The defendant, James Lee Clark, has filed a successive habeas petition in this case, arguing that the evidence presented to the state courts established that he suffers from significantly sub-average intellectual functioning to the point of mental retardation and thus may not be executed. The district court concluded that the state court did not err in finding that Clark had failed to establish that he had significantly sub-average general intellectual functioning, the first element of the Texas test for mental retardation, and did not proceed further to the other elements. Clark argues that the district court erred in upholding the state court’s findings on that element and erred in refusing to consider Clark’s arguments as to the other elements of the test. Upon our review, we determine that the district court did not err in affirming the state court as to the “significantly sub-average intellectual functioning” element of mental retardation and that the district court did not err in considering only that element.
Standard of Review
This case is governed by AED-PA. For questions of law and mixed questions of law and fact, habeas relief may not be granted unless the adjudication of the claim resulted in a decision that was contrary to, or 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). A state court decision is contrary to clearly established federal law if it “applies a rule that contradicts the governing law set forth in [Supreme Court] cases” or “confronts facts that are materially indistinguishable from a relevant Supreme Court precedent” and arrives at an opposite result. Williams v. Taylor, 529 U.S. 362, 405-06, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000). A state court decision unreasonably applies clearly established federal law if it “identifies the correct governing legal rule” from Supreme Court cases, but unreasonably applies it to the facts of the particular case, or if it unreasonably extends a principle to a new context where it should not apply or unreasonably refuses to extend that principle where it should apply. Id. at 407-09, 120 S.Ct. 1495. For questions of fact, relief may not be granted unless the decision was based upon an unreasonable determination of the facts in light of the evidence presented in the State court proceedings. 28 U.S.C. § 2254(d)(1); Moore v. Johnson, 225 F.3d 495, 501 (5th Cir.2000). A factual determination made by a state court must be rebutted by clear and convincing evidence. 28 U.S.C. § 2254(e)(1).
Analysis
Initially, we note that Clark is incorrect in arguing that the district court *444erred in failing to address the other elements of mental retardation under the Texas definitions after it had determined that the state court did not err in finding that Clark did not have significantly sub-average intellectual functioning.
The Texas Court of Criminal Appeals adopted two definitions of mental retardation in the aftermath of Atkins, both of which contain the same substantive elements. The first, the AAMR definition, defined mental retardation as a disability characterized by “(1) ‘significantly subav-erage’ general intellectual functioning; (2) accompanied by ‘related’ limitations in adaptive functioning; (3) the onset of which occurs prior to the age of 18.” Ex parte Briseno, 135 S.W.3d 1, 7 (Tex.Crim.App.2004). The second, from the Texas Health and Safety Code, requires “significantly subaverage general intellectual functioning that is concurrent with deficits in adaptive behavior and originates during the developmental period.” Id. It is plain from the use of the words “accompanied by” and “concurrent” that both of these definitions require that all three elements exist to establish mental retardation. It therefore was not in error for the district court to determine that Clark could not prevail once it had already held that the state court had not erred in holding that Clark failed to meet the first element. If the state court correctly found that Clark failed to meet any of the three elements, he cannot demonstrate mental retardation under the Texas definitions.
We agree with Clark’s contention that the question of whether he suffers from significantly subaverage intellectual functioning is a question of fact, and not a mixed question of law and fact as determined by the district court. See United States v. Webster, 162 F.3d 308, 351-52 (5th Cir.1998) (discussing, in the context of the Federal Death Penalty Act, that the judge may act as fact-finder on the issue of mental retardation). However, Clark also raises the separate, legal question of whether federal law permits the state court the discretion to choose as the relevant score the base IQ score or the low point on the range that the score represents.
We review questions of law to determine whether the State court adjudication “resulted in a decision that 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). But we review questions of fact for whether the state court decision was based upon “an unreasonable determination of the facts in light of the evidence presented in the State court proceedings.” 28 U.S.C. § 2254(d)(2). The burden is on Clark to rebut the state court’s determination with clear and convincing evidence. 28 U.S.C. § 2254(e)(1).
Clark raises issues with respect to each of the three elements of mental retardation in Texas. With respect to the first element, the question of whether Clark had significantly subaverage intellectual functioning, Clark argues that the state court erred in considering the numerical IQ scores of Clark’s tests instead of the “confidence band,” or range of potential “true” scores someone with Clark’s score falls within. This confidence band is designed to , account for the measurement error inherent in intelligence testing, and indicates the upper and lower scores between which a psychologist conducting the test can be ninety-five percent confident that the “true” score lies.
