Because Rogers was prejudiced by the circuit court's erroneous ruling in refusing to allow Rogers to impeach LW with her prior theft conviction, I dissent from the majority opinion and would reverse and remand the matter for a new trial. Here, the majority holds that the circuit court's ruling was harmless because the evidence of guilt was overwhelming:
[T]he circuit court's error fits the harmless-error mold.... Here, the evidence that Rogers raped LW is overwhelming. LW testified in specific detail to multiple occurrences of rape. TB testified that she observed a video of Rogers engaged in sex with LW. Moreover, all four victims described what this court emphasized in Kelley when affirming for harmless error as "remarkably similar conduct on the part of [the defendant]." Kelley , 2009 Ark. 389, at 20, 327 S.W.3d at 383. The girls similarly described their sexual encounters with Rogers, including the color of the condom, the dead-end road where he took two of them, and Rogers's suicide threats.
Additionally, if Rogers experienced any prejudice, it was slight. Whether an error is slight hinges on the degree to which the defendant was prejudiced. Id. The proposed impeachment testimony of LW, unlike the victim's testimony in Scamardo , did not directly relate to the allegation at hand. Thus, although there was error, it was harmless.
This position is not supported by the record. The State's case in this matter hinged upon one thing-the credibility of the victims, including LW. There is no other evidence in this case. Simply put, it was a "he said she said" battle between Rogers, his family and neighbors on one side and the victims and their mother on the other. Yet, the majority holds that the evidence is overwhelming as to Rogers's guilt even though Rogers was erroneously prevented from impeaching LW with her prior theft conviction. The impeachment evidence Rogers sought to introduce concerning LW directly relates to her credibility and LW's credibility was critical to the question before the jury. Therefore, to say that either the error was slight or that the evidence was overwhelming is simply wrong.
Further, the error cannot be slight because the circuit court's ruling tainted the entire trial. The circumstances of this case and the testimony of the victims are intertwined as is demonstrated by the record. Rogers's defense was based on the premise that the four girls, along with Bryant, had conspired to concoct their rape allegations against him. He claimed that they had hatched the plan and made up the charges to get even with him for moving out of the home and ending the relationship with Bryant. In the State's closing argument, the State acknowledged the defense's theory of the case when the prosecutor stated: "The Defendant wants you to believe the victims colluded and planned their stories in order to frame him because [Bryant] was mad." Because all of the girls' allegations were intertwined, the circuit court's error in excluding the evidence cannot be considered slight, this particularly true since the only evidence to support Rogers's conviction was the victims' testimony. Thus, the victims' credibility was presumably a major consideration for the jury. See Scamardo v. State , 2013 Ark. 163, 426 S.W.3d 900. Therefore, the error was not harmless. Finally, I must note that in holding the error here was harmless, the majority seems to expand the breadth of harmless error and avoid addressing the constitutional violations. The "harmless *395error doctrine has been called a '[c]onstitutional [s]neak [t]hief' for the way it enables courts to take away with one hand what they have given with the other.... Recognizing that harmless error is inexorably tied up with the way constitutional rights are defined makes visible how judges can deploy harmless error doctrine to expand, contract, or even eliminate constitutional rights. And thus it avoids sweeping difficult normative questions under the rug." Daniel Epps, Harmless Errors and Substantial Rights , 131 Harv. L. Rev. 2117, 2125 (2018).
Kemp, C.J., and Hart, J., join.
Josephine Linker Hart, Justice, dissenting.
The majority's decision to excuse the circuit court's erroneous evidentiary ruling as "harmless" is not grounded in either law or fact. It ignores clear precedent from the Supreme Court of the United States that this court is required to follow. Further, the decision is unsupported by the very cases that it claims to rely on, and it fails to consider all of the evidence in arriving at its remarkable conclusion that the evidence was "overwhelming." Mr. Rogers was denied a fair trial and proper consideration of his appeal.
Denying Mr. Rogers his right to use LW's conviction for a crime involving dishonesty impaired his rights under the Confrontation Clause to fully challenge her credibility.2 The State presented no corroborating physical or scientific evidence. Accordingly, the issue of whether a complaining witness was credible was of pivotal importance in this case.
