Michigan's Proposed Law Admitting Prior Crimes in Child Sex Abuse Cases:
Why It Has Broader Implications for the Law of Evidence

By SHERRY F. COLB
Wednesday, Sep. 21, 2005

The Michigan state senate has passed, and the state's governor has promised to sign, a bill that would alter the rules of evidence in criminal trials. The bill provides that in criminal trials for sexual abuse of a minor, a defendant's prior convictions for similar acts would be admissible as evidence that the defendant is guilty of the offense charged in the current case.

The bill will likely soon become law in Michigan, and an increasing number of states have passed comparable statutes. These statutes are significant, for the decision to admit such evidence against a criminal defendant represents a sharp departure from the usual approach, an approach that reflects the concern that hearing about prior offenses could unfairly prejudice a jury against the accused.

The Michigan statute thus confronts us with a broader question about character evidence at trials: Should we reconsider the old assumption that juries cannot hear about a defendant's prior crimes without losing their way?

If we conclude upon reexamination that juries truly are unable to handle such evidence, then the law has a far bigger problem on its hands than the handling of cases that involve the sexual abuse of minors.

The Rule Against the Admission of Propensity Evidence

The Federal Rules of Evidence, as well as most states' rules, provide that in general, evidence of a person's character may not be offered at trial to prove that on the particular occasion giving rise to the litigation, a person behaved in a manner consistent with his overall character.

Consider the following example. John Doe (a patient) brings a lawsuit against Jane Smith (a medical doctor) alleging that Smith negligently failed to diagnose Doe's throat infection before it could spread to his lungs and cause serious complications. As it turns out, Dr. Smith was previously found liable for malpractice against a different patient -- Luke Roe -- for surgically removing a healthy portion of his lung, instead of the diseased section that she was supposed to remove. John Doe wishes to introduce evidence of this prior malpractice against Jane Smith in the current suit.

If permitted to do so, John Doe would offer the evidence of Jane Smith's prior negligence for the following set of inferences: Dr. Smith's negligence in treating Luke Roe tends to support the conclusion that Jane is a negligent doctor. If Jane is a negligent doctor, then her failure to diagnose John's throat infection can more readily be attributed to negligence than it could be if the same error were made by a doctor without a demonstrated propensity for negligence in the practice of medicine.

I suspect that these inferences will strike most readers as eminently reasonable. Indeed, if one (say, an insurance company) were trying to predict whether a doctor was likely to commit malpractice in the future, one would probably want to know whether the doctor had done so in the past.

Propensities really do seem to have some meaningful predictive power, and therefore, knowledge about a person's track record can be illuminating. If we accept the reality of propensities, moreover, it would appear no less useful to learn of them in a context in which one is attempting to ascertain what happened in the past, where the past is unknown to the fact-finder.

Yet evidence of prior negligence is generally not admissible in a negligence case to support the inferences outlined above: that the defendant is a negligent practitioner and that he or she was therefore more likely to have been negligent in this particular case.

The same principle holds true in the criminal context. A prosecutor cannot offer evidence that a criminal defendant previously committed armed robbery, to help prove that he is guilty of the armed robbery with which he stands accused at the current trial. This is true, again, notwithstanding the probability that a person who is prepared to commit armed robbery once, is more likely to commit the crime again in the future than a person who has never before committed armed robbery - as recidivism statistics bear out.

Stated differently, propensity evidence, in the criminal and civil contexts, seems relevant to the determination of what happened on a particular occasion. Yet it is not generally admissible. Why?

The Reason for the Propensity Rule

The rationale behind the general rule against admitting propensity evidence at trial is that juries will place far too much weight on such evidence. Propensities may be relevant, in other words, but not as relevant as juries will assume that they are. A jury will thus systematically overvalue propensity evidence.

Consider the case of the armed robber. A person who stands accused of armed robbery claims that he is innocent. The fact that he has committed armed robbery in the past, however, does make the innocence story slightly less likely to be true than if there were no prior history.

