Serology Evidence in Court: Interpretation and Testimony

Forensic serology proceeds through a hierarchy of inferential steps, each more discriminating than the last. The first confirms the biological nature of the stain: is it blood? The second confirms its origin: is it human? The third, historically performed by ABO and other blood grouping, places the donor into a class. The fourth, now standard in most jurisdictions, produces an individual DNA profile with a match probability.

1. Identification (is it blood?)
   Presumptive tests screen; confirmatory tests such as the Takayama or Teichmann crystal tests establish haem-derived material. Result: consistent with blood.
2. Species identification (is it human?)
   Anti-human antibody tests or now PCR-based species identification. Result: consistent with human blood.
3. Blood group typing (class level)
   ABO, Rh, MN, and secretor status. Result: consistent with a proportion of the population, for example 40% are type A.
4. DNA profiling (individual level)
   STR profiling at multiple loci produces a match probability typically expressed as 1 in billions or more. Result: with a given random match probability, the evidence strongly supports the inference that the donor is the named individual.

Even with a very strong DNA match, the inference does not end at identification. The serology tells you who the blood came from. It does not tell you whether the person was present during the crime, deposited the blood personally, or was a victim rather than a perpetrator. The activity-level question, what was happening when the blood was deposited, is not answered by serology and is often not addressed at all in the laboratory report. Distinguishing these levels of inference clearly is the foundation of defensible testimony.

When a serology result is not a definitive match or exclusion but falls somewhere in between, the expert must communicate that uncertainty without turning it into either a distorted positive or a meaningless negative. The Association of Forensic Science Providers (AFSP) verbal scale and similar frameworks in use by ENFSI members provide a structured vocabulary: 'provides strong support for', 'is consistent with', 'provides limited support for', 'neither supports nor contradicts'. Each phrase maps roughly onto a likelihood-ratio range.

In older ABO blood grouping testimony, where the result places the donor in a class covering 40% of the population, the evidential value is genuinely modest. A type A stain at a scene where the suspect is type A is consistent with the suspect's presence but equally consistent with the presence of 40% of the general population. The expert should not overstate this as placing the suspect at the scene; they should state that the blood type is consistent with the suspect's and provide the population frequency so the trier of fact can weigh it appropriately.

The prosecutor's fallacy is a confusion of conditional probabilities. When a laboratory reports that a blood type occurs in 1 in 10 people and the defendant matches, the fallacy is to interpret this as meaning there is only a 1-in-10 chance the defendant is innocent. The 1-in-10 figure is the probability of seeing that blood type in a randomly selected innocent person, not the probability of innocence given the evidence. These are different quantities; treating them as equivalent inverts the logic of probabilistic inference.

The correct framing is the likelihood ratio. The expert compares two propositions: the evidence if the suspect is the source, versus the evidence if a randomly selected person from the relevant population is the source. In the blood-type example, if the type occurs in 1 in 10 people, the likelihood ratio is 10: the evidence is 10 times more probable if the suspect is the source than if a random person is. This says nothing about prior probabilities of guilt; it provides the jury with a multiplier to apply to whatever prior belief they formed from other evidence.

The defender's fallacy runs in the opposite direction. If a blood type occurs in 1 in 10 people and the adult population is 60 million, approximately 6 million individuals share that type. A defence argument that the match is therefore meaningless because 6 million people could be the source commits the defender's fallacy. It treats the match as if the suspect were drawn at random from the entire adult population, ignoring the prior investigation that narrowed the suspect pool, the location, time, and circumstances of the crime, and any other evidence already in the case.

An expert witness has a duty to resist pressure from both sides to distort their evidence. If the prosecution attempts to overstate a weak match, the expert must correct this on direct examination and in their report. If the defence attempts to characterise a strong match as meaningless, the expert must explain the evidential weight accurately under cross-examination. The expert is not an advocate; they are an adviser to the court on matters within their expertise.

In the United States, Daubert v Merrell Dow Pharmaceuticals Inc (1993) replaced Frye as the federal standard and is now applied in most states. Under Daubert, the trial judge acts as a gatekeeper and assesses four factors: whether the theory or technique can be and has been tested; whether it has been subjected to peer review; whether it has a known or potential error rate; and whether it is generally accepted in the relevant scientific community. All four are relevant, but none is individually determinative.

Conventional serology methods, ABO blood grouping, species identification by immunoprecipitation, and DNA STR profiling, are well past any Daubert challenge at this point. They have been validated, peer-reviewed, and used extensively for decades. A Daubert challenge is more likely to target a novel method, such as a new probabilistic genotyping software or an untested variant of a stain test, than a well-established serological technique.

The O.J. Simpson trial (1995) is the most widely analysed case involving contested serology and DNA evidence. The defence successfully raised doubts about chain-of-custody procedures, contamination risks in the Los Angeles Police Department laboratory, and the handling of samples. The case prompted widespread reform of forensic quality standards in US crime laboratories. The scientific results, which showed DNA consistent with Simpson at the scene and the victims' DNA at his residence, were technically sound; the challenge was to the reliability of the laboratory procedures, not the chemistry itself.

In England and Wales, the case of R v Deen (1994) was among the first to result in a successful appeal on the grounds of misuse of DNA statistics. The Court of Appeal found that the expert had fallen into the prosecutor's fallacy by equating the random match probability with the probability of the defendant's innocence. The Court laid down guidance on how DNA statistics should be presented to a jury, guidance that shaped subsequent practice.

More recently, challenges to probabilistic genotyping software in mixed-profile DNA cases, such as concerns raised about the TrueAllele and STRmix programs in US proceedings, illustrate that the contested frontier has moved from basic serology to the interpretation of complex DNA mixtures. The underlying serology remains well-established; the probabilistic modelling layer is where active scientific debate continues.

A serology expert appearing in court operates under a duty that overrides their relationship with the instructing party. In England and Wales, CPR Part 35 states that the expert's primary obligation is to help the court, not to advance the case of the party that instructed them. The US Federal Rules of Evidence, Rule 702, require the expert's testimony to be based on sufficient facts, a reliable method, and reliably applied principles. In both jurisdictions, an expert who departs from their honest opinion under pressure commits a professional breach.

• Know your limitations: be explicit in your report about what the evidence does not address. A serology report that says nothing about what it cannot establish is incomplete.
• Prepare for the fallacy challenge: anticipate that cross-examination will probe whether you committed the prosecutor's fallacy and have a clear explanation ready of what a match probability does and does not mean.
• Address alternative propositions: structure your opinion around competing hypotheses, not just the prosecution's case, so that you can demonstrate you have considered both.
• Distinguish source from activity: be ready to explain why a source-level finding (this is the suspect's blood) does not automatically answer the activity-level question (how and when it got there).

Cross-examination in major serology cases often targets four areas: the reliability of the analytical method, the handling and documentation of the sample, the statistical interpretation of the match, and the expert's impartiality. Preparation means having the case notes, the chain-of-custody records, the method validation data, and the statistical workings available and understood thoroughly enough to explain them clearly under adversarial questioning.