Levasseur v. Pepe

STAHL, Circuit Judge. In January of 1990, petitioner Mark Le-vasseur was convicted in a Massachusetts state court of rape, indecent assault and battery, and assault and battery. After exhausting his state court remedies, Levasseur sought a writ of habeas corpus in the United States District Court for the District of Massachusetts, pursuant to 28 U.S.C. § 2254, claiming that his state convictions were obtained in violation of the United States Constitution. Specifically, Levasseur asserted that an improper admission of hearsay testimony violated his Sixth Amendment right to confrontation, an unduly suggestive identification procedure violated his Fourteenth Amendment right to due process of law, and his trial counsel’s deficient performance deprived him of his Sixth Amendment right to the effective assistance of counsel. The district court denied the habeas application, and this appeal followed. For the reasons discussed below, we affirm. I. BACKGROUND A Pretrial Events 1 At some time between the late evening and early morning hours of June 30 and July 1, 1988, a man driving a big, dark-colored, “4 x 4” pickup truck approached the victim (“Jane Doe”) as she was walking on the sidewalk and asked her if she wanted a ride. Undeterred by Doe’s initial rejection, the driver returned and Doe relented. Instead of dropping Doe at her destination, the driver took *190 her to a deserted area behind a factory and raped her on the truck seat. Her attacker, still restraining Doe, drove away from the scene and then pushed her out of the truck. Ronald Ralls, driving behind the truck, saw Doe fall from the truck and stopped to render aid. Ralls drove Doe to his home and notified the police. Officer Mendes arrived, but Doe rebuffed his attempts to approach her. The officer detected the smell of alcohol on Doe’s breath. Emergency paramedics transported Doe to the hospital. Doe’s continued refusal to let anyone touch her prevented the hospital staff from administering a “rape kit.” Eventually, Doe, in the presence of Officer Mendes, related the incident in general terms to a female hospital administrator. Five days later at the police station, Doe further recounted the incident to Inspector Boutselis. She described her assailant’s pickup truck as a big, dark “4 x 4” sitting high off the ground and her attacker as a white male with a medium-type build, thin face, high cheekbones, dark hair, dark eyelashes, and a small moustache. Boutselis showed Doe over six thousand photographs of white males. A 1983 photograph of Le-vasseur was among the six thousand. Although Doe said several photographs, including Levasseur’s, looked similar to her assailant, she was unable to make an identification. On July 13, 1988, because of a new lead, Inspectors Boutselis and Guilfoyle began to focus on Levasseur. On their way to Levass-eur’s residence in an unmarked car, Boutselis and Guilfoyle noticed a pickup truck fitting Doe’s description. Boutselis and Guilfoyle observed the truck pull over to the side of the road, saw the driver beckon to a young blond-haired girl walking on the sidewalk, and heard him say “Come on.” After the girl continued walking and the truck drove on, Guilfoyle questioned the girl, who said that the man in the truck had offered her a ride but she did not know him. Learning from the police dispatcher that the truck belonged to Levasseur, Boutselis and Guil-foyle stopped Levasseur and asked about the girl. Levasseur told them that she was his cousin and he had offered her a ride, but he did not volunteer her name. Levasseur agreed to go to the police station, where Boutselis and Guilfoyle questioned and photographed him. The next day, Inspector Boutselis showed Doe a group of sixty photographs including Levasseur’s photograph from the day before. Doe stated that Levasseur’s recent photograph looked like her assailant and asked Boutselis for more photographs of Levass-eur. Upon examining Levasseur’s 1983 photograph, Doe said that it looked more like her assailant than the recent photograph. Given the seriousness of the charges, Doe refused to make a positive identification of Levasseur based solely on the photographs and requested to see him in person. Using an unmarked car, Inspector Boutsel-is and Doe began field identifications. 2 Over the next five months, they made nine excursions lasting two to three hours each to locations near Levasseur’s place of work and home. On five occasions, they sat outside the Textron plant in Wilmington thinking that Levasseur was employed there. Le-vasseur, it was later discovered, worked at Digital Equipment Corporation, not Textron. On the remaining four occasions, Boutselis and Doe sat at an intersection in North Chelmsford one-half mile from Levasseur’s home. Doe never positively identified anyone on the first eight excursions. During the ninth trip, on December 16, 1988, however, Doe saw Levasseur’s truck approach and exclaimed, “Wow, this looks like it.” As the truck came within six feet of Doe and Le-vasseur looked in her direction, Doe screamed, “It’s him, it’s him. Get me ... out of here right now.” B. The Trial On January 24, 1990, Levasseur