Appellate Review of Child Sexual Abuse Convictions: The Importance of Character Evidence

Character Evidence in the Law

Character evidence describes evidence in which either “character witnesses” or “other acts” are offered to prove that a person is or is not the kind of person who has the propensity to commit an act. Character witnesses testify to their opinion of a person’s propensity, or to a person’s reputation for having a propensity. For example, in a child sexual abuse case, a witness could testify that in their opinion the defendant does not have sexual interest in children. Other acts describe a person’s behavior other than the charged act. The prosecutor could offer other acts of sexual abuse by the defendant in order to prove aberrant sexual interest.

Character evidence is generally inadmissible in both state and Federal courts in the United States (Mueller and Kirkpatrick, 2021). This is because character evidence offered against a defendant can unfairly prejudice the jury in two ways. First, the jury may give the character evidence more weight than it deserves. Second, the jury may convict the defendant for their propensity, or for their other acts, rather than for the charged crime. This violates the basic principle in criminal law that defendants are on trial for what they did and not for the kind of person they are (Mueller and Kirkpatrick, 2021).

The rules of character evidence are complicated because there are a number of exceptions to the rule against admissibility, and because other acts may be admissible on the grounds that they prove something other than propensity. Furthermore, even if evidence has a non-propensity use, the court may exclude it because of the dangers that the jury will misuse it as propensity evidence. The U.S. Supreme Court called the common law rules regarding character evidence a “grotesque structure” that was “archaic, paradoxical, and full of compromises and compensations by which an irrational advantage to one side is offset by a poorly reasoned counter-privilege to the other” (Michelson v. U.S., 1948, p. 486). Nevertheless, the common law rules were for the most part codified in the Federal Rules of Evidence when they were enacted in 1975, and most states (including Arizona) have evidence codes that are similar to the Federal rules (Graham, 1990).

There are a number of standard exceptions to the rules against character evidence, and those most relevant to CSA prosecutions are noted here. One standard exception is that defendants in criminal cases can present character witnesses on their behalf. For example, in a CSA case, a witness could testify for the defense that in their opinion the defendant was not sexually interested in children, or that the defendant did not have a reputation as a molester. Furthermore, defendants are often allowed to offer character evidence against victims. If the defendant offers character evidence, however, the prosecutor can then rebut with their own character evidence. Examining final arguments in criminal trials of CSA in California, Stolzenberg and Lyon (2014) found that character evidence offered by the defense on the defendant’s behalf was presented in at least 20% of the cases, and the defendant offered character evidence against the victim in at least 14% of the cases.

Another standard exception is that both the defense and the prosecution can offer character evidence regarding the opponent’s witnesses, in which a witness will testify about another witness’ character for truthfulness (referring either to their opinion or to the reputation of the other witness). A “witness” includes anyone who is called to testify or whose hearsay statements are offered into evidence. Stolzenberg and Lyon (2014) found that the defendant offered character evidence against prosecution witnesses in at least 27% of the cases, and the prosecutor offered character evidence against defense witnesses in at least 10% of the cases.

In CSA cases, the most significant exception to character evidence is the prosecutor’s ability to introduce other acts of abuse to prove that the defendant has the propensity to sexually abuse children. This exception exists in the Federal Rules (Federal Rules of Evidence 414), 23 states (including Arizona and California), and the District of Columbia (Tchividjian, 2012). Stolzenberg and Lyon (2014) found that prosecutors offered character evidence against defendants in at least 29% of the cases.

As noted above, the sexual propensity exception is the source of much controversy. Speaking of the exception, Mueller and Kirkpatrick (2021) argued that it represents “a breach in a bedrock principle of evidence law that is closely connected to a version of our idea of due process.” Indeed, a few state supreme courts have struck down legislation enacting the exception (Lave & Orenstein, 2012).

Arizona has long had the sexual propensity exception, first under its common law (State v. McFarlin, 1973 [Hereafter Arizona cases are cited by last name of the defendant, hence McFarlin]) and, as of 1997, under the Arizona Rules of Evidence (Rule 404(c)). This type of proof is subject to heightened requirements. The trial court must make findings in the record that the defendant committed the prior act by clear and convincing evidence, that the prior act demonstrated a propensity to commit the charged act, and that the evidentiary value of the act was not substantially outweighed by the unfairly prejudicial effect it would likely have on the jury. In assessing these factors, the trial court considers the similarity of the other acts to the charged acts, their remoteness in time, and frequency.

