On appeal, the Tenth Circuit approved the admission at Merritt’s trial of three other DUI-related acts. Two of the other acts were prior convictions. The third was a drunk-driving arrest that occurred while Merritt was on bond in the federal case.
The Tenth Circuit reasoned that other acts tended to prove that Merritt had the requisite awareness of the serious risk of harm associated with drunk driving, rendering his decision to drive reckless and wanton, thereby supporting a finding of malice aforethought. The Tenth Circuit rejected Merritt’s argument that driving drunk does not evince an awareness of the riskiness but rather suggests the opposite.
In assessing the admissibility of the drunk driving incident while on bond, the Tenth Circuit considered the government’s argument that the similarity of that incident to the charged crime implicated the “doctrine of chances.” Under this doctrine, the similarity of the charged conduct to other conduct increases the likelihood that the incidents are not innocent random events. After seemingly signaling that the doctrine might apply, the Court decided not to decide, reasoning that any error in the admission of the evidence was harmless.
Applying the doctrine of chances to prove mens rea is fraught with the danger that the jury will misuse the other-crimes evidence. As with other-crimes evidence generally, the problem is one of dual relevance: the evidence supports the improper inference of bad character as well as the proper inference of intent. The improper chain of inference, prohibited by FRE 404(b), has two steps. Step One involves inferring from a past crime that the defendant has a bad character. Of course, this inference is itself empirically weak. Beyond that, the inference creates the risk that the jury might convict just to punish the defendant for her criminal past. Step Two involves inferring from this bad character that the defendant must be guilty. This creates the risk that the jury will overvalue the bad character evidence, failing to recognize that bad character is empirically a poor predictor of behavior on a particular occasion.
When used to prove the actus reus (as opposed to the mens rea) of a crime, the doctrine of chances does not necessarily implicate character concerns. A classic use of the doctrine is to prove that an initially unexplained death that is discovered to be similar to other deaths associated with the defendant was actually a homicide. The warrant for this inference is that common sense tells us that these similar deaths are very unlikely to represent mere coincidences. This use of the doctrine certainly implies the defendant’s bad character, but that implication is a side effect, not a necessary link in the inferential chain. But when the doctrine of chances is ostensibly used to prove mens rea, the close connection between a person’s intent and their character makes it unlikely that a lay jury could follow a limiting instruction—even crafting an intelligible instruction conveying this distinction would be difficult. If nothing else, Merritt at least flags the issue, allowing defense lawyers to anticipate and prepare to battle its use.