State v. Lopez

This opinion is subject to revision before final publication in the Pacific Reporter 2018 UT 5 IN THE SUPREME COURT OF THE STATE OF UTAH STATE OF UTAH, Appellee, v. KOMASQUIN LOPEZ, Appellant. No. 20151094 Filed February 9, 2018 On Direct Appeal Third District, Salt Lake The Honorable Judge Paul Parker No. 141900304 Attorneys: Sean D. Reyes, Att’y Gen., John J. Nielsen, Asst. Solic. Gen., Salt Lake City, for appellee Teresa L. Welch, Andrea Garland, Nick A. Falcone, Salt Lake City, for appellant JUSTICE PEARCE authored the opinion of the Court, in which CHIEF JUSTICE DURRANT, ASSOCIATE CHIEF JUSTICE LEE, JUSTICE HIMONAS, and JUDGE DIREDA joined. Having recused herself, JUSTICE DURHAM did not participate; DISTRICT JUDGE MICHAEL D. DIREDA sat. JUSTICE PETERSEN became a member of the Court on November 17, 2017, after oral argument in this matter, and accordingly did not participate. JUSTICE PEARCE, opinion of the Court: INTRODUCTION ¶ 1 On the day she died, Shannon Lopez picked her husband Komasquin Lopez up from work. According to Lopez, he drove STATE v. LOPEZ Opinion of the Court home while he and Shannon argued in the cab of his truck. Lopez claimed that the argument continued until Shannon shot herself with a gun Lopez had in his truck. A jury disbelieved Lopez and convicted him of murder. On appeal, Lopez argues the district court erred in two ways. First, Lopez challenges the admission of expert testimony that assessed Shannon’s risk of suicide. Second, Lopez contends that the district court erred by admitting evidence that he had, on one occasion, pointed a gun at Shannon’s head, and that on another, he had leveled a gun at an ex-wife and threatened to kill her. Lopez also argues that the errors were harmful both individually and cumulatively, and that insufficient evidence existed to convict him. ¶ 2 We conclude that the State did not lay a sufficient foundation to demonstrate that the theory its expert employed could be reliably used to assess the suicide risk of someone who had died. We also conclude that the district court erred by admitting the evidence of Lopez’s prior actions. The errors were harmful. We reverse. BACKGROUND ¶ 3 Lopez and Shannon, who were married at the time of Shannon’s death, both enjoyed shooting guns. Lopez used a gun throughout his career in the military and law enforcement. Shannon, who had been introduced to firearms at the age of nine, was a recreational shooter. Lopez and Shannon kept multiple guns in their home. Lopez usually carried a gun in his truck and another on his person. ¶ 4 On the night of Shannon’s death, Shannon picked Lopez up from work. Shannon had consumed methamphetamine in a quantity that the medical examiner described as “toxic.” Lopez also had methamphetamine in his system. During their commute, Lopez and Shannon argued about Shannon’s methamphetamine use and their financial problems. Lopez said during a police interview “that Shannon’s last words were . . . that she would take the kids and go to her father’s[.]” He “repeat[ed] . . . several times during [one of the] interview[s]” that “[s]he said she’d take the kids, she’s already packed, and she’ll leave . . . .” He further stated that he “told her . . . during the argument he was going to leave her also.” During his trial testimony, Lopez maintained that he said he would leave her, but denied hearing Shannon say that she would leave him. ¶ 5 Lopez testified that as he was making a left hand turn, he heard the sound of breaking glass. Lopez turned to see that Shannon 2 Cite as: 2018 UT 5 Opinion of the Court was “slumped forward.” Lopez tried to turn the truck around to take Shannon to a hospital, but crashed into another car and then into a fence. Witnesses saw Lopez jump out of the truck and lie on the ground while saying “sorry mommy” or “sorry mama” repeatedly. ¶ 6 Shannon had been shot in her left ear. When police arrived, they found that Shannon’s legs were crossed at the ankles. Shannon’s right hand—Shannon was right handed—was hidden in her jacket sleeve. Officers found a gun and holster on the floor of the driver’s side of the cab. 1 ¶ 7 Months before she died, Shannon sent a text message to Lopez expressing a desire to end her life. Shannon had spoken to her son, M.N., about suicide in the past. A month before she died, Shannon had threatened to shoot herself. ¶ 8 The State charged Lopez with criminal homicide murder. Because the medical examiner’s report ruled out the possibility that the gun had accidentally discharged, the key dispute was who fired the shot that killed Shannon. ¶ 9 The physical evidence was inconclusive. The medical examiner testified that the location of the wound was “atypical” for a suicide, but that he could not determine the manner of death conclusively. Examiners found gunshot residue on Lopez’s hand, but all that could be gleaned from this was that Lopez was “in proximity when [the] firearm was discharged.” A blood spatter analyst was unable to conclude whether Shannon’s wound was self-inflicted. ¶ 10 To prove that Shannon did not shoot herself, the State offered expert testimony from Dr. Craig Bryan, a clinical psychologist. Dr. Bryan specializes in the treatment of suicide patients using the Fluid Vulnerability Theory of Suicide (FVTS). FVTS is the “most commonly used theory and approach to developing treatment and understanding suicide risks.” The theory is based on “scientific evidence gained from clinical care of suicide patients as well as multidisciplinary scientific efforts internationally.” _____________________________________________________________ 1There was some dispute about whether witnesses had moved Shannon’s body and if the gun and holster were moved when the car was towed. Because these disputes are not relevant to our analysis, we do not discuss them here. 3 STATE v. LOPEZ Opinion of the Court ¶ 11 FVTS assesses two different types of risk: baseline and acute. Predispositions—including demographic factors, “[d]ifficulty managing emotions,” and “history of psychiatric disorders”— increase the baseline risk, meaning that “[i]ndividuals with many predispositions . . . . experience