concurring, in part, and dissenting, in part.
I.
I concur in Part I and Part 11(A) of the opinion and in the holding that the trial judge erred by excluding two journals in which the older girl chronicled her sexual activities. This evidence was probative of Cairns’s claim “that the complaining witness had a motive to fabricate the charges,” Code § 18.2-*29767.7, and was relevant as tending to prove for impeachment purposes the claim of fabrication. I disagree, however, with the conclusion that the error in excluding the evidence was harmless and, therefore, I dissent from Part 11(B).
The Supreme Court of Virginia has “adopt[ed] the Kotteakos [v. United States, 328 U.S. 750, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946),] harmless-error test” for measuring error under Code § 8.01-678. Clay v. Commonwealth, 262 Va. 253, 260, 546 S.E.2d 728, 732 (2001). Applying that test, the United States Supreme Court recently held that “the principle of Kotteakos [means] that when an error’s natural effect is to prejudice substantial rights and the court is in grave doubt about the harmlessness of that error, the error must be treated as if it had a ‘substantial and injurious effect’ on the verdict.” O’Neal v. McAninch, 513 U.S. 432, 444, 115 S.Ct. 992, 998, 130 L.Ed.2d 947 (1995) (citing Kotteakos, 328 U.S. at 764-65, 776, 66 S.Ct. at 1247-48, 1253). Moreover, when a trial error has been shown on direct appeal from a conviction, the government bears the burden of proving harmlessness under this standard. See O’Neal, 513 U.S. at 437, 115 S.Ct. at 995. Indeed, the Supreme Court of Virginia has held that “error will be presumed to be prejudicial unless it plainly appears that it could not have affected the result.” Caldwell v. Commonwealth, 221 Va. 291, 296, 269 S.E.2d 811, 814 (1980).
Although I disagree with the majority opinion’s characterization that “overwhelming evidence ... supports] the convictions,” I believe it is important to note, as the United States Supreme Court has observed, that an “emphasis and perhaps overemphasis, upon the [concept] of ‘overwhelming evidence,’ ” has the effect of clouding the relevant question “ ‘whether there is a reasonable possibility that the evidence complained of might have contributed to the conviction.’ ” Chapman v. California, 386 U.S. 18, 23, 87 S.Ct. 824, 827, 17 L.Ed.2d 705 (1967) (footnote and citations omitted). Indeed, the principle is well established that a harmless error analysis is entirely distinct from a sufficiency of the evidence analysis. “The inquiry cannot be merely whether there was enough to *298support the result, apart from the phase affected by the error. It is rather, even so, whether the error itself had substantial influence. If so, or if one is left in grave doubt, the conviction cannot stand.” Kotteakos, 328 U.S. at 765, 66 S.Ct. at 1248. Consistent with these principles, the Supreme Court of Virginia has held that even if “the other evidence amply supports the ... verdicts, [error is not harmless when] the disputed [evidence] may well have affected the ... decision.” Cartera v. Commonwealth, 219 Va. 516, 519, 248 S.E.2d 784, 786 (1978). See also Hooker v. Commonwealth, 14 Va.App. 454, 458, 418 S.E.2d 343, 345 (1992) (holding that “a harmless error analysis ... [is not] simply a sufficiency of the evidence analysis”).
The record contains statements Cairns made during his voluntary interview with the police before they arrested him. At all times Cairns denied he engaged in sexual contact with his stepdaughter and his daughter. Although Cairns admitted playing poker with the girls, he denied playing strip poker with them. He told the police that both girls were angry with him because he “grounded” them after he “got up in the middle of the night, and [found] seven guys ... in their room.” He explained that the girls were still angry because he removed their privileges for “a long, extended” period; he “grounded [one] ‘till Halloween and this has been in effect since last month,” and he “grounded [the other one] ‘till the end of the year ... [so that she] can’t go out the house and she can’t have no friends over.”
Cairns did not tell the police that he and his wife intentionally exposed his daughters to their marital sexual activity.
[Detective]: Has your daughters ever been in the room when you and your wife have had sex?
[Cairns]: Probably so.
[Detective]: Okay.
[Cairns]: I mean, they fall asleep in the room, and we go in the other room, but still, they’re....
*299[Detective]: Okay. Was there a night when [the older girl] was ... was in your bed when you and your wife had sex?
I mean, she could have just been layin’ there.
[Cairns]: I don’t know. Truthfully, I don’t know.
*:< # % # % *
[Detective]: Um, you and your wife have sex?
[Cairns]: Um, do we?
[Detective]: No, with [the older girl] in there.
[Cairns]: I ain’t sayin’ it might not have ever happened.
[Detective]: Right.
[Cairns]: I’m sayin’ the best what I can remember, I’m not sure, so I won’t answer.
