OPINION

delivered the opinion for a unanimous Court.

The trial judge excluded the testimony of Richard Vela, Jr.’s witness, Cheryl Hartzendorf, who would have testified that[*130] if there is no physical evidence, then no rape occurred. The court of appeals ruled that Hartzendorf should have been allowed to testify because her experience qualified her as an expert. ^[1] Because we hold that the court of appeals failed to conduct a proper analysis, we remand this case to the court of appeals.

Facts and Procedural History

At Richard Vela’s trial for three counts of sexual assault and one count of aggravated assault, the State called Sonia Ed-dleman, the sexual assault nurse examiner who examined the victim after the rape. Eddleman stated that the examination revealed no injuries to the victim’s mouth or genitals but that there was an “oozing tear” in the victim’s anus and bruising on her body. Based on her examination, Ed-dleman claimed the victim had been sexually assaulted.

Vela later called Cheryl Hartzendorf, a certified legal nurse consultant, and the State requested a hearing outside the jury’s presence. Before trial, Hartzendorf had reviewed the victim’s hospital records and the Department of Public Safety lab report containing the results of Eddle-man’s examination. Hartzendorf intended to testify that, in her opinion, because there was no DNA or physical evidence linking Vela to the alleged rape, no sexual assault occurred. She further stated that her opinion was based on her general nursing experience and that she had not written, nor was aware of, any published articles supporting that theory. The trial judge allowed the State to recall Eddle-man to rebut Hartzendorfs testimony. The State then challenged Hartzendorfs credentials and methodology, and the trial judge sustained the State’s objection. The jury later acquitted Vela on two of the sexual-assault counts but convicted him of the sexual assault alleged in count three and the aggravated assault charge.

The Thirteenth Court of Appeals determined that Hartzendorf was a qualified expert witness, ^[2] and after performing a harm analysis, the court found that her testimony could have had a significant effect on the outcome of the sexual-assault conviction. ^[3] Asserting that “the trial judge, in excluding Hartzendorfs testimony, acted without reference to any guiding rules and principles,” the court of appeals held that the trial judge abused his discretion in excluding Hartzendorfs testimony. ^[4] As a result, the court of appeals reversed Vela’s sexual-assault conviction but affirmed the aggravated assault conviction because Hartzendorf did not intend to testify about the alleged choking of the victim. ^[5]

We granted review to determine whether the Thirteenth Court of Appeals erred in holding that the trial judge abused his discretion in not permitting an expert witness’s testimony to the effect that where there is no physical evidence, there is no rape.

Expert Testimony

The Texas Rules of Evidence set out three separate conditions regarding admissibility of expert testimony. First, Rule 104(a) requires that “[preliminary questions concerning the qualification of a person to be a witness ... be determined by[*131] the court....” ^[6] Second, Rule 702 states: “If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education may testify thereto in the form of an opinion or otherwise.” ^[7] And third, Rules 401 and 402 render testimony admissible only if it “tend[s] to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence.” ^[8]

These rules require a trial judge to make three separate inquiries, which must all be met before admitting expert testimony: “(1) the witness qualifies as an expert by reason of his knowledge, skill, experience, training, or education; (2) the subject matter of the testimony is an appropriate one for expert testimony; and (3) admitting the expert testimony will actually assist the fact-finder in deciding the case.” ^[9] These conditions are commonly referred to as (1) qualification, (2) reliability, and (3) relevance. Only the first two conditions are at issue in this case, and we consider each in turn.

A. Qualifications of an Expert Witness

In Rodgers v. State, we noted that an appellate court should consider three criteria when determining whether a trial court abused its discretion in evaluating a witness’s qualifications as an expert: (1) “is the field of expertise complex?” ^[10] ; (2) “how conclusive is the expert’s opinion?” ^[11] ; and (3) “how central is the area of expertise to the resolution of the lawsuit?” ^[12]

Qualification is distinct from reliability and relevance and, therefore, should be evaluated independently. Although this Court has touched on the qualification analysis in prior cases, we have never discussed it in depth. We therefore look to Texas Supreme Court opinions for additional guidance. As that Court recognized in Broders v. Heise, the mere fact that a witness “ ‘possessed] knowledge and skill not possessed by people generally ... ’ does not in and of itself mean that such expertise will assist the trier of fact regarding the issue before the court.” ^[13] And because a witness will not always qualify as an expert merely by virtue of a general background, ^[14] qualification is a two-step inquiry. A witness must first have a sufficient background in a particular field, but a trial judge must then determine whether that background “goes to the very matter on which [the witness] is to give an opinion.” ^[15]

