District Court of Appeal of Florida, Second District.
*698 Larry Hoffman, Clearwater, for Appellant.
Robert Butterworth, Attorney General, Tallahassee, and Patricia E. Davenport, Assistant Attorney General, Tampa, for Appellee.
PER CURIAM.
William Farrill has appealed from his convictions for two counts of capital sexual battery in violation of section 794.011(2), Florida Statutes (1997). His sole contention is that the trial court's erroneous admission of collateral sex crime evidence pursuant to Williams v. State, 110 So. 2d 654 (Fla.1959), amounted to reversible error. The charged and collateral offenses were similar, but not strikingly so, and they did not possess unique characteristics to distinguish them from other similar sexual batteries on children. The evidence should not have been admitted. Moreover, after an examination of the entire record, we have concluded that the error affected the verdict, and we reverse.
Mr. Farrill was charged with two counts of sexual battery on M.W., one involving union of his mouth with her vagina and the other involving union of his finger with her vagina. He was also charged in the original information with sexual battery on C.R. by union of his mouth with her vagina. Prior to trial, however, the judge severed count 3 on the basis that the charges involving the two girls were inconsistent and involved inconsistent defenses. The State then gave notice of its intent to use evidence of the incident involving C.R. in the prosecution of the M.W. case. The defense attorney filed a motion to strike the notice; after a hearing, the judge ruled in favor of the State to allow the admission of the evidence.
At the hearing on the motion to strike the notice of intent to use similar fact evidence, the judge began by saying, "I don't think this is a familial custody issue. I think the State is gonna argue that. I don't think there's a reason they would be using Williams' Rule other than-what turns out to be the prohibitive purpose for propensity of the bad acts." The judge then heard argument only of the assistant state attorney and defense counsel. At the conclusion, without any explanation of his decision, the judge ruled that the evidence could be admitted.
This perfunctory procedure became important when the trial ensued, because a new judge presided over the trial. At every appropriate juncture in the trial the defense attorney objected to admission of the similar fact evidence, and after the defendant was convicted he did so in his motion for new trial. Yet the trial judge, admittedly not having had the benefit of the transcript of that hearing, gave complete deference to the ruling of his predecessor, even though the initial decision was based only on argument of the lawyers and not on a proffer of the evidence through the witnesses themselves. Furthermore, even though the pretrial judge indicated no foundation for his ruling, the attorneys at trial acted with the assumption that the pretrial judge had ruled that both girls were in a familial or custodial relationship with Mr. Farrill, a conclusion that the trial judge rejected in the hearing on the motion for new trial. The successor judge nevertheless denied that motion.
The trial testimony revealed that Mr. Farrill became involved with M.W. because he was a friend of her parents. Mr. Farrill, who had known M.W.'s parents for some years before these incidents were *699 alleged to have happened, would from time to time stay with the Farrill family. M.W. testified that Mr. Farrill treated her "like an uncle" and that she "grew up with him." Her parents would occasionally leave M.W. and her brother in Mr. Farrill's charge when they needed to run to the store. Later on, when M.W.'s mother separated from her father, the mother would occasionally ask Mr. Farrill to babysit the children. Based upon these facts, the trial judge ruled at the hearing on the motion for new trial that the State had established a familial or custodial relationship between M.W. and Mr. Farrill.
The situation with C.R. was quite different, however. C.R. was a next door neighbor of Mr. Farrill's at a time when he was not staying at M.W.'s home. Instead, his household consisted of himself, a young daughter, and the mother of his child. C.R., who was approximately 9 years old when the Williams rule incident occurred, spent a lot of time in the Farrill home, but almost exclusively with the mother and daughter and not Mr. Farrill. Occasionally she would babysit for the little girl. Absolutely no testimony was adduced from which one could conclude that a familial or custodial relationship existed between C.R. and Mr. Farrill, which occurs when the "relationship [is] one in which there is a recognizable bond of trust with the defendant, similar to the bond that develops between a child and her grandfather, uncle, or guardian." State v. Rawls, 649 So. 2d 1350, 1353 (Fla.1994).
