EDWARD LEON DANIELS, APPELLANT,

v.

STATE OF FLORIDA, APPELLEE

WALLACE, Judge.

Edward Leon Daniels appeals his judgment and life sentence for committing a sexual battery on a child less than twelve years of age, a violation of section 794.011(2), Florida Statutes (2005 and 2006). The trial court abused its discretion in permitting a medical witness to give expert opinion testimony that lacked a proper factual basis. Because we are unable to conclude that this error was harmless, we reverse Mr. Daniels’ judgment and sentence and remand for a new trial.

I. THE FACTS

The State charged Mr. Daniels with sexually battering K.D., a six-year-old child, during a three-month period beginning in May 2006 and ending in August 2006. K.D. testified that the sexual batteries had occurred at a mobile home in Hillsborough County where Mr. Daniels was residing with K.D. and her mother. After Mr. Daniels and K.D.’s mother separated, K.D. reported the alleged sexual abuse to her mother. K.D.’s mother promptly notified the authorities.

K.D. was taken to the Child Protection Team clinic in Tampa where she was examined by Sandra Shulman, an advanced registered nurse practitioner. Ms. Shul-man has extensive experience in conducting physical examinations in cases involving allegations of sexual abuse against children. During her physical examination of K.D., Ms. Shulman noted the child’s atypical reaction to a particular procedure used in testing for the presence of sexually transmitted diseases. Ms. Shulman was concerned about K.D.’s reaction to the procedure and concluded that it was “evidence that tends to suggest that child sexual abuse occurred.”

II. THE PRETRIAL AND TRIAL PROCEEDINGS

Mr. Daniels filed a pretrial motion in limine seeking to exclude “[a]ny testimony from [Ms.] Shulman that there is evidence of penetration based on” K.D.’s reaction to the procedure. At a pretrial hearing on the motion, Mr. Daniels’ counsel argued that Ms. Shulman’s opinion was not based on any scientific evidence. The trial court denied the motion and suggested that defense counsel could cross-examine Ms. Shulman concerning the basis of her opinion at trial and address the issue further during closing argument.

*747The case went to trial before a different judge, and the parties addressed the issue of the admissibility of Ms. Shulman’s opinion again during the trial. Defense counsel conducted a voir dire examination of Ms. Shulman directed to the basis for her opinion. During this examination, Ms. Shulman explained that she had never seen a child react to the procedure in a manner similar to the way in which K.D. had reacted. Defense counsel questioned Ms. Shulman about the factual basis for her opinion that KD.’s atypical reaction to the procedure was “evidence that tends to suggest that child sexual abuse occurred” as follows:

[DEFENSE COUNSEL]: Is your opinion about the [procedure] based on any kind of medical documents or medical— [?]

[MS. SHULMAN]: It’s based on my experience.

[DEFENSE COUNSEL] Okay. And the only thing you are able to testify about that is that [K.D.] was compliant and most children aren’t or have a difficult time?

[MS. SHULMAN]: I would say 99.9 percent of children have some reaction to that [procedure] at the age of six.

[DEFENSE COUNSEL]: But you weren’t able to make any specific finding based on that?

[MS. SHULMAN]: Correct.

At the conclusion of the voir dire examination, defense counsel argued that it would be permissible for Ms. Shulman to testify about her observations but not her conclusion because her conclusion was not based on facts. The trial court disagreed and ruled that Ms. Shulman would be allowed to express her opinion.

In her trial testimony, Ms. Shulman drew a distinction between findings that were “diagnostic” or conclusive evidence of child sexual abuse and other findings that were only “suggestive.” Ms. Shulman cited pregnancy and the presence of semen as examples of “diagnostic” findings because they constituted conclusive evidence that child sexual abuse had occurred. Ms. Shulman characterized other findings that were associated with child sexual abuse but were not conclusive indicators as merely “suggestive.”

On direct examination, Ms. Shulman testified that she had never observed a six-year-old child react to the procedure in the same manner that K.D. had reacted. Then Ms. Shulman opined that KD.’s atypical reaction was “suggestive” of child sexual abuse. Ms. Shulman explained that KD.’s reaction was not “diagnostic” or conclusive evidence that sexual abuse had occurred, but it tended to suggest that it had. On cross-examination, Ms. Shulman conceded that she had never examined K.D. before and could not determine if K.D.’s reaction was unusual for her. Ms. Shulman also admitted that she had found no conclusive evidence of penetration or child sexual abuse. In other words, Ms. Shulman’s examination of K.D. was ultimately inconclusive for evidence of sexual abuse.

