Dr. Edmund Chein was an expert medical witness in an automobile accident trial in California state court. He was also involved in a suit with a former business associate concerning the distribution of fees paid by patients. In both lawsuits he provided evidence — in the first instance trial testimony, in the second an interrogatory answer — that was misleading, at the least, concerning his medical credentials. At the instigation of the trial judge in the personal injury trial, he was charged in California state court with four counts of perjury and convicted of three. This ha-beas case raises various questions concerning the propriety of his conviction, of which we address only one.
Before plunging into the details of this perjury case, it is worth recalling “the traditional Anglo-American judgment that a prosecution for perjury is not the sole, or even the primary, safeguard against errant testimony.” Bronston v. United States, 409 U.S. 352, 360, 93 S.Ct. 595, 34 L.Ed.2d 568 (1973). Underlying this limited role of perjury prosecutions is the “one consideration of policy [that has] overshadowed all others during the years when perjury first emerged as a common-law offense: ‘that the measures taken against the offense must not be so severe as to discourage witnesses from appearing or testifying.’ ” Id. at 359, 93 S.Ct. 595 (citation omitted). Quoting a nineteenth century treatise to the effect that “the obligation of protecting witnesses from oppression, or annoyance, by charges, or threats of charges, of having borne false testimony, is far paramount to that of giving even perjury its deserts,” Bronston held that intent to mislead, or actually misleading, a jury or other factfin-der is not sufficient to make out the crime of perjury. Id. at 359, 93 S.Ct. 595. California law is the same. In re Rosoto, 10 Cal.3d 939, 949, 112 Cal.Rptr. 641, 647, 519 P.2d 1065, 1071 (1974) (holding that “failure to volunteer testimony to avoid the misleading impression does not constitute perjury,” and citing Bronston approvingly)-
These cautions apply with particular force to expert witnesses such as Chein. Although paid, usually well, for their efforts, such witnesses generally appear because they freely choose to do so, often with considerable immunity from subpoena. See generally Janet Fairchild, Annotation, Right of Independent Expert To Refuse To Testify as to Expert Opinion, 50 A.L.R.4th 680 (1986). Unless the strict requirements governing perjury convictions developed by the common law and applied by California are carefully applied, the willingness of experts to assist factfin-ders with the specialized knowledge needed to decide many cases, see Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), may atrophy.
As will appear, Chein undoubtedly did calculate the answers for which he was convicted in the hope that he would succeed in misleading the jury in the personal *981injury case and the 'Opposing lawyer in the monetary dispute case. But, on careful review of the record, we conclude that no reasonable jury could have concluded that all the elements of the crime of perjury were made out, and therefore reverse the denial of the habeas petition.
I.
Background
The allegedly perjurious statements that lie at the heart of this case were made during the course of two civil suits. The first was Lopez v. Consolidated Freightways, L.A. County Sup.Ct. Case No. WEC 105580 (1990), a personal injury case in which Chein testified as an expert witness for the plaintiffs. Chein was convicted for falsely testifying in Lopez that (1) “he was a specialist in orthopedic surgery” (the “specialist count”); and (2) “he had one office location on October 9, 1989 and on October 30, 1990 to November 2, 1990” (the “office count”).
The information charging Chein with perjury details the relevant portions of his Lopez testimony concerning the specialist count:
A. When one graduates from medical school, they cannot obtain a license to practice unless they have one year of internship in a hospital. And after one finishes the one year of internship, then they are entitled to get the California state license. When one gets that license, he’s called a general practitioner or family practitioner. He can practice medicine but with no specialty. If a doctor wants to specialize in a field, then they have to spend another four to five years, depending on the specialty, in a program called residency, in a medical residency or specialty training program. Then he goes out and practices as a specialist.
Q. Did you receive your license to practice medicine?
A. Yes.
Q. You received your specialty qualifications?
A. Yes.
Q. What field is that specialty, sir?
A. Board certified in December 1988 by the American Board of Orthopedic and Neurological Surgery.
Q. Very good, Sir. And would you indicate if you — whether or not you engage in a regular day-to-day practice of medicine?
A. Yes, I do.
Q. And in what field, sir?
A. In orthopedics and neurology.
Q. Then when the residency ends, am I correct, in layman’s terms, you’re a regular doctor; right?
A. No. You’re a specialist.
Q. You were a specialist following that—
A. Training period.
Q. —-Training period? And your specialty, what would be the correct designation?
A. Physical medicine and orthopedic surgery.
The state has concentrated on Chein’s statement that he was a specialist in orthopedic surgery at the time he finished his training as the perjurious one.
The second count of perjury relating to Chein’s Lopez testimony, the office count, dealt with the number of offices Chein “had.” He was charged with perjury for his responses to the following two questions:
Q. Doctor, how many office locations do you have now?
*982A. One.