I am mindful that the Supreme Court has not foreclosed finding harmless error in cases such as the one before us where a criminal defendant was denied his full right of cross-examination. In Delaware v. Van Arsdall , 475 U.S. 673, 106 S.Ct. 1431, 89 L.Ed.2d 674 (1986), the Supreme Court rejected a per se reversal rule in cases where a criminal defendant was denied an opportunity to impeach a witness. The Van Arsdall Court held that a reviewing court could find that the error was harmless beyond a reasonable doubt based on a review of factors that include
the importance of the witness' testimony in the prosecution's case, whether the testimony was cumulative, the presence or absence of evidence corroborating or contradicting the testimony of the witness on material points, the extent of cross-examination otherwise permitted, and, of course, the overall strength of the prosecution's case.
475 U.S. at 684, 106 S.Ct. 1431. However, the majority has not undertaken this analysis. This court is bound by the law to follow and apply Supreme Court precedent.3
Had the majority faithfully undertaken its duty to apply Supreme Court precedent, it could only have concluded that LW's testimony was crucially important-without *396her testimony, the State could not have presented evidence regarding all the elements of the charges that involved Mr. Rogers's alleged sexual conduct with her. As previously mentioned, there was no physical or scientific evidence to corroborate LW's testimony.
Further, there were significant conflicts and inconsistencies in the testimony. For instance, TB testified that she saw a video of Mr. Rogers having sex with LW, yet LW did not mention any videotaping. Finally, regarding the strength of the State's case, the testimony was weak, uncorroborated, lacking in significant detail, and substantially different from the reports that the witnesses made to police. Significantly, the jury acquitted Mr. Rogers of the charges that involved TB; so obviously, the jury found her testimony not credible. Conversely, the defense case presented consistent proof that the alleged victims' mother, Tia Bryant, had orchestrated the accusations to punish Mr. Rogers for leaving her and breaking his promise to help support her and the girls.
Not surprisingly, none of the cases cited by the majority actually support its decision. In Scamardo v. State , 2013 Ark. 163, 426 S.W.3d 900, this court held that an erroneous decision to exclude impeachment evidence was not harmless when "the main evidence supporting [a] conviction was the victim's testimony," which makes the victim's credibility "presumably a major consideration for the jury." Accordingly, the error cannot be considered "slight." Id. at 9, 426 S.W.3d at 905.
Likewise, in Winfrey v. State , 293 Ark. 342, 738 S.W.2d 391 (1987), this court refused to find harmless a circuit court's erroneous decision to prohibit a criminal defendant from using for cross-examination prior inconsistent statements by the victims. The statements were contained in reports generated by members of the prosecution's staff after they had interviewed the alleged victims, and the circuit court excluded the evidence as "work product." The Winfrey court held that the circuit court erred because prosecutor work product was not privileged material that could be excluded from being used at trial and that the error was not harmless because the ruling impaired the defendant's ability to cross-examine the victims, which violated the Confrontation Clause.
Citing Delaware v. Fensterer , 474 U.S. 15, 106 S.Ct. 292, 88 L.Ed.2d 15 (1985), Pennsylvania v. Ritchie , 480 U.S. 39, 107 S.Ct. 989, 94 L.Ed.2d 40 (1987), and Davis v. Alaska , 415 U.S. 308, 94 S.Ct. 1105, 39 L.Ed.2d 347 (1974), the Winfrey court held that an exclusion of evidence that is useful in cross-examination is an error of constitutional dimension. Winfrey , 293 Ark. at 343-44, 738 S.W.2d at 393-94. It noted further that "the main and essential purpose of confrontation is to secure for the opponent the opportunity of cross-examination." Id. at 344, 738 S.W.2d at 394 (quoting Davis , 415 U.S. at 315-316, 94 S.Ct. 1105 ).
Winfrey , like the case before us, turned on whether the jury found credible the testimony of an alleged victim. The Winfrey court analyzed the harmless-error issue as follows:
Here, the State could not make its case without the testimony of the victims. Their credibility was the key factor in the determinations of guilt. Yet, the trial court, by ruling that appellant's attorney was prohibited from using the prior inconsistent statements for impeachment purposes, prevented appellant from questioning the witnesses' credibility. Accordingly, we cannot say the error was harmless beyond a reasonable doubt.
293 Ark. at 350-51, 738 S.W.2d at 395. Here, as in Winfrey , the State could not *397make its case without LW's testimony. Denying Mr. Rogers the use of impeachment evidence was therefore not harmless.
