dissenting:
Because I would adopt a clear error standard of review that would give deference to the trial judge’s decision regarding the existence or non-existence of a prima facie case of discriminatory jury selection, I respectfully dissent. Applying that standard to this case, I would conclude that the trial court did not commit clear error in finding that the defendant failed to make out a prima facie showing under Batson, and I would therefore affirm the court of appeals.
I.
Defining and adopting a standard of review is a critical part of the appellate function. Absent statutory directive or rulemak-ing, the standard is an outgrowth of the nature of the decision being reviewed— namely, whether it is a decision characterized by findings of fact or whether it is a conclusion of law.
Findings of fact are generally reviewed under a clear error or abuse of discretion standard, whereas conclusions of law are generally reviewed under a de novo standard. The reasons are straightforward. De novo means “anew; afresh; a second time.” Black’s Law Dictionary 392 (5th ed.1979). Indeed, appellate courts can and should review anew the question of whether a trial court reached the correct conclusion of law, or the right of appeal would be essentially meaningless. On the other hand, the appellate courts defer to the factual findings of the trial court because the trial judge is in the courtroom, and is charged with the duty to find facts. Appellate courts may not undertake fact-finding.
The tension arises between these two types of review when the issue before the appellate court is a mixed question of fact and law. Under those circumstances, the court may take a number of different approaches.1 The court may treat the ultimate conclusion as one of fact for purposes of review and apply the clear error standard. See, e.g., Rascon v. U.S. West Communications, Inc., 143 F.3d 1324, 1333 (10th Cir.1998); Holl, 54 F.3d at 648; Doss v. Doss, 77 Colo. 262, 266, 236 P. 129, 131 (1925). Alternatively, the court may conclude that a mixed question of fact and law demands de novo review. See, e.g., Lewis v. Colorado Rockies Baseball Club, Ltd., 941 P.2d 266, 271 (Colo.1997). Lastly, the court may review the findings of fact for clear error and still look de novo at the legal conclusions that the trial court drew from those factual findings. See, e.g., People v. Smith, 926 P.2d 186, 188 (Colo.App.1996).
*599II.
With that background in mind, I turn to the issues before us in this case. The procedures set out by the United States Supreme Court in Batson v. Kentucky, 476 U.S. 79, 106 S.Ct. 1712, 90 L.Ed.2d 69 (1986), protect a defendant against the exercise of peremptory challenges by the prosecutor in a manner that violates the Equal Protection Clause.2 For purposes of this discussion, the offending challenges are those that are racially motivated.
The three-part test adopted in Batson both instructs and empowers the trial judge to safeguard the neutrality of the process by which peremptory challenges are exercised. In step one, the defendant must make a prima facie showing to the trial court of purposeful race-based discrimination in jury selection. Specifically, the defendant bears the burden of showing that (1) the prosecution struck from the jury a member of a cognizable racial group and (2) the totality of the relevant facts and circumstances give rise to an inference of purposeful discrimination. See Batson, 476 U.S. at 96-98, 106 S.Ct. 1712. The trial court must determine at step one whether the defendant has succeeded in establishing such a prima facie ease. Second, the burden shifts to the prosecution to state a race-neutral explanation for the challenge. The reasons need not be persuasive, and the trial court must accept the reasons as true. Only if the prosecutor’s discriminatory intent is apparent in the explanation will the court cut off the process and grant the Batson request. See Purkett v. Elem, 514 U.S. 765, 768, 115 S.Ct. 1769, 131 L.Ed.2d 834 (1995). Third, the defendant has the opportunity to rebut the prosecutor’s race-neutral explanations, showing them to be mere pretext used to obscure discrimination, and then the trial court rules on Batson objection. Batson is grounded in the assumption that the trial judge presiding in the courtroom is in the best position to guard against discrimination. Indeed, the Supreme Court stated that “[w]e have confidence that the trial judges, experienced in supervising voir dire, will be able to decide if the circumstances concerning the prosecutor’s use of peremptory challenges creates a prima facie case.” Batson, 476 U.S. at 97, 106 S.Ct. 1712.
Discrimination is as sly as it is insidious. It lives in inference, tone, and gesture as much as in action. The trial judge is the judicial officer who watches and listens as voir dire unfolds, and who can discern the presence or absence of discriminatory intent. Indeed, whether a prima facie case of discriminatory exercise of peremptory challenges exists is a question of fact, or, at the very least, “fact-intensive.” United States v. Moore, 895 F.2d 484, 485 (8th Cir.1990).