Q. How many did you have on October the 9th 1989?
A. One.
Finally, Chein was convicted of one count of perjury (the “university count”) related to his answer to an interrogatory in Kancilia v. Chein, L.A. County Sup.Ct. Case No. LC012300 (1992), a case in which Chein’s former medical partner sued for an accounting of funds after the termination of their business agreement. The interrogatory read:
State:
a) the name and address of each school or other academic or vocational institution you have attended beginning with high school
b) the date you attended;
c) the highest grade level you have completed;
d) the degree received.
Chein’s response was: “American University School of Medicine, Florida, 1979-1980 M.D.”
Chein appealed his conviction to the California Court of Appeal, claiming (1) the expert witness testimony of Judge Robert Altman, the sitting state court judge who had presided over the Lopez case, amounted to a directed verdict and denied Chein a fair trial; (2) the prosecutor committed misconduct; (3) there was insufficient evidence to support his conviction; and (4) the jury instructions were improper. The Court of Appeal, in an unpublished opinion, affirmed the conviction. With regard to the sufficiency of the evidence issues we here address, the court held that the trial statements in Lopez were false, and were material because they “had a tendency to enhance the credibility of defendant,” while the interrogatory answer in Kancilia was false and “was material to the outcome of the case because it went to his credibility both as a party to the lawsuit and as a physician in business with Dr. Kancilia.” People v. Chein, No. B113514, at 18 (Cal. Ct.App. Nov. 25, 1998). Chein then filed a petition for review in the California Supreme Court, which was summarily denied.
Chein’s timely habeas petition in federal district court pursuant to 28 U.S.C. § 2254 raised substantially the same claims presented to the California Court of Appeal. The habeas petition was denied.
On review of the district court’s denial of habeas relief, we conclude that there was constitutionally insufficient evidence to support a guilty verdict on any of the perjury counts for which Chein was tried. We therefore grant Chein’s habeas petition. We need not and do not address Chein’s other challenges.
II.
Standard of Review
We review de novo the rejection of the sufficiency of the evidence challenge that Chein raises in his habeas petition. See Melendez v. Pliler, 288 F.3d 1120, 1124 (9th Cir.2002).
Chein attacks his state conviction on the ground that there was insufficient evidence to find him guilty of perjury beyond a reasonable doubt, a challenge that was first made available to state prisoners seeking federal habeas review in Jackson v. Virginia, 443 U.S. 307, 319, 99 S.Ct. 2781, 61 L.Ed.2d 560 (1979). Under Jackson, on habeas, “the relevant question is whether, after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.” Id. (emphasis in original). Put another way, the dispositive question under Jackson is “whether the record evidence could reasonably support *983a finding of guilt beyond a reasonable doubt.” Id. at 318, 99 S.Ct. 2781. On direct review, the California Court of Appeal considered and rejected Chein’s sufficiency of the evidence challenge, citing the Jackson standard.
Because Chein’s habeas petition was . filed after April 24,1996, the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA) governs our review. See Arredondo v. Ortiz, 365 F.3d 778, 781 (9th Cir.2004). Under AEDPA, where a state court decides the merits of a petitioner’s constitutional challenge, a federal court may grant habeas relief with respect to that claim if the state decision “was contrary to, or involved an unreasonable application of, clearly established federal law, as determined by the Supreme Court of the United States,” 28 U.S.C. § 2254(d)(1) (2000), or “resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding,” 28 U.S.C. § 2254(d)(2).
Two Ninth Circuit post-AEDPA cases have applied Jackson as written. They have not inquired whether the state court reasonably decided that a rational jury could have found the elements of the crime beyond a reasonable doubt. See Davis v. Woodford, 333 F.3d 982 (9th Cir.2003) (applying Jackson to determine whether there was sufficient evidence to find premeditation beyond a reasonable doubt, id. at 992-94, even though the California Supreme Court had previously applied Jackson as to premeditation, People v. Davis, 10 Cal.4th 463, 509, 41 Cal.Rptr.2d 826, 850, 896 P.2d 119, 143 (1995)); Turner v. Calderon, 281 F.3d 851, 881-884 (9th Cir.2002) (applying the Jackson standard in a case in which the state court did not expressly rely on Jackson). But see Mitchell, 107 F.3d at 1339 n. 3 (holding that AEDPA did not apply, but assuming that if it did, deferential review of the state court’s application of Jackson would have been appropriate).
