SMITH, Presiding Judge.

A jury found Gary Goldstein guilty of child molestation and aggravated sexual battery. His amended motion for new trial was denied, and he now appeals, raising several enumerations of error. In addition to raising the general grounds, Goldstein contends that the trial court erred by clearing the courtroom for the victim's testimony and that his trial counsel provided ineffective assistance. Upon review, we find that the evidence presented at trial was sufficient to authorize the jury to find Goldstein guilty and that the trial court did not err in temporarily closing the courtroom during the young victim's testimony. For several reasons, however, we conclude that although Goldstein's trial counsel was experienced and generally able, his assistance in this trial was ineffective, particularly because he inexplicably failed to present evidence known to him that bears on several important issues and which could have changed the outcome of the trial. We therefore reverse the trial court's denial of Goldstein's motion for a new trial.

1. We first address Goldstein's contention that the evidence presented at trial was insufficient to support his convictions. Viewed to support the jury's verdict, the evidence showed that the incident upon which the charges against Goldstein were based occurred at a family dinner at Goldstein's home on April 11, 2003. A number of other family members were present, including Goldstein's wife Joanne and their two young children, Joanne's sister Esther, Esther's seven-year-old daughter (the victim), Esther's four-year-old daughter, Joanne's parents, and Goldstein's parents. The chronology of some events is disputed, but it is clear that at some point the children left the dinner table and went upstairs to play. At some point, Goldstein went upstairs for a short time to help one of his sons with a broken toy, and he returned shortly thereafter. At one time during the evening, the victim sat on Goldstein's lap at the dinner table. Also at some point, the victim wet her pants and told her mother. According to her mother's subsequent testimony, it was unusual for the victim to wet her pants. The victim then changed into sweat pants belonging to one of her cousins, and she and her family left shortly thereafter.

At home, while in the bathtub the victim told her mother that Goldstein "stuck his finger in my vagina tonight," demonstrating with her mother's finger. Her mother quickly withdrew her finger, and the victim told her that Goldstein's finger had gone "much further than that." When the victim's father came home from work later in the evening, the victim told him as well. To her father, she added that Goldstein's act "felt kind of good." Her mother testified that since that time, the victim's behavior had changed. She wets her pants often and has nightmares.

A Fulton County police officer testified that Esther reported the incident ten days later, on April 21, 2003, telling him that she learned of the incident when she discovered her daughter in an upstairs bathroom at home upset and crying. A videotaped interview with the victim was played for the jury.

A pediatrician testified that she saw the victim on April 22, 2003 and found no signs of trauma. According to the pediatrician, neither the absence of indications of trauma nor the victim's professed lack of pain was surprising, and these results did not indicate that the victim had not been digitally penetrated. Nancy Aldridge, a forensic interviewer, testified that she spoke with the victim's parents before interviewing the victim because it was important to know what the parents told the victim when the victim made her disclosure. The victim also testified, and before her testimony began, the trial court closed the courtroom over Goldstein's objection. The victim identified Goldstein as the person who placed his finger in her vagina. She testified that after he did so she went back to playing with the other children and forgot about the penetration. She went downstairs and rejoined the adults, at some point sitting on Goldstein's lap.

Other family members testified that Goldstein was never out of their sight that evening until after the victim wet her pants. Goldstein took the stand in his own defense, testifying that he was at the dinner table all evening before the victim wet herself. According to Goldstein, the first time he saw the victim in the upstairs bedroom of one of his sons, she had already changed into his other son's sweat pants. Goldstein stated that this occurred after he went upstairs to repair a toy for one of the boys, who told him that his other son was in his room. He testified that when he opened the bedroom door to check on his son, he saw the victim with the sweat pants slightly lowered, standing only a few inches from his son and with the other girl in the room. After he opened the door, the victim turned around and said, "Hey, you can't see my vagina." When he reprimanded her in front of the other children, she became embarrassed. He denied touching the victim inappropriately.