more suicidal crises more often and take longer to ‘recover’ from crises and discrete periods of emotional distress.” This baseline risk can be “offset[]” by “protective factors,” such as an optimistic outlook, a strong support network, or motherhood. “Acute risk,” on the other hand, “entails the emotional, physiological, behavioral, and cognitive factors associated with an active suicidal episode.” Taking baseline and acute risk together, the model posits that “a triggering event will only lead to suicide among individuals with sufficient [baseline risk].” When applying FVTS, Dr. Bryan conducts interviews with his patients. Sometimes, Dr. Bryan employs testing “designed to . . . identify the risk and protective factors in a way that might not be obvious to the respondent.” ¶ 12 Lopez challenged the admission of Dr. Bryan’s testimony on various grounds, including that it was neither helpful nor reliable. The district court admitted “Dr. Bryan’s opinion as to whether Shannon Lopez’s behavior prior to her death was inconsistent with suicide” into evidence. The district court excluded, however, “[a]ny testimony that Dr. Bryan’s opinions are definitive or based on scientific certainty.” ¶ 13 In response to Dr. Bryan’s testimony, the defense called an expert who testified that Shannon’s death was a “classic suicide,” noting that “[t]ypically someone doesn’t hold their head still while you shoot them.” ¶ 14 The State also sought to offer evidence of several prior acts involving Lopez threatening a family member with a gun and/or pointing a gun at their head. The court allowed the admission of two of those acts. 2 The first described Lopez and Shannon talking with their coworker about how to kill effectively. To demonstrate, Lopez pulled out a gun and pointed it at Shannon’s head near her left ear. The district court found that “the act of someone describing the best place to shoot someone, in fact even demonstrating that, is very _____________________________________________________________ 2The district court rejected the State’s argument that those acts, as well as the others, could be admitted under the doctrine of chances. 4 Cite as: 2018 UT 5 Opinion of the Court relevant to the identification of someone who may have done that at another time.” ¶ 15 The second prior act involved an argument Lopez had with an ex-wife where he “hit her in the stomach,” “pointed a gun at her head,” and verbally threatened to kill them both if she ever sought a divorce. The district court reasoned this evidence was admissible because “the identity of the shooter is the issue, and therefore, it is relevant to that, what he did to a prior spouse, under a prior circumstance when she indicated she was leaving him.” ¶ 16 The jury found Lopez guilty of murder. The jurors additionally found Lopez had used a dangerous weapon. The district court sentenced Lopez to sixteen years to life. Lopez appeals. ISSUES AND STANDARD OF REVIEW ¶ 17 Lopez first argues that Dr. Bryan’s testimony should not have been admitted because, among other things, it lacked an adequate foundation. Next, Lopez contends that the character evidence should not have been admitted because it was not relevant, not offered for a proper purpose, and was prejudicial. Lopez also argues cumulative error and claims there was insufficient evidence to convict him. ¶ 18 We review the admission of expert testimony and character evidence under an abuse of discretion standard. State v. Maestas, 2012 UT 46, ¶ 154, 299 P.3d 892; State v. Thornton, 2017 UT 9, ¶ 56, 391 P.3d 1016. Because we agree with Lopez on the first two grounds, and conclude they constitute harmful error, we do not reach the issues of cumulative error or sufficiency of the evidence. ANALYSIS I. The District Court Abused Its Discretion by Admitting Dr. Bryan’s Testimony Without an Adequate Foundation Establishing Its Reliability ¶ 19 Lopez challenges the admissibility of Dr. Bryan’s testimony. Dr. Bryan used the Fluid Vulnerability Theory of Suicide (FVTS) to assess the likelihood that Shannon had taken her own life. Lopez asserts that the FVTS testimony was not helpful, that it impermissibly addressed the ultimate issue, and that there was not an adequate threshold showing of its reliability. We agree with Lopez that the State did not make the threshold showing of FVTS’s reliability when applied to a decedent, and that therefore the district 5 STATE v. LOPEZ Opinion of the Court court abused its discretion by admitting Dr. Bryan’s testimony under Utah Rule of Evidence 702. 3 ¶ 20 We rely on our district court judges to act as “gatekeeper[s]” to “screen out unreliable expert testimony.” UTAH R. EVID. 702 advisory committee’s note. This requires our judges to view proposed expert testimony with “rational skepticism.” Id. Utah Rule of Evidence 702 details the framework a judge should employ to perform her gatekeeping function. First, the judge must determine that the expert is qualified “by knowledge, skill, experience, training, or education” and that the proposed expert testimony will “help the trier of fact to understand the evidence or to determine a fact in issue.” UTAH R. EVID. 702(a). Second, the judge inquires as to whether the “[s]cientific, technical, or other specialized knowledge” underlying the expert’s testimony meets a threshold showing that the “principles or methods . . . underlying . . . the testimony (1) are reliable, (2) are based on sufficient facts or data, and (3) have been reliably applied to the facts.” Id. 702(b). ¶ 21 The threshold showing is satisfied if “the underlying principles or methods, including the sufficiency of facts or data and the manner of their application to the facts of the case, are generally accepted by the relevant expert community” or if the testimony meets a threshold showing of reliability. Id. 702. If the expert testimony clears those hurdles, the district court may admit the evidence. And we afford the district court discretion in its decision to admit or exclude expert testimony. State v. Maestas, 2012 UT 46, ¶ 122, 299 P.3d 892. The district court abuses its discretion when “no reasonable person would take the view the trial court adopted.” Id. A. There Was No Threshold Showing that the Fluid Vulnerability Theory of Suicide Is a Reliable Method to Evaluate the Risk of Suicide of a Deceased Person ¶ 22 Although Lopez raises several challenges to Dr. Bryan’s testimony, we focus our analysis on the question of threshold reliability. As noted above, the State could have met this threshold showing by demonstrating that the method Dr. Bryan used is “generally accepted by the relevant expert community” or by establishing that the principles underlying his testimony are _____________________________________________________________ 3Because we resolve this matter under Utah Rule of Evidence 702, we do not reach Lopez’s alternative arguments. 