dissenting.
In this case, acknowledging the continuing validity of Bohnert v. State, 312 Md. 266, 539 A.2d 657 (1988)1 and, despite *714declaring it distinguishable2 in its application to the facts sub judice, the majority holds that the testimony of Edward Anderson, a critical witness for the State, was inadmissible. Anderson was permitted to testify in front of the jury, the trier of fact, that certain law enforcement officers, ie., a prosecutor with the Anne Arundel County State’s Attorney’s office and a police detective, testified at his sentence reduction hearing that “[he] was being truthful in bringing them this information [about] the petitioner’s involvement in the murders with which he was charged.” The petitioner challenged Anderson’s credibility by attempting to prove that the motivation for his testimony was leniency. Thus, the petitioner’s cross examination of Anderson concerned Anderson filing a motion for modification of sentence prior to the petitioner’s first trial, his bringing to the attention of the sentencing judge that he was a State’s witness in the petitioner’s trial, and the fact that a prosecutor and a police detective involved in the investigation of the murder with which the petitioner was charged were called to testify at Anderson’s modification hearing. Consequently, Anderson’s testimony in this regard was designed to, and did, bolster his credibility for the jury.
Having determined that Anderson’s testimony concerning the prosecutor and the police detective was inadmissible and, thus, that its admission was error, the majority then holds that the error was harmless. It reasons that the distinction it has *715drawn reduces the prejudicial impact of the evidence. 360 Md. at 679, 759 A.2d at 779. Then it asserts:
“Anderson made a self-serving statement to the effect that certain persons not present once affirmed, on an unknown basis, his truthfulness in making the statements he again made at trial. Such a statement, by a witness whose credibility is in question, is far less weighty than the expert in Bohnert, and its effect on the jury was likely to be insignificant. Moreover, it is implicit that the police believed Anderson or they would not have gone to bat for him at the hearing on his motion to reduce his sentence.”
Id. at 679, 759 A.2d at 779. I do not agree and, so, dissent.
Once again, I find myself engaged in what sometimes seems like a never-ending battle to protect the integrity and vitality of the harmless error rule. See my dissenting opinions in Jensen v. State, 355 Md. 692, 718, 736 A.2d 307, 320 (Md.1999); Evans v. State, 333 Md. 660, 711, 637 A.2d 117, 136 (1994); Bruno v. State, 332 Md. 673, 696, 632 A.2d 1192, 1204 (1993); Rubin v. State, 325 Md. 552, 591, 602 A.2d 677, 696 (1992). This is so despite the clarity of the statement of the rule in Dorsey v. State, 276 Md. 638, 350 A.2d 665 (1976), and the frequency with which it has been repeated with approval and thus reaffirmed, since. But as this case and those that precede it, to some of which I have just referred, demonstrate, to state the rule accurately is not necessarily to apply it appropriately.
The test of harmless error focuses on the effect of erroneously admitted, or excluded, evidence on the verdict rendered by the jury. As enunciated in Dorsey, the rule is:
“When an appellant, in a criminal case, establishes error, unless a reviewing court, upon its own independent review of the record, is able to declare a belief, beyond a reasonable doubt, that the error in no way influenced the verdict, such error cannot be deemed “harmless” and a reversal is mandated. Such reviewing court must thus be satisfied that there is no reasonable possibility that the evidence com*716plained of — whether erroneously admitted or excluded— may have contributed to the rendition of the guilty verdict.”
276 Md. at 659, 350 A.2d at 678 (footnote omitted). Once it has been determined that error was committed, reversal is required unless the error did not influence the verdict; the error is harmless only if it did not play any role in the jury’s verdict. The reviewing court must exclude that possibility “beyond a reasonable doubt.”