was brought to trial on charges of rape, indecent assault and battery, and assault and battery. The prosecution called six witnesses in its direct case and one witness on rebuttal, with Levasseur calling himself, his wife, and his mother-in-law as his defense witnesses. *191 In Ms rebuttal case, the district attorney called Inspector Guilfoyle to testify about the July 13, 1988, incident with the blond-haired girl (we use the language used at trial and refer to “the blond-girl incident”). In recounting Ms conversation with the blond girl, Inspector Guilfoyle uttered the following hearsay testimony: Q. Sir, ... what was the nature of the conversation with her [the blond girl]? A. We pulled up; we were in an unmarked vehicle. We identified ourselves to the young lady. We asked her if she knew the gentleman in the truck; she said no. We asked her if she could tell us what he wanted and she said he offered her a ride. (emphasis added). After a three-day trial, the jury found Le-vasseur guilty of rape, indecent assault and battery, and assault and battery. Levasseur was sentenced to concurrent state prison terms of ten to fifteen years and four to five years. C. Postr-Conviction Proceedings After the jury’s verdict, Levasseur filed a motion for a new trial wMch the trial court denied. The Massachusetts Appeals Court affirmed the convictions, and the Supreme Judicial Court of the Commonwealth of Massachusetts denied Levasseur’s petition to obtain further appellate review. The United States Supreme Court denied Levasseur’s petition for a writ of certiorari. Having exhausted his state remedies, Le-vasseur filed a petition for habeas corpus in the United States District Court for the District of Massachusetts. The district court denied the petition for the following reasons. Levasseur’s trial counsel had procedurally defaulted on the claim of unduly suggestive identification by his failure to object. The court held that Levasseur could not excuse this procedural default because he could not show cause for that failure via a claim of meffective assistance of counsel. Counsel’s failure to object did not constitute deficient performance because the identification method was not unduly suggestive. With respect to the hearsay claim, the court held that Levasseur’s Confrontation Clause rights were violated, but the error was harmless. Finally, according to the district court, Le-vasseur’s remaining ineffective-assistance-of-counsel claims failed because trial counsel’s overall performance was not deficient and Levasseur suffered no prejudice. II. DISCUSSION On appeal, Levasseur makes the following four arguments: (1) he did not procedurally default on Ms Due Process and Confrontation Clause claims, (2) the violation of his Sixth Amendment right to confrontation was not harmless error, (3) the pre-trial identification procedure was unduly suggestive and violated his Fourteenth Amendment right to due process of law, and (4) he was denied effective assistance of counsel. We discuss each argument in turn. A Procedural Default Levasseur asserts that the district court erred in holding that he procedurally defaulted on his Due Process and Confrontation Clause claims. Levasseur admits that Ms trial counsel did not object to the pre-trial identification procedure and the hearsay statement as each was introduced and acknowledges that, to consider such claims on appeal, Massachusetts requires contemporaneous objection to their admission as evidence at trial, Mass.R.Crim.P. 24(b). He contends, however, that his filing a motion for new trial and the trial court’s consideration of the issues presented acted to resurrect and preserve these claims for review. We disagree. Under Massachusetts case law, issues previously lost for appeal may be resurrected and preserved for appellate review if a trial judge exercises Ms discretion and considers them in ruling on a motion for a new trial. See Commonwealth v. Harrington, 379 Mass. 446, 399 N.E.2d 475, 478 (1980); Commonwealth v. Gagne, 367 Mass. 519, 326 N.E.2d 907, 911 (1975); Commonwealth v. Buckley, 17 Mass.App.Ct. 373, 458 N.E.2d 781, 783 review denied, 391 Mass. 1103, 461 N.E.2d 1219 (1984). Whether the trial court *192 considered Levasseur’s Due Process and Confrontation Clause claims and thereby waived his procedural default is a legal question subject to our plenary review. See Burks v. Dubois, 55 F.3d 712, 716 (1st Cir.1995) (reviewing de novo question of whether state court’s limited review of petitioner’s claim for miscarriage of justice effected a waiver); Puleio v. Vose, 830 F.2d 1197, 1200 (1st Cir.1987), cert. denied, 485 U.S. 990, 108 S.Ct. 1297, 99 L.Ed.2d 506 (1988) (same). In ruling on Levasseur’s motion for new trial, the trial judge did not consider Levass-eur’s Due Process and Confrontation Clause claims individually but rather only as part and parcel of Levasseur’s ineffective-assis-tanee-of-counsel claim. The court concluded that the attorney’s failure to object to the identification process and the hearsay testimony was not ineffective assistance of counsel. Because the trial judge never considered the Due Process and Confrontation