RYAN, J., delivered the opinion of the court, in which KENNEDY, NELSON, BOGGS, NORRIS, SUHRHEINRICH, SILER, and BATCHELDER, JJ., joined. MERRITT, J. (pp. 389-396), delivered a separate dissenting opinion, in which MARTIN, C. J., DAUGHTREY, COLE, CLAY, and GILMAN, JJ., joined. MOORE, J. (pp. 396-397), delivered a separate dissenting opinion.
OPINION
RYAN, Circuit Judge.Lawrence DeLisle was convicted in a Michigan state court of the premeditated murder of his four children and the attempted murder of his wife. When his appeals within the Michigan appellate courts proved fruitless, he filed an application for a writ of habeas corpus, pursuant to 28 U.S.C. § 2254, alleging numerous constitutional defects in his conviction. The district court denied De-Lisle’s application. DeLisle now renews his argument that he was denied due process, alleging: (1) some of the members of the jury that convicted him were biased; (2) the evidence against him was insufficient to prove premeditation and intent; and (3) the trial court improperly denied his request for a bench trial. For the reasons that follow, we affirm the judgment of the district court.
I.
A.
On August 3, 1989, shortly after 9:00 p.m., with his wife and four children as passengers, Lawrence DeLisle drove the family’s station wagon down Eureka Road in Wyan-dotte, Michigan, at an increasingly high rate of speed, through a barrier of two wooden posts, and into the Detroit River. Although DeLisle and his wife, Suzanne, survived the incident, all four of their children — Bryan, Melissa, Kathryn, and Emily — drowned.
A few days later, at the request of Wyan-dotte police, DeLisle agreed to take a polygraph test, and accordingly, at about 7:30 a.m. on August 10, he was taken to a Michigan State Police facility. The exam began shortly after 10:00 a.m., and continued until about 5:30 p.m., with a break of an hour or so for lunch. These sessions were videotaped. At about 6:00 p.m., DeLisle was arrested and taken back to the Wyandotte police station. An audiotaped interrogation then began at approximately 10:45 p.m., and lasted until 1:00 a.m. During this interrogation, DeLisle “made statements which purportedly consti*373tute a confession that he intentionally drove the family car into the river.” People v. DeLisle, 183 Mich.App. 713, 455 N.W.2d 401, 402 (Mich.Ct.App.1990). His statements were rambling and confused, but representative of the most arguably inculpatory portions are the following:
I had a little cramp and I just egged it on____ As I was going down I just couldn’t slam on the brakes. I didn’t want to....
I just wanted to give, um, just scare my wife enough to slam on the brakes, come to a skidding halt and get her all upset— Just let me be. I couldn’t stop accelerating. I didn’t.
When asked what he thought “should happen to [him] ... now,” DeLisle replied, “Electrocution. Gas chamber, hang me. I don’t care. I don’t deserve to live.”
In addition to his statements regarding the August 3, 1989, incident, DeLisle told police that eight years earlier he had attempted to blow up his home by leaving a candle burning near a gas leak in the basement while his wife and son were asleep. He stated that he wanted to “[j]ust blow up everything in [his] past.” His statements about this incident included the following exchange:
[Q]: Did you know your little baby was in there?
[A]: Yes.
[Q]: And you know that the thing would' have blown your baby up with it ... ?
[A]: That’s why I stayed for a half hour. Because I wanted to go.
On the day following his interrogation, De-Lisle was arraigned on four counts of first-degree murder and one count of attempted first-degree murder, pursuant to Mich. Comp. Laws §§ 750.316, 750.91. DeLisle’s admissions were widely publicized by the media, which based the stories on the accounts of Wyandotte police representatives. On August 12, 1989, newspapers around the country reported that DeLisle had confessed to intentionally driving his family into the Detroit River, and had been arraigned on multiple murder charges. Local newspapers later reported the story under such headings as “Prosecutor: DeLisle tried twice.” And an article published in the November 10, 1989, edition of the Detroit Free Press stated that, according to the prosecution and the police, DeLisle had confessed both “to trying to blow up his house in 1981 while his wife and infant son slept,” and to “purposely driving] into the river to ease himself of financial burdens.”
At the preliminary examination, in making the decision to bind DeLisle over for trial, the district judge relied in part on the videotape and audiotape of the August 10 interrogations. The judge, however, did not make those exhibits part of the record, as would usually have been done. Instead, the judge reviewed them in camera.
Not satisfied with secondhand police accounts of DeLisle’s statements, certain members of the media filed suit demanding access to these exhibits. Both the prosecution and DeLisle resisted this demand, but in September 1989, one judge of the Detroit Recorder’s Court' — which at the time had jurisdiction— concluded that the First Amendment required that the media have access to the exhibits in question, and that a release would not pose “any great potential risk” to the defendant’s right to a fair trial.
DeLisle’s attorney immediately obtained a stay of the Recorder’s Court order from the Michigan Court of Appeals, which court then remanded the case to the Detroit Recorder’s Court to allow a record to be developed. The order further provided that the state district court judge in Wyandotte who had conducted the preliminary examination should make specific factual findings as to the extent to which he relied on the video- and audiotapes in binding DeLisle over for tidal, and as to whether a release would prejudice DeLisle’s right to a fair trial. In October 1989, accordingly, the state district court issued findings that it had relied on the exhibits “in determining whether or not the crimes charged had been committed and whether there was probable cause to believe the Defendant committed them,” but that the tapes “contained] statements prejudicial to the Defendant,” which statements “may be inadmissible at the trial.” It further concluded that “[t]here is a substantial probabili*374ty that the Defendant’s right to a fair trial would be prejudiced by public airing of [the tapes] before trial,” and that “[Reasonable alternatives to closure would not adequately protect the Defendant’s fair trial rights because of the high degree of publicity of this case and because the material which the Court has sealed are statements or confessions of the Defendant ... which may be inadmissible at trial.”
Again, however, the Recorder’s Court judge disagreed, holding that the district court’s order was “conclusory” and not sufficiently specific to merit a prior restraint. It was the Recorder’s Court judge’s view that “[notwithstanding extensive pretrial publicity, empirical data demonstrates that there remains a high probability of seating a jury capable of rendering a fair and impartial verdict.”
DeLisle appealed, and in late November 1989, the Michigan Court of Appeals vacated, in part, the Recorder’s Court’s order. There was, the court of appeals held, no constitutional requirement that the media have unlimited access to or opportunity to copy the tapes or transcripts of the tapes. However, there was a right on the part of the media “to attend criminal proceedings, including preliminary examinations, and to report what they have observed.” Further, the court concluded that DeLisle had not shown “that there is a substantial probability that his right to a fair trial w[ould] be prejudiced by publicity that closure would prevent, or that reasonable alternatives to closure would not adequately protect his fair trial rights.” Therefore, the court of appeals ruled, “The Recorder’s Court shall ... allow limited public access to the exhibits in a manner reasonably calculated to put interested members of the public in the position they would have been in if the district court had reviewed the exhibits during the preliminary examination in open court.” The Michigan Supreme Court denied leave to appeal.
While the media lawsuit was moving between the Recorder’s Court and the Michigan Court of Appeals, the Wayne County Circuit Court assumed jurisdiction of the prosecution, pursuant to the state district court bind-over. DeLisle had moved in a timely fashion to suppress the fruits of the inteiTogation, and on December 21, 1989, more than six months before the ultimate June 1990 trial date, the circuit judge agreed that DeLisle’s statements to the police had not been made voluntarily. Accordingly, he ordered their suppression. In a pretrial appeal taken in early January 1990, the trial court’s decision was affirmed by the Michigan Court of Appeals. DeLisle, 183 Mich.App. 713, 455 N.W.2d 401. The court noted that “the length of defendant’s interrogation strongly suggests that his statement was not made voluntarily”; that DeLisle’s “emotional state going into the interrogation was very poor due to the death of his four children just seven days before”; and that DeLisle “had only a tenth grade education and no previous experience with the criminal justice system.” Id. at 403. The court further “found that the techniques used by the interrogators were likely to produce a false or untrustworthy statement,” although declining to go so far as to find that DeLisle “was actually hypnotized,” as DeLisle had argued. Id. In sum, the totality of the circumstances led the court of appeals to conclude that DeLisle’s statements on August 10 were untrustworthy. The Michigan Supreme Court denied leave to appeal. People v. DeLisle, 447 Mich. 987, 525 N.W.2d 445 (Mich.1994).
