George Lindstadt v. John P. Keane, Superintendent

JACOBS, Circuit Judge: Petitioner-appellant George Lindstadt (“Lindstadt”) has moved for a writ of ha-beas corpus pursuant to 28 U.S.C. § 2254, challenging his 1989 conviction in New York Supreme Court, Suffolk County, for the sexual abuse of his daughter. Lind-stadt challenges the conviction on the grounds (A) that his criminal trial counsel was constitutionally ineffective; and (B) that the use of closed-circuit television to take the testimony of his nine-year old daughter violated his rights under the Confrontation Clause (and that counsel’s failure to object to this arrangement was another instance of ineffectiveness). The United States District Court for the Eastern District of New York (Sifton, J.) denied Lindstadt’s motion, Lindstadt v. Keane, No. CV-97-1843 (CPS) (E.D.N.Y. Oct. 20, 1998), and denied his pro se motion for reconsideration, Lindstadt v. Keane, No. CV-97-1843 (CPS) (E.D.N.Y. May 18,1999). We agree with Judge Sifton that the closed-circuit televising of the child’s testimony did not under the circumstances violate Lindstadt’s rights under the Confrontation Clause; but we reverse nevertheless, and grant the writ, because Lindstadt has established (i) that trial counsel’s errors, considered cumulatively, amounted to constitutional ineffectiveness, and (ii) that Lindstadt was prejudiced by those errors. Lindstadt has served about eleven years of his 1256 to 25 year sentence. He was convicted on the basis of testimony from his daughter and estranged wife, and on testimony from two experts. Lindstadt testified in his own defense that he committed *194 no abuse, and asserted that the allegations were fabricated by his embittered wife. Lindstadt claims that his lawyer at trial committed numerous errors serious enough to render the defense services constitutionally ineffective. We base our decision on the cumulative effect of four errors: (i) Defense counsel (along with the judge and the prosecution) failed to notice a one-year error in the date of the alleged abuse. Lindstadt’s wife and daughter testified that the first incident of abuse (Lindstadt was convicted on the basis of two incidents) took place in December 1986. It is now conceded that the first incident must have happened, if at all, in December 1985. Since the daughter’s account of the event reflects that Lindstadt was at the time of the incident a member of the household (doing cooking, bathing the child, present at night, etc.), and since Lindstadt lived with his wife and daughter in December 1986 but not in December 1985 (and can prove it), a lawyer who conducted an adequate investigation could have elicited testimony by Lindstadt that he did not live with his daughter at the time of the alleged abuse, and in that way (and others) shaken the credibility of both key prosecution witnesses. (ii) Defense counsel made no effective challenge to the only physical evidence of sexual abuse. Dr. Milton Gordon testified that his observations of the child’s private parts confirmed that she was abused. This testimony was based on unnamed studies that were (a) not requested by defense counsel, (b) not produced by the prosecution, (c) never produced or usefully identified since, (d) essentially unchallenged at trial, and (e) controverted by other easily available, published studies. (iii) Defense counsel announced in his opening that, after the close of the state’s case, Lindstadt and counsel would decide “whether [the prosecutors] have proven their case,” and only “if they have made their case” would Lindstadt testify. Lind-stadt’s testimony in his own defense therefore became an implicit concession that the prosecutor had “made [its] case.” (iv)Defense counsel proffered testimony from two law enforcement witnesses that, before the daughter made allegations of abuse, Mrs. Lindstadt had attempted several times to put her husband in jail for a variety of alleged crimes. This testimony was essential to bolster Lindstadt’s only defense: that his wife had fabricated the allegations of abuse and coached the child to remember them. The testimony was excluded, however, after counsel failed to make the obvious relevance argument to the court. Because we find that these errors, in combination, fall outside the “wide range of professionally competent assistance” and that they “prejudiced the defense,” we conclude that Lindstadt’s petition should be granted. Strickland v. Washington, 466 U.S. 668, 687, 690, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). The government moved for an order allowing Lindstadt’s daughter to testify via closed-circuit television. Lindstadt’s counsel did not object, and the motion was granted. Lindstadt contends that the trial court thereby violated his Sixth Amendment right to confront the witnesses against him, and that his counsel’s failure to object further demonstrates counsel’s ineffectiveness. We disagree. The trial court’s grant of the unopposed motion did not violate Lindstadt’s rights, and the decision to acquiesce is not ineffective because the decision was supportable for tactical reasons. BACKGROUND On June 17, 1988, Suffolk County Grand Jury indicted Lindstadt on one count of first degree rape, three counts of first degree sodomy, and one count of first degree sexual abuse. All counts involved Lindstadt’s alleged abuse of his daughter. *195 A. Pre-Trial Motion for Closed-Circuit Testimony Before trial, the prosecution moved to have the child declared a “vulnerable witness,” pursuant to New York Criminal Procedure Law § 65.20, so that she could testify via closed-circuit television. In support of its motion, the prosecution cited five factors: 1) the alleged crimes were heinous (sexual abuse of a six-year-old girl); 2) the witness was nine years old; 3) the defendant was the witness’s father, and had lived previously with her in the same home; 4) the alleged abuse was part of a course of conduct over two years; and 5) according to a psychologist, the child would be vulnerable to psychological harm if called to testify in open court. The defense filed no opposition. Justice D’Amaro (not the trial judge) granted the motion. B. Suffolk County Trial The trial commenced on March 30, 1989 before Justice Weissman. The prosecution called four witnesses: Lindstadt’s wife; his daughter; Dr. Don Lewittes (an expert child psychologist); and Dr. Milton Gordon (the doctor