259 Conn. 831
Conn.
2002

Opinion

BORDEN, J.

The respondent, the state of Connecticut (state), appeals1 from the judgment of the trial court granting the petitioner, Tyrone Adams, a new trial. After the unsuccessful appeal of his conviction for aiding and abetting manslaughter in the first degree, the petitioner had petitioned for a new trial based on the newly discovered testimony of Crystal Greene, which he claimed could be used to impeach the only eyewitness to the crime. The dispositive issue raised on appeal is whether the trial court improperly concluded that its decision on the petition was controlled, in large part, by our decision in State v. Valentine, 240 Conn. 395, 398, 692 A.2d 727 (1997), wherein we held that the trial court’s exclusion of Greene’s testimony in the trial of Daryl Valentine, the petitioner’s codefendant, constituted reversible error. More specifically, the state claims that by relying on our decision in Valentine, wherein we held that Greene’s testimony, if produced during the course of Valentine’s trial and believed by the jury, more likely than not would have affected the result therein, *833the trial court improperly failed to undertake an independent evaluation of Greene’s credibility. We agree with the state and, therefore, we reverse the judgment granting a new trial.

The petitioner in this case originally was charged with two counts of aiding and abetting murder in violation of General Statutes §§ 53a-82 and 53a-54a,3 and one count of aiding and abetting assault in the first degree in violation of General Statutes §§ 53a-8 and 53a-59 (a) (l).4 5Following a jury trial, he was convicted of two lesser included offenses of aiding and abetting manslaughter in the first degree in violation of General Statutes §§ 53a-8 and 53a-55 (a) (1).5

The petitioner filed a direct appeal to the Appellate Court, which affirmed the judgment of conviction. State *834v. Adams, 36 Conn. App. 473, 482, 651 A.2d 747 (1994). He then petitioned this court for certification to appeal from the judgment of the Appellate Court, which we initially granted. State v. Adams, 232 Conn. 913, 654 A.2d 357 (1995). We ultimately dismissed the appeal, however, after determining that certification had been improvidently granted. State v. Adams, 235 Conn. 473, 476, 667 A.2d 796 (1995).

Thereafter, the petitioner brought this petition for a new trial pursuant to General Statutes § 52-270,6 alleging that the newly discovered testimony of Greene likely would result in a different verdict in the event of a retrial. Following oral arguments on the petition, the trial court granted the requested relief and ordered a new trial. Thereafter, the state requested, and received, the trial court’s permission to appeal pursuant to General Statutes § 54-96.7 This appeal followed.

The following facts, as stated by the Appellate Court in the petitioner’s direct appeal, are pertinent to our resolution of this appeal. “At approximately 3 a.m. on September 21, 1991, police responded to a report of shots, fired at the Athenian Diner on Whalley Avenue *835in New Haven. At the scene, the police found a hostile crowd of fifty to 100 people surrounding two wounded men lying on the ground. These two men, Hury Poole and Andrew Paisley, were later pronounced dead at a hospital. Another victim, Christopher Roach, had been shot in the arm.

“An altercation had originated in the diner and continued outside. A crowd gathered, including the defendant; his friend Daryl Valentine, who was the shooter; and Byron McFadden, who would later testify at trial.

“As Roach, Poole and Paisley joined the crowd, [McFadden] heard the [petitioner] shout, ‘Shoot him, shoot him, f— it, shoot him.’ Valentine shot Poole and Paisley, ran to a car parked in the driveway of the diner and got in on the passenger side. Roach ran after him. As Roach approached the car, Valentine fired at him, hitting him in the arm. The car with Valentine in it sped away, followed shortly thereafter by the [petitioner], who drove off in his own car.” State v. Adams, supra, 36 Conn. App. 474-75.

At the hearing on this petition for a new trial, the petitioner called one witness, public defender Thomas Ullmann, who had represented Valentine in his trial. Ullmann testified that the exculpatory information provided by Greene first became available during Valentine’s trial, which took place after the petitioner had been convicted. Ullmann also testified that he had been contacted by Valentine’s sister, who told him that Greene “had information regarding a key witness in the case.” Ullmann testified further that he had interviewed Greene, who revealed that she had engaged in two separate conversations regarding the incident with her cousin, Roach, the only victim to have survived, and who identified Valentine as the shooter at trial. According to Greene, Roach had admitted to her that he had not seen the shooting, but stated that “somebody *836has to pay the price, somebody has to go.” In Valentine’s trial, Ullmann had submitted Greene’s testimony to the trial court in an offer of proof for the purpose of impeaching Roach’s prior statement that he had never spoken to Greene about the incident at the Athenian Diner. The trial court excluded the testimony, however, characterizing it as collateral.

