OPINION
delivered the opinion of the Court
in which PRICE, WOMACK, JOHNSON, and HOLCOMB, JJ., joined.
Appellant was originally convicted in 1990 of capital murder for the shooting deaths of his wife’s mother, father, and brother. Based upon the jury’s answers to the special punishment issues, the trial judge sentenced him to death. This Court upheld his conviction and sentence on direct appeal.1 In 2007, the Fifth Circuit Court of Appeals granted habeas relief and remanded the case for a new trial on punishment.2 On retrial in 2008, a second jury sentenced appellant to death. Appellant raises twenty-five points of error. Finding no reversible error, we affirm the judgment and sentence.
Factual Background
Karen Vicha was appellant’s third wife. They were married in July 1988 and lived in a house down the road from her brother and across the street from her parents. Appellant was almost forty years old. The marriage quickly disintegrated,3 and, after *262a year, Karen told appellant to move out. She wanted a divorce. Appellant attempted to talk her out of this decision and would randomly call her and show up at her work place.
Appellant then kidnapped Karen as a further effort to dissuade her from divorcing him. He hid in the trunk of her car while she was at a bar one evening with a girlfriend. When Karen started to drive home, appellant folded down the back seat and “popped out of the trunk with a knife.” He jumped over the console, halfway into the front seat, and stuck the knife against Karen’s ribs. He told her to keep driving until they came to a field. Karen stopped the car, and appellant said that he if couldn’t have her, then no one else could. He pulled out a roll of black electrical tape, but Karen kept talking, and, after about two hours, she convinced him that she would reconsider the divorce issue. He let her go, and she called her brother, Bobby, who was a police officer. Bobby told Karen to report the kidnapping.
After he arrested appellant for kidnapping Karen, Officer James Head looked in his patrol-car mirror and saw appellant staring at him with a look that “made the hair on the back of [his] head stand up.” He got “the heebie-jeebies.” Appellant muttered something like “They’re going to be sorry.” Officer Head called Karen’s brother, Bobby, and warned him about appellant. When appellant was released on bail for the kidnapping charge, Bobby got Karen a German shepherd for proteetion. A few days later, appellant told Karen, “Oh, I see you — you’ve got a dog now.... [TJhat’s a big mean dog you’ve got.” Shortly thereafter, Karen found the dog lying dead in front of her house.
Nine days after he had kidnapped Karen, appellant went to her house in the early afternoon. As Karen’s three daughters each came home from school along with Bobby’s son,4 appellant handcuffed them, tied up their feet, and taped their mouths closed. Karen’s oldest daughter testified that she heard appellant cut the telephone lines. Then he left to ambush and shoot Karen’s father, mother, and brother Bobby as each of them came home.5
Appellant returned to Karen’s house after the triple killings and waited for his wife to come home from work. He told the children, “I wish I had blown you away like I intended to.” When Karen arrived, appellant came out of one of the bedrooms with a gun. Appellant said, “Karen, I’ve killed your momma and your daddy and your brother, and they are all dead, and nobody is going to come help you now.” She didn’t believe him, so appellant showed her Bobby’s gun lying on the kitchen table and pulled the curtains so she could see her father’s truck parked behind the house. He showed her $1,000 in cash that he had taken from her mother. Appellant told Karen that she was lucky that he hadn’t molested her daughters, and he told her to kiss them good-bye. She *263did. He made her put on handcuffs. Karen talked appellant into leaving the house and taking her with him.6 He said he was going to take her away for a few weeks and torture her.7
As appellant drove, Karen tried to escape by freeing one hand from the handcuffs and grabbing at the steering wheel, making the car swerve into a ditch. She grabbed one of appellant’s guns, pointed it at his stomach, and pulled the trigger, but nothing happened. Then Karen and appellant fought over the gun, with appellant repeatedly pulling the trigger, but still the gun did not fire. Appellant pistol-whipped Karen until she couldn’t see for all of the blood on her face. A woman passerby started shouting at appellant, “[W]hat are you trying to do to that woman,” so appellant drove the car out of the ditch as Karen lay in the passenger seat. He shouted at her that if she got blood on his clothes, he would kill her. But he was also rubbing her between her legs as he drove. He told her that his reputation was ruined because she had had him arrested and his name was in the papers.
He drove to a deserted field in Bosque County where he threatened to rape her. After dark, he drove out of the field, but they passed a sheriffs patrol car which turned around to follow them. Appellant grabbed a knife and started stabbing Karen’s chin, forehead, and nose, as he was driving. Appellant said that he did not want to die in prison, so he “floored it” and rammed into a parked car. After the crash, appellant turned to Karen and said, “I guess now you’ll get a new car.” Both appellant and Karen were injured in the crash. Officers had to cut the car door open to get Karen out. Appellant was found with Karen’s father’s watch and wallet, as well as .37 and .38 caliber revolvers.
Although appellant was forty years old when he committed this triple murder, the State’s evidence showed that he was no stranger to violence. He had a long history of brutalizing and molesting women. Appellant beat both of his former wives and molested several young girls, including relatives.
His first wife, Pam Woolley, testified that they were married in 1970 when appellant was twenty-two. They had two children, but their marriage started downhill after two years. By 1974, appellant had become violent, and he used to beat her on the head so that her hair would hide the marks.8 Pam said that appellant could go from normal to extremely angry in a split second, and he always blamed her for his violent acts. Appellant told her that if she ever filed for divorce, he would “fix her” so no other man would look at her again.
During this ten-year marriage, appellant molested Pam’s younger sister and *264punched her on the mouth, “busting” her lip. He molested his children’s thirteen-year-old babysitter while teaching her how to water ski. He groped the breast of another neighborhood girl. In 1979, when appellant was thirty, he raped his cousin who was about fifteen at the time. When appellant’s niece was fifteen, he grabbed her ankles as she sat in a chair wearing a nightgown, spread open her legs, and gestured with his tongue as if he were performing oral sex on her. Later that same day, he forcibly kissed her and then threw her a $5 bill.
Appellant married Candy Ryan, his second wife, when he was thirty-five and she was eighteen. After one year of marriage, appellant started physically abusing her. He regularly hit her on the head. Once he grabbed her by the hair and repeatedly hit her against the cabinet and floor. After she dared to throw something at him, he hit her with a sledge hammer. Candy said that appellant had a “switch-type” personality — changing from sweet to nasty in a split-second. He stalked her, both during and after their marriage. He would sit in his car outside the gas station where Candy worked, and, if a customer stayed inside too long, appellant came in and gave the customer an intimidating look. After Candy left appellant, he would call her late at night and tell her where she had been, whom she had been with, and what she had been doing. Appellant threatened Candy’s father when he tried to help Candy leave.
Appellant’s childhood did not augur well for his future. His earliest years were spent in the custody of an alcoholic stepfather who worked only periodically and a sickly, withdrawn, and depressed mother. When appellant was four, his mother was institutionalized in the Austin State Hospital with a psychoneurotic disorder. Appellant, his brother, and his older sister were sent to the Corsicana State Home for Children. Because of her promiscuous acting-out, appellant’s older sister was sent to a convent school, and his problematic older brother was placed under the supervision of the Waco Probation Department. Appellant remained at the Home for twelve years.
When appellant was fifteen, a psychiatrist, Dr. Hodges, evaluated him and concluded that he was paranoid, distant, and impulsive; he showed poor self-control, displayed hostility to women, and blamed others for his own bad conduct. Dr. Hodges’s impression was that appellant “represented] a sociopathic personality disturbance of the dissocial type.” People with this diagnosis gratify their own desires without regard for the cost to others. Appellant’s “long term prognosis [did] not look good.”
At age seventeen, appellant joined the Marine Corps and was sent to Vietnam. Although he received an honorable discharge after his four-year tour of duty, he was not recommended for re-enlistment because of a series of violations and convictions. He married his first wife shortly after he left the Marines.
