OPINION
Hervey, J.,delivered the opinion of the Court
in which Keller, P.J., Johnson, Keasler, Alcala, Richardson, Yeary, and Newell, JJ., joined.The issue in this ease is whether the court of appeals erred when it affirmed the judgment of the trial court holding that Appellant failed to establish the reliability of proffered expert testimony. The testimony at issue was about the weapon-focus effect, which postulates that the accuracy of an eyewitness identification can be detrimentally impacted when, during the commission of an offense, a weapon is used by the assailant and seen by the complainant. We will affirm the judgment of the court of appeals.
Facts and pRocedural histoRY
Because we have already discussed the facts of this case in a prior opinion, we recite only the facts necessary to resolve this appeal.1 Appellant was charged with aggravated robbery. The only dispute at trial was the identity of the assailant. To prove that the eyewitness misidentified Appellant, the defense called a forensic psychologist named Dr. Steven Rubenzer to testify about the weapon-focus effect and its possible impact in this case. The trial court, however, excluded that testimony as irrelevant.2 Appellant was subsequently convicted.
On appeal, the Ninth Court of Appeals agreed that the testimony was not relevant and affirmed the ruling of the trial court.3 See Blasdell v. State, No. 09-09-00286-CR, 2010 WL 3910586 (Tex. App.— Beaumont Oct. 6, 2010) (mem. op.) (not designated for publication) (Blasdell I). In Appellant’s out-of-time petition for discretionary review, he argued that the court of appeals erred because that testimony was, in fact, relevant. We agreed with Appellant, reversed the judgment of the court of appeals, and remanded the cause for further proceedings. See Blasdell II, 384 S.W.3d at 831. On remand, the court of appeals, once again, affirmed the judgment of the trial court, but this time on the basis that the scientific principles of the weapon-focus effect were not proven to be reliable by clear and convincing evidence. See Blasdell v. State, 420 S.W.3d 406 (Tex. App.— Beaumont 2014, pet. granted) (Blasdell III). Appellant filed a petition for discretionary review, which we granted on two grounds:
(1) Eyewitness misidentification is a hallmark of wrongful conviction, and
(2) Whether the court of appeals has ’ decided an important question of federal law in a way that conflicts with the applicable decisions of the Court of Criminal Appeals or the Supreme Court of the United States.
*62ARGUMENTS OF THE PARTIES
Appellant offers two arguments as to why the trial court abused its discretion (and why the court of appeals erred to hold otherwise).- 'He first asserts that the trial court erred when it required Appellant to prove that Rubenzer’s proposed weapon-focus-effect testimony was reliable but did not hold him to the same burden for- other aspects of Rubenzer’s testimony. Second, ■ he contends that the trial court erred in excluding the weapon-focus-effect testimony because it “forestalled” any discussion of the reliability of that testimony in favor of the qualifications of Rubenzer and the relevancy of his other proposed testimony.
' The State responds that Appellant bore the “burden — not the trial court[ ] — to make a sufficient offer of proof to ensure the record was fully developed for appellate purposes” and that he failed to meet that burden as to Rubenzer’s qualifications4 and the reliability of his proposed weapon-focus-effect testimony. It also asserts that Appellant “has essentially admitted that the current record establishes neither Rubenzer’s qualification to testify about the weapon-focus effect nor the reliability of his proffered testimony on the topic.” We agree with the State.
Discussion
A. The principles of expert testimony.
A trial judge’s decision to admit expert testimony is reviewed for an abuse of discretion and may not be reversed unless that ruling fell outside the zone of reasonable disagreement. See Blasdell II, 884 S.W.3d at 829; Weatherred v. State, 15 S.W.Bd 540, 542 (Tex. Crim. App. 2000).
Rule 702 of the Texas Rules of Evidence governs the admissibility of expert evidence:
A witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the form of an opinion or. otherwise if the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue.
Tex. R. Evid. 702. We have interpreted this rule to require a proponent to prove by clear and convincing evidence that the proffered’ testimony is reliable and helpful to the jury in deciding the ultimate issue. See Blasdell II, 384 S.W.3d at 829; see also Kelly v. State, 824 S.W.2d 568, 572 (Tex. Crim. App. 1992). However, because only the reliability of the proposed expert testimony is at issue in this case, we address only that requirement.
