If a veteran’s service record does not mention any reports that the veteran was sexually assaulted while in service, can the Department of Veterans Affairs (VA) use this absence of records as evidence that no sexual assault occurred? In AZ v. Shinseki, a recent decision by the Court of Appeals for the Federal Circuit, the court wrote that the VA may not make this inference; the absence of a contemporaneous report should have no value as evidence.
These consolidated cases arose when the veterans sought compensation for their post-traumatic stress disorder (PTSD), caused by having been sexually assaulted by fellow service members while they were in service. In two separate decisions addressing the veterans’ claims, the VA had weighed the absence of contemporaneous reports as evidence against the veterans’ claims, and the Court of Appeals for Veterans Claims accepted the VA’s factual findings.
The Federal Circuit appeals court found that these findings were improper. Judge Dyk, writing for the panel, examined the common law of hearsay. Although not binding law in administrative proceedings, hearsay rules are a useful guide to what evidence should be considered. The court found that an absence of an entry about the incidents of sexual assault was not admissible under traditional hearsay principles, because such reports are not reliably and regularly made; in fact, various studies (including the Defense Department’s SAPRO reports, which the court frequently referenced) have shown that at least 85% of these incidents go unreported within the chain of command.
The court went even further because the veterans in these cases admitted that they had not reported the incidents to their superiors. So the court was also faced with the question of whether an admitted failure to report has any value as evidence, regardless of hearsay rules. The court again ruled that this should not be considered as evidence against the veterans’ claims. Again turning to the common law, the court found that even the traditional (and somewhat outdated) common law principles regarding failure to report sexual assault were satisfied here; the common law looked to whether the victim told anyone about the incident, and in this case, the veterans had confided in other people, just not their commanding officers.
The court also noted, referencing DoD’s reports, that sexual assault between service members is a major problem for the military (as anyone who reads the news lately is aware.) Several unique factors in the military chill reports of such incidents, and this backdrop is relevant and inescapable.
Judge Moore dissented from the panel’s opinion. While she might have disagreed with the VA’s weighing of the evidence, she argued that the panel had no authority to decide that the evidence of non-reporting was totally irrelevant. Relevance, she wrote, is a pretty low threshold: a piece of evidence is relevant if it tends to make any material fact just slightly more or less probable than it would be otherwise. Even if 90% of sexual assaults go unreported in the military, under reasonable assumptions, the failure to report does have some very small evidentiary value.1 It’s a question of how the evidence should be weighed, not whether it should be admitted, in her view.
So Judge Moore is probably right, as a technical matter, under the general low threshold of relevance she described from Federal Rule of Evidence 401. But, to throw my two cents in, the general rules of evidence also have another principle that the panel did not note. After Rule 401 on relevance, there is Rule 403: Relevant evidence may be excluded if its value is substantially outweighed by the danger of unfair prejudice or confusion of the issues. On the facts of these cases, I think the VA demonstrated both unfair prejudice and some degree of confusion in attaching great weight to the failure to report; perhaps the court was too polite to say so. Given how rarely these incidents are reported, the absence of a report is of extremely little value, and if the VA can’t recognize that, they shouldn’t be allowed to consider it at all.
And from now on, as a result of this opinion, they won’t be.
- “Under reasonable assumptions” are my words, not hers: fans of Bayesian inference will note that, if there are any false reports, the evidentiary value of the records would be even smaller, or perhaps even negative. This is akin to the problem of false positives in medical tests: even a very small rate of false positives can mess up the evidentiary value of a test; it depends on a number of variables unknown here. Without going into equations, I don’t believe there would be a significant number of false reports in this context, but some people might hold this belief. Ironically, I think many of these people are also likely to attach undue weight to a victim’s failure to report, though their beliefs contradict each other as a matter of logic. But I digress! [↩]