Among the other cases cited by the majority, Pigg , Buford , Johnston , and Kelley , only Pigg involved the exclusion that a criminal defendant sought to have admitted. In Pigg , the defendant challenged a pretrial ruling under the rape-shield statute, so the issue is not analogous. Id. However, it is worth noting that the Pigg court did illustrate what might be considered "overwhelming" evidence:
[The victim's] testimony, in detail, revealed that Pigg had sexual relations with her for five to six years beginning when she was eleven years old. Pigg's daughter and W.S. witnessed some of the sexual activity, which they described in their testimony. In addition, expert testimony disclosed that [the victim] had a deep notch in her hymen, which was suggestive of sexual abuse or penetrating trauma. Moreover, the jury heard the testimony of Pigg's niece who said that Pigg had molested her when she was eight years old.
2014 Ark. 433, at 4-5, 444 S.W.3d at 865-66.
For obvious reasons, the case before us differs substantially from Pigg. As previously mentioned, Pigg involved the rape-shield statute, and the case before us is simply an obvious evidentiary error. The testimony was clear and directly corroborated in Pigg , while in the case before us the evidence is conflicting and uncorroborated. Moreover, physical evidence corroborated the rape allegations in Pigg. No physical evidence was introduced in Mr. Rogers's trial. In fact, the State's witnesses, including Detective Ashley Noel and Arkansas Children's Hospital physician Dr. Kristen Long admitted that they did not even attempt to collect physical evidence.
None of the remaining cases cited by the majority concerned the exclusion of impeachment evidence or the denial of the right of cross-examination regarding this evidence. Johnston , Buford , and Kelley involved evidence admitted in the State's case-in-chief that was unnecessary to prove the specific elements of the charges. The issue in those cases was whether erroneously admitted evidence was so prejudicial as to require a new trial.
In Johnston , the victim was raped by her father. The evidentiary error at issue was the admission of pornographic images found on Johnson's laptop computer. While all of the images were obtained from incest-themed websites,4 not all of the females in the images were actually minors. As in Pigg , the court did not hold that there was a clear evidentiary error, but concluded that the evidence against Johnson was "overwhelming." Id. In addition to the victim's "detailed testimony recounting the many years of rapes by her father," Johnson's semen that was found on the crotch area of five pairs of the victim's underwear, including the pair that she had worn home following the visitation, during which Johnson had raped her three times, a vaginal swab from the sexual-assault exam performed the next day that tested positive for sperm, and medical testimony that the child-victim had a "major *398tear" in her hymen, which was consistent with sexual abuse involving penetration, 2014 Ark. 433, at 7-8, 431 S.W.3d at 899. Conversely, in Mr. Rogers's trial, the State introduced no physical or scientific evidence whatsoever.
Buford involved a circuit court's erroneous decision to allow an "expert" to opine as to whether the victim was telling the truth. The Buford court found that the error was harmless because there was "overwhelming" evidence; the victim provided "graphic details" of the rape, which was independently corroborated by three witnesses who observed the rape through a window, and testimony from the arresting officer that when the defendant was found with the victim in his vehicle, he gave a false name and falsely identified the victim as his nephew. In the case before us, the only "graphic detail" that the majority cites as being remarkable is that Rogers allegedly used a gold-colored condom. While I must defer to the majority's conclusion that gold is an unusual color for a condom, it is not a fact worthy of judicial notice.
Finally, in Kelley , during the guilt phase of the trial, the circuit court erroneously admitted certified copies of out-of-state convictions for sexual indecency with a child. "Overwhelming" evidence consisted of two victims who testified to "remarkably similar conduct," and a recording of a phone call made at the jail in which the defendant tried to arrange for one of the victims to be unavailable to testify at trial. In the case before us, testimony by the alleged victims was anything but "remarkably similar." For instance, all the alleged victims admitted that their statements to police were different than their trial testimony. In Mr. Rogers's case, the testimony was remarkably inconsistent. The alleged victims' trial testimony did not even match the statements that they gave to police. Additionally, the Kelley court found that the error was "slight" because the certified copies of the convictions would have been admissible during the penalty phase.
Not only is the majority's decision unsupported by the law, but also the trial transcript supports even less the majority's finding that the evidence was "overwhelming." As noted previously, the State introduced no physical or scientific evidence. Furthermore, the victims' testimony included few details, and accounts differed from the statements that they gave to police.
Unlike the victims in Kelley, whose testimony was "remarkably similar," the alleged victims in Mr. Rogers's case did not corroborate each other's stories. For example, the majority finds significance in TB's testimony that she viewed a video of Rogers having sex with LW. However, LW did not testify about being videoed. It is also noteworthy that the State did not produce the video at trial. Similarly, TB claimed that one morning, she heard a "clapping noise" and found Rogers having sex with MiB. MiB did not recall the incident in her testimony. Finally, it is significant that Rogers was acquitted of the rape of TB. If TB's testimony was worthy of belief, the jury necessarily should have convicted Rogers of raping her.