Whereas many states and the Federal courts liberally allow other acts to be admissible against criminal defendants in order to demonstrate a propensity to commit sexual crimes, Rape Shield Laws protect alleged victims in sex crime cases from the use of their prior sexual history because of the likelihood it will prejudice the jury against the alleged victim, and in order to protect the privacy of victims and encourage them to come forward (Federal Rule of Evidence 412; Arizona Rule of Evidence 412). There are narrow exceptions to these rules, and defendants can argue that they have a due process right to introduce sexual history when it is not offered to prove propensity and is highly relevant in proving a defense argument.

An additional reason that the rules regarding character evidence are complicated is that other acts may be admissible to prove something other than propensity. First, if other acts directly prove an element of abuse, or if they occur at the same time as the abuse and facilitate the abuse, then they are considered “intrinsic” to the crime and not “other” acts (State v. Ferrero, 2012). For example, the prosecution could prove that the defendant physically restrained the child in order to sexually abuse them. Second, other acts can be offered to prove something other than propensity, such as “motive, opportunity, intent, preparation, plan, knowledge, identity, or absence of mistake or accident” (Federal Rule of Evidence 404(b); Arizona Rule of Evidence 404(b)). For example, the prosecution could offer evidence that the defendant had given the alleged victim marijuana in order to prove that the defendant groomed the victim; under the legal rules this would constitute evidence of a plan (State v. Garcia, 2001).

The ways in which other acts and opinion or reputation evidence might be admissible in CSA prosecutions is of obvious importance to attorneys, but it is also important for professionals who question children and other potential witnesses about child sexual abuse. Sexual abuse by the defendant in other jurisdictions, or at remote times, either against the primary victim or against other victims, is often admissible to prove propensity to abuse or for another purpose. Knowing that actions by the defendant that fall short of prosecutable sexual abuse might nevertheless be admissible in a sexual abuse prosecution encourages investigators to question witnesses about the defendant’s actions more broadly.

Character Evidence in Appeals of Criminal Convictions

Of course, if character evidence often leads to reversal on appeal, this is also of importance to attorneys and others who assess CSA. Anecdotal evidence suggests that character evidence is an important issue in criminal appeals across crimes in general. Imwinkelried (1990) noted that “[i]n many jurisdictions, alleged errors in the admission of uncharged misconduct evidence are the most common ground for appeal in criminal cases,” and citing an Arizona source, added that “[i]n some jurisdictions, errors in the introduction of uncharged misconduct are the most frequent basis for reversal in criminal cases” (p. 577; see also McClennen, 1990). The reference to “uncharged” refers to the fact that the other acts are not part of the current criminal charge. Similarly, Gottsfield (1997), an Arizona trial court judge, argued that “based on personal experience and an unscientific poll of some trial and appellate judges,” improper admission of other acts was the most common evidence issue leading to reversal (p. 24).

Only a few studies have systematically examined criminal appeals, and most do not discuss the types of evidence underlying the legal challenges (Buller, 2015; Foley, 1999; Heise et al., 2017). We have identified two studies that mention character evidence, but their results are difficult to interpret. Chapper and Hanson (1989) examined whether appellate courts found reversible error (other than sentencing errors), such that one or more of the criminal charges would have to be retried. Reversible error was found in 13% of the cases and the authors noted that “[e]rrors occur most frequently during the government’s direct examination and involve (in decreasing order of frequency) character testimony, opinion evidence (problems both with expert testimony and opinions proffered by nonexperts), evidence of uncharged acts and prior convictions, and hearsay in general” (p. 15). However, it was unclear whether the frequency of reversal for character testimony and uncharged acts was due to the number of times the issue was raised or the success rate of such challenges, or both. Waters et al. (2015) reported that “character testimony” was only the seventh most common issue on appeal and led to reversal in only 4% of the cases. However, they included claims (such as sentencing issues) that are unrelated to specific types of evidence. Furthermore, some of the issues that they classified as distinct from character evidence (e.g., pretrial evidence issues, prosecutorial misconduct) may have involved arguments about character evidence (e.g., the admissibility of other acts might have been addressed pretrial), such that character evidence issues may have been more common and challenged more successfully than the numbers suggest.