Moreover, an appellate court reviewing a trial court verdict must apply the harmless error rule consistently with its role; it should not take on the role of the trier of fact and substitute its judgment for that of the jury or the trial court whose verdict is under review. That would be to usurp the function of the trier of fact and that is not allowed.3 See, e.g., Daniels v. State, 24 Md.App. 1, 7, 329 A.2d 712, 716 (1974) (“We may not usurp the function of the jury by holding that the eyewitnesses should be believed over the alibi evidence.”). See also Shelton v. State, 198 Md. 405, 412, 84 A.2d 76, 80 (1951) (“This Court will not inquire into or measure the weight of the evidence, and will not reverse the judgment if there is any proper evidence before the jury on which to sustain a conviction.”); Alexander v. Tingle, 181 Md. 464, 467, 30 A.2d 737, 738 (1943) (“The Court had not the authority to direct the jury that the evidence established a certain fact even though the evidence was uncontradicted and highly persuasive. The Court could not thus usurp the function of the jury to weigh the credibility of the evidence.”); Collins v. State, 14 Md.App. *717674, 679, 288 A.2d 221, 224 (1972) (“The weight of the evidence and the credibility of witnesses are matters within the realm of the jury.”); Wilkins v. State, 11 Md.App. 113, 127, 273 A.2d 236, 243 (1971) (“The weight of evidence and the credibility of the witnesses [are] for the jury.”). In this regard, what I said in dissent in Rubin has relevance here:
“No matter how strong a case for conviction the State may present, even when the defense presents no evidence, the court may not direct a verdict for the State. See Maryland Rule 4-324, which, while providing that a defendant may move for judgment of acquittal, Rule 4-324(a), and the court may direct the entry of judgment in his or her favor if there is insufficient evidence, as a matter of law, Rule 4-324(b), makes no provision for the making of a motion for judgment by the State. Compare Maryland Rule 2-519, the civil counterpart. Lyles v. State {State v. Lyles}, 308 Md. 129, 135, 517 A.2d 761, 764 (1986). This is so because it is the trier of fact, whether the court or a jury, that must determine if the State has met its burden of proof. To make that determination, the trier of fact is required to find the facts and when, as is usually the case, there are credibility issues, to resolve them. That, in turn, involves weighing the evidence. Appellate courts do not find facts or weigh evidence, ‘what evidence to believe, what weight to be. given it, and what facts flow from that evidence are for the jury ... to determine.’ Dykes v. State, 319 Md. 206, 224, 571 A.2d 1251, 1260-[6]1 (1990). See Gore v. State, 309 Md. 203, 214, 522 A.2d 1338, 1341 (1987); Wilson v. State, 261 Md. 551, 566, 276 A.2d 214, 221 (1971); Jacobs v. State, 238 Md. 648, 650, 210 A.2d 722, 723-[2]4 (1965). Even when an appellate court assesses the sufficiency of the evidence, it does not weigh it, see Clemson v. Butler Aviation-Friendship, 266 Md. 666, 671, 296 A.2d 419, 422 (1972); Gray v. Director, Patuxent Institution, 245 Md. 80, 84, 224 A.2d 879, 881 (1966), it only determines if any evidence exists, on the basis of which a rational trier of fact could find the elements of the crime beyond a reasonable doubt. See *718Jackson v. Virginia,[4] 443 U.S. 307, 319, 99 S.Ct. 2781, 2788, 61 L.Ed.2d 560, 573 (1979); Bloodsworth v. State, 307 Md. 164, 167, 512 A.2d 1056, 1057 (1986). There is no reason that a harmless error analysis should permit it to do more.”