The Texas Supreme Court examined the second step of the qualification inquiry in Broders. The Court considered “whether[*132] the trial court abused its discretion in excluding the testimony of an emergency physician that the conduct of the three defendant emergency physicians and the defendant hospital was a cause in fact of a patient’s death.” ^[16] The proponents argued that “merely because [the witness was] a medical doctor,” the witness was qualified to testify about “all medical matters.” ^[17] Rejecting that argument, the Court explained that “there is no validity ... to the notion that every licensed medical doctor should be automatically qualified to testify as an expert on every medical question.” ^[18] If a medical degree carried automatic expert qualification in all medical matters, a trial judge could no longer fulfill his gatekeeping duty and “ensur[e] that those who purport to be experts truly have expertise concerning the actual subject about which they are offering an opinion.” ^[19]

The Court explained that the proponent must “establish that the expert has ‘knowledge, skill, experience, training, or education’ regarding the specific issue before the court which would qualify the expert to give an opinion on that particular subject.” ^[20] Given the facts in Broders, the Court determined that the proponents “simply did not establish that [the witness’s] opinions on cause in fact would have risen above mere speculation to offer genuine assistance to the jury.” ^[21]

In Christophersen v. Allied-Signal Corporation, the Fifth Circuit acknowledged the same problem with medical experts:

[Although credentials can be significant, they alone are not necessarily determinative. The questions, for example, do not stop if the expert has an M.D. degree. That alone is not enough to qualify him to give an opinion on every conceivable medical question. This is because the inquiry must be into actual qualification.... ^[22]

Similarly, in Gammill v. Jack Williams Chevrolet, Inc., the Texas Supreme Court again addressed the second part of the qualification inquiry and examined the qualifications of witnesses to testify as experts in an automobile-related products-liability suit. ^[23] Examining the background of one witness, the Court concluded that the witness “was shown to be experienced in designing and testing fighter planes and missiles, but he was not shown to have any training or experience in the design or manufacture of automobiles or their relevant components.” ^[24] Because “his only experience with automobiles at all was while working part-time as a mechanic doing general repairs,” the witness did not “have any expertise that would qualify him to testify about design defects in a vehicle’s accelerator or restraint system ... [or] as to the cause of [the victim’s] injuries or death.” ^[25] Notwithstanding his extensive background, the Court held that the witness was not qualified to render an opinion on the specific questions regarding automobiles that were at issue in that case. ^[26]

[*133] “The focus, then, is on the ‘fit’ between the subject matter at issue and the expert’s familiarity therewith, and not on a comparison of the expert’s title or specialty with that of the defendant or a competing expert.” ^[27] We discussed the “fit” requirement in Jordan v. State and explained that the issue under the reliability and relevance conditions “is whether the expert’s testimony took into account enough of the pertinent facts to be of assistance to the trier of fact on a fact in issue.” ^[28] But “fit” is not just a component of reliability and relevance — it is also a component of the qualification inquiry. Just as the subject matter of an expert’s testimony should be tailored to the facts of a case, the expert’s background must be tailored to the specific area of expertise in which the expert desires to testify.

The court of appeals failed to do an adequate inquiry into Hartzendorfs qualifications to testify about physical evidence of rape. Rather, the court simply referenced Hartzendorfs education and experience as a nurse and deemed that general background sufficient to qualify her as an expert witness. So the court made no meaningful inquiry into Hartzendorfs qualifications in that specific area of expertise.

B. Reliability of an Expert’s Testimony

The court of appeals claimed that “the reason for the trial judge’s decision ... [was] that the testimony was not reliable primarily because Hartzendorf had no scientific theory to support her opinion, and therefore, was not qualified.” ^[29] This assertion illustrates the court’s substitution of “reliability” for “qualification.” While qualification deals with the witness’s background and experience, reliability focuses on the subject matter of the witness’s testimony. The court of appeals muddled the qualification and reliability analyses and, therefore, failed to consider the reliability of Hartzendorfs testimony at all.