The Current Study

The current study examined appellate court opinions of CSA trials ending in conviction in Maricopa County, Arizona from 2005–2015. We identified, read, and coded 168 appellate opinions identifying legal issues raised by the defendants, the courts, and the state. We classified issues as involving specific types of evidence (our primary interest), not involving specific types of evidence, and sentencing issues (which did not contest the verdict). We assessed whether the appellate court identified an error and whether there was reversible error, such that one or more of the charges had to be retried or that the sentence should be changed. Because of the lack of prior research examining CSA appeals, we treated our analyses as exploratory and did not make predictions.

Sample

The current study examined appellate issues in 168 appeals filed by defendants following a conviction in a CSA trial. We obtained case information on all CSA cases that went to trial from January 2005 to August 2015 (n = 252) in Maricopa County, Arizona from the Maricopa County District Attorney’s Office. Maricopa County includes the city of Phoenix, is the largest county in Arizona, and is the fourth largest county in the United States. CSA cases included one or more charges of Sexual Conduct with a Minor (A.R.S.13–1405), Child Molestation (A.R.S. 13–1410), or Sex Abuse (A.R.S. 13–1404). Of the 252 cases, 184 resulted in the conviction of the defendant on at least one charge (73%). In two cases, the defendant was convicted only of non-CSA offenses (burglary in one case and kidnapping in another). In Arizona, as in most states, criminal defendants have a right to appeal convictions (Waters et al., 2015). We searched for appellate opinions reviewing these convictions on the Arizona Court of Appeals’ website, Westlaw, and Google search, identifying 168 appellate decisions (91% of convictions). We were unable to find appeals for 16 cases. In eight of these, court dockets suggested that defendants declined to appeal the conviction or accepted a plea and forfeited their right to an appeal. No information was available on the remaining eight cases.

Coding

Before coding began, and based on informal review of the appellate opinions, the research team identified and defined 36 unique issue types, including 15 types of evidence (e.g., character evidence; victim’s testimony; defendant admission or confession), 20 issues unrelated to specific evidence types (e.g., general assertions of the insufficiency of evidence, jury issues, ineffective assistance of counsel), and sentencing issues. We separately considered issues raised by the defendant, the courts, and the state, because we expected the frequency of issues raised by each entity, and their success, to vary. Defendants would raise as many issues in their favor as possible. The appellate courts, by contrast, would likely only raise issues on their own initiative when they aroused concern, which may lead to a higher rate of error. The state would raise few issues because they would be procedurally limited in appeal (e.g., the state cannot argue on post-conviction appeal that evidence in its favor was unfairly excluded).

Two research team coders independently and reliably (K >.80) coded the 168 appellate opinions to identify if any of the 36 issues were raised by the defendant, court, and/or state; in the case of evidentiary issues, the legal objection; and in all cases the appellate court’s response. Coding was an iterative process: each appeal decision was independently examined by both coders; coders then compared codes, clarified the coding guide, independently re-coded appeals, and compared again until reliability assessments revealed they agreed on codes 80% of the time. Then, all discrepancies were discussed and resolved by the coders agreeing on the final code. During this coding process we identified four issues that we had not anticipated—two related to types of evidence and two unrelated to types of evidence—and added them to the coding scheme. Character evidence and its subtypes are defined in Table 1, and the definition of the other issue types are in the online appendix (Supplemental Table 1).

OPEN IN VIEWER

Table 1.

Definition of Character Evidence and Subtypes.

Character evidence	Any evidence offered to prove that a person is the kind of person (or has the propensity) to commit an act; includes other acts evidence offered by the prosecution or the defense, expert testimony about character, and non-expert opinion/reputation testimony
Subtypes
Other acts offered by prosecution	Any act other than the charged offense, by any person, offered by the prosecution to demonstrate someone’s character; includes uncharged sexual acts, criminal behavior, or noncriminal behavior by defendant indicating a propensity for sexual misconduct.
Other acts offered by defense	Any act other than the charged offense, by any person, offered by the defense to demonstrate someone’s character; includes behavior by victim to demonstrate non-credible character.
Expert testimony about character	Expert testimony offering an opinion regarding someone’s propensity, including the defendant’s propensity to commit the charged act.
Opinion/Reputation	Testimony about the character of defendant, victim, or other witnesses by someone other than an expert, referring to the witness’ opinion of the other person’s character or knowledge of the other person’s reputation.

With respect to the appellate court’s response, we assessed whether the court concluded that an error had occurred, and whether the error was reversible. Errors would not be reversible (1) if they were harmless, meaning the appellate court concluded beyond a reasonable doubt that the error would not have influenced the jury’s verdict (State v. McVay, 1980); or (2) when the defendant waived the error by failing to object at trial, and the appellate court failed to find fundamental error, which requires that the error went to the heart of defendant’s case and rendered the trial unfair (State v. Valdez, 1989).