Dr. Richard Coons, a psychiatrist, had testified at appellant’s 1990 trial that he would be a future danger. Dr. Coons testified at the 2008 retrial that appellant would still be a future danger even though appellant did not have a single disciplinary report for the eighteen years that he had been on death row. Dr. Coons explained this discrepancy by stating that all those on death row have an incentive to behave because their convictions are on appeal, and thus they are less violent than they would be in the general prison population.
Appellant called several witnesses to attest to his prison reformation and lack of violence for the entire time that he had been on death row. According to one fel*265low inmate, appellant was well liked by everyone; he was always even-tempered and had the ability to “talk sense” into some of the more violent inmates. He said that appellant had organized a sports league at the Ellis Unit and that he helped inmates write letters and would read them their letters from family members. After Death Row was moved to the Polunsky Unit, appellant’s behavior was the same; he was always helpful and upbeat.
Another inmate testified that appellant would take people “under his wing” and help the “agitated” ones. He stated that, while at the Ellis Unit, appellant was an SSI, which was like a trustee, and would often walk around with female officers. A third inmate testified that appellant was generous and gave commissary items to other inmates. A fourth inmate said that appellant helped him to learn English and to file a federal habeas petition. Appellant helped mentally-retarded inmates and was known for his respect for the law and God.
Appellant’s older sister testified about their childhood and how appellant changed for the worse after coming home from Vietnam. She said that, shortly before the triple murders, she saw appellant throwing away many of his most prized possessions, and he began talking about his experiences in Vietnam, something he had never done before. On the day of the murders, appellant threw his truck keys at her and said that, if anything happened, the truck was hers. Appellant’s son testified that appellant taught him welding, and he described his father as loving and helpful to others.
Dr. Cunningham, a forensic psychologist, testified that he had reviewed appellant’s prison record which contained no disciplinary write-ups. Dr. Cunningham conducted a violence risk assessment of appellant. In his opinion, appellant had a very low probability of committing acts of violence while in prison.
Sufficiency of the Evidence to Prove Future Dangerousness
In his first and second points of error, appellant asserts that the evidence is legally and factually insufficient to support the jury’s finding that there is a probability that he would commit criminal acts of violence in the future. As appellant acknowledges, we have consistently held that we lack authority to conduct a factual sufficiency review of the jury’s future-dangerousness verdict.9 Appellant’s arguments do not persuade us otherwise.
In assessing the legal sufficiency of the evidence to support future dangerousness, we “view the evidence in the light most favorable to the jury’s findings and determine whether any rational trier of fact could have found beyond a reasonable doubt that there is a probability that [the defendant] would commit criminal acts of violence that would constitute a continuing threat to society.”10 Only if, after reviewing all of the record evidence, we conclude that a rational jury would necessarily have entertained a reasonable doubt about the defendant’s future dangerousness, will we find that the evidence is legally insufficient.11
Appellant does not suggest that the evidence of his gruesome triple murder and his life-long history of violence toward women and young girls is — viewed in a *266vacuum — insufficient to support the jury’s finding. Clearly it is sufficient. Instead, he argues that, like Saul on the road to Damascus, he has experienced a character conversion while spending the last eighteen years in prison with a spotless disciplinary record. He has proven that he no longer poses any realistic threat of violence. This is, at first blush, a compelling argument.
Appellant notes that he was almost sixty years old and in poor health12 at the time of the present trial. Appellant points to the evidence that shows that he has not merely stayed out of trouble for eighteen years in prison, but that he has made positive contributions to his prison society. He worked in the prison garment factory when he was housed in the Ellis Unit; he helped diffuse potential conflicts by talking “sense” into frustrated inmates; he formed a prison sports league; he gave commissary items to inmates who did not have money; he helped an inmate learn English and draft legal papers. Dr. Cunningham, his expert forensic psychologist, placed appellant in the lowest risk group for violence in prison. But, as the prosecutors pointed out, appellant had done many of these same positive things before the murders as well: he coached one of Karen’s daughter’s baseball teams; he fixed things around the house; he tended the garden; he praised Karen; he repaired their car; he helped organize a school sports banquet. Appellant’s son, Gordon, testified that his father helped him with sports and took him fishing and hunting. He taught Gordon welding, electrical work, and a good work ethic. He was a very patient teacher and friendly, talkative, happy, and helpful to others.
Appellant agrees with the proposition that “the past is the best predictor of the future,” and he relies upon a spotless, positive prison record as a realistic predictor of the future. Appellant concludes that, “[i]n light of [his] age and his prison record, ... the only rational finding in this case is that he would not be a continuing threat to society. For that reason, his sentence must be vacated.”13
This is the same argument that appellant made during the trial, and the jury must have taken it seriously because it asked for just three pieces of evidence during its deliberations, evidence that was directly relevant to this argument:
(1) Dr. Hodges’s Austin State School psychiatric report from 1964 when appellant was 15. That report stated that appellant seemed paranoid and distant and “extremely hostile to women”; Dr. Hodges’s impression was that “this boy represents a so-eiopathic personality disturbance of the dissocial type.”14 He concluded, *267“The long term prognosis does not look good.”
(2) The military medical record from appellant’s 1967 self-inflicted stabbing wound in his thigh after he had a fight with his girlfriend.15 According to the military doctor, appellant “revealed evidence of lifelong maladjustment.” On the hospital ward he was “hostile and belligerent” and only slowly “began to conform to the ward milieu.”
(3) The pictures and cards that appellant had in his death row cell. These included numerous pictures of scantily clad young women and girls — young gymnasts and skaters — as well as romantic cards and photographs from a female pen pal.
The jury had also heard from several different sources about appellant’s mercurial moods: one moment calm and sweet, the next moment in a towering rage.16
The jury also heard evidence that appellant, after all his time on death row, was still hostile to women. Karen testified that when she appeared for a hearing in 1998, almost ten years after appellant’s original conviction, appellant “turned around and watched me sit down. And then, after that, he kept turning around and looking at me and grinning.” It was a “weird evil grin.” Karen “called it to the Judge’s attention, and then he told him to stop. And then, he did it again. And [the judge] told him — I think his words were, I have to admonish you for that and I’ll have to call you in contempt if you don’t stop it.” According to Karen, appellant had that same grin when she testified at the 2008 trial.
Appellant’s attorney explained at trial that, “I’m not saying Bill Coble is a different person — okay—than he was in 1989. But you can see that he’s made changes. You can see that he has adapted himself to prison environment, that[ ] he’s adapted himself to institutional life. That’s very clear.” That is clear; appellant has adapted very well to prison life, but that fact, by itself, does not resolve the special issue:
Is there a probability that the defendant would commit criminal acts of violence that would constitute a continuing threat to society?17
This question is essentially a normative one as the Legislature declined to specify a particular level of risk or probability of *268violence.18 But the “future dangerousness” special issue ensures that no defendant, regardless of how heinous his capital crime, will be sentenced to death unless the jury finds that he poses a real threat of future violence.
The special issue focuses upon the character for violence of the particular individual, not merely the quantity or quality of the institutional restraints put on that person.19 As we recently stated in Estrada v. State,20 “This Court’s case law has construed the future-dangerousness special issue to ask whether a defendant would constitute a continuing threat ‘whether in or out of prison’ without regard to how long the defendant would actually spend in prison if sentenced to *269life.”21 That is, this special issue focuses upon the internal restraints of the individual, not merely the external restraints of incarceration. It is theoretically possible to devise a prison environment so confining, isolated, and highly structured that virtually no one could have the opportunity to commit an act of violence, but incapacitation is not the sole focus of the Legislature or of our death penalty precedents.22
The Supreme Court has stated that “a state capital sentencing system must: (1) rationally narrow the class of death-eligible defendants; and (2) permit a jury to render a reasoned, individualized sentencing determination based on a death-eligible defendant’s record, personal characteristics, and the circumstances of his crime.”23 Thus, juries appropriately focus upon the defendant’s individual character for violence and the probability that he would commit acts of violence in whatever society he found himself.24 Obviously, the likelihood that a defendant does not or will not pose a heightened risk of violence in the structured prison community is a relevant, indeed important, criterion, but it is not the exclusive focus of the “future dangerousness” issue.