Notwithstanding the suppression of DeLi-sle’s statements for purposes of his trial, the result of the earlier ruling of the Michigan Court of Appeals in the parallel media lawsuit was that on January 10, 1990, excerpts from tapes of DeLisle’s interrogation were played to the interested public, including media representatives. The parties give no particular attention in their briefs to an account of this proceeding. And contrary to the dissent’s characterization of this proceeding as a “judicial press conference,” it appears that nothing occurred beyond a straightforward playing of the tapes. We find no record of any commentary rendered by any member of the Michigan judiciary in connection with the tapes; indeed, there is no indication that any judge was even in attendance during the proceeding. And the playing of the tapes did not occur in the court in which DeLisle was eventually tried; the Michigan Court of Ap*375peals order directing that “interested members of the public [be put] in the position they would have been in if the district court had reviewed the exhibits during the preliminary examination in open court,” was addressed to the Detroit Recorder’s Court, not to the Wayne County Circuit Court, which presided over DeLisle’s criminal trial. Finally, the playing of the tapes occurred on a single occasion only and, as the Michigan Court of Appeals had ordered, copies were not made available, thus assuring that the tapes could not be broadcast by the electronic media at all. As far as we can glean from the record, the Detroit Recorder’s Court did precisely what the Michigan Court of Appeals directed it to do, and disseminated the information in a circumscribed manner.
Nevertheless, on January 11, 1990, the day following the playing of the tapes, The Detroit News and the Detroit Free Press published lengthy transcripts of DeLisle’s interrogation in articles headlined, “I don’t deserve to live” and “DeLisle tells of torment.”
B.
In late January 1990, the trial court made its first attempt to seat a jury for DeLisle’s trial. Recognizing that the pretrial publicity had been widespread and potentially prejudicial to the defense, the trial court permitted prosecution and defense attorneys to prepare a lengthy questionnaire to be completed by prospective jurors. On January 25, 1990, after a panel of jurors had completed the questionnaire, the defense filed a motion for change of venue. The trial court deferred decision on the motion, pending a determination whether an impartial jury could be seated. The initial jury venire that had been summoned was dismissed, however, when the state’s appeal of the trial court’s suppression ruling caused a stay of DeLisle’s trial.
Four months later, on May 31, 1990, a second panel of jurors convened to complete questionnaires. The questionnaire contained the following caveat:
Upon completing this questionnaire, there is a possibility that you will be selected to be a juror in this ease. There has been, and continues to be, a substantial amount of news reports and other publicity about this case. You are instructed not to read anything or listen to or watch any broadcasts concerning Mr. DeLisle or pertaining to this trial until such time as you are dismissed as a potential juror in this case. Media accounts may be inaccurate or incomplete and may contain matters which are intended to influence you as a prospective juror.
The defense again moved for a change of venue, and the trial court again deferred a ruling on the motion until it could determine whether an impartial jury could be seated. The trial court then conducted an extensive four-day voir dire of potential jurors. Each juror was questioned individually, rather than en masse, by the court, the prosecutor, and defense counsel as to the prospective juror’s knowledge of the case from media reports, and as to the answers given to the juror questionnaire. Prosecution and defense attorneys were given virtually unlimited opportunity to question prospective jurors about their knowledge of the ease, and their ability to render a fair and impartial verdict. Occasionally, the court instructed prospective jurors, after the voir dire, “Don’t read any newspaper stories about this case. Don’t watch or listen to any TV accounts or reports about this case.”
This continued until 41 potential jurors were passed for cause, a number sufficient to encompass all of the peremptory challenges to which the parties were entitled plus an additional 14 jurors — 12 plus two alternates — to constitute the petit jury. A total of 68 potential jurors were questioned over the four days. The court excused 16 venire-persons for cause, sua sponte. Of these 16, seven were excused because they admitted that they had made up them mind or believed that DeLisle was guilty. Five were excused for hardship; two were excused because they said they could not be fair when children were involved; one was excused because he had made plans to visit the scene of the accident; and one was excused because he did not believe he could be fair because he had been a police officer for 28 years. The comb also excused one juror the prosecution had challenged for cause, who could not presume DeLisle innocent.
*376Over the course of the voir dire, DeLisle’s attorney challenged an additional 15 venire-persons for cause. The trial court granted 10 of these 15 challenges. Of those excused on DeLisle’s challenges for cause, only one clearly expressed a belief that DeLisle was guilty; a few doubted that the tragedy could have been an accident; one was excused because he could not accept the burden of proof; and a few admitted that they did not think they could be fair. Most of the 10 excused on the defense’s challenges were excused because they indicated that they would have some trouble putting aside what they had heard or read about the case. De-Lisle’s counsel passed 37 jurors without challenge for cause, including all 12 of the jurors who would eventually hear the case and participate in the verdict.
At the beginning of the third day of voir dire, on June 8, 1990, DeLisle renewed his request for a change of venue. His attorney noted that 23 of the 35 jurors that had been passed to that point, including those passed over DeLisle’s challenge, had some knowledge of the suppressed statements. The trial court restated its desire to “reserve ruling on [the] motion.” DeLisle’s attorney then stated:
I want to make it clear on the record that I would challenge for cause any juror who has knowledge of the confession. My problem is, if I identify them now and challenge them by name solely on then-knowledge, the Court will deny that challenge and I will disclose my list before the Prosecutor exercises a peremptory challenge. Those jurors that I feel have the knowledge in this case not to be fair [sic]. So, I would ask if I be permitted to file that list subsequent to jury selection if we proceed to that. And identify those jurors that I am challenging on that basis.
(Emphasis added.) The court replied that DeLisle would “be permitted to do that.”
On June 11, 1990, the defense exercised 10 of its 15 peremptory challenges and the state exercised six of its 12.
After the jury was seated, the trial court denied DeLisle’s motion for change of venue. In denying the motion, the court noted the extraordinary effort it had exerted to seat an impartial jury, and further noted:
There is no evidence that I have been able to find to indicate there is strong community feeling or a bitter prejudice towards the Defendant. I think the majority of the people that we interviewed indicated that they had an open mind on the issue.
It was interesting to me that of the people that we interviewed, the 68 people we interviewed to get to our number of 41, only about 10 indicated that there was a prejudice that would prevent them from deciding the case.
The other people were excused for reasons like, it would be a financial hardship for them to be on this jury, or they didn’t understand the principles of law that they would be required to apply to this particular case.
In addition I think it’s important to note that there has been some time since this incident occurred. We’re now ten months after the fact. A lot of the jurors noted that their memories had faded with respect to the events about this case, so I think that the time factor is an important consideration here.
The court also noted that the defense had not used all of its peremptory challenges. We note in passing that under Michigan law, a party’s failure to exercise all of its peremptory challenges may constitute a waiver of the party’s right to challenge the impartiality of the jury. See People v. Taylor, 195 Mich.App. 57, 489 N.W.2d 99, 101 (Mich.Ct.App.1992). However, the Michigan Court of Appeals did not treat DeLisle’s failure to use all of his peremptory challenges as a procedural default, see People v. DeLisle, 202 Mich.App. 658, 509 N.W.2d 885, 891 n. 4 (Mich.Ct.App.1993), and, accordingly, neither will we, see McBee v. Grant, 763 F.2d 811, 813 (6th Cir.1985).
All 12 of the jurors who eventually determined DeLisle’s guilt admitted that they had heard about the case before being called for jury service. Significantly, however, none of the jurors had an opinion regarding DeLi-sle’s guilt. Indeed, each of the jurors specifically denied that he or she had formed any opinion, even a tentative one, regarding De-*377Lisle’s guilt. All of the eventual jurors assured the court that they could presume DeLisle innocent, and that they would base them decision on the evidence presented at trial.