The evidence was in conflict, and the jury obviously believed the testimony of the child victim and not that of Goldstein. The victim's testimony alone was sufficient to authorize the jury to find Goldstein guilty under the standard of Jackson v. Virginia, 443 U. S. 307 (99 SC 2781, 61 LE2d 560) (1979). See Pate v. State, 269 Ga. App. 684, 687 (2) ( 605 SE2d 90) (2004).

2. Goldstein maintains that the trial court committed reversible error in clearing and closing the courtroom before the victim testified. We do not agree. The State requested that the courtroom be cleared, and Goldstein objected on the ground that he had a right to a public trial and that the victim, although a child, had told her story before, in interviews and in therapy. The trial court granted the State's request, noting that on those occasions only a few individuals from the victim's family were present. Those interviews were unlike the courtroom situation, in which many strangers would be present, comprising a 12-person jury and a courtroom full of other people. The trial court allowed experts from the State and the defense to remain in the courtroom during the child's testimony, and the courtroom was reopened immediately following the child's testimony.

Contrary to Goldstein's argument, such limited closure did not violate his Sixth Amendment right to a public trial. Moore v. State, 151 Ga. 648, 653 ( 108 SE 47) (1921). Although a strong presumption exists in favor of openness, reasonable limitations may be imposed for good reason. Waller v. Georgia, 467 U. S. 39, 45 (II) (A) (104 SC 2210, 81 LE2d 31) (1984).

The presumption of openness may be overcome only by an overriding interest based on findings that closure is essential to preserve higher values and is narrowly tailored to serve that interest. The interest is to be articulated along with findings specific enough that a reviewing court can determine whether the closure order was properly entered.

(Citation and punctuation omitted.) Id. "[T]he party seeking to close the hearing must advance an overriding interest that is likely to be prejudiced, the closure must be no broader than necessary to protect that interest, the trial court must consider reasonable alternatives to closing the proceeding, and it must make findings adequate to support the closure." Id. at 48 (II) (B).

It is clear that under the proper circumstances, a courtroom may be closed. Indeed, Georgia has a statute that provides specifically for partial closure of the courtroom when a person under the age of 16 testifies about a sex offense. OCGA § 17-8-54. The statute was "based upon a legislative determination that there is a compelling state interest in protecting children while they are testifying concerning a sex offense." (Citation and footnote omitted.) Hunt v. State, 268 Ga. App. 568, 571 (1) ( 602 SE2d 312) (2004). This court has held that the partial closure under Georgia's statute does not violate a defendant's Sixth Amendment right to a public trial. Id. at 569-571 (1). Here, as in Hunt, we conclude that the trial court's temporary partial closure, effected to protect the young victim, was justified. Id. at 571 (1).

3. Goldstein's remaining arguments focus on his assertion that his trial counsel provided ineffective assistance. To prevail on a claim of ineffective assistance of counsel, a defendant must show that his attorney's performance was deficient and that the deficiency so prejudiced him that a reasonable likelihood exists that but for counsel's deficiency the outcome of the trial would have been different. Strickland v. Washington, 466 U. S. 668 (104 SC 2052, 80 LE2d 674) (1984). A strong presumption exists that counsel's conduct falls within the broad range of reasonable professional conduct. Harris v. State, 279 Ga. 304, 306 (3) ( 612 SE2d 789) (2005). As a reviewing court, we must uphold the trial court's findings of fact unless they are clearly erroneous. Morris v. State, 280 Ga. 184, 185 (2) ( 625 SE2d 391) (2006). We review the trial court's legal conclusions de novo. Id. Goldstein cites a number of instances of alleged ineffectiveness.

(a) Goldstein maintains that his trial counsel was ineffective in failing to cross-examine the victim's mother regarding her many prior allegations of child molestation. At the hearing on the motion for new trial, a number of witnesses testified that these allegations were false. The victim's mother first made such an allegation when she was a child and has continued to do so into her adulthood. Witnesses, including her own mother (the grandmother of the victim), her sister (Goldstein's wife), her two brothers (who were themselves accused), and family friends (one of whom was also accused), testified at the hearing on the motion for new trial.