There is no denying appellant’s impressive history of nonviolence in prison. Nor did the prosecutors at trial try to minimize that record. They noted that appellant has always done some good things in his life. The issue, however, is whether he is the same person — with the same character for sudden explosive violence — that he was when he was diagnosed at age 15 as having a “sociopathic personality disturbance of the dissocial type.” Has his character changed since he was again diagnosed as having a lifelong history of maladjustment, belligerence and violence, when he was hospitalized at the age of 19 after fighting with his fiancee and stabbing himself in the thigh as a Marine? Was the “evil grin” Karen said that he gave her in court when appellant was fifty years old, and then again when he was sixty, indicative of a continuing animosity and character for brutality toward women? And did the pictures in his death row cell indicate an unnatural interest in young, athletic, scantily clad women for a sixty-year-old man *270with a heart condition? It was the jury’s duty to assess appellant’s present character for future dangerousness, and there was ample evidence to support its finding, beyond a reasonable doubt, that appellant had not experienced a conversion on the road to Damascus; rather, he had the same character for violence at age 60 that he did at ages 15, 19, and 40, despite his spotless prison record.25
The evidence is legally sufficient to support the jury’s finding on the future dangerousness special issue. We overrule points of error one and two.
The Admissibility of Dr. Coons’s Expert Testimony
In points of error three and four, appellant contends that Dr. Richard Coons’s expert testimony concerning future dangerousness was not admissible under Rule 70226 because it was insufficiently reliable. We agree. In point of error five, appellant asserts that this type of evidence fails to meet the heightened reliability requirement of the Eighth Amendment, but the United States Supreme Court, in Barefoot v. Estelle,27 rejected this argument, and we are required to follow binding precedent from that court on federal constitutional issues.28
A. The Daubert/Kelly Hearing.
At trial, appellant objected to Dr. Coons’s proposed testimony and requested a Daubert/Kelly29 hearing outside the presence of the jury. At that hearing, Dr. Coons testified that he is board certified in general psychiatry and has been practicing forensic psychiatry for thirty-one years. He has evaluated the competency or sanity of between 8,000 to 10,000 people, has performed 150 evaluations of “future dangerousness,” and has testified in fifty trials as an expert.
Dr. Coons testified that psychiatric principles are commonly used when making determinations of a person’s danger to himself or others in the context of involuntary psychiatric commitments. He said that he also relies upon psychiatric principles when he evaluates defendants for “future dangerousness” for capital murder trials. He repeatedly stated that “the best predictor of the future is the past” and noted that
there are certain trends in people who are, in other words, habit patterns or personality patterns that — that we rely *271on. Um, and then, of course, there’s the experience one has, the training and then the experience that one has in seeing quite a number of people and, uh— uh — watching classifications within various jails and so forth. Uh, those are kind of the principles or the things that are — opinions are based on.
Dr. Coons noted that there are some psychiatric diagnoses that are listed in the DSM,30 such as antisocial personality disorder, that might indicate that a person is dangerous. But in this case, Dr. Coons relied on materials supplied by the District Attorney’s Office.
Dr. Coons explained his standard methodology in assessing the issue of future dangerousness. For at least the past twenty years he has relied upon several different factors:
(1) The person’s history of violence;
(2) The person’s attitude toward violence;
(3) The particulars of the criminal offense;
(4) The person’s personality and general behavior;
(5) The person’s conscience; and
(6) Where the person will be — in or out of prison.
He assesses these factors based on the information that he has been given. This is his own personal methodology. He does not know whether others rely upon this method, and he does not know of any psychiatric or psychology books or articles that use his factors. But “[t]hese are matters that are discussed commonly at — at forensic meetings and among forensic psychiatrists .... [B]ut generally speaking, those are the — are the kinds of things that, uh, forensic psychiatrists would take into consideration in reaching an opinion.” He doubts that his methodology is shared by everyone because different psychiatrists construct their own methodologies.
Dr. Coons stated that multiple psychiatrists would not necessarily agree on what is important in the first factor — looking to past conduct to predict future conduct. “I’m the one who’s making the decision— about whether it means something to me in terms of what I — my education or experience or background is.” It is a subjective evaluation. When assessing past violence, Dr. Coons looks at its nature and context.
The same subjectivity is true for the second factor, a person’s attitude about violence, as well as the third factor, the circumstances of the offense. Two different psychiatrists may come to different conclusions based on the same facts. Dr. Coons said that forensic psychiatrists develop an experiential body of knowledge and information and approach that helps them make their decisions. But Dr. Coons disagreed that it was “just a gut feeling.”
When it comes to the fourth factor of personality and behavior, Dr. Coons looks to whether the crime was an aberration or whether that person has had a problem looking out for other people. Is he controlling? Manipulative? With the fifth factor, “conscience is involved in — in helping people control their behavior. And, I mean, really, I guess almost everybody knows that.” There is no yardstick to measure it. With the final factor, Dr. Coons stated that if the person is on death row he will be less violent because “everybody that’s on death row is on appeal by definition. And they tend to be on their good behavior. Uh, because if they — on their bad behavior and they get another trial or punishment, they uh — they know *272they’ll hear about it again. Their violence on death row or threats or whatever.”31
All of these factors overlap and blend, but Dr. Coons knows of no book or article that discusses these factors or their overlap. He is not aware of any studies in psychiatric journals regarding the accuracy of long-term predictions into future violence in capital murder prosecutions or of any error rates concerning such predictions. Nor is he aware of any psychiatric studies which support the making of these predictions. Dr. Coons has never gone back and obtained records to try to check the accuracy of the “future dangerousness” predictions he has made in the past. He cannot tell what his accuracy rate is.
On redirect, the prosecutor asked Dr. Coons to read from a legal brief containing the names and titles of some articles on future dangerousness that had been filed in a different case, but Dr. Coons was not familiar with any of those articles.
Based on this testimony, the trial judge found that Dr. Coons qualified as an expert witness, that the subject matter of his testimony was an appropriate one for experts, and “that admitting the expert testimony will actually assist the factfinder in deciding this case.”
Dr. Coons then testified before the jury and, in response to a lengthy hypothetical setting out the salient features of appellant’s life and crimes, opined that there was a probability that appellant would commit future acts of violence.
B. Legal Principles Concerning the Admission of Expert Psychiatric or Psychological Testimony Concerning Future Dangerousness.