who examined the child). Lindstadt’s wife testified briefly that there was acrimony in her marriage; that in April 1988 her daughter told her about “sexual acts involving her father”; and that she immediately ordered Lindstadt out of the house and took her daughter to Child Protective Services. Mistakenly, she testified that in December 1986, the first alleged instance of abuse (Counts 1-4 of the indictment), Lindstadt lived with her and their daughter at 62 Fairfax Drive in Co-ram. (As discussed later in some detail, the family did not live in Coram in 1986.) Defense counsel mounted no useful cross-examination, focusing instead on a curious series of questions, including a line of questioning concerning her name. Lind-stadt’s wife had several previous husbands and was apparently thought by counsel to have had an inordinate choice of last names. The child testified as to the nature of the abuse. She too mistakenly testified that the first incident took place in December 1986. On that occasion, Lindstadt is alleged to have performed various sexual acts with his daughter, including intercourse. She testified that, on instructions from her father, she reported this to no one. The child testified to a second incident in March 1988 (Count 5 of the indictment) in which her father performed a sexual act on her until she stopped him. Lindstadt’s lawyer conducted a lengthy cross-examination, limited by numerous objections sustained by the court. He did get out that the child rode horses and was active in gymnastics, activities he cited in his closing as alternative causes of the physical phenomena later described by the prosecution’s medical expert. The prosecution then called a child psychologist, Dr. Don Lewittes, to explain why a child might conceal sexual abuse for years. Dr. Lewittes described the effect of child abuse on the victim, specifically addressing sexual abuse syndrome, a form of post traumatic stress disorder. Dr. Lew-ittes had never examined the child. Defense counsel asked a single question on cross: “Is [Lindstadt’s daughter] your client?” It is not contended that more was required. The next prosecution witness was Dr. Milton Gordon, the pediatrician who examined the child after her mother brought her to Child Protective Services. The doctor gave her general medical history, followed by a discussion of the genital examination. Dr. Gordon found no obvious signs of trauma (bruising, swelling, etc.); her hymen was intact. The doctor did find “bumps and clefts” on the hymen at the “two o’clock and three o’clock and extending down to the five o’clock positions,” a phenomenon that he testified was consistent with sexual abuse. On cross examination, the doctor indicated that he arrived at this critical conclusion based on what he referenced as a “Boston study.” *196 The doctor also discussed “toluidine blue dye” testing, which he claimed can disclose otherwise sub-dermal scarring of the genitalia. When Dr. Gordon applied the dye, it indicated some scarring of the child’s posterior fourchette. The doctor testified that this is consistent with intrusion into or rubbing of the vaginal area due to sexual abuse. That conclusion was based on a “review” conducted by a Dr. McCaully from Johns Hopkins University. Defense counsel tried to elicit an admission that the damage might have been caused by horseback riding, gymnastics, masturbation, or infection; the doctor’s rejection of each of these alternative causes was based largely on the “Boston study” and the “McCaully review” he had referenced earlier. With that, the prosecution rested. The defense put two witnesses on the stand, George Lindstadt and George Lindstadt’s mother, and unsuccessfully proffered the testimony of two probation officers who had supervised Lindstadt following his conviction for drunk driving. Lindstadt testified that he never engaged in sexual conduct with his daughter. He detailed the turmoil of his marriage, which he summarized as “hell.” As to the family finances, he testified that he and his wife had trouble paying the rent, were sometimes evicted, and moved often from dwelling to dwelling. Lindstadt described a good relationship with his daughter, involving games, hiking, horseback riding, and the like. He testified that he regularly helped her with her bath, and that on one occasion helped to clean the girl’s private parts after he noticed her scratching at a mild infection. Lindstadt also suggested that the sexual abuse allegations had been fabricated by his wife out of rancor, though the prosecution successfully objected to much of his testimony along these lines. Lindstadt’s mother testified that Lind-stadt’s relationship with his daughter was happy and appropriate. She also testified that Lindstadt’s wife had attempted to cajole her into repeating any admissions Lindstadt may have made to his mother. Defense counsel then attempted to call Lindstadt’s two probation officers to testify as to their visits to the Lindstadt household. The prosecutor objected on the ground of relevance. The trial judge decided to hear the testimony outside the presence of the jury before ruling on the motion. The two officers testified that as part of their supervision of Lindstadt they observed the Lindstadt household from February 1987 through the summer of 1988. They testified to acrimony in the household, Mrs. Lindstadt’s animosity and vindictiveness toward her husband, and her refusal to help him with the basic tasks of supporting the household (such as driving him to work when the terms of his probation prohibited him from doing so). Significantly, they testified that, prior to the reporting of the sexual abuse, Lindstadt’s wife repeatedly tried to convince the officers that Lindstadt had violated his probation and should be put in jail. Defense counsel made several arguments in support of admissibility, but failed to make the obvious, pivotal argument: the testimony would have corroborated Lindstadt’s defense that his wife fabricated the sexual abuse. The trial judge precluded the testimony. After the court delivered the jury charge, defense counsel asked for the first time that the jury be charged on lesser included offenses. The court rejected counsel’s requests, which were apparently read from a digest, and characterized counsel as “thoroughly unprepared”: I would think in representing your client and as appropriate to the Court as well, you