During closing argument, the prosecutor told the jury that “the two main things you need to focus on in your deliberations is [sic] the testimony of the victim and the testimony of [Ms. Shulman].” The prosecutor emphasized that Ms. Shulman’s opinion testimony corroborated KD.’s testimony about the alleged sexual abuse.

The jury found Mr. Daniels guilty of capital sexual battery as charged, and the trial court sentenced him to life in prison. This appeal followed.

III. THE PARTIES’ ARGUMENTS

According to Mr. Daniels, the trial court improperly admitted Ms. Shulman’s opin*748ion testimony that K.D.’s reaction to the procedure was suggestive of sexual abuse. Mr. Daniels argues that this testimony was based on Ms. Shulman’s “speculation that [K.D.] would not have acted as she did unless she had been penetrated.” Mr. Daniels also contends that the admission of Ms. Shulman’s opinion testimony was not harmless because it corroborated KD.’s credibility, which was the main issue at trial. In response, the State argues that Ms. Shulman’s testimony was admissible as opinion testimony because it was based on her experience as a child abuse investigator.

IV. DISCUSSION

A. The Admissibility of the Expert’s Opinion

An expert’s opinion is admissible if it is “based on valid underlying data which has a proper factual basis.” Carnival Corp. v. Stowers, 834 So.2d 386, 387 (Fla. 3d DCA 2003). However, if an “expert’s opinion is based on speculation and conjecture, not supported by the facts, or not arrived at by [a] recognized methodology,” it should not be admitted into evidence. See M.A. Hajianpour, M.D., P.A. v. Khosrow Maleki, P.A., 932 So.2d 459, 464 (Fla. 4th DCA 2006). “ ‘[T]he basis for a conclusion cannot be deduced or inferred from the conclusion itself. The opinion of the expert cannot constitute proof of the existence of the facts necessary to the support of the opinion.’ ” Schindler Elevator Corp. v. Carvalho, 895 So.2d 1103, 1106 (Fla. 4th DCA 2005) (quoting Arkin Coustr. Co. v. Simpkins, 99 So.2d 557, 561 (Fla.1957)).

The crux of Mr. Daniels’ argument is that Ms. Shulman’s opinion was inadmissible in evidence because it lacked a proper factual basis. Ms. Shulman explained that (1) children typically reacted in one way during the procedure in question and (2) K.D. did not exhibit a similar reaction. From these facts, Ms. Shulman inferred that KD.’s behavior “suggested” that she had been sexually abused. But as Mr. Daniels observes, Ms. Shulman’s logic is flawed because she did not testify that in her experience, children who did not exhibit a typical reaction during the procedure were likely to have been sexually abused. In fact, Ms. Shulman could not establish the critical missing link because she had never seen a child who reacted to the procedure in the same way that K.D. had reacted. Therefore, Ms. Shulmaris unstated assumption that children who did not exhibit a typical reaction to the procedure are likely to have been sexually abused constituted the only foundation for the existence of the facts necessary to the support of her stated opinion.

Here, Ms. Shulmaris opinion that was based on facts not in evidence permitted her to express “ ‘unstated and perhaps unwarranted factual assumptions concerning the event.’ ” Young-Chin v. City of Homestead, 597 So.2d 879, 882 (Fla. 3d DCA 1992) (quoting Nat Harrison Assocs., Inc. v. Byrd, 256 So.2d 50, 53 (Fla. 4th DCA 1971)). Because Ms. Shulmaris opinion lacked a proper factual basis, the trial court abused its discretion when it admitted her expert opinion that K.D.’s reaction to the procedure was suggestive of sexual abuse. See Stano v. State, 473 So.2d 1282, 1287 (Fla.1985) (holding that “[a]n expert’s opinion must be based on facts in evidence or within his [or her] knowledge” and that the admission of an expert’s opinion is reviewed for an abuse of discretion); Schindler Elevator Corp., 895 So.2d at 1106-08; Carnival Corp., 834 So.2d at 387.