The Texas Court of Criminal Appeals held when adopting its tests for mental retardation that scores gathered through intelligence testing are necessarily imprecise and must be interpreted flexibly. Bri *445 seno, 135 S.W.3d at 7, n. 24. The testing error, coupled with the differences between various IQ tests, mean that in many cases an individual who tests as having an IQ above 70, the rough cut-off for mental retardation, may still be diagnosed as mentally retarded, and vice versa. Id. Under this approach, courts should not rigidly consider an IQ score to be determinative of the defendant’s intellectual functioning.
To support his argument that the state court erred in choosing his base IQ score as the relevant IQ score rather than the lowest number in the confidence band, Clark argues that the Texas courts must apply the approach articulated by the American Association on Mental Retardation (“AAMR”), which dictates that IQ examiners account for the appropriate confidence band. He argues that the AAMR approach is the proper standard for determining whether an individual has subaver-age intellectual functioning. Clark is thus asking us to consider whether the Texas courts were properly given the discretion to choose between a base IQ score and a confidence band, see Briseno, 135 S.W.3d at 14 n. 53.
The Supreme Court, in Atkins v. Virginia, 536 U.S. 304, 122 S.Ct. 2242, 153 L.Ed.2d 335 (2002), “[left] to the State[s] the task of developing ways to enforce the constitutional restriction upon [their] execution of sentences.” Id. at 317, 122 S.Ct. 2242. Although the Court did refer to the clinical definitions of mental retardation promulgated by the AAMR and the American Psychiatric Association (“APA”), it did not dictate that the approach and the analysis of the State inquiry must track the approach of the AAMR or the APA exactly. It also did not mention the portion of the AAMR Manual upon which Clark relies in his argument. Therefore it is not “clearly established Federal law as determined by the Supreme Court of the United States” that state court analysis of subav-erage intellectual functioning must precisely track the AAMR’s recommended approach. See 28 U.S.C. § 2254(d)(1).
Under the definition of mental retardation as established by the Texas Court of Criminal Appeals, the state court did not unreasonably determine the facts in light of the evidence relating to Clark’s intellectual functioning. The court was confronted with multiple IQ tests taken by Clark. The first, taken in November of 1983 when Clark was 15, measured his IQ at 74. The “true” IQ range indicated by the confidence band for this test was 69 to 79. On April 17, 2003, after the Atkins decision, Clark took another test which measured his IQ at 65. The confidence band of this test indicated that his IQ ranged from 60 to 70. On June 5, 2003, Clark took a third IQ test which measured his IQ at 68, with a confidence range of 64 to 72.
Clark contends that the state court acknowledged the confidence bands, but “simply chose not to apply the standard error of measurement to Clark’s IQ score of 74.” This is incorrect. Regarding the 1983 IQ test, the state court noted that the test showed Clark’s IQ to be “in the range of 69 to 79.” More important to the state court were its findings based on its evaluation of expert testimony on the IQ tests. The court evaluated both of the tests administered to Clark in 2003, and it found that these specific tests were subject to manipulation that would result in lower scores and that in 2003 Clark had motivation to lower his scores deliberately. The court also found that there was an unusual discrepancy in Clark’s subscores on one of the 2003 tests indicating a higher IQ. The court further found - that the 1983 score was considered reliable by Clark’s experts and that the psychologist who conducted the test noted based on his observations *446that Clark’s intellectual functioning fell between the dull average and average range.
Clark has not challenged any of these findings, and instead asserts only that it was error for the state court to find mental retardation where the lowest potential score in the confidence band of the 1983 test was below 70. The explanation provided by the Texas Court of Criminal Appeals in adopting its definition of mental retardation plainly forecloses this argument. Briseno, 135 S.W.3d at 7, n. 24. The state court was required to evaluate the intelligence testing and make a flexible determination based on those tests as to whether Clark had “significantly subaver-age general intellectual functioning.” The court was not required to find Clark to be mentally retarded merely because the low end of Clark’s confidence band was below 70, just as it would not be required to find that Clark could be executed on the basis that the high end of this band fell above 70. Clark thus has not shown that the state court made an unreasonable determination of the facts in light of the evidence as to the first element of mental retardation.
While, as discussed above, all three elements of mental retardation must be shown in order to meet the Texas definition, we review Clark’s arguments as to the other elements out of an abundance of caution and because the evidence as to these elements is informative as to the state court’s belief that Clark may have attempted to manipulate his 2003 IQ scores.
Clark has challenged the state court’s finding that he did not have adaptive deficits. Clark argues here that “limitations often coexist with strengths” and that the state court erred in dismissing adaptive behavior testing and relying “exclusively upon the court’s own interpretation of lay testimony about Clark’s adaptive strengths.”
