Defendant Micheál Lee Wilson1 was convicted of first-degree murder and robbery with a dangerous weapon in Oklahoma state court and sentenced to death. On direct appeal the Oklahoma Court of Criminal Appeals (OCCA) ordered dismissal of his robbery conviction but affirmed his murder conviction and death sentence. See Wilson v. State, 983 P.2d 448, 463, 473 (Okla.Crim.App.1998) (Wilson I). Defendant sought a writ of habeas corpus under 28 U.S.C. § 2254 in the United States District Court for the Northern District of Oklahoma, but the district court denied his application. See Wilson v. Sirmons, No. 00-CV147CVEFHM, 2006 WL 2289777 (N.D.Okla. Aug. 8, 2006) (Wilson II). We affirmed in part, but vacated and remanded for an evidentiary hearing on Defendant’s claims that he received ineffective assistance of counsel at the sentencing phase of his trial. See Wilson v. Sirmons, 536 F.3d 1064 (10th Cir.2008) (Wilson III), reinstated sub nom., Wilson v. Workman, 577 F.3d 1284 (10th Cir.2009) (Wilson TV) (en banc). After holding the evidentiary hearing, the district court ruled that Defendant had failed to establish that his trial counsel had performed deficiently or that counsel’s alleged failures had affected the outcome of the penalty phase, and it again denied the writ. See Wilson v. Workman, No. 00-CV-0147-CVE-FHM, 2011 WL 744661 (N.D.Okla. Feb. 23, 2011) (Wilson V). The district court granted Defendant a certificate of appealability (COA) on his ineffective-assistance-of-counsel claim, and Defendant appealed. See 28 U.S.C. § 2253(c)(1)(A) (requiring a COA to appeal the denial of a § 2254 application).
We affirm the denial of Defendant’s habeas application. In light of the evidence presented at the hearing before the district court, he has not shown that he was prejudiced by the alleged deficiencies in his counsel’s performance at trial.2
1. BACKGROUND
A. The Murder
In 1995 Defendant worked at a QuikTrip convenience store in Tulsa, Oklahoma. He, along with accomplices Billy Alverson, Darwin Brown, and Richard Harjo, planned to rob the store. In the early morning hours of February 26, 1995 (two days before Defendant’s 20th birthday), the four men entered the QuikTrip, loitering for about an hour while Defendant conversed with the victim, Richard Yost, the employee on duty. In one chilling exchange, Defendant, just 14 minutes before the assault began, asked Yost how long he planned to work at QuikTrip. When Yost answered that he hoped to become store manager someday, Defendant responded “For real?” R., Vol. 3 Tr. Feb. 20, 1997, at 29-30. The conversation and other events in the store were captured on the store’s surveillance-camera *1289recording, obtained by law-enforcement officers from Alverson’s home.
While Yost was cleaning the store’s coolers, the four men attacked him and dragged him into a back room. Alverson and Harjo briefly left the store while Yost screamed for help. The two men returned with a black aluminum baseball bat and went to the back room, where the robbers beat Yost to death with the bat. Yost was handcuffed during the beating: a piece of handcuff later recovered from his skull indicated that he was conscious and attempting to ward off blows for at least part of the fatal attack.
During the beating Defendant left the back room, donned a QuikTrip jacket, and began attempting to remove the store’s safe from its position under the counter. As customers entered the store, Defendant greeted them, rang up their transactions, and wished them a good day. After dislodging the safe, Defendant and his three accomplices fled the store with the safe, the contents of the cash drawer, and the surveillance video. Yost’s body, lying in a pool of blood, milk, and beer, was discovered by a customer within a few hours.
All four culprits were arrested later that day. Under police interrogation Defendant confessed to participating in the crime. He stated that the robbery and Yost’s killing had been planned for two weeks. Corroborating this admission was evidence that the body alarm typically worn by QuikTrip employees on overnight shifts had been found missing from the store the week before the murder. The safe, the surveillance tape, and several other items from the QuikTrip were recovered from Alverson’s home.
B. The Penalty Phase at Trial
Evidence was presented during the guilt phase of Defendant’s trial on February 11, 12, and 13, 1997. He did not present any evidence other than the tape-recorded statement of codefendant Brown. Given the irrefutable evidence of guilt, the penalty phase, which began on February 18, was the true contest.
1. The State’s Case
The State alleged three aggravating factors to justify the death penalty: (1) that the murder was committed for the purpose of avoiding or preventing arrest or prosecution, (2) that the murder was especially heinous, atrocious, or cruel, and (3) that Defendant was a continuing threat to society because of the probability that he would commit future crimes of violence. See Okla. Stat. tit. 21, § 701.12 (2011). The first aggravator — murder to avoid arrest or prosecution — was obvious from the evidence at the guilt stage of trial. For the other aggravators, the State offered additional evidence.
To support the second aggravator, a forensic pathologist testified that in his opinion Yost would have suffered during the attack unless he had been rendered immediately unconscious, and he described Yost’s injuries, including wounds to Yost’s hand and scalp, that indicated his efforts to defend himself. See Willingham v. State, 947 P.2d 1074, 1084 (Okla.Crim.App. 1997) (For a jury to find that the murder was heinous, atrocious, or cruel, “the State must prove conscious serious physical abuse or torture prior to death.”).
To prove the continuing-threat aggravator, the State called police witnesses who testified about two recent offenses. First, Defendant had been convicted as an accessory after the fact to the murder of a woman in September 1994 (five months before the Yost murder). The conviction was based on his taking possession of a gun at the principal’s request, although, as noted by the OCCA, “the facts revealed that he may have been more involved in this drive-by shooting by providing ammu*1290nition for the gun on the day of the murder.” Wilson I, 983 P.2d at 466. Second, 10 days before Yost’s murder, police had seized a loaded revolver from Defendant’s car during a traffic stop. Codefendant Brown was also in the car. On cross-examination Defendant’s counsel pointed out that Brown was the one charged with possession of the gun and suggested that Defendant was not aware of the gun’s presence.
The State also presented victim-impact statements from Yost’s wife and mother, both of whom testified to how their lives had been tragically affected by Yost’s murder.
2. The Defense Mitigation Case
a. The Lay Witnesses
In the mitigation case the defense put on five lay witnesses and one expert. Two of Defendant’s former teachers and two fellow church members offered similar testimony: that they had known Defendant as a polite, respectful, well-behaved, and intelligent young man and that the murder for which he had been convicted did not represent “the Mike Wilson [they] knew.” R., Vol. 3 Tr. Feb. 19, 1997, at 11. On cross-examination the prosecutor elicited that three of the witnesses had last seen Defendant between two and five years earlier and that his criminal conduct suggested that his behavior and character might have changed in the interim.
The other lay witness was Defendant’s mother, Patricia Taylor. She testified that his father frequently used crack cocaine, was often absent, and was not a positive influence on Defendant’s life; that she had cultivated a loving relationship with Defendant and had attempted to inculcate positive values in him; and that she had visited Defendant in jail almost every week. She also testified that she turned over to police some physical evidence of the murder, including the murder weapon, explaining that she had acted to follow “the rules,” and that she expected that doing so would help her son in the long run. Id. at 99. She further testified that immediately after the murder (but before his apprehension), Defendant had returned home in a disturbed emotional state. He said that a man had been killed in a robbery at the QuikTrip, although he did not know how it happened. They then prayed together about the crime. Ms. Taylor concluded with a plea for her son’s life.
b. Dr. Reynolds
Defendant’s expert witness was Dr. Allan Eugene Reynolds, a clinical psychologist. On direct examination Dr. Reynolds testified to some aspects of his pretrial preparation, stating that he had reviewed information about Defendant’s background, including his school, medical, and criminal records; that he had interviewed Defendant’s mother; and that he had reviewed statements by lay witnesses. In addition, he had met with Defendant on three occasions at the Tulsa County Jail, and had administered several psychological tests to Defendant: the Slossen Intelligence test; two tests designed to screen for organic brain damage, the Bender Gestalt test and the Memory for Designs test; and two personality tests, the MMPI-2 and the MCMI-III. Defendant also completed a psychosocial questionnaire, and Dr. Reynolds observed and interviewed Defendant in addition to the formal testing.
Dr. Reynolds was not asked about a failure in the testing. He had administered the MMPI-2 on February 6, 1997, and sent the responses to be scored by computer. The PhD psychologist who scored the test reported that the test was probably invalid, and Dr. Reynolds informed Defendant’s counsel of the invalidity by February 10, the day before trial *1291testimony commenced. But no retest was administered.