The parties to this case and the original three judge panel assumed, without analysis, that Jackson applies as written. There has therefore been no briefing in this case, and no judicial analysis, concerning the proper intersection of Jackson and AEDPA, or whether it is § 2254(d)(1) or § 2254(d)(2) that applies to Jackson substantial evidence review under AEDPA. As will appear, we would reach the same result in this case were we to follow Davis and Turner or, instead, ask whether the California Court óf Appeal reasonably applied the Jackson standard. We therefore do not consider the question concerning the impact, if any, of AEDPA on the Jackson standard.
III.
Analysis
A.
General Principles
The Jackson standard “must be applied with explicit reference to the substantive elements of the criminal offense as .defined by state law.” Jackson, 443 U.S. at 324 n. 16, 99 S.Ct. 2781. Under California law, the elements of perjury are: “a willful statement, under oath, of any material matter which the witness knows to be false.”1 Cabe v. Sup.Ct., 63 Cal.App.4th *984732, 74 Cal.Rptr.2d 331, 333 (1998) (internal quotation marks omitted). Chein challenges the sufficiency of the evidence used to convict him on materiality and falsity grounds.
1. Materiality
The jury at Chein’s perjury trial was instructed that the legal standard for materiality is whether the statement in question “could probably have influenced the outcome of the proceedings.” This standard is the one generally used by California trial courts, see CALJIC § 7.21 (1996 Revision), and finds support in California case law, see People v. Wade, 39 Cal.App.4th 1487, 46 Cal.Rptr.2d 645, 650 (1995) (quoting People v. Pierce, 66 Cal.2d 53, 61, 56 Cal.Rptr. 817, 823, 423 P.2d 969, 975 (1967), overruled on other grounds by People v. Kobrin, 11 Cal.4th 416, 45 Cal.Rptr.2d 895, 903 P.2d 1027 (1995) (holding that materiality is an element of the crime to be determined by the jury beyond a reasonable doubt, and not a question of law for the court)).
Importantly, when applying the materiality test, California law focuses not on whether, as a matter of historical fact, the false statement probably did influence the outcome of the proceedings, but instead on whether the false statement, at the time it was made, had the tendency to probably influence the outcome of the proceedings. See, e.g., People v. Poe, 265 Cal.App.2d 385, 71 Cal.Rptr. 161, 164-65 (1968). In other words, under California law, materiality is evaluated from an ex ante, not an ex post perspective. This fundamental rule of perjury law is codified in California Penal Code § 123:
Witnesses’ knowledge op mateRiality of HIS TESTIMONY NOT NECESSARY. It ÍS nO defense to a prosecution for perjury that the accused did not know the materiality of the false statement made by him; or that it did not, in fact, affect the proceeding in or for which it was made. It is sufficient that it was material, and might have been used to affect such proceeding.
Id. (emphasis added).
As the dissent notes, Kobrin, an opinion directed only at the question whether the judge or jury determines materiality, stated the materiality standard in the introduction to the opinion in somewhat different language than that used in the jury instruction:
[W]e conclude that a constitutionally valid perjury conviction under Penal Code section 118 requires the jury, not the court, to determine all elements of the charge including materiality, i.e., whether the statement or testimony “might have been used to affect [the proceeding in or for which it was made].” (Pen.Code, § 123; see People v. Pierce, supra, 66 Cal.2d at p. 61, 56 Cal.Rptr. 817, 423 P.2d 969.) We therefore reverse the contrary judgment of the Court of Appeal.
11 Cal.4th at 420, 45 Cal.Rptr.2d at 896, 903 P.2d at 1028 (second alteration in original). It does not appear that this dicta in Kobrin, derived from California Penal Code § 123, purported to overrule earlier formulations of the materiality standard. The issue of the applicable standard (as opposed to the proper decisionmaker) was not before the court, and Kobrin cites approvingly to Pierce, at the very spot *985where the “could probably have influenced the outcome of the proceedings” standard is stated.2
People v. Weinberg, 51 Cal.App.4th 1566, 60 Cal.Rptr.2d 323 (1997), a post-Kobrin decision of the California Court of Appeal, confirms that Kobrin, at most, textually reformulated the materiality inquiry without substantively lowering the standard:
The test for whether a statement is material has been stated as “whether the statement or testimony ‘might have been used to affect [the proceeding in or for which it was made]’ ” [quoting Kob-rin] or “whether the statement could probably have influenced the outcome of the proceedings,” [quoting Pierce ].
Id. at 329 (first alteration in original). We are therefore satisfied that Pierce’s “could probably have influenced the outcome of the proceedings” standard was unchanged by Kobrin.
2. Falsity
The falsity element of the crime of perjury requires that a statement be literally false. Misleading and nonresponsive testimony that is literally true cannot support a perjury conviction. See Cabe, 74 Cal.Rptr.2d at 336. As the California Supreme Court has stated:
[W]hen ... a witness’ answers are literally true he may not be faulted for failing to volunteer more explicit information. Although such testimony may cause a misleading impression due to the failure of counsel to ask more specific questions, the witness’ failure to volunteer testimony to avoid the misleading impression does not constitute perjury because the crucial element of falsity is not present in his testimony.