should have a down charge request in writing before the charge.... I admonish you, sir, of that fact right now. This is the first time in 23 years when this has occurred where down charges haven’t been presented to me in the form wanted by their attorney — by defense attorney. *197 The jury found George Lindstadt guilty on counts 1, 3, and 4 (one count of first degree rape and two counts of first degree sodomy) concerning December 1986, and on count 6 (first degree sodomy) concerning April 1988. The jury found Lindstadt not guilty on count 2 (first degree sexual abuse). C. State Appeal With new counsel, Lindstadt appealed to the Appellate Division challenging the trial court rulings that allowed: the daughter’s testimony via closed-circuit television; expert testimony on “sexual abuse syndrome”; and Dr. Gordon’s testimony. The Appellate Division affirmed the judgment. As to the televising of the child’s testimony, the appellate court initially noted that the issue was unpreserved because no objection had been made, but ruled on the merits that no hearing was required because the prosecution’s submission was sufficient. Lindstadt’s other claims were summarily rejected. Lindstadt subsequently moved in Suffolk County Court to vacate his conviction pursuant to New York Criminal Procedure Law § 440.10 on the ground that trial counsel was ineffective. Lindstadt cited the confusion over the date of the first incident, and explained that the error drew into question the testimony of his daughter and wife on the circumstances surrounding the offense. Lindstadt also argued that the child’s testimony did not match her mother’s concerning the date and location of the fifth count. The motion further cited trial counsel’s failures to challenge the closed-circuit testimony, to challenge the admission of Dr. Lewittes’s testimony, and to effectively cross-examine Dr. Gordon. The Suffolk County Court denied the motion on the ground that Lindstadt received “good and meaningful representation” and blamed Lindstadt for any chronological “confusion in the evidence,” as it was “his obligation to advise his counsel.” The court treated Lindstadt’s motion as one seeking to raise “newly discovered information,” and denied it because Lindstadt “knew where he had lived over the years together with his family and the victim.” D. Habeas Petition Lindstadt moved pro se for a writ of habeas corpus pursuant to 28 U.S.C. § 2254 on the grounds that he was denied his Sixth Amendment rights to confront witnesses and to enjoy effective assistance of counsel. The district court denied the writ. See Lindstadt v. Keane, No. CV-97-1843 (CPS) (E.D.N.Y. Oct. 20, 1998) (Sifton, J.); Lindstadt v. Keane, No. CV-97-1843 (CPS) (E.D.N.Y. May 18, 1999) (denying pro se motion for reconsideration). The district court concluded that Lind-stadt’s Confrontation Clause claim was procedurally defaulted in state court, but went on to address (and reject) the claim on the merits. The court held that trial counsel’s errors with regard to the expert testimony did not amount to deprivation of a fundamentally fair trial. The court found that most of the trial decisions attributed by Lindstadt to ineffectiveness could be attributed instead to litigation tactics and strategy. As to counsel’s failure to notice the one-year error, and the ensuing confusions in the testimony of the prosecution witnesses, the district court considered that counsel’s performance may have fallen outside the wide range of professionally competent behavior, but that Lindstadt failed to make the necessary showing of prejudice. Lindstadt moved pro se for reconsideration, adding several affidavits in aid of his claims. This, too, was denied. DISCUSSION A. Ineffectiveness of Counsel 1. Standard of Habeas Review Lindstadt’s ineffectiveness of counsel claim was heard on the merits by way of his N.Y.C.P.L. § 440 motion in the state appellate court. The Supreme Court has ruled that under the Antiterrorism and *198 Effective Death Penalty Act, codified at 28 U.S.C. § 2254, [a] writ may issue only if one of the following two conditions is satisfied — the state-court adjudication resulted in a decision that (1) “was contrary to ... clearly established Federal law, as determined by the Supreme Court of the United States,” or (2) “involved an unreasonable application of ... clearly established Federal law, as determined by the Supreme Court of the United States.” Under the “contrary to” clause, a federal habeas court may grant the writ if the state court arrives at a conclusion opposite to that reached by this Court on a question of law or if the state court decides a case differently than this Court has on a set of materially indistinguishable facts. Under the “unreasonable application” clause, a federal habeas court may grant the writ if the state court identifies the correct governing legal principle from this Court’s decisions but unreasonably applies that principle to the facts of the prisoner’s case. Williams v. Taylor, 529 U.S. 362, 120 S.Ct. 1495, 1523, 146 L.Ed.2d 389 (2000) (quoting 28 U.S.C. § 2254) (O’Connor, J., concurring, writing for the majority in this part) (omissions in original). “The threshold question ... is whether [the petitioner] seeks to apply a rule of law that was clearly established at the time his state-court conviction became final.” Id. at 1511. In a petition for habeas relief alleging ineffective counsel, the question as to whether the matter is governed by existing Supreme Court precedent “is easily answered because the merits of [such] claim[s] are squarely governed by [the Supreme Court’s] holding in Strickland v. Washington.” Id. (citation omitted). The state court decision denying Lind-stadt’s appeal on the basis of ineffective counsel does not reference Strickland; instead, it relies on a standard articulated in People v. Baldi, 54 N.Y.2d 137, 444 N.Y.S.2d 893, 898, 429 N.E.2d 400 (N.Y.1981): “[T]rial tactics which terminate unsuccessfully do not automatically indicate ineffectiveness. So long as the evidence, the law, and the circumstances of a particular case, viewed in totality and as of the time of the representation, reveal that the attorney provided meaningful representation, the constitutional requirement will have been met.” (citations omitted) The standard applied by the state court is not “diametrically different, opposite in character or nature, or mutually opposed” to the standard articulated in Strickland. Williams, 120 S.Ct. at 1519 (internal quotation marks omitted). Nor is Lindstadt’s case governed by any case in which the Court encountered a “materially indistinguishable” set of facts. See id. Relief is appropriate therefore only if the state “unreasonably applie[d] [the governing legal] principle to the facts of the prisoner’s case.” 1 Id. at 1523. 