Dr. Reynolds testified to the results of the valid tests. He reported Defendant’s IQ as 126, placing him in the “superior” range of intelligence. Id. at 55. He said that there was no indication of organic brain dysfunction but that the personality tests showed that Defendant suffered from a “severe personality disturbance.” Id. at 57. He explained:
[Tjhere were responses and there were scores that indicate that [Defendant] has some very unusual, bizarre types of thinking. That would suggest that at times he’s not or has not periodically been in touch with reality. That he basically does not necessarily function at times in a normal state, but that he has a great deal of emotional pathology.
Id. Dr. Reynolds also said that in reviewing the statements of those who knew Defendant and in observing Defendant himself, he found it “remarkable” that
no one seemed to have any indication that he would engage in the types of behavior that he did.
And then on the other hand, the psychological tests show that he has that propensity to engage in that type of behavior. And so there’s a big conflict in terms of what people observed of him and maybe what was going on inside of him.
Id. at 58.
When asked if his psychological analysis could explain “the two Michael Wilsons,” Dr. Reynolds spoke of the differences between Defendant’s two parents. Id. at 59. He said:
Well, the social history questionnaire indicates that — and his — his past that he grew up in a family where the father left, who was involved in drugs and alcohol, and pretty much was not involved in Michael’s life.
The mother was very strong, caring, disciplinarian and tried to keep the family together. Michael was involved in church. And then, I believe one of the persons that he identified with very closely, a Sunday school teacher, died of cancer. From my interviews with him, he indicated to me that this was very distressing and very upsetting to him.
There was a lot of gang activity in the neighborhood where he grew up. The mother sent him off to North Carolina for a while with her sister. And Mike did very well out there in a different environment.
So on the one hand, here’s a young man who has an uninvolved father, who’s a role model, was involved in drugs and alcohol, and not particularly caring. And on the other hand, you have a highly structured mother, who provides church and this sort of thing.
So you have the two pictures of Mike. On the one hand, you have the picture of the Sunday school-going child. On the other hand, you have the picture of the gang and uninvolved father, who did not set a particularly good role model.
When he got out of both of those environments and went off to North Carolina, apparently there he did very well. I believe that his sister and his brother-in-law were involved in the Marines, provided a lot of structure for him.
So if we look at the environments to which he was exposed to does explain somewhat of the two type of Michael’s [sic] that you have, depending where he was at as to who he identified with.
Id. at 59-61.
Dr. Reynolds testified about the gang violence to which Defendant had been exposed from a young age. He identified several particularly traumatic experiences: Defendant’s being shot in a drive-by attack *1292when a young adolescent, the torching of his home by rival gang members, and the death of his Sunday-school teacher from cancer.
Dr. Reynolds also suggested that Defendant could be rehabilitated because of his superior intelligence. He explained:
[Defendant’s high level of intelligence] provides him, I think, with the intelligence to do something with himself, as well as be a contributor to others in terms of understanding his world, understanding what he’s done, what other persons have done. And therefore, with this level of intelligence, he has the ability to make contributions that may be helpful to those who have lesser intelligence.
Id. at 63.
The prosecutor’s cross-examination of Dr. Reynolds elicited a much more disturbing picture of Defendant. It began with the suggestion that Defendant’s personality disturbance was that of a psychopath:
Q: Doctor, in your experience are there psychopathic criminals who have superi- or intelligence?
A: Yes.
Q: Makes them all the better, doesn’t it?
A: It can.
Q: You referenced a severe personality disturbance and mental disorder, I believe you testified?
A: Yes.
Q: Doesn’t that mean he’s a psychopath?
A: No.
Q: It doesn’t?
A: No.
Q: The sharp contrast he exhibits that you testified to, aren’t those the classic designs of a psychopath? ‘Tes” or “no”?
A: It can be.
Q: And aren’t psychopaths the most likely to re-offend, based on the studies?
A: Yes.
Id. at 65. Later, Dr. Reynolds agreed that “a 19-year-old with superior intelligence know[s] right from wrong.” Id. at 66.
The prosecutor used published studies to attack the validity of the intelligence and personality tests employed by Dr. Reynolds. But he also elicited some of their conclusions. Dr. Reynolds conceded that the MCMI-III interpretive report stated that “[t]he guiding principle of [Defendant] is to outwit others, exerting power over them before they can exploit him,” id. at 69 (internal quotation marks omitted), and that Defendant was “easily provoked” and “may express sudden and unanticipated brutality,” id. at 70 (internal quotation marks omitted). He also acknowledged that Defendant had responded “True” to the following statements on the test questionnaire: “Lately, I have begun to feel like smashing things”; “I often get angry with people who do things slowly”; “I have had to be really rough with some people to keep them in line”; and “I sometimes feel crazy-like or unreal when things start to go badly in my life.” Id. at 70-71 (internal quotation marks omitted).
Later in the cross-examination the prosecutor returned to the psychopath theme, referring to Dr. Reynolds’s prior testimony about the sharp contrast between Defendant’s behavior before family and friends and his violent criminal conduct:
Q: [I]sn’t this consistent with the characteristics of a psychopath?
A: As I stated before, it can be.
Q: Well, aren’t superficial charm and good intelligence, coupled with cunning and manipulative lack of implusivity [sic] behavior characteristics of a psychopath?
A: Yes, they are.
*1293Q: And that’s what Mr. Wilson has, isn’t it?
A: Some of those characteristics, he has.
Q: And you told this jury that psychopaths are the most likely to re-offend based on the studies?
A: Yes.
Id. at 76.
Dr. Reynolds acknowledged that he did not know what mental-health treatment options would be available to Defendant in prison, and he agreed that without treatment, Defendant would represent a continuing threat to others. Dr. Reynolds also conceded that past violent behavior is the best predictor of future violence.
On redirect examination Dr. Reynolds defended his use of the psychological instruments challenged by the prosecutor. He also expanded on the significance of Defendant’s answer on the MCMI-III that he felt “crazy-like”:
Q: What things were you referring to, sir, whenever you told us there were evidences of severe mental disorder in this person?
A: Well, I think probably the most classic example would be in the Noteworthy Responses when [the prosecutor] asked me to read one of the questions, which “Sometimes I feel crazy-like or unreal when things start to go badly in my life,” he answers true.
Id. at 82. Dr. Reynolds went on to note that “[u]sually, unless someone’s pretty disturbed, they’re not going to answer that true, they’re going to answer that false.” Id.
Dr. Reynolds further testified that the tests were mere components of his overall evaluation of Defendant, to be interpreted in light of clinical experience, interviews, observations, and Defendant’s social history. He stated that Defendant’s mental disorders were treatable, reiterated his opinion that Defendant’s high intelligence made him a good prospect for rehabilitation, and agreed that in light of Defendant’s upstanding behavior in school and church, a structured setting such as prison might have a positive influence on Defendant’s conduct and mental state.
During closing argument the prosecutor repeatedly referred to the cross-examination of Dr. Reynolds, describing Defendant as a “psychopath” and a “psychopathic killer.” Id., Tr. Feb. 20, 1997, at 46. The defense attorneys had little to counter the argument. They did not discuss Defendant’s mental disorder during closing, except to suggest that the “psychopath” label obscured Defendant’s humanity and positive characteristics. Their main themes in closing were that there was no evidence proving that Defendant personally participated in the beating of Yost, and that Defendant could reform his life and be of benefit to others if allowed to live.
The jury found all three aggravating circumstances beyond a reasonable doubt and sentenced Defendant to death.
C. The § 2254 Evidentiary Hearing
From the time of his original appeal to the OCCA to the present, Defendant has contended that he was denied effective assistance of counsel during the penalty phase of his trial. Adopting the language of a member of this court, Defendant has referred to the cross-examination of Dr. Reynolds as a “ ‘train wreck.’ ” E.g., Aplt. Br. at 11 (quoting Wilson III, 536 F.3d at 1076 (McConnell, J.)). He attributes Dr. Reynolds’s poor performance to tardy and inadequate preparation by defense counsel. He emphasizes that counsel did not retain Dr. Reynolds until shortly before trial and failed to arrange for Defendant to retake the MMPI-2 personality test when the initial results turned out to be invalid. He contends that trial counsel also failed to conduct adequate interviews of family *1294members that would have enabled Dr. Reynolds to arrive at a diagnosis that could have changed the jury’s decision to impose the death penalty.