Valentine ultimately was convicted of murder. He then appealed from that judgment to this court. We held improper the ruling excluding Greene’s testimony, on the grounds that “the identity of the shooter and whether Roach knew that identity, [were] material to the central issue in the case and, therefore, was not collateral.” State v. Valentine, supra, 240 Conn. 404. Because this court determined that the improper exclusion of Greene’s testimony constituted harmful error, we reversed Valentine’s conviction and ordered a new trial. Id., 404-405.

At Valentine’s new trial, Greene initially testified that, in her first conversation with Roach, he refused to answer when she asked whether Valentine had anything to do with the shooting, stating that he could not discuss it. It was only after Ullmann refreshed her recollection with the prior offer of proof that Greene testified that Roach had told her that he “didn’t see who did it.” With respect to their second conversation, Greene again testified that Roach consistently had told her that he could not discuss matters related to the trial, and that it was only after being “pushed” on the issue that Roach said, “somebody has to pay the price, somebody has to go.” After hearing this, and other relevant testimony, the jury on the retrial again convicted Valentine on two counts of murder, one count of attempted assault in the first degree, and one count of carrying a pistol without a permit.8 *837As previously stated, only Ullmann testified on behalf of the petitioner during the hearing on the petition for a new trial. The trial court did not hear live testimony from Greene, but did have the benefit of reviewing the transcript of her testimony from Valentine’s second trial. After considering the transcript, together with this court’s first decision in State Valentine, supra, 240 Conn. 395, the trial court granted the petition.

The dispositive issue raised by the state on appeal is whether the trial court improperly concluded that its decision was dictated, in large part, by this court’s decision in State v. Valentine, supra, 240 Conn. 395. Specifically, the state claims that, in adhering to our appellate ruling in Valentine, the trial court improperly failed to evaluate Greene’s credibility in order to determine whether her testimony was likely to produce a different result in the event of a new trial. The petitioner concedes that it was necessary for the court to engage in such a credibility analysis, but counters with the argument that the trial court fulfilled its obligation in this respect. We agree with the state.

As a preliminary matter, we set forth the standard of review that governs our analysis on appeal. Typically, we review a trial court’s actions with respect to a petition for a new trial for an abuse of discretion. See Kubeck v. Foremost Foods Co., 190 Conn. 667, 670, 461 A.2d 1380 (1983). This particular case, however, requires that we determine whether the trial court applied the appropriate standard in evaluating the petition at issue. Because this is a question of law, our review is plenary. See Olson v. Accessory Controls & Equipment Corp., 254 Conn. 145, 156, 757 A.2d 14 (2000).

Our analysis is founded, in large part, on our holding in Shabazz v. State, 259 Conn. 811, 792 A.2d 797 (2002), also decided today. In Shabazz, the petitioner appealed *838from the trial court’s denial of his petition for a new trial, claiming that the court improperly had engaged in a credibility analysis of certain newly discovered eyewitness testimony that was alleged to justify a new trial. Id., 812-13. We affirmed the judgment denying the petition, and held that: (1) a trial corn! must review a petition for new trial based on newly discovered evidence in accordance with the four-part test set forth in Asherman v. State, 202 Conn. 429, 521 A.2d 578 (1987);9 Shabazz v. State, supra, 818; and (2) both past precedent and the various interests implicated by petitions for a new trial mandate that a court assess the credibility of the newly discovered evidence at issue when deciding, under Asherman’s fourth prong, whether it is likely to produce a different result on retrial. Id., 822. We articulated the trial court’s burden as follows: “The trial court must always consider the newly discovered evidence in the context of the evidence presented in the original trial. In so doing, it must determine, first, that the evidence passes a minimum credibility threshold. That is, if, in the court’s opinion, the newly discovered evidence simply is not credible, it may legitimately determine that, even if presented to a new jury in a second trial, it probably would not yield a different result and may deny the petition on that basis. ... If, however, the trial court determines that the evidence is sufficiently credible so that, if a second jury were to consider it together with all of the original trial evidence, it probably would yield a different result or otherwise avoid an injustice, the fourth element of the Asherman test would be satisfied.” (Citation omitted.) Id., 827-28.