2. Defense Counsel’s Errors The guarantee of counsel in criminal trials protects the fundamental right to a fair trial afforded by the Sixth Amendment. See Strickland, 466 U.S. at 684-85, 104 S.Ct. 2052. To give substance to this right, counsel must be reasonably effective. See, e.g., McMann v. Richardson, 397 U.S. 759, 771 n. 14, 90 S.Ct. 1441, 25 L.Ed.2d 763 (1970) (“It has long been recognized that the right to counsel is the right to the effective assistance of counsel.”) (citations omitted). To demonstrate constitutional ineffectiveness, “[f]irst, the defendant must show that counsel’s performance was deficient.” Strickland, 466 U.S. at 687, 104 S.Ct. 2052. To determine whether a counsel’s conduct is deficient, “[t]he court must ... determine whether, in light of all of *199 the circumstances, the identified acts or omissions were outside the wide range of professionally competent assistance.” Id. at 690, 104 S.Ct. 2062. In gauging the deficiency, the court must be “highly deferential,” must “consider[ ] all the circumstances,” must make “every effort ... to eliminate the distorting effects of hindsight,” and must operate with a “strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance....” Id. at 688-89, 104 S.Ct. 2052. The Strickland standard is rigorous, and the great majority of habeas petitions that allege constitutionally ineffective counsel founder on that standard. Nonetheless, we count four errors made by Lindstadt’s defense counsel that can be said to fall outside the “wide range of professionally competent assistance.” We need not decide whether one or another or less than all of these four errors would suffice, because Strickland directs us to look at the “totality of the evidence before the judge or jury,” keeping in mind that “[s]ome errors [ ] have ... a pervasive effect on the inferences to be drawn from the evidence, altering the entire evi-dentiary picture.... ” Id. at 695-96, 104 S.Ct. 2052. We therefore consider these errors in the aggregate. See Moore v. Johnson, 194 F.3d 586, 619 (5th Cir.1999) (holding that court should examine cumulative effect of errors committed by counsel across both the trial and sentencing); Stouffer v. Reynolds, 168 F.3d 1155, 1163-64 (10th Cir.1999) (“Taken alone, no one instance establishes deficient representation. However, cumulatively, each failure underscores a fundamental lack of formulation and direction in presenting a coherent defense.”); cf. Rodriguez v. Hoke, 928 F.2d 534, 538 (2d Cir.1991) (dismissing case for failure to exhaust claims, but noting, “[sjince Rodriguez’s claim of ineffective assistance of counsel can turn on the cumulative effect of all of counsel’s actions, all his allegations of ineffective assistance should be reviewed together.”). The four errors are as follows: (i) In her statement to the police, Lind-stadt’s daughter said that the first incident of abuse occurred when she was in the first grade, “about December 1986”. The indictment repeats December 1986 as the date. She testified to this date at trial. So did Lindstadt’s wife. In short, everything presented by the state indicated that the first four counts of abuse occurred in December 1986. The prosecution now concedes that this was an error-that the abuse could not have occurred at the time specified in all the evidence introduced at trial. Lindstadt’s daughter stated in her police report that the first abuse occurred at the family home located at 62 Fairfax Drive, the town of Coram. So too say the indictment, the child’s testimony, and the wife’s testimony. However, the Lindstadts did not live at 62 Fairfax during December 1986 (or anywhere else in Coram.) The government concedes that the alleged abuse must have occurred in December 1985 (at 62 Fairfax Drive). The apparent explanation for this error is that though the child’s first-grade year (academic year 1985-86) ended in 1986, the December of her first grade year fell (of course) in 1985. The principal evidence against Lindstadt was the testimony of his daughter and wife. Both made the same one-year error in dating the incident alleged in the first four counts. And though the prosecution now dates the incident in December 1985, the child’s testimony fits the circumstantial details to the erroneous December 1986 date. An effective lawyer who worked out the chronology of the events could have argued convincingly that the child’s errors evidenced manipulative coaching by an adult. Lindstadt’s daughter testified that the residence in which she was abused had a second floor. But the residence at 62 Fair-fax — where mother and child lived in 1985 — had no second floor. More impor *200 tant, the child’s testimony described the household at the time of the abuse as a conventional domestic household, with father, mother, and daughter living together. So, too, does her statement to the police. Lindstadt’s wife testified similarly. In December 1985, however, the defendant was living apart from his wife and daughter, lodging with his mother and sister, as they have both attested. The prosecution has not tried to explain these incongruities. “[C]ounsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.” Strickland, 466 U.S. at 691, 104 S.Ct. 2052. In Williams v. Washington, 59 F.3d 673, 679-682 (7th Cir.1995), a case of alleged sexual abuse of a child by her foster parent, counsel’s failure to investigate was held to constitute ineffective counsel where an investigation would have disclosed information bolstering his client’s credibility and information “indicating] that, given the layout of the home ... the alleged assault could not have taken place as claimed.” See also Williams, 120 S.Ct. at 1514-15 (holding that a failure to investigate and present mitigating evidence during sentencing hearing constituted ineffective assistance, even when doing so would have admitted some unfavorable evidence); Hart v. Gomez, 174 F.3d 1067, 1070-71 (9th Cir.1999) (holding that failure to introduce evidence supporting defendant’s contention