To give Defendant an opportunity to show the inadequacy of his counsel’s performance and what could have been accomplished through the additional steps that he believes counsel should have taken, this court ordered the district court to conduct an evidentiary hearing. See Wilson III, 536 F.3d at 1096. The hearing was held on July 28, 2010. Testifying were Dr. Reynolds; Defendant’s lead trial counsel, Joe Paul Robertson, who was director of the Oklahoma Indigent Defense System at the time of the hearing; and Defendant’s second-chair trial counsel, Kent Hudson. For our purposes, we need focus only on Dr. Reynolds’s testimony.3
.1. Dr. Reynolds’s Testimony on Direct
The thrust of Dr. Reynolds’s direct examination was to show that if defense counsel had asked him to do before trial what he did only after the jury had sentenced Defendant to death, he could have arrived at a diagnosis of mental disease that would have explained Defendant’s evil misconduct in a way that would generate sympathy from the jury. Dr. Reynolds was purportedly able to arrive at that diagnosis-schizophrenic paranoid personality disorder or paranoid schizophrenic disorder — because he now had a valid MMPI test and had interviews (of Defendant’s family and girlfriend) establishing the delusions and hallucinations necessary for the diagnosis. The new diagnosis would also have enabled him to fend off suggestions by the prosecutor that Defendant is a psychopath.
Defense counsel began by marching Dr. Reynolds through his pretrial preparation and, his meetings with Defendant. In answer to a question by the court, he then explained the meaning of the invalidity of the MMPI-2 test:
[Wjhen the individual takes the test in a manner that shows that there may be a variety of reasons that would not allow the test results to be used simply because he answered and his responses were in such a way, that, let’s say, for example, maybe he had a low reading level, he was confused, or he tried to manipulate the test, or let’s say he was inconsistent with his answers.
R., Vol. 2 at 62. Even though the test was invalid, Dr. Reynolds thought that the results “indicated that there was a severe disturbance,” id. at 63, and he felt that he needed more information to determine what was going on. (The computer-scored report on the test said, “Many individuals with this profile are considered to have severe Personality Disorders; however, the possibility of Schizophrenia, Paranoid type, or of a Bipolar Affective Disorder should also be considered.” Id., Vol. 1 pt. 5 at 912. And the report’s section on “Symptomatic Patterns” said that Defendant “endorsed a number of extreme and bizarre thoughts, suggesting the presence of delusions and/or hallucinations.” Id. at *1295910.) Dr. Reynolds thought that he would have told defense counsel “that the test was invalid, we can’t use this information, but there is further information here that is important and that is not totally supported by the MCMI but there’s a need to really know more.” Id., Vol. 2 at 65.4 He explained:
The MCMI really never reported the auditory hallucinations, okay? But where I began to get a tie between the two was the MCMI talked about the paranoid personality. Then I hear the MMPI, the way he answered those questions, talking about voices, et cetera, et cetera, et cetera, and then that began to give me information that there’s something more here that I need to investigate.
Id. He said that if he had had more time he “would have redone the MMPI again and probably given him instructions in terms of how, if he didn’t understand a question, or if he was unable to comprehend what it was asking, to make a note of it and I would assist him with it so that he would be able to give a more valid result.” Id. at 67. He added that he had now done those things.
Dr. Reynolds then testified about the valid pretrial MCMI-III test. The results indicated “that there was a severe personality disturbance, and that the test gave a variety of diagnoses that could be eonsideced.” Id. at 68. These included narcissistic personality disorder with passive-aggressive personality traits, schizotypal personality features, generalized anxiety disorder, bipolar disorder, posttraumatic stress disorder, and paranoid personality disorder.5
As for the disorder he now diagnosed Defendant as having — schizophrenic paranoid personality disorder — Dr. Reynolds explained that “it would have to have with it auditory hallucinations, and that’s where they’re hearing things, they’re hearing things to be told to do, or they’re hearing things, for example, that they’re a special person, or maybe there’s evil spirits.” Id. at 78-79. He said that people with the disorder “have the paranoid ideas that people are out to get them, that they don’t really trust people, that they’re suspicious of others, and he met that diagnosis.” Id. at 79. But, he added, people with the disorder can function in society:
[Everybody kind of thinks of schizophrenia as somebody who is dysfunctional, completely can’t talk straight, or anything like that. But on the paranoid— schizophrenic paranoid personality disorder, the person still appears quite able to function, and Mr. Wilson did, so he met that criteria as well. It was the combination of his paranoid ideas, the *1296delusions, and the auditory hallucinations that fit that diagnosis.
Id. (emphasis added).
Explaining how he arrived at his diagnosis, Dr. Reynolds testified that the statements by lay persons were critical:
[W]hat clued me in is when he indicated to me about the voices, but when I then got collaborative [sic] information that verified that from the girlfriend, his brother, and some sense of it from his mother, that somehow solidified my essence that now I felt more on the right track, but that’s not what I had before.
Id. at 71-72.
In response to the court’s question about how the diagnosis would have affected the jury, Dr. Reynolds said that the mental illness could explain Defendant’s motivations:
Well, I don’t know how a jury thinks, but I think it may have helped them understand that the motivation for what he did, once they knew about the delusions, once they knew about the voices, and once they knew about his mental illness, would have helped them know a little bit more that maybe he was influenced by his mental illness as averse to being just a raw, bad guy that’s just out there for the fun of it, but that the mental illness had some impact as opposed to being a normal guy that just goes out there and does something heinous, as he did. So I think it had to do with them and it would depend on what their opinion of mental illness is. So I think just knowing that the motivation may have been driven somewhat by his mental illness as opposed to not having that disorder.
Id. at 72. Defense counsel later pursued the matter further, asking what was “the most crucial information” about Defendant that Dr. Reynolds could have given the jury but did not. Id. at 89. Dr. Reynolds replied, “I think the discussion of his mental disorders,” and went on to assert that this information “was crucial in allowing the jury to know every aspect of Mr. Wilson in this part of the trial, to be able to understand him, possibly understand some of his motivations, understand his behavior, and that his behavior may have had something to do with his mental illness.” Id. Dr. Reynolds said that the schizophrenia diagnosis would have presented “a different picture” of Defendant had it been available at trial. Id. at 109. He explained that “my part of the trial, it was basically his intelligence and that he was disturbed. And the picture that we have now is someone who is very seriously disturbed with a psychiatric diagnostic disorder that is very, very severe.” Id. He stated that the posttrial diagnosis “superseded the antisocial diagnosis,” id. at 96, “simply because of the delusions, the auditory hallucinations, and the test data.” Id.
Dr. Reynolds also testified that he could have corrected the prosecutor’s use of the term psychopath if he had been properly questioned on redirect examination:
Q: Do you believe that Mr. Wilson’s attorney allowed you to explain the term “psychopath”?
A: No, he did not.
Q: Okay. Could you explain the term “psychopath” to the jury had you been asked?
A: Yeah. Yes.
Q: What is the meaning of a psychopath?
A: Well, there is no meaning because it doesn’t exist.
Q: [W]ell, is it outdated, that it used to exist?
A: It’s outdated, yes.
Q: Okay. Whether it doesn’t exist now or is outdated or not, does that term carry a stigma?
*1297A: Probably to the lay persons it does, yes. Q: And what would that stigma be?
A: Well, I guess I’ll just put it very simply, in that this is a very dangerous, bad person who has no regard for life of others.
Q: Did you feel the jury was left with an inaccurate depiction of Mr. Wilson?
A: Well, I think they got part of Mr. Wilson, but I don’t think they got more of what was there, but I can’t say that it was inaccurate. [I]f the part maybe you’re referring to is being defined as a psychopath, that was inaccurate. But they did get other parts of him.
Q: Is being depicted as a psychopath harmful?
A: Certainly, yes.
Id. at 91-92.