The admission of expert testimony is reviewed on appeal for an abuse of discretion.32 However, trial judges must act as a true “gatekeeper” when addressing the reliability and relevance of expert testimony.33 In Daubert, the United States Supreme Court held that when the subject of the expert’s testimony is “scientific knowledge,” the basis of his testimony must be grounded in the accepted methods and procedures of science.34 As that court explained,
[I]n order to qualify as “scientific knowledge,” an inference or assertion must be derived by the scientific method. Proposed testimony must be supported by appropriate validation — i. e., “good grounds,” based on what is known. In short, the requirement that an expert’s testimony pertain to “scientific knowledge” establishes a standard of eviden-tiary reliability.35
*273Four “general observations” guide the inquiry into scientific reliability: (1) falsifiability; (2) peer review and publication; (3) the existence of methodological standards, including the error rate; and (4) general acceptance within the relevant scientific field.36 The goal of these “flexible” guidelines is to evaluate the admissibility of expert testimony by the standards that comparable experts within the same scientific field use in evaluating each other’s professional work.37
In Kelly v. State,38 this Court adopted several procedural and substantive limitations upon the admission of expert scientific testimony to ensure that unreliable expertise would be excluded from the jury’s consideration.39 Under Kelly, a trial judge must, upon request, conduct a “gatekeeping” hearing outside the presence of the jury to determine whether scientific evidence is sufficiently reliable40 and relevant41 to help the jury in reaching an accurate result. Then the judge must decide whether, on balance, that expert testimony might nonetheless be unhelpful or distracting for other reasons.42 To be considered reliable, evidence from a scientific theory must satisfy three criteria: “(a) the underlying scientific theory must be valid; (b) the technique applying the theory must be valid; and (c) the technique must have been properly applied on the occasion in question.”43 The trial court’s essential gatekeeping role is to ensure that evidence that is unreliable because it lacks a basis in sound scientific methodology is not admitted.44
Forensic psychiatry is certainly a science;45 as Dr. Coons stated, it is prac*274ticed solely by those with a medical degree.46 It may be a “soft science,” but trial courts, in their gatekeeping function, must ensure that the expertise is not only soft, but that it is science as well.47 “Soft” science does not mean soft standards.48 When “soft” sciences are at issue, the trial court must inquire “(1) whether the field of expertise is a legitimate one, (2) whether the subject matter of the expert’s testimony is within the scope of that field, and (3) whether the expert’s testimony properly relies upon and/or utilizes the principles involved in the field.”49
This inquiry is somewhat more flexible than the Kelly factors applicable to Newtonian and medical science.50 “The general principles announced in Kelly (and Daubert) apply, but the specific factors outlined in those cases may or may not apply depending upon the context.”51 Un-der either Daubert/Kelly or Nenno, reliability should be evaluated by reference to the standards applicable to the particular professional field in question.52
Appellant does not quarrel with the first prong — the legitimacy of the field of forensic psychiatry, nor, apparently, with the second prong — Dr. Coons’s testimony is within the scope of forensic psychiatry, but he contends that Dr. Coons’s testimony did not properly rely upon the accepted principles of forensic psychiatry, at least as far as those principles apply to the prediction of long-term future dangerousness.
While the United States Supreme Court (as well as other American courts) has recognized the fallibility of psychiatric assessments of future dangerousness, it nevertheless acknowledged the necessary reliance on psychiatry to assist *275in judicial decisionmaking.53 We reaffirm that such expert testimony may, in a particular case, be admissible under Rule 702 and helpful to the jury in a capital murder trial.54 However, the burden is on the *276proponent of such psychiatric testimony to establish its admissibility in each individual case.55 Science is constantly evolving and, therefore, the Rule 702-703 “gatekeeping” standards of the trial court must keep up with the most current understanding of any scientific endeavor, including the field
of forensic psychiatry and its professional methodology of assessing long-term future dangerousness.56 The objective of the “ga-tekeeping” requirement is to make certain that an expert employs the same professional standards of intellectual rigor in the courtroom as is expected in the practice of the relevant field.57 The validity of the *277expert’s conclusions depends upon the soundness of the methodology.58
C. The Application of Daubert/Kelly and Nenno Principles in This Case.
As the Seventh Circuit observed in Rosen v. Ciba-Geigy Corp.,59 “under the regime of Daubert a district judge asked to admit scientific evidence must determine whether the evidence is genuinely scientific, as distinct from being unscientific speculation offered by a genuine scientist.”60 Here, there is no question that Dr. Coons is a genuine forensic psychiatrist with a lengthy medical career, but the issue under Rule 702 is whether his “future dangerousness” testimony is based upon the scientific principles of forensic psychiatry.
From this record, we cannot tell what principles of forensic psychiatry Dr. Coons might have relied upon because he cited no books, articles, journals, or even other forensic psychiatrists who practice in this area.61 There is no objective source material in this record to substantiate Dr. Coons’s methodology as one that is appropriate in the practice of forensic psychiatry. He asserted that his testimony properly relied upon and utilized the principles involved in the field of psychiatry, but this is simply the ipse dixit of the witness.62 Dr. Coons agreed that his methodology is idiosyncratic and one that he has developed and used on his own for the past twenty to thirty years. Although there is a significant body of literature concerning the empirical accuracy of clinical predictions versus actuarial and risk assessment *278predictions,63 Dr. Coons did not cite or rely upon any of these studies and was unfamiliar with the journal articles given to him by the prosecution.
Dr. Coons stated that he relies upon a specific set of factors: history of violence,64 attitude toward violence, the crime itself, personality and general behavior, conscience, and where the person will be (i.e., the free community, prison, or death row). These factors sound like common-sense ones that the jury would consider on its own,65 but are they ones that the forensic psychiatric community accepts as valid?66 *279Have these factors been empirically validated as appropriate ones by forensic psychiatrists? And have the predictions based upon those factors been verified as accurate over time?67 Some of Dr. Coons’s factors have great intuitive appeal to jurors and judges,68 but are they actually accurate predictors of future behavior? Dr. Coons forthrightly stated that “he does it his way” with his own methodology and has never gone back to see whether his prior predictions of future dangerousness have, in fact, been accurate. Although he had interviewed appellant before the first trial in 1990, Dr. Coons had lost his notes of that interview in a flood and apparently had no independent memory of that interview. He relied entirely upon the documentary materials given to him by the prosecution, including his 1989 report. Dr. Coons, therefore, did not perform any psychiatric assessment of appellant after his eighteen years of nonviolent behavior on death row, nor did he refer to any psychological testing that might have occurred in that time frame.
Based upon the specific problems and omissions cited above, we conclude that the prosecution did not satisfy its burden of showing the scientific reliability of Dr. Coons’s methodology for predicting future dangerousness by clear and convincing evidence during the Daubert/Kelly gatekeep-ing hearing in this particular case.69 We *280conclude that the trial judge therefore abused his discretion in admitting Dr. Coons’s testimony before the jury.70
D. Did Dr. Coons’s Inadmissible Expert Testimony Affect Appellant’s Substantial Rights to a Fair Sentencing Trial?
Having found error in the admission of Dr. Coons’s expert testimony, we must decide whether that error affected appellant’s substantial rights to a fair sentencing trial.71 A substantial right is affected when the error had a substantial and injurious effect or influence in determining the jury’s verdict.72 But if the improperly admitted evidence did not influence the jury or had but a slight effect upon its deliberations, such non-constitutional error is harmless.73 In making a harm analysis, we examine the entire trial record and calculate, as much as possible, the probable impact of the error upon the rest of the evidence74 We consider overwhelming evidence supporting the particular issue to which the erroneously admitted evidence was directed— here, the “future dangerousness” special issue — but that is only one factor in our harm analysis.75 It is the responsibility of the appellate court to assess harm after reviewing the record, and the burden to demonstrate whether the appellant was harmed by a trial court error does not rest on either the appellant or the State.76
*281In his Brief, appellant cites articles that note the high persuasive value of “scientific” expert testimony, especially clinical psychological testimony concerning future dangerousness.77 Indeed, some studies have shown that juror reliance on an expert’s credentials is directly proportional to the complexity of the information represented: the more complex the information, the more the jury looks to the background, experience, and status of the expert himself rather than to the content of his testimony.78 There is also some evidence that jurors value medical expertise higher than other scientific expertise; thus, even when the information is identical, jurors find evidence from a doctor more persuasive than the very same testimony from a psychologist.79 Furthermore, evidence that corresponds to firmly held beliefs may be particularly persuasive to jurors.80 Thus, an expert’s appeal to the juror’s own common sense may be considerably more persuasive than a counterintuitive and complex, but empirically verified, theory.
These studies and articles would support a determination that the erroneous admission of a psychiatrist’s unreliable testimony concerning the defendant’s future dangerousness affects a substantial right to a fair sentencing hearing under Tex.R.App. P. 44.2(b). However, each case must be examined on its own facts, taking into account the specific evidence and the probable impact of the erroneously admitted expert evidence upon the jury’s decision-making in the particular case.