6 Cite as: 2018 UT 5 Opinion of the Court “reliable, . . . based upon sufficient facts or data, and . . . have been reliably applied to the facts.” UTAH R. EVID. 702. ¶ 23 The State attempted to meet its rule 702 burden with evidence that FVTS 4 is generally accepted in the psychological community. 5 When asked if FVTS was generally accepted, Dr. Bryan _____________________________________________________________ 4 FVTS assesses an individual’s risk for suicide based on his or her baseline and acute risk. Baseline risk factors include genetics, demographics, and “a history of suicidal behavior,” among other things, and determine someone’s risk of suicide over time. Acute risk is “fluid” and takes into account someone’s physical and mental health as well as her thought process and emotional state. Protective factors—parenting, a sense of purpose, and a strong support system, to name a few—also play an important role because they mitigate someone’s risk factors. 5 The parties argue that a number of cases should inform our analysis. See Halvorsen v. Plato Learning, Inc., 167 F.App’x. 524 (6th Cir. 2006); State v. Guthrie, 627 N.W.2d 401 (S.D. 2001); Foster v. Globe Life & Accident Ins. Co., 808 F. Supp. 1281 (N.D. Miss. 1992); Bethley v. Keller Constr., 836 So. 2d 397 (La. Ct. App. 2002). These cases are not on point because they discuss psychological autopsies. Psychological autopsies seek to establish the cause of a person’s death. For example, in Guthrie, the court allowed testimony that “included an account of the common factors for persons at risk for suicide [and] a comparison of those factors to [that] case.” 627 N.W.2d at 414. And, although the court took issue with an expert’s conclusory opinion, the court permitted the expert to testify concerning the “typical characteristics or profiles of suicidal persons” and whether the decedent “met a suicidal profile.” Id. at 415–17. The cases addressing psychological autopsies are inapposite here. Although the testimony itself is similar, the method is not. Both parties have distinguished FVTS from a psychological autopsy. As Dr. Bryan said during the preliminary hearing, A psychological autopsy has many similarities to the methods I’ve used; however, it’s different in that the psychological autopsy is often intended to arrive at a psychiatric diagnosis for the decedent. In this case my purpose was not to try to diagnosis the individual. I was trying to make a determination about the likelihood of suicide. (continued . . .) 7 STATE v. LOPEZ Opinion of the Court replied that it was, elaborating that “when you look particularly at researchers who are treatment developers, [and] they are testing therapies to reduce suicidal behaviors, this particular model is the leading and the most commonly used theory and approach to developing treatment and to understanding suicide risks.” (Emphasis added). And, “[w]hen . . . use[d] [for] treatment[] . . . based on the principles contained within the theory, [there is] typically [a] 50 to 60 percent reduction[] in suicidal behaviors.” ¶ 24 Dr. Bryan’s testimony addressed the acceptance of FVTS as a tool to evaluate patients. The foundation the State attempted to lay did not speak to the precise question the district court needed to answer: Is FVTS generally accepted as a means of assessing the risk of suicide in someone who has passed away? With respect to that question, Dr. Bryan provided no information on how effective the theory was at determining suicide risk of those who are deceased, nor did he indicate whether there was any peer-reviewed literature regarding this application. Indeed, it appears that Dr. Bryan never even addressed whether FVTS had ever been used to assess the risk of someone who had died. ¶ 25 The district court nevertheless concluded that the FVTS “methodology . . . is generally accepted in the psychological community.” As suggested above, the problem with this conclusion is that the State placed no evidence before the court that FVTS was generally accepted in the psychological community for the purpose of determining whether any particular death was the result of suicide. The evidence before the district court demonstrated that FVTS is based on “scientific evidence gained from clinical care of suicide patients as well as multidisciplinary scientific efforts internationally.” (Emphasis added). Dr. Bryan’s published works on suicide are similarly based on “[c]linical trials, epidemiological research as well as laboratory-based research.” ¶ 26 Demonstrating that FVTS was generally accepted to assess suicide risk in live patients is not the same as demonstrating that it is Because our analysis rests on the reliability of a method other than a psychological autopsy, and none of those cases otherwise speak to the use of FVTS to assess the likelihood of suicide, we assess FVTS’s reliability without reference to the cases that have analyzed the admission of psychological autopsy evidence. 8 Cite as: 2018 UT 5 Opinion of the Court generally accepted to gauge whether a decedent died by her own hand. On the record before the district court, there were reasons to question whether FVTS is reliable as a post-mortem tool. Dr. Bryan referenced interviews with his patients to assess their suicide risk. And Dr. Bryan’s description of FVTS demonstrated its reliance on accurately identifying baseline and acute factors. Dr. Bryan never discussed whether he could accurately identify those factors in someone he had never interviewed, let alone someone who was incapable of being interviewed. For example, Dr. Bryan testified that Shannon’s acute risk for suicide was low because of her positive outlook, social engagement, and improved mood. Dr. Bryan never testified, however, that he could accurately assess those risks without actually interviewing Shannon. 