*839In light of our holding in Shabazz, both parties in the present case are correct in their assertion that the trial court was obligated to evaluate Greene’s credibility as a necessary part of the analysis under Ashemnan. The question we must resolve on appeal, therefore, is whether the trial court properly fulfilled its duty in this regard. Moreover, we note that “[n]ew trials [typically] are not granted upon newly discovered evidence which discredits a witness unless the evidence is [both] vital to the issues and . . . strong and convincing . . . .” Turner v. Scanlon, 146 Conn. 149, 163, 148 A.2d 334 (1959). “The rule restricting the right to a new trial when one is claimed on the basis of newly discovered evidence merely affecting the credibility of a witness is necessary because scarcely has there been an important trial . . . [after which a] diligent search would not have discovered evidence [to impeach] some witness .... Without such a rule, there might never be an end to litigation.” (Citation omitted; internal quotation marks omitted.) Lancaster v. Bank of New York, 147 Conn. 566, 578, 164 A.2d 392 (1960).

Our review of the trial court’s memorandum of decision discloses that the trial court failed to assess properly the credibility of the proffered newly discovered evidence. The court first correctly noted that “for the petitioner to be entitled to a new trial on the basis of newly discovered evidence he must establish the following: ‘(1) the proffered evidence [namely, Greene’s testimony] is newly discovered, such that it could not have been discovered earlier by the exercise of due diligence; (2) it would be material on a new trial; (3) it is not merely cumulative; and (4) it is likely to produce a different result in a new trial.’ Asherman v. State, [supra, 202 Conn. 434].” The court then applied the first three of these elements to the evidence before it, and determined that: (1) Greene’s testimony did, in fact, constitute newly discovered evidence because her con*840versations with Roach did not take place until after the petitioner had been convicted; (2) the question of whether Greene’s testimony would be material in the event of a new trial was answered by this court’s evidentiary ruling in Valentine-, and (3) Greene’s testimony was not cumulative.

The remainder of the trial court’s memorandum of decision was devoted to an analysis under the fourth prong of the Asherman test, namely, whether Greene’s testimony was likely to produce a different result on retrial. In this context, the court expressly articulated its “opinion . . . [that] the outcome of [the] petition is to a large extent governed by the Supreme Court’s holding in State v. Valentine [supra, 240 Conn. 404] . . . that the exclusion of Greene’s testimony was harmful error.” In State v. Valentine, supra, 404, we had stated that “Greene’s testimony, if believed by the jury, seriously would have undermined Roach’s in-court testimony identifying the defendant as the shooter” and, therefore, would have been more likely than not to affect the outcome of the trial. (Emphasis added.) The trial court noted that, in this respect, the harmful error standard “is not unlike the fourth element [of Asherman] which must be established [before the petitioner is] entitled to a new trial, i.e., that the evidence in question is likely to produce a different result in a new trial.” Without further discussion of, or analysis under, Asherman’s fourth prong, the trial court concluded that it was bound to provide the petitioner with a new trial because we had accorded the same relief to his codefendant in Valentine based on the harmful exclusion of Greene’s testimony.

The trial court’s reliance on our conclusion of harmful error in Valentine as a substitute for an independent credibility assessment under the fourth prong of the Asherman test was misplaced. First, in relying on the transcript of Greene’s testimony from Valentine’s sec*841ond trial, rather than requiring that Greene testify as a witness in the hearing on the petition, the trial court in the present case acted solely as an evidentiary gatekeeper, rather than as a fact finder. This was improper. See Shabazz v. State, supra, 259 Conn. 824, 826 (in hearing petition for new trial “ ‘it [is] the duty of the trial judge to try the facts’ otherwise, trial court would be “relegated to the role of gatekeeper of the evidence ... a result [that] would render judgments of conviction unduly susceptible to collateral attacks”).10 Moreover, implicit in our recognition in Shabazz that the trial court sits as fact finder in a hearing on a petition for a new trial is the principle that, absent extraordinary or extenuating circumstances, the court should make its credibility assessments—both initial and thereafter—on the basis of the presentation of live testimony, rather than on the basis of a printed record. It was, therefore, improper for the trial court in the present case to limit its analysis of the petition to Greene’s offer of proof from Valentine’s first trial and the transcript of her testimony from his second trial.