Rosoto, 10 Cal.3d at 949, 112 Cal.Rptr. at 647, 519 P.2d at 1071.
To avoid confusion, we note before proceeding to apply these standards to Chein’s case that our embedded application of California perjury law within federal habeas law creates the need to reference two distinct juries. Under Jackson, we ask whether any rational trier of fact could have found every element of the crime established beyond a reasonable doubt. In contrast, California perjury law inquires, for the purpose of establishing materiality, whether the allegedly false statement “could probably have influenced the outcome of the proceedings.” Pierce, 66 Cal.2d at 61, 56 Cal.Rptr. at 823, 423 P.2d at 975. The state’s theory of liability for all three counts is that Chein’s allegedly false statements “could probably have influenced the outcome” of the Lopez and Kancilia cases by unduly augmenting his credibility with the juries in those previous cases. We therefore find ourselves looking through the eyes of rational juries everywhere (under Jackson) to decide how the Lopez and Kancilia juries would have evaluated Chein’s credibility. •
With these understandings of California perjury law and of the task before us in mind, we turn to the specific counts on which Chein was convicted.
B.
The Specialty Count
On the specialty count, Chein was charged with falsely testifying that the *986specialty in which he was trained was “physical medicine and orthopedic surgery.” To evaluate the accuracy of Chein’s testimony, the following relevant facts were adduced at trial: Chein is a practicing doctor. The title of Chein’s residency program at the University of Southern California was “physical medicine and rehabilitation.” Although physical medicine includes training in orthopedic surgery, and Chein received such training during his residency, the focus of Chein’s residency was not orthopedic surgery, and, as Chein himself testified, he has not performed surgery since his residency. Chein is certified by the American Board of Orthopedic and Neurological Surgery, an institution that has not been approved by the American Medical Association.
Several doctors specializing in orthopedics and orthopedic surgery testified as to whether Chein was a “specialist” in orthopedic surgery. The expert testimony was generally consistent on two points: (1) doctors would not, as a matter of practice, have referred to Chein as a specialist in orthopedic surgery upon completion of his residency; and (2) Chein satisfied one American Medical Association (AMA) definition of a specialist, namely, “a physician with advanced training and knowledge of a particular branch of medicine or surgery.”
Given this evidence presented at trial, there are at least two respects in which Chein’s “specialty” testimony may have been literally true under Rosoto: He has been certified by the American Board of Orthopedic and Neurological Surgery, and, as the prosecution experts generally conceded, he satisfied the AMA definition of “specialist” quoted above, which only requires advanced training in a particular field.
Supporting the jury’s finding that at least one of Chein’s specialty statements was not literally true, however, is the convention among doctors to describe the spe- ' cialty in which one is trained according to the designation of the residency, which in Chein’s case, was physical medicine and ' rehabilitation, not orthopedic surgery. A rational jury could quite possibly have found the falsity element of the crime of ¡ perjury satisfied, because Chein’s statement that when he completed his residency he was a specialist in orthopedic surgery breached professional convention.
We need not, in the end, decide whether the jury’s conclusion regarding literal truth on this count is inconsistent with the Jackson standard, as Chein’s statement that he was a specialist in orthopedic surgery could not reasonably be found material beyond a reasonable doubt. Granting the element of falsity for purposes of our materiality analysis, Chein testified that he was a specialist in orthopedic surgery when he completed his residency, when he was in fact a physical medicine and rehabilitation specialist at that time. The distinction had no material significance in the context of the Lopez , trial. !
Under Jackson, we ask whether any rational trier of fact could find that materiality was proved beyond a reasonable doubt. 443 U.S. at 319. Again, as Jackson requires us to perform this analysis with explicit reference to California law, id. at 324, 99 S.Ct. 2781, our inquiry, precisely stated, is whether any rational trier of fact could find beyond a reasonable doubt that Chein’s testimony that he was a specialist in orthopedic surgery “could probably have influenced the outcome” of the Lopez trial. CALJIC § 7.21 (1996 Revision).