6 ¶ 27 The State argues that the “question of whether someone is at risk to commit suicide is the same question as whether someone was at risk—it merely focuses on a different time.” And the State may be correct about that, but the question for the district court was whether the difference in that timing undercut FVTS’s reliability. We do not rule out the possibility that FVTS could be used to determine whether a decedent was at risk of suicide, but the State needed to lay a foundation that it could. It failed to do so. ¶ 28 The district court held that FVTS was generally accepted, but did not draw a distinction between how FVTS had been used to gain general acceptance in the psychological community and the manner in which Dr. Bryan proposed to use it here. On the record before it, the district court abused its discretion by admitting the testimony as the product of a generally accepted scientific method. ¶ 29 The State could have also satisfied the threshold showing by offering evidence that the “principles or methods . . . underlying . . . the testimony (1) are reliable, (2) are based on sufficient facts or _____________________________________________________________ 6 Or, indeed, anyone. It does not appear that Dr. Bryan interviewed anyone in connection with his opinion, opting instead to rely on materials the State selected for him. These materials included transcripts of police-conducted interviews, police reports, the medical examiner’s report, Shannon’s medical records, cell phone records, and something Shannon posted on Facebook on the day she died. Dr. Bryan testified that the interviews gave him all the information that he needed, but this statement was not sufficient to lay a foundation for this seemingly novel use of FVTS. 9 STATE v. LOPEZ Opinion of the Court data, and (3) have been reliably applied to the facts.” UTAH R. EVID. 702(b). For the same reasons we have just discussed, we conclude that there was an insufficient basis for the district court to conclude that FVTS is reliable when used to assess suicide risk post-mortem. 7 B. Dr. Bryan’s Testimony Likely Swayed the Jury and Was Therefore Harmful Error ¶ 30 “[A]n [evidentiary] error requires reversal only if there is ‘a reasonable likelihood of a more favorable result’ for the accused had the error not occurred.” State v. Tuttle, 780 P.2d 1203, 1213 n.12 (Utah 1989) (citation omitted). “A reasonable likelihood of a more favorable outcome exists if our confidence in the result of the trial is eroded.” Id. (citation omitted). Because the admission of Dr. Bryan’s testimony has eroded our confidence in the jury’s verdict, we conclude that the error was harmful. ¶ 31 As the State said multiple times, this case hinged on whether Lopez or Shannon fired the weapon. Some evidence pointed to Lopez, including the fact that Shannon was right-handed and was shot on the left side of her head. Lopez “admitted . . . that [Shannon] was not familiar with [the gun].” The gun and holster were also found on the driver’s side of the car (where Lopez had been) and Shannon’s right hand remained hidden in her sleeve. ¶ 32 Other evidence suggested Shannon took her life. An expert testified that this was a “classic suicide” and “[t]ypically someone doesn’t hold their head still while you shoot them.” 8 There was also evidence that Shannon was experienced with firearms and could shoot with her left hand, albeit with some difficulty. And there was evidence that Shannon had mentioned suicide on occasion. ¶ 33 Given this conflicting evidence, Dr. Bryan’s opinion that Shannon’s death was inconsistent with suicide likely swayed the jury. It was the strongest statement the State introduced to demonstrate that Shannon’s death was not a suicide. Without Dr. Bryan’s testimony, the most direct evidence speaking to whether _____________________________________________________________ 7 Lopez also contends that Dr. Bryan’s opinion was not based on sufficient facts and data and that he did not reliably apply FVTS to the facts of the case. We offer no opinion on these contentions. 8 The State has not challenged the admission of this expert’s testimony. 10 Cite as: 2018 UT 5 Opinion of the Court Shannon had killed herself was a State expert opining that the location of the wound was “atypical” for suicide. ¶ 34 Dr. Bryan’s testimony drove to the heart of the matter such that we conclude that the admission of Dr. Bryan’s testimony on the foundation presented was prejudicial. II. The District Court Abused Its Discretion by Admitting Evidence of the Prior Acts to Attempt to Prove that Lopez Fired the Weapon ¶ 35 Lopez also argues that the district court abused its discretion by admitting evidence of prior instances of Lopez pointing guns at family members. The State initially sought the admission of six instances where Lopez pointed a gun at a member of his family. The State argued that this evidence could be admitted for a number of non-character purposes, including lack of mistake, knowledge, intent, and identity. Lopez countered by arguing, among other things, that the State was seeking to admit the testimony solely to show that he had a propensity to point firearms at family members. ¶ 36 The district court excluded four of the prior acts but allowed the State to introduce evidence that Lopez had discussed the best way to commit suicide or kill someone, and that in the course of that discussion, he had placed his gun behind Shannon’s left ear. The district court also admitted testimony that during an argument with a former spouse, Lopez had pointed a gun at her and suggested that if she were to leave him, he would kill them both. The district court reasoned that both of these acts could be admitted for the non- character purpose of showing “the identity of the shooter.” ¶ 37 Lopez challenges the district court’s decision to admit these prior acts. Specifically, Lopez argues that: (1) admission of the prior acts was not relevant to a proper non-character purpose and (2) the probative value of such evidence was substantially outweighed by its ability to unfairly prejudice the jury. ¶ 38 Generally, “[e]vidence of a crime, wrong, or other act is not admissible to prove a person’s character in order to show that on a particular occasion the person acted in conformity with the character.” UTAH R. EVID. 404(b)(1). We recognize an exception to that rule where prior acts are “relevant[,] offered for a genuine, noncharacter purpose,” and not unduly prejudicial. State v. Lucero, 2014 UT 15, ¶ 13, 328 P.3d 841, abrogated on other grounds by State v. Thornton, 2017 UT 9, 391 P.3d 1016. “[G]enuine, noncharacter purpose[s],” id., include but are not limited to “proving motive, 11 STATE v. LOPEZ Opinion of the Court opportunity, intent, preparation, plan, knowledge, identity, absence of mistake, or lack of accident.” UTAH R. EVID. 404(b). The district court admitted the two prior acts reasoning that they were relevant to show the identity of the person who fired the weapon. ¶ 39 We have generally upheld the admission of prior acts to show identity in two circumstances. First, we have upheld the admission of prior acts to prove identity because the evidence reveals that the defendant has a modus operandi. See Lucero, 2014 UT 15, ¶ 15. In this context, we have spoken of modus operandi as an “intermediate inference” that demonstrates identity. Id. (“In seeking admission of prior acts for the purpose of proving ‘identity,’ parties are most often actually seeking to admit evidence of an intermediate inference, such as modus operandi, that bears on the ultimate issue of identity.” (footnote omitted)). ¶ 40 To use a prior act to show modus operandi, the prior act must bear a “very high degree of similarity” to the charged act and demonstrate “a unique or singular methodology.” Id. (citation omitted). For example, in Lucero, we concluded that a prior act of child abuse was “remarkably similar” to the charged crime where a mother was accused of fatally injuring her toddler. Id. ¶¶ 1–8, 16. “Both injuries occurred along the spinal column and were caused by the spine being bent unnaturally.” Id. ¶ 16. The two acts also “occurred within days of each other.” Id. Because of the similarity of the injuries, methods, and “temporal proximity,” we held that the evidence was properly admitted as evidence of a modus operandi. Id. ¶ 41 We have also affirmed the introduction of 404(b) evidence where the perpetrator had a particular method of robbing Payless Shoe stores. State v. Decorso, 1999 UT 57, ¶¶ 27–28, 993 P.2d 837, abrogated on other grounds by Thornton, 2017 UT 9. Each time, the accused pretended to be a customer and waited until the store was closed and locked to rob it. Id. ¶ 27. The victim in both instances was a female store clerk. Id. The perpetrator took two pairs of shoes the first time, and had three pairs ready to take the second time. Id. The accused also wore or planned to wear rubber gloves in both instances and cut or removed the telephone cord. Id. Based on these similarities, we concluded that the “trial court did not abuse its discretion in concluding that these were ‘signature-like’ crimes.” Id. ¶ 42 In contrast, here, the admitted events are not similar enough to demonstrate a modus operandi. To find that the prior 12 Cite as: 2018 UT 5 Opinion of the Court incident of Lopez pointing a gun at Shannon to demonstrate how to kill effectively was admissible, the court emphasized “the similarity between the position of the gun as demonstrated and the position of the gun in the crime that was committed.” And the district court was correct that the positioning was indeed similar: in the prior act and the crime, the gun was pointed at or near Shannon’s left ear. But the commonality of the positioning of the gun on a single occasion is not enough to establish that Lopez had a modus operandi when it came to pointing a gun. The similar positioning does not establish a “signature” like the combination of commonalities in Decorso did; pointing a gun in an approximate location is not the same as entering the same chain of shoe stores right before closing and using the same method to rob it. See id. ¶¶ 27–28. There simply are not enough factual similarities between the gun pointing incidents to constitute a modus operandi. ¶ 43 The other admitted act—Lopez threatening to shoot his ex- wife if she left him—is not sufficiently similar to the alleged behavior to be considered a modus operandi. The district court noted that the “similarities between the crimes are very close” because both of them involved “[p]ointing a gun at a spouse in the context of . . . the discussion that the spouse was leaving.” But threatening one’s wife with a gun on two occasions is not a “unique or singular methodology.” Lucero, 2014 UT 15, ¶ 15 (citation omitted). Indeed, neither the context of the act nor its execution is similar enough to the charged conduct to suggest a modus operandi. ¶ 44 We have also upheld the admission of 404(b) evidence when it supported an intermediate inference that tied the defendant to the charged crime. For example, in State v. Reece, we allowed prior act evidence “showing that [the defendant] had access to the type of gun investigators determined was likely the murder weapon.” 9 2015 UT 45, ¶ 58, 349 P.3d 712. The evidence was “offered for the noncharacter purpose of identi[ty.]” Id. The defendant had “claimed that he could not have committed the murder because he did not even have access to [the murder weapon] when the victim was killed.” Id. The State was able to rebut this defense by demonstrating that the defendant had stolen the firearm and therefore had access to the murder weapon. Id. _____________________________________________________________ 9Intermediate inferences include access to the murder weapon, motive, or common plan. 1 MCCORMICK ON EVID. § 190 (7th ed. 2016). 