¶1dissenting.
¶2Fundamental to our notions of a fair trial is the right of the defendant to be convicted only on evidence produced in court, and not on evidence contained in newspaper stories. The newspaper story that appeared less than one week before the trial, to which two-thirds of the venire apparently were exposed, provided the information that Barton had previously been convicted of this murder by a jury in another county, but that his conviction had been overturned, that the victim was his landlady, and that he had been evicted from her trailer court. The latter is a “fact” that would have supplied a motive for the killing. At the time of voir dire examination of prospective jurors, the trial court and counsel knew that information as to motive was not going to be admitted into evidence, and defense counsel said the information • was false. In the circumstances of this case, I believe the voir dire was inadequate to assure that Barton would be tried only on properly admitted evidence; therefore, I respectfully dissent.
¶3In another, previous trial Barton had not been convicted of this killing because the jury was unable to agree on a verdict. Much of the certainty that his most recent trial is afforded, as well as evidence of aggravating circumstances supporting imposition of the death penalty, came from ever-helpful fellow prisoners. Perhaps the evidence of guilt may be subject to non-frivolous debate; if so, we should give particular scrutiny to assure that facts outside the courtroom did not assist in the conviction of this defendant. If we are unsure, a new trial should be afforded.
¶4It may seem wasteful to try Barton again, since 24 jurors have unanimously found him guilty in two of his three trials. On the other hand, there were 12 who were unable to agree on Barton’s guilt. Where the stakes are life and death, we do not hesitate to give trials careful review. It is impossible to guarantee a capital defendant a perfect trial, but he is entitled to one that is more than merely good enough. Since 1976, the death penalty has been reinstated in most states. It has been reported that nationally, since 1976, 77 death row prisoners found guilty by unanimous juries have been set free; the number of death row inmates later found to have been wrongfully convicted is thus about one-seventh of the number of prisoners executed.
¶5The voir dire examination conducted in this case resembles that upheld in Mu’Min v. Virginia, 500 U.S. 415, 111 S.Ct. 1899, 114 L.Ed.2d 498 (1991). There, as here, the trial court divided the prospective jurors into small groups, but denied questions as to the source and content of the pre-trial publicity. However, in Mu’Min the pre-trial publicity was extensive and the evidence of the defendant’s guilt overwhelming. Here, the pre-trial publicity was not extensive - it was intensive or targeted, in that these Benton County jurors would have had no exposure to media accounts of the killing when it actually occurred some years before in a different county. Rather, these prospective jurors were exposed to a particular story in the local newspaper (and perhaps in other sources) immediately prior to the trial and apparently after the time the venireper-sons were called for jury service. Of the initial venire of 92 persons, 63 or 64 had heard about the case. After initial excu-sáis of venirepersons for cause, 40 prospective jurors were questioned, 17 had already formed an opinion about the case, and 27 had discussed the case with other persons and/or had heard someone else express an opinion about the case.
¶6Opinions, of course, are based upon “facts”, at least in part. When a juror is asked if he can set aside what he has heard and his own opinions, and render a fair verdict, most will respond affirmatively. In Mu’Min v. Virginia, only one of the many jurors exposed to pre-trial publicity indicated an inability to do so. In Barton’s case, there were more who expressed that opinion and were excused.
¶7But if motive is a critical question in the minds of jurors, and the only evidence of motive is in the pre-trial publicity to which a substantial number of jurors were exposed, it is impossible to ascertain on this record that such a fact has truly been set aside. Missouri cases hold that a defendant’s right to an impartial jury has been sufficiently safeguarded if the venireper-son is questioned appropriately as to bias and states that his or her decision can be made based on the evidence presented at trial. The trial judge, of course, must believe the statement and believe the prospective juror to be unbiased. State v. Nicklasson, 967 S.W.2d 596, 611-612 (Mo. banc 1998).