In this case, there was ample evidence that there was a probability that appellant would commit future acts of violence quite apart from Dr. Coons’s testimony. And, as noted above, it was some of that independent evidence that the jury requested to see during its deliberations. First, the psychiatric interview and evaluation done by Dr. Hodges more than twenty years before the offense and forty years before the trial reached the same basic conclusion as Dr. Coons did concerning appellant’s character and his animosity toward women. Dr. Hodges’s 1964 interview and clinical evaluation was completed long before any possible motive to view the facts and events of appellant’s later life through any “future dangerousness” litigation prism had arisen. Expertise that is developed entirely independent of litigation by professionals acting in their normal field is more likely to be considered reliable than expertise developed especially for trials.81 The same is true with the *2821967 military medical report which noted appellant’s “lifelong maladjustment” and his jealous violent rage when he thought that his fiancée was having an affair with someone else. Significantly, the jury asked to see these two reports during its deliberations; it did not ask to see Dr. Coons’s 1989 report. We have often held that erroneously admitting evidence “will not result in reversal when other such evidence was received without objection, either before or after the complained-of ruling.”82 Although neither Dr. Hodges nor the military doctor specifically opined on whether there was a probability in 2008 that appellant would commit acts of future violence, their psychiatric and medical assessment of appellant’s character for violence is remarkably similar to that of Dr. Coons.
Furthermore, Dr. Coons’s testimony was rebutted and refuted by appellant’s expert, Dr. Mark Cunningham, a forensic psychologist. Although Dr. Cunningham is not a medical doctor, he did win the 2005 Texas Psychology Association award for his outstanding contribution to science and, in 2006, he was awarded the American Psychological Association (APA) award for distinguished contributions to research in public policy. Both awards were for his research concerning factors that predict violence in prison and his research in capital sentencing. He is also among the 2,000-3,000 psychologists elected as a Fellow of the APA out of the 155,000 members. He has published a significant number of peer-reviewed studies and articles. He testified, with a PowerPoint slide presentation to illustrate,83 about the violence risk assessment factors that he uses to assess the probability of future dangerousness in prison. His factors are based on research data from prisons, as well as other research and scholarly writings. He explained how his research is “scientific,” replicable, and less subjective: “It’s not based on my gut feeling about something. It’s based on what the data tells me. And so, it’s accurate. It’s reliable.”
After explaining the various studies, data, and statistical analysis, Dr. Cunningham concluded that appellant fell within the lowest risk-of-violence category.84 He *283criticized “the hypothetical inference” mode of predicting future dangerousness as
entirely speculative.... That’s just blind guessing unless those factors have been demonstrated to be predictive of violence in prison. Critically important. ... That’s the problem with not knowing the literature, without knowing anything about the scientific studies that have been done in this area is then you have no idea whether the factors that you’re looking at are predictive of anything or not.
According to Dr. Cunningham, if “what you’re doing is basing it on your own gut and you haven’t done anything to check whether your gut reaction is correct or not, then your accuracy level never improves.”
He pointed to appellant’s first trial as an example of the “tea-leaf-reader” school of subjective clinical assessments. In that trial, Dr. James Grigson,85 who used the same subjective methodology as Dr. Coons, testified that, in his opinion, appellant posed no risk of future violence: “[H]e said, the ladies and gentlemen of the jury are more likely to kill somebody in the future” than appellant. Dr. Coons, using that very same methodology and facts concerning appellant, came to exactly the opposite conclusion.86 Dr. Cunningham also told the jury that the major psychological associations had criticized Dr. Coons and his methodology as “unreliable and inconsistent with the standard of practice.” In sum, Dr. Cunningham refuted Dr. Coon’s expertise and the whole “tea-leaf-reader” notion of clinical psychiatric predictions of future dangerousness.
Furthermore, the prosecution did not rely heavily upon Dr. Coons’s testimony during its closing arguments. Instead, the prosecutor emphasized his position that appellant was exactly the same person that he was when he killed Karen’s parents and her brother back in 1989. He had not changed a bit.87 The prosecutor then went *284on to recount evidence from the murders themselves and appellant’s bragging to Karen afterwards. The prosecutor then turned to the topic of predicting future dangerousness:
Can we predict a person’s future? Well, we absolutely can. You heard what Dr. Hodges said. He made an analysis of it. He talked about how he had a dislike of women, how he had a low opinion of them. Did he? His conduct was absolutely borne out. And Dr. Hodges said the prognosis is poor, and it was, because this person ultimately cares only about himself.
He then recounted how Bobby’s fellow officer had predicted that appellant would commit some violent act after he had been arrested for kidnapping Karen before the murders. He then turned briefly to Dr. Coons:
Dr. Coons examined — first he talked to [appellant] personally before the first trial in 1990. He interviewed him, then he gave his assessment of him. And the assessment is — sure, you deal with medical predictions and the training of psychiatrists — but it’s just common sense. If you don’t have a conscience and you’ve committed dangerous violent acts and you’ve shown that you have no regard for human beings in any form unless it’s something that serves you, of course, there’s a probability that you will commit criminal acts of violence.
The prosecutor then referred to appellant’s expert psychologist, Dr. Cunningham, and how he had called Karen biased for saying that appellant twice gave her an evil grin in two different court appearances. The prosecutor then returned to his theme of appellant’s lack of conscience and how he had simply been restrained, not changed, in prison.
The defense, in its closing argument, quickly focused on the future dangerousness issue as well. Counsel argued that the statistical evidence that Dr. Cunningham had presented made it very difficult for the prosecution to prove that appellant would commit future acts of violence. He compared the two experts:
I want to talk about Dr. Coons versus Dr. Cunningham, because it really does sort of come down to Dr. Coons versus Dr. Cunningham. Dr. Coons is a likable guy. Dr. Coons does an excellent job of testifying. He seems to have a lot of *285horse sense. Okay. He seems to have a lot of common sense. That’s totally true. Dr. Cunningham is extremely long-winded. Okay. He has a hard time sort of answering a question directly. I recognize those facts. But that’s because he is a scientist.88 All right. And Dr. Cunningham talked about being a scientist and what that means. What that means is, I don’t just look at the evidence and make a wild guess. Okay. I’m not a tea-leaf reader. I’m not a guy who says, well, I’m just going to depend on my — my experience and say this person is a future danger — okay—without going back and checking my work, without quantifying things, without being able to say, you know, I’m correct to this certain quantum of correctness.... So he’s a scientist. A scientist comes up with an idea, a theory. Okay. He tests that theory. He doesn’t just test the theory, but he also gives his data to other scientists to look at, so they can test his theory. Then he goes back and double-checks his work. Then he thinks, now, maybe there’s a weakness in my own argument that I’ve already made. Let’s go back and double-check that weakness and see it that changes our numbers or does that reinforce our numbers. All right. So that’s what a scientist is supposed to do.
Do you remember Dr. Coons’s testimony? Dr. Coons, do you check your work? Not really. Dr. Coons, do you remember going back and looking at the records of people that you have predicted are going to be a future danger to see if they really were? Well, I’m sure I’ve done it, but I can’t tell you who I’ve done it with. In other words, he’s a guy who is completely uninterested in whether he’s correct or not....
... How can he ever get better? How can he establish for the jury that his opinion is reliable? He can’t because he’s not a scientist. He’s a tea-leaf reader.
The defense then recapped Dr. Cunningham’s testimony which had been that appellant posed an extremely low risk of committing future acts of violence because (1) he is well-adjusted to institutional life; (2) he is sixty years old and thus has “aged out” of his violent years; (3) he has performed many positive acts and developed a positive attitude toward fellow inmates; (4) he is serving a very long sentence and “40 years of tests” show that long-term inmates are statistically less likely to commit acts of aggression than are “short-term-ers;” (5) he has a GED and additional work certificates; and (6) he has continuing family ties to the community. All of these factors are supported by “the numbers that exist in reality. These are the official numbers. So he’s not making up the numbers. He’s a scientist. He’s just reporting what the data is.” The defense concluded its discussion of “future dangerousness,” with the statement that “Dr. Cunningham’s conclusions are very appropriate and very reasonable and scientific and provable as opposed to Dr. Coons’s conclusions, which are nonscientific and not provable. Okay. And even if they were provable, he hasn’t bothered to go out and try to prove them.” Counsel then moved on to the mitigation issue and appellant’s miserable childhood and youth.