Seven of the jurors admitted to no more than some recollection of the accident or of DeLisle’s claim that the accident was caused by a leg cramp or a problem with his ear. For the most part, these jurors indicated that they had not heard much about the case, that they had forgotten about the case, or that their memories of the case had faded. Five jurors, however, indicated that they had some knowledge of the suppressed statements. One of these jurors remembered only that DeLisle had changed his mind regarding whether he “did it.” Three jurors remembered that DeLisle had been pressured to confess, and one was aware of DeLi-sle’s statements regarding both the incident for which he was being tried and the 1981 “candle incident.”
Once the jury was empaneled, after instructing the jurors not to talk to anyone about the case, the court firmly and unambiguously instructed the jury to avoid any press coverage of the trial:
The second thing that I really want you to be very conscious of is newspaper articles and television and radio comments about this case. Now, I suggest to you that you not even read the newspaper or listen to radio news reports or TV news during the trial. And that way you can be certain that you won’t even be tempted to hear or see something you should not hear or see.
But, again, this is a very important rule that you all must follow, because as the attorneys have explained to some of you and I’ll explain to you again, what the media reports is not evidence. And it is not subject to certain rules that we have for evidence^] that is[,] cross-examination.
So if you read or listen to or watch media reports of this trial, you may be getting misleading information that is not fair to one or both sides, and to which both sides cannot respond. So please, again, this is a very important rule and I want you to follow it.
C.
DeLisle’s trial began on June 11, 1990. The state’s first witness, Beverly Lake, testified that she lived in an llth-floor condominium unit at the eastern end of Eureka Road in Wyandotte. She explained that Eureka Road runs west to east and dead-ends to the east at the Detroit River. She also noted that, at the eastern end of Eureka Road, between the pavement and the bank of the river, two wooden posts served as a barricade to traffic.
Lake testified that she was on her balcony with her parents on August 2,1989, the night before the drownings, when she first saw the DeLisles’ station wagon. She testified that she saw the DeLisles’ car being driven “at an extremely slow pace” to the very end of Eureka, and then watched it turn around and head away from the river toward Biddle Avenue, which runs north to south and intersects Eureka to the west. Lake took particular notice of the DeLisles’ car because (1) it was moving so slowly; (2) there had been a rash of car thefts and vandalism in thé area; and (3) Eureka is not a through-street, and she therefore usually recognized passing cars. Lake noted at least two occupants in the car:, a heavyset white woman in the passenger seat and a short-haired white male driver.
The next evening, August 3, 1989, at approximately 9:00 p.m., Lake was again on her balcony with her parents when she saw the same car with the same driver and passenger return. She testified that she “recognized them specifically.” She stated that the De-Lisles’ car again drove “extremely slowly ... to the wooden barricade” at the very end of the street, and after “a considerable pause, probably 30 to 60 seconds,” the car turned around, and headed west away from the river. According to Lake, the car returned about 20 minutes later, accelerated rapidly down the street, crashed through the posts, and disappeared into the river. Lake testified that she heard the car “shift[] into passing gear”; that the car proceeded in a straight path between the barricades without “swerving” or “deviating]”; and that there were no apparent signs of braking.
*378The state presented several other witnesses who confirmed that the DeLisles’ station wagon drove straight down the middle of Eureka Road, at a high rate of speed, accelerating, and not swerving or braking. Additionally, Lake’s mother, Goldie Hunt, confirmed Lake’s account of the events on August 2, 1989. She agreed that the car she observed on that evening was the same car that plunged into the Detroit River the following night.
The DeLisles’ car struck the “right third” of the left of the two barricade posts at the end of Eureka. It also clipped a sign attached to the right post, though not the post itself. Irving Rozian, an engineer and an expert in accident reconstruction, testified that a slight indentation on the left front wheel suggested that the car struck the curb at the end of the road with a slight angle to the right. He explained that this angle was the likely result of a slight change of direction which would not have been observed as a “swerve.” Rozian testified that this change of direction would be consistent with an attempt to drive between the barricade posts if the DeLisles’ car was toward the left side of the road as it approached the end of the street. Although he stated that it could have been accomplished by taking a more perfectly straight path, Rozian acknowledged that someone trying to split the protective posts at the end of Eureka could not hope to be much more successful than DeLisle.
Another expert in accident reconstruction, Sergeant Weldon Greiger, testified that the car traveled some 40 feet from the curb to impact in the river, and that, in order to do so, the car would have had to have been traveling at least 40 to 47 miles per hour. Greiger estimated that, at maximum acceleration, it would have taken 1% seconds and 300 feet for the DeLisles’ car to reach a speed of 45 miles per hour. Greiger further testified that, in post-accident tests, the DeLisles’ car brakes functioned properly. Specifically, he stated that while maintaining full acceleration at 45 and 55 miles per hour, the brakes had succeeded in bringing the car to a stop. Finally, Greiger added that he had found no skid marks at the scene of the accident. This fact was confirmed by Sergeant Daniel Galeski.
Bryan Ross was seated in his boat on the Detroit River and saw the DeLisles’ car go into and under the water. Ross testified that the ear went under in a matter of seconds and that DeLisle surfaced quickly “right above where the ear had gone under.” Ross added that DeLisle did not initially say anything and that he never went under the water again. Rather, “[DeLisle] was just sitting there treading water.” Suzanne De-Lisle, on the other hand, “popped up downriver” and was “hysterical.” Ross testified that “[s]he was spitting water and started screaming, ‘My babies,’ and then she’d go back under again.”
Eric Stoneburner, another boater, eventually pulled DeLisle and Suzanne from the river. He testified that DeLisle did not say anything when he was pulled into the boat, but that Suzanne “was screaming for her babies.” He also remembered that, after Suzanne was pulled into the boat, the DeLi-sles embraced and DeLisle “said something to th[e] effect, ‘Oh my God, what have I done.’ ”
D.
DeLisle’s defense was that the death of his children was a tragic accident caused by a leg cramp and an engine throttle that became stuck. DeLisle did not testify, and his account of the accident was introduced primarily through the testimony of his wife, Suzanne.
Suzanne’s testimony concerning the events of August 2 and 3, 1989, was as follows: On August 2, 1989, the family had been out for a pleasure drive and had driven down Eureka Road in an attempt to reach a nearby park. The next evening, August 3, 1989, the family was out running errands and, at the request of their daughter Melissa, decided to go to the river to watch the boats. Although the DeLisles had not been able to reach the park by way of Eureka on the previous evening, DeLisle again drove down Eureka. After crossing Biddle, he pulled over and stopped the car because his leg was bothering him. He rubbed his right calf for a few minutes *379and the family decided to return home because their infant was crying loudly.
DeLisle then drove toward the water, east down Eureka, turned around, and proceeded to a drug store at the corner of Biddle and Eureka. After leaving the drug store, DeLi-sle drove down an alley and, finding it difficult to make a right turn to head home, again pulled onto Eureka heading toward the river.
Suzanne reminded her husband that he could turn around in a nearby driveway. De-Lisle suddenly grabbed his leg and screamed that he was having a cramp. The car began to accelerate, and DeLisle grabbed his calf with his right hand to “try[ ] to pull his right foot off of the accelerator.” Both Suzanne and her husband tried to steer the car with their left hands, and the “wheel was moving-back and forth.” Although Suzanne claimed that DeLisle succeeded in pulling his foot off of the accelerator, she did not remember the car ever slowing down.
Suzanne testified further that the car’s engine had “raced” in the past, including some times while the car was in drive. She adhered to this testimony even when she was reminded of her previous deposition testimony that the “racing” had only occurred when the car was in neutral. Both parties presented evidence regarding the mechanical condition of the car and, in particular, the throttle. Two expert witnesses, Weldon Greiger and Jon Yager, testified that during one test of the DeLisles’ car, the engine “raced” while in neutral. The engine eventually developed significant problems and the “racing” could not be duplicated. One defense witness, mechanic James Cokewell, testified that the car’s accelerator cable was bent, its engine mounts were cracked, and one of its throttle plates appeared to stick. According to Coke-well, the DeLisles’ vehicle was “dangerous” because these problems could contribute to excess or uncontrollable acceleration.