that sexual abuse could not have occurred under the circumstances alleged rendered counsel ineffective); United States ex rel. Williams v. Brown, 721 F.2d 1115, 1119—121 (7th Cir.1983) (holding that counsel’s inadequate preparation for trial, including counsel’s failure to sufficiently consult with defendant, constituted ineffective counsel); United States v. Tucker, 716 F.2d 576, 579-586, 595 (9th Cir.1983) (same). If defense counsel had investigated the case and ascertained that the child was in the first grade in December 1985 rather than 1986, Lindstadt would have been able to demonstrate that the corroborative details in the child’s testimony concerning the first sexual assault — her father’s presence in her home cooking meals, bathing her and running the household; her memory of him ascending the stairs of the home on the date of the molestation — reference the household in December 1986 (when, as everyone now concedes, the assault did not take place) and is incompatible with the family’s arrangements in December 1985 (the month in which the assault is now alleged to have taken place), at which time mother and daughter lived in a one-story house, and the defendant was residing apart from the pair, with his own parents. Counsel could have exploited these holes in the prosecution’s case to shake the testimony of the only percipient witnesses; to argpe that the mother was inventing impossible details and circumstances; to suggest that the child has absorbed those misimpressions by coaching from her mother; to elicit testimony from the defendant that he was not living with his wife and daughter when (the daughter testified) she was abused by her stay-at-home father; and to gather the easily available corroborative testimony of the defendant’s mother and sister, with whom he was living at the critical time. The information needed to accomplish these defense objectives was available from a variety of sources, including Lind-stadt himself and the police report. In fact, during defense counsel’s examination of Lindstadt on unrelated points, Lindstadt testified that, in early 1986 the family lived in Ridge and then moved to Middle Island later in the year. Counsel’s errors prevented Lindstadt from offering something akin to an alibi: Lindstadt was not living with his daughter in December 1985. Cf. Brown v. Myers, 137 F.3d 1154, 1156-57 (9th Cir.1998) (taking it as given that a failure to adequately investigate alibi claim or witnesses constitutes ineffective counsel); Bryant v. Scott, 28 F.3d 1411, 1418 (5th Cir.1994) (holding that a failure to adequately investigate alibi witnesses constitutes ineffective coun *201 sel); Nealy v. Cabana, 764 F.2d 1173, 1177-78 (5th Cir.1985) (same); Noble v. Kelly, 89 F.Supp.2d 443, 463 (S.D.N.Y.2000) (holding that, in the absence of a strategic explanation, the failure to properly call an alibi witness constituted constitutionally ineffective counsel). For obvious reasons, there is no record evidence that Lindstadt had access to the house or the child in December 1985, when the offense is now said to have happened; and if the prosecution had had an opportunity to augment its evidence to show parental access, no one can say what the particulars of that showing would be. The prosecution emphasizes that defense counsel was not alone in failing to notice these errors, a point also made by the state appellate court and the district court. We do not think that the failure of others to notice establishes that the point is so subtle that an effective counsel would miss it. A trial judge has a full plate without undertaking an independent study of the facts on which the parties seem to agree. The prosecutor may (though should not) have other priorities. And it is understandable that the defendant — who was being tried for child rape on the word of his own wife and daughter — was distracted enough to leave the basics of his defense to his counsel. Defense counsel investigating a case of child molestation should be expected to ascertain the child’s age at the time of the event, to consider the how many years elapsed before the trial testimony, to read the child’s prior statements closely, and to check school performance for signs of psychological trauma. These elementary steps would have uncovered the prosecution’s error. (ii) Defense counsel made no challenge to the only physical evidence that the prosecution introduced. The only physical evidence of abuse was the testimony of Dr. Gordon, consisting of two observations: (i) bumps and clefts on the child’s hymen; and (ii) sub-dermal scarring of her posterior fourehette that became detectable when a special dye was applied. The only link between the “bumps and clefts” identified by Dr. Gordon and the sexual abuse charged against Lindstadt was a “Boston study” referenced during the doctor’s testimony. Defense counsel never requested a copy of this study. It was not introduced into evidence. It was never referred to with any greater specificity than the “Boston study.” Lindstadt’s appellate counsel could not locate the study. The prosecutor admitted at oral argument before this Court that the study is not in his file and that he has no knowledge of it. Defense counsel therefore could not cross-examine Dr. Gordon about where or whether the study was published, who prepared it, whether it was peer reviewed, or even what it found. We cannot say that such cross-examination would have been fruitful. But counsel’s failure to seek the study was an amazing dereliction. Moreover, there is no evidence that defense counsel contacted an expert, either to testify or (at least) to educate counsel on the vagaries of abuse indicia. See generally Beth A. Townsend, Defending the “Indefensible”: A Primer to Defending Allegations of Child Abuse, 45 A.F.L.Rev. 261, 270 (1998) (“It is difficult to imagine a child abuse case ... where the defense would not be aided by the assistance of an expert.”). Such an expert could have brought to light a contemporaneous study, accepted for publication at the time of Lindstadt’s trial, that found similar irregularities on the hymens of girls who were not abused. See John McCann et al., Genital Findings in Prepu-bertal Girls Selected for Nonabuse: A Descriptive Study, 86 Pediatrics, No. 3, at 428-439 (Sept.1990); see also Townsend, 45 A.F.L.Rev. at 269-270 (discussing studies that indicate the presence of clefts on the hymen