Texas Rule of Evidence 705(c) governs the reliability of expert testimony and states that “[i]f the court determines that the underlying facts or data do not provide a sufficient basis for the expert’s opinion under Rule 702 or 703, the opinion is inadmissible.” ^[30] “[Rjeliability depends upon whether the evidence has its basis in sound scientific methodology. This demands a certain technical showing.” ^[31] And that showing gives a trial judge the opportunity to “weed out testimony pertaining to so-called ‘junk science.’ ” ^[32] Thus, just because “ ‘junk science’ or otherwise inadequately tested scientific theories might be shown to relate to the facts of a case,” it will not always have a sufficiently reliable basis. ^[33]

Unlike qualification, the reliability condition has been thoroughly discussed in prior opinions. In Kelly v. State, we explained that scientific evidence must meet three criteria to be reliable: “(a) the underlying scientific theory must be valid; (b) the technique applying the theory must be valid; and (c) the technique must have been properly applied on the occasion in ques[*134] tion.” ^[34] We then outlined a list of nonexclusive factors that could affect a trial judge’s decision on reliability:

(1) the extent to which the underlying scientific theory and technique are accepted as valid by the relevant scientific community, if such a community can be ascertained; (2) the qualifications of the experts testifying; (3) the existence of literature supporting or rejecting the underlying scientific theory and technique; (4) the potential rate of error of the technique; (5) the availability of other experts to test and evaluate the technique; (6) the clarity with which the underlying scientific theory and technique can be explained to the court; and (7) the experience and skill of the person(s) who applied the technique on the occasion in question. ^[35]

And even if the traditional Kelly reliability factors do not perfectly apply to particular testimony, the proponent is not excused from proving its reliability. As the Texas Supreme Court recognized, “The court in discharging its duty as gatekeeper must determine how the reliability of particular testimony is to be assessed.” ^[36] The reliability inquiry is, thus, a flexible one. In some cases, the reliability of scientific knowledge will be at issue; in others, “the relevant reliability concerns may focus upon personal knowledge or experience.” ^[37] But the proponent must establish some foundation for the reliability of an expert’s opinion. ^[38] “Experience alone may provide a sufficient basis for an expert’s testimony in some cases, but it cannot do so in every case.” ^[39]

“Without more than credentials and a subjective opinion, an expert’s testimony that ‘it is so’ is not admissible.” ^[40] In Viterbo v. Dow Chemical Company, the Fifth Circuit excluded an expert’s testimony because it was “no more than Viterbo’s testimony dressed up and sanctified as the opinion of an expert.” ^[41] The district court had examined the data on which Viterbo’s expert’s opinion was based and found it to be lacking in reliability and probative value. ^[42] The court held that the expert “lacked objectivity in that he diagnosed Viterbo’s condition ... based only on the patient’s oral history and without the benefit of medical tests.” ^[43] The expert also “had no scientific literature to support his position and the tests which [the expert] performed did not establish a causal link between Viterbo’s symptoms and [the drug manufactured by the defendant].” ^[44]

The Fifth Circuit noted that the dispute was not over qualification, because the expert was properly qualified. ^[45] Rather, “[t]he dispute centered] on the source and basis of the expert opinion....” ^[46] The court acknowledged that the basis of an expert’s opinion may sometimes be “of such little weight that the jury should not[*135] be permitted to receive that opinion.” ^[47] But the court also recognized that even though trial judges may give experts “wide latitude” in selecting their sources, the judge must still evaluate those sources’ reliability. ^[48]

In this case, the court of appeals asserted that if “a particular type of scientific evidence is well established as reliable,” the proponent may avoid the burden of producing supportive evidence by asking that the trial court take judicial notice. ^[49] And the court claimed that “[cjlearly, medical testimony in sexual assault cases, whether by a nurse, sexual assault examiner, or physician, is the type of scientific evidence that is well established as reliable.” ^[50] The court of appeals, therefore, failed to make any inquiry into the reliability of Hartzendorfs theory. Instead, the court of appeals deemed Hartzendorf qualified to testify as an expert and then simply asserted that the trial court erred in excluding her testimony because medical testimony is “well established as reliable.” ^[51]

C. Analysis

The court of appeals performed an incorrect analysis of Hartzendorfs qualifications as an expert witness and failed to perform any analysis of the reliability of the subject matter of her testimony. The court of appeals gave three reasons why the trial court erred in sustaining the State’s objection to Hartzen-dorfs qualifications. ^[52] First, the court stated that “since practical experience alone is sufficient to qualify one as an expert, there is necessarily no requirement that the witness have written or read any articles subject to peer review before stating an opinion.” ^[53] While perhaps true in some circumstances, such a blanket assertion is inaccurate. To qualify a witness as an expert by practical experience alone, a trial judge must fully explore that witness’s experience in the particular field in which the witness intends to give an expert opinion. A cursory reference to the witness’s credentials is insufficient to support expert status. Moreover, the State’s argument at trial regarding Hartzendorfs lack of published articles referred to the reliability of her scientific theory, not to her qualifications as an expert witness. Therefore, the court of appeals misunderstood the State’s argument.