13 STATE v. LOPEZ Opinion of the Court ¶ 45 Similarly in State v. Shaffer, we permitted prior act evidence that the defendant had stolen a wallet. 725 P.2d 1301, 1308 (Utah 1986). The State used the evidence to show access to the murder weapon because the defendant had used the identification in the stolen wallet to purchase the gun ten days prior to the murder. Id. By demonstrating access to the identification, and thus the murder weapon, the State sought to “establish[] the identity of the defendant as the person in possession of the gun the killed [the victim].” Id. ¶ 46 And our court of appeals has also affirmed use of evidence to show identity by linking the defendant to the murder weapon. In State v. Clark, the court permitted evidence of a prior shooting where the same gun was used in the charged crime. 2014 UT App 56, ¶¶ 21–22, 322 P.3d 761. The court reasoned that, because the defendant assert[ed] that he was not present at the [subsequent] shooting, identity of the perpetrator was clearly at issue in this case. . . . Evidence that [the] [d]efendant was carrying the . . . handgun when he arrived at the scene of the [prior] shooting, that he displayed a familiarity with the handgun’s operation, and that he actually used the handgun in that incident all tend to show that the handgun was actually his. Id. ¶ 22. Therefore, the use of the gun in the prior shooting tended to show that the defendant was the perpetrator in the later shooting by demonstrating that he likely had access to the gun. Id. ¶ 47 The evidence at issue here did not support any intermediate inference of identity like those in Reece, Shaffer, or Clark. The State did not, for example, seek to show the identity of the shooter with evidence that Lopez had access to a firearm like the one that killed Shannon. Rather, the district court said it would admit the evidence that Lopez had pointed a gun at Shannon to prove “identification” and the evidence that Lopez had aimed a firearm at an ex-wife and threatened to kill her to show the “identity of the shooter.” In other words, the district court admitted the evidence because it tended to show that Lopez had in the past pointed guns at family members. And the fact that he had done this in the past suggested that he had done it on the night Shannon died. This is nothing more than the propensity evidence 404(b) excludes, and the district court abused its discretion by admitting it. See 1 MCCORMICK 14 Cite as: 2018 UT 5 Opinion of the Court ON EVID. § 190 (7th ed. 2016) (“[T]he need to prove identity should not be, in itself, a ticket to admission.”). ¶ 48 On appeal, the State raises an argument that the district court rejected below. The State argues that the district court could have admitted the evidence because it established identity through the doctrine of chances. The district court rejected that argument, and the State did not cross-appeal that determination. As such, the State asks us to affirm the district court on an alternative ground apparent from the record. Bailey v. Bayles, 2002 UT 58, ¶ 20, 52 P.3d 1158 (“[A]n appellate court may affirm the judgment appealed from if it is sustainable on any legal ground or theory apparent on the record.”). We decline the State’s invitation, as it is not apparent from the record that the district court abused its discretion by refusing to admit the evidence through the doctrine of chances. ¶ 49 As an initial matter, we have not previously applied the doctrine of chances to show identity. We have held that the doctrine of chances can “rebut a charge of fabrication.” State v. Verde, 2012 UT 60, ¶ 47, 296 P.3d 673, abrogated on other grounds by Thornton, 2017 UT 9. We have also concluded that the doctrine can be used to show the requisite mens rea or lack of consent in a rape case. State v. Lowther, 2017 UT 34, ¶¶ 25–26, 398 P.3d 1032. And we have noted that the doctrine might be used for other purposes, including to “rebut defenses based on mistake, coincidence, . . . accident,” and intent. Id. ¶ 23. But we have not applied the doctrine to establish identity. We need not resolve whether such application would be proper because the doctrine is inapplicable to the set of facts presented here. ¶ 50 The doctrine of chances is “a theory of logical relevance that ‘rests on the objective improbability of the same rare misfortunate befalling one individual over and over.’” Verde, 2012 UT 60, ¶ 47 (citation omitted); See also Edward J. Imwinkelried, An Evidentiary Paradox: Defending the Character Evidence Prohibition by Upholding a Non-Character Theory of Logical Relevance, the Doctrine of Chances, 40 U. RICH. L. REV. 419, 448 (2006) (“The claim is based on the disparity between the expected and actual values: How many incidents would we expect the average person to be involved in, and how many incidents was the defendant involved in?” (footnote omitted)). ¶ 51 United States v. Woods ably illustrates the doctrine’s application. 484 F.2d 127 (4th Cir. 1973). There, the defendant was tried for the murder of a child in her care. Id. at 128. The child died from respiratory problems, though when taken to the hospital for 15 STATE v. LOPEZ Opinion of the Court those same troubles on prior occasions, the doctors sent him home because he was perfectly healthy. Id. at 129–30. The State’s expert testified that there was a three in four chance that the child was smothered as opposed to dying from natural causes. Id. at 130. In order to prove that Woods smothered the child, the State presented evidence that nine other children in the defendant’s care suffered from twenty similar respiratory episodes, and that seven of those children died. Id. The court concluded that the evidence was admissible because, when “considered collectively, . . . an unmistakable pattern emerges. That pattern overwhelmingly establishes defendant’s guilt.” Id. at 135. ¶ 52 Much like the unexpected death of otherwise healthy children, other doctrine of chances cases involve rare events happening with unusual frequency. See id. In a classic English case, for example, the defendant was convicted of murdering his wife, who was found dead in her bathtub. Rex v. Smith, 11 Crim. App. 229, 84 L.J.K.B. 2153 (1915). The court permitted evidence that two other wives of the defendant had also died in their bathtubs “under nearly the same circumstances.” Id. Similarly, in United States v. York, the Seventh Circuit allowed evidence of a prior crime under the doctrine of chances. 