of non-abused girls); J. Gardner, Descriptive Study of Genital Variation in Healthy, Nonabused Premenarchal Girls, 120 J. Pediatrics 258-260 (Feb.1992). Dr. Gordon linked the sub-dermal scarring to sexual abuse, relying on a “review” conducted by a doctor from Johns Hop *202 kins, identified in the transcript (apparently phonetically) as “Dr. McCaully.” Again, defense counsel did not request this material, nor was it introduced into evidence, nor was it referred to in detail. Counsel had no way to test or evaluate the scientific basis of Dr. Gordon’s assertions. Dr. Gordon admitted that the study was “based on someone else’s work,” though it is unclear what that means. And again, Lindstadt’s appellate counsel was able to locate contemporaneous studies that cast doubt on any link between (i) the scarring of the posterior fourchette disclosed by the toliudine dye test and (ii) sexual abuse. See David Muram, Child Sexual Abuse: Relationship Between Sexual Act and Genital Findings, in 13 Child Abuse and Neglect 211-16 (1989); R.G. Teixeira, Hymenal Colposcopic Examination in Sexual Offenses, 2 Am. J. Forensic Med. Pathology 209-215 (1980); see also M.D. Dowd et al., The Interpretation of Urogenital Findings in Children with Straddle Injuries, 29 J. Pediatric Surgery, No. 1, at 7-10 (Jan. 1994). In sum, defense counsel’s failure to consult an expert, failure to conduct any relevant research, and failure even to request copies of the underlying studies relied on by Dr. Gordon contributed significantly to his ineffectiveness. See Holsomback v. White, 133 F.3d 1382, 1387-89 (11th Cir.1998) (holding that failure to conduct adequate investigation into medical evidence of sexual abuse was ineffective); Knott v. Mabry, 671 F.2d 1208, 1212-13 (8th Cir.1982) (noting that counsel may be found to be ineffective for failing to consult an expert where “there is substantial contradiction in a given area of expertise,” or where counsel is not sufficiently “versed in a technical subject matter ... to conduct effective cross-examination.”); cf. Tucker, 716 F.2d at 581 (holding that, in complex fraud case, “it should have been obvious to a competent lawyer that the assistance of an accountant [was] necessary.”). The prosecution and the district court’s opinion point to the cross-examination of Dr. Gordon that defense counsel did conduct as evidence that his performance was satisfactory. This effort was hamstrung, however, by counsel’s lack of familiarity with the studies upon which Dr. Gordon was presumably relying; the effect was ruinous because Dr. Gordon testified that the (unidentified) studies ruled out every theory of innocent injury (such as horseback riding) posited by the defense. The two remaining errors of counsel, which bear upon the case presented by the defense, are less compelling and would not alone suffice. But when they are added to the first two, the cumulative weight of error convinces this Court that the ineffectiveness of counsel reached the constitutional threshold. (iii) Lindstadt’s counsel announced in his opening that Lindstadt was under no obligation to testify; that, after hearing the state’s evidence, Lindstadt and counsel would decide “whether [the prosecutors] have proven their case”; and that, only “if they have made their case,” Lindstadt would testify, (emphasis added) This gratuitous comment compelled Lindstadt to pay a heavy price for the exercise of his constitutional right to give testimony in his defense: by mounting the witness stand, he was conceding the prosecution had “proven their case.” And the only way a prosecutor can prove a case is beyond a reasonable doubt. See United States v. Swanson, 943 F.2d 1070, 1075 (9th Cir.1991) (holding that “[o]nce [counsel] told the jury that there was no reasonable doubt regarding his client’s identity as the perpetrator of the crime charged against him, he ceased to function as defense counsel.”). In the circumstances of this case (Lind-stadt was not defending on the basis of entrapment or insanity, for instance), the only tactical justification for such a statement would be a prior irrevocable decision by the defendant to avoid the stand. But such a decision is not for counsel to make; and here the defendant was the only possible witness to his conduct available to the defense, and was not burdened by a histo *203 ry of similar acts or convictions, or other damaging circumstances. We think that this error reflected no tactical considerations and contributed to defense counsel’s constitutional ineffectiveness. Cf. Stouffer, 168 F.3d at 1163 (holding that defense counsel’s failure to make an opening statement “serve[d] to underscore the lack of any discernible effort by Petitioner’s counsel to present a defense.”). (iv) Defense counsel called to the stand Lindstadt’s two probation officers. (Lind-stadt was then on probation following a conviction for drunk driving.) Both testified^ — outside the presence of the jury— that in the period immediately prior to the reporting of the abuse, Lindstadt’s wife repeatedly tried to convince the officers that he had violated his probation and should be jailed. Mrs. Lindstadt accused him of drinking (in violation of the conditions of his parole), kidnaping his daughter, stealing his wife’s car, and sexually abusing children of his wife by a previous marriage. There is no corroboration for any of these accusations, but they show that Lindstadt’s wife wanted him in jail— even before the daughter is said to have reported the sexual abuse. The prosecution objected to this testimony; defense counsel argued that the testimony was needed to show the “general atmosphere in [the household]” and the “battle between this husband and wife to the point that the wife now turns around and uses the child to get back at the husband.” The trial judge ruled that the testimony was insufficiently relevant; husband and wife had by then both testified that the household atmosphere was bad. But the obvious point, which defense counsel failed to make, is that the testimony of the parole officers would show that Lind-stadt’s wife wanted him in jail even before she claims to have learned of the sexual abuse. 