The opinion
I concur. I write specially only to note that Houston v. State , 354 So.2d 825 (Ala.Crim.App. 1977), relied on in Mixon v. State , 596 So.2d 605 (Ala.Crim.App. 1991), as primary authority for its three-pronged test to determine the existence of harmless error, does not stand for the proposition for which it was cited in Mixon . In Houston , the defendant's exculpatory story was not transparently frivolous, the evidence of his guilt was not overwhelming, and the prosecutor made multiple references to the fact of the defendant's silence. Nonetheless, this Court stated:
"On four separate occasions throughout the course of the trial reference was made to the fact of the defendant's silence. Only on one of those occasions did the trial judge sustain defense counsel's objection and instruct the jury. Had the trial court sustained the objections, instead of overruling, and admonished the jury that the argument was improper, the prosecutor's comments would not have constituted reversible error. Adkins v. State , 265 Ala. 666 , 93 So.2d 522 (1957)."
354 So.2d at 829 (emphasis added).
It is therefore apparent to me that the Houston Court did not intend to cite Chapman v. United States , 547 F.2d 1240 (5th Cir. 1977), for the same proposition that the Mixon Court apparently cited it — to establish a three-pronged test that would govern every case. Accordingly, I agree with the main opinion's conclusion that the determination as to whether a Doyle v. Ohio , 426 U.S. 610 , 96 S.Ct. 2240 , 49 L.Ed.2d 91
(1976), violation is harmless should be made on a case-by-case basis using as a guide the three factors discussed in Houston
and Mixon .