During his final argument, the prosecutor mentioned Dr. Coons very briefly by reminding the jury that another psychia*286trist, Dr. Hodges, had talked to appellant back in 1964 and “he looked at him and listened to his answers” and reached the conclusion that appellant was “extremely hostile to women, very low opinion of women, has poor control, very low self-esteem. Projects a great deal of responsibility for his own actions on other people. It was Karen’s fault that she got kidnapped. It was Karen’s fault because she stood up to him. And it ruined his life. So it was her fault. And he had to extract revenge on her and he did it in the most brutal, the most selfish way he possibly could.” The prosecutor then referred to the military doctor’s assessment from 1967, with both doctors reaching the “same common sense assessment” of appellant.
Based upon the complete record of this ease, we find that the error in admitting Dr. Coons’s testimony did not affect appellant’s substantial right to a fair sentencing hearing because
(1) There was ample other evidence supporting a finding that there was a probability that appellant would commit future acts of violence;89
(2) The same basic psychiatric evidence of appellant’s character for violence was admissible and admitted, without objection, through other, entirely objective, independent medical sources — the reports by Dr. Hodges and the military doctor years before appellant committed these murders; 90
(3) Dr. Coons’s opinion was not particularly powerful, certain, or strong;91 his opinion, coming after an extremely long and convoluted hypothetical was simply that “there is a probability that” appellant would be a continuing threat to society by committing criminal acts of violence;
(4) Dr. Coons’s testimony was effectively rebutted and refuted by Dr. Cunningham, who not only relied upon *287specifically listed scientific materials and data during his testimony, but who also noted that Dr. Coons and his methodology had been criticized by both the American and Texas Psychological Associations; and
(5) The State barely mentioned Dr. Coons during closing argument and did not emphasize him or his opinions.
Given these particular circumstances, we conclude that the error in admitting Dr. Coons’s testimony did not have a “substantial and injurious” effect upon the jury’s deliberations concerning the future dangerousness special issue.92 We therefore overrule points of error three, four, and five.
The Admissibility of A.P. Merillat’s Testimony
In his sixth point of error, appellant claims that the trial court erred in admitting the testimony of A.P. Merillat, an investigator for the Special Prosecution Unit, about the Texas prison classification system and violence in prison. Appellant argues that: (1) Mr. Merillat’s testimony was irrelevant as it did not relate to appellant personally, and (2) this witness testified to information that was already common knowledge among jurors. The State argues that Mr. Merillat’s rebuttal testimony was relevant to refute Dr. Cunningham’s statistical data and to impeach the accuracy of his “low risk” future dangerousness prediction.93 We agree that Mr. Merillat’s testimony was admissible as rebuttal “educator-expert” evidence.
On voir dire, Mr. Merillat stated that his testimony is based on his specialized knowledge of Texas prisons and prison violence during his nineteen years as a criminal investigator with the Special Prosecution Unit. He proposed to testify concerning the under-reporting of prison violence in official data compilations, the prison classification system, and the opportunities for violence inside prison.
The trial judge allowed Mr. Merillat’s testimony, although he granted appellant’s motion in limine to avoid mention of any specific instances of misconduct by other inmates except for one anecdote concerning an inmate’s forced starvation death which served as “a great example for un-derreporting of violence.”
Mr. Merillat then testified before the jury about the inmate classification system and the under-reporting of violence in prison. He also described administrative segregation and how it is used as “punitive housing” for recalcitrant inmates. Mr. Merillat explained why the official prison statistics used by Dr. Cunningham are not completely reliable: (1) the prison reporting system does not match the penal code definitions of “violent” behavior;94 and (2) *288not all incidents of inmate-on-inmate incidents of violence are reported. Finally, he told the jury that, in the last few years, his unit had prosecuted 94 inmates who were serving life sentences for capital murder for both assaultive and non-assaultive felonies.
On cross-examination, Mr. Merillat agreed that he knew nothing about appellant except that his office had never prosecuted him. He agreed that he was not qualified to express any opinion regarding appellant’s “future dangerousness.” He also explained how death row inmates “had the run of the row” and could work in the garment factory when death row was in the Ellis Unit. Mr. Merillat agreed that the point of his testimony was that there are abundant opportunities for inmates to be either violent or good, depending upon their own decisions.
Appellant asserts that the primary subject of Mr. Merillat’s testimony— opportunities for violence in prison — is within the common knowledge of the jurors. Indeed, most jurors probably have some understanding that violence can and does occur in prison, but a trial court need not exclude expert testimony when the general subject matter is within the comprehension of the average juror, as long as the witness has some specialized knowledge on the topic that will “assist” the jury.95 It is only when the expert offers no appreciable aid that his testimony fails to meet the Rule 702 standard.96 The question under Rule 702 is not whether the jurors know something about this subject, but whether the expert can expand their understanding in a relevant way.
In this case, Mr. Merillat confined his testimony to specific information about the operations of the Texas prison system and the opportunities for violence or productive behavior. His expert testimony was intended to (1) educate the jury about an area in which it lacked a thorough understanding; 97 and (2) cast doubt upon the official prison data that Dr. Cunningham relied upon. Mr. Merillat acted “as an advisor to the jury, much like a consultant might advise a business[.]”98 Because Mr. Merillat’s testimony was edueator-exper-*289tise information designed to “assist” the jury under Rule 702, the trial judge did not abuse his discretion in admitting it. Point of error six is overruled.
In point of error seven, appellant contends that the trial court erred by allowing Mr. Merillat to testify to hearsay information in violation of the Confrontation Clause99 and of the Texas Rules of Evidence. Out of six instances in which appellant claims that Mr. Merillat testified to hearsay information, we have found only three trial objections based on hearsay or the Confrontation Clause. We will address only those three instances:
(1) Appellant objected on the basis of hearsay to Mr. Merillat’s statement that 78 serious staff assaults were documented in the official prison report that Dr. Cunningham had used as the basis for his statistical analysis;
(2) Appellant objected on the basis of hearsay and the Confrontation Clause to Mr. Merillat’s explanation of why inmate-on-inmate violence is under-reported — nobody wants to be a “snitch” which is the “very lowest form of life in the penitentiary”; and
(3) Appellant objected on the basis of hearsay and an inability to confront and cross-examine when Mr. Meril-lat cited the story of an inmate who had been beaten and starved to death by his stronger, gang-member cellmate, as an example of why and when fellow inmates fail to report acts of violence.
The trial judge properly overruled these three hearsay and confrontation objections.
Hearsay is an out-of-court statement by a person offered for the truth of the matter asserted.100 None of these three pieces of testimony fits that definition. In the first, Mr. Merillat was not offering his statement of the official prison data compilation of “78 serious staff assaults” for the truth of the matter asserted — that there were 78 serious staff assaults in the previous year. Quite the reverse. Mr. Meril-lat’s point was that the official number of 78 was significantly lower than the actual number of serious assaults and thus the official prison statistics that Dr. Cunningham used as the basis for his expert opinion were inaccurate.101 In the second, the testimony concerning why assaults upon inmates aren’t reported “because by telling on the person who did it, they are going to be much worse off[,]” Mr. Merillat did not disclose any out-of-court statement.102 He was simply explaining, as a general proposition, why inmates do not “snitch” on each other. In the third, the inmate who was beaten and starved to death, appellant does not point to any out-of-court statement. There is none. Mr. Merillat was recounting an event, not a verbal or writ*290ten statement. He may have first heard of the event by someone telling him of it,103 but he did not recite or imply any out-of-court statements.104 Because the trial judge did not abuse his discretion in overruling appellant’s hearsay and confrontation objections, we overrule point of error seven.