The grandmother testified that when Esther was about nine years old, she accused a family friend of molesting her sister. She also testified that Esther had repeated this allegation quite recently, once at a family dinner, and again when she and Esther encountered each other in a grocery store in 2004. The grandmother testified that Esther also claimed that her mother (the grandmother) herself had been molested as well. Esther did not deny making the accusations, testifying that her mother and her mother's two sisters all had been molested. Joanne also testified at the hearing on the motion for new trial. She recalled a family gathering at her home at which her sister Esther accused a family friend of having molested Joanne. Joanne testified that she was never molested by the family friend.

Esther and Joanne's two brothers also testified at the hearing. One brother testified that Esther had accused her father of molesting her. He testified that in addition to accusing their father and a family friend, she also accused him and their other brother of molesting her. He also related an incident at a dinner table, when he saw their father touch his sister Esther. At her father's touch, Esther immediately jumped up and fled the house, screaming that their father was a molester. According to this brother, he was himself falsely accused of molestation by Esther when he was 11 or 12 years of age. Esther's other brother testified that allegations of molestation leveled against him by Esther were also false. The male family friend accused by Esther testified that he was a friend of the family for approximately 20 years and that he had never molested Joanne. A female friend of the family testified that during a telephone conversation in August 2005, Esther told her that, given that a family friend had molested her sister Joanne when she was a child, it was surprising that Joanne did not believe that Goldstein had molested the victim. This witness also testified that Esther told her that Goldstein had molested the victim on two or three prior occasions and that this time was only "the last time and the final straw."

Most of these witnesses testified that they told Goldstein's trial counsel about what they claimed were false allegations made by Esther; almost all of them testified they had never been contacted to testify at trial. This court has held that situations exist in which such allegations are only marginally relevant or in which trial strategy calls for omitting evidence of them. See, e.g., Wand v. State, 230 Ga. App. 460, 463 (2) (c) ( 496 SE2d 771) (1998) (physical precedent only). But see id. at 466 (Smith, J., concurring specially) (disagreeing that in that case the mother's testimony was "marginal" because prior false allegations could impact on mother's "credibility and propensity for making such allegations").

In other situations, however, an outcry witness's prior false allegations are highly relevant. See In the Interest of M. G., 239 Ga. App. 787 ( 521 SE2d 918) (1999). In M. G., the victim did not report to anyone that she had been molested until after her grandfather "repeatedly asked what was wrong with her." Id. at 790. The trial court refused to permit the defendant to cross-examine the grandfather. We noted that

[t]he purpose of cross-examination is to provide a searching test of the intelligence, memory, accuracy, and veracity of the witnesses, and it is better for cross-examination to be too free than too much restricted. Whenever the purpose is to impeach or discredit the witness, great latitude should be allowed by the court in cross examinations.

(Citations omitted.) Id. Citing with approval the special concurrence in Wand, this court held in M. G. that "the trial court abused its discretion when it refused to allow the defendant to cross-examine [the grandfather] about a prior false allegation of molestation. [Cit.]" Id.

In M. G., although refusal to permit cross-examination of the outcry witness regarding prior false allegations was error, we concluded that it was harmless. Here, however, we conclude that the number of prior complaints made by the victim's mother against so many different individuals, who all claim that such complaints are false, and her persistence in this regard from childhood through motherhood cannot be seen as "marginally" relevant. These complaints are highly relevant, both to her own credibility and also to that of the young victim, whose mother's prior and persistent claims of molestation could surely be expected to affect the victim's own attitude, behavior, and outlook. This evidence was readily available to trial counsel. We must agree with Goldstein that trial counsel's failure to investigate this evidence, present it, and use it in cross-examining Esther was a major deficiency. A reasonable probability exists that had trial counsel done so, it could have influenced the outcome of the trial.

(b) We also agree with Goldstein that his trial counsel provided ineffective assistance by failing to present expert medical and/or psychological testimony. The State presented the testimony of four such experts: Dr. Terese DeGrandi, a pediatrician who was also medical director of the Center for Advocacy and Protection at Children's Healthcare of Atlanta; Dr. Danielle Levy, a psychologist; Shirley Gee, a clinical coordinator at Children's Healthcare of Atlanta; and Dr. Nancy Aldridge, a forensic interviewer. Because trial counsel did not call a single expert witness to testify for the defense, the opinions expressed by the State's expert witnesses went completely unchallenged except by cross-examination.