Searching Westlaw, we identified which of the appellate opinions were published, which essentially requires a majority of the appellate court panel to view the opinion as legally significant. Unpublished opinions are more common, and although they are publicly available, they usually cannot be cited by attorneys or other courts as precedent (though this was modified somewhat for cases after 2015). Indeed, technically speaking, they are not even “opinions” but “memorandum decisions” (Az. Rule of Supreme Court 111).

Overview of Issues Raised

Issues with Specific Types of Evidence

Issues not Tied to Specific Types of Evidence

Sentencing Issues

The defendant, court, and state raised issues about the imposed sentence in 35% (n = 50) of cases, raising 64 sentencing issues (Table 4). At first glance, sentencing issues appear highly successful, suggesting that the defendants obtained substantial relief. However, defendants in cases with successfully argued sentencing issues received an original average sentence of 146 years (SD = 209 yrs, Median = 72 years). In nine out of the 32 cases in which the courts found error, the court modified the sentencing minute entries (a record of the judgment available to the parties), which had no effect on the sentence length or fines. In another nine cases, the courts acknowledged errors in calculating presentence incarceration credit (also known as time served), which resulted in a mean sentence reduction of 190 days (SD = 548 days, Median = 3 days). In seven cases, defendants’ fines were reduced; six of these vacated the requirement to pay for DNA testing. In one case, the defendant’s sentence was increased. The appellate court agreed with the state that the sentencing judge incorrectly sentenced the defendant to four concurrent terms of 16 years when they were legally required to impose consecutive terms. As a result, the defendant’s sentence was increased by 48 years.

In six cases, the court vacated one or more of the defendant’s sentences, potentially reducing the sentence an average of 57 years (SD = 74 years, Median = 35 years). However, because these cases were remanded to the trial court for recalculation, the final actual reduction in sentence was unclear.

Character Evidence in the Published Cases

Ferrero involved application of the special rules for other acts by criminal defendants proving a propensity to abuse children (Az. Rule of Evidence 404(c)), which was also at issue in another of the four published cases (State v. Vega, 2011), as well as in several of the unpublished cases in which the appellate courts identified trial court errors (State v. Bucholtz, 2016; State v. Galvez, 2018). The reversal meant that the defendant would be freed from prison (but lifetime probation for a third count he had admitted) unless and until the state refiled charges. As noted above, Ferrero narrowed the admissibility of other acts by finding, contrary to prior cases, that the requirements of 404(c) applied to acts committed by the defendant against the alleged victim in the case (as also held in Vega), and applied to acts that had previously been considered “intrinsic” to the crime.

Another published opinion, State v. Rhodes (2008) involved the standard exception to the character evidence rules that defendants can introduce character witnesses to testify that the defendant does not have a propensity to molest children. The case involved repeated misreading of the character evidence rules by the trial judge, first in excluding the evidence at trial, and second in thinking that propensity was an “essential element” of the crime. Because the appellate court upheld the trial court’s decision to grant a new trial, the case constituted the only other case we examined in which the defendant was freed unless and until he was retried. The fourth published case was Montgomery v. Duncan (2011), yet another character evidence case. Here, the state successfully argued that the alleged victim’s prior sexual history should not be admitted.

Illusory Success for Defendants on Appeal

Of course, the published cases present an unbalanced view of appellate decision-making, because the great majority of cases are unpublished, and the published cases predominantly reflect the rare wins for defendants. The overall rates at which appellate courts identified errors were low to begin with, consistent with prior research examining appellate review (e.g., Heise et al., 2017; Waters et al., 2015), and closer examination revealed that reversible errors usually constituted small wins for the defendants. Summing across the cases in which non-evidence-based issues led to findings of reversible error, defendants obtained relief from only seven of 73 guilty counts. Similarly, although sentencing claims led to higher percentages of rulings favoring the defense, also consistent with prior research (e.g., Heise et al., 2017; Waters et al., 2015), substantial reductions in sentences (i.e., other than time served) occurred in less than 20% (6/32) of the cases finding error.