2. Cross-Examination of Dr. Reynolds
The prosecutor’s cross-examination of Dr. Reynolds at the evidentiary hearing was, if anything, more devastating to Defendant’s case than the trial cross-examination of Dr. Reynolds had been. Rather than reporting the cross-examination in chronological order, we organize it by subject matter.
a. Use of the Term Psychopath
The prosecutor began by pointing out that the term psychopath was not as clinically inappropriate or outdated as Dr. Reynolds had suggested. He noted that the American Psychiatric Association’s widely used and authoritative Diagnostic and Statistical Manual of Mental Disorders (4th ed. Text Revision 2000) (DSM-IV-TR), itself says that the term has been used as another name for antisocial personality disorder. Indeed, Dr. Reynolds himself had used the term shortly after the trial. The affidavit that he drafted for Defendant’s state-court appeal stated that “psychopath is a term associated with antisocial personality disorder.” Id., Vol. 1 pt. 5 at 1011. (The statement was deleted from the affidavit filed with the OCCA.) Dr. Reynolds responded that it would nevertheless be incorrect to use the term because “it’s not a DSM-IV diagnosis any longer,” id., Vol. 2 at 114, and, in any event, as his affidavit quoted the DSM-IV, “ ‘Antisocial behavior that occurs exclusively during the course of schizophrenia or a manic episode should not be diagnosed as antisocial personality disorder’ or psychopathic personality.” Id., Vol. 1 pt. 5 at 1011.
b. Effect of Mental Disorder on Defendant’s Participation in the Murder
The prosecutor challenged whether Dr. Reynolds’s diagnosis of Defendant, even if correct, could help explain his behavior at the QuikTrip:
Q: Doctor, I believe you mentioned on direct examination that you believe your diagnosis of paranoid schizophrenia may have helped the jury understand Mr. Wilson’s motivation for what he did. Is that what you testified to?
A: Something to that effect.
Q: And, Doctor, you’ve reviewed the surveillance tape from the Quik Trip depicting Mr. Wilson’s actions during the commission of the murder; correct?
A: I have.
Q: All right. And will you agree, sir, that that videotape depicts a number of logical, goal-oriented behaviors on Mr. Wilson’s behalf?
A: Yes.
Q: So there’s no question that when he goes from the cooler, after Mr. Yost has been herded in there by Mr. Wilson and his accomplices, Mr. Wilson goes to the front checkout stand and assumes the role of clerk; right?
*1298A: Yes.
Q: Logical, goal-oriented behavior?
A: Yes.
Q: Okay. And when Mr. Wilson is conversing with the individuals, the various customers who came into the Quik Trip store that early morning time period, did he seem to have any trouble conversing with them, sir?
A: No.
Q: The tape you had had the audio portion of it; right?
A: Yes.
Q: He’s able to strike up conversations; right?
A: Yes.
Q: He’s able to help these individuals find whatever it was they supposedly were looking for; right?
A: Yes.
Q: Okay. And he was able to help move them along and get them on out; right?
A: Yes.
Q: And that’s logical, goal-oriented behavior under the circumstances; would you agree?
A: Yes.
Q: And the same is true with Mr. Wilson removing the safe from the cabinet; right?
A: Yes.
Q: And I could go on and on. Doctor, that’s a pretty good indication, is it not, of what Mr. Wilson was motivated to do there; right?
A: Yes.
Id., Vol. 2 at 127-29.
c. Consistency of Dr. Reynolds’s Diagnosis with the Test Results
The prosecutor also challenged Dr. Reynolds’s posttrial diagnosis. He first questioned whether it was consistent with the pretrial MCMI-III, which was valid:
Q: [I]sn’t it true the MCMI-III has several clinical scales that measures for the sort of thought disorders and delusional disorders that would be relevant to a diagnosis of schizophrenia?
A: Yes.
Q: [DJespite that, the MCMI-III did not suggest, as a likely diagnosis, paranoid schizophrenia, did it?
A: That’s correct.
Q: A diagnosis of paranoid personality disorder suggests the absence of paranoid schizophrenia; is that correct?
A: Well, it would be only the absence of the hallucinations and delusions.
Id. at 123.
The prosecutor then asked about the consistency of the diagnosis with the valid posttrial MMPI-2:
Q: [UJnder “Diagnostic Considerations,” where the report states, “Individuals with this MMPI-2 clinical profile are usually diagnosed as having a personality disorder, antisocial type.”
A: Uh-huh.
Q: Okay. And it goes on to say a lot of other things; right?
A: Correct.
Q: Basically bolstering the antisocial features of Mr. Wilson’s personality; is that fair to say?
A: Well, again, you have to understand that this is a hypothesis that you then correlate with collateral data. So the collateral data then supported the bizarre, unusual thinking, things — and it said that I need to take those into consideration, which I did, and therefore, as a clinician, I felt the diagnosis was more clearly of the paranoid schizophrenic than it was the antisocial. As I explained before, there are antisocial characteristics here, for sure.
Id. at 126-127 (emphasis added). (The record shows that the computer-scored report of the initial but invalid MMPI-2 had *1299actually been more supportive of Dr. Reynolds’s ultimate diagnosis. Its first sentence under “Diagnostic Considerations” stated: “Many individuals with this profile are considered to have severe Personality Disorders; however, the possibility of Schizophrenia, Paranoid type, or of a Bipolar Affective Disorder should also be considered.” Id., Vol. 1 pt. 5 at 912. The same section in the report of the valid test did not suggest consideration of schizophrenia, paranoid type.)
More dramatically, the prosecutor pointed out that the posttrial MMPI-2 emphatically supported a picture of Defendant that correlated closely with the psychopath image elicited at trial. He asked about a section of the interpretive report for the test that described Defendant as posing a continuing threat to society:
Q: MMPI-2, the second that you administered here, the valid one, sir, refers to the Megargee scale; right?
A: Yes.
Q: All right. And it classified Mr. Wilson as a type C offender, correct?
A: Yes.
Q: What does type C offender mean? [Discussion of where to find the section of the report]
A: It indicates that these are difficult and criminal offenders.
Q: It also goes on to say that, in fact, type C inmates typically have difficulty adjusting to prison life; right?
A: Yes.
Q: It may be necessary to segregate them from weaker or more vulnerable inmates; correct?
A: Yes.
Q: And when it refers to treatment considerations in the last paragraph on the page, the MMPI-2, the second one, the valid one that you generated, it says that, in fact, Mr. Wilson — or individuals with this profile type tend not to be very successful with treatment or rehabilitation programs; is that correct?
A: Yes.
Q: And it says it’s because they usually do not seek treatment on their own and have little motivation to alter their behavior. Okay. Well, could that extend to the taking of medication?
A: You mean whether they’ll take medication or not?
Q: Correct.
A: Well, medication is treatment, so it could.
THE COURT: And it says ... “Their suspicious attitudes and deep-seated hostility toward others make them a difficult case for rehabilitation. ” That would be counterproductive to a strategy of having an intelligent human being that could do well in a prison environment and be rehabilitated; correct?
THE WITNESS: Correct.
THE COURT: It’s inconsistent?
THE WITNESS: Yes.
Id., Vol. 2 at 129-131 (emphasis added).6
d. Factual Support for Dr. Reynolds’s Diagnosis
As previously discussed, the basis of Dr. Reynolds’s rejection of the diagnosis of *1300“personality disorder, antisocial type” (which was suggested by the valid posttrial MMPI-2 report) in favor of his diagnosis of paranoid schizophrenic disorder was his finding that Defendant suffered from delusions and auditory hallucinations. On cross-examination of Dr. Reynolds the prosecutor took aim at that finding. The finding relied in large part on statements by Defendant’s family and girlfriend.
First, the prosecutor asked whether Dr. Reynolds had spoken with any of the witnesses other than Defendant’s mother. Although Dr. Reynolds had testified on direct examination that he had conducted interviews with at least two others, he retreated from that assertion on cross-examination, saying that he may have just • relied on their affidavits. The exchange continued:
Q: Okay. And would it have been important to you to have sat down and spoken with these family members who signed these affidavits?
A: I felt like the information that I got with them on an affidavit under oath was adequate ... and that it was valuable.
Q: Okay. And who—
THE COURT: How do you know whether that information is true or not?
THE WITNESS: I don’t.... It’s not my job to investigate. I take them for what they tell me. But what I saw was consistency across the board, which gave me some sense that there was truthfulness to it.
THE COURT: Unless someone drafts the affidavits for them?
THE WITNESS: True.
Id. at 118.
Later the prosecutor asked about the specifics of the statements by the various witnesses. He began by questioning Dr. Reynolds about his reliance on three statements by Defendant’s girlfriend, Tonya Holt. The first was her statement that he had told her that his father was dead:
Q: Okay. And one of the instances from the affidavits that you refer to to support your diagnosis of paranoid schizophrenia is Tonya Holt’s affidavit, her statement that the defendant told her that his father was dead; right?