E.
On the second day of the trial, DeLisle moved to waive the jury in favor of a bench trial. The prosecution objected and the trial court formally denied the motion. Following less than two days of deliberation, the jury convicted DeLisle on all counts and he was sentenced to five concurrent terms of life imprisonment.
On direct appeal to the Michigan Court of Appeals, DeLisle alleged numerous due-process violations. In particular, he argued that the evidence against him was insufficient to prove premeditation or intent to kill. He also challenged the trial court’s denial of his motion for a bench trial, his motion for a change of venue, and his motion to challenge for cause all jurors with knowledge of the suppressed statements. The court of appeals readily concluded that the evidence was sufficient to sustain DeLisle’s conviction, DeLisle, 509 N.W.2d at 888, and on the authority of People v. Kirby, 440 Mich. 485, 487 N.W.2d 404 (Mich.1992), it rejected DeLisle’s argument that the trial court erred by not granting DeLisle’s request for a bench trial, DeLi-sle, 509 N.W.2d at 893. The court of appeals did not specifically address DeLisle’s argument regarding the blanket challenge for cause. It did, however, consider at length the denial of DeLisle’s motion for change of venue. The court understood DeLisle’s argument to be, essentially, that “the jury ... was exposed to [his] inadmissible ‘confession’ through the media to such a great extent that they [sic] could not be expected to set aside such knowledge.” Id. at 889. The court rejected this argument, finding, “after a thorough review of the record, ... no reason to disregard the jurors’ assurances that they could render a fair and impartial verdict in [the] case.” Id'In particular, the court found that only 21 of 68 potential jurors were excused for bias, and noted that all the jurors who heard the case indicated “that their memories of the case were vague because of the lapse of time.” Id. at 891. The court also found, after reviewing the pretrial publicity that has been made part of the record, that, although “[t]he tone of the news articles became inflammatory shortly after the incident” and “[t]here was ... considerable coverage regarding defendant’s confession,” the “inflammatory news coverage had basically stopped after January 17,1990,” id. at 892, more than four months before the trial.
*380After the Michigan Supreme Court denied DeLisle’s application for leave to appeal, DeLisle, 447 Mich. 987, 525 N.W.2d 445, DeLi-sle filed the present application for a writ of habeas corpus in the United States District Court for the Eastern District of Michigan. In his petition, DeLisle renewed his challenge to the sufficiency of the evidence and to the trial court’s denial of his request for a bench trial. Although DeLisle did not persist with his challenge to the trial court’s denial of his motion for change of venue, and has abandoned it for purposes of this appeal, he reasserted his claim that he was denied due process because the trial court denied a blanket challenge for cause to any venireper-son with knowledge of the suppressed pretrial statements.
The district court agreed with the Michigan Court of Appeals that the evidence against DeLisle was sufficient to permit a rational jury to find DeLisle guilty as charged. It also agreed that DeLisle had no constitutional right to a bench trial. Finally, the district court rejected DeLisle’s argument that any prospective juror with knowledge of his suppressed statements should have been dismissed for cause. The district court found that DeLisle’s right to a fair trial had been adequately safeguarded by the thorough voir dire and the use of the questionnaire. It explained that “each juror[] indicat[ed] that they [sic] would base[] the verdict on the evidence, ... no juror expressed] a predisposition to find [DeLisle] guilty,” and “only 15% of the prospective jurors interviewed in this case indicated a level of prejudice which would prevent them from deciding the case.” The district court concluded that “[b]ased on the record in this case, [it] [could not] find that the jurors ... ‘had such fixed opinions that they could not judge impartially’ [DeLisle’s] guilt.”
II.
A.
A federal court’s consideration of an application for a writ of habeas corpus, filed by a prisoner incarcerated pursuant to the judgment of a state court, is governed by 28 U.S.C. § 2254. When reviewing a district court’s disposition of such an application, we review the district court’s legal conclusions de novo, and any findings of fact for clear error. McQueen v. Scraggy, 99 F.3d 1302, 1310 (6th Cir.1996), cert. denied, — U.S. -, 117 S.Ct. 2422, 138 L.Ed.2d 185 (1997).
The provisions of section 104 of the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), Pub.L. No. 104-132, 110 Stat. 1214 (1996), served to amend section 2254. See 28 U.S.C. § 2254(d), (e)(1) (Supp.1997). These amendments were signed into law on April 24,1996, and they do not apply to “applications in noncapital cases that were already pending when the Act was passed.” Lindh v. Murphy, 521 U.S. 320, -, 117 S.Ct. 2059, 2061, 138 L.Ed.2d 481 (1997). DeLisle’s application for a writ of habeas corpus was filed on March 8, 1995. Accordingly, we consider his application under pre-AEDPA standards.
B.
The state argues that all of the issues DeLisle has presented in his application for a writ of habeas corpus were raised in his direct appeal to the Michigan Court of Appeals and that “he may not seek to relitigate [on habeas ] the points resolved against him on direct appeal without violating this Court’s obligation to give full faith and credit to the judgment of a state court.” We understand the state’s argument to be an attempt to extend the rule of Stone v. Powell, 428 U.S. 465, 96 S.Ct. 3037, 49 L.Ed.2d 1067 (1976), to DeLisle’s challenge to the sufficiency of the evidence and the impartiality of his jury.
The United States Supreme Court has explained that Stone was a “prudential” decision, based in large part on the fact that the exclusionary rule “ ‘is not a personal constitutional right,’ ” and it has “repeatedly declined to extend the rule in Stone beyond its original bounds.” Withrow v. Williams, 507 U.S. 680, 686-87, 113 S.Ct. 1745, 123 L.Ed.2d 407 (1993) (citation omitted). Specifically, the Court has expressly refused to extend Stone to due-process challenges to the sufficiency of the evidence. See Jackson v. Virginia, 443 U.S. 307, 323, 99 S.Ct. 2781, 61 *381L.Ed.2d 560 (1979). We have no doubt, following the decisions in Withrow, 507 U.S. at 687-95, 113 S.Ct. 1745, Kimmelman v. Morrison, 477 U.S. 365, 375-83, 106 S.Ct. 2574, 91 L.Ed.2d 305 (1986), and Rose v. Mitchell, 443 U.S. 545, 559-66, 99 S.Ct. 2993, 61 L.Ed.2d 739 (1979), that, where the fundamental and personal constitutional right to an impartial jury trial is at issue, federal habeas corpus review remains available notwithstanding a full and fair opportunity to litigate the issue in state court. Accordingly, we now turn to the merits of DeLisle’s application.
C.
1.
DeLisle argues that the district court erred in holding he was not denied due process when the state trial court denied his blanket challenge for cause to all venireper-sons with knowledge of the suppressed statements. Although the Michigan Court of Appeals failed to address this argument, it is clear that DeLisle raised the argument both in his brief on direct appeal and in his application for leave to appeal to the Michigan Supreme Court. Accordingly, he has met the exhaustion requirement of the habeas statute. See Smith v. Digmon, 434 U.S. 332, 333-34, 98 S.Ct. 597, 54 L.Ed.2d 582 (1978); Harris v. Rees, 794 F.2d 1168, 1174 (6th Cir.1986).
Nonetheless, as another preliminary matter, we note that it is not at all clear that DeLisle actually properly preserved in the trial court his present argument for appeal. In support of his claim that he made a blanket challenge to all venirepersons with knowledge of the suppressed statements, De-Lisle refers this court to his attorney’s June 8,1990, statement “that [he] would challenge for cause any juror who has knowledge of the confession,” and his attorney’s request that he “be permitted to file ... subsequent to jury selection if we proceed to that,” a list of “[tjhose jurors that ... have the knowledge in this case not to be fair [sic].” (Emphasis added.) We are satisfied, on a careful reading of the transcript, that DeLisle’s attorney never actually made this blanket challenge; rather, he merely requested that he be permitted to do so later in the trial. Although the trial court acceded to DeLisle’s request, we find no indication in the record that DeLi-sle ever followed through with a motion or filed a list of the jurors he wished to challenge. We are, however, handicapped in coming to any conclusion on this point because the Michigan Court of Appeals did not address this important factual issue, and the district court appears to have simply assumed that DeLisle’s attorney preserved this argument. Nonetheless, we need not resolve this undeveloped issue of procedural default because we are satisfied that DeLisle’s argument regarding the blanket challenge must be rejected on the merits.