Lopez was a personal injury case in which Chein testified as an expert witness that the plaintiffs would require orthopedic surgery. The Lopez plaintiffs had not un*987dergone surgery at the time their personal injury suit went to trial. Thus, Chein’s testimony reflected his medical prediction about the need for surgery. As Dr. Cail-liet testified on direct examination, “[A]l-though we [non-surgeon orthopedic doctors] know the needs for, the demands for and the specificity of surgery, we per se could not perform surgery.” Chein, quité obviously, did not perform surgery on the Lopezes, nor did he testify concerning the performance of surgery, as would have been required, for example, in a malpractice ease concerning surgery gone bad. He was simply asked whether surgery was indicated.3
Dr. Cailliet testified that physicians of physical medicine and rehabilitation can also be identified as physiatrists or non-surgeon orthopedic doctors, and that a further difference between such doctors and orthopedic surgeons is that orthopedic surgeons decide what type of surgical procedure should be performed. However, there is no evidence suggesting that the precise type of surgical procedure among the available options for the Lopezes’ supposed maladies was a material issue during their personal injury trial, particularly given the Lopez defendant’s position that no injury was sustained at all, see infra.
Whatever differences may exist between the apparent credibility of orthopedic surgeons and non-surgeon orthopedic doctors must be viewed in the context of the Lopez trial, and, particularly, in light of Chein’s testimony that he had not practiced surgery in the five years since his residency. Thus, the more appropriate materiality inquiry is whether the difference between (1) a person trained as an orthopedic surgeon who has not operated since his residency, and (2) a non-surgeon orthopedic doctor who participated in twenty orthopedic surgeries during his residency but has not participated in any surgeries since his residency is sufficient to support a perjury conviction.4 Yfe believe it is not.
Put another way, the Lopez jury could not “probably” have determined a dispute regarding the relative credibility of two medical experts testifying about the need for future surgery based on the difference between Chein’s actual credentials and his *988testimony concerning the nature of his residency training. Rather, with or without Chein’s embellishment, the opposing expert, with many years of practice as an orthopedic surgeon, was much better qualified if, but only if, expertise in performing surgeries were deemed pertinent by the Lopez jury.
Further adding to the reasonable doubt regarding the materiality of any literal falsity in Chein’s testimony concerning his training as a specialist that a rational jury would have encountered at Chein’s trial was Judge Robert Altman’s testimony that the need for surgery was not really at issue during the Lopez litigation. The Lopez plaintiffs alleged that the car accident that had given rise to the lawsuit caused them substantial physical injury. By contrast, the Lopez defendants adopted the position that the plaintiffs in that case suffered no injury at all, and that the entire lawsuit was a sham. Given this substantial discrepancy in the positions taken by the parties in the Lopez case, the relatively insignificant difference between the injury diagnosis of an orthopedic surgeon and the injury diagnosis of a physical medicine and rehabilitation doctor who had advanced, albeit limited, training in orthopedic surgery cannot sustain a perjury conviction. No rational trier of fact could have found materiality beyond a reasonable doubt.5
'C.
The Office Count
, On the office count, the state’s.position is that during the Lopez trial, Chein falsely testified that he “had” only one office. The cross-examining attorney impeached Chein’s testimony on that point by reading from the transcript of a prior deposition. In that testimony, Chein stated that he practiced out of several offices: He was the sole practitioner at a Beverly Hills location, and he practiced with other doctors and chiropractors at locations in Southgate, Carson, and Long Beach.
Chein maintains that his statement that he only “had” one office was literally true because he only “owned” one office, although he practiced out of several. A jury could quite reasonably have disbelieved this contention. The business records admitted at trial strongly suggest that Chein actually owned the other offices, and the deposition transcript with which he was impeached at the Lopez trial documents Chein as having previously testified that the Southgate, Carson, and Long Beach offices “are under my own name.” While some doubt persists as to whether Chein actually ran the other offices or just worked at them, a rational jury could have found beyond a reasonable *989doubt that Chein’s testimony about the number of offices he “had” was false.
The only evidence introduced by the prosecution to demonstrate the materiality of Chein’s statement that he only “had” one office was the testimony of California Superior Court Judge Robert Altman, who presided over the Lopez case. Judge Altman testified not only as a percipient witness, but also as the state’s “expert” on materiality.6 When asked during direct examination about the materiality of Chein’s statement regarding the number of offices he “had,” Judge Altman stated:
The materiality is-the question of how many offices a doctor owns or runs is relevant to the kind of practice that that doctor has. A doctor can certainly have one office in Beverly Hills, a doctor can have another office and maybe he’s busy. But at some point, the argument is or the image is that the doctor runs a PI [personal injury] mill and that the doctor isn’t just in medicine just making money and that the doctor is handling all personal injuries and workmen’s compensation and then churning cases through the courtroom.
That’s the image that the lawyer was trying to bring out as to Dr. Chein. That’s why the number of locations is material to the qualifications.
As far as appears in the record, Judge Altman had no basis for determining “why” the lawyer in the Lopez case was interested in the number of offices Chein “had.” Judge Altman was neither a percipient nor an expert witness as to the mental state of the Lopez defense lawyer. Nor did the prosecution provide any objective basis for divining the defense’s reason for demonstrating that Dr. Chein, like many doctors, practiced from several offices. Lacking any basis other than Judge Altman’s bare, unsubstantiated assertion for drawing a connection between lack of credibility and “having” several offices, no reasonable juror would rely on the quoted language alone as evidence beyond a reasonable doubt that a doctor who practices out of several offices is less credible than a doctor who practices from a single office.