A: Yes.
Q: Do you remember that? Well, Doctor, do you recall what kind of a relationship Mr. Wilson had with his father?
A: It was rather distant and difficult.
Q: Okay. Basically a non-entity in his life?
A: Yes. Pretty much so.
Q: Do you recall, was his father, you know — was he a homeless person? Did he have a house? Do you recall what his social status was in life?
A: I believe that he had an issue with drugs and alcohol.
Q: Okay. So is it possible, sir, that Mr. Wilson told Tonya Holt that—
[Defense counsel]: Objection, Your Honor.
THE COURT: Let him ask the question first.
Q: Is it possible, sir, that Mr. Wilson told Ms. Holt that his father was dead because he was embarrassed by his father?
*1301[Defense counsel]: Objection, Your Honor; there’s no—
THE COURT: Calls for speculation?
[Defense counsel]: Exactly.
THE COURT: Sustained.
Q: Is there a possible nonpsychotic explanation for Mr. Wilson’s statement to Ms. Holt that his father was dead?
A: None that I have.
Q: None that you have that you received from the sources you read, or none that you’ve come up with?
A: I have no information to indicate that there was a different reason that he might make that statement.
Id. at 131-32.7 Dr. Reynolds apparently had forgotten what Holt had said in her statement: “Micheál never explained to me why he said his father was dead. I thought Micheál meant his father Oscar was not being a father.” Id., Vol. 1 pt. 5 at 854.
The second Holt statement about which the prosecutor questioned Dr. Reynolds was her statement that Defendant had said his name was Tom, not Micheál:
Q: You cite the fact that Mr. Wilson— and this is according to Ms. Holt’s affidavit — Mr. Wilson denied his name was Michael and introduced himself as Tom. You cite that as a basis to support your diagnosis of paranoid schizophrenia; correct?
A: That’s one of the delusions that I cited, yes.
Q: Okay. And you’ve reviewed the recording of Mr. Wilson’s statement to the police; correct?
A: Yes.
Q: He gives a confession?
A: Gave a confession, yes.
Q: Right?
And do you recall in that audio statement he tells the police, in response to questioning, that he used the name Tommy when he identified himself to the newspaper man who came into the store? Do you recall that?
A: I don’t. I don’t recall that.
Id., Vol. 2 at 132-33. The prosecutor did not refer to Holt’s statement in her affidavit that when Defendant used the name “Tom,” she thought “he was just playing.” Id., Vol. 1 pt. 5 at 855.
Third, the prosecutor asked Dr. Reynolds about Holt’s suggestion that Defendant had auditory hallucinations:
Q: All right. Well, you cite Mr. Wilson’s statement to Tonya Holt that he hears voices and you have to fight them away and pray them away as an example of psychotic behavior supporting your paranoid schizophrenia diagnosis; correct?
A: Correct.
Q: Okay. And in this situation, Mr. Wilson basically states that he can control the voices; is that right?
A: I believe that he tries to control the voices through prayer.
Q: Okay. Because he states you can pray them away if you just pray hard enough, fight hard enough; is that right?
A: Well, no, that’s not what he said. I don’t think he said that you can pray them away or if you pray hard enough, but that was an attempt which is very common for people with this disorder to make these voice [sic] goes away. They usually try to do whatever they can, and in this case he used religion to attempt to make the voices go away because they told him to do things.
Id., Vol. 2 at 133-34. Dr. Reynolds’s recollection was faulty. The affidavit of Holt in the record states that Defendant told her: “I hear voices & its OK. You just *1302have to fight them, you just have to pray them away and they will go away.”8 Id., Vol. 1, pt. 5 at 856.
Next, the prosecutor asked Dr. Reynolds whether his diagnosis was supported by the statement of Defendant’s sister that he had told school officials that his mother was white. Dr. Reynolds said that he did not recall how old Defendant was when he made the statements. The prosecutor reminded Dr. Reynolds that the sister said that it was when they were in school together. (The sister’s affidavit states that Defendant was “just a kid” at the time. Id. at 864.) He then continued:
Q: Okay. Sir, would it be unusual for paranoid schizophrenia to onset in a young child?
A: The onset usually begi — it’s not unusual. I mean, it does occur, it’s not common, but the onset is usually the late teens to the early forties, but it is not uncommon for it to occur in children.
Q: Okay. And Doctor, you’ve reviewed medical records from Children’s Medical Center in preparation for your trial testimony; is that correct?
A: That is correct.
Id., Vol. 2 at 134-35. Dr. Reynolds could not recall when Defendant had been seen at the center, but he was given records showing that Defendant was 16. Dr. Reynolds then acknowledged that the doctors at the center had found no indication of psychotic behavior, specifically reporting the absence of delusions or hallucinations.
Similarly, the prosecutor suggested that another supposed delusion cited by Dr. Reynolds was not indicative of mental illness:
Q: Okay. You cite Mr. Wilson’s belief, firm belief that he will be released from prison as an example of unusual behavior supporting your diagnosis of paranoid schizophrenia; is that right?
A: I believe that it was along the lines of not the diagnosis but that was a delusional comment.
Q: Okay. Well, Doctor, have you had any experience with inmates in the correctional system in the course of your career as a psychologist?
A: Yes.
Q: And are there a lot of inmates who, in fact, may have an irrational belief that they’re going to be released no matter how strong the evidence is against them?
A: I think your word was correct, irrational.
Q: Are they all psychotic?
A: Of course not.
[Defense counsel]: Objection, Your Honor; irrelevant.
THE COURT: Sustained.
Id., at 136.9
*1303The prosecutor also challenged Dr. Reynolds’s statement in his affidavit that Defendant may have been out of touch with reality at the time of trial. Dr. Reynolds admitted that he had not interviewed the trial attorneys on the matter.
Later the prosecutor inquired whether Defendant’s alleged paranoia might be rational, rather than delusional:
Q: All right. In your report, you cite the fact that the defendant believes he’s being plotted against as evidence supporting a diagnosis of paranoid schizophrenia; correct?
A: Yes.
Q: Is it your understanding, sir, that Mr. Wilson was involved in gang life?
A: Yes.
Q: Both before — well, at least before he went to jail; right?
A: Yes.
Q: And do you recall, sir, that, in fact, he was having gang trouble even in jail?
A: I understand there was some fights in jail. I didn’t know if it was gang-related or not but I know there were some altercations.
Q: All right. I mean, gang life, the sort of gang life Mr. Wilson had, is that a possible nonpsychotic explanation for his belief that he was being plotted against?
A: Well, it would certainly fit into gang life, that is correct. And so, therefore, that would be consistent with the development of a paranoid disorder, whether how accurate it was that he was being plotted against or not, but he would probably have, most likely with this disorder, an exaggerated view of how he is out to be gotten or killed. But there had some relative — there was some relevance to the idea that he was definitely being plotted against, yes.
Q: I mean, you recall the situation where his mother’s house was burned down; right?
A: Yes.
Id. at 138-39.
We also note a matter not explored during the hearing. Present in the record submitted to the district court was a telling discrepancy between Defendant’s responses on the pretrial invalid MMPI-2 and the posttrial valid MMPI-2. In the pretrial MMPI-2, under the “Critical Items” section for “Mental Confusion,” the interpretive report indicated that Defendant had answered “True” when asked to respond to the statement “I often hear voices without knowing where they come from.” Id., Vol. 1 pt. 5 at 922. (Dr. Reynolds’s handwritten notes on the pretrial report reflect that Defendant told him that when he was in jail he heard two voices telling him to hang himself. Dr. Reynolds did not recall when he wrote the notes.) But on the posttrial MMPI-2 under the same Critical Items section, no such response is noted, meaning that Defendant apparently did not answer this item “True” during the re-test. When administering the re-test, Dr. Reynolds had instructed Defendant to ask for clarification or assistance if he had difficulty understanding any of the questions.10
This discrepancy may partly account for the fact that while the invalid MMPI-2 suggested that “Schizophrenia, Paranoid type” should be considered as a possible diagnosis, id. at 912, the valid MMPI-2 *1304stated that “individuals with this MMPI2 clinical profile are usually diagnosed as having a Personality Disorder, Antisocial type,” id. at 1017, and did not mention a possible diagnosis of schizophrenia. Dr. Reynolds testified that it was collateral data from the other witnesses’ affidavits and the additional records furnished him by appellate counsel that enabled him to arrive at the schizophrenia diagnosis.