Cairns did admit, however, that he and his wife were in the room when the older girl “performed oral sex on her boyfriend,” but he explained this is a lapse in judgment that occurred because he and his wife were intoxicated. Although Cairns’s conduct in this instance was despicable, it was not one of the offenses charged and it pales in comparison to the charges in the indictments.
Cairns claimed that the girls had a motive to fabricate their accusations and that he needed to introduce the journals to establish (1) the older girl was having sexual relations with a substantial number of people during the period at issue, (2) she explicitly described those sexual encounters and did not name Cairns in the journal as one of those sexual partners, (3) she was untruthful when she told a detective the journal would confirm her allegation, and (4) she had a cause to be angry and to fabricate her accusation after Cairns discovered five or six young men secretly visiting her bedroom late at night.
The evidence clearly proved Cairns and his wife acted inappropriately in the presence of their children and may have contributed to the older girl’s sexually inappropriate conduct. The evidence that he committed the charged offenses, however, was based solely on the testimony of the children and was unconfirmed by any other evidence. Significantly, the girls did not dispute that Cairns “grounded” them because five or *300six boys were in the girls’ bedroom at night when all the family had retired to their beds for the night. Moreover, the journals tend to establish the truth of Cairns’s claim that the girls entertained boys in the bedroom late at night and to provide a basis upon which a trier of fact might tend to credit Cairns’s claim of a retaliatory motive by the girls.
The journals establish that by the time the older girl had reached the age of fourteen, she had engaged in various sex acts, including fellatio, cunnilingus, and sexual intercourse, on ninety-six different occasions with twenty-four partners, both male and female. Those journals mention her parents but do not identify them as engaging in any sexual activity with her. Indeed, one entry records the older girl’s fear that her father would awaken in the night and discover her having sexual intercourse in her bedroom with a boy. The admission of the evidence could have had its intended effect of impeaching the testimony of the older girl and proving her complaint to be a fabrication. Thus, the evidence went to the heart of Cairns’s defense and its exclusion prejudiced his right to cross-examine witnesses and to prove evidence in his defense. We cannot say its exclusion played no significant part in the verdict.
“ ‘[A] fair trial on the merits and substantial justice’ are not achieved if an error at trial has affected the verdict.” Lavinder v. Commonwealth, 12 Va.App. 1003, 1005, 407 S.E.2d 910, 911 (1991) (quoting Code § 8.01-678). Thus, we have held that “in determining if an error is harmless, a reviewing court ... determines ... whether as a matter of law, this decision of the fact finder was affected by the error.” Id. at 1006, 407 S.E.2d at 911. I believe this record proves as a matter of law that the judge’s decision was affected by the error.
Furthermore, I disagree that the trial judge ruled, as a matter of fact, that the guilty verdict would have been the same if the excluded ■ evidence had been admitted. At the sentencing hearing, when the issue before the judge was the appropriate sentence to impose, Cairns’s attorney argued that the evidence did not prove aggravating circumstances of force and that the journals indicated in mitigation that the older girl *301“has had significant sexual contact.” In that context and in response to this argument, the trial judge reminded Cairns’s attorney that the journals had been ruled inadmissible and “that had it been considered as part of the evidence in the case, it wouldn’t have caused me to change my mind or rule any differently.” Because this discussion concerned the issue of sentencing, Cairns’s attorney reminded the judge that “the court has received victim impact statements which address exactly that.” In short, the judge’s comment, which was made at sentencing and in response to the argument that the older girl’s conduct was a mitigating factor, should not be construed to suggest that the trial judge was commenting upon the impact the excluded evidence potentially might have had at the guilt phase of the trial. Viewed in the context of the discussion about sentencing factors, the judge was not suggesting that the guilt verdict would have been the same if the improperly excluded evidence had been admitted at the guilt phase and had had its proffered effect.
The exclusion of the journals affects a substantial credibility determination and belies the majority’s determination that the exclusion was harmless. A trier of fact might normally be inclined to disbelieve a fourteen-year-old child would have an independent basis to be familiar with fellatio, cunnilingus, and sexual intercourse as described in her testimony and, as a consequence, to disbelieve she could have fabricated the detailed sexual activities she says her parents committed upon her. Moreover, it is not unreasonable to believe that a trier of fact, after learning of the child’s ninety-four episodes of prior sexual contact with twenty-four different partners, might have concluded as a matter of fact that the girl’s own conduct tended to render her less susceptible to intimidation. The erroneous exclusion of this evidence also substantially negated Cairns’s ability to prove his claim of fabrication. Thus, the exclusion raises a grave concern .that the probative value the trier of fact gave the Commonwealth’s evidence would have been qualitatively diminished if the journals had been admitted. This error was not insignificant; it had a “substantial *302and injurious effect” on the trier of fact’s credibility determination and verdict.