The relationship between defendant and victim has become preeminently significant in the recent analyses of the admissibility of Williams rule evidence in cases of sexual battery on children. In sexual battery cases, as in any criminal proceeding, admission of collateral crime evidence is highly prejudicial to the defendant, and the court must be alert to avoid the risk that the jury will convict a defendant based upon propensity alone rather than upon proof that he committed the charged crimes. See Heuring v. State, 513 So. 2d 122 (Fla.1987). Traditionally, as the Heuring court noted, the similar fact evidence must meet a "strict standard of relevance. The charged and collateral offenses must be not only strikingly similar, but they must also share some unique characteristic or combination of characteristics which sets them apart from other offenses." 513 So.2d at 124. Generally, as set out in section 90.404(2)(a), Florida Statutes (1997), the evidence must be relevant to a material fact such as identity, intent, motive, opportunity, plan, knowledge, or absence of mistake or accident. As the Heuring court noted, identity is rarely an issue when a child is sexually assaulted or battered by a family member or other person who stands in loco parentis to the victim, but courts have allowed collateral evidence to show "modus operandi, scheme, plan, or design, even though the distinction between sexual design and sexual disposition is often tenuous." 513 So.2d at 124. In those instances, the similar fact evidence is simply relevant for corroboration, and its probative value is held to be greater than the prejudicial effect to the defendant. Id. at 125.
The Florida Supreme Court further refined the standards for admissibility of collateral crime evidence in the familial relationship arena in Saffor v. State, 660 So. 2d 668, 672 (Fla.1995), finding that the familial context provides one marker of significant similarity between the charged and collateral offenses but that further scrutiny of the parallels is always required:
The additional showing of similarity will vary depending on the facts of the case and must be determined on a case-by-case basis. Thus, we do not eliminate the requirement of similarity which undergirds the Williams rule. However, the strict similarity in the nature of the offenses and the circumstances surrounding their commission which would be required in cases occurring outside the familial context is relaxed by virtue of the evidence proving that both crimes were committed in the familial context.
*700 Because the incident with C.R. was not committed in a familial or custodial setting, the State was not entitled to the benefit of this relaxed standard in this case. A comparison of the charged and collateral offenses, furthermore, reveals that the circumstances surrounding them were not strikingly similar.
M.W. was 16 when she testified at trial about events that had occurred eight to ten years earlier. The first incident allegedly occurred when she was about 6 years old. Mr. Farrill came into her room when she was sleeping, took off her clothes, and put his mouth and tongue on her privates. Some time later, M.W.'s family moved to another house where Mr. Farrill was also a frequent visitor. She described his practices as follows: "In my bedroom at night when everyone else was asleep he would come in and undress himself and me and spread my legs and perform oral sex, and then sometimes he would, like a little after, he would start to put his fingers in my vagina." Mr. Farrill reportedly did this "whenever he had the chance," or "most of the time when I was alone with him." By promising her gifts such as dolls and cookies, Mr. Farrill persuaded M.W. not to tell anyone.
On cross examination she described one specific incident in more detail. One night after a 4th of July party, when M.W. was about 9, she fell asleep in a big bed with her 6-year old brother and a neighbor boy of about the same age. Mr. Farrill then came into the room and performed oral sex on her in the bed with the other two children present.
C.R. was 17 when she testified to an incident that occurred when she was 9. She went next door to babysit for Mr. Farrill's daughter and found Mr. Farrill and his girlfriend discussing what it meant to be a virgin. Because C.R. was not familiar with the meaning, Mr. Farrill offered to explain it. After the child's mother had left the home, Mr. Farrill took C.R. to a nearby Circle K convenience store to buy a magazine to use as a visual aid. C.R. was under the impression that Mr. Farrill paid $12 or $20 for the magazine, and she described at least one of the pictures in great detail. When the two returned to the Farrill home and were looking at the magazine, C.R. questioned why the models had "hair down there" but she did not. Mr. Farrill acted surprised to hear that she did not have pubic hair, so he offered her $20 to show him. C.R. then went into the child's room, removed her shorts, and sat on the small bed. Mr. Farrill came in and put his hands on her knees. At first C.R. thought that Mr. Farrill was just going to look, but then he "leaned down and, like, licked it." C.R. then got up, put her clothes on, left the house, and never returned again.
To determine whether the charged and collateral offenses are strikingly similar, our court has suggested in Moore v. State, 659 So. 2d 414 (Fla. 2d DCA 1995), that it is appropriate to compare and contrast the victims, the alleged acts of abuse, the events, and factors affecting reliability-in that order-to determine if the similarity standards have been met. In this case, the primary similarities are that both victims were young girls between the ages of six and nine and that the offense in both was oral/vaginal union. As explained in Moore, "[a]lthough sexual battery on an underage child is a reprehensible offense, it is not so unique in itself that it should be uniformly admissible under section 90.404(2)," 659 So.2d at 414 (quoting Feller v. State, 637 So. 2d 911, 916 (Fla.1994)). The Florida Supreme Court in Feller had reversed the judgment because those were the only two common elements in the charged and collateral offenses. In this case, however, both offenses occurred during the same basic time frame.
