dissenting:
¶ 25 I respectfully disagree with my colleagues for several reasons. First, I believe Arizona’s existing legal framework for evaluating harmless error, articulated by our supreme court in numerous cases, provides appropriate and adequate standards for resolving the legal issue remanded to us. See, e.g., State v. Valverde, 220 Ariz. 582, ¶ 11, 208 P.3d 233, 236 (2009); Anthony, 218 Ariz. 439, ¶ 39, 189 P.3d at 373; Bible, 175 Ariz. at 588, 858 P.2d at 1191; State v. McVay, 127 Ariz. 450, 453, 622 P.2d 9, 12 (1980). I therefore see no need for the fragmented harmless error analysis adopted by the majority, particularly on the facts of this case.
¶ 26 The majority insists that it does not depart from our supreme court’s established standards by employing its own multi-factor approach in evaluating harmless error, but I question that view and the necessity of such an approach here. Though Bible cited Wein-stein, it did so in support of its declaration that “[t]here is no bright line statement of what is and what is not harmless error,” and only noted generally that Weinstein “list[ed] factors courts examine in determining whether error was harmless.” 175 Ariz. at 588, 858 P.2d at 1191. It did not, however, mention, let alone expressly adopt, any of those factors. See id. Moreover, the court further clarified that “[d]ue to th[e] case-specific factual inquiry [involved], an error may be harmless in one ease but require reversal in another.” Id. In my view, the supreme court’s “broad statements” are neither con-elusory nor accidental; instead, it appears the court has intentionally articulated flexible standards in order to account for the great number of variables and factual considerations that often arise in multifaceted criminal cases.
¶ 27 Second, I would conclude under either approach that the state has established beyond a reasonable doubt that the exclusion of Romero’s expert did not influence the jury’s verdict. See Leteve, 237 Ariz. 516, ¶ 25, 354 P.3d at 401-02. As previously stated by my colleague who specially concurred in our first decision in this case:
Notwithstanding the relevance of [the precluded] testimony to significant evidence against Romero, I would also conclude the trial court’s error was harmless beyond a reasonable doubt. Haber’s testimony was brought exclusively to challenge the weight the jury could place on Powell’s opinion that only Romero’s gun could have fired the fatal shots. But there was other circumstantial evidence connecting Romero to the scene of the crime. [He] was both connected to a cell phone found at the scene and a truck observed leaving it. Given that the gun in question was found with the very person otherwise connected to the crime by two other items of evidence, the results of Powell’s testing rendered the proposition that another gun had fired the bullets unlikely in the extreme. Put another way, it would be an extraordinary coincidence if a weapon creating such similar markings as the murder weapon, but not involved in the murder, would happen to be found with Romero. Haber’s testimony—that Powell’s methodology could not scientifically exclude every other handgun in circulation as having fired the weapon— would not have altered that stark fact.
Romero I, 236 Ariz. 451, ¶ 69, 341 P.3d at 511 (Eckerstrom, C.J., specially concurring). The evidence in this case has not changed since those astute observations were made.
¶ 28 Third, viewed in its entirety, overwhelming evidence, albeit circumstantial, supported Romero’s conviction, including DNA linking him to the gun, as well as to the cell phone found on the ground beside the victim; a billing address also linking him to the phone; eye-witness testimony that immediately after the shots were fired one of the murder party said, “damn it, Joe, or something to that effect,” using Romero’s first name; and Romero’s possession and, upon being followed by police, immediate disposal *513of a pistol of the same caliber used in the shooting—another highly significant link, even if the match between the gun and the shell casings at the scene were not confirmed. Thus, as suggested in my colleague’s previous concurrence, although the weight of one expert’s testimony is disputed, “[t]he other evidence, points with unerring consistency to one inarguable conclusion.” Bible, 175 Ariz. at 588, 858 P.2d at 1191.
¶ 29 The majority reasons that the evidence only linked Romero to the scene, where at least two people were present, the jury was not instracted on accomplice liability, and the match of the gun to the shell casings was the strongest evidence Romero was the one who pulled the trigger. However, King, cited by the majority, does not say or suggest a jury cannot find guilt as an accomplice unless so instructed, 226 Ariz. 253, ¶¶ 14-20, 245 P.3d at 943-44, and a jury is not ordinarily required to state the basis for its verdict, cf. State v. Hansen, 237 Ariz. 61, ¶¶ 20, 22, 345 P.3d 116, 123 (App.2015) (noting courts do not second-guess jury’s verdicts nor inquire into deliberative process; “[a] court must simply accept the verdicts without probing into the jurors’ thought processes or demanding adherence to its instructions”).
¶ 30 Here, a question from the jury during its deliberations asked if Romero must have “pulled the trigger” to be convicted, to which the trial court only referred the jury to the instructions provided. The court later noted it had omitted an accomplice instruction through oversight. See State v. Rhymes, 129 Ariz. 56, 60, 628 P.2d 939, 943 (1981) (accomplice to murder liable under A.R.S. § 13-303(A)(3) for substantive crime); State v. Rios, 217 Ariz. 249, ¶ 10, 172 P.3d 844, 846 (App.2007) (accomplice with requisite mental state “considered as liable as if he had personally committed the offense”).
¶ 31 Finally, whether or not Romero might have been convicted as an accomplice, once it was established that he was present for the murder, additional evidence demonstrated his role as a shooter: an eyewitness reported that “both” people he had seen were shooting at the victim, and it is manifestly evident that Romero did not just wait in the truck because his cell phone was found on the ground beside the victim—Romero surely did not throw it out the window at him, either before or after he was shot. None of these facts or the other circumstantial evidence has any connection to the reliability of the firearms testing, but all go to the “one inarguable conclusion,” id. arrived at by the jury, that Romero was responsible for the murder.
¶ 32 Accordingly, applying our supreme court’s well-established standards for harmless error to the issue and evidence on remand, I would conclude beyond a reasonable doubt that the exclusion of Romero’s proffered expert “had no influence on the jury’s judgment,” Bible, 175 Ariz. at 588, 858 P.2d 1152; “the guilty verdict actually rendered in this trial was surely unattributable to the error,” Anthony, 218 Ariz. 439, ¶ 39, 189 P.3d at 373; and for those reasons would uphold Romero’s murder conviction.