While other evidence amply supports the verdicts, “[ojther evidence of a disputed fact, standing alone, does not establish that an error is harmless.” Hooker, 14 Va.App. at 458, 418 S.E.2d at 345. I would hold, as a matter of law, the disputed evidence affected the trier of fact’s credibility decisions and, thus, the verdicts. Accordingly, I conclude that the error in excluding evidence was prejudicial to Cairns and not harmless.15
II.
I also dissent from Part III of the opinion.
The United States Supreme Court announced a prophylactic rule in North Carolina v. Pearce, 395 U.S. 711, 89 S.Ct. 2072, 23 L.Ed.2d 656 (1969), that “placed a limitation on the power of a sentencing authority to increase a sentence after reconviction following a new trial.” Texas v. McCullough, 475 U.S. 134, 137, 106 S.Ct. 976, 978, 89 L.Ed.2d 104 (1986). The Court held as follows:
Due process of law ... requires that vindictiveness against a defendant for having successfully attacked his first conviction must play no part in the sentence he receives after a new trial. And since the fear of such vindictiveness *303may unconstitutionally deter a defendant’s exercise of the right to appeal or collaterally attack his first conviction, due process also requires that a defendant be freed of apprehension of such a retaliatory motivation on the part of the sentencing judge.
In order to assure the absence of such a motivation, we have concluded that whenever a judge imposes a more severe sentence upon a defendant after a new trial, the reasons for his doing so must affirmatively appear. Those reasons must be based upon objective information concerning identifiable conduct on the part of the defendant occurring after the time of the original sentencing proceeding. And the factual data upon which the increased sentence is based must be made part of the record, so that the constitutional legitimacy of the increased sentence may be fully reviewed on appeal.
Pearce, 395 U.S. at 725-26, 89 S.Ct. at 2080-81 (emphasis added).
This ruling clearly denotes that the Pearce presumption “operate[s] in the absence of any proof of improper motive.” United States v. Goodwin, 457 U.S. 368, 373, 102 S.Ct. 2485, 2488, 73 L.Ed.2d 74 (1982). Moreover, the Supreme Court has ruled that “the reach of Pearce is best captured by [its] statement in United States v. Goodwin, 457 U.S., at 374[, 102 S.Ct., at 2489] ... [that,] ‘[i]n sum, the Court [in Pearce ] applied a presumption of vindictiveness, which may be overcome only by objective information ... justifying the increased sentence.’ ” McCullough, 475 U.S. at 142, 106 S.Ct. at 981.
The Supreme Court has not overruled Pearce. In McCullough, the Supreme Court explained that “[t]he facts of this case provide no basis for a presumption of vindictiveness.” 475 U.S. at 138, 106 S.Ct. at 979. There, the retrial occurred because the trial judge who presided at the initial jury trial “granted McCullough’s motion for a new trial on the basis of prosecutorial misconduct.” 475 U.S. at 136, 106 S.Ct. at 977. In imposing a higher sentence on retrial than the twenty years *304fixed by the jury at the first trial, the judge “found that ... she relied on new evidence about the murder that was not presented at the first trial and hence never made known to the sentencing jury [at the first trial].” Id. In addition, the judge “explained that she learned for the first time on retrial McCullough had been released from prison only four months before the later crime had been committed.” Id. In view of this record, the Supreme Court reasoned as follows:
In contrast to Pearce, McCullough’s second trial came about because the trial judge herself concluded that the prosecutor’s misconduct required it. Granting McCullough’s motion for a new trial hardly suggests any vindictiveness on the part of the judge towards him. “[U]nlike the judge who has been reversed,” the trial judge here had “no motivation to engage in self-vindication.” In such circumstances, there is also no justifiable concern about “institutional interests that might occasion higher sentences by a judge desirous of discouraging what he regards as meritless appeals.” In granting McCullough’s new trial motion, [the trial judge] went on record as agreeing that his “claims” had merit. Presuming vindictiveness on this basis alone would be tantamount to presuming that a judge will be vindictive towards a defendant merely because he seeks an acquittal.
Id. at 138-39, 106 S.Ct. at 979 (citations omitted). In short, based on the record in McCullough, the Supreme Court held that “not even ‘apprehension of ... a retaliatory motivation on the part of the sentencing judge’ could be present.” 475 U.S. at 139, 106 S.Ct. at 979 (citation omitted).
Significantly, the Court also ruled that “the second sentencer provide[d] an on-the-record, wholly logical, nonvindictive reason for the sentence.” Id. at 140, 106 S.Ct. at 980. Indeed, the resentencing judge “made findings of fact as to why the sentence was longer ... [and] found that ... she relied on new evidence ... that was not presented at the first trial.” 475 U.S. at 136, 106 S.Ct. at 977-78. Noting these circumstances, the Supreme Court “read Pearce to require no more.” Id. at 140, 106 S.Ct. at 980. The Court held “that the careful explanation by the trial judge for the sentence imposed [in *305McCullough ] fits well within [its] prior holdings.” Id. at 143, 106 S.Ct. at 981.