The purpose of the reliability inquiry is to “separate the wheat’ from the chaff.” That is, the trial court in its role as gatekeeper should exclude unreliable “junk” science while allowing other permissible types of reliable expert testimony. See Jordan v. State, 928 S.W.2d 550, 555 (Tex. Crim. App. 1996). We have identified two types of permissible scientific evidence under Rule 702, “hard” science evidence, such as physics or mathematics and “soft” science evidence, such as “social sci-*63enees or fields that are based primarily upon experience and training as opposed to the scientific method....” Nenno v. State, 970 S.W.2d 549, 561 (Tex. Crim. App. 1998). Psychology is a soft science, and eyewitness identification is an established subject within the scope of psychology. Tillman v. State, 354 S.W.3d 425, 436-37 (Tex. Crim. App. 2011). To establish the reliability of a soft science, the proponent must show that “the field of expertise involved is a legitimate one,” “the subject matter of the expert’s testimony is within the scope of that field,” and “the expert’s testimony properly relies upon or utilizes the principles involved in that field.”5 Nenno, 970 S.W.2d at 561. Although the proponent of the evidence has to establish its reliability, we have stated that the proponent “need not always present expert testimony, treatises, or other scientific material to satisfy the [reliability] test.”6 Hernandez v. State, 116 S.W.3d 26, 28-29 (Tex. Crim. App. 2003). Instead, to meet his burden, the proponent can establish that the “scientific principle [in question] is generally accepted in the professional community and has been accepted in a sufficient number of trial courts through adversarial Daubert/Kelly hearings.... ” Id. at 29. This alternative method of proving the reliability of a scientific principle is to avoid re-inventing the “scientific wheel” in every trial. Id.
Because two prior decisions of this Court inform our reliability discussion, we review them now. In Tillman, we decided that the appellant had proven'the relevancy and reliability of proposed expert ‘testimony on the subject of eyewitness identification. Tillman, 354 S.W.3d at 443. There, the witness intended to testify about various theories within the study of eyewitness identifications and how those theories could have affected the identifications in that case,7 but the trial court excluded his testimony. Id. at 442. We reversed, holding that the expert’s testimony was relevant and reliable. Id. at 443. In reaching that conclusion, we discussed the qualifications of the expert in the area of psychology and within the field of eyewitness identification, the existence of literature supporting the theories about which he proposed to testify, the fact that the results of testing the hypotheses could be repeated and error rates calculated, the acceptance of his theories within the relevant scientific community, and the clarity with which his scientific theories could be explained to the court. Id. at 437-38. Specifically, we noted that the expert was a psychologist with extensive experience in the field of eyewitness identifications and that the relevant scientific community had accepted his theories through his peer-reviewed literature. Id. We also thought it important that he referred to empirical experiments supporting his theories and from which error rates could be ascertained, and that he clearly articulated .to *64the court the principles governing the theories at issue and gave specific examples illustrating the application of those principles to hypothetical fact patterns. Id.
In Coble v. State, 330 S.W.3d 253 (Tex. Crim. App. 2010), an expert testified regarding the future dangerousness of the appellant, which the appellant challenged on reliability grounds. We agreed that the expert’s testimony was not reliable because it was not clear “what principles of forensic psychiatry [the. expert] might have relied upon because he cited no books, articles, journals, or even other forensic psychiatrists who practice[d]” in the area. Id. at 277. We also noted that his methodology was admittedly “idiosyncratic,” and was developed without reference to the “significant body of literature” on the topic, and that there was no objective evidence in the record to validate his methodology. Id. at 277-78. In addition, although there was an ascertainable scientific community, the expert did not produce any peer-reviewed literature or materials supporting his theory or show that the scientific community accepted his theory. Id. at 278-79.
B. Appellant failed to prove the reliability of the proposed weapon-focus-effect testimony.
We agree with the court of appeals that Appellant failed to establish the reliability of Rubenzer’s proposed weapon-focus-effect testimony because he did not meet the criteria under Nenno nor did he show that the trial court could take judicial notice8 of the general acceptance of the weapon-focus-effect theory or Rubenzer’s methodology.
Rubenzer has a doctorate in psychology and is a board-certified forensic psychologist. He began studying eyéwitness identifications in 2001, and he testified that the area of psychology dealing with eyewitness identifications is “a major topic.” He also stated that the body of knowledge hinges on peer-reviewed literature and the replication of test results. When asked about peer-reviewed articles in the area of eyewitness identifications, Rubenzer respond*65ed that there are “literally thousands” and that, in his fourteen-year experience, he had read 85 to 90 articles and seven books, as well as attended a symposium every year or two for the past five years. He has also appeared in court as an eyewitness-identification expert in one or two cases,9 but he conceded that he had not published any peer-reviewed articles or conducted any studies of his own about eyewitness identifications,10 much less the weapon-focus effect. Based on the foregoing, Rubenzer considers himself “basically self-educated.”