Clause claims severally, it did not resurrect these proeedurally defaulted claims, and we consider them on appeal only as they constitute part of Levasseur’s ineffeetive-assis-tance-of-counsel claim. Levasseur invites this court to rule that an appellant claiming ineffective assistance of counsel in a motion for a new trial before a state trial court preserves for habeas corpus review not only the ineffective-assistance-of-counsel claim but the substantive claims subsumed within the ineffeetive-assistance-of-counsel claim. In this case and typically, however, the substantive claims underlying the ineffective-assistance-of-counsel claim are constitutional claims that were proeedurally defaulted due to the ineffective performance of counsel. In Coleman v. Thompson, 501 U.S. 722, 750, 111 S.Ct. 2546, 2564-65, 115 L.Ed.2d 640 (1991), the Supreme Court ruled that federal habeas review of proeedurally defaulted claims is barred unless the petitioner demonstrates (1) cause for the default and “actual prejudice as a result of the alleged violation of federal law” or (2) that failure to consider the claims will result in “a fundamental miscarriage of justice.” By proposing that proeedurally defaulted claims should also be preserved if they form part of an ineffeetive-assistance-of-counsel claim, Le-vasseur suggests that we create a third way to excuse procedural default, where Coleman purposefully provides only two. Recognizing Levasseur’s suggestion as an attempt to make an end run around Coleman, we refuse his invitation. Having determined that the trial court did not resurrect Levasseur’s otherwise proee-durally defaulted Due Process and Confrontation Clause claims, we turn our attention to determining whether we nonetheless may review the proeedurally defaulted claims on habeas. We begin by considering the Confrontation Clause claim. B. Inadmissible Hearsay Where a petitioner, like Levasseur, has proeedurally defaulted on his claim, we reach the merits on habeas corpus review only if the default can be excused by establishing either cause for the default and actual prejudice resulting therefrom or that failure to consider the claim will result in a “substantial miscarriage of justice.” Coleman, 501 U.S. at 750, 111 S.Ct. at 2565. Should the petitioner surmount this hurdle and excuse his default, we may then reach the merits of his claim and determine whether a constitutional error has occurred and, if so, whether the error was harmless. 3 In considering Levasseur’s proeedurally defaulted Confrontation Clause claim, the district court assumed arguendo that Levass-eur had excused his default and limited its analysis to the merits of Levasseur’s Con *193 frontation Clause claim. Specifically, the district court held that although Inspector Guil-foyle’s testimony about the blond girl’s statements was inadmissible hearsay that violated Levasseur’s Sixth Amendment right to confront his accusers, it was harmless error. Similarly we limit our analysis to the determination of harmless error and affirm for the reasons that follow. We first set forth the legal framework. The standard here for determining whether habeas relief must be granted because of unconstitutional trial error is the Brecht/Kotteakos standard of harmless error. Under that standard, an unconstitutional trial error is harmless unless the court finds with fair assurance that the error, considered in light of the record as a whole, “ ‘had [a] substantial and injurious effect or influence in determining the jury’s verdict.’ ” Brecht v. Abrahamson, 507 U.S. 619, - -, 118 S.Ct. 1710, 1721-22, 128 L.Ed.2d 353 (1993) (quoting Kotteakos v. United States, 328 U.S. 750, 776, 66 S.Ct. 1239, 1253, 90 L.Ed. 1557 (1946)); see Kotteakos, 328 U.S. at 765, 66 S.Ct. at 1248 (holding that requisite degree of certainty is “fair assurance”). Assessments of harmless error are necessarily context-specific. Kotteakos, 328 U.S. at 762, 66 S.Ct. at 1246-47. The following factors, however, are relevant to our determination of whether the jury was substantially swayed by the tainted hearsay evidence: (1) the extent to which the error permeated the proceeding, (2) the centrality of the issue affected by the error to the ease as actually tried and (3) the relative strength of the properly admitted evidence of guilt. See Brecht, 507 U.S. at -, 113 S.Ct. at 1722 (considering infrequency of state’s references to constitutional error and strength of evidence of guilt as factors relevant to assessing whether error “substantially influence[d]” jury); Shaw v. Collins, 5 F.3d 128, 132-33 (5th Cir.1993) (considering centrality of the issue affected by the error to the case as tried and finding that Confrontation Clause violation was not harmless because tainted testimony “was the linchpin in the State’s case”). After reciting the standard of review, we consider each factor in turn. Because a harmless-error determination on habeas corpus review is a mixed question of law and fact, we examine this issue de novo. Brecht, 507 U.S. at -, 113 S.Ct. at 1724 (Stevens, J., concurring); Scarpa v. Dubois, 38 F.3d 1, 9 (1st Cir.1994) (holding that “[i]n federal courts, mixed questions of law and fact arising in section 2254 cases are ordinarily subject to de novo review”), ce rt. denied, — U.S. -, 115 S.Ct. 940, 130 L.Ed.2d 885 (1995). 