The Supreme Court, thus, has indicated in McCullough and its other decisions since Pearce that, “in each case, we [must] look to the need, under the circumstances, to ‘guard against vindictiveness in the resentencing process.’ ” McCullough, 475 U.S. at 138, 106 S.Ct. at 979 (citation omitted). For example, the Court has analyzed the circumstances and the need to guard against vindictiveness when a jury resentences following a new trial. See Chaffin v. Stynchcombe, 412 U.S. 17, 93 S.Ct. 1977, 36 L.Ed.2d 714 (1973). There, the Court pointed to at least two circumstances that do not exist when a jury sentences. First, the jury typically is unaware of the prior sentence, which is the “first prerequisite for the imposition of a retaliatory penalty.” Id. at 26, 93 S.Ct. at 1982. Second, the jury “is unlikely to be sensitive to the institutional interests that might occasion higher sentences by a judge desirous of discouraging ... meritless appeals.” Id. at 27, 93 S.Ct. at 1983.
I believe Cairns’s case is one that presents circumstances “where [the Pearce ] ‘objectives are ... most efficaciously served.’ ” McCullough, 475 U.S. at 138, 106 S.Ct. at 979. The sentence was dramatically increased by the trial judge following the second trial. Although the sentencing judge noted that he was basing his sentence on Cairns’s egregious conduct, the record indicates the evidence at the second trial was not substantially different than the evidence at the first trial. Moreover, the trial judge acknowledged that the sentence he imposed was not based on new information concerning conduct after the first trial. Thus, this case is dramatically different than McCullough, where “the second sentencer provide[d] an on-the-record, wholly logical nonvindictive reason for the sentence.” 475 U.S. at 140, 106 S.Ct. at 980. See also Macomber v. Hannigan, 15 F.3d 155, 156 (10th Cir.1994) (noting that several federal “circuits have determined that no presumption' of vindictiveness is present when the second sentence was delivered by a different judge and the record indicates nonvindictive reasons supporting the harsher sentence”).
*306The record indicates that, when the trial judge sentenced Cairns, he “arrived at the sentences based on what [he] thought was the appropriate disposition in each of those cases.” The judge, however, was aware of the previous reversal and had some knowledge of the prior sentence. He “knew generally, not the specific number of years, but ... had a general idea of the range of punishment imposed but did not know the specific counts.” Indeed, the record reflects that the sentences contain some stark similarities. Both judges imposed a fifty-year sentence for the rape conviction, which has a statutory maximum of life imprisonment. Code § 18.2-61. Unlike the first judge, who suspended thirty years of the rape sentence, the resentencing judge suspended none. Both judges imposed an active sentence of ten years on each of the sodomy convictions, which also has a statutory maximum of life imprisonment for each conviction. Code § 18.2-67.1. Based on different sentences, both judges imposed no active sentence for the pornography offense, which has a statutory maximum of ten years imprisonment. The two sentences appear to reflect a similar methodology. The increase in the sentence on retrial, however, is unexplained.
It bears repeating, as the Court explained in Pearce, that the due process concern requires the “defendant be freed of apprehension of ... a retaliatory motivation” by an affirmative showing on the record of objective information to explain the increased sentence. 395 U.S. at 726, 89 S.Ct. at 2081. In this case, the institutional interests that might intrude on a judge’s resentencing after a successful appeal remain unnegated and the due process concern expressed in Pearce is unsatisfied. Therefore, I would hold that the Pearce presumption applies in this case and that the record does not rebut it. The record contains no “objective information” to justify the increased sentence “beyond the naked power to impose it.” 395 U.S. at 726, 89 S.Ct. at 2081.
For these reasons, I would reverse the convictions and remand for a new trial.
. The exclusion, of this evidence also could be said to violate Cairns’s right to confront a witness against him as protected by the Confrontation Clause of the Sixth Amendment. "[A] major reason underlying the constitutional confrontation rule is to give a defendant charged with crime an opportunity to cross-examine the witnesses against him.” Pointer v. Texas, 380 U.S. 400, 406-07, 85 S.Ct. 1065, 1069, 13 L.Ed.2d 923 (1965). As the United States Supreme Court has noted, a trial error may "rise to the level of a constitutional violation ... if it results in a prejudice so great as to deny a defendant his Fifth Amendment right to a fair trial.” United States v. Lane, 474 U.S. 438, 446 n. 8, 106 S.Ct. 725, 730 n. 8, 88 L.Ed.2d 814 (1986).
Because, in my view, the Commonwealth cannot even establish the lower standard of harmless error under the Kotteakos analysis, I do not address the more onerous standard required by Chapman for constitutional error.