In considering DeLisle’s arguments, it is crucial to understand what he does not argue. First, he expressly disavows his previous claim that the trial court’s denial of his motion for change of venue resulted in a denial of due process. He explains that he has abandoned this argument because, he says, he does not want this court to be distracted from his claim that five jurors had knowledge, to some degree or other, of his suppressed statements. Second, DeLisle does not now argue, and has not at any point argued, that particular jurors actually revealed a disqualifying animus or unfavorable predisposition toward him during voir dire. Indeed, in his primary brief on appeal, DeLi-sle does not make a single, specific reference to the voir dire examination of the venireper-sons who eventually served on his jury. Rather, he relies only indirectly on the voir dire examination to establish which jurors had knowledge of the suppressed statements. He then asks this court to presume that the “ability [of the jurors with such knowledge] to fairly and impartially consider only the evidence presented at trial was fatally tainted.” Thus, DeLisle argues that we must hold as a matter of law that no juror with knowledge of the suppressed statements was capable of being impartial. He argues, in essence, that we must presume bias on the part of these jurors, and he takes the position that the absence of demonstrable evidence of bias is not fatal.
That DeLisle’s argument is based on a presumption of prejudice, rather than any *382evidence of juror hostility, is clear from the fact that DeLisle’s claim is premised on the denial of a blanket challenge for cause. As we have indicated, during voir dire DeLisle made successful use of a number of challenges for cause, on various grounds including bias. In fact, the trial court granted 10 of DeLisle’s 15 challenges. Obviously, DeLi-sle quite clearly understood his right to challenge venirepersons he thought displayed bias. And, given that DeLisle passed each and every member of the eventual jury for cause, we are not surprised that he now takes a “presumptive,” rather than an “in fact,” approach to his allegation of juror bias. We observe, however, that “[t]he fact that petitioner did not challenge for cause any of the jurors ... selected is strong evidence that he was convinced the jurors were not biased and had not formed any opinions as to his guilt.” Beck v. Washington, 369 U.S. 541, 557-58, 82 S.Ct. 955, 8 L.Ed.2d 98 (1962).
In any event, we have no question that DeLisle’s claim implicates due-process concerns. “[T]he right to jury trial guarantees to the criminally accused a fair trial by a panel of impartial, ‘indifferent’ jurors. The failure to accord an accused a fair hearing violates even the minimal standards of due process.” Irvin v. Dowd, 366 U.S. 717, 722, 81 S.Ct. 1639, 6 L.Ed.2d 751 (1961). Our standard of review in this habeas proceeding does not permit us to substitute our view of possible juror bias for the state court’s view; we may only overturn the state court’s findings of juror impartiality if those findings were manifestly erroneous. See Mu’Min v. Virginia, 500 U.S. 415, 428, 111 S.Ct. 1899, 114 L.Ed.2d 493 (1991); see also Patton v. Yount, 467 U.S. 1025, 1031, 104 S.Ct. 2885, 81 L.Ed.2d 847 (1984); Irvin, 366 U.S. at 723, 81 S.Ct. 1639.
Beginning with Irvin v. Dowd, the United States Supreme Court has on occasion presumed prejudice on the part of the jury when a defendant has been able to “show[] the actual existence,” albeit unexpressed, “of such an opinion in the mind of the juror as will raise the presumption of partiality.” 366 U.S. at 723, 81 S.Ct. 1639. But even when a defendant’s due-process claim is grounded on an allegation that bias must be presumed, as opposed to being grounded on actual evidence of bias, our inquiry is the same: “Of course there could be no constitutional infirmity ... if [the defendant] actually received a trial by an impartial jury. Hence, our inquiry is addressed to that subject.” Beck, 369 U.S. at 556, 82 S.Ct. 955.
The category of cases where prejudice has been presumed in the face of juror attestation to the contrary is extremely narrow. Indeed, the few cases in which the Court has presumed prejudice can only be termed extraordinary, see generally Murphy v. Florida, 421 U.S. 794, 798-99, 95 S.Ct. 2031, 44 L.Ed.2d 589 (1975), and it is well-settled that pretrial publicity itself — “even pervasive, adverse publicity — does not inevitably lead to an unfair trial,” Nebraska Press Ass’n v. Stuart, 427 U.S. 539, 554, 96 S.Ct. 2791, 49 L.Ed.2d 683 (1976). As the Court observed in Murphy, “indicia of impartiality” on the part of a jury have thus far been disregarded only in those cases “where the general atmosphere in the community or courtroom is sufficiently inflammatory.” 421 U.S. at 802, 95 S.Ct. 2031. Thus, it is beyond question that mere prior knowledge of the existence of the case, or familiarity with the issues involved, or even some preexisting opinion as to the merits, does not in and of itself raise a presumption of jury taint; such a standard would be certainly unsalutary, and likewise impossible to achieve:
It is not required ... that the jurors be totally ignorant of the facts and issues involved. In these days of swift, widespread and diverse methods of communication, an important case can be expected to arouse the interest of the public in the vicinity, and scarcely any of those best qualified to serve as jurors will not have formed some impression or opinion as to the merits of the case. This is particularly true in criminal cases. To hold that the mere existence of any preconceived notion as to the guilt or innocence of an accused, without more, is sufficient to rebut the presumption of a prospective juror’s impartiality would be to establish an impossible standard. It is sufficient if the juror can lay aside his impression or opinion and *383render a verdict based on the evidence presented in court.
Irvin, 366 U.S. at 722-23, 81 S.Ct.. 1639.
Although cases of alleged juror bias, particularly cases warranting an across-the-board presumption of bias, are sui generis, a review of the kind of cases in which the Supreme Court has made a presumption of bias is instructive. We look to four decisions of the Supreme Court to guide us in our task of determining whether it would be proper to presume prejudice on the part of the jury under the circumstances of DeLisle’s case: Irvin, 366 U.S. 717, 81 S.Ct. 1639; Rideau v. Louisiana, 373 U.S. 723, 83 S.Ct. 1417, 10 L.Ed.2d 663 (1963); Estes v. Texas, 381 U.S. 532, 85 S.Ct. 1628, 14 L.Ed.2d 543 (1965); and Sheppard v. Maxwell, 384 U.S. 333, 86 S.Ct. 1507, 16 L.Ed.2d 600 (1966). Taking these cases as illustrative of the circumstances the Supreme Court has found sufficiently exceptional to merit an imputation of prejudice, we conclude, as we shall explain, that these cases are so wholly distant from the one before us that we are compelled to conclude that DeLisle’s claim fails.
In hvin, the defendant was accused of the murder of a family of six, a crime that the Court described as “ ‘arous[ing] great excitement and indignation,’ ” 366 U.S. at 719, 81 S.Ct. 1639, and which became a cause célebre in a small community, id. at 725, 81 S.Ct. 1639. As the Court described it, the pretrial “build-up of prejudice,” due to an abundance of detailed newspaper articles, was “clear and convincing.” Id.
These stories revealed the details of his background, including a reference to crimes committed when a juvenile, his convictions for arson almost 20 years previously, for burglary and by a court-martial on AWOL charges during the war. He was accused of being a parole violator. The headlines announced his police line-up identification, that he faced a lie detector test, had been placed at the scene of the crime and that the six murders were solved but petitioner refused to confess. Finally, they announced his confession to the six murders and the fact of his indictment for four of them in Indiana. They reported petitioner’s offer to plead guilty if promised a 99-year sentence, but also the determination, on the other hand, of the prosecutor to secure the death penalty____
Id. at 725-26, 81 S.Ct. 1639.