Indeed, Judge Altman himself did not stand on the proposition that doctors who, for the convenience of their patients, arrange to see them in multiple locations are not to be believed with regard to their medical evaluations. Rather, Judge Altman later clarified that his testimony regarding the materiality of several offices was intended to bolster a quite independent assertion: that Chein lacked credibility because he had never in fact examined the Lopez plaintiffs. Thus, Judge Altman testified as follows:
[Dr. Chein] said he had different offices that were in his name and he practiced at different locations in his name. And then, again, on cross-examination, when he realized that the defense was armed for bear as they were armed for bear, they-he started talking about he just worked there, even though they were in his name, he didn't see patients.
In fact, as far as the present record shows, Dr. Chein never said he did not see *990patients at the other locations. Instead, after confrontation with his deposition testimony regarding his other offices, Dr. Chein testified: “I do not have other offices. I work for those people. When I am short of money, I do extra work including emergency room. If I work for Cedars, don’t tell me I own Cedars.” So on this point, Judge Altman’s testimony was squarely refuted by material from the Lopez transcript read to the Chein jury.
That Judge Altman’s assertion regarding the materiality of the number of offices depended on his understanding that the additional offices were not used for seeing patients became more patent on cross examination:
Q. Now, Let’s talk about the number of offices.
A. Uh-huh.
Q. Now, is this a snobby thing like what law school did you go to or what medical school or is this a class thing that—
A. No. I think that there comes a point where people realize that if a doctor has multiple offices, that certainly it’s arguable that the doctor is running a mill, that the doctor is just seeing or not seeing patients and writing reports and billing people and sending them to court because that is the way the world works.
Q. But there are doctors who have multiple locations that are perfectly legitimate; isn’t that correct?
A. Yes, there are. Not too many that have four and five that I know of. I mean I’ve heard of doctors having one or two, but when it reaches a point of three and four, you know, you read about some of their credentials, they have multiple offices, I don’t know what they do, but one has to suspect.
Q. Some doctors — let’s take internists. They may have three or four offices; isn’t that correct?
A. Not too many that I know. I don’t know any.
Q. For example, Michael Gottlieb—
A. I’ve heard the name and I know there are doctors that have multiple offices who are legitimate. I think there are some doctors, yeah, like Dr. Gottlieb that can have multiple offices who are legitimate.
Q. Not only legitimate, but world rek-nown [sic ], right?
A. I don’t know the answer to that question.
Q. Dr. Gottlieb discovered the AIDS disease?
A. I don’t know. I’ll just take you on face.
Q. Would it surprise you to know he’s got offices in Santa Ana, Pasadena and Van Nuys?
A. Okay.
Q. Would that make him some kind of flake?
A. No, it doesn’t. I’m always a little troubled when doctors start to lend their name out to offices and clinics and they’re not there. It bothers me. Do I think the doctors themselves are unethical, no, but it troubles me quite honestly when doctors start to merchandise their names.
Also pertinent is the following passage from the prosecutor’s redirect examination:
Q. Does the number of offices have any bearing on the quality of care that a doctor can render to his patients?
A. I think there are certainly doctors who can practice at multiple offices and see patients. In this case, there *991was — there were no medical records. There’s a question of whether the patient had ever been seen.
(Emphasis added.)
Thus, both after cross examination and with an opportunity to clarify on redirect, Judge Altman returns to the same bottom line: the actual concern about Chein’s credibility at the Lopez trial was not that he practiced out of many offices, but rather, that he had never actually examined the Lopez plaintiffs.
This clarification of the ultimate import of Judge Altman’s testimony allows us to properly frame the materiality inquiry: Is the number of offices out of which a physician practices material to the question of whether he actually examined the patients about whom he testified? Any rational jury would respond in the negative.
The only evidence in the record that arguably supports the proposition that physicians practicing out of multiple offices are more likely than single office practitioners to lie about having examined their patients is Judge Altman’s testimony that the former group is more likely to “run[ ] a personal injury mill.” Judge Altman was neither a percipient nor an expert witness with regard to that question, and no rational jury would have thought otherwise.
Nor is Judge Altman’s assertion such common knowledge that the Chein jury could have concluded, based on their own life experience, that the Lopez jury would have regarded Dr. Chein as more likely to lie if he practiced from more offices. Absent some competent evidence to that effect — for example, an expert on medical practice — a reasonable lay jury would not perceive any connection between the number of offices a doctor practices from and his competence, general credibility, or likelihood to claim he has examined patients when he has not.
