FLAUM, Circuit Judge.
Appellant Brian Grancorvitz petitioned the district court for a writ of habeas corpus pursuant to 28 U.S.C. § 2254. He claimed his Sixth Amendment right to an impartial jury and his Fifth Amendment privilege against self-incrimination were violated during his trial and conviction by the State of Wisconsin. The district court denied the petition. Grancorvitz appeals, raising the same two claims he presented to the district court. We affirm.
I.
Appellant was charged with first degree murder. Prior to trial, he moved for a change of venue or a jury outside of Vernon County, Wisconsin, which was denied by the trial court. The court did grant appellant’s motion in limine to prevent references to certain “other crimes” evidence. At trial, appellant admitted he stabbed the victim, but claimed self-defense. The jury convicted appellant of first degree murder, and the court sentenced him to life imprisonment. Appellant unsuccessfully moved for a new trial, partly on the ground that the trial court improperly denied his motion for a change of venue. Appellant then unsuccessfully appealed through the state courts, claiming a denial of his right to a fair and impartial jury; the court of appeals affirmed this conviction and the supreme court denied review. Appellant was also unsuccessful in his pursuit of state post-conviction relief, in which he alleged a violation of his privilege against self-incrimination from the prosecutor’s repeated references to his post-arrest silence.
Appellant then filed his habeas corpus petition in district court. He first argued, as he does on appeal, that he did not have an impartial jury because of the publicity regarding his arrest, the type of community in which he was tried, and the type of defense he presented. The district court held that “the trial court’s determination that the jury was impartial is fairly supported by the record.” It reasoned that the nature and timing of the pre-trial publicity did not suggest juror impartiality, that the voir dire sufficiently ruled out impartiality, and that the trial court was justified in giving little weight to the results of a survey submitted by appellant (and the accompanying expert witness testimony) which suggested community prejudice against him. Appellant also argued and continues to argue that the seven references by the prosecutor to appellant’s post-arrest failure to complain of injuries that would have occurred during his altercation with the victim and his attempts at self-defense, or to ask for medical treatment for those injuries, violated his right to remain silent invoked after receiving his
Miranda
warnings. The district court rejected this argument as well, holding that the prosecutor’s comments were intended to show appellant’s lack of physical injury, as relevant to appellant’s asserted claim of self-defense, rather than as an inference of guilt. The court held that at any rate any error was harmless beyond a reasonable doubt.
II.
TRIAL COURT’S DENIAL OF MOTION TO CHANGE VENUE
A.
The Sixth Amendment guarantees criminal defendants the right to “an impartial jury of the State and district wherein the crime shall have been committed.” Defen
dants can establish the existence of a partial jury either by showing that pretrial publicity rendered the trial setting inherently prejudicial or by showing that the publicity created actual juror prejudice.
Willard v. Pearson,
823 F.2d 1141, 1146 (7th Cir.1986). Actual prejudice, which is the issue here, is established where a defendant shows “that a juror cannot lay aside any preconceived impression or opinion of the case and decide the case solely on the evidence.”
Id.; see also United States v. Reynolds,
821 F.2d 427, 432 (7th Cir.1987). Jurors need not be totally ignorant of the facts and issues in a case to be impartial.
Irvin v. Dowd,
366 U.S. 717, 722, 81 S.Ct. 1639, 1642, 6 L.Ed.2d 751 (1961). The fact that jurors can recall hearing about the case and any unfavorable publicity does not establish unconstitutional partiality. “The relevant question is not whether the community remembered the case, but whether the jurors at [defendant’s] trial had such fixed opinions that they could not judge impartially the guilt of the defendant.”
Patton v. Yount,
467 U.S. 1025, 1035, 104 S.Ct. 2885, 2891, 81 L.Ed.2d 847 (1984) (citing
Irvin,
366 U.S. at 723, 81 S.Ct. at .1643). Even a “preconceived notion as to the guilt or innocence of an accused” is insufficient to rebut the presumption of impartiality, if “the juror can lay aside his impression or opinion and render a verdict based on the evidence presented in court.”
Irvin,
366 U.S. at 723, 81 S.Ct. at 1643. However, juror’s assurances during voir dire that they can judge the case impartially “cannot be dis-positive of the accused’s rights,” if the defendant demonstrates actual prejudice sufficient to raise the presumption of partiality.