The result of all this was a venire that exhibited far more indications of prejudice against the defendant than did the venire in this case against DeLisle. “The panel consisted of 430 persons. The court itself excused 268 of those on challenges for cause as having fixed opinions as to the guilt of petitioner. ...” Id. at 727, 81 S.Ct. 1639. Further, 90% of the prospective jurors had some opinion about guilt, “ranging in intensity from mere suspicion to absolute certainty.” Id. Indeed, eight of the 12 seated thought the defendant was guilty. See id.
These facts and figures bear no resemblance to DeLisle’s case. The trial court specifically found that of “the 68 people ... interviewed ..., only about 10 indicated that there was a prejudice that would prevent them from deciding the case.” That is, fewer than 15% had any fixed opinion about DeLi-sle’s guilt. Adding in all those challenged for cause who expressed more ambiguous sentiments arguably suggestive of preconceptions about DeLisle’s guilt, as determined by the Michigan Court of Appeals, the total reaches only 21, or fewer than a third. See DeLisle, 509 N.W.2d at 891; cf. Murphy, 421 U.S. at 796, 95 S.Ct. 2031; Beck, 369 U.S. at 556, 82 S.Ct. 955. Either way, it is a far cry from the 90% in Irvin. Moreover, and what is of greatest importance, none of those actually seated on DeLisle’s jury was challenged for cause, and none expressed a fixed opinion about DeLisle’s guilt, let alone the type of “fixed opinion” rightly decried in Irvin.
We are well aware that a “juror’s assurances that he is equal to this task cannot be dispositive of the accused’s rights.” Murphy, 421 U.S. at 800, 95 S.Ct. 2031. That is, the mere fact that the DeLisle venire failed to express the strength of sentiment that the venire under consideration in Irvin did, does not, of itself, mean that DeLisle’s claim necessarily fails. It does, however, tellingly distinguish this case from Irvin. As the Murphy Court rightly observed:
*384In a community where most veniremen will admit to a disqualifying prejudice, the reliability of the others’ protestations may be drawn into question; for it is then more probable that they are part of a community deeply hostile to the accused, and more likely that they may unwittingly have been influenced by it.
Id. at 803, 95 S.Ct. 2031. That is, if, as was the situation in Irvin, 90% of DeLisle’s panel had expressed fixed opinions about his guilt, while those actually selected to be jurors did not, the former fact would cast some doubt on the veracity of the jurors. On the other hand, the fact that the 12 jurors who tried the case here expressed no such belief is entirely consonant with the beliefs of the venire as a whole. And the Supreme Court has made clear that we cannot, as a general mattei", simply disregard a juror’s own assurances of his impartiality based on a cynical view of “the human propensity for self-justification.” See Smith v. Phillips, 455 U.S. 209, 215, 102 S.Ct. 940, 71 L.Ed.2d 78 (1982).
We turn next to Rideau, in which the defendant was araested a few hours after allegedly robbing a bank. “The next mox-n-ing a moving picture film with a sound track was made of an ‘interview’ in the jail between Rideau and the Sheriff.... This ‘intexview’ lasted approximately 20 minutes. It consisted of interrogation by the sheriff and admissions by Rideau that he had pex’petrated the bank robbery, kidnapping, and murder.” 373 U.S. at 724, 83 S.Ct. 1417. This “interview” was then televised on three occasions over the next three days. Id. The Court held that “due process of law ... required a trial before a jury drawn from a community of people who had not seen and heard Rideau’s televised ‘interview.’ ” Id. at 727, 83 S.Ct. 1417.
To some extent, Rideau appears, at least superficially, helpfxxl to DeLisle; the community in Rideau was made aware, befox*e trial, of the defendant’s confessioxx, and the Court held that this violated due process. But a closer reading of Rideau x’eveals that the controlling factor in the decision was the fact that the public viewed the confession in a televised format, which of coux'se was not the case here. As the Rideau Court explained, it was the power of the communicating medium that was significant; actually seeing and hearing the confession, as one would in a courtroom, would create a certainty of belief that would be difficult for the public to lay aside:
[W]e hold that it was a denial of due process of law to refuse the request for a change of venue, after the people of Calca-sieu Parish had been exposed x-epeatedly and in depth to the spectacle of Rideau personally confessing in detail to the crimes with which he was later to be charged. For anyone who has ever watched television the conclusion cannot be avoided that this spectacle, to the tens of thousands of people who saw and heard it, in a very real sense was Rideau’s trial' — at which he pleaded guilty to murder. Any subsequent court proceedings in a community so pervasively exposed to such a spectacle could be but a hollow formality.
Id. at 726, 83 S.Ct. 1417. Thus, it was only “the televising of a defendant in the act of confessing to a cx-ime” that the Rideau Coxxrt held “was inherently invalid under the Due Process Clause of the Fourteenth Amendment.” Estes, 381 U.S. at 538, 85 S.Ct. 1628 (emphasis added). The Rideau Court did not purport to create a rale that the dissemination of the fact of a defendant’s confession through some other, less dramatic and compelling medium, is equally a violation of the Due Process Clause.
We turn, finally, to Estes and Sheppard, and conclude that these cases speak very little to the circumstances at issue hex-e. There is no allegation by DeLisle that his trial took place under the conditions of total chaos that px-evailed in Estes and Sheppard, and a review of those cases leaves no room for doubt that it was that chaos that drove those decisions. See Murphy, 421 U.S. at 799, 95 S.Ct. 2031. Indeed, the proceedings in those cases were nothing short of astonishing. As the Murphy Court explained:
The trial in Estes had been conducted in a circus atmosphere, due in large part to the intrusions of the press, which was allowed to sit within the bar of the court and to ovex'run it with television equipment. Similarly, Sheppard arose from a trial in*385fected not only by a background of extremely inflammatory publicity but also by a courthouse given over to accommodate the public appetite for carnival. The proceedings in these cases were entirely lacking in the solemnity and sobriety to which a defendant is entitled in a system that subscribes to any notion of fairness and rejects the verdict of a mob. They cannot be made to stand for the proposition that juror exposure to information about a state defendant’s prior convictions or to news accounts of the crime with which he is charged alone presumptively deprives the defendant of due process.
421 U.S. at 799, 95 S.Ct. 2031.
To be sure, the Sheppard Court also considered at length “the trial judge’s failure to protect Sheppard sufficiently from the massive, pervasive and prejudicial publicity that attended his prosecution.” 384 U.S. at 335, 86 S.Ct. 1507. But the Court nonetheless explicitly held that it “c[ould not] say that Sheppard was denied due process by the judge’s refusal to take precautions against the influence of pretrial publicity alone.” Id. at 354, 86 S.Ct. 1507. It was, instead, “[t]he carnival atmosphere at trial” — the “bedlam” that “reigned at the courthouse” — that animated the Sheppard decision. Id. at 358, 355, 86 S.Ct. 1507. It is, moreover, the intense media coverage of the trial itself, combined with a jury that had not been adequately instructed not to read or listen to anything concerning the case, see id. at 347, 353, 86 S.Ct. 1507, that separates Sheppard from the DeLisle trial. As we have said, the trial here was well-run and tightly controlled by a skilled and careful presiding judge; there is no suggestion to the contrary by DeLisle, and indeed, the record indisputably reveals the great pains to which the court went to ensure that the jurors were insulated from media coverage of the trial.
There are two further significant distinctions between the circumstances here and those in which the Supreme Court has presumed partiality on the part of the jury. First, the Court has repeatedly suggested that a cessation of publicity for some period prior to trial will go a long way toward undoing the damage of a previous media blitz. See Stroble v. California, 343 U.S. 181, 192, 195, 72 S.Ct. 599, 96 L.Ed. 872 (1952); see also Murphy, 421 U.S. at 802, 95 S.Ct. 2031. “That time soothes and erases is a perfectly natural phenomenon, familiar to all.” Patton, 467 U.S. at 1034, 104 S.Ct. 2885. Here, as the Michigan Court of Appeals explicitly found, media attention to the case largely ceased shortly after publication of the transcript of DeLisle’s statements to the police in January 1990; indeed, prior to that time, interest in the case had already begun to diminish. Thus, 37 articles were published in local papers in August 1989; 15 in September; 8 in October; and 5 in November. Twelve articles were then published in December, largely in connection with the suppression proceedings, and 19 in January 1990. See DeLisle, 509 N.W.2d at 891-92. However, after January 1990, coverage was minimal; indeed, DeLisle does not complain of a single article appearing after January, and the record contains no such articles. There was, accordingly, more than a four-month lapse between the intense coverage and the time of trial, at which time public attention to the case began anew. However, once the venire had filled out the questionnaire, they were instructed not to read, watch, or listen to any news coverage about the case, and DeLisle does not suggest that coverage of the trial itself posed a problem.