2 Such testimony would have supported an inference that Lindstadt’s wife fabricated the abuse charges and coached her daughter to lie in order to jail her husband. Nothing could have been more relevant, for “[i]n a credibility contest, the testimony of neutral, disinterested witnesses is exceedingly important.” Williams, 59 F.3d at 681-82 (holding that counsel’s failure to investigate and call witnesses who could bolster client’s defense of innocence in a case alleging sexual abuse constituted ineffective counsel). The district court found that there was a sound strategic consideration for counsel to withhold an obvious and compelling relevancy argument, because evidence that Lindstadt’s wife accused him of other crimes might lead the jury to believe that he actually committed them. This analysis makes sense, but it was not counsel’s strategic consideration. If it were, counsel would not have called these witnesses in the first place or (if he developed misgivings) could have withdrawn his initiative to put them on the stand. Taken together, ineffectiveness permeated all the evidence. Counsel’s inept opening statement assured that, before Lindstadt opened his mouth to testify, he admitted that the prosecution had made its case. Mrs. Lindstadt’s testimony was demonstrably wrong and was given by a spouse who (other witnesses could have shown) wanted her husband in jail even before she claimed to have learned of the abuse. The child’s testimony had signifi *204 cant factual errors that made it vulnerable to cross-examination tending to establish that she had been coached. The only physical evidence was offered by an expert relying on unnamed studies; but counsel did not seek or review the studies, search the available literature for rebuttal, or retain an expert to give advice or to testify. We assess the impact of these errors in the aggregate. See Strickland, 466 U.S. at 695-96, 104 S.Ct. 2052. In that light, counsel’s services to Lindstadt fell “outside the wide range of professionally competent assistance.” Id. at 690, 104 S.Ct. 2052. 3. Prejudice “An error by counsel, even if professionally unreasonable, does not warrant setting aside the judgment of a criminal proceeding if the error had no effect on the judgment.” Strickland, 466 U.S. at 691, 104 S.Ct. 2052. To merit habeas relief, “the defendant must show that the deficient performance prejudiced the defense.” Id. at 687, 104 S.Ct. 2052. The level of prejudice the defendant need demonstrate lies between prejudice that “had some conceivable effect” and prejudice that “more likely than not altered the outcome in the case.” Id. at 693, 104 S.Ct. 2052. Thus “[t]he defendant must show that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Id. at 694, 104 S.Ct. 2052. The Court defined “reasonable probability” as one that “undermine[s] confidence in the outcome.” Id. Even serious errors by counsel do not warrant granting habeas relief where the conviction is supported by overwhelming evidence of guilt. For example, in Scarpa v. DuBois, 38 F.3d 1, 16 (1st Cir .1994), the court concluded that, “despite [defense counsel’s] ineptitude,” the uncontroverted eyewitness testimony against the defendant overcame any possible prejudice caused by counsel’s errors. Likewise, in United States v. Reiter, 897 F.2d 639, 645 (2d Cir.1990), counsel was so frequently late and absent from the trial that the trial court judge jailed her for contempt to ensure her appearance. We nonetheless found that the requirements of Strickland had not been met, as there was “overwhelming record support” for defendant’s guilt. See also Wise v. Smith, 735 F.2d 735, 739 (2d Cir.1984) (“overwhelming” eyewitness testimony and physical evidence indicating that the defendant committed the crime overcame any possible prejudice). On the other hand, “a verdict or conclusion only weakly supported by the record is more likely to have been affected by errors than one with overwhelming record support.” Strickland, 466 U.S. at 696, 104 S.Ct. 2052. Here, for reasons already stated in this opinion, we conclude that counsel’s errors prejudiced the defendant to the extent that they “undermine confidence in the outcome.” Id. at 694, 104 S.Ct. 2052. There were only two people with any direct knowledge as to whether Lindstadt sexually abused his daughter. Counsel’s statement that Lindstadt would testify only if the prosecution had “proven its case” was a boomerang. And Counsel’s failure to investigate prevented an effective challenge to the credibility of the prosecution’s only eyewitness, Lindstadt’s daughter. See Williams v. Washington, 59 F.3d 673, 682-685 (7th Cir.1995) (“Counsel’s lack of familiarity with the case, combined with his failure to investigate, provided [the defendant] with a trial significantly different than she might have received if represented by a competent attorney.”); United States ex rel. Williams v. Brown, 721 F.2d 1115, 1120 (7th Cir.1983) (holding that counsel’s inadequate preparation for trial prejudiced defendant); United States v. Tucker, 716 F.2d 576, 592-95 (9th Cir.1983) (same). Only two witnesses were offered to corroborate the prosecution’s account of the abuse: Lindstadt’s wife and Dr. Gordon. For reasons stated, counsel’s errors impaired his ability to mount an effective cross examination of these witnesses. In *205 the case of Dr. Gordon, counsel’s failure to use an expert or take other preparatory measures allowed the prosecution to dress up the weak and inconclusive physical evidence in the trappings of Dr. Gordon’s expertise. See Holsomback, 133 F.3d at 1388 (holding that, in a case turning on credibility, the failure to adequately investigate the physical evidence concerning whether sexual abuse occurred prejudiced the defendant); cf. Chatom v. White, 858 F.2d 1479, 1487 (11th Cir.1988) (holding that failure to make warranted motion to exclude damaging circumstantial evidence prejudiced the defendant). Only two witnesses could corroborate Lindstadt’s defense: his probation officers. The probation officers, as outside, impartial observers of the Lindstadt household, could have provided essential corroboration for Lindstadt’s testimony that his wife was vindictive; and their testimony would have put the case in the context of her previous attempts to get her husband jailed. This is a case of underwhelming evidence. See, e.g., Hart v. Gomez, 174 F.3d 1067, 1072-73 (9th Cir.1999) (holding that, even with the defendant’s confession to abusing child on previous occasions, there was not overwhelming evidence that defendant abused child on charged occasion). All of the evidence against Lindstadt was affected by