The common elements between the charged and collateral offenses in this case would indicate only that Mr. Farrill had a propensity to molest young girls some eight to ten years prior to trial. The other circumstances of the events do not bear uniquely factual characteristics, however, *701 and differ in several significant ways. M.W. was allegedly abused at least ten times over a three year period, whenever Mr. Farrill was alone with her. All of the incidents occurred at night when she had been in bed, asleep, even the situation with the two boys in the same bed. C.R.'s testimony suggested that her encounter with Mr. Farrill occurred in the daytime. She described an elaborate plan by Mr. Farrill to induce her into a vulnerable situation. For M.W., however, Mr. Farrill's modus operandi was simply to enter her room and molest her. M.W. alleged that Mr. Farrill undressed himself and her, whereas C.R. described removing her own shorts while Mr. Farrill apparently remained clothed. C.R. also described a specific gift of money, which was really a bribe, to procure her cooperation. In contrast, M.W. stated that Mr. Farrill gave her cookies and dolls to insure her silence. Although M.W. apparently did not fully cooperate with Mr. Farrill, she did not describe a struggle or any rejection of his advances. C.R., however, left the Farrill home and never returned after the abuse occurred.
The record before us reflects many more differences than similarities between the charged and collateral offenses. Consequently, the trial court should not have admitted C.R.'s testimony in M.W.'s trial. See Saffor; Corpus v. State, 718 So. 2d 1266 (Fla. 2d DCA 1998); Sullivan v. State, 713 So. 2d 1023 (Fla. 2d DCA 1998); Gutierrez v. State, 705 So. 2d 660 (Fla. 2d DCA 1998); Moore.
Finally, we note that the assistant state attorney emphasized the propensity aspect of the Williams rule testimony in closing arguments with such comments as, "Where there's smoke, there's fire," and "Lightning doesn't strike twice." The suggestion of propensity was particularly bolstered by the vivid detail of C.R.'s testimony in contrast to the less specific descriptions of the events given by M.W. At least from a cold record, C.R.'s testimony emerges as convincing and memorable, and it undoubtedly influenced the jury. Accordingly, we cannot say that the erroneous admission of this Williams rule evidence did not affect the verdict. See State v. DiGuilio, 491 So. 2d 1129 (Fla.1986). The error cannot be considered harmless, and we reverse and remand for a new trial.
GREEN and CASANUEVA, JJ., Concur.
ALTENBERND, A.C.J., Concurs specially.
ALTENBERND, Acting Chief Judge, Concurring.
I fully concur in this opinion. I write only to express, or perhaps confess, an increasing concern that our Williams rule analysis in capital sexual battery cases may have little scientific validity or reliability.[1] If judges, as experts in the science of jurisprudence, were required to justify their selection of factors under a Frye analysis, I doubt that our opinions would withstand that scrutiny. See Frye v. United States, 293 F. 1013 (D.C.Cir.1923); Brim v. State, 695 So. 2d 268 (Fla.1997) (stating Florida still follows Frye standard despite Supreme Court's adoption of more lenient standard in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993)). Especially when identity is not the issue authorizing the admissibility of the Williams rule evidence and the evidence is introduced to establish one of the other issues under section 90.404(2)(a), Florida Statutes (1999), I frankly doubt that I have *702 the expertise to determine what prior specific behavior is validly predictive of subsequent sexual misconduct.[2] No matter whether we justify the introduction of this evidence as "corroborative" of a witness's testimony, or relevant to "motive," "intent," or "plan," I think we ought to admit that it is actually introduced, or at least relied upon by jurors, because it is prior behavior that we believe, with or without scientific basis, to be validly predictive of subsequent sexual misconduct by the defendant.
My concern goes beyond the lack of scientific validity underpinning our judicial tests of sexual misconduct; I also suspect that they lack reliability. In cases such as this one, where the Williams rule evidence consists of other child victims' corroborative testimony, the case law requires substantial factual similarity between the prior instances of sexual misconduct and the crime charged before the Williams rule evidence is admissible. See Saffor v. State, 660 So.2d at 669 (Fla.1995); Heuring v. State, 513 So. 2d 122 (Fla.1987). The outcome of this comparison depends significantly upon which factors the court decides to analyze. See Charles W. Ehrhardt, Florida Evidence, § 404.18, at 207 (1999). Compare Shipman v. State, 668 So. 2d 313 (Fla. 4th DCA 1996) with Moore v. State, 659 So. 2d 414 (Fla. 2d DCA 1995). That is to say, a court might find factual similarity if it compared the age, gender, and location of the attacks, but not find similarity if it compared, for example, the time of day of the attacks, the method of the attacks, or the body parts involved in the attacks.