A second consideration is that the Court has consistently emphasized the deleterious effect of pretrial publicity when that publicity has amounted to an “out-of-court campaign to convict,” reflecting “ ‘inflamed public sentiment,’ ” Shepherd v. Florida, 341 U.S. 50, 52, 53, 71 S.Ct. 549, 95 L.Ed. 740 (1951) (Jackson, J., concurring), such as when a defendant is persistently labeled in incendiary terms: “a ‘werewolf,’ a ‘fiend,’ a ‘sex-mad killer,’ and the like,” Stroble, 343 U.S. at 192, 72 S.Ct. 599; see generally Sheppard, 384 U.S. at 354-57, 86 S.Ct. 1507. As the Beck Court observed, coverage that consists of “straight news stories rather than invidious articles which would tend to arouse ill will and vindictiveness,” 369 U.S. at 556, 82 S.Ct. 955, is not so troubling. Accord Murphy, 421 U.S. at 800 n. 4, 95 S.Ct. 2031.
*386There is no question that the coverage of DeLisle’s case was, at least initially, intense, and the Michigan Court of Appeals described it as “inflammatory shortly after the incident.” DeLisle, 509 N.W.2d at 892. But we do not think that the publicity can fairly be described as ‘Virulent.” Sheppard, 384 U.S. at 354, 86 S.Ct. 1507. Indeed, DeLisle makes no such claim; he does not complain that the media ever labeled him in the fashion of the Stroble case, or portrayed him as an inhuman monster, or beat the drum for conviction.
Instead, DeLisle’s argument is simply that the fact-based coverage of his “confession” is what transforms this ease into one in which prejudice should be presumed. Any juror who read of his “confession,” he contends, must be presumed partial, because a confession is unlike any other piece of information. We must disagree, both as a legal and factual matter. First, we do not find support in the case law for the proposition that juror knowledge of an inadmissible confession leads per se to a presumption of prejudice. Indeed, the Supreme Court’s holdings “squarely foreclose[ ]” the argument that “jurors who read that [the defendant] had confessed to the murder [for which he is being tried] should [be] disqualified as a matter of law.” Mu’Min, 500 U.S. at 432, 111 S.Ct. 1899 (O’Connor, J., concurring). Thus, in Patton, the Court upheld the trial court’s decision to seat jurors who had read reports that the defendant had confessed. See 467 U.S. at 1029, 104 S.Ct. 2885. Further, we note, there certainly are other types of evidence, such as DNA evidence or particularly compelling uncharged-misconduct evidence, that would be equally damning for a defendant as would be a confession.
In any event, DeLisle’s statements did not constitute the type of full and unambiguous confession that it would be unrealistic to expect a jury to disregard. They were confused statements, not susceptible of easy interpretation. Unquestionably, they were of an inculpatory tone, and unquestionably, they bespoke a man riddled with guilt. But there was nothing in the statements that was fundamentally inconsistent with DeLisle’s defense at trial; any parent whose involuntary actions resulted in the death of his four children would certainly be expected to express deep remorse and self-reproach. Furthermore, media coverage of the statements also gave heavy emphasis to the Michigan Court of Appeals ruling that the statements would be suppressed because they had been extracted by coercive police tactics; in fact, all five of the jurors who were aware of DeLisle’s statements were aware of the suppression, and aware that he had recanted.
DeLisle here argues that “we can never be sure” that the printing of his statements did not taint the process. That is, of course, a truism. But the Supreme Court has cautioned that those cases in which “pretrial publicity presents [an] unmanageable threat[ ]” to a defendant’s right to an impartial jury trial are “relatively rare.” Nebraska Press Ass’n, 427 U.S. at 551, 554, 96 S.Ct. 2791. On the record before us, we find no extraordinary or utterly corrupting circumstances that give us reason to impute partiality to jurors who explicitly and firmly stated their lack of bias. What threats there were to DeLisle’s right to a fair trial were capably managed by the “ever watchful” trial court, Smith, 455 U.S. at 217, 102 S.Ct. 940, which took great care at every step of the way to ensure DeLisle received every due-process protection that he was due.
2.
The dissent here is concerned almost exclusively with the impropriety of what it terms a “judicial pi'ess conference.” This phrase denotes, for the dissent, the Detroit Recorder’s Court’s playing of the video- and audiotapes of DeLisle’s interrogation, on a single occasion some five months prior to his tidal in the Wayne County Cii’cuit Court, to an audience composed of interested members of the public, including the media. Using such hyperbolic phraseology as “collusion” and “unholy alliance,” (diss. op. at 391, 396), and asserting that the state courts “deliberately” and “intentionally” trampled on DeLi-sle’s rights, (diss. op. at 392, 393), the dissent suggests that the Michigan judiciary’s decision to play the tapes was not merely erroneous from a First Amendment pei-spective, but actually based, sub silentio, on a mali-*387eious desire to ensure DeLisle’s conviction. For this emotion-charged and fairly outrageous proposition, there is an utter dearth of support in the record. It is, moreover, not even a contention that has been urged by DeLisle; indeed, at oral argument, DeLisle’s attorney explicitly declined to take the position that the court’s role in the dissemination of DeLisle’s statements bore my significance for the due-process analysis. But we shall nonetheless address this theory, lest the failure to do so be interpreted as a nod to the theory’s persuasiveness.
The dissent relies on a line of cases beginning with Mattox v. United States, 146 U.S. 140, 13 S.Ct. 50, 36 L.Ed. 917 (1892), for the assertion that because there was a “judicial source,” (diss. op. at 391), for the publicity regarding DeLisle’s statements — that is, because the tapes of the statements were made available to the public pursuant to a valid court order, not because a particular judge improperly leaked the information to the media — the publicity necessarily resulted in prejudice. The dissent undertakes virtually no inquiry into whether the evidence in the record shows that the defendant received an unfair trial because the jury was irredeemably tainted; this inquiry is bypassed in favor of a presumption to that effect.
Mattox and its progeny, however, do not authorize that presumption. First, as a factual matter, Mattox, along with Turner v. Louisiana, 379 U.S. 466, 85 S.Ct. 546, 13 L.Ed.2d 424 (1965), and Parker v. Gladden, 385 U.S. 363, 87 S.Ct. 468, 17 L.Ed.2d 420 (1966), all concerned personal contact between bailiffs and the jury during the jury’s deliberations. In Mattox, the bailiff discussed the case with the jurors, conveying information about the defendant that had not been put in evidence, and attempted to sway the jury to convict. In Turner, the two sheriffs who were in charge of the jury, and who were thereby placed in “continuous and intimate association” with the jury, were also key witnesses for the prosecution. Turner, 379 U.S. at 473, 85 S.Ct. 546. And in Parker, the bailiff made statements to members of the jury, attempting to convince them of the defendant’s guilt.
The distinctions between these cases and that of DeLisle are manifest. Among other things, the publicity regarding DeLisle’s statements took place not during the jury’s deliberations but five months jn advance of trial. And obviously, the release of the statements involved no personal contact between any officer of the state judiciary and the eventual members of the jury. Indeed, there is nothing whatsoever in the record to suggest that in publicizing DeLisle’s statements, the media suggested that their accounts bore the official imprimatur of the Michigan courts; the courts’ role in making the statements available was not one that was itself publicized or of any apparent interest or importance to the reading public. Finally, and significantly, there is no basis for impugning the motives of the Michigan judiciary. The court of appeals ruling that the First Amendment required some limited release of DeLisle’s statement, given that those statements had been a strong evidentiary factor in DeLisle’s bind-over, contains no hint that anti-DeLisle animus was a key ingredient in the decision. While there was ample evidence that the bailiffs in Mattox, Turner, and Parker all harbored personal convictions against the defendants in question, there is no such evidence here.