3. Redirect Examination of Dr. Reynolds
On redirect examination Dr. Reynolds reiterated his belief that he reasonably relied on the witness affidavits in forming his clinical opinion, and that he had not needed to interview the witnesses to obtain adequate information. He clarified that although the valid MMPI-2 had contained indications of antisocial behavior, he thought that those patterns were “part of a paranoid schizophrenic process.” Id., Vol. 2 at 146. He testified that the actions of Type C offenders can be controlled in a prison. And he stated that the absence of delusions and hallucinations reported in Defendant’s medical records from age 16 did not negate the schizophrenia diagnosis, because the onset of the disorder is typically later in life.
II. DISCUSSION
A. Standard of Review
Generally, when an applicant seeks relief from a state court conviction or sentence under 28 U.S.C. § 2254, we must apply the deferential standard of review established by the Antiterrorism and Effective Death Penalty Act (AEDPA). If the state court adjudicated the federal claim that the applicant now presses on the merits,
we may only grant federal habeas relief if the habeas petitioner can establish that the state court decision “was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States,” or “was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.”
Saiz v. Ortiz, 392 F.3d 1166, 1175 (10th Cir.2004) (quoting 28 U.S.C. § 2254(d)(1)).
On direct appeal from his conviction and sentence, Defendant pressed his ineffective-assistance claim before the OCCA, supported by affidavits from Dr. Reynolds and his family and girlfriend. Defendant requested an evidentiary hearing on the claim under Rule 3.11, Rules of the Oklahoma Court of Criminal Appeals, Title 22, Ch. 18, App. (1995). The OCCA denied relief on the claim and Defendant’s request for an evidentiary hearing. See Wilson I, 983 P.2d at 471-72. Although noting Defendant’s “attempt to supplement the record with material not found in the record,” id. at 472, the OCCA held that on the basis of its “review of the trial record,” Defendant had failed to demonstrate either deficient performance or prejudice on the mental-health issue, id.
On Defendant’s prior appeal we held that AEDPA deference was not required because the OCCA’s denial of Defendant’s request for an evidentiary hearing was not an adjudication “on the merits” within the meaning of 28 U.S.C. § 2254(d). We noted (1) the OCCA’s apparent failure to consider Defendant’s nonrecord evidence in denying his request for an evidentiary hearing, and (2) the higher threshold for triggering an evidentiary hearing under the OCCA’s Rule 3.11 than permitted under the federal standard stated in Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). See Wilson IV, 577 F.3d at 1290-1300.
The State argues that we should “restore” AEDPA deference to the OCCA’s *1305decision in Wilson I, because the later decision of the OCCA in Simpson v. State, 230 P.3d 888, 905-06 (Okla.Crim.App. 2010), shows that we misinterpreted OMahoma’s threshold for an evidentiary hearing. Recently, this court has adopted the State’s view. See Lott v. Trammell, 705 F.3d 1167, 1211-13 (10th Cir.2013). But we need not decide whether the OCCA’s decision on the record before it requires that we affirm under AEDPA, because we deny relief based on the evidence that Defendant asks us to consider.
In evaluating Defendant’s claim that his Sixth Amendment rights were violated by ineffective assistance of counsel at the sentencing proceedings, we apply the standard laid out in Strickland, 466 U.S. at 687, 104 S.Ct. 2052:
A convicted defendant’s claim that counsel’s assistance was so defective as to require reversal of a conviction or death sentence has two components. First, the defendant must show that counsel’s performance was deficient. This requires showing that counsel made errors so serious that counsel was not functioning as the ‘counsel’ guaranteed the defendant by the Sixth Amendment. Second, the defendant must show that the deficient performance prejudiced the defense. This requires showing that counsel’s errors were so serious as to deprive the defendant of a fair trial, a trial whose result is reliable.
In conducting this two-part inquiry, we need not address the performance and prejudice prongs in that order. See id. at 697, 104 S.Ct. 2052 (Because defendants must establish both components to prevail, “a court need not determine whether counsel’s performance was deficient before examining the prejudice suffered by the defendant as a result of the alleged deficiencies.”). Here, we need address only prejudice.
“An error by counsel, even if professionally unreasonable, does not warrant setting aside the judgment of a criminal proceeding if the error had no effect on the judgment.” Id. at 691, 104 S.Ct. 2052. In a challenge to a capital sentence, “the question is whether there is a reasonable probability that, absent the errors, the sentencer ... would have concluded that the balance of aggravating and mitigating circumstances did not warrant death.” Id. at 695, 104 S.Ct. 2052. We “must ask if the defendant has met the burden of showing that the decision reached would reasonably likely have been different absent the errors.” Id. at 696, 104 S.Ct. 2052.
To resolve whether there was prejudice, we do not consider omitted mitigation evidence in a vacuum. In Wong v. Belmontes, 558 U.S. 15, 130 S.Ct. 383, 175 L.Ed.2d 328 (2009), the defendant was sentenced to death after his counsel presented evidence of his difficult family background during the sentencing hearing but declined to introduce certain additional mitigating evidence (including expert testimony to explain his behavior, see id. at 388-89, and evidence of a serious illness that caused “emotional instability, impulsivity, and impairment of the neurophysiological mechanisms for planning and reasoning,” id. at 389 (internal quotation marks omitted)) out of concern that it would open the door for the prosecution to introduce evidence of a prior murder committed by the defendant, see id. at 385-86. Counsel’s concern that this extremely harmful evidence would be admitted if he introduced certain mitigating evidence was supported by the trial court’s warning that it would allow the prosecution to put on the evidence if the defense presented substantial evidence of the defendant’s nonviolent character. See id. at 386. Rejecting the defendant’s ineffective-assistance-of-counsel claim for failure to demonstrate prejudice, the Supreme *1306Court held that in assessing prejudice a reviewing court must “consider all the relevant evidence that the jury would have had before it if [the defendant] had pursued the different path — not just the mitigation evidence [the defendant] could have presented, but also the [evidence of the other murder] that almost certainly would have come in with it.” Id. Although some of the omitted mitigation evidence might have helped the defendant, “the worst kind of bad evidence would have come in with the good.” Id. at 390. Hence, the defendant had failed to show prejudice. Following the holding of Belmontes, we must consider not just the mitigation evidence that Defendant claims was wrongfully omitted, but also what the prosecution’s response to that evidence would have been.
In reviewing the district court’s conclusion that Defendant was not prejudiced by his counsel’s alleged deficiencies, “we accept the district court’s underlying factual findings unless clearly erroneous, and we review de novo ... whether any deficiencies prejudiced the defendant.” United States v. Rodriguez-Rivera, 518 F.3d 1208, 1216 (10th Cir.2008) (internal quotation marks omitted). The court ruled that introduction of the additional mental-health evidence submitted by Defendant at the court’s hearing on remand might well have been a double-edged sword, “supporting] the prosecution’s portrayal of [Defendant] as a dangerous and continuing threat to society.” Wilson V, 2011 WL 744661, at *26. Because of the “distinct possibility that additional mental health evidence might have been counterproductive and harmed Defendant’s mitigation case,” the district court could not “conclude that additional evidence of [Defendant’s] mental health problems would have affected the jury’s imposition of the death penalty.” Id. We agree with the district court that Defendant has not shown that trial counsel’s alleged deficiencies prejudiced him.
1. The Failure to Call Additional Family Witnesses
We first address Defendant’s argument that the witnesses who provided affidavits for his appeal, including his sister, brother, and girlfriend, should have been called during the mitigation case and that their testimony may have convinced jurors to vote against death.
We look to the posttrial affidavits as proffers of how these witnesses would have testified. These affidavits assert that Defendant was affected by gang violence from his youth, stating that he sustained a gunshot injury in a drive-by shooting; that he was frequently targeted by rival gang members, who shot at him, vandalized his car, and burned the house he shared with his mother; and that he was attacked by gang members, including members of his own gang, while incarcerated. But the jury heard about the drive-by shooting and arson incident during Dr. Reynolds’s testimony on Defendant’s social history. Although Dr. Reynolds did not speak from first-hand knowledge about either incident, neither, so far as the record shows, could any of the omitted witnesses. The additional matter would not have qualitatively altered the picture of Defendant’s exposure to violence and could have merely emphasized his gang involvement.