Ultimately, it appears that the prosecution at Chein’s perjury trial pursued a conviction for an immaterial falsehood— that he practiced out of only one office— because it was unable to prove a material one — that Dr. Chein had lied about whether he had actually examined the Lopez plaintiffs. Wise as he may be as a jurist, Judge Altman’s testimony on the practices of the medical profession was simply insufficient to generate materiality where it did not exist.7
Habeas relief is therefore warranted as to the office count.
D.
The University Count
Finally, Chein was prosecuted for his response to an interrogatory concerning his education. He responded that he attended “American University School of Medicine, Florida, 1979-1980,” when he in fact was enrolled in the American University of the Caribbean School of Medicine, which is located in the West Indies but has an office in Florida. (While -enrolled at the University, Chein’never actually attended classes in the West Indies, but instead did his hospital rotation in the United States.) The name of the school was written incorrectly on the interrogatory response, and there was evidence both that offshore *992medical schools are less prestigious than domestic ones and that Chein had used the same erroneous name for the University-on a resume, indicating that the error was purposeful. A rational trier of fact could have found that the interrogatory response was intentionally false beyond a reasonable doubt.
This interrogatory was propounded for the Kancilia case, which arose out of a dispute between Chein and Dr. Kancilia, a chiropractor, his business associate. Kan-cilia sued Chein, after the termination of their business agreement, for an accounting of funds. Additionally, there were allegations that Chein had committed fraud by hiding patient accounts. As a result of Chein’s alleged wrongdoing, Kancilia asked for punitive damages.
That the medical school from which Chein graduated has little pertinence to a dispute between business associates is evident. The business partner was a chiropractor; Chein was a doctor of medicine and a graduate of a residency program in physical medicine and rehabilitation at the University of Southern California. In a business dispute over money, it is extremely unlikely that medical credentials would be pertinent to credibility determinations, and, if they were, the fact that Chein graduated from an offshore medical school could not alone affect his credibility vis-avis that of Kancilia, a chiropractor, should credibility have become an issue.8
To fill this gap, the state’s theory of materiality, at trial, on appeal, and here on habeas review — although not relied upon by the California Court of Appeal — is that by suing for fraud, and by claiming punitive damages, Kancilia placed Chein’s credibility at issue, and that Chein’s false answer could be used to show him to be generally untrustworthy. Although a number of fundamental defects plague this theory of criminal liability, we will state only the most obvious.
The materiality inquiry does not focus on whether the trier of fact’s knowledge that a witness’s statements are false would impact the proceedings, but rather, on whether the outcome of the proceedings could probably have been affected if the trier of fact had known the truth in lieu of what it was told. In this ease, the pertinent question is whether it is probable that the outcome of the Kancilia proceedings, a fraud and contract case between business associates, could probably have been influenced by the knowledge that Chein attended a medical school in the Caribbean, as opposed to a medical school in Florida.
The state’s theory instead emphasizes the falsity of the statement actually made as a basis for impeaching the speaker’s credibility, maintaining that the falsehood is material because knowledge that the statement was false could be a basis for disbelieving the speaker at trial. But perjury law aims to discourage false statements, not to encourage such statements for use to impeach credibility. California case law does recognize that falsehoods going to the credibility of witnesses can be material, see, e.g., People v. Gamble, 8 Cal.App.3d 142, 87 Cal.Rptr. 333, 334-35 (1970). But in Gamble and similar cases, a witness either impairs the credibility of another witness by making a false statement that renders that witness materially less credible than would the truth, or materially embroiders his own credibility by making a false statement. Such cases do not regard false statements as material to credibility simply because, at trial, its very falsity, were it discovered, could be used for impeachment purposes. See, e.g., Peo*993ple v. Dunstan, 59 Cal.App. 574, 211 P. 813, 816 (1922).
The state’s theory would largely eliminate the materiality element of perjury. Falsity can always be used for impeachment purposes if the truth is discovered, yet opposing litigants have no opportunity to exploit this tactic if there is no falsehood in the first instance. On the state’s theory, that is, it might not matter what Chein lied about in his interrogatories, in deposition, or at trial. On that theory, any lie can be used for impeachment purposes in a fraud case if the lie is discovered, so any lie, including one having nothing to do with the issues in the case, would be material. California perjury law has not so abandoned the element of materiality.
Thus, evidence showing that Chein lied in one interrogatory answer and that Kan-cilia was a fraud case where punitive damages were claimed was insufficient under Jackson to establish the materiality of an otherwise immaterial false statement beyond a reasonable doubt.
IV.
Conclusion
As there was insufficient evidence under Jackson to sustain any of Chein’s three counts of conviction, the district court erred in denying habeas relief.