We find it noteworthy, too, that the dissent does not fault the actions of the trial judge— nor, we believe, could it do so. But the dissent fails to acknowledge, or perhaps to recognize, that the publicizing of DeLisle’s confession occurred as a result of a judicial proceeding entirely independent of the criminal prosecution. It is obvious, we would have thought, that this court has no authority, in this habeas appeal, to attack or defend the Michigan judiciary’s understanding of the demands of First Amendment jurisprudence. But whatever that court’s understanding of that subject, there is no question that the conduct of the criminal proceeding itself was beyond reproach.
Even if there were a closer factual similarity between the facts of Mattox, Turner, and Parker and those here, those cases simply do not create a per se rule that the official, albeit improper, dissemination of suppressed evidence necessarily results in a due-process *388violation. These cases are plain that there must in all events be an inquiry into the existence of actual prejudice, or some basis for concluding that prejudice should be presumed. And,the dissent’s analysis simply begs the question of whether prejudice occurred. It states, in conclusory fashion, that knowledge of a confession “is inherently prejudicial,” (diss. op. at 394), and that this court “may not indulge in the fiction that the jury can somehow put this evidence aside,” (diss. op. at 395). It is of no minor importance, we note, that DeLisle’s statements were not “evidence.” Nevertheless, to be sure, the jury’s knowledge of a suppressed confession necessarily raises due-process concerns. But as we have discussed in detail, existing Supreme Court precedent gives no support for the dissent’s per se approach.
Likewise, there simply is no case law— and, of course, the dissent cites none — for the proposition that a jury is presumed to have been tainted when a body of previously suppressed information is released at the direction of the state judiciary, rather than having been made public without judicial action. The due-process analysis should be no different if the tapes of DeLisle’s statements wex*e obtained by some illicit media initiative, as opposed to hex'e, where the tapes were released as a result of the state courts’ arguably misguided, yet unquestionably valid, First Amendment analysis. Indeed, even if thex-e were reason to believe that the conduct of the “Michigan courts,” that is to say, three court of appeals panelists and one obedient Recorder’s Court judge, was as ill-intentioned as the dissent insinuates, that fact would be a red herring. As the Supreme Court has observed in another context, “the touchstone of due process analysis ... is the fairness of the trial, not the culpability” of the particular state actor under scrutiny. Smith, 455 U.S. at 219, 102 S.Ct. 940. “[T]he aim of due px-ocess ‘is not the punishment of society for the misdeeds of the [state] but avoidance of an unfair trial to the accused.’ ” Id. (citation omitted). Therefore, as we have already noted, “there could be no constitutional infirmity ... if [the defendaixt] actually received a trial by an impartial jury,” and “our inquiry [should thex'efore be] addressed to that subject.” Beck, 369 U.S. at 556, 82 S.Ct. 955. It is, however, a subject that is all but ignored by the dissent in favor of groundless hypothesizing about the bad faith of the Michigan judiciary.
We note, in closing, that this court does not, in this habeas case," exercise the sox't of supervisory power over the Michigan courts that led the Supreme Court in Marshall v. United States, 360 U.S. 310, 79 S.Ct. 1171, 3 L.Ed.2d 1250 (1959), to reverse a federal-eoux-t conviction when, during trial, the jury was exposed to two newspaper articles containing prejudicial information that had been excluded from evidence. We have no power hex-e “to intervene to protect the integrity of the [state court] system,” Frazier v. Heebe, 482 U.S. 641, 647 n. 7, 107 S.Ct. 2607, 96 L.Ed.2d 557 (1987); we are instead “limited to enforcing the commands of the United States Constitution,” Mu’Min, 500 U.S. at 422, 111 S.Ct. 1899. We must not be so distracted by preoccupying personal outrage toward the conduct of the state courts that we ignore our limited constitutional mandate.
Throughout the life of this constitutional democracy, courts have struggled with the necessary tension between “the right of the accused to trial by an impartial jury and the rights guaranteed others by the First Amendment.” Nebraska Press Ass’n, 427 U.S. at 551, 96 S.Ct. 2791. Federal courts have decided a long line of cases accommodating this tension — acting always, of necessity, after the fact. But the federal judiciary's distaste for media frenzies, and our dissatisfaction with how the Recorder’s Court handled these competing constitutional rights in this case, do not empower us to make broad superintending pronouncements that would result in the reversal of a murder conviction validly obtained iix a state court.
D.
DeLisle next argues that the district court erred when it held that the evidence presented against him is sufficient to prove that he acted with both premeditation and intent to kill when he drove into the Detroit River on August 3,1989. We disagree.
*389In order to prove that a defendant is guilty of first-degree murder or first-degree attempted murder under Michigan law, the state must prove that the defendant acted with premeditation and intent to kill. See Mich. Comp. Laws §§ 750.316, 750.91; see also People v. Youngblood, 165 Mich.App. 381, 418 N.W.2d 472, 475 (Mich.Ct.App.1988); People v. Ng, 156 Mich.App. 779, 402 N.W.2d 500, 503-04 (Mich.Ct.App.1986). Premeditation and intent to kill may be inferred from circumstantial evidence. See, e.g., Youngblood, 418 N.W.2d at 475; Ng, 402 N.W.2d at 504.
An applicant for a writ of habeas corpus is entitled to relief if, “after viewing the evidence in the light most favorable to the prosecution,” the court concludes that no “rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.” Jackson, 443 U.S. at 319, 99 S.Ct. 2781.
In this case, the evidence of premeditation and intent to kill is entirely circumstantial, a situation that is not unusual in murder cases generally. Rarely is direct evidence of premeditation and intent to kill available in a murder case, except where a confession is received in evidence. DeLisle’s jury was required to examine the evidence of DeLisle’s conduct and infer what he thought and what he intended to do from what he did and failed to do—the quintessential jury responsibility in a circumstantial-evidence case. Our duty is to determine whether, viewing the evidence in the record in the light most favorable to the prosecution, we are able to conclude that no rational juror could be convinced beyond a reasonable doubt that before driving into the Detroit River, DeLisle intended with premeditation to drown his wife and children. We have carefully and deliberately considered the record in this case, and, like the trial court, the Michigan Court of Appeals, and the United States District Court, we are not able to reach such a conclusion.
DeLisle’s intent to drown his family, and perhaps even himself, could reasonably be inferred from the fact that he drove slowly to the end of Eureka Road on two occasions shortly before the incident; that on the evening of the drownings, the station wagon accelerated rapidly down Eureka Road and into the Detroit River, splitting the protective posts at the end of the street nearly perfectly, without any apparent swerving or braking, and despite the fact that the ear’s brakes were capable of bringing the car to a stop; and that DeLisle surfaced quickly and did not call for help from boaters nearby, and made no attempt whatever to save his family.
That some or most of this evidence may also be consistent with DeLisle’s defense is ultimately of no consequence. The jurors were the final judges of the credibility of DeLisle’s defense and they were entitled to reject the story of a stuck accelerator and leg cramp. We, of course, are not free to reject the jurors’ credibility determination, and we may not now upset the jurors’ verdict simply because the evidence does not “rule out every hypothesis except that of guilt beyond a reasonable doubt.” Jackson, 443 U.S. at 326, 99 S.Ct. 2781.
E.
DeLisle’s final argument on appeal is that he was denied due process when the trial court denied his request to waive the jury in favor of a bench trial. Having already concluded that DeLisle’s right to an impartial jury was adequately safeguarded, it is clear that DeLisle had no federal constitutional right to waive his jury over the objection of the prosecution. Singer v. United States, 380 U.S. 24, 34-36, 85 S.Ct. 783, 13 L.Ed.2d 630 (1965).
III.
For all the foregoing reasons, we AFFIRM the judgment of the district court.