Murphy v. Florida,
421 U.S. 794, 800, 95 S.Ct. 2031, 2036, 44 L.Ed.2d 589 (1975).
A federal habeas court’s standard of review over a state court’s determination that a jury was impartial has not yet been definitively established. Both parties here ultimately argue under both a manifest error standard and a presumption of correctness standard requiring fair support for the finding in the record. In
Irvin,
the Supreme Court held that the issue of jury impartiality was a mixed question of law and fact, that a federal habeas court had an independent duty to evaluate the voir dire testimony, and that a determination of juror impartiality should only be set aside if the error was manifest. 366 U.S. at 723-24, 81 S.Ct. at 1642-44. However, the Court has since held that whether an individual juror was impartial is a finding entitled to the presumption of correctness of 28 U.S.C. § 2254(d).
Patton,
467 U.S. at 1036, 104 S.Ct. at 2891. However, the Court failed to determine whether this presumption is applicable to the question of whether the whole jury was impartial in violation of the Sixth Amendment.
Id.
at 1031 n. 7, 104 S.Ct. at 2889 n. 7.
In
Patton
the Court noted that its decision in
Irvin,
which required a federal court to independently evaluate voir dire testimony and view the question of juror impartiality as a mixed question of law and fact, came before the enactment of § 2254(d). However, the Court did not resolve the issue of the appropriate standard, since it found that the § 2254(d) standard was not less stringent than the manifest error standard of
Irvin,
and since there were no grounds to grant habeas under the less deferential manifest error standard, there would be no basis for granting habeas under the § 2254(d) presumption of correctness standard.
Id.
Under § 2254(d)’s presumption of correctness, a federal habeas court will reverse a state trial court determination only if it is not fairly supported by the record. Although a reviewing court’s independent review into whether another court’s decision was manifestly erroneous is also a deferential standard of review, we do not believe it involves the almost complete deference required under § 2254(d).
The Supreme Court itself has applied the two standards in a way that supports the argument that the standards involve differing levels of deference to state courts. The manifestly erroneous standard used by the Court in both
Irvin
and
Patton
regarding the issue of whole jury impartiality involved an independent examination and interpretation by the Court into the trial court record and voir dire testimony.
Irvin,
366 U.S. at 725-28, 81 S.Ct. at 1644-46;
Patton,
467 U.S. at 1032-35, 104 S.Ct. at 2889-91. The Court’s analysis in
Patton
in applying a presumption of correctness to the issue of individual juror impartiality, on the other hand, deferred to the judge on the procedures and inquiry employed during voir dire, and to the judge’s belief in the jurors’ assurances of impartiality as primarily a credibility determination.
Patton,
467 U.S. at 1038-40, 104 S.Ct. at 2892-94.
We believe that the slightly less deferential manifest error analysis should be applied in this ease.
Although the Supreme Court’s language in
Patton
clearly requires deference from reviewing courts on determinations of jury impartiality, the degree of deference required is not necessarily that of § 2254(d)’s presumption of correctness. Indeed, the Court itself in
Patton
analyzed the facts before it under the manifest error standard, and arguably that standard should apply until the Court states otherwise.
Moreover, the manifest error standard is sufficiently deferential to remain well within the realm of a federal habeas court reviewing state court determinations, and the burden remains on the petitioner to show manifest error.
United States v. Smith,
748 F.2d 1091, 1094 (6th Cir.1984). Finally, it remains true that any underlying credibility determinations as to the partiality of an individual juror are accorded a presumption of correctness.
Cummings v. Dugger,
862 F.2d 1504, 1510 (11th Cir.1989).
We believe manifest error is the appropriate test here because the determination of whole jury impartiality is less purely factual and less credibility based than the determination of individual juror impartiality, where the Court requires a presumption of correctness. Individual juror impartiality is determined on the basis of voir dire and the credibility of individual jurors. In determining whole jury impartiality, whether voir dire sufficiently dealt with preconceptions is one consideration, but there are larger inquiries into whether the type and timing of the publicity and the
nature and posture of the community infringed on the Sixth Amendment right to an impartial jury. These additional factors are ones which a state trial court is not necessarily in a better position to judge than a federal court on habeas review. Finally, as stated above,
Patton
itself utilized a different standard for individual juror impartiality than for whole jury impartiality, stating that
Irvin’s
analysis for whole jury impartiality could not be extended to an individual juror impartiality determination, because the latter decision is “plainly one of historical fact”, and “not one of mixed law and fact.” 467 U.S. at 1036,104 5.Ct. at 2891.