Second, harmful error is an appellate construct, used to determine, in cases such as Valentine, whether an improper evidentiary ruling by the trial court likely affected the outcome of the trial. See State v. Valentine, supra, 240 Conn. 404 (“[t]o require the reversal of a conviction because of evidentiary error, the defendant must show that the error was harmful, i.e., that it is more probable than not that the erroneous ruling of the trial court affected the result”). When judging the likely effect of such a trial court ruling, the reviewing court is constrained to make its determination on the *842basis of the printed record before it. The appellate court does not decide whether the jury would have believed the evidence and whether the verdict therefore would have been different; instead, the court decides whether the evidence, if believed by the jury, likely would have produced a different result. Id. The reviewing court thus abstains from drawing any independent conclusions regarding the credibility of the evidence, leaving that responsibility to the jury on a new trial, who may observe, firsthand, the witness’ conduct, demeanor and attitude. See State v. Edwards, 247 Conn. 318, 323, 721 A.2d 519 (1998); State v. Mejia, 233 Conn. 215, 224, 658 A.2d 571 (1995).

In deciding a petition for a new trial, however, the trial court sits as fact finder in place of the jury and examines the newly discovered evidence independently, in order to determine whether it is likely to result in a different verdict in the event of a retrial. See Shabazz v. State, supra, 259 Conn. 824; People v. Shilitano, 218 N.Y. 161, 180, 112 N.E. 733 (1916). To this end, the trial court necessarily undertakes some sort of credibility assessment; for, if the newly discovered evidence is not worthy of belief, a jury sitting on the new trial would not be persuaded by it to reach a different result. See Shabazz v. State, supra, 825; Smith v. State, 141 Conn. 202, 208, 104 A.2d 761 (1954). The trial court, moreover, is obliged to make such a credibility determination, not on the basis of a printed record, as a reviewing court must do in undertaking a harmful error analysis, but rather on the basis of live testimony.

In substituting our finding of harmful error in Valentine for its analysis under the fourth prong of the Asherman test, the trial court thus assumed, without deciding, the believability of Greene’s testimony. Because such a credibility evaluation is necessary to a determination of whether Greene’s testimony was likely to result in an acquittal in the event of a new trial, the *843trial court’s finding with respect to this element of the Asherman test was incomplete.

The petitioner maintains, contrary to this conclusion, that the trial court’s analogy of Asherman’s fourth prong with the appellate test for harmful error did not amount to a “lock-step blind adherence to Valentine . . . . Quite the contrary, it [was] the well-reasoned and logical conclusion to the trial court’s thorough assessment of all the evidence, including the Valentine I offer of proof.” This argument, however, disregards several references in the trial court’s memorandum of decision to the binding nature and effect of our decision in Valentine. As previously stated, the trial court explicitly pronounced its belief that the outcome of the petition was largely dictated by our holding in Valentine, even going so far as to conclude that “[i]n the last analysis the Supreme Court has held that . . . Valentine was entitled to a new trial in which the jury would hear from . . . Greene. This court believes it is therefore bound to follow that holding and provide the same opportunity to [the petitioner].” (Emphasis added.) Moreover, the trial court twice emphasized our holding that Greene’s testimony, if believed by the jury, would likely affect the result of the trial. This reiteration of the appellate standard for harmful error suggests that, in the absence of any explicit finding by the trial court that Greene was a credible witness, the court relied on our holding in Valentine in lieu of making an independent determination as to whether her testimony was sufficiently believable as to warrant a new trial.

The petitioner further contends that, irrespective of any reference to Valentine, the trial court nonetheless fulfilled its obligation to evaluate the credibility of Greene’s testimony. In support of this assertion, the petitioner first points to that portion of the memorandum of decision in which the trial court responded to the state’s characterization of Greene’s testimony at *844Valentine’s second trial as “weak, ambiguous and blatantly biased.” The court noted that, although “Greene and Valentine had a relatively close relationship, it cannot be said for that reason alone that she was not worthy of belief.” In this context, the court also stated its belief that Greene’s testimony would not “necessarily be considered by a jury as weak and ambiguous.”