The omitted witnesses might also have testified to some of Defendant’s unusual behaviors and thought processes, which Dr. Reynolds thought indicative of a mental disorder. The value of this evidence, however, depends on the value of Dr. Reynolds’s use of it, a matter we address in the next section of this opinion. Suffice it to say now that Defendant gives no reason to suppose that the evidence would have been more effective if presented by the witnesses in person (and subject to cross-examination), rather than through Dr. Reynolds’s testimony.
*13072. Deficiencies in Dr. Reynolds’s Trial Preparation and Testimony
Defendant’s chief argument regarding prejudice relates to Dr. Reynolds’s testimony. According to Defendant, had Dr. Reynolds been given sufficient time and information to conduct all necessary interviews and psychological tests properly, he would have arrived at his diagnosis of schizophrenic paranoid personality disorder before trial and, if properly prepared by counsel, could have presented that diagnosis to the jury and explained how Defendant’s background contributed to his mental illness, demonstrated that Defendant’s mental illness affected his behavior and caused him to have difficulty conforming his conduct to the law, and helped the jury understand the motivation for Defendant’s actions, thereby calling into question Defendant’s moral culpability. Defendant contends that conveying this information would have caused at least one juror to believe that Defendant was not deserving of capital punishment. And he asserts that had counsel performed properly, the prosecutor would not have been able to elicit testimony from Dr. Reynolds that supported the prosecutor’s repeated characterizations of Defendant as a psychopath.
In evaluating these claims of prejudice, we look to the testimony at the evidentiary hearing, together with the exhibits offered in evidence at that hearing, to see what likely would been presented at trial if Defendant’s counsel had done what he contends they should have. Our review indicates that Defendant would have been no better off with the evidence presented at the hearing, and in significant ways would have been worse off.
We start with the MMPI-2. Since his direct appeal, Defendant has laid great stress upon the failure to obtain a valid MMPI-2 before trial. Dr. Reynolds’s affidavit submitted to the OCCA in support of an evidentiary hearing states: “There was some evidence for a diagnosis of schizophrenia but because his first MMPI-2 was invalid, I needed additional testing, and further collateral data to support this diagnosis. Unfortunately, there wasn’t enough time to obtain this information before the trial.” R., Vol. 1 pt. 5 at 845.
But the evidence from the hearing demonstrates that having Defendant retake the test did not help his case. The valid test contradicted more than it supported Dr. Reynolds’s transition to a diagnosis of paranoid schizophrenic disorder. The interpretive report for the invalid test stated: “Many individuals with this profile are considered to have severe Personality Disorders; however, the possibility of Schizophrenia, Paranoid type, or of a Bipolar Affective Disorder should also be considered.” Id. at 912 (emphasis added). But the valid test’s interpretive report said: “Individuals with this MMPI-2 clinical profile are usually diagnosed as having a Personality Disorder, Antisocial type.” Id. at 1017. It did not suggest that schizophrenia be considered as a possible diagnosis.
Worse for Defendant, the interpretive report of the valid MMPI-2 emphatically bolstered the prosecution’s position that Defendant was a continuing threat to society. The invalid test’s interpretive report said that Defendant “fits equally well into more than one classification according to the Megargee classification,” which was described as a “system for classifying male criminal offenders.” Id. at 912. The report did not say which classifications Defendant fit within, nor did it describe the characteristics of those classifications. The valid test’s interpretive report, however, painted Defendant as being the most despicable type of criminal. It said:
[Defendant’s] profile matches those of Type C offenders in the Megargee typol*1308ogy. Individuals matching this profile type are among the 'most difficult criminal offenders. They are often viewed as distrustful, cold, irresponsible, and unstable. They tend to have antisocial, aggressive, and hostile attitudes toward others. They engage in violent crimes against other people and usually have an extensive criminal record. They tend to come from deviant and stressful home environments and typically have a great deal of difficulty adjusting to society. They are viewed by others as alienated, bitter, rigid, and dogmatic. Their interpersonal relationships are quite disruptive; their suspicious attitudes and deep-seated hostility toward others make them a difficult case for rehabilitation. Research has supported the view that Type C inmates typically have difficulty adjusting to'prison life. It may be necessary to segregate them from weaker or more vulnerable inmates during incarceration.
Id. at 1017 (emphasis added). This is a roadmap for one seeking to portray Defendant as a dangerous criminal who could not safely be allowed to live.
When confronted at the evidentiary hearing with these statements contradicting his diagnosis of Defendant, Dr. Reynolds stated that the test’s suggested diagnoses merely provided “a hypothesis that you then correlate with collateral data.” Id., Vol. 2 at 126. But the value of the collateral data on which he relied — the statements by members of Defendant’s family and his girlfriend that Dr. Reynolds found to be evidence of Defendant’s delusions and hallucinations — was undermined by the cross-examination of Dr. Reynolds at the § 2254 evidentiary hearing. Indeed, Dr. Reynolds’s insistence that the statements provided clear support for the diagnosis suggested sloppy analysis and bias, raising serious question about his credibility.
We note a few examples. First, as evidence of Defendant’s delusions, Dr. Reynolds referred to the statement by Defendant’s sister that years earlier, when she and Defendant had attended the same school, Defendant had told school officials that the two of them had different mothers and that his mother was white. But the statement was made when Defendant was “just a kid,” id., Vol. 1 pt. 5 at 864, and a later mental-health evaluation of Defendant administered by Children’s Medical Center when he was 16 reported no delusions or hallucinations. After being confronted by this information on cross-examination, Dr. Reynolds tried to explain away the mental-health evaluation during his redirect examination by noting that “usually the onset [of paranoid schizophrenia] is late teens to the early forties.” Id., Vol. 2 at 146. The explanation, however, merely emphasizes the irrelevance of the “white mother” incident during Defendant’s childhood.
Second, Dr. Reynolds said that the presence of delusions and hallucinations was supported by the untrue statement of Defendant to his girlfriend, Tonya Holt, that his father was dead. Dr. Reynolds testified on cross-examination that he had no “possible nonpsychotic explanation” for Defendant’s statement, id. at 132, even after he had conceded that Defendant’s father was a man with drug and alcohol issues who was “pretty much” a non-entity in Defendant’s life, id. at 131, and the prosecutor had suggested that Defendant was simply embarrassed by his father. Most remarkably, the very statement by Holt relied on by Dr. Reynolds continued: “[Defendant] never explained to me why he said his father was dead. I thought [Defendant] meant his father Oscar was not being a father.” Id., Vol. 1 pt. 5 at 854.
*1309Third, another statement from the Holt affidavit that Dr. Reynolds cited in support of the presence of delusions and hallucinations described Defendant’s habit of sometimes introducing himself by a false name, Thomas, and adopting different mannerisms. The prosecutor pointed out that Defendant had told police (in his murder confession) that he had used the name “Tommy” during the robbery in order to conceal his identity. Id., Vol. 2 at 133. Dr. Reynolds did not recall this information. He also apparently ignored Holt’s statement that “[she] always thought it was strange [that Defendant used the name “Tom”], but [she] also thought that he was just playing.” Id., Vol. 1 pt. 5 at 855.
The best of this evidence of hallucinations was Holt’s statement that Defendant had told her that he heard voices. But even on this point Dr. Reynolds misremembered facts that may have been relevant to his diagnosis. At the evidentiary hearing the prosecutor asked whether “[Defendant] state[d] you can pray [the voices] away if you just pray hard enough, fight hard enough.” Id., Vol. 2 at 133. Dr. Reynolds responded: “Well, no, that’s not what he said. I don’t think he said that you can pray them away or if you pray hard enough, but that was an attempt which is very common for people with this disorder to make these voice goes [sic] away.” Id. at 133-34. The prosecutor was not wrong. According to Holt, “[Defendant] told me, ‘I’ve heard them too, I hear voices & its OK. You just have to fight them, you just have to pray them away and they will go away.’ ” Id., Vol. 1 pt. 5 at 856.
Another point on which Dr. Reynolds looked foolish was his finding that Defendant suffered from paranoia. He said in his affidavit: “[Defendant’s] delusions appear to be of a grandiose and paranoid type. He believes ... that he is being plotted against.” Id. at 848. But Defendant was a gang member, and a rival gang had shot him and set his mother’s home ablaze. Dr. Reynolds’s struggle to construct a finding of delusional paranoia can be illustrated by the statement in his affidavit that Defendant’s sister “described him as very suspicious and paranoid,” id. at 847, when her full statement was:
I am familiar with the gang environment that [Defendant] grew up in and I know it is no exaggeration to say that he was shot at on a daily basis. I believe [Defendant] was being pulled into the gang scene by at least his ninth grade year of high school. I remember [Defendant] got real suspicious and paranoid after he joined the gang. He knew it was a game of survival and he was always checking to see if someone was in the house or near the car.