1. Prevalence of the error The constitutional error in this case consists of two discrete statements uttered at the end of the trial. During his rebuttal case, the district attorney violated Levass-eur’s Sixth Amendment right to confront his accusers by eliciting from Inspector Guilfoyle the hearsay testimony about the blond girl’s statements to him. Specifically, Guilfoyle testified that the blond girl said Levasseur had offered her a ride and that she had said no. In his closing argument, the district attorney paraphrased the offending statement: One can look at this case, and if one adds up the facts that were proven beyond a reasonable doubt and then add any inferences that can be drawn from the account with the young female on the bridge on July 13th, with almost the same words that were spoken to [Jane Doe]: Do you want a ride? (emphasis added). Viewing these two constitutionally erroneous statements in the context of the trial as a whole, we find that they did not permeate the trial proceedings. During his opening argument and throughout the presentation of his direct case, the district attorney carefully avoided any hearsay testimony about the blond girl’s statements. The district attorney purposefully excluded Guil-foyle — the only witness with personal knowledge of the blond girl’s statements — from the six witnesses who testified as part of his direct case. Only after Levasseur took the stand and stated that the blond girl was his cousin did the district attorney offer Guil-foyle’s hearsay evidence on rebuttal and paraphrase it in his closing argument. At least *194 on the facts of this case, an error occurring in the rebuttal case and repeated once in closing argument — comprising three lines in a 513-page transcript — is not prevalent in the context of a three-day trial in which the opening argument and direct case (the first two days) are free of error. See Brecht, 507 U.S. at -, 113 S.Ct. at 1722 (finding that error comprising less than two pages of a 900-page trial transcript was not prevalent). 4 2. Centrality of the issue affected, by the error The issue primarily affected by the erroneous admission of Guilfoyle’s hearsay statement was the prosecution’s modus oper-andi theory: because Levasseur attempted to pick up the blond girl two weeks after Doe’s rape using the same modus operandi he had used to lure Doe into his truck, Levasseur must have raped Doe. 5 Had mo-dus operandi been the prosecution’s central theory to support Levasseur’s conviction, the constitutional error in admitting the hearsay would likely have been fatal to the prosecution’s ease and necessarily harmful. The prosecution, however, only used the modus operandi evidence to corroborate Doe’s otherwise strong identification of Levasseur. After considering the trial in its entirety, we find that Doe’s clear identification of Levass-eur was the linchpin in the prosecution’s ease, and the modus operandi evidence was merely collateral thereto. In his opening argument, the district attorney concentrated exclusively on Doe’s identification of Levasseur; he did not mention the blond-girl incident. The bulk of the prosecution’s direct case concerned Doe’s identification of Levasseur, to which four of the six prosecution witnesses testified. The testimony of the prosecution’s first witness, Doe, centered on her ability to view her assailant and his truck, the effect of her consumption of two drinks on her perception, her description of her assailant and his truck, and her identification of Levasseur in photo arrays and in the field. Officer Mendes, the prosecution’s second witness, testified to Doe’s condition closely following the rape, concluding that she was not intoxicated, and to her initial description of the incident. Mr. Ralls, the third witness, gave an eyewitness description of the truck from which he saw Doe thrown. He also described Doe’s condition immediately following the rape, detecting no odor of alcohol. The sixth and final witness, Inspector Boutselis, recalled Doe’s detailed description of her assailant, the truck, and the incident, recounted five days after the rape. He also recalled Doe’s reactions to the photo arrays and the field identifications. He then provided the only account of the blond-girl incident in the prosecution’s direct case. As previously stated, this account was limited to what Boutselis saw and heard transpire between Levasseur and the blond girl. In response to Levasseur’s assertion that the blond girl was his cousin, the district attorney presented additional evidence illustrating modus operandi in his rebuttal case. Finally, in his closing argument, the district attorney referred to the blond-girl incident only twice, focusing his comments instead on Doe’s identification of Levasseur. The district attorney’s tactical decision not to introduce the blond-girl incident in his *195 opening argument and to wait until the last witness in his direct case to do so reveals the diminished importance the prosecution