Still, there are lots of people who could have committed this robbery, other than the defendant, notwithstanding his prior bad acts. The likelihood of his innocence, then, is very great, even after we know that he has committed armed robbery before. Most armed robbers in the country, after all, are innocent of any one particular robbery charged, yet the fact that they each have a history of armed robbery does nothing to distinguish one from another.

Okay, the reader might say. That may well be true, but the jury understands this reality - it is simply part of how a jury evaluates evidence. For example, if the particular person who robbed a liquor store was wearing a Yankees shirt at the time, we would want the jury to know that the defendant on trial owned a Yankees shirt, even though that evidence alone does not distinguish him from a large group of other, innocent, people who also own Yankees shirts.

One piece of evidence does not need to do all of the work. It will either add up when considered as a whole, or, if it does not, then the jury will acquit the defendant. Therefore, the exclusion of prior bad acts evidence cannot rest entirely on the limited relevance of such evidence; it must rest instead on the perceived disparity between the actual value of the evidence, on the one hand, and the jury's probable perception of the value of the evidence, on the other.

The Fear that Juries Will Attribute Too Much Value to Propensity Evidence

The conventional view of jurors is that they are unable to comprehend that propensity evidence is not all that relevant. As a result, once they learn that a defendant has committed a crime before, they will be inclined to convict him this time, regardless of the strength or weakness of the other evidence in this case. By contrast, when they learn that a defendant owns clothing that matches the culprit's clothing, they will not have a similar reaction.

Perhaps this view of jurors reflects, in part, the idea that jurors might no longer care very much whether the defendant is innocent once they learn that he has acted in this way before. If the defendant is a killer or an armed robber, the jury might feel, then it is not such a terrible thing if he is convicted incorrectly of the crime charged -- he is guilty of something, after all.

It may be in part the fear of such thinking that drives the ordinarily-applicable rule against propensity evidence. But is this fear well-founded?

Trusting Juries In Cases Involving Sex Crimes Against Minors: Rationales

Michigan will probably soon have formalized its decision to trust juries to value propensity evidence appropriately, at least with respect to sexual criminal propensity in the context of sex crimes against minors.

This trust may reflect one of two views: The first view is that, in the case of sex crimes against children, the likelihood of recidivism is far greater than in the case of other sorts of crimes and therefore, placing great weight on propensity evidence -- which jurors are inclined to do anyway -- is now appropriate.

The second view is that the youth of their victims makes sex crimes against children especially difficult to prosecute, and propensity evidence -- while perhaps subject to distortion -- is necessary in such cases if anyone is ever to be held accountable for violating children.

On this second view, the risk of convicting some innocent defendants (who have a history of molesting children) is not as troubling as the risk of never convicting any guilty defendants, given that young witnesses are often incapable of providing useful testimony.

One could challenge these two views. Some have suggested, for example, that although recidivism may be high among child molesters, it is also high among people who commit robberies (though not among murderers). In view of this fact, they say, child molestation should not be singled out for distinct evidentiary treatment.

As to the second argument, our system seems generally to embrace the view that acquitting an innocent person is so important that ensuring such an acquittal is worth the release of ten (or a hundred, or more, depending on one's source) guilty people. If the introduction of character evidence is enough to convict a person (against whom the evidence is otherwise weak) of child molestation, then the risk of convicting the innocent becomes quite grave.

Michigan's decision to maintain its general rule against propensity evidence and lift it only in the case of child molestation trials, moreover, demonstrates that Michigan continues to distrust juries' evaluation of propensity evidence -- in most cases -- much as other states do. If that were not so, it would have lifted the ban on propensity evidence more generally.

Why the Jury Often Gets to Hear Propensity Evidence

But is such distrust consistent with the manner and circumstances in which propensity evidence often can make its way into a trial?

Though evidence of character is generally inadmissible at trial, it is not invariably so. Character evidence is admissible, for example, to attack witness credibility. Once a person testifies at a trial, the opponent may accordingly offer prior bad acts (including criminal convictions) that tend to show that the witness is dishonest.

The credibility exception comes up most frequently in the case of the criminal defendant who takes the witness stand. A criminal defendant has a constitutional right to testify on his own behalf, just as he has the right to refrain from testifying. Once a defendant does take the stand, however, his credibility becomes an issue for the jury, and he is therefore subject to evidentiary attempts to portray him as a liar.