Id. at 865 (emphasis added).
In sum, Defendant has utterly failed to show that additional testing and interviews would have produced a plausible diagnosis of paranoid schizophrenic disorder. Nor has Defendant shown that better preparation of Dr. Reynolds could have eliminated or weakened the prosecutor’s success at trial in characterizing Defendant as a psychopath. The term is used in the DSM-IV. It was even used by Dr. Reynolds in his draft affidavit for use in the appeal to the OCCA, before it was deleted in the editing process. And little would have been gained by prohibiting use of the term because the description in the valid MMPI-2 of the Defendant’s profile — a Type C offender in the Megargee typology — explicitly describes the vision of evil evoked by the word psychopath.
Defendant’s evidence from the hearing also failed to establish any connection between Dr. Reynolds’s diagnosis of Defendant and his commission of the murder. Although Dr. Reynolds testified that *1310knowing that Defendant suffered from schizophrenic paranoid personality disorder might help a jury understand Defendant’s “motivation” for committing the crime, id, Vol. 2 at 72, there was no credible evidence that Defendant acted as a result of delusions or hallucinations. Dr. Reynolds suggested in his affidavit: “It is possible that [Defendant] could have been delusional at the time of the crime; for example, he went to his own place of employment knowing he would be observed by his coworker, and customers. His thinking must have been delusional to believe that he would not be easily identified.” Id., Vol. 1 pt. 5 at 848. But this ignores the planning for the crime, the efforts to allay any suspicion by customers who entered before the perpetrators could remove the safe, and the murder of the one witness who had known Defendant previously.
Finally, insofar as Defendant contends that his counsel were ineffective in not even presenting adequately what Dr. Reynolds had determined from the information obtained before his trial testimony, Defendant has not satisfied his burden of showing what more Dr. Reynolds could usefully have said. He argues that the jury should have been presented “a complete picture of [Defendant’s] mental health,” Aplt. Br. at 40; but he does not describe what that picture was. At most, from what we can tell from the record, Dr. Reynolds had several possible diagnoses stated in the interpretive report on the valid MCMI-III test; but we have no way of assessing how that would have played out at trial, particularly when Dr. Reynolds has never stated that he had adopted one of those diagnoses at the time of his trial testimony. Defendant also states that Dr. Reynolds could have testified at trial that those with mental disease “have less capability of controlling their behavior in regards to the law.” Id. at 38 (quoting R. Vol. 2 at 88). Such a vague, generalized statement, however, is not the sort of mental-health evidence that courts rely on to show prejudice; indeed, this statement seems as probative of future dangerousness as of diminished moral culpability.
We are not denigrating the value of mental-health evidence to support mitigation in a capital case, nor are we departing from our precedent addressing such evidence. We do not dispute that mental retardation and organic brain damage are well-recognized grounds for mitigation. See Hooks v. Workman, 689 F.3d 1148, 1205 (10th Cir.2012) (organic brain damage); Anderson v. Sirmons, 476 F.3d 1131, 1144 (10th Cir.2007) (brain damage) Smith v. Mullin, 379 F.3d 919, 941-42 (10th Cir. 2004) (mental retardation and brain damage); see also Atkins v. Virginia, 536 U.S. 304, 306, 122 S.Ct. 2242, 153 L.Ed.2d 335 (2002) (prohibiting execution of the mentally retarded, partly because they “do not act with the level of moral culpability that characterizes the most serious adult criminal conduct”). And if (1) the record had supported a plausible diagnosis of paranoid schizophrenia and (2) Defendant’s behavior during the crime could have been tied to that disorder, this would be a different case.
We conclude that Defendant has not satisfied his burden of showing that the jury’s sentence of death “would reasonably likely have been different absent the errors” of his counsel. Strickland, 466 U.S. at 696, 104 S.Ct. 2052. Accordingly, we must deny relief.
III. CONCLUSION
We AFFIRM the district court’s denial of Defendant’s claims under § 2254.
. Although he has been referred to as "Michael Lee Wilson” throughout this case, we are informed and the record reflects that the correct spelling of Mr. Wilson's first name is "Micheál.”
. We thank counsel for both Defendant and the State for their skillful advocacy in this appeal.
. Mr. Robertson had little specific recollection of his representation of Defendant. He did not recall being told by Dr. Reynolds that more time was needed to test Defendant, or that the results of the MMPI-2 personality test were invalid. He suggested that the late hiring of Dr. Reynolds was due to budgetary constraints and to a desire not to turn over Dr. Reynolds's raw data to the State until absolutely necessary.
Mr. Hudson testified that he did not recall any coordination of strategy between the lay witnesses (whom he handled) and the expert testimony. He also testified that during his many interactions with Defendant, he had never noticed any signs or symptoms of mental illness.
. Dr. Reynolds believed that he had said this to defense counsel during a meeting on February 17; but it appears that he probably conveyed this information at least a week earlier, because defense counsel informed the court on February 10 (the day before trial testimony began) that the test was unreliable.
. Dr. Reynolds later described these disorders as follows: A person with a narcissistic personality disorder believes the world revolves around him and may exploit others; a passive-aggressive personality trait manifests by redirecting hostile or angry feelings from their source to another person or object; persons with schizotypal personality features are not really schizophrenic but are socially awkward and tend to withdraw from interactions with others; generalized anxiety disorder involves physiological and psychological symptoms of anxiety unconnected to any obvious stressor; bipolar disorder is characterized by mood swings, sometimes dramatic, with the subject cycling from manic energy and carefreeness to deep depression; and posttraumatic stress disorder involves hypervigilance, depression, or irritability after exposure to a dangerous situation. See R., Vol. 2 at 82, 83, 84, 86, 87.
. The full description of Type C offenders from the interpretive report is chilling:
This client’s profile matches those of Type C offenders in the Megargee typology. Individuals matching this profile type are among the most difficult criminal offenders. They are often viewed as distrustful, cold, irresponsible, and unstable. They tend to have antisocial, aggressive, and hostile attitudes toward others. They engage in violent crimes against other people and usually have an extensive criminal record. They tend to come from deviant and stressful home environments and typically have a great deal of difficulty adjusting to society. They are viewed by others as alienated, bitter, rigid, and dogmatic. Their interpersonal relationships are quite disruptive; *1300their suspicious attitudes and deep-seated hostility toward others make them a difficult case for rehabilitation. Research has supported the view that Type C inmates typically have difficulty adjusting to prison life. It may be necessary to segregate them from weaker or more vulnerable inmates during incarceration.
R., Vol. 1 pt. 5 at 1017.
. Although the objection to the question about embarrassment was sustained as speculative, the question later became relevant to probing Dr. Reynolds’ bias.
. The entirely of the paragraph from Holt’s affidavit stated:
After I was raped I had some severe mental problems that finally forced me to seek counseling. I now know that I have been diagnosed with Post Traumatic Stress Disorder (PTSD). I started worrying when I would hear things — voices. I told Micheál about the voices and he told me, 'its O.K. to hear voices, I hear voices.' I said, no Micheál this is not like hearing your conscience tell you not to do bad things, they are telling me things. Micheál told me, 'I've heard them too, I hear voices & its OK. You just have to fight them, you just have to pray them away and they will go away.’ That is when I started thinking he is really crazy, he has mental problems. I asked him what kind of voices do you hear? He told me he didn't want to talk about it and he wouldn’t tell me anymore. I have sought help and my doctor prescribed Zoloft and later on Xanax. I am not on any medication now, and I do not hear the voices anymore.
R., Vol. 1 pt. 5 at 856.
. Although the objection was sustained, the prosecutor could have elicited the same point through a properly phrased question.
. Defendant’s responses on the "Mental Confusion” section were otherwise nearly identical. On the second MMPI-2, Defendant responded "True” to the statement “I find it hard to keep my mind on a task or job.” R., Vol. 1 pt. 5 at 1026. On the first MMPI-2, no response to this statement is indicated in the “Mental Confusion” section. Id. at 921-22. Other than the different responses to that statement and the statement about voices, Defendant’s responses in the Mental Confusion section were the same on both reports.