placed on the issue of modus operandi Conversely, the primacy of Doe’s identification of Levasseur is evident throughout the trial, being the district attorney’s focus in both his opening and closing arguments and his direct case. 8. Relative strength of the properly admitted evidence The final factor tests the strength of the properly admitted evidence of guilt in deciding whether the error substantially affected the jury. Was the properly admitted evidence so strong that it overwhelmed the impact of the erroneously admitted evidence? Or was the case very close, with evidence evenly balanced on both sides, enhancing the effect of the error on the jury’s verdict? In answering these questions, we heed Brecht’s warning not to assess the potential strength of the properly admitted evidence if offered at a new trial but rather to assess the strength of the properly admitted evidence as that evidence actually was presented at the trial. Brecht, 507 U.S. at -, 113 S.Ct. at 1724 (Stevens, J., concurring) (“The habeas court cannot ask only whether it thinks the petitioner would have been convicted even if the constitutional error had not taken place. Kotteakos is full of warnings to avoid that result.”). We find that this factor also militates against Levass-eur and that the prosecution’s properly admitted evidence was strong enough to minimize the impact of the erroneously admitted hearsay testimony. The prosecution’s evidence of Levasseur’s guilt, stripped of the erroneously admitted evidence, reduces to two broad areas of incrimination: (1) Doe’s identification of Le-vasseur and (2) Levasseur’s encounter with the blond girl. 6 We will consider the strength of each area of incriminating evidence separately, then cumulatively, and finally we will compare the prosecution’s evidence of guilt to Levasseur’s defense. a. Doe’s identification of Levasseur As the linchpin in the prosecution’s case, Doe’s identification of Levasseur is strong inculpatory evidence. It readily satisfies all but one of the traditional indicia of reliability that the Due Process Clause requires of identifications. See Manson v. Brathwaite, 432 U.S. 98, 114, 97 S.Ct. 2243, 2253, 53 L.Ed.2d 140 (1977) (setting out the factors for assessing reliability of identification testimony). Doe had adequate opportunity to view her assailant throughout the incident; her degree of attention during a traumatic experience is presumed to have been acute, despite her having consumed two drinks earlier that evening; her prior description of her assailant accurately fits Levasseur; when she saw Levasseur in his truck, Doe demonstrated a high degree of certainty in identifying him as the perpetrator (and this comes from a woman who appreciated the gravity of the charges as demonstrated by her refusal to identify Levasseur by his photo alone and request to see him in person); and she never positively identified anyone other than Le-vasseur throughout the six-month investigation. Although the six-month lapse between the rape and Doe’s identification of Levass-eur is troubling, its effect is outweighed by the sheer strength and number of the other factors. Officer Mendes’ and Inspector Boutselis’ testimony corroborated Doe’s identification testimony on certain events, such as her initial and subsequent descriptions of her assailant and the truck, her assertion that she was not intoxicated, and her reactions to the photo arrays and field identification. In addition, Mr. Ralls’ eyewitness testimony describing the assailant’s truck was consistent with Doe’s. The only inroads that Levasseur’s counsel was able to make against Doe’s credibility was highlighting her failure to notice a scar on Levasseur’s neck and tattoos 7 on his arm *196 despite her testimony that Levasseur was wearing a tanktop when he raped her. Such an oversight, however, is understandable given the circumstances 8 and would be unlikely to undermine Doe’s otherwise detailed description. b. Levasseur’s encounter with the blond girl After removing the tainted hearsay evidence from our consideration, we conclude that the remaining evidence about the blond-girl incident is still moderately probative of Levasseur’s guilt. Before assessing its probative strength, we recount the evidence as it came in at trial. Inspector Boutselis introduced the blond-girl incident by testifying that two weeks after Doe’s rape he saw Levasseur drive up to an approximately sixteen-year-old blond girl walking on the sidewalk, beckon to her, and say “Come on”; the young girl kept walking and did not get in the truck. Bout-selis also testified that when he and Guilfoyle subsequently pulled Levasseur over, Levass-eur told Boutselis that the girl was his cousin and he was going to give her a ride. When Boutselis asked Levasseur her name, Levass-eur did not answer. Levasseur’s testimony provided a different account of the blond-girl incident. Levass-eur testified that he had been swimming with his aunt and cousin, that he had taken his aunt home first, and that he was dropping his cousin off when Inspectors Boutselis and Gu-ilfoyle arrived on the scene. He testified that his cousin’s name is Tina