325 Md. at 596-97, 602 A.2d at 698-99.
Citing Dorsey, the majority declares itself “satisfied that there is no reasonable possibility that the evidence complained of ... may have contributed to the rendition of the guilty verdict.” 276 Md. at 659, 350 A.2d at 678. Although mouthing the proper test, its rationale for so concluding belies its proper application. To the majority, it is significant and, indeed obvious, that Anderson’s testimony, because Anderson’s credibility is in issue, was “far less weighty” when compared with the expert testimony in Bohnert and, more important, “its effect on the jury was likely to be insignificant.” In so concluding, rather than determining whether the evidence likely, or could have, influenced the jury verdict, the majority places itself in the shoes of the jury and, in that posture, determines the weight that it would give the evidence, attributing that weight to the jury that decided the case. Thus, rather than apply the Dorsey test, it applies the test as explained in Ross v. State, 276 Md. 664, 674, 350 A.2d 680, 686-87 (1976). In so doing, the majority expands the harmless error rule in a manner that is both unwarranted and unsupportable.
As I explained in Rubin,
“The essence of this test is the determination whether the cumulative effect of the properly admitted evidence so out*719weighs the prejudicial nature of the evidence erroneously admitted that there is no reasonable possibility that the decision of the finder of fact would have been different had the tainted evidence been excluded (quoting Ross, 276 Md. at 674, 350 A.2d at 686-87).
“Thus excising the evidence admitted in error, and focusing on what remains, it concludes that a reasonable doubt simply could not have been generated from that evidence. Under that formulation, the test for harmless error is: whether, excluding the offending evidence, that which remains is sufficient to sustain the conviction and/or is, in fact, such that the case for conviction is “overwhelming.” It requires, in addition to the threshold determination of sufficiency, that an appellant court weigh the evidence. So, the majority reasons, where the evidence is sufficient to convict and it is also strong enough to meet its definition of “overwhelming,” only one verdict, guilty, is possible, as a matter of law; hence, that evidence cannot generate a reasonable doubt.
“Although the Ross formulation of the test has been stated in subsequent cases, see Trusty v. State, 308 Md. 658, 668-69, 521 A.2d 749, 754 (1987), and even applied, see Mills v. State, 310 Md. 33, 48-49, 527 A.2d 3, 10 (1987) (dicta), I have found no case which has provided a reasoned justification for it. It is significant, I think, that the error in Ross was held not to be harmless and that a factor influencing that holding was that the case was tried by a jury. 276 Md. at 674, 350 A.2d at 686-87. See also State v. Fuller, 308 Md. 547, 554, 520 A.2d 1315, 1318 (1987) (where trier of fact considered erroneously submitted evidence, case remanded to determine if, without that evidence trier of fact would reach same conclusion). In any event, that approach, in my opinion, is contrary to the role of an appellant court.”
Rubin, 325 Md. at 595-96, 602 A.2d at 698 (Bell, J. dissenting).
Furthermore, applying the harmless error rule as the majority does fails to recognize that the harmless error inquiry is not simply a sufficiency of the evidence inquiry, although when the exclusion of the evidence erroneously admitted results in *720an insufficiency of the evidence to convict, the error can never be harmless. To be sure, such an inquiry is required to be made whenever there is a reversal of a conviction on appeal, its function is to determine if a new trial is required, not, except as indicated above, to determine if the error is sufficiently egregious to warrant reversal of the conviction. Moreover, because it is the trier of fact, in this case, the jury, and not an appellate court, that must find the facts and resolve credibility issues,
“what appears, on the cold record, to be an insurmountable case for the State, when viewed from the jury’s perspective, having seen it unfold through live witnesses, in the dramatic atmosphere of the courtroom may be quite a close case or result in a defense verdict. How, or why, a jury may decide to resolve credibility or fact issues in a particular manner is a matter only it knows. One thing is certain, the jury is under no obligation to decide any case consistently with what is, objectively, the strongest case.”
Rubin, 325 Md. at 593, 602 A.2d at 697 (Bell, J. dissenting). Furthermore, harmless error is the exception to the general rule; it was never intended to be the general rule. In that regard, we have said that the test of harmless error “has been and should be carefully circumscribed.” Younie v. State, 272 Md. 233, 248, 322 A.2d 211, 219 (1974). And the continuing validity of the majority’s harmless error analysis makes what was said in People v. Jablonski, 38 Mich.App. 33, 195 N.W.2d 777, 780 (1972):
“Continued expansion of the harmless error rule will merely encourage prosecutors to get such testimony in, since they know that, they have a strong case, such testimony will not be considered to be reversible error, yet if they have a weak case, they will use such testimony to buttress the case to gain a conviction and then hope that the issue is not raised on appeal,”
more significant and a matter of concern.