¶8However, the fact that prospective jurors say they can set aside what they have heard or seen should not end the inquiry. In Irvin v. Dowd, 366 U.S. 717, 81 S.Ct. 1639, 6 L.Ed.2d 751 (1961), for instance, the court held that the evidence of deep and bitter prejudice that pervaded the community and was reflected in voir dire questioning was so prejudicial that a new trial was warranted, even though the jurors stated they could decide the case on the evidence presented at trial. Irvin was charged with six murders that had generated substantial local publicity and indignation. Eight of the twelve jurors admitted thinking the defendant was guilty, but each claimed they could remain impartial. In a much less extreme case than Irvin, the court in Marshall v. United States, 360 U.S. 310, 79 S.Ct. 1171, 3 L.Ed.2d 1250 (1959), found that the exposure of some jurors during the trial to newspaper articles with facts about Marshall not admissible in evidence was so prejudicial as to entitle Marshall to a new trial. During trial for unlicensed dispensing of drugs, the prosecutor sought to introduce Mar*32shall’s prior convictions for practicing medicine without a license. The trial judge refused to admit the prior convictions into evidence, but two newspapers containing the information got before seven of the jurors. The trial judge questioned the jurors individually, and each assured the court that they could decide the case only on the evidence presented at trial. See also, Sheppard v. Maxwell, 384 U.S. 333, 86 S.Ct. 1507, 16 L.Ed.2d 600 (1966), and Patton v. Yount, 467 U.S. 1025, 104 S.Ct. 2885, 81 L.Ed.2d 847 (1984).
¶9One of the weaknesses of the jury selection standards articulated since Mu’Min v. Virginia, supra, is that jurors instructed to disregard something will often do the opposite, though perhaps not consciously disregarding the court’s admonitions. Kalvin and Zeisel, The American Jury (University of Chicago Press, 1971) reported that juries that had previous knowledge of a criminal defendant, such as criminal record, were more likely to convict. The same series of empirical studies of jury behavior found that jurors instructed to disregard a particular fact apparently did the opposite. See, Broeder, The University of Chicago Jury Project, 38 Nebraska Law Review 744 at 754 (1959). Even though there is “no doubt each juror was sincere when he said he would be fair,” as the court said in Irvin v. Dowd, supra,“The influence that lurks in an opinion once formed is so persistent that it unconsciously fights detachment from the mental processes” of the average person. 366 U.S. at 727, 728, 81 S.Ct. 1639.
¶10Our cases largely leave it to the trial judge to determine a prospective juror’s bias, which is “often a question of demean- or.” State v. Storey, 901 S.W.2d 886, 894 (Mo. banc 1995) (citing State v. Schneider, 736 S.W.2d 392, 403 (Mo. banc 1987), cert. denied, 484 U.S. 1047, 108 S.Ct. 786, 98 L.Ed.2d 871 (1988). This standard makes the trial judge’s discretion virtually undis-turbable because demeanor is not subject to appellate review. Especially with this deferential standard, we ought to review carefully not just the general notion of bias, but whether the trial court properly ascertained whether jurors had in their heads facts about the case that might form part of the basis for their verdict. To be specific, we cannot tell from this record whether some of the jurors came to court with the information that Barton is the man who killed his former landlady because she evicted him. The jurors were asked if they could set aside what they had heard or read - without inquiry as to what it was. Common sense tells us that it is probably not humanly possible to set aside these facts, especially where a person has no reason to believe that the “fact” about motive is false.
¶11In this case, the risk that the trial process was tainted with extraneous evidence was sufficiently great that I would find it an abuse of discretion not to allow, at the very least, individual questioning of prospective jurors to ascertain the extent of their knowledge of matters that properly were not evidence in the case so as to ensure a panel as free as possible of factual taint and prejudging disposition. Such questioning would also have provided a basis for soundly determining whether the defense motion for change of venue or continuance should have been granted. Barton should be given a new trial.
¶12. Viveca Novak, The Cost of Poor Advice, Time, July 5, 1999, at 38. See also, Carolyn Tuft, Ex-death Row Inmates Attack Capital *31Punishment, Courts, St. Louis Post Dispatch, November 16, 1998, at A-l.