In his eighth point of error, appellant asserts that Mr. Merillat’s testimony was inadmissible because of the Eighth Amendment’s “heightened reliability” requirement in capital murder prosecutions. Appellant fails to cite any authority for increasing the admissibility requirements for evidence in a capital murder sentencing trial. Indeed, some state and federal courts have suggested that the Confrontation Clause, the Rules of Evidence, and the rule against hearsay do not apply with full force in capital murder sentencing trials.105 We express no opinion on that matter, but we reject appellant’s Eighth Amendment claim and therefore overrule point of error eight.
Emotional Outbursts by Two Witnesses
In points of error nine and ten, appellant claims that the trial court erred in denying his motion for a mistrial when Karen Vicha and Lorna Sawyer made separate emotional outbursts during the punishment trial. In each case, the trial judge sustained appellant’s objection to the outburst and instructed the jury to disregard the remarks. We conclude that the trial court did not abuse his discretion in these rulings.
During Karen Vicha’s testimony describing what appellant told her about how he had chased and shot her brother, she explained that
[Appellant] started talking about — he told me, you’re pretty tough, you put up a good fight with that gun. And he said, *291your brother thought — he said, your brother thought he was tough too. He said, all cops think they’re tough, but he thought he was really tough. He said— he said — he told me he said, all I was trying to do was keep you away from my sister. And he said, I finally had to shut him up and blow a hole that big in his neck.
At that point, Karen broke down crying and said, “And I hate you for making me go through this again and my kids. You’re mean.” The trial judge immediately called a recess.106 When the jury returned, the trial judge instructed them: “Ladies and gentlemen of the jury, at this time I am going to give you an instruction to disregard the last comment of the witness and not consider it for any purpose whatsoever.” He denied appellant’s motion for mistrial, and the prosecutor continued with his questions.
After six more witnesses had testified, the State called Lorna Sawyer, appellant’s cousin. As soon as she had been sworn in, but before any questioning, she burst out, “Evil piece of shit.” The defense immediately responded: “Judge we object. Call for a mistrial. Request an instruction to disregard. Call for a mistrial.” The Judge said, “I’ll instruct the jury to disregard the last comment of the witness,” and Ms. Sawyer said, “Sorry.” When appellant then requested that she be excluded as a witness, the trial judge took a recess.107
"When the jury returned, Ms. Sawyer testified that appellant offered her a job at a drive-in theater when she was 16. She had worked there for about two and a half weeks when appellant picked her up for work, but he took her to his house and raped her instead. She was so scared that she had never told anyone, except her sister, about this experience. When the prosecutor asked Ms. Sawyer if she was still afraid of appellant, the following occurred:
Witness: Actually, uh, without being ugly, I’d like to go there and just knock the shit out of him.
Defense: Judge, I’m, going to have to object, Your Honor. I think that was an inappropriate comment.
Witness: It is not inappropriate.
Defense: I’d object to the sidebar from the witness. Judge. I’d ask that— the jury to disregard.
*292Court: The jury is instructed to disregard the last comment of the witness and not consider it for any purpose.
Defense: I’d ask for a mistrial, You Honor.
Court: That’s denied.
State: Lorna—
Witness: I’m sorry, I’m sorry.
And the testimony then continued.
Appellant argues that these outbursts, individually or collectively, unfairly influenced the jury and that influence could not be limited by an instruction to disregard.
A trial judge’s denial of a motion for mistrial is reviewed under an abuse of discretion standard,108 and his ruling must be upheld if it was within the zone of reasonable disagreement.109 We have held that an outburst from a bystander or witness “which interferes with the normal proceedings of a trial will not result in reversible error unless the defendant shows that a reasonable probability [exists] that the conduct interfered with the jury’s verdict.”110 In the context of such outbursts, the trial judge’s instructions to disregard are generally considered sufficient to cure the impropriety because it is presumed that the jury will follow those instructions.111
Appellant relies upon Stahl v. State112 for his claim that the judge’s instruction to disregard the spontaneous outbursts could not have cured their prejudicial effect and those outbursts must have interfered with the jury’s verdict. But Stahl was decided upon the basis of prosecutorial misconduct, not merely the witness’s emotional outburst. In Stahl, the prosecutor called the victim’s mother to the stand, knowing that she was prone to emotional outbursts, and asked her to identify a photograph of her dead son.113 She burst into tears and yelled, “Oh, my god. My baby. My God.... May he rest in hell. May he burn in hell. Oh, my baby.”114 The judge instructed the jury to disregard, but the prosecutor “exacerbated” the impact by repeatedly referring to the incident in closing argument.115 The prosecutor’s “deliberate” and “persistent” conduct, “in direct contravention of prior rulings by the judge” indicated “a desire to impermissibly sway the jury.”116 Indeed, the court of appeals had suggested that the Stahl prosecutor “actually orchestrated the original *293outburst.”117
In this case, however, there is no suggestion that the prosecutor anticipated the short emotional outburst by Ms. Vicha in the middle of her lengthy testimony or the entirely inappropriate start of Ms. Sawyer’s testimony. In the first instance, the prosecutor agreed with the correctness of an instruction to disregard and, in the second, he did not attempt to justify Ms. Sawyer’s outburst.118 The trial judge immediately instructed the jury to disregard those outbursts, and we must presume that the jurors followed these instructions.119 The prosecution did not refer to, or attempt to capitalize upon, the outbursts during closing arguments. Furthermore, they occurred during the sentencing stage of a capital murder trial, not the guilt stage as in Stahl. At the punishment hearing, evidence of the defendant’s character is both relevant and admissible as is the opinion testimony concerning good or bad character traits by those who know him.120 Obviously, character evidence must be offered in a proper form and be responsive to specific questions, so these outbursts were not proper, but their potential for prejudice was less than had they occurred during the guilt phase of a trial.
Because we conclude that nothing in the record suggests that the outbursts were of such a nature that the jury could not follow the trial judge’s instructions to disregard them,121 we overrule appellant’s points of error nine and ten.
The Admission of the Hearsay Statement by a Witness’s Sister
In his eleventh point of error, appellant claims that the trial judge erred in admitting Amy Zuniga’s testimony that her sister, Karen, told her that appellant was looking in her bedroom window as she was dressing. Appellant objected to hearsay, but the trial judge admitted Amy’s testimony as both an excited utterance and a present sense impression. We conclude that the trial judge did not abuse his discretion in finding that Karen’s out-of-court statement was admissible as an excited utterance exception to the hearsay rule.122
Amy Zuniga testified that appellant was her uncle and, when she was young, she thought he was a model of how a parent should be because he was so nice to his own son. However, Amy changed her mind about appellant when she was fifteen. She explained that one day she was sitting in a rocking chair in her nightgown, when appellant came in and pulled her legs apart; then he made “a vulgar display like he was licking me” between her legs. After that, she avoided him. But right before appellant moved into Amy’s mother’s home shortly before the murders, Amy came out of the shower and was changing her clothes in her bedroom when she heard a “commotion from the kitchen, a beating on the window.” Then Karen ran through Amy’s bedroom door from the kitchen, went outside through her back bedroom door, and started yelling. Amy *294peeked out through the curtains and saw appellant driving off in his truck. Then Karen came back inside, “very mad and frustrated. She was red, angry.” Karen told Amy that she had seen appellant outside looking through the curtains as Amy was dressing.