Acceptance of the distinction the majority draws between Bohnert and the instant case does not warrant a different *721result with regard to the harmless error analysis. It may well be that the quality of the testimony of a witness seeking to buttress his own credibility is less than that of a witness deemed to be independent, testifying as an expert. I might also agree that the effect of the latter on the jury might be greater. Neither of these propositions answers the issue in this case, however. There can be a disparity in the quality of the testimony and in impact and yet the testimony of both could, and, I submit, probably would, affect the jury’s verdict. Who knows what factors guide the resolution of credibility issues by individual jurors? That respected members of the law enforcement community attest to the witness’ truthfulness likely will be viewed by the jury as important and worthy of credit. The source of the testimony, i.e., Anderson, may well be, or, in this case, could have been considered by the jury to be, less important than the fact that was the substance of that testimony, i.e., a prosecutor and a police detective believed what Anderson had to say about the petitioner’s involvement in the crime with which the petitioner was charged.
. Bohnert v. State, 312 Md. 266, 278, 539 A.2d 657, 663 (1988) makes clear "that a witness, expert or otherwise, may not give an opinion on whether he believes a witness is telling the truth. Testimony from a witness relating to the credibility of another witness is to be rejected as a matter of law.” We also said in Bohnert, "[i]t is . . . error for the court to permit to go to the jury a statement, belief, or opinion of another person to the effect that a witness is telling the truth or lying.” Id. at 277, 539 A.2d at 662.
. In Bohnert, a social worker, testifying as an expert, was allowed to say, based only on the child’s statements, that a child was the victim of sexual abuse. We held that testimony invaded the province of the jury because it was in effect a declaration that the child was telling the truth and that the defendant in the case was lying. See 312 Md. at 276, 278-79, 539 A.2d at 662-63. The majority distinguishes Bohnert and this case, as follows:
"In Bohnert, the State called a witness before the jury to bolster the credibility of another witness. In this case, we merely have the self-serving statement of the witness saying that others thought he was being truthful. Here we have Anderson basically saying: I am telling you the truth. I am telling you also that two others said I am telling you the truth.”
360 Md. 650, 679, 759 A.2d 764, 779 (2000).
. There was a time when the Court of Appeals refused to pass upon the question of the sufficiency of evidence to establish the crime with which the accused is charged for the reason that such action would usurp the constitutional function of the jury. See, e.g., Berger v. State, 179 Md. 410, 416-17, 20 A.2d 146, 149 (1941). At that time, it was well settled that, because, "[ujnder the Maryland Constitution, art. 15, sec. 5, the jury are the judges of both the law and the facts in the trial of all criminal cases in the State,” "the question of the legal sufficiency of evidence in a criminal case to sustain a conviction is exclusively for the jury to determine.” Id. at 416, 20 A.2d at 149 (citing Deibert v. State, 150 Md. 687, 695, 133 A. 847, 851 (1926)); Willie v. State, 153 Md. 613, 617, 139 A. 289, 291 (1927).
. In Jackson, the Supreme Court states:
"The sufficiency of the evidence to support a criminal conviction must be not simply to determine whether the jury was properly instructed, but to determine whether the record evidence could reasonably support a finding of guilt beyond a reasonable doubt (citing Woodby v. INS, 385 U.S. 276, 87 S.Ct. 483, 17 L.Ed.2d 362 (1966)). Instead, the relevant question is whether, after reviewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt (citing Johnson v. Louisiana, 406 U.S. 356, 92 S.Ct. 1620, 32 L.Ed.2d 152 (1972)).’’