Whether or not the jury believes the defendant's sworn statements will turn, in part, on whether they consider him to be an honest person. Evidence rules thus allow the jury to hear information about propensity that might assist in the determination of whether or not the witness is dishonest (and is therefore more likely to be lying in the proceeding before the jury).

One way to judge whether a witness is dishonest is to consider prior actions that suggest dishonesty. Evidence rules, moreover, tend to consider criminal convictions relevant to the issue of honesty and dishonesty. Therefore, if the criminal defendant on trial for armed robbery opts to testify, he may find the prosecutor educating the jury about his prior convictions for larceny. Only if he does not testify can he feel confident that the jury will remain ignorant of these prior convictions.

Once the defendant takes the stand, the prosecutor will argue -- perhaps plausibly -- that committing larceny is an act that tends to show that a person is dishonest, and therefore that the person is currently more likely than he would be in the absence of such a propensity, to be lying to the jury. If the prior conviction is admitted, as it often will be, the judge will tell the jury that it may only consider the conviction as proof bearing on the witness's credibility, not as proof on the question of whether he in fact committed the offense with which he is charged.

But think about what this means. A person is charged with robbery. The jury wants to know whether the person committed the robbery. The defendant swears that he is innocent. The prosecutor gives the jury evidence about a larceny (theft) of which the defendant was previously convicted. The judge then explains to the jury that it may not draw the following inference: the defendant committed larceny, so that makes it more likely than it would be absent the larceny conviction, that he committed the robbery he is now charged with committing.

The judge would tell the jury that it may, however, draw a different inference: the defendant committed larceny, so that makes it more likely that he would tell a lie while under oath -- which, in turn, makes his statement "I am innocent" less persuasive than it would otherwise be.

Therefore, we can draw on the propensity evidence in order to choose not to believe him when he says that he did not commit the robbery, but we cannot draw on the same evidence to find that, in fact, he did commit the robbery. Get it?

If you do not, you are in good company. It is difficult to imagine that a jury can - as the judge instructs it -- limit its consideration of prior crimes to their reflection on the defendant's likelihood of lying as a witness (to which prior crimes are only marginally relevant), but at the same time ignore the implications of a criminal history for whether he committed the offense with which he is charged (to which it is arguably more relevant).

The many defendants who fail to testify because they have a prior criminal record have voted with their feet on the question of whether juries can handle the limiting instruction at issue.

A More Logical Direction for the Law of Character Evidence

Many people find quite frustrating the rule that character evidence may not be offered to prove behavior in conformity with that character. It seems logical to them that a robber is more likely to commit robbery than a person who has never robbed. Nonetheless, we recognize that a prior bad act is not powerful evidence that out of the whole universe of suspects, the defendant -- and not someone else -- committed this particular crime.

Yet the law does not trust jurors to digest these fairly basic concepts. And then we turn around and ask for the impossible when the defendant-witness testifies on his own behalf - demanding that the jury make the fine distinction described above, between using past crimes to evaluate the credibility of a denial, and using them to assess guilt.

A better approach would be to admit character evidence in criminal cases but also explain our concerns to jurors -- or allow the attorneys to explain them. This would relieve the frustration of those who trust juries to handle relevant information and not overvalue it.

It would also relieve the pressure on criminal defendants (whether in child molestation cases or other types of cases) to stay off the witness stand if they have a prior record.

Defendants should feel free to testify, and ironically, if their character is fair game no matter what they do, then they will feel free to testify. Juries can then hear of prior convictions, hear testimony from the defendant, and decide the cases with more, rather than less, information relevant to their task.

Alternatively, if we truly believe that jurors are incapable of doing that, then we must seriously rethink the limiting instructions that permeate our law of evidence and that allow jurors to hear all sorts of prejudicial information that they are subsequently asked not to consider for its full effect.


Sherry F. Colb, a FindLaw columnist, is Professor and Frederick B. Lacey Scholar at Rutgers Law School in Newark. Her columns on criminal law and procedure, among other subjects, may be found in the archive of her work on this site.