The trial judge admitted Karen’s out-of-court statement to Amy based on it being both an excited utterance and a present sense impression. We need examine its admissibility only under the excited utterance exception. An excited utterance is a statement that relates to a startling event or condition, and it is made when the declarant is still under the stress of excitement caused by the event or condition.123
Appellant contends that Karen’s statement was not admissible as an excited utterance because there was no showing that she was in the “grip of violent emotion, excitement or pain.”124 She was angry, not excited. The critical question, however, is not the specific type of emotion that the declarant is dominated by — anger, fear, happiness — but whether the declarant was still dominated by the emotion caused by the startling event when she spoke.125 Appellant also argues that Karen’s statement was not admissible because there was no independent evidence of the startling event — appellant’s “Peeping Tom” conduct. Appellant cites to a Texas Supreme Court case, Richardson v. Green,126 which applied the common-law “res gestae” rule. But Rule 803(2) changed the common law; the current rule does not require independent evidence of the exciting event before the trial judge may admit the declarant’s statements relating to that event.127 The trial judge decides, under Rule 104(a), whether there .is sufficient evidence to prove an exciting event, and he may consider the statement itself in making that decision.128 Here, for example, the trial judge could consider the *295evidence that Amy said that she (1) heard her sister banging on the kitchen window, (2) saw Karen run through her bedroom and out the door, and (3) saw appellant driving off just before Karen returned to tell her that appellant was peeping in her bedroom window. That evidence, when combined with Karen’s statement, would support a finding of the startling event— appellant’s “Peeping Tom” conduct.129
Because we conclude that the trial court did not abuse his discretion in admitting Karen’s excited utterance, we overrule appellant’s eleventh point of error.
Miscellaneous Claims
A. Limitation of Voir Dire Questions
In his twelfth point of error, appellant claims that the trial court erred in limiting his voir dire by refusing to allow him to question the jurors about the mitigation value of specific facts, including evidence of a troubled childhood, mental illness or extreme emotional distress, community service, age, kindness to others, work ethic, or military service. The State objected, citing Standefer v. State,130 Sells v. State,131 and Wingo v. State,132 and stated that these were commitment questions. The trial judge sustained the State’s objections, but allowed more general mitigation questions about whether there was anything that the jurors “could consider under the circumstances of having found [appellant] a future danger to society which might merit a life penalty.”
In Raby v. State, we rejected appellant’s claim that he is entitled to ask potential jurors in a death penalty case about what specific evidence that juror could or would consider as mitigating.133 We stated that “[a] trial court does not abuse its discretion by refusing to allow a defendant to ask venire members questions based on facts peculiar to the ease on trial (e.g. questions about particular mitigating evidence).” 134 Appellant does not persuade us that Raby was wrongly decided. We therefore overrule appellant’s twelfth point of error.
B. The Mitigation Instruction
In his thirteenth point of error, appellant claims that the trial court erred by refusing to instruct the jurors that they need not unanimously agree on what particular evidence supports an affirmative finding on the mitigation issue. Appellant invokes Mills v. Maryland135 in support of *296his argument. We addressed and rejected this same argument in Segundo v. State,136 and appellant has not persuaded us that Segundo was wrongly decided. We over-rale his thirteenth point of error.
C. The Definition of Mitigating Evidence
In his fourteenth and fifteenth points of error, appellant argues that the trial court should not have given the jury the statutory definition of mitigating evidence “as evidence that a juror might regard as reducing the defendant’s moral blameworthiness.”137 This same claim was rejected in Roberts v. State,138 and appellant’s arguments do not persuade us to overrule that case. He also contends that the trial judge should have instructed the jury that there need be no nexus between the mitigating evidence and the capital murder because the mandatory statutory definition, he argues, implies that there must be a connection between the reduced moral blameworthiness and the capital offense itself. We do not see any “nexus” requirement in the statutory definition. Appellant relies upon Tennard v. Dretke,139 but the Supreme Court, in that case, simply chastised the Fifth Circuit for requiring a nexus between the crime and the mitigating evidence.140 It never suggested that a jury can, should, or must be instructed not to consider any nexus between the crime and the mitigating evidence. Such an instruction would be necessary only if the jury would be reasonably likely to infer a nexus requirement from the statutory words.141 That is not the case. We overrule appellant’s fourteenth and fifteenth points of error.
D. The Mitigation Issue
In his sixteenth point of error, appellant relies on Apprendi142 and Ring143 to argue that Article 37.0711 is unconstitutional because it fails to require the State to prove beyond a reasonable doubt that there are no mitigating circumstances that warrant a life sentence. He fails to mention that this Court has rejected that claim in numerous cases,144 and- he fails to persuade us that our prior decisions were mistaken.
In his eighteenth point of error, appellant claims that the Texas death-penalty scheme is unconstitutional under Penny II,145 because the mitigation issue sends *297“mixed signals” to the jury, thus rendering any finding reached on that special issue unreliable. Penry II is distinguishable because, in that case, the jury was given a judicially crafted nullification instruction.146 Here, the jury was given the statutorily mandated mitigation question, which does not contain a nullification instruction. No error exists, and we have repeatedly rejected this claim.147 We overrule appellant’s eighteenth point of error
E. Constitutional Challenges to Art. 37.0711
In his seventeenth point of error, appellant contends that the Texas death-penalty statute gives the jury too much discretion and therefore permits arbitrary and inconsistent application of the ultimate penalty. We have repeatedly rejected this claim and appellant does not persuade us to overrule these prior eases.148 In fact, he fails to mention them.
In his nineteenth and twentieth point of error, appellant argues that the jurors should have been instructed on the consequences of a hung jury so that they would immediately stop deliberating if a single juror voted in appellant’s favor on an issue that required unanimity. He also argues that the jurors should not have been instructed on the “10-12” Rule. Although appellant fails to mention controlling precedent from this Court, we have repeatedly rejected these claims.149 We do so again.
In his twenty-first through twenty-third points of error, appellant argues that the statutory “future dangerousness” special issue is unconstitutional because the terms “probability,” “criminal acts of violence,” and “society” are not defined. We have repeatedly rejected these claims,150 and, although counsel suggests that we should revisit this precedent, we decline to do so.
In his twenty-fourth point of error, appellant contends that the statutory “future dangerousness” special issue violates the Eighth Amendment because no one can reliably predict whether another person will commit acts of violence in the future and therefore this is an arbitrary factor. The “future dangerousness” aggravating factor has been recognized by the Supreme Court as properly narrowing the jury’s consideration to ensure individualized sentencing as recently as two years ago in Kennedy v. Louisiana.151 Although *298appellant asserts that only one other state, Oregon, requires a finding of future dangerousness in his effort to prove that a “national consensus” has developed against imposing the death penalty based on that factor, he fails to note that twenty-one other states include a defendant’s possible future dangerousness among the aggravating circumstances to be considered at the sentencing stage of a capital case.152 Furthermore, we have previously rejected this claim,153 and we are not persuaded by appellant’s arguments that our precedent should be overruled.
In his final point of error, appellant claims that the “future dangerousness” statutory scheme violates the Texas constitutional ban on cruel or unusual punishment. As appellant acknowledges, we have already rejected this argument.154 Appellant asserts that we now have before us his “evidence” that the “future dangerousness inquiry results in inaccurate, unreliable, arbitrary and disproportionate determinations.” His evidence is the citation to an article written by the Texas Defender Service,155 an advocacy group that represents inmates on death row. This is not the type of “evidence” upon which we can base a finding that the “future dangerousness” special issue is necessarily an unreliable factor to use in determining whether a life or death sentence is appropriate. Evidence proves historical facts; the “future dangerousness” special issue is a normative assessment mandated by the legislature and determined by the jury. Questions about its “appropriateness” as a factor in determining a life or death sentence should be addressed to the legislature. Furthermore, the article speaks of psychiatric predictions, not of the unreliability of jury verdicts. We overrule appellant’s twenty-fifth point of error.
Having found no reversible error, we affirm the trial court’s judgment and sentence.
MEYERS, J., joined except for points of error 3 and 4.
KELLER, P.J., filed a concurring opinion in which MEYERS and KEASLER, JJ., joined.
HERVEY, J., concurred.