Although we applied a presumption of correctness standard to the “inherently prejudicial” prong
of the whole juror impartiality issue in
Willard,
823 F.2d at 1146, we did not state whether such a presumption was applicable to the “actual prejudice” prong as well. Moreover, we only applied the § 2254(d) presumption to those factual findings underlying the state court’s determination that the trial setting was not inherently prejudicial; we did not defer to the state court on the ultimate issue, and we evaluated the record and the publicity independently during our review.
Id.
at 1146-47. Our holding in the present case employing the manifest error standard is therefore not inconsistent with
Willard.
B.
Appellant claims that the combination of the pre-trial publicity, the nature of the community (a small rural community which he claims is biased against “outsiders”), and the nature of his defense (a self-defense claim heavily dependent on credibility determinations) deprived him of an impartial jury. He relies on various newspaper articles and radio reports, on a survey conducted by two University of Wisconsin faculty members about community attitudes toward appellant and the charges against him (based on 251 telephone interviews),
and on the testimony of four expert witnesses, testifying before trial and at appellant’s post-trial motion for a new trial, interpreting this survey, the pretrial publicity, and the likelihood of obtaining an impartial jury.
In its denial of appellant’s motion for a change of venue, the state trial court determined that voir dire would sufficiently protect appellant’s right to an impartial jury:
By attempting to find out during voir dire the extent of a potential juror’s knowledge of the case, including the drug charges, the Court will be able to exclude those persons where it appears they may have heard about those charges regardless of whether or not they could decide the ease impartially without reference to their previous knowledge. To this extent, the Court would allow individual Voir Dire of the potential jurors so that counsel may inquire specifically what they have heard about the case.
The court rejected the conclusion of the survey and appellant’s expert witnesses that “there was a reasonable probability of prejudice in the panel from which a jury would be drawn,” reasoning that the survey failed to inquire whether the 43% of the respondents who thought appellant was probably or definitely guilty could put aside what they had heard and decide the case on the facts given in court. The court emphasized that, contrary to appellant’s witnesses’ testimony, courts presume persons testify truthfully and that it is possible for them to “put aside outside factors and decide a case only on the evidence given in Court.” As for the survey’s suggestion of prejudice against persons with defendant’s characteristics and appearance,
the court found that no showing was made that such prejudices would not also exist elsewhere in the state, and the trial judge believed these biases could be brought out in voir dire. The trial court concluded that “a reasonable probability of prejudice does not exist in Vernon County.”
Under the manifest error standard, we must independently evaluate the testimony given to the state trial court to determine whether it was manifest error for that court to hold that the jury was sufficiently impartial to avoid a Sixth Amendment violation.
See Irvin,
366 U.S. at 724, 81 S.Ct. at 1643;
see also Patton,
467 U.S. at 1032-35, 104 S.Ct. at 2889-91 (court independently evaluated record below and voir dire testimony). Three Supreme Court decisions govern our inquiry.
In
Patton,
467 U.S. 1025, 104 S.Ct. at 2885, the Court held that the trial court’s determination of jury impartiality was not manifestly erroneous. There, the defendant had been retried, and most of the publicity occurred before the first trial, four years prior to the selection of the jury at issue.
Id.
at 1032, 104 S.Ct. at 2889. In the year and a half prior to the second trial, there was an average of less than one newspaper article per month. During voir dire, articles appeared daily, but these were factual articles primarily discussing the jury selection process. The Court’s evaluation of the voir dire led it to conclude that the lapse in time between the pretrial publicity and the trial had softened public opinion and caused the jurors to forget the details of the articles. Finally, as to those members of the venire who retained fixed negative opinions about the defendant, the Court found that they were properly disqualified and that “the voir dire resulted in selecting those who had forgotten or would need to be persuaded again.”
Id.
at 1034, 104 S.Ct. at 2890.
The Court also upheld the state court’s determination of jury impartiality in
Murphy,
421 U.S. 794, 95 S.Ct. 2031. There, the pre-trial publicity had ceased only seven months before jury selection.
Id.
at 802, 95 S.Ct. at 2037, Seventy-eight potential jurors were questioned, and twenty of those were excused after expressing a preconceived opinion on the defendant’s guilt.