933 F.2d 1343, 1349–52 (7th Cir. 1991), overruled on other grounds by Wilson v. Williams, 182 F.3d 562 (7th Cir. 1999). The defendant was tried for insurance fraud after taking out policies on his colleague and their bar. Id. at 1345–47. Two homemade explosives destroyed the bar; his colleague was also killed, though the evidence was unclear regarding whether she had been killed before the explosion. Id. at 1346. The court permitted the State to introduce evidence that the defendant had taken out a policy on his wife and later killed her to demonstrate that the defendant intended to defraud the insurance company when he took out the policy (which was required to establish the crime). Id. at 1349–50. ¶ 53 Here, the State offered several prior acts under the doctrine of chances, including the two that were admitted. One of these prior acts involved Shannon: the incident described in detail above. Two involved Lopez’s ex-wife: the incident described above and another where Lopez pointed a gun at her head while she was in a closet. The other three involved Lopez’s son, K.L. K.L. would have testified that after an altercation over pizza, he bit Lopez which caused Lopez to run to his gun safe to obtain a firearm which he then racked as 16 Cite as: 2018 UT 5 Opinion of the Court K.L. ran downstairs. 10 K.L. would have also testified that Lopez once pointed a gun at his head and tapped his head with the weapon. And K.L. would have testified that “the family would routinely point guns at one another in an ‘educational’ context.” ¶ 54 We have opined that for evidence to be admitted under the doctrine of chances, it must meet four foundational requirements: materiality, similarity, independence, and frequency. Verde, 2012 UT 60, ¶¶ 57–61; Lowther, 2017 UT 34, ¶ 40 n.66 (“We note that [Verde’s] foundational requirements are requirements within the context of rule 404(b). A court must find that each of the requirements has been satisfied to admit doctrine of chances evidence for purposes of a proper, non-character statistical inference.”). The evidence the State sought to introduce fails to clear this hurdle. ¶ 55 To be material, “[t]he issue for which the uncharged misconduct evidence is offered ‘must be in bona fide dispute.’” Verde, 2012 UT 60, ¶ 57 (emphasis omitted) (citation omitted). We agree with the State that the shooter’s identity is in “bona fide dispute.” ¶ 56 But the district court could have legitimately questioned the independence of the prior acts. See id. ¶ 60. The independence requirement helps ensure there is no collusion between the victims and that the victims have not influenced each other’s recollections of what occurred. All but one of the prior acts involved either K.L. or Lopez’s ex-wife (K.L.’s mother). And K.L. and Lopez’s ex-wife lived together for over a year after all five of the prior acts involving them occurred. This provided ample opportunity for ex-wife and K.L. to discuss the incidents and compromise the independence of their recollections. ¶ 57 The remaining two factors, similarity and frequency, interact with each other to become a safeguard against the doctrine of chances becoming a work-around for the admission of otherwise improper propensity evidence. For doctrine of chances purposes, frequency does not mean just how many times a prior act has _____________________________________________________________ 10 K.L.’s testimony was confused regarding whether Lopez pointed a firearm at him. At one point, K.L. stated that he “didn’t see [Lopez] point the gun at [him],” but later said—regarding what seems to be the same incident—that Lopez aimed the gun at him at close range. 17 STATE v. LOPEZ Opinion of the Court occurred, but whether “[t]he defendant [has] been accused of the crime or suffered an unusual loss ‘more frequently than the typical person endures such losses accidentally.’” Id. ¶ 61 (emphasis omitted) (citation omitted). Similarity assumes importance in this inquiry because a district court could logically conclude that the more similarities repeated events share, the less likely they are to occur frequently by accident. ¶ 58 And here, Lopez’s prior acts did not share a great deal of similarity. Although each involved Lopez drawing a firearm in the presence of family members, they happened in very different contexts. The two incidents where Lopez threatened his ex-wife with a gun are the most similar to one another but because his ex-wife “could not recall the [closet] incident with great detail[,]” the district court may not have been well-positioned to accurately assess the degree of similarity of those events. And the prior act in which K.L. thought he heard Lopez rack a gun bears some resemblance to those two. But the other prior acts describe different actions. Two of the other events—pointing a gun to “educat[e]” his family and pointing a gun at Shannon in front of a co-worker to show the most effective way to kill, do not bear great similarity to the instances where Lopez pointed a firearm during a domestic argument. The dissimilarity between the actions makes it difficult to see how the district court would have abused its discretion in finding that these prior acts were not suitable candidates for admission under the doctrine of chances. ¶ 59 And the lack of similarity made it more difficult for the State to meet its burden of demonstrating frequency. 11 Had the prior _____________________________________________________________ 11 The proponent must establish that, together with the uncharged incident, the charged incident would represent an extraordinary coincidence. In some cases, that will be obvious. [Rex] is a case in point. In a fact situation such as [Rex], the jury hardly needs an expert’s testimony to appreciate that, on average, finding one’s spouse drowned in the family bathtub is at most a ‘once in a lifetime’ experience. In other cases, though, the proponent may need to introduce independent evidence to establish the ordinary incidence of the type of event in which the defendant was involved. (continued . . .) 