Far from definitively establishing that the trial court found Greene to be a believable witness, as the petitioner contends, these portions of its memorandum of decision simply illustrate the court’s conclusion, based solely on the transcript of Greene’s testimony, that she was not incredible. As we stated in Shabazz, such a finding constitutes only the first, or threshold, step in a trial court’s analysis of a petition for a new trial under the fourth prong of the Asherman test. See Shabazz v. State, supra, 259 Conn. 824. The trial court is also required to determine whether the newly discovered evidence is sufficiently credible such that, if admitted in a new trial and reviewed by a second jury together with all of the evidence presented at the original trial, it is likely to produce a different result. Notably absent from the trial court’s memorandum of decision is any finding that might satisfy this additional obligation. Moreover, as we previously have stated, whether the proffered newly discovered evidence surmounts an initial determination of credibility and, thereafter, whether it is sufficiently credible to justify a new trial, must be decided on the basis of the trial court’s own assessment of credibility, not on the type of cold transcript utilized by the trial court in the present case.

The petitioner also points to the trial court’s request that the parties brief the issue of whether it was necessary to hear Greene’s live testimony as evidence that “the court was well aware that it had to make a threshold determination as to the impact of . . . Greene’s evidence.” To the contrary, this request, taken both by *845itself and in the context of the colloquy between the court and counsel during the hearing on the petition, clearly reveals that the trial court was unsure of its responsibilities with respect to assessing the credibility of Greene’s testimony.

The trial court first touched upon the issue of the witness’ credibility in the context of ruling on an objection to Ullmann’s recitation of what was told to him by Greene in their initial interview. The court stated: “I think this is being offered merely to establish what was said because obviously what was said by this person at any stage of these proceedings is relevant to the issue here, and again, as we discussed in chambers, issues of credibility, truth of what was said, those things may or may not come into play.” (Emphasis added.) Subsequently, at the conclusion of the hearing, the court informed counsel that it would review their briefs together with Greene’s transcript from Valentine’s second trial and notify them as to a date for the presentation of Greene’s testimony, if, in the court’s opinion, such testimony was necessary. In this context, the court stated: “[I]t could well be that the testimony of . . . Greene with cross-examination might satisfy that requirement. It will give me a better sense as to at least what was said. Certainly I can’t assess her demeanor, but as we discussed in chambers, I am not so sure as to the extent I should be getting into that in any event.” (Emphasis added.)

Taken together, these two statements by the trial court reveal its ambivalence regarding whether, and to what extent, it should evaluate Greene’s credibility. Had the court been aware of its obligation in this respect, as the petitioner contends, it would not have expressly referred to the issue in such equivocal terms, noting both that credibility “may or may not come into play,” and that it was “not . . . sure as to the extent” to which it should assess Greene’s demeanor. Moreover, the *846mere fact that the court asked the parties to brief the issue of whether live testimony was necessary, coupled with its ultimate decision that such testimony was not required for a resolution of the petition, indicates that it was unaware of its duty to engage in an analysis of Greene’s credibility. As previously discussed, a trial court reviewing a petition for a new trial must hear live witness testimony, where such testimony constitutes the newly discovered evidence at issue, in order to determine properly whether the fourth prong of the Asherman test is satisfied. In other words, absent extraordinary circumstances, the trial court must have the opportunity, not only to evaluate the substance of the newly discovered testimony, but also the attitude, demeanor and candor with which it is communicated; otherwise, the court will be unable to assess adequately whether the testimony is sufficiently credible such that it is likely to produce a different result in the event of a new trial. The trial court recognized as much when it acknowledged, in its memorandum of decision, that it “did not have the benefit of seeing Greene testify so as to evaluate her credibility . . . .”11

Finally, there is no indication in the trial court’s memorandum of decision that it gave any consideration to the heightened standard for granting a new trial based on newly discovered impeachment evidence set forth *847in Turner v. Scanlon, supra, 146 Conn. 163. The absence of any finding that Greene’s testimony was sufficiently strong and convincing so as to warrant a new trial further convinces us that the trial court misunderstood its responsibilities with respect to this petition.