Id.
at 803, 95 S.Ct. at 2037. Some selected jurors had knowledge of past crimes and some recollection of the publicity on the current charge,
id.
at 800, 95 S.Ct. at 2036, but because the jurors did not reveal a belief that this past activity was relevant to the current charge, the Court found that these jurors remained sufficiently impartial.
In
Irvin,
366 U.S. 717, 81 S.Ct. 1639, the Court found manifestly erroneous the state court’s decision that the jury was impartial. Most of the pre-trial publicity there occurred six to seven months before trial. The petitioner presented forty-six articles that had appeared in newspapers regularly delivered to 95% of the dwellings of the County in which he was tried, and evidence that the television and radio stations in the County had extensively covered the story as well.
Id.
at 725, 81 S.Ct. at 1644. The information revealed through the publicity included references to juvenile and adult convictions, statements that petitioner had been identified in a police line-up and placed at the scene of the crime, and the fact that petitioner had later confessed to six other murders and twenty-four burglaries.
Id.
at 726, 81 S.Ct. at 1644. The Court noted that the tone of the articles very strongly encouraged prejudice against the petitioner. The jury was chosen from a panel of 430 persons, of which 268 were excused for cause based on fixed opinions that petitioner was guilty. Of the 430, 370 believed petitioner was guilty, in a range from merely suspicious to absolutely certain.
Id.
at 727, 81 S.Ct. at 1645. Eight of the twelve seated jurors believed petitioner was guilty, “some going so far as to say that it would take evidence to overcome their belief.”
Id.
at 728, 81 S.Ct. at 1645.
We hold that the circumstances of appellant’s trial are more akin to those of the defendants in
Patton
and
Murphy
than to the defendant in
Irvin.
In this case, there were 180 persons in the initial juror pool. Seventy-nine of these individuals were questioned. Sixteen were excused for personal reasons. Sixty (95%) of the remaining sixty-three were familiar with the case,
and the court denied appellant’s motion to exclude all jurors who had heard about the case. The court and the lawyers then conducted individual sequestered voir dire. Nineteen persons were excused because they admitted strong pre-conceived notions about the case. Two individuals were excused because of their knowledge that appellant was allegedly involved with drugs. Appellant unsuccessfully moved to exclude two other prospective jurors who had prior knowledge of the case from reading a newspaper article that referred to excluded evidence. Of the thirteen persons selected for the panel, all but one person had prior knowledge of the case, five had exposure to publicity that referred to inadmissible evidence, and none could recall details of the publicity or of the case in general. The process of jury selection took one day.
While there are factors in this case which favor appellant’s argument that he was denied an impartial jury, the factors against appellant’s claim of an impartial jury outweigh those in his favor. First, the publicity at issue here was mostly factual. “[Ljargely factual” publicity does not present the same potential for prejudice as publicity which is “invidious or inflammatory.”
Murphy,
421 U.S. at 800 n. 4, 95 S.Ct. at 2036 n. 4. Such publicity is not prejudicial enough to disbelieve the juror’s individual assurances that they would base their decision solely on the evidence at trial. Second, the extent of the publicity in
Irvin
was much more extreme than that surrounding appellant. Third, although the three month period between the publicity and the jury selection is shorter than the periods in
Patton
and
Murphy,
it was enough time to diffuse publicity that was at most only slightly prejudicial. Fourth, the trial court here allowed individual, sequestered voir dire and an opportunity, without much restriction by the court, to weed out jurors with prejudicial pre-disposi-tions. Fifth, the survey on which appellant relies did not measure the possibility of respondents putting aside preconceived notions, or give its respondents the choice of stating that they had no opinion or pre-dis-position of appellant’s guilt. Sixth, as for the reference to “other crimes evidence” in some newspaper articles, the Supreme Court has held that “juror exposure to information about a state defendant’s prior convictions” does not presumptively deprive a defendant of a fair trial.
Murphy,
421 U.S. at 799, 95 S.Ct. at 2035. Also, the jurors actually chosen for the panel were those who could not recall the details of the news reports and therefore presumably did not remember mention of these prior crimes. As stated above, it is not prior exposure to facts of the case that creates unconstitutional juror partiality; it is an inability to decide the case solely on evidence presented at trial. Absent a showing of such inability, we have refused to hold that a jury was unconstitutionally partial.