Although we suggested that the relevant scientific community may have generally accepted the weapon-focus effect as an established phenomenon as early as 1992, we disagree with the court of appeals that we directly decided the issue of general acceptance in Blasdell II}11 Rather, we expressly declined to resolve Blasdell II on the ground that the weapon-focus effect has been established as a scientific phenomenon, and the three studies we cited in that case regarding the weapon-focus effect reached conflicting results.12 Blasdell II, 884 S.W.3d at 828 n.3. Rubenzer testified that eyewitness identification is a field of study with thousands of peer-reviewed articles on the subject, but he never stated whether any of the symposia he attended or any of the literature he had read specifically addressed the weapon-focus-effect theory. As a result, we are unable to discern the potential error rate of the hypothesis or even whether literature in the scientific community supports or rejects *66the theory. Based on the limited information in this record, we cannot say that the weapon-focus-effect theory has been generally recognized as a valid hypothesis so as to be admissible here.13
In addition, although Rubenzer explained the weapon-focus effect — “a tendency, when there is a weapon involved, particularly in brief encounters, for the weapon to essentially attract, attention away from the perpetrator’s face and, by doing so, result in lesser accuracy for the identification” — and he stated that the theory “possibly” was applicable to the facts of this case, the problem is that he did not describe any principles of the theory or the methodology by which he would apply those principles. See Blasdell III, 420 S.W.3d at 411. For example, Appellant asked Rubenzer if it was significant to him that the complainant saw a revolver, and he responded that, “the witness apparently did look at the gun long enough to determine what type of gun it was.” However, Rubenzer never indicated what was significant about the victim being able to identify the model of the gun used during the .offense or any other potentially applicable principles such as whether proximity between the victim and the weapon is important or whether the type'of firearm (e.g., a large gun. versus a small gun) affects how the theory should be applied. In short, because Appellant has neither met the criteria for admissibility under Nenno in this case, nor directed us to a case in which the weapon-focus effect and Rubenzer’s methodology have been proven reliable such that we can take judicial notice of their general acceptance, we affirm the judgment of the court of appeals.
Meyers, J., filed a dissenting opinion.
. See Blasdell v. State, 384 S.W.3d 824, 826-29 (Tex. Crim. App. 2012) (Blasdell II).
. The trial court allowed Rubenzer to testify regarding potential problems with cross-racial identifications, deteriorating memory of eyewitnesses as time passes, police-administered admonishments for eyewitnesses, and the tendency of eyewitnesses to identify the wrong person when the "foils” provided by the police look substantially different from the described suspect.
.Because the court of appeals disposed of the case on relevancy grounds, it did not decide whether the weapon-focus effect is based on valid .science. See Blasdell I, 2010 WL 3910586 at ⅜3.
. We do not address the State’s claim that Rubenzer was not qualified as an expert because we did not grant review of that issue. Moreover, the State did not properly raise the qualification issue in a cross-petition for dis: cretionary review. See Keith v. State, 782 S.W.2d 861, 863 n. 4 (Tex. Crim. App. 1989) (holding that when "the State intends to claim an error by the Court of Appeals' rejection of an argument, that claim should be presented to this Court in a petition for discretionary review or a cross-petition for discretionary review. Since the State 'won' before the Court of Appeals in the instant case, a cross-petition would have been both appropriate and necessary’to preserve this issue for review by this Court.”).
. Because the first two elements of reliability are not at issue, we address only the third requirement that the expert’s testimony must properly rely upon or use the principles involved in the field of eyewitness identifications.
. While it is true that in Hernandez we discussed when trial courts can take judicial notice of the validity of a scientific principle or methodology in the context of Kelly and hard sciences, that logic is equally applicable to the soft sciences, as evidenced by our decision in Morris v. State, 361 S.W.3d 649, 668-69 (Tex. Crim. App. 2011) (talcing judicial notice that “grooming” is an established phenomenon subject to. expert testimony under Nenno).
; Specifically, he proposed to testify about cross-racial identification issues and to discuss the manner in which the lineup and photo spread were used. Tillman, 354 S.W.3d at 429.