his counsel’s failures; and but for those failures, the prosecution’s case quite likely would have collapsed altogether. The district court observed that “[t]he testimony regarding what petitioner did lay at the heart of his conviction. Where and when he did it did not.” However, aside from the testimony of Dr. Gordon, the only evidence of abuse came from his wife and daughter; their credibility lay at the heart of the prosecution’s case. Had defense counsel mounted a proper challenge to their testimony as to where and when the abuse occurred, it would have naturally called into question whether the abuse occurred at all. Lindstadt has demonstrated that his counsel’s services fell “outside the wide range of professionally competent assistance” and that “there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Strickland, 466 U.S. at 690, 694, 104 S.Ct. 2052. Accordingly, the state appellate court’s decision to deny Lindstadt’s claim for relief on the ground that he was afforded ineffective assistance of counsel “involved an unreasonable application of clearly established Federal law, as determined by the Supreme Court of the United States.” 28 U.S.C. § 2254(d)(1). B. Spillover The prosecution argues that, since the factual errors in the testimony bear upon the first incident of alleged abuse (the December 1985 episode) and not to the second (the March 1988 episode), the jury could still have found Lindstadt guilty on the count relating to the second incident (for which he received the same concurrent sentence). Although counsel’s error bears directly only on Counts 1-4, a court that overturns fewer than all criminal counts must consider the spillover effect on the remaining counts. See, e.g., United States v. Morales, 185 F.3d 74, 82 (2d Cir.1999). Three factors guide the inquiry on impermissible spillover: 1) whether the evidence on the vacated counts was inflammatory and tended to incite or arouse the jury to convict the defendant on the remaining counts; 2) whether the evidence on the vacated counts was similar to or distinct from that required to prove the remaining counts; and 3) the strength of the government’s case on the remaining counts. United States v. Naiman, 211 F.3d 40, 50 (2d Cir.2000). *206 As to the first factor: sexual abuse of a child is a heinous crime and those who commit it are thought by many to be serial offenders and incorrigible; a jury that finds a defendant guilty of three counts of sexual abuse on one occasion is primed to find the defendant guilty of another. As to the second factor: to the extent the evidence differs, there is less to support Count 5, and all of it is infected by the same errors affecting the other counts. True, not all of counsel’s errors bear on Count 5; but we see no good reason to refuse vacatur of Count 5 simply because all of the evidence supporting it is infected by less than all of the errors affecting the other counts. As to the third factor: Count 5 was the weakest of the lot; it relied solely on the word of Lindstadt’s wife and daughter, and was not buttressed by physical evidence or expert testimony of any kind. C. Closed-Circuit Testimony of the Child The state appeals court ruled that Lind-stadt’s claim concerning the alleged violation of his Confrontation Clause rights was unpreserved for appellate review because counsel failed to object at trial, but proceeded to reject the claim on its merits, finding sufficient evidence of vulnerability to warrant presenting the child’s testimony by closed-circuit television. The court’s analysis on the merits applied the New York Court of Appeals decision in People v. Cintron, 75 N.Y.2d 249, 552 N.Y.S.2d 68, 551 N.E.2d 561 (1990), which anticipated the holding of Maryland v. Craig, 497 U.S. 836, 110 S.Ct. 3157, 111 L.Ed.2d 666 (1990) (holding that the Confrontation Clause does not categorically prohibit the presentation of a child’s testimony via closed-circuit testimony in cases alleging sexual assault). For the reasons stated above, the state court’s application of Cintron and Craig is reviewed on habeas for “unreasonable application” of established Federal law, as determined by the Supreme Court. Because we grant Lindstadt’s motion on an unrelated ground, it is unnecessary to reach the merits of his Confrontation Clause claim. Were we to reach this matter, we would likely uphold the district court’s decision to deny Lindstadt’s petition for habeas relief on this ground on substantially the same reasons articulated by the district court. Lindstadt claims on appeal that his counsel’s failure to oppose the motion was evidence of counsel’s ineffectiveness. But it is easy to think of tactical reasons why a defense counsel might wish a witness to testify via closed-circuit television: the witness might become emotional in court; the live witness may prove more compelling or sympathetic; etc. We do not think that this decision contributed to counsel’s ineffectiveness. CONCLUSION Accordingly, we reverse the decision of the District Court and remand for the entry of judgment conditionally granting the writ and ordering George Lindstadt’s release unless the State provides him a new trial within 90 days. The mandate will issue forthwith. Footnotes [1] The Court has not defined the term "unreasonable” application, because "it is a common term in the legal world and accordingly, federal judges are familiar with its meaning.” Williams, 120 S.Ct. at 1522. The Court nevertheless noted that the term differs from "erroneous” or "incorrect,” and sets a higher bar to relief than either of those terms. See id. [2] See generally Jacqueline M. Beckett, The True Value of the Confrontation Clause: A Study of Child. Sex Abuse Trials, 82 Geo. L.J. 1605, 1635-36 (1994) ("[Ajlmost all situations involving child sex abuse accusations promote the reward-seeking behavior that leads to lying. These allegations are most common when parents involved in marital disputes accuse their spouses of abusing their children. In these situations, the child may be willing to exchange her testimony for the love and attention that she badly craves.”); Brian L. Schwalb, Child Abuse Trials and the Confrontation of Traumatized Witnesses, 26 Harv. C.R.-C.L.L.Rev. 185, 185 n. 4, 185-86, (1991) (discussing studies showing false claims of sexual abuse motivated by vindictive spouses, noting that "false allegations of abuse have [ ] become disturbingly common”).