See, e.g., Reynolds,
821 F.2d at 432 (rejecting a claim that voir dire was so inadequate that prejudice could not be determined);
United States v. Kimberlin,
805 F.2d 210, 224 (7th Cir.1986) (almost all of fifty-one prospective jurors had heard of defendant, twelve were excused for cause, twelve of the sixteen selected had heard about the crime, several were aware of the accusations against the defendant, but none could recall details),
cert. denied,
483 U.S. 1023, 107 S.Ct. 3270, 97 L.Ed.2d 768 (1987);
United States v. Garza,
664 F.2d 135, 139 (7th Cir.1981) (although many of the potential jury members were familiar with the case, the voir dire filtered out the biased individuals, and only four of the jurors seated had read or heard of the case),
cert. denied,
455 U.S. 993, 102 S.Ct. 1620, 71 L.Ed.2d 854 (1982).
Finally, although appellant makes an interesting argument in claiming that receiving an impartial jury was particularly important because his case came down to whether the jury found his defense of self-defense credible or whether it instead believed the prosecution’s theory that the defense was fabricated, reliance on such an argument would allow reviewing courts to determine the appropriateness of a motion for a change of venue through hindsight. Trial courts do not know beforehand what issues will be determinative at trial. Moreover, an accused’s right to an impartial jury does not vary in degree by how crucial
the jury’s credibility determinations were to the conviction. While such an analysis may be relevant to a harmless error inquiry, this type of inquiry is inapplicable here.
Coleman,
778 F.2d at 1540-41 & n. 24.
III.
PROSECUTOR’S COMMENTS ON POST-ARREST SILENCE
Appellant’s second claim is that the state prosecutor unconstitutionally commented upon his post-arrest silence. The Supreme Court has held that a prosecutor violates a defendant’s due process right by seeking to “impeach a defendant’s exculpatory story, told for the first time at trial, by cross-examining the defendant about his failure to have told the story after receiving
Miranda
warnings at the time of his arrest.”
Doyle v. Ohio,
426 U.S. 610, 611, 96 S.Ct. 2240, 2241, 49 L.Ed.2d 91 (1976).
Such comment is “fundamentally unfair because
Miranda
warnings inform a person of his right to remain silent and assure him, at least implicitly, that his silence will not be used against him.”
Anderson v. Charles,
447 U.S. 404, 407-08, 100 S.Ct. 2180, 2181-82, 65 L.Ed.2d 222 (1980) (citations omitted). Thus, a prerequisite for a
Doyle
violation is some type of implicit assurance from the government that silence will not used against him; a
Miranda
warning is such an assurance.
See, e.g., Greer v. Miller,
483 U.S. 756, 763, 107 S.Ct. 3102, 3108, 97 L.Ed.2d 618 (1986).
There are seven references to appellant’s post-arrest silence at issue here.
Appellant points to the prosecutor’s one reference during the case-in-chief to appellant’s failure to discuss any injuries with the jailors following his arrest,
to the five references during cross-examination of appellant regarding his failure to seek medical help,
and to the one reference during closing argument to appellant’s failure to
complain of injuries.
Appellant claims that the prosecutor’s use of this silence was unconstitutional because it inferred that since appellant did not complain of serious injuries he probably was not injured during his fight with the victim, and therefore he could not have been acting in self-defense when he killed the victim. The state claims that appellant’s failure to complain about the injuries or to request medical assistance was not an invocation of his
Miranda
rights, since those rights guarantee a right not to respond to custodial interrogation, and not a generalized right to silence. Appellant counters that argument with reference to his affidavit which states that his silence was such an invocation because he was abiding by his counsel's advice that he not talk to anyone. Finally, the government claims that even if the references constituted a
Doyle
violation, such error was harmless.
The first prerequisite of
Doyle
was met here — appellant was given his
Miranda
warnings and thus there were government assurances that he could remain silent without penalty.
Cf. Fletcher v. Weir,
455 U.S. 603, 606, 102 S.Ct. 1309, 1311, 71 L.Ed.2d 490 (1982);
Jenkins v. Anderson,
447 U.S. 231, 239-40, 100 S.Ct. 2124, 2129-30, 65 L.Ed.2d 86 (1980). However, for the reasons discussed below, we do not believe this case is within the realm of due process violations contemplated by the Court in
Doyle.