As an example, the victim testified that the penetration of her vagina was deep and that it felt "kind of good." No witnesses testified that the victim showed any distress after her alleged molestation. The State sought to counter any impression that the victim would have been hurt or distressed had she been molested by introducing expert medical and psychological testimony. State's expert Dr. De-Grandi testified that she examined the victim ten days after the alleged molestation and that the examination was normal. She testified that normal findings are found in 95 percent of such cases because the hymen in a child the victim's age is very flexible and elastic. She demonstrated the elastic quality of a seven-year-old's hymen by comparing it to an elastic hair "scrunchie," testifying that the prepubescent hymen "has tremendous elastic capability." Because of that, according to the doctor, the lack of abnormal findings was nonetheless completely consistent with digital penetration. The State also introduced testimony showing that deep digital penetration of a seven-year-old's vagina, such as that Goldstein was alleged to have made, could be pleasurable to the victim, rather than painful. Dr. Danielle Levy, a psychologist, was called for the State, and she testified that it is "not that unusual" for a molested child to report that the molestation "feels good to their bodies. Those body parts are meant to feel good when stimulated."

At the hearing on the motion for new trial, new defense counsel presented two witnesses whose testimony completely contradicted that of the State's experts. Dr. Kim Collins, a forensic pathologist and Director of Forensic Pathology at the Medical University of South Carolina in Charleston, testified that a prepubescent hymen and vaginal lining are extremely delicate, sensitive, and easily traumatized. She explained that the hymen in a child is very thin, and contrary to the description given by Dr. DeGrandi, it is neither wrinkled or folded nor thick and elastic like a hair "scrunchie." It is so sensitive to touch that even a medical examination requires general or local anesthesia. Dr. Collins explained that not only would a seven-year-old feel pain at the time of penetration, the pain could continue afterward while bathing and urinating. Moreover, Dr. Collins opined that the victim's claim that the digital penetration "felt kind of good" was inconsistent with what a child her age would experience.

The deposition of Dr. Betty Spivack was also introduced at the hearing on the motion for new trial. Dr. Spivack is a forensic pediatrician employed in the Medical Examiner's Division of the Justice Cabinet for the State of Kentucky. She is also Co-Director of the Living Forensics Program, involving forensic evaluations of injured and abused children, and she specializes in sexual abuse evaluations. She lectures frequently nationally to physicians and law enforcement agencies on the techniques of examining and collecting evidence in child sexual assault cases. Although Dr. Spivack has testified approximately 50 times for the prosecution in such cases over the last 20 years, this was the only time she had ever been called to testify for the defense in a sexual abuse case. She received no compensation personally for her review or her opinion.

According to Dr. Spivack, a hair "scrunchie" illustrates well the typical appearance of an adult, well-estrogenized hymen, but a prepubescent, "unestrogenized" hymen is simply not elastic. It is sensitive even to light touch, and digital penetration such as that described by the victim in this case would cause extreme pain. She described the hymen of a child the victim's age as "a delicate, thin membrane which isn't very elastic and doesn't have folds and doesn't have an ability to stretch and return with no evidence of trauma." In her opinion, the child's statement that the penetration "felt kind of good" was inconsistent with the acts charged in the indictment. Furthermore, Dr. Spivack testified that the opinions of the State's experts in this case are inconsistent with any child she has ever examined or any report she has read in the medical literature.

Goldstein's wife Joanne testified at the hearing on the motion for new trial that she specifically asked trial counsel whether medical doctors would testify for the defense to rebut the prosecution experts, and trial counsel promised her that they would. In fact, none testified. When questioned at the hearing, trial counsel admitted that he had received notice from the State that Dr. DeGrandi and Dr. Levy would testify and information as to what their opinions would be. Nevertheless, he did not consult with medical doctors to determine the validity of those opinions. He testified expressly that no strategic reason existed for this failure; it simply never crossed his mind.

