## Facts

- Joseph M. Spann, a corrections officer at the Salem County Jail, was accused of having intercourse with a female detainee held at the jail.
- Spann was charged with second-degree sexual assault under N.J.S.A. 2C:14-2(c)(3), which criminalizes sexual penetration by a person with supervisory or disciplinary power over a detainee; for this offense, intercourse itself constitutes the crime and issues of consent or force are irrelevant.
- The State contended the detainee became pregnant while incarcerated and that Spann was the father, making paternity evidence central to proving intercourse occurred.
- The State introduced blood and tissue testing (including HLA testing) and expert testimony stating there was a 96.55% “probability” Spann was the father.
- The expert reached 96.55% by combining test results with a Bayesian calculation that assumed a 50% prior probability of paternity.
- On cross-examination, the expert conceded the 50% prior was not based on case facts or record evidence, but was selected to “equally” treat the parties’ positions; the expert also described the percentage as meaning it was “very likely” Spann was the father.
- The jury convicted Spann; the Appellate Division reversed due to improper admission of the probability opinion.

## Issues

1. Whether it was error to admit an expert opinion assigning a 96.55% probability of paternity when the calculation depended on an assumed, non-evidence-based 50% prior probability.
2. If admission was error, whether the error was harmless in light of the other evidence.

## Decision

- The Supreme Court of New Jersey affirmed the Appellate Division.
- The Court held the 96.55% probability-of-paternity opinion was improperly admitted because it rested on an assumed prior probability not grounded in evidence.
- The Court held the error was not harmless because the numerical probability and its qualitative “very likely” characterization were highly prejudicial on the case’s central factual question.
- The conviction was reversed and the matter remanded for a new trial.

## Legal Principles

- Expert statistical probability testimony is inadmissible when it depends on a prior-probability assumption that is not supported by record evidence and is presented as “neutral” or “objective.”
- In a prosecution where proof of intercourse is the only material factual dispute, paternity-probability testimony can be outcome-determinative and carries a substantial risk of undue jury weight.
- Improper admission of a highly persuasive probability opinion on the case’s key fact is not harmless error when it may have influenced the verdict, even if other evidence of guilt is strong.

## Conclusion

The court ordered a new trial because the State’s expert paternity probability opinion incorporated an unsupported 50% prior assumption and, given the dispositive role and persuasive force of the resulting percentage and “very likely” characterization, its admission could not be treated as harmless.


State v. Spann, 617 A.2d 247 (N.J. 1993)