Next, the court of appeals stated that “we are not aware of any authority permitting the party against whom the expert testimony is offered to use its own expert to challenge the qualifications of the opponent’s expert.” ^[54] Even if this was an accurate statement, the State called Eddleman to rebut the reliability of Hartzendorfs testimony, not to challenge her qualifications.

Finally, the court of appeals stated that “to the extent the trial judge based his decision to exclude Hartzendorfs testimony on the basis that he personally did not agree with her opinion or did not find her credible was an abuse of discretion.” ^[55] But credibility and reliability are not the same. A jury should evaluate credibility,[*136] but unreliable evidence should never make it to the jury. As the court of appeals below acknowledged, “[t]he trial judge as gatekeeper is to determine the reliability, relevancy, and admissibility of scientific evidence.” ^[56] And acting as “gatekeeper,” the trial judge found Hartzendorf s scientific theory unreliable and, therefore, refused to admit her testimony.

We recognize that the “focus, of course, must be solely on principles and methodology, not on the conclusions that they generate.” ^[57] But we also agree that a trial judge need not “admit opinion evidence which is connected to existing data only by the ipse dixit of the expert. A court may conclude that there is simply too great an analytical gap between the data and the opinion proffered.” ^[58] Here, the court of appeals did not give proper deference to the trial judge’s ruling. “[Bjecause the possible spectrum of education, skill, and training is so wide, a trial court has great discretion in determining whether a witness possesses sufficient qualifications to assist the jury as an expert on a specific topic in a particular case.” ^[59] And when a trial judge determines that a witness is or is not qualified to testify as an expert, “appellate courts rarely disturb the trial court’s determination.” ^[60]

Conclusion

Because the court of appeals improperly evaluated Hartzendorfs qualifications, failed to evaluate the reliability of her proposed testimony, and failed to give proper deference to the trial judge, we vacate the judgment of the court of appeals and remand this case to the court of appeals for proceedings consistent with this opinion.

filed a concurring opinion, in which JOHNSON, J., joined.

I join the majority opinion. I add these comments only to emphasize that when an expert offers an opinion which is so outside the general mainstream of a particular scientific field as to be extraordinary, the proponent of that expertise must provide a greater-than-usual foundation for its reliability.

In this case, the complainant, J.S., testified that she started dating appellant while she was estranged from her husband. She moved into his garage apartment. Appellant owed people a lot of money, and he was frequently angry. On August 25th, appellant was drinking heavily and got angry at J.S. He told her that he needed to teach her a lesson “in respect.” He said that she was a whore, and if she wanted to be treated like a whore, he would do so. According to J.S., that is when the sexual assault started.

Appellant grabbed her by the hair, pushed her down on the couch, and tried to force her to perform oral sex on him. He slapped her, turned her over, pulled down her pants and began to anally rape her. J.S. said she

was crying and I told him to stop and I said, “You’re hurting me. Why are you doing this?” And he wouldn’t stop, and[*137] then he — he raped me vaginally, and then he stopped and he said — he said, “I’m so mad,” he said, “I can’t even finish.” He said, “I’m so mad at you.” And then he took his — he took his belt off from his pants and he started beating me with it across my legs and I tried to turn around and I was, you know, blocking myself and then he hit me in the butt and the legs.

J.S. repeatedly testified that she did not then agree, and never before had agreed, to have anal sex with appellant. She said that she did not agree to have vaginal or oral sex with him on this occasion, although she had agreed to do so on prior occasions, under different circumstances.

When J.S. went into the bathroom after the incident, she was bleeding profusely from the anus. Appellant came into the bathroom and told her that he had done this to teach her a lesson about respect. He said, “And if you act up anymore, next time I’ll tear you from one end to the other.” She was in shock. She went into the bedroom and cried until she fell asleep.

The next morning, appellant woke her up and said, “Bitch, get up and go make me some breakfast.” J.S. was scared, but obeyed, all the while planning to “buy some time” and then leave. Two days later, appellant got mad at J.S. again — “he went crazy” — and started to choke her. She tried to call 911, but appellant pulled the phone out of the wall. As appellant stepped outside to block J.S. from running away, she closed the door on him and locked him out. He then came around to the back of the apartment to climb through a window, so she ran out the front door and escaped, banging on people’s doors for help.

