Opinion
KING, J.This case presents the same issue as in In re Amber B., post, page 682 [236 Cal. Rptr. 623] whether the trial court erred in admitting expert opinion testimony that appellant had committed child sexual abuse. We conclude the court erred, but the error was harmless.
John C. appeals from an order declaring his twin children Christine and Michael to be dependent children of the juvenile court (Welf. & Inst. Code, § 300). We affirm.
Two separate petitions, filed by the Marin County Department of Social Services and later amended to conform to proof, alleged that John had sexually molested Christine and Michael when they were three years old. At the hearing on the petitions the department presented testimony by a psychiatrist, Dr. David Corwin, who had examined the children on one occasion. Over John’s objection the court permitted Dr. Corwin to testify he believed the twins’ claims that John had molested them.
Dr. Corwin’s conclusion was based primarily on “what the children told me” and “the way in which they told me.” Specifically, he relied on the consistency of their reports of abuse and the manner in which they reported the abuse, their demonstrations of emotion during the interview, the absence of any deprecation of their father, their demonstrated ability to discriminate between events that did or did not occur, their behavior with anatomically correct dolls, and the general “air of realism” in their assertions.
The department also presented opinion testimony by a therapist and two social workers that the twins’ claims of abuse by John were true. The social workers’ opinions were based on the manner in which the twins reported the abuse; the therapist’s opinion was based on her assertion that “children don’t lie.”
*679The children, who were five and one-half years old at the time of the dependency hearing, also testified. They described numerous acts of sexual abuse by John.
The court ruled that the twins had been sexually abused. The court’s dispositional order included findings that (1) John had stuck his finger in Christine’s vagina and anus, had put his penis on or near her vaginal and anal openings, and had made Christine play with his penis, and (2) John had put his penis on or near Michael’s anus, had played with Michael’s penis, and had had Michael play with John’s penis. The court declared the children to be dependent children of the juvenile court, placed them in the custody of their mother under the department’s supervision, afforded John supervised visitation, and required counseling for all parties.
I
John challenges the admission of the expert opinion testimony that the children’s claims of abuse were truthful. He asserts two reasons why the testimony should have been excluded, First, he argues the testimony should have been subjected to the Kelly-Frye test for admission of evidence based on a new scientific method of proof, requiring a showing that the procedure has been generally accepted as reliable in the scientific community in which it was developed. (People v. Kelly (1976) 17 Cal.3d 24 [130 Cal.Rptr. 144, 549 P.2d 1240]; Frye v. United States (D.C. Cir. 1923) 293 Fed. 1013 [34 A.L.R. 145].)1 Second, he argues the admission of the testimony identifying him as the perpetrator contravened In re Cheryl H. (1984) 153 Cal.App.3d 1098, 1118-1120 [200 Cal.Rptr. 789], which prohibits such testimony because it draws inferences about conduct by a third party based primarily on hearsay.
Our opinion in In re Amber B., supra, post, page 682, concludes that the practice of detecting child sexual abuse by (1) observing a child’s behavior with anatomically correct dolls, and (2) analyzing the child’s reports of abuse, is a new scientific method of proof which is subject to the requirements of Kelly-Frye. In the present case the experts relied on these same two factors to detect sexual abuse (although there does not appear to have been as much reliance on play with dolls as in Amber B.).2 For the reasons we state *680in Amber B., the trial court erred when it admitted the expert opinion testimony without first requiring a showing of general acceptance in the relevant scientific community in accordance with Kelly-Frye.
There was also error under Cheryl H. to the extent the experts identified John as the person who abused the children. As in Cheryl Ft., this testimony impermissibly drew inferences about conduct by a third party based primarily on hearsay. (In re Cheryl H., supra, 153 Cal.App.3d at pp. 1118-1120.)
Unlike in Amber B., however, the error was harmless. This case is different because the children testified and the trial judge provided a detailed oral explanation of her ruling.
Three elements of the court’s explanation demonstrate the error was harmless. First, the court said flatly that it found the children to be “believable.”
Second, the court accepted the basic premise of John’s defense, yet still found abuse. John asserts the children’s mother, Connie, along with the therapist and two social workers, were the sources of the abuse claims. According to John, Connie’s “bizarre and unusual preoccupation and attitude toward sexual matters,” combined with “leading and suggestive examinations” by the therapist and two social workers, caused the children to believe they had been abused when in fact they had not. The court said it agreed Connie was “obsessed ... with her interest in sex,” and theorized Connie had been “tantalated [szc] by what happened.” The court also agreed that some of the children’s testimony incorporated “things that adults had said to them.” In spite of this, however, the court still found there had been sexual abuse. The court’s comments indicate its acceptance of the basic premise of John’s defense would not have benefitted him had the expert opinion testimony been excluded.
Third, the court did not accept all of Dr. Corwin’s testimony. Dr. Corwin testified that the “most reliable” and “believable” of the children’s claims were Michael’s description of anal intercourse and Christine’s claims of vaginal and digital intercourse and fellatio. The court ruled, however, that there was insufficient evidence to conclude John had performed anal or vaginal intercourse with either child. This indicates the court relied on the generally “believable” testimony of the children, rather than the improperly admitted expert opinion testimony, to determine there had been sexual abuse.
In view of the court’s characterization of the children as believable, its acceptance of the basic premise of John’s defense without ruling in his favor, *681and its rejection of Dr. Corwin’s most forcefully asserted conclusions, it is not reasonably probable a result more favorable to John would have been reached in the absence of the erroneous admission of the expert opinion testimony. (People v. Watson (1956) 46 Cal.2d 818, 836 [299 P.2d 243].) The error does not require reversal.
II-IV*
The order is affirmed.
Low, P. J. and Haning, J. concurring.
A petition for a rehearing was denied May 21, 1987, and appellant’s petition for review by the Supreme Court was denied July 30, 1987.
Although John did not specifically assert the Kelly-Frye issue below, his attempt to limit the scope of the opinion testimony was sufficient to preserve the issue for appellate review. (People v. Bledsoe (1984) 36 Cal.3d 236, 245-246 [203 Cal.Rptr. 450, 681 P.2d 291].)
The expert witnesses occasionally described behavior by Christine and Michael which appears similar to some of the elements of “child sexual abuse accommodation syndrome” (see Amber B., supra, post, p. 689, fn. 3), but none of the witnesses asserted this theory or described its elements.
See footnote, ante, page 676.