The other witnesses' credibility was also impugned. MiB admitted in court that she "would lie whenever she felt like it." LW's testimony that Rogers was harassing her at work was contradicted by a coworker, Cleo Hackett, who testified that Rogers helped LW get her job and approached LW at work only to use her employee-discount card. Nonetheless, the jury was not allowed to know that LW had committed a crime involving dishonesty.
Further, it must be noted that most of the alleged sexual activity supposedly took place in a 1000-square-foot house while the *399alleged victims' mother occupied a room just a few feet away. Logic suggests that discovery of this activity by Tia Bryant had to approach certainty. In fact, contrary to the representations made by the majority, this circumstance was a substantial part of Mr. Rogers's factual predicate for his challenge to the sufficiency of the evidence. He asserted that because of this circumstance, the victims' testimony had to be rejected on appeal because it was "inherently improbable, physically impossible, or so clearly unbelievable" that reasonable minds could not accept it as worthy of belief.5 While this argument may not have compelled the overturning of a jury verdict, it remains part of the evidence that must be considered before concluding that the evidence of guilt was overwhelming.
The evidence adduced at trial also included the testimony of Tia Bryant. The defense's theory of the case was that Bryant concocted the allegations against Mr. Rogers because he had left her for another woman. It was proved that Bryant went to the North Little Rock police with her allegations only after she caught Mr. Rogers "hugged up" with another woman at a neighborhood barbecue. Further, much of Bryant's testimony was exposed as half-truths, evasions, and likely outright fabrications.
Mr. Rogers also testified and denied all of the allegations. His testimony was corroborated by neighbors, friends, and relatives. Bryant's testimony was not similarly corroborated and was contradicted in several important respects. Accordingly, it is disingenuous to call the evidence "overwhelming."
It is apparent that the majority has confounded the sufficiency-of-the-evidence standard of review with what is required to establish "overwhelming" evidence in a harmless-error analysis. In reviewing the sufficiency of the evidence, we do not consider evidence that does not support the verdict. Conway v. State , 2016 Ark. 7, 479 S.W.3d 1. Accordingly, under the majority's reasoning, if there is substantial evidence to support a verdict, all evidentiary errors would be harmless. However, suffice it to say that reviewing a case for "overwhelming" evidence necessarily requires consideration of all the evidence, even evidence that does not support the verdict.
Our system of justice depends on adversary proceedings in which both sides must be allowed to test the evidence in accordance with the limits set by our rules of evidence. When one side is prevented from fully presenting its case, we cannot have confidence in the outcome of the trial. This is not "slight" error. It is not disputed that the circuit court failed to follow Rule 609 of the Arkansas Rules of Evidence, and the Supreme Court of the United States has given this court clear guidance on how to conduct harmless-error review. The majority has utterly failed to follow the Supreme Court's guidance. Mr. Rogers deserves a new trial.
I respectfully dissent.
The majority's footnote to the effect that Mr. Rogers argument that he was denied his right to fully cross-examine LW did not apprise the circuit court-or this court of the constitutional dimension of this error is puzzling. The right to cross-examine a witness is synonymous with the most important guarantees found in the Confrontation Clause of the Sixth Amendment. Davis v. Alaska , 415 U.S. 308, 94 S.Ct. 1105, 39 L.Ed.2d 347 (1974). Furthermore, the majority's reliance on Roston v. State , 362 Ark. 408, 208 S.W.3d 759 (2005) as support for this proposition is dubious at best. The Roston court merely held that a hearsay objection does not preserve a Confrontation-Clause argument for appeal. Id.
The majority should be aware of the requirements of Van Arsdall as they are recited in Roston v. State , 362 Ark. 408, 208 S.W.3d 759 (2005), a case relied on by the majority.
According to the arresting officer, all of the approximately 2200 images on Johnson's computer had been downloaded in a nineteen-minute time span on June 2, 2011. Police prepared a report containing about 188 of these images watermarked with names such as "whole family incest," "dads permindaughters," "home incest videos," "daughterdestruction.com," "tryincest.com," and "incestsexcite.net." The arresting officer conceded that the images did not necessarily feature underage females, although many of the actors were dressed or posed in such a manner as to appear young.
Interestingly, the majority has failed to address the sufficiency-of-the-evidence argument that Mr. Rogers actually makes. It has instead substituted its own straw-man argument, one that it can succinctly dispose of. However, because I believe that Mr. Rogers's argument is without merit, I accept the majority's disposition of this issue.