REVERSED.
. The California perjury statute, Cal. Penal Code § 118, provides, in pertinent part:
(a) Every person who, having taken an oath that he or she will testify, declare, depose, or certify truly before any competent tribunal, officer, or person, in any of the cases in which the oath may by law of the State of California be administered, willfully and contrary to the oath, states as true any material matter which he or she knows to *984be false, and every person who testifies, declares, deposes, or certifies under penally of perjury in any of the cases in which the testimony, declarations, depositions, or certification is permitted by law of the State of California under penalty of perjury and willfully states as true any material matter which he or she knows to be false, is guilty of perjuty.
. In section 123, the "might have been used to affect the proceeding” language seems not to be a definition of materiality but an additional requirement that materiality be judged ex ante, rather than as a question of how the false statement actually affected the trial. The statute, after all, says that a lie must both be "material, and might have been used to affect such proceeding.” Cal. Pen. Code § 123 (emphasis added) (quoted in full in text, infra ). Kobrin had no reason to delve into the precise import of the just-quoted portion of section 123 and did not do so.
. We reject the dissent’s attempt to marry California perjury law with California’s law governing expert witness testimony to develop a novel per se rule that "under California law, an expert witness’s credentials are inherently material to his or her testimony.” Post at 995. There is no suggestion in the California cases that the normal materiality inquiry is held in abeyance when dealing with expert witnesses.
The dissent then proceeds to suggest that Chein may have been precluded from offering his opinion had he accurately stated his credentials. There is no basis for this suggestion. Judge Altman did not so testify, even though, on that point, he was a percipient witness as the judge responsible for the decision. Moreover, the case did not concern the performance of any orthopedic surgery. Dr. Chein's actual, quite considerable specialty credentials were adequate to qualify him as an expert on the extent of physical injury and the likely need for surgery-
. The dissent’s "Nobel Prize in Medicine” example is therefore not helpful. Our analysis hinges, in part, on the .fact that the credibility enhancement Chein received, if any, was insignificant. Claiming to be a Nobel Prizewinning doctor, as opposed to the ordinary, albeit specialist, doctor he was could have been material. A jury is considerably more likely to believe a doctor in a pertinent field who has achieved the pinnacle of recognition available in a career in medicine than one who has no awards or prizes attesting to his unique brilliance. On the other hand, if it were made clear that the Nobel Prize was for work bearing no relationship to the issues in the case — if the award was for psychiatry or dermatology, for example — then the assertion would indeed be immaterial, as would be a false claim by a medical expert that he had won an Oscar for acting.
. Our result would be no different if we viewed this case through the § 2254(d)(1) prong of AEDPA. The California Court of Appeal decided otherwise in one sentence: "The jury could reasonably conclude defen- ■ dant's testimony was material to the outcome of the Lopez case because it related to his ability to convince the jury that his testimony was credible regarding the plaintiffs’ alleged injury.” This analysis is unreasonable in two respects: First, the state court addressed none of the uncontested evidence discussed in the text that shows that in the context of the issues before the jury, the difference between his actual credentials and his stated credentials were minor. Second, the California court also did not address Judge Altman's testimony that Chein's credibility as an expert witness was only marginally at issue, because the Lopez defendants had taken the position that no accident had occurred at all. As a result, the Court of Appeal's opinion cannot and does not explain how a rational jury could have found that Chein's alleged falsehood "could probably have influenced the outcome” of the Lopez trial. Pierce, 66 Cal.2d at 61, 56 Cal.Rptr. at 823, 423 P.2d at 975.
. That Judge Altman, a sitting state court judge, testified not only as a percipient witness to the events that transpired before him but also as the state’s expert witness on the element of materiality is the basis of a separate due process challenge raised by Chein. Although we do not reach the merits of that challenge, the likely impact on the jury of a sitting state court judge pronouncing the existence of an essential element of a crime, while vigorously denouncing the defendant and his credentials, is difficult to ignore. Without deciding whether the admission of Judge Altman's testimony amounted to a due process violation, we note that this highly unusual testimony at least explains why the jury returned a guilty verdict on sparse, constitutionally insufficient evidence.
. The California Court of Appeal’s conclusion to the contrary is not reasonable. In the same way that an eye-witness’ in-court identification is meaningless if the witness concedes on cross-examination that he had not been at the crime scene after all, Judge Altman's initial testimony must be analyzed in the context of the subsequent modification of his testimony. Judge Altman conceded on cross and redirect examination that the material issue at the Lopez trial was not the number of offices Chein had but rather whether he had actually examined the Lopez plaintiffs.
. If the § 2254(d)(1) prong of AEDPA applies, the California Court of Appeal's conclusion to the contrary was an unreasonable application of Jaclcson.