The record shows that the state court made a number of findings of fact that support its finding that Clark did not have significant limitations in his adaptive skills. The court found that records from Clark’s youth showed that he completed his GED with improving grades after initial problems, that he completed a welding program at Cooke County College in 1985, that he was employed by the Gainesville State School with numerous duties and positive reports from supervisors, and that he was able to get along with other people.
The state court also heard testimony from Clark’s former landlord that he was a tenant of a mobile home park, where he was paying his own bills, doing chores in exchange for rent reduction, playing cards, and successfully socializing with others. The landlord testified that he followed the rules of the park, kept his mobile home clean and cut the nearby grass, and was able to both drive a car and follow the speed limits of the park. The landlord’s daughter testified to similar activities by Clark, and added that they had conversed on a number of occasions without difficulty.
A Texas Ranger who investigated the case testified that he interviewed Clark for several hours without noticing any difficulty by Clark in understanding the questions. He stated that Clark was able to think on his feet, modifying his story in response to inconsistencies that were pointed out to him. He further testified that Clark’s actions in the crime included several that showed adaptive functioning, including removing the butt stock of his gun to make it easier to conceal, purchasing ammunition for the gun, practicing with the gun, and removing evidence from the scene and concealing it. He testified that his investigation of Clark showed no *447evidence of adaptive limitations or problems meeting basic human needs.
The court listened to testimony from several other officers and prison officials who had interacted with Clark. It also heard testimony from the warden’s secretary, responsible for handling inmate requests, and found that several of Clark’s written requests indicate adaptive skills ranging from average to sophisticated.
Clark has not challenged the accuracy of any of these findings, but instead argues only that they do not support the state court’s determination that Clark did not have adaptive deficits because they are evidence of strengths and not limitations. This is incorrect — evidence of a strength in a particular area of adaptive functioning necessarily shows that the defendant does not have a weakness in that particular area. Even if, as Clark argues, adaptive limitations rather than strengths often define mental retardation, the evidence in this case shows primarily adaptive strengths and does not show limitation in any significant area. The evidence in this case showed that prior to being incarcerated, Clark functioned normally across a broad range of adaptive behaviors. Our review of the evidence of Clark’s behavior in prison casts serious doubts on his claims of adaptive limitation, as evidence collected from his cell along with his handwritten requests include complaints that he needed a technician to fix his television as it had been several “weeks now of no reception via my coaxial cable hooked up to the jack on the wall;” a handwritten diet plan entitled “Eat to Beat Stress” noting that he should “eat small meals and snacks several times a day to keep blood sugar from fluctuating” as well as notes about the effects of various chemicals such as folic acid, pyridoxine, and thiamine; handwritten puzzles including the decipherment of several extremely complicated codes; and complaints about delays in approving his request for a legal visit with another inmate in which Clark planned to assist the inmate in obtaining parole.
Clark faults the failure of the state court to credit an adaptive behavior assessment administered at age 34 which attempted to retroactively determine his abilities at age 25. The court found that the test was unreliable because it relied on Clark’s self-reporting of his adaptive limitations coupled only with his ex-wife’s memories about what he could and could not do at age 25. The court found that this testing was unreliable because it did not account for the incentive of Clark and his ex-wife to misreport Clark’s adaptive skills and did not take into account Clark’s prior employment and the written materials he produced during prison. The court heard the testimony of the State’s expert claiming Clark did not display limitations in adaptive functioning. This expert reviewed the testing results along with the various testimony and documents and was judged more credible by the state court. The findings by the state court on the issue of adaptive functioning have not been rebutted by clear and convincing evidence, and its determination that the evidence of Clark’s actual behavior was more credible than the adaptive behavior assessment administered to Clark was not unreasonable in light of the evidence presented.
Finally, Clark objects that as to the third element of retardation, onset before the age of 18, the state court again erred in its factual finding that Clark did not show onset before the age of 18. Clark points to the opinions of his expert witnesses, his failure of several grades, and his participation in special education classes in school. While the state court provides less support for its determination on this issue, pointing only to documents about Clark produced by the Texas Youth *448Council when he was age 15, Clark has not rebutted these findings by clear and convincing evidence. The TYC documents describe Clark as a troubled child with intellectual potential between the dull average and average range. The TYC conducted an IQ test which, as the state court determined under the first element, showed that at age 15 Clark was not mentally retarded. The state court did not make an unreasonable determination of the facts in light of the evidence as to the age of onset element.
Because Clark has not shown by clear and convincing evidence that the state court made unreasonable factual determinations in light of the evidence presented, we AFFIRM the decision of the district court.