Some neighbors allowed her to use the phone, and she called the police. When the police arrived, they took her back to the apartment where appellant was already in custody. She told the police about the physical assault, but not about the sexual assault two days earlier because she was ashamed and embarrassed. She asked the officers to drive her to her estranged husband’s home. Once there, she told her husband about appellant’s sexual assault two days earlier and he took her to the hospital the next morning.

Although it was three days after the sexual assault, a S.A.N.E. nurse, Sonia Eddleman, examined J.S., noting an “oozing” anal tear, purple and yellow bruises on her left breast, left hip, left thigh, and left buttocks, as well as scrapes on her right buttocks. Nurse Eddleman said that J.S. did not have any other genital injuries.

After the State rested, appellant called his expert witness, Nurse Cheryl Hartzen-dorf, outside the presence of the jury, on a Daubert ^[1] hearing. Nurse Hartzendorf testified that she had examined J.S.’s medical records and found that the “rape kit” materials, gathered more than two days after the incident, did not uncover any sperm or fingernail scrapings that linked appellant to J.S. From these physical findings, she concluded that no sexual assault had occurred. She said that she was testifying pro bono because she saw “a right that needs to be wronged”: ^[2]

There is no evidence to indicate that this particular person, no sperm, no head hair combings, no vaginal swabs, fingernail swabs, oral smears, vaginal smears,[*138] anal smears that — there’s no DNA evidence linking this particular — Mr. Vela to the alleged rape of [J.S.].... I think it is consensual sex.

There may well be a scientifically reliable basis for this opinion, but if so, the basis for that opinion is nowhere in this record. Of course, Nurse Hartzendorf was certainly qualified to testify to the absence of vaginal trauma, and to the lack of other physical evidence linking appellant to the complainant. ^[3] However, there is nothing at all in the record that supports any scientific connection between the absence of vaginal trauma and the conclusion that therefore any sexual encounter was consensual. Nor is there any explanation for concluding that this encounter was “consensual sex” from the fact that there was no DNA evidence linking appellant to the vaginal swabs taken from J.S. DNA evidence usually indicates ejaculation of sperm or the presence of some human portion of anatomy. The record contains no scientific support for an opinion that a lack of DNA evidence indicates consensual sex.

Here, Nurse Hartzendorf admitted that she had not consulted any scientific articles, nor was she aware of any that supported her hypothesis. She agreed that hers was not a widely accepted belief with persons in her profession. The defense reiterated that she was a nurse and had testified “in cases like this” once before. Therefore, she was qualified to express this particular opinion in this particular case without any further foundational showing of scientific reliability. When questioned by the trial court, Nurse Hart-zendorf stated:

According to the medical records, there’s no trauma to this particular client that she’s claiming to of the neck bruising and — or fingerprints. According to the medical records of Christus Spohn Memorial or Doctor’s Regional or Ms. Eddleman’s report, it’s not indicated that there is any marks around the— [J.S.’s] neck. There is no evidence that was collected in terms of hair samples, sperm, wet mounts, vaginal swabs or anything that indicate there was any evidence linking this particular client to the rape, alleged rape.

After thanking the witness, listening to the arguments of counsel, hearing Ms. Eddle-man testify that she had never heard of this theory or of any scientific articles or other experts that supported the defense expert’s position, the trial court sustained the State’s objection to Nurse Hartzen-dorf s testimony.

Appellate courts must examine the trial court’s Daubert ruling within the context of the specific factual situation, the specific “fit” between those facts and the expert’s opinion, the expert’s qualifications concerning the specific area of proffered testimony, and the scientific basis for any proposed expert opinion. Although the general theory of “no rape if no physical trauma or physical findings” may have a scientific basis, it is such a scientifically unusual theory that a trial court would not abuse its discretion in excluding testimony about such a theory unless the expert provided ample scientific documentation to support its reliability and applicability in a situation such as the present one.

Furthermore, the defense made no showing that Nurse Hartzendorf was aware of controlling Texas law concerning sexual assault. ^[4] Unless an expert is using[*139] the same legal standards as the jury must use, he cannot give his opinion on a legal term because his opinion would be of no assistance to the jury. ^[5] First, of course, Texas law does not require ejaculation, only penetration, for a sexual assault. ^[6] Second, Texas law does not require the use of physical force; sexual assault occurs when a person penetrates the sexual organ of another without that person’s consent. ^[7] Further, even a layman might conclude that recovering physical evidence from a sexual assault that had occurred more than two days earlier might be a difficult proposition if the complainant had performed normal hygiene.

With these comments, I join the majority.