Even the character evidence victories appear less favorable to criminal defendants upon closer examination. It is unlikely that Ferrero and Rule 404(c) substantially reduced the number of cases in which prosecutors are able to admit other acts against defendants. In the cases at issue, the state sought to introduce other acts of abuse involving the same victim. State v. Vega (2011), which found harmless error by conducting its own 404(c) assessment, and holding that the evidence would have been admissible under 404(c), illustrates how the other acts can usually (1) be proven by the victim’s statements, satisfying the clear and convincing standard; (2) prove propensity precisely because they involve the same child victim; and (3) are not prejudicial as long as they are not more serious than the charged acts. With respect to Ferrero’s narrowing of what constitutes “intrinsic” acts, the Arizona Supreme Court itself noted that most, if not all, of the acts that had been admitted as intrinsic would henceforth be admissible to prove something other than character, such as the defendant’s plan.

Rhodes involved a straight-forward application of the rule that criminal defendants can offer character witnesses in their defense, and thus helped future defendants only to the extent that other trial courts were similarly ill-informed about settled law. Moreover, the appellate court upheld the trial court’s grant of a new trial, and thus offered no new relief. Had the trial court rejected the motion for a new trial, and let the defendant seek an appeal of the conviction, it is not obvious that he would have won. In State v. Dahnad (2009) the trial court similarly erred in refusing to allow the defendant to present a character witness on his behalf, and the appellate court held it was harmless error.

Limitations and Future Directions

We only examined one county (albeit one of the nation’s largest), and each of the 50 states (as well as the Federal government) has its own rules and its own politics. Because this is the first study to examine appeals of CSA cases, and one of the few to examine character evidence, definitive statements are premature. It is possible that character evidence fares less well on appeal in other jurisdictions, given Arizona’s liberal admissibility of propensity evidence against defendants accused of CSA. When state law does not allow other uncharged acts to prove propensity in sex crime cases, then prosecutors are likely to seek admission of the other acts under the provisions allowing non-propensity use of other acts (e.g., to prove motive, intent, or plan; Bryden and Park, 1994). This specific issue—admission of uncharged acts to prove something other than propensity—has generated the most commentary in the legal literature on character evidence (Reed, 2005). In the future, researchers can determine if CSA defendants in the 27 states that do not allow sexual propensity evidence are more likely to prevail.

We also limited our focus to character evidence, though defendants often raised other issues, such as the victim’s testimony or the victim’s out of court statements (i.e., hearsay). These issues will be of interest to future researchers, particularly given concerns regarding the accuracy of child witnesses. Furthermore, we limited our results to issues that were raised by the defendants in more than 10% of the cases, reflecting the heterogeneity and uniqueness of CSA appeals. In the future, researchers may want to compare the type, frequency, and success of issues raised in CSA appeals to appeals of other crimes.

Implications for Practice

Although character evidence led to challenges on appeal, it only very rarely led to any substantive changes in convictions. The appellate courts were not skeptical of character evidence, but corrected trial courts’ errors in interpreting highly complex rules. In general, had attorneys more clearly established the legal foundation for admissibility, most of the errors would have been avoided. Pre-trial motions could clearly identify both witnesses qualifying as character witnesses and other acts that prove propensity, or another relevant issue (such as plan), and point to the specific legal provision allowing admissibility.

Forensic interviewers may also benefit from understanding character evidence and when it may be admissible. For example, interviewers are often challenged to identify specific information about individual acts of abuse. It is particularly difficult for children to provide temporal information and enumerate events (Wandrey et al., 2012). Understanding that individual acts may either form the basis for a specific charge, which requires more information, or for admission as other acts supporting specific charges, which requires less detail, enables interviewers to know when it is most necessary to attempt to probe for specifics.

As a general rule, identifying the most serious act with specificity is most important, because it should form the basis for a charge. Other acts are most likely admissible when they are less serious than the charged act, because the jury is less likely to be prejudiced by less serious acts. If the other acts are as serious or more serious than the charged act, then the jury is more likely to improperly conclude that because the defendant committed the other acts, it does not matter if he committed the charged act.

Conclusion

Character evidence, which includes other acts offered to prove propensity to commit the charged act and character witnesses testifying to reputation or opinion, plays an important role in CSA prosecutions (Myers, 2021; Stolzenberg and Lyon, 2014), yet has received little attention in research. Character evidence has been posited to be an important issue in appellate decision-making, yet research has neglected to examine what happens to CSA cases on appeal. This study showed that in one large county, character evidence was an important issue in CSA appeals, and prominently featured in the highest profile appellate cases. However, it ultimately led to few decisions giving defendants substantial relief, consistent with a general tendency for the appellate courts to uphold virtually all CSA convictions, counts, and sentences.