Guillemette, she was twenty-seven or twenty-eight years old, and he does not know where she lives now. Although she was not a witness to the blond-girl incident, Levasseur’s wife, Judith Levasseur, testified to historical facts underlying Levasseur’s explanation of the incident. Judith Levasseur testified that Tina Guillem-ette is Levasseur’s cousin, she was approximately twenty-eight years old on July 13, 1988, and does not look sixteen years old. Judith Levasseur also testified that Tina Guillemette is not a good friend of her husband’s, it would be unusual for Levasseur to go swimming with Tina Guillemette and his aunt, and she knows how to get to Tina Guillemette’s house in Lowell. To rebut the Levasseurs’ testimony, the district attorney called Inspector Guilfoyle. Consistent with Boutselis, Guilfoyle testified to seeing a young blond girl walking on the sidewalk, Levasseur’s truck pulling up next to her, Levasseur motioning to her, and the girl continuing to walk on. After having spoken with the blond girl in close proximity, Guilfoyle presumed that she was sixteen or seventeen years old. We begin our assessment of the strength of the blond-girl incident by noting that although the Inspectors and Levasseur provided competing explanations of the blond-girl incident, consideration of the nature, source and extent of the contradictions between the stories reveals the weakness of Levasseur’s explanation. Boutselis’ and Guilfoyle’s testimony, Levasseur’s prior statements to the police, and his wife’s testimony all contradict Levasseur’s explanation on significant details. We briefly note each of the contradictions. After talking with the blond girl face-to-face, Guilfoyle testified that the blond girl was sixteen or seventeen years old, and Boutselis estimated that she was sixteen after seeing her from the police car. Levass-eur, however, testified that Tina Guillemette was twenty-eight years old and Levasseur’s wife testified that Tina Guillemette does not look sixteen. Boutselis and Guilfoyle both testified that Levasseur had beckoned and called “Come on” to the blond girl, consistent with offering her a ride, whereas Levasseur testified that he was dropping his “cousin” off. Similarly, Guilfoyle testified that the blond girl was walking along the sidewalk before Levasseur pulled over, which is also inconsistent with Levasseur’s claim that he was dropping her off. Levasseur’s statement to Boutselis immediately following the blond-girl incident contradicted his trial testimony. Boutselis testified that when he questioned Levasseur on *197 the day of the blond-girl incident, Levasseur explained his encounter with the blond girl as offering to give his cousin a ride and he did not provide his cousin’s name when asked. However, Levasseur testified at trial that he was dropping his cousin off and his cousin’s name is Tina Guillemette. Most telling, perhaps, is that even Levass-eur’s wife, testifying to seemingly innocuous background information, indirectly contradicted Levasseur. While Levasseur claimed that he had gone swimming with his aunt and his cousin on July 13,1988, Judith Levasseur testified that Levasseur and Guillemette were not good friends and such an event would have been an unusual occurrence. Similarly, after Levasseur testified that he did not know where Tina Guillemette now lives, Judith Levasseur testified that she knows how to find Guillemette. This testimony raises the question why Levasseur would not have contacted Guillemette to have her testify and corroborate his testimony if his wife knows where she lives. 9 Given her importance to Levasseur’s explanation of the blond-girl incident, Tina Guillemette is notably absent. Having noted the contradictions discounting Levasseur’s explanation, we must assess how strong Inspectors Boutselis’ and Guil-foyle’s testimony about the blond-girl incident is as evidence of Levasseur’s guilt. Even without the hearsay evidence, Boutsel-is’ and Guilfoyle’s testimony presents circumstantial evidence that the blond girl did not know Levasseur and he had offered her a ride. For instance, that Levasseur did not know the blond girl can be inferred from Boutselis’ and Guilfoyle’s testimony that the blond girl kept walking, did not stop and talk with Levasseur and did not get in his truck. Similarly, that Levasseur offered the blond girl a ride can be inferred from the testimony that Levasseur beckoned to her and said “Come on.” These two inferential facts support the prosecution’s suggestion that, two weeks after Doe’s rape, Levasseur followed the same modus operandi with the blond girl as was used to lure Doe into the truck. Because this evidence makes it more probable that Levasseur' raped Doe, it is moderately probative of Levasseur’s guilt. 