. Although courts may take judicial notice of the reliability of a scientific theory or methodology applying that theory; we need not determine if the weapon-focus effect is such a phenomenon because Appellant does not argue that it is. Hernandez, 116 S.W.3d at 29; cf. Morris, 361 S.W.3d at 668-69. We do note though that in Morris, we recognized via judicial notice for the first time on appeal that "grooming” is a reliable experience-based field subject to expert testimony, and we upheld the ruling of the trial court under that alternative theory of law. However, these two cases are distinguishable. First, the evidence in Morris was experience based, unlike the evidence here that is scientific in nature. Additionally, the State in Morris argued that we should judicially notice the reliability of "grooming,” but Appellant has never made such an argument regarding the weapon-focus effect. Finally, the procedural posture of Monis and this case are quite different. In Morris, we employed a familiar rule of appellate law that the ruling of a trial court should be upheld under any applicable theory of law, but that rule is inapplicable here because we would' have to reverse the trial court’s ruling under an alternative theory of law, i.e., judicial notice. See State v. Mercado, 972 S.W.2d 75, 77 (Tex.Crim.App.1998) (per curiam) ("Although this Court has approved of appellate courts considering alternative theories of law applicable to the facts of the case which support the trial court’s decision, we have not afforded the courts of appeals latitude to reverse a trial court's decision on new theories of law not previously presented to that court for its consideration.”); see also Hernandez, 116 S.W.3d at 34 (Keller, P.J., concurring) (arguing that judicial notice on appeal should be able to serve as the sole source of support concerning scientific reliability if there is an adequate basis to do so "just as an appellate court would in any case in which a trial court's decision was correct on an applicable theory of law that was not in fact litigated before the trial court”).
. Rubenzer testified as an eyewitness-identification expert two times, according to reported appellate opinions. However, neither of these cases dealt with the weapon-focus effect. See Baldree v. State, 248 S.W.3d 224, 228-30 (Tex. App.— Houston [1st Dist] 2007, pet. ref d) (eyewitness-identification testimony dealing with photo arrays excluded because it was not relevant); Garcia v. State, No. 14-06-00570-CR, 2007 WL 2447301, at *4-5 (Tex. App.— Houston [14th Dist.] Aug. 30, 2007, pet. ref d) (mem. op.) (not designated for publication) (testimony about the reliability of eyewitness-identification testimony excluded because Rubenzer was not qualified).
. At one point Rubenzer testified that he had begun a project but had not collected any data. Addressing that comment, the trial judge stated, "And his comment was interesting, that ‘I’ve begun a project but haven’t collected data.’ Well, that’s sort of like saying 'I haven’t begun a project.’ ”
. The court of appeals stated that this Court in Blasdell III "noted that 'a substantial portion of the relevant community of forensic psychologists has regarded the weapon focus effect to be sufficiently well-established phenomenon to merit expert testimony.' ” Blasdell III, 420 S.W.3d at 411 (citing Blasdell II, 384 S.W.3d at 828 n. 3). However, while we did make that statement, we also "decline[d] the State’s invitation to resolve the issue of the scientific validity of the weapon focus effect....” Blasdell II, 384 S.W.3d at 828 n. 3. We do, however, agree with the court of appeals that Rubenzer did not discuss the practices of the relevant scientific community and that Appellant’s trial took place more than two years before this Court suggested that the scientific community may have recognized the weapon-focus effect as a valid psychological theory. See Blasdell III, 420 S.W.3d at 411 n. 1.
.In Tillman, this Court, quoting the Supreme Court of New Jersey, noted that a 2001 survey by sixty-four experts showed that 87% of respondents believed research on the weapon-focus effect was reliable. Tillman, 354 S.W.3d at 436-37 (quoting New Jersey v. Henderson, 208 N.J. 208, 27 A.3d 872 (2011)). The following year in Blasdell II, we suggested that the weapon-focus effect may have been accepted within the relevant scientific community as early as 1992. See Blasdell II, 384 S.W.3d at 828 n. 3. However, we have also cited a 2011 article questioning the validity of the hypothesis. See id. (citing George Vallas, A Survey of Federal and State Standard for the Admission of Expert Testimony on the Reliability of Eyewitnesses, 39 Am. J.Crim. L. 97 (Fall 2011)) (declining to dispose of the appeal on the grounds that the weapon-focus effect had not been generally accepted).
. Although Appellant bore the burden to establish that Rubenzer's proposed weapon-focus-effect testimony was reliable, he failed to proffer any meaningful evidence on this point. The only admitted documentary evidence was Rubenzer’s curriculum vitae and resume, and the only supporting testimony was that the weapon-focus effect "possibly” could have af- • fected the identification in this case. While it can be sufficient for an expert to demonstrate through hypotheticals how the scientific phenomenon in question may have affected the case at hand, Appellant's evidence was limited to stating the premise of the theory and explaining that it possibly may have affected the identification in this case. Moreover, we cannot take judicial notice of a scientific theory or hypothesis without evidence of a sufficient number of adversarial hearings at which the reliability of scientific principle was examined, assessed, and approved. Thus, Appellant cannot now “rely on the appellate courts to become independent scientific sleuths to ferret out the appropriate scientific materials ...” to prove the reliability of the weapon-focus-effect hypothesis for him. Hernandez, Í16S.W.3d at 30-31.