18 Cite as: 2018 UT 5 Opinion of the Court acts been multiple incidents of Lopez shooting a person by accident, or several occasions of placing a firearm to someone’s ear during a fight, the district court possibly could have concluded that those actions would not occur repeatedly by accident. 12 But on these facts, the district court could readily conclude that Lopez had not suffered the “same rare misfortune” repeatedly. 13 Verde, 2012 UT 60, ¶ 47 (citation omitted). ¶ 60 Because of this, it is not apparent on the record that the district court abused its discretion by refusing to allow this evidence in under the doctrine of chances. Bailey, 2002 UT 58, ¶ 20; Verde, 2012 UT 60, ¶¶ 57–61. ¶ 61 Although the 404(b) evidence was admitted in error, “an [evidentiary] error requires reversal only if there is ‘a reasonable Imwinkelried, An Evidentiary Paradox, supra ¶ 50, at 437 (footnotes omitted). 12 What distinguishes the doctrine of chances from character evidence is the intermediate inference the jury makes. With character evidence, the jury would infer from the prior misconduct that the defendant must have “bad character,” and thus must be guilty because the defendant “acted ‘in character.’” Imwinkelried, An Evidentiary Paradox, supra ¶ 50, at 426. Using the doctrine of chances, however, the jury makes a decision about the weight of prior act evidence based on “the objective improbability of so many [incidents].” Id. at 436. Similarity and frequency are both important inputs for determining this improbability; the less similar the acts, the more probable it is that they would occur in the general population. And the less frequently they occur in the general population, the more it is “objective[ly] improbabl[e]” that so many incidents would occur randomly. See id. Thus, to determine probability, similarity and frequency may be examined in tandem. 13 “If the judge has no satisfactory basis for determining the frequency of such accidental occurrences among the general populace, the judge may not admit the uncharged misconduct evidence under the aegis of the doctrine of chances.” Edward J. Imwinkelried, The Use of Evidence of An Accused’s Uncharged Misconduct to Prove Mens Rea: The Doctrines Which Threaten to Engulf the Character Prohibition, 51 OHIO ST. L. J. 575, 592 (1990). There was no such basis here. 19 STATE v. LOPEZ Opinion of the Court likelihood of a more favorable result’ for the accused had the error not occurred.” State v. Tuttle, 780 P.2d 1203, 1213 n.12 (citation omitted). We have concluded that admission of less probative character evidence was harmless in State v. Hamilton. 827 P.2d 232, 233 (Utah 1992). The prior act evidence there concerned the defendant physically abusing his girlfriend. Id. at 239–40. Even without the prior act, the remaining evidence was strong: the defendant was seen with the victim in his truck after she went missing, her hairs were found in his truck, and his fingerprints were found on beer cans near the location he left the body. Id. at 234–35. As a result, we held the admission of the prior act evidence harmless. Id. at 233. ¶ 62 In State v. Webster, the court of appeals reversed a conviction based upon the improper admission of character evidence because—after omitting the impermissible character evidence—the court of appeals opined that it was “not confident” that the jury would have convicted on the remaining evidence. 2001 UT App 238, ¶ 39, 32 P.3d 976. The defendant, a car salesman, was accused of stealing a car off the lot. Id. ¶¶ 1–8. The court found harmful error where evidence had been admitted that Webster had stolen a car previously along with a statement from his wife confirming that he had driven the stolen car around. Id. ¶ 39. The court found the remaining evidence, though sufficient to convict, too insubstantial in comparison to the erroneously admitted evidence. Id. The defendant was employed as a salesman at the lot where the car was stolen, someone had seen him drive off the lot with the car even though salesmen were not allowed to do so, the defendant quit his sales job after only two weeks, and the car was ultimately found in front of the apartment complex that the defendant lived at. Id. ¶ 39. The court reversed because “the damning statements . . . and his apparent history of taking cars that did not belong to him could easily have been the deciding factors . . . .” Id. ¶ 63 Here, the remaining evidence is less conclusive than that in either Webster or Hamilton. Unlike the fingerprint and hair evidence in Hamilton, the physical evidence here was inconclusive. Compare Hamilton, 827 P.2d at 234–45, with supra ¶ 9. The State introduced the testimony of the medical examiner who opined that the wound location was strange for a suicide. But, Lopez called an expert who testified the death was a “classic suicide.” The jury also heard evidence that Shannon had previously discussed suicide. On this record, there is a “reasonable likelihood” that the character evidence 20 Cite as: 2018 UT 5 Opinion of the Court could have likely swayed the jury toward conviction. Tuttle, 780 P.2d at 1213 n.12 (citation omitted). ¶ 64 The error in admitting the 404(b) evidence is reversible error because there was “a reasonable likelihood of a more favorable result” had the character evidence been excluded. Id. (citation omitted). CONCLUSION ¶ 65 We conclude that Dr. Bryan’s expert testimony was inadmissible because the State did not demonstrate that it met a threshold of reliability as Utah Rule of Evidence 702 requires. And because Dr. Bryan offered significant testimony on the key dispute in a case where much of the other evidence was ambiguous, this error was harmful. Furthermore, the district court abused its discretion in admitting evidence of Lopez’s prior acts for the purpose of showing identity. This error was also harmful. Accordingly, we reverse Lopez’s conviction and remand for proceedings consistent with this opinion. 21