florida’s 4th District Court of Appeal) ruled against Thomas, denying his appeal. The court reasoned that the 50 percent assumption was reasonable, as it did not err on the side of concluding Thomas was the father nor exclude it. The court reasoned: 

“The presumption of innocence does not require a jury to assume it was impossible for a defendant to commit the crime charged. Rather, it requires the jury to assume as a starting proposition that the defendant did not commit the crime, until proven otherwise. The probability of paternity … is merely a way of expressing and interpreting the actual DNA test results. Thus, the statistic itself does nothing to shift the burden of going ahead to the defendant.”

Though no court in Florida had squarely addressed the use of Bayes’ Theorem, Thomas noted that it was in conflict on this issue with the Connecticut Supreme Court’s 1994 Skipper decision. State v. Skipper, 228 Conn. 610, 637 A.2d 1101 (1994). There, the court ruled in a very similar situation that the use of a 50 percent starting probability of guilt in a Bayesian DNA analysis violates the presumption of innocence.

However, others have argued that a 50 percent starting assumption is permissible on the grounds that it is “neutral.” State v. Hartman, 145 Wis.2d 1, 426 N.W. 2d 320, 326 (1988)(Steinmetz, J., dissenting). Though the Hartman majority ruled that a 50-50 presumption violated Hartman’s presumption of innocence, a dissenting justice (Steinmetz) wrote a dissenting opinion that the Thomas court embraced.

So, what does this all mean for the presumption of innocence in Florida DNA cases? It is not entirely clear yet. However, Thomas v. State has the force of law. The Pardo rule states that a Florida District Court of Appeal’s (DCA) ruling on a matter of first impression binds all courts until another DCA conflicts with it or it is overruled by the Florida or U.S. Supreme Court.

As a result of this decision, an expert may be permitted to use (and potentially misuse) Bayes’ theorem to override the presumption of innocence of a defendant in a case involving DNA, by assuming there is a 50 percent chance that the defendant is guilty when assessing the odds of a random match.

By inferring a 50 percent probability of the defendant being the DNA’s source, this may skew the odds that a jury will find someone guilty based on DNA evidence. This seems to simply be a more scientifically complicated version of the “prosecutor’s fallacy” discussed earlier. But now, Florida courts must allow such expert testimony to be heard.

Though this is an issue that is not often addressed, it may be significant in DNA cases going forward. Will other District Courts of Appeal issue rulings that conflict with Thomas? Will the Florida Supreme Court overrule it? This remains to be seen, but it is worth keeping an eye on.

If someone is arrested and formally charged in Florida and concerned about whether they qualify as a family member or custodian for enhancement purposes, or are challenging DNA evidence, it is critical to find experienced and trusted legal representation as soon as possible. This decision could make the difference in whether or not someone faces a lengthy prison term and hefty fines.

Criminal Defense Attorney in Tallahassee, FL

Don Pumphrey, Jr. is a Former Prosecutor, Former State Police Officer, Lifetime Member of the Florida Association of Criminal Defense Lawyers; for over 25 years as a private defense attorney who is Trusted, Experienced, Aggressive in Criminal Defense as a Trial Attorney, Criminal Lawyer, Criminal Defense Lawyer for the accused in Florida State Courts located in Tallahassee, Florida but handling cases throughout the State of Florida.

Don Pumphrey, Jr. and the Tallahassee criminal defense lawyers at Pumphrey Law have decades of experience fighting drug charges on behalf of clients and winning. Call Pumphrey Law now at (850) 681-7777 to learn more about what we can do for you. Our lawyers will be happy to provide you with a free consultation.

  Social Share