In assessing trial counsel's performance, we must be mindful that "counsel's function is to make the adversarial testing process work in the particular case." (Citation and punctuation omitted.) Jowers v. State, 260 Ga. 459, 462 (2) ( 396 SE2d 891) (1990). Trial counsel was aware of the substance of the testimony of the State's experts. Testimony from credible, objective scientists was available to trial counsel to completely refute the State's experts' opinions and substantially undermine the State's case. Indeed, Dr. Amy Morton, a psychologist specifically trained in the forensic interviewing of abused children, was actually sitting in the courtroom during the trial. She was the only potential defense witness permitted to remain when the courtroom was closed. For no strategic reason, the jury was left with the impression that the opinions of the State's expert witnesses were unassailable. We must agree that this omission fell below the standard of acceptable professional conduct.

As with trial counsel's failure to bring out the mother's history after he learned it from several members of Goldstein's family, trial counsel's failure to investigate the medical issues and seek and present expert testimony to rebut that presented by the State was a crucial omission. The State's evidence, other than opinions from their experts, was far from overwhelming. No evidence existed showing any physical trauma to the victim, no witnesses were present during the alleged molestation, the victim appeared to be happy and undisturbed after the alleged digital penetration, and the first outcry witness had a history of making accusations of molestation. A reasonable possibility — even a strong possibility — exists that had trial counsel's performance not been deficient, the outcome of the trial would have been different. Trial counsel simply "did not `make the adversarial testing process work.'" Jowers, supra at 462 (2). "Adjudged under the foregoing criteria, we hold that the omitted evidence in this case is of sufficient probative value to require a new trial." Id.

(c) Goldstein also points out that trial counsel failed to request a mistrial when a police officer testified that he believed the victim was telling the truth. This court has "repeatedly held that a witness, even an expert, can never bolster the credibility of another witness as to whether the witness is telling the truth. Credibility of a witness is not beyond the ken of the jurors[,] but, to the contrary, is a matter solely within the province of the jury. OCGA § 24-9-80." (Citations and punctuation omitted.) Mann v. State, 252 Ga. App. 70, 72 (1) ( 555 SE2d 527) (2001).

Here, a police officer testified that the victim "was very consistent. . . . I was very convinced that this had happened to her. I don't believe that she could have made it up." Trial counsel did object, his objection was sustained, and the jury was instructed "to disregard," without further explanation. Goldstein maintains that trial counsel should have sought a mistrial or admonition of the prosecutor. Trial counsel testified that no strategic reason existed for his failure to move for mistrial.

Of course, the officer's statement was improper and damaging. Roberson v. State, 214 Ga. App. 208, 210 (4) ( 447 SE2d 640) (1994). We agree with trial counsel's assessment of the statement as "outrageous." "The credibility of a witness, including a victim witness, is a matter for the jury's determination under proper instruction from the court. OCGA § 24-9-80. It is well established that in no circumstance may a witness's credibility be bolstered by the opinion of another, even an expert, as to whether the witness is telling the truth." (Citations and punctuation omitted.) Orr v. State, 262 Ga. App. 125, 128 (2) ( 584 SE2d 720) (2003). But here, trial counsel did object and the trial court gave a brief curative instruction. "A trial court has wide discretion in deciding whether to grant a mistrial." (Citation and footnote omitted.) Stokes v. State, 272 Ga. App. 721, 722 ( 613 SE2d 225) (2005). Under these circumstances, it is not clear that a mistrial would have or should have been granted, even had trial counsel moved for one.

4. In light of our holdings in Division 3, we need not reach Goldstein's remaining claim of ineffective assistance of counsel. Further, we have reviewed Goldstein's remaining enumerations of error and conclude that they are unlikely to recur at retrial.

Judgment reversed. Ruffin, C. J., and Phipps, J., concur.

SMITH, Presiding Judge.

DECIDED NOVEMBER 1, 2006 — RECONSIDERATION DENIED DECEMBER 15, 2006 — CERT. APPLIED FOR.