The judgment is reversed and the case is remanded for a new trial on the petition.

In this opinion the other justices concurred.

Notes

The state appealed from the judgment of the trial court to the Appellate Court, and we transferred the case to this court pursuant to General Statutes § 51-199 (c) and Practice Book § 65-1.

General Statutes § 53a-8 (a) provides: “A person, acting with the mental state required for commission of an offense, who solicits, requests, commands, importunes or intentionally aids another person to engage in conduct which constitutes an offense shall be criminally liable for such conduct and may be prosecuted and punished as if he were the principal offender.”

General Statutes § 53a-54a (a) provides: “A person is guilty of murder when, with intent to cause the death of another person, he causes the death of such person or of a third person or causes a suicide by force, duress or deception; except that in any prosecution under this subsection, it shall be an affirmative defense that the defendant committed the proscribed act or acts under the influence of extreme emotional disturbance for which there was a reasonable explanation or excuse, the reasonableness of which is to be determined from the viewpoint of a person in the defendant’s situation under the circumstances as the defendant believed them to be, provided nothing contained in this subsection shall constitute a defense to a prosecution for, or preclude a conviction of, manslaughter in the first degree or any other crime.”

General Statutes § 53a-59 (a) provides in relevant part: “A person is guilty of assault in the first degree when: (1) With intent to cause serious physical injury to another person, he causes such injury to such person or to a third person by means of a deadly weapon or a dangerous instrument . . . .”

General Statutes § 53a-55 (a) provides in relevant part: “A person is guilty of manslaughter in the first degree when: (1) With intent to cause serious physical injury to another person, he causes the death of such person or of a third person . . . .”

General Statutes § 52-270 (a) provides: “The Superior Court may grant a new trial of any action that may come before it, for mispleading, the discovery of new evidence or want of actual notice of the action to any defendant or of a reasonable opportunity to appear and defend, when a just defense in whole or part existed, or the want of actual notice to any plaintiff of the entry of a nonsuit for failure to appear at trial or dismissal for failure to prosecute with reasonable diligence, or for other reasonable cause, according to the usual rules in such cases. The judges of the Superior Court may in addition provide by rule for the granting of new trials upon prompt request in cases where the parties or their counsel have not adequately protected their rights during the original trial of an action.”

General Statutes § 54-96 makes the requirements set forth in General Statutes § 54-95 (a) applicable to the state, and provides: “Appeals from the rulings and decisions of the Superior Court, upon all questions of law arising on the trial of criminal cases, may be taken by the state, with the permission of the presiding judge, to the Supreme Court or to the Appellate Court, in the same manner and to the same effect as if made by the accused.”

Thereafter, we affirmed Valentine’s convictions. State v. Valentine, 255 Conn. 61, 78, 762 A.2d 1278 (2000).

In Asherman v. State, supra, 202 Conn. 434, we stated that a court may grant a petition for a new trial “when it is satisfied that the evidence offered in support thereof: (1) is newly discovered such that it could not have been discovered previously despite the exercise of due diligence; (2) would be material to the issues on a new trial; (3) is not cumulative; and (4) is likely to produce a different result in the event of a new trial.” Shabazz v. State, supra, 259 Conn. 820-21.

We recognize that neither the petitioner nor the state alerted the trial court to this procedural lacuna, perhaps because of the uncertainty in our prior case law, which Shabazz has now resolved. Furthermore, of course, the trial court did not have the benefit of our decision in Shabazz when it ruled on the petition.

The petitioner points to the trial court’s next statement, namely, that “[t]he jury must make the ultimate decision as to the weight to give [Greene’s] testimony,” as indicating, not that the court failed to make an independent determination regarding credibility, but rather that the court recognized the inherent difference between weight and credibility. We disagree. Coupled with the court’s acknowledgment that it did not have the benefit of Greene’s live testimony so as to evaluate her credibility, this statement further reveals the court’s misconception that it was solely the function of the jury on retrial, rather than the court in ruling on the petition, to evaluate whether the testimony was worthy of belief. In placing the burden of evaluating Greene’s testimony on the shoulders of the jury on retrial, the trial court improperly abdicated its responsibility to make the initial, requisite credibility determination outlined in Shabazz.