In other criminal cases, it is most common for Williams rule evidence to be used to identify the perpetrator. Thus, if a defendant is known to have committed three other robberies of drugstores in the last month-all occurring in the evening while armed with a chrome revolver and while covering his face with a ski mask-intuitively it seems probable that another such robbery of a drugstore is also his handiwork. In such a case, the fact that crimes are so factually similar suggests that the perpetrator had a particular modus operandi, and we permit the introduction of evidence establishing such prior crimes because it is relevant to prove identity. In other words, we permit this evidence because, as a matter of experience and psychology, it is unlikely that someone else committed the crime in the same manner. This is why such similar evidence is frequently described as "fingerprint" evidence. See State v. Savino, 567 So. 2d 892 (Fla.1990); State v. Smith, 586 So. 2d 1237 (Fla. 2d DCA 1991).
In this case, identity is not an issue. Significantly, Mr. Farrill does not admit his alleged behavior with C.R., and no prior jury has determined the accuracy of those allegations. Assuming that the corroborative testimony of M.W. and C.R. is true, their testimony could lead a jury to believe that M.W.'s testimony concerning the charged offenses is also truthful. This is so, in part, because it is less likely that M.W. is fabricating her story if C.R. has a similar story. See Richard J. Sanders, "A Dangerous Bend in an Ancient Road" The Use of Similar Fact Evidence for Corroboration, 74 Fla. B.J. 40, 45-46 (Feb.2000). But is it "less likely" only because humans trust propensity logic?
I personally will not deny the frailty of my own humanity; I also reach this conclusion, in part, because I intuitively believe this "corroborative" testimony is probative or relevant on the ultimate material issue of whether Mr. Farrill committed the illegal conduct. Intuitively, I suspect that the psychological make-up of sexual predators leads them to prefer children of a certain age and gender over others and to *703 chose a preferred method of attack. However, I have no scientific evidence to support or reject my intuitive belief. I am convinced that the typical juror is also likely to believe that a person who would perform one sexual assault on a child is likely to perform another.
I recognize that my intuition in this respect is in essence nothing more than propensity logic. Our human willingness to use propensity logic is what makes Williams rule evidence both useful and dangerous.[3] It is possible that it is appropriate, scientifically speaking, to draw a valid conclusion concerning alleged sexual misconduct by a particular person based upon such propensity logic. See David M. De La Paz, Sacrificing the Whole Truth: Florida's Deteriorating Admissibility of Similar Fact Evidence in Cases of Child Sexual Abuse, 15 N.Y.L. Sch. J. Hum. Rts. 449 (Spring 1999). It is also likely many courts are drawing conclusions based, in part, upon this very type of propensity logic, whether explicitly stated or not. See David J. Kaloyanides, The Depraved Sexual Instinct Theory: An Example of the Propensity for Aberrant Application of Federal Rule of Evidence 404(b), 25 Loy. L.A. L.Rev. 1297 (1992); John McCorvey, Corroboration or Propensity? An Empty Distinction in the Admissibility of Similar Fact Evidence, 18 Stetson L.Rev. 171 (Fall 1988). This is all the more reason why we ought to carefully examine the scientific reliability and validity of the factors we consider when deciding whether to admit this potentially highly prejudicial evidence.
Other victim's corroborative evidence should have a proper, but limited role, in criminal cases. This is especially true in capital sexual battery cases when the defendant challenges a youthful victim's credibility or competency. Nevertheless, I remain concerned that the factors we consider to determine admissibility under Heuring and Saffor may not actually measure whatever it is we think we should measure.
[1] I use the terms "validity" and "reliability" in the context of the scientific method. See John Monahan and Laurens Walker, Social Science in Law: Case and Materials 41-43 (3d ed.1994). A scientist wants to know whether his test measures that which he seeks to measure. To be "valid" as a matter of social science, a test for sexual predators must identify and measure conduct that accurately predicts incidents of sexual predatory behavior. To be "reliable," the test must reach the same outcome upon repetition by different testers.
[2] I make this confession despite my own effort to create a list of relevant factors in Moore v. State, 659 So. 2d 414 (Fla. 2d DCA 1995). After five years of reflection, I agree with Professor Ehrhardt that the district courts' efforts to identify relevant factors in this analysis have added little to the existing case-by-case method. See Charles W. Ehrhardt, Florida Evidence, § 404.18, at 207 (1999).
[3] My neighbor's dog bit me once. I avoid that dog. No one thinks my use of propensity logic in the matter of the dog is inappropriate.