10 Considering the prosecution’s case cumulatively, we find Doe’s strong identification of Levasseur corroborated by moderately strong evidence of modus operandi. Levass-eur’s defense, in contrast, is too unsound to weaken the prosecution’s ease. c. Levasseur’s defense Levasseur’s defense was misidentification, and it entailed his refutation of the charges and an alibi. Levasseur’s alibi and his refutation testimony are weak exculpatory evidence, however, because Levasseur’s alibi witness was sleeping during the time of the disputed event, and the district attorney impeached Levasseur’s credibility. Levasseur’s wife, Judith Levasseur, was his alibi witness. She testified that on the night of the rape, she and Levasseur obtained take-out pizza (as was their custom on Thursday nights), drove home, ate it, watched television and went to bed around 10:00 p.m. On cross-examination, the district attorney elicited the following damaging information: Judith Levasseur is a heavier sleeper than her husband and she frequently does not awaken when he gets out of bed during the night. Even accepting Judith Levasseur’s testimony as true, she was asleep during the hours surrounding the inci *198 dent and could not consciously account for Levasseur’s presence in bed at that time. The rape occurred sometime after midnight, leaving Levasseur ample time to get out of bed undetected, get dressed and reach the scene of the incident in nearby Lowell. On direct examination, Levasseur resolutely denied that he had raped Doe and testified consistently with his wife that on June 30, 1988, the night in question, they had taken pizza home and- gone to bed. On cross-examination, however, the district attorney impeached Levasseur’s refutation. The district attorney elicited from Levasseur that on July 13, 1988, at the police station he had told Inspector Boutselis that at 9:00 p.m. on June 30, 1988, he could have been driving around downtown Lowell, but he was not sure. 11 Levasseur admitted that he only became sure about his whereabouts after he spoke with his wife and her immediate family. Levasseur conceded, however, that his in-laws were not with Levasseur and his wife on June 30,1988. Therefore, their assistance in reconstructing his whereabouts is dubious. In addition to the impeachment of Levass-eur’s testimony about the blond-girl incident, the district attorney also impeached Levass-eur’s credibility more generally with another logical inconsistency. Levasseur testified on direct that he had offered to give Inspector Boutselis a sperm sample while he was at the police station on July 13, 1988. On cross-examination, however, Levasseur conceded that as of July 13, 1988, the police had not specifically informed him that he was suspected of rape; they had only made vague references to abusing some girl and claims of sexual harassment. With this line of questioning, the district attorney implied that Le-vasseur disclosed a guilty mind by offering a sperm sample when all that was charged was battery and sexual harassment, neither of which necessarily involve the ejaculation of semen. Given the inability of Levasseur’s alibi witness to vouch for his whereabouts and the widespread impeachment of Levasseur’s testimony, Levasseur’s shaky defense case helped the prosecution’s case more than it hurt it. After comparing the prosecution’s case to Levasseur’s defense, we conclude that the strength of the prosecution’s remaining evidence of guilt, relative to Levasseur’s defense, minimizes the impact of the improperly admitted evidence on the jury’s verdict. Having determined that the error did not permeate the record, the issue affected by the error was not central to the prosecution’s case, and the prosecution’s case minus the inadmissible evidence was strong, we find with fair assurance that, in light of the record as a whole, the Confrontation Clause violation did not substantially influence the jury’s verdict, and therefore it was harmless. We now turn to Levasseur’s remaining arguments. C. Pre-trial Identification Levasseur also contends that the district court erred by holding that Doe’s December 17, 1988, identification of him in his truck one-half mile from his home was not unduly suggestive and did not violate the Due Process Clause. We disagree and affirm for the reasons set forth in the well-reasoned opinion of the district court below. Levasseur v. Pepe, No. 93-10832-DPW, slip op. at 14-22 (D.Mass. Feb. 28, 1995). Careful review of the record convinces us that the district court was correct in finding that Doe’s identification of Levasseur, under the totality of the circumstances, possessed sufficient indicia of reliability, as we discussed in section B.3.a. supra, and that counsel’s failure to object to the identification was not ineffective assistance of counsel. D. Ineffective Assistance of Counsel Levasseur contends that the district court erred in concluding that Levasseur was not denied effective assistance of counsel when his trial counsel failed to object to the hearsay testimony of the blond girl, failed to object to the testimony of Levasseur’s prior bad acts offered solely to impugn his character, failed to object to the jury instruction on *199 identification, and failed to call a rebuttal witness claiming to be the blond girl. We disagree and affirm for the reasons set forth in the opinion of the district court below. Levasseur v. Pepe, No. 93-10832-DPW, slip op. at 27-32 (D.Mass. Feb. 28, 1995). Careful review of the record convinces us that the district court was correct to