First, the type of silence at issue here is not the kind of silence assured protection under the
Miranda
warnings. Since the due process right established in
Doyle
is dependent upon the government giving assurances to an individual, such as those given under the
Miranda
warnings, the individual must be invoking the rights inherent in those assurances.
The
Doyle
holding stems from notions of “induced detrimental reliance” and “fundamental unfairness” because the warnings invite the arrestee to exercise the cited rights.
United States ex rel. Saulsbury v. Greer,
702 F.2d 651, 655 (7th Cir.),
cert. denied,
461 U.S. 935, 103 S.Ct. 2104, 77 L.Ed.2d 310 (1983). “What is impermissible is the evi-dentiary use of an individual’s
exercise of his constitutional rights
after the State’s assurance that the invocation of those rights will not be penalized.”
Wainwright v. Greenfield,
474 U.S. 284, 295, 106 S.Ct. 634, 640, 88 L.Ed.2d 623 (1986) (emphasis added).
We do not believe that appellant’s failure to complain of injuries was an exercise of his
Miranda
right to remain silent. The type of silence addressed in
Doyle
was that defendant “did not speak about the facts of the case at that time, as he was told he need not do ...”
Doyle,
426 U.S. at 619, 96 S.Ct. at 2245 (quoting
United States v. Hale,
422 U.S. 171, 182-83, 95 S.Ct. 2133, 2139-40, 45 L.Ed.2d 99 (1975)). There is a difference between refusing to speak about the facts of a case or explain one’s role, if any, in the crime, and not asking for medical treatment or failing to complain of injuries. The right to remain silent in the face of accusations of illegal activity only includes the former.
Some examples of silence which are protected by the
Miranda
warnings are a defendant’s failure to give the police his alibi,
Smith v. Rowe,
618 F.2d 1204, 1206 (7th Cir.), vacated and remanded on other
grounds sub nom.,
Franzen v. Smith,
449 U.S. 810, 101 S.Ct. 57, 66 L.Ed.2d 13 (1980), a defendant’s statement that he will remain silent and not answer police officer’s questions until consulting with an attorney,
Greenfield,
474 U.S. at 286, 106 S.Ct. at 636, a defendant’s failure to tell police officers he had been framed,
Doyle,
426 U.S. at 613, 96 S.Ct. at 2242;
United States v. Shue,
766 F.2d 1122, 1130-31 (7th Cir.1985), and a defendant’s failure to tell officers that he acted in self-defense.
United States ex rel. Allen v. Franzen,
659 F.2d 745, 747 (7th Cir.1981),
cert. denied,
456 U.S. 928, 102 S.Ct. 1975, 72 L.Ed.2d 444 (1982) (although court held use of that silence permissible). We do not believe appellant’s failure to complain of injuries or seek medical attention is akin to these other protected instances of silence.
Although appellant’s affidavit from his state post-conviction motion states that his silence was the result of his attorney’s advice that he speak to no one about the case, that does not automatically transform silence into constitutionally protected silence. A conclusion to the contrary would imply that many instances of a defendant’s failure to do something involving verbal communication following receipt of his
Miranda
warnings could not be used by the prosecutor at trial, even though such failure was not a constitutionally protected act.
Second, the prosecutor’s use of appellant’s silence differs from the use of silence in cases where this court and the Supreme Court have found
Doyle
violations. The precedent in this area makes it clear that the government cannot use protected silence to impeach an exculpatory story or an explanation of the facts of the crime told at trial,
Greenfield,
474 U.S. at 291, 106 S.Ct. at 638 (quoting
South Dakota v. Neville,
459 U.S. 553, 565, 103 S.Ct. 916, 923, 74 L.Ed.2d 748 (1983));
Fencl v. Abrahamson,
841 F.2d 760, 765 (7th Cir.1988), but that such silence can be used to impeach the credibility of a defendant. “ ‘Comment on the lack of credibility of the defendant is plainly proper; it is not proper, however, for the prosecutor to ask the jury to draw a direct inference of guilt from silence — to argue, in effect, that silence is inconsistent with innocence.’ ”
Saulsbury,
702 F.2d at 656 (quoting
Doyle v. Ohio,
426 U.S. 610, 634-35, 96 S.Ct. 2240, 2252-53, 49 L.Ed.2d 91 (1976) (Stevens, J., dissenting)). Thus, what is impermissible is the government’s inference that silence is inconsistent with innocence. Appellant argues that the state was making such an inference here: that the state argued that defendant’s silence meant that he did not sustain serious injuries as he claimed and that therefore his claim of self-defense was probably fabricated and therefore he probably was guilty. Even were we to view this as an indirect attack upon defendant’s innocence, we think the government’s use was primarily one attacking defendant’s credibility and not defendant’s innocence.