We have carefully reviewed the cases cited by the State in support of an affir-mance. These cases do not address Mr. Daniels’ argument and are otherwise inap-posite. Tingle v. State, 536 So.2d 202, 205 *749(Fla.1988), and Quintero v. State, 889 So.2d 1013, 1013-14 (Fla. 1st DCA 2004), describe the types of opinions that an expert may offer in a child sexual abuse case. However, these cases do not address the nature of the factual basis necessary to support the expert’s opinion. Russ v. State, 934 So.2d 527, 531 (Fla. 3d DCA 2006), does not address the factual basis for the expert’s opinion offered in response to the defendant’s claim in that case that the child had fabricated her story. The defendant in Oliver v. State, 977 So.2d 673, 677 (Fla. 5th DCA 2008), review denied, No. SC08-674, 2008 WL 4489844 (Fla. Oct. 6, 2008) (table decision), unsuccessfully argued that the expert improperly vouched for the victims’ credibility when the expert “offered observations from his experience regarding behaviors of child sex abuse victims.” But the expert in Oliver “did not directly testify about the victims,” id., and the challenged opinion testimony was simi- • lar to the testimony deemed admissible in Ritss.

B. Harmless Error Analysis

The improper admission of expert opinion testimony is subject to a harmless error analysis. See Nardone v. State, 798 So.2d 870, 874 (Fla. 4th DCA 2001); Shaw v. State, 557 So.2d 77, 78 (Fla. 1st DCA 1990). Here, the State has not argued that the error was harmless. Nevertheless, we will apply the harmless error test to the improper admission of Ms. Shulman’s expert opinion testimony. See §§ 59.041, 924.33, Fla. Stat. (2005 and 2006); see also Heuss v. State, 687 So.2d 823, 824 (Fla.1996); Thomas v. State, 838 So.2d 1192, 1196 n. 2 (Fla. 2d DCA 2003). The application of the harmless error test requires a review of the entire record, “ ‘including a close examination of the permissible evidence on which the jury could have legitimately relied, and in addition an even closer examination of the impermissible evidence which might have possibly influenced the jury verdict.’ ” Brooks v. State, 918 So.2d 181, 194 (Fla.2005) (quoting State v. DiGuilio, 491 So.2d 1129, 1135 (Fla.1986)). “ ‘If the appellate court cannot say beyond a reasonable doubt that the error did not affect the verdict, then the error is by definition harmful.’ ” Blanton v. State, 978 So.2d 149, 156 (Fla.2008) (quoting DiGuilio, 491 So.2d at 1139).

The only evidence of the crime presented by the State was KD.’s testimony and KD.’s statement to Ms. Shulman. The State did not present evidence of any admissions by Mr. Daniels, and there was no physical evidence linking him to the crime. During closing argument, the prosecutor correctly told the jury that the case hinged on whether or not they believed K.D. Mr. Daniels’ defense was that KD.’s mother had fabricated the allegations against him with the child after the relationship between Mr. Daniels and the mother had ended. Therefore, Ms. Shulman’s testimony was crucial to the State’s case.

In this context, testimony from Ms. Shulman was properly offered to help the jury assess KD.’s truthfulness “ ‘by summarizing the medical evidence and expressing [her] opinion as to whether it was consistent with [K.D.’s] story that she was sexually abused.’ ” Tingle, 536 So.2d at 205 (quoting United States v. Azure, 801 F.2d 336, 340 (8th Cir.1986)). Although Ms. Shulman conceded that the medical evidence was inconclusive, she offered impermissible opinion testimony that KD.’s reaction to the procedure was suggestive of sexual abuse. Ms. Shulman is highly qualified as an expert in the area of child sexual abuse. Her opinion testimony tended to corroborate K.D.’s story. During closing argument, the prosecutor emphasized Ms. Shulman’s extensive experience in the field and her opinion testimony that the child’s reaction to the procedure *750was suggestive of child sexual abuse. Thus it is probable that the jury viewed KD.’s atypical reaction to the procedure as tending to prove that K.D. had in fact been sexually abused. Under these circumstances, we are unable to conclude beyond a reasonable doubt that Ms. Shul-man’s expert opinion testimony that K.D.’s reaction was suggestive of sexual abuse did not affect the verdict.

V. CONCLUSION

The trial court abused its discretion in admitting Ms. Shulman’s expert opinion testimony into evidence. Because it does not appear that the improper admission of this testimony was harmless beyond a reasonable doubt, we reverse Mr. Daniels’ judgment and sentence and remand for a new trial.

Reversed and remanded for a new trial.

CASANUEVA and KHOUZAM, JJ., Concur.