find that only one of the alleged errors amounted to deficient performance — the failure to object to the hearsay testimony of the blond girl — but that no actual prejudice resulted therefrom. Ill CONCLUSION In sum, we find that Levasseur procedurally defaulted on his Due Process Clause and Confrontation Clause claims; the Confrontation Clause violation was harmless error; the pre-trial identification procedure was not unduly suggestive and counsel’s failure to object to the identification was not ineffective assistance of counsel; and Levasseur was not denied effective assistance of counsel. For the reasons articulated, the district court did not err in denying the application for habeas relief. The district court’s judgment is affirmed. Footnotes [1] For a more comprehensive statement of the facts, see Commonwealth v. Levasseur, 32 Mass.App.Ct. 629, 592 N.E.2d 1350, 1351-53, review denied, 413 Mass. 1104, 600 N.E.2d 171 (1992), cert. denied, - U.S. -, 113 S.Ct. 978, 122 L.Ed.2d 132 (1993). [2] Because Levasseur was no longer cooperating with the investigation and probable cause to arrest him did not exist, Inspector Boutselis could not require Levasseur to participate in a line-up. [3] We note that the Fourth and Eighth Circuits have held that once cause and prejudice is found sufficient to excuse the procedural default and an error is found on the merits, no additional harmless-error review is necessary. Hill v. Lockhart, 28 F.3d 832, 839 (8th Cir.1994) ("[I]t is unnecessary to add a separate layer of harmless-error analysis to an evaluation of whether a petitioner in a habeas case has presented a constitutionally significant claim for ineffective assistance of counseh”), cert. denied, — U.S. -, 115 S.Ct. 778, 130 L.Ed.2d 673 (1995); Smith v. Dixon, 14 F.3d 956, 974, 976 (4th Cir.) (en banc), cert. denied, - U.S. -, 115 S.Ct. 129, 130 L.Ed.2d 72 (1994). Because we review the Confrontation Clause issue as the district court decided it, assuming arguendo that Levasseur excused his procedural default, we need not decide whether we agree with our sister circuits. [4] Levasseur would have us consider every reference to the blond-girl incident in determining prevalence. Although such evidence is relevant to the next factor we consider, determining the centrality of the issue affected by the error, it is not relevant here. To determine prevalence, we need only ask how often the error occurred. Because the other references to the blond-girl incident in the record were free of constitutional error, we do not consider them in assessing the error's prevalence. For instance, when Inspector Boutselis testified in the prosecution’s direct case about the blond-girl incident, he limited himself to what he saw and heard transpire between Levasseur and the blond girl. He stated that he saw Levasseur beckon to the blond girl and heard Levasseur call, "Come on.” Because this evidence is free of constitutional error and clearly admissible, we do not consider it in assessing the error's prevalence. [5] Levasseur’s trial counsel failed to object under Massachusetts' rules of evidence to the introduction of the blond-girl incident as a prior bad act offered to show identity (modus operandi). Because 28 U.S.C. § 2254 limits habeas corpus review of a state prisoner's conviction to violations "of the Constitution or laws or treaties of the United States,” we consider this failure only as it implicates the Sixth Amendment right to effective assistance of counsel. [6] No scientific evidence of Levasseur's guilt, such as a semen specimen, exists. Doe's refusal to let anyone touch her following the rape prevented the hospital staff from administering a "rape kit.” [7] Although Levasseur showed his scar to the jury, he did not show his tattoos to the jury. [8] Discussing Doe's failure to notice Levasseur's tattoos, the district court found it reasonable to assume that a rape victim would focus on her assailant's face rather than on his arms. [9] Although a similar question may be asked of the prosecution regarding its failure to call the blond girl to testify, its failure to do so seems more excusable: Inspectors Boutselis and Guil-foyle did not get the blond girl's name and address. After learning from the blond girl that she did not know Levasseur and he had offered her a ride and deducing that Levasseur was following the same modus operandi used to lure Doe into the truck, Boutselis and Guilfoyle immediately pursued Levasseur, who had driven away from the scene moments before. Under these circumstances, one can excuse Boutselis and Guilfoyle for not taking the additional time needed to remain and obtain the blond girl’s name and address. [10] This evidence becomes stronger evidence of guilt, however, considering that the inconsistencies likely damaged Levasseur’s credibility overall and affected the other elements of his defense, i.e., that he did not know Doe and that he was misidentified. [11] In contrast, Boutselis testified that Levasseur had told him that “he was in the Lowell area, the downtown area from 9:00 p.m. on, but he couldn't recall his exact whereabouts or account for his exact whereabouts.”