In this sense, this ease is similar to
Saulsbury, supra,
where the prosecutor’s comments during closing argument were primarily an attack on defendant’s credibility, but were also indirectly an attack on defendant’s innocence. We held there that where a slight inference of guilt is inextricably intertwined with an attack on credibility, there is not necessarily reversible error.
Id.
Here, the indirect attack on defendant’s underlying self defense theory is similarly not sufficient for reversible error.
We have noted before that since
Doyle
was based on flexible due process considerations, rather than those underlying the Fifth Amendment, “the rule prohibiting the government’s use of a defendant’s post-arrest silence depends upon a balancing of defense and prosecution interests and upon considerations of fairness within the context of the truth-seeking function of trials.”
United States v. Mavrick,
601 F.2d
921, 933 (7th Cir.1979). We therefore must consider the degree of the state’s interest and the defendant’s involvement in creating that interest. In
Mavrick,
601 F.2d at 932, we found it permissible for the prosecutor to refer to defendant’s
post-Miranda
invocation of his right to remain silent to rebut the defendant’s assertion on direct examination that he was not given an opportunity to explain his story to the authorities.
Although the defendant’s testimony would not permit the government to argue that his post-arrest silence was inconsistent with his claim of innocence, we hold it did permit the government to attempt to discredit his testimony by showing that he was given such an opportunity and did not take advantage of it. We find nothing fundamentally unfair to the defendant in permitting the government to correct an impression that the defendant created by his own testimony.
Id.
at 933. As in
Mavrick,
the appellant here raised on direct examination the issue the prosecutor was attempting to counter during cross-examination: that he had been attacked with a bottle and the victim’s fists.
In
United States v. Shue,
766 F.2d 1122, 1130-31 (7th Cir.1985), we found a
Doyle
violation in the prosecutor’s use of defendant’s failure to tell police officers that he was framed for the crime as evidence that he was not framed.
We found that the prosecutor’s cross-examination “was not limited to the permissible use of impeaching appellant’s credibility”, nor was the government’s closing argument, which stated: “He refused to talk to the FBI, refused. And no one ever heard of this preposterous, incredible story of a frame until he hit the witness stand.”
Id.
Such use of a defendant’s
post-Miranda
silence is clearly inconsistent with a defendant's claim of innocence in a way which the present case is not, and was more than an “inextricably intertwined” inference of guilt.
Id.
at 1131.
Where a defendant claims self-defense at trial, it would be a violation of
Doyle
for the state to point out that the defendant, after receiving
Miranda
warnings, did not tell the police officers that he had acted in self-defense, and use that to infer that defendant had indeed not acted in self-defense. In
Allen,
659 F.2d at 746-47, we found unconstitutional the prosecutor’s use of defendant’s
post-Miranda
silence in failing to tell police officers he had acted in self-defense in shooting his wife as evidence that defendant did not act in self-defense.
See also Saulsbury,
702 F.2d at 655. Such a scenario is not present here, however. The prosecutor here made no reference to appellant’s failure to tell bystanders, arresting officers, or his jailors that he acted in self-defense when he stabbed the victim. Rather, the prosecutor only referred to appellant’s failure to seek medical attention, and argued that such failure contradicted appellant’s testimony that he had been attacked by the victim’s fists and possibly a bottle in the victim’s possession.
Allen
and
Saulsbury
are therefore clearly distinguishable.
In sum, we do not believe it is fundamentally unfair for a prosecutor to make reference to a defendant’s failure to complain of injuries following his arrest and
Miranda
warnings, even where such reference may indirectly undermine defendant’s theory of self-defense. The prosecutor’s references
here therefore did not amount to a
Doyle
due process violation.
IV.
The district court properly rejected appellant’s constitutional challenges to the state court’s denial of his motion for a change of venue and to the prosecutor’s use of his post-arrest silence. The district court’s denial of appellant’s petition for a writ of habeas corpus is therefore
Affirmed.