The Supreme Court has defined the standard of review governing the trial court’s ultimate determination of whether the peremptory challenge was motivated by discriminatory intent as clear error. The Court states that treating “intent to discriminate as a pure issue of fact, subject to review with a deferential standard, accords with our treatment of that issue in other equal protection cases.” Hernandez v. New York, 500 U.S. 352, 364, 111 S.Ct. 1859, 114 L.Ed.2d 395 (1991). Hence, only if the appellate court is convinced that the trial court’s conclusion was clearly erroneous will it overturn the decision. Id. In reaching that conclusion, the Court rejected a mixed law and fact approach:
Petitioner advocates “independent” appellate review of a trial court’s rejection of a Batson claim. We have difficulty understanding the nature of the review petitioner would have us conduct. Petitioner explains that “independent review requires the appellate court to accept the findings of historical fact and credibility of the lower court unless they are clearly erroneous. Then based on these facts, the appellate court independently determines whether there has been discrimination.” But if an appellate court accepts a trial court’s find*600ing that a prosecutor’s race-neutral explanation for his peremptory challenges should be believed, we fail to see how the appellate court nevertheless could find discrimination.
Id. at 366-67, 111 S.Ct. 1859 (citation omitted).
Later, in Purkett v. Elem, 514 U.S. 765, 115 S.Ct. 1769, 131 L.Ed.2d 834 (1995), the Court addressed a circumstance in which a defendant raised a Batson challenge to which the prosecution offered a race-neutral explanation relating to the challenged juror’s hair length and facial hair. The trial court overruled the defendant’s Batson objection. On habeas corpus appeal, the Court of Appeals for the Eighth Circuit reversed and remanded on the basis that the factors articulated by the prosecution were pretextual and not legitimate. The Supreme Court reversed the court of appeals, holding that the race-neutral explanation need not be related to the specific case or even legitimate. The Court clarified that the credibility determination was a factor to be taken into account at the third stage of the Batson analysis, not at the second stage where even an implausible or fantastic reason would serve to shift the burden of persuasion back to the opponent of the strike.3 The Court reaffirmed that the ultimate determination of racial motive was a finding of fact presumed to be correct only if not fairly supported by the record.
Hence, the two themes of the Supreme Court pronouncements in the Batson context are that discrimination is principally an issue of fact, and that deference to the trial court is therefore appropriate.
For all of these reasons, in my view, the appropriate standard of review for appellate oversight of a trial court’s determination of a prima facie case is clear error. In reaching that conclusion, I subscribe to the reasoning of numerous other courts that have addressed the issue. See, e.g., United States v. Stewart, 65 F.3d 918 (11th Cir.1995), cert. denied sub nom. Daniel v. United States, 516 U.S. 1134, 116 S.Ct. 958, 133 L.Ed.2d 881 (1996); United States v. Bergodere, 40 F.3d 512, 516 (1st Cir.1994), cert. denied, 514 U.S. 1055, 115 S.Ct. 1439, 131 L.Ed.2d 318 (1995); United States v. Vasquez-Lopez, 22 F.3d 900 (9th Cir.1994), cert. denied, 513 U.S. 891, 115 S.Ct. 239, 130 L.Ed.2d 162 (1994); United States v. Branch, 989 F.2d 752 (5th Cir.), cert. denied sub nom. Thompson v. United States, 509 U.S. 931, 113 S.Ct. 3060, 125 L.Ed.2d 742 (1993); United States v. Casper, 956 F.2d 416 (3d Cir.1992); United States v. Moore, 895 F.2d 484 (8th Cir.1990); and United States v. Grandison, 885 F.2d 143 (4th Cir.1989). Indeed, the Tenth Circuit is the only federal circuit that has expressly embraced the de novo standard of review on Batson step one, and even the Tenth Circuit’s standard is tempered by a clear-error standard for issues of fact observed by the trial court. See United States v. Hartsfteld, 976 F.2d 1349, 1355-56 (10th Cir.1992) (reviewing Batson step one de novo, but “giving deference to the trial court’s first-hand observation of the circumstances of each case”).
The seven federal circuits that have rejected the de novo standard of review have expressed similar reasoning, based on the fact-intensive nature of Batson step one and on the advantages of allowing the trial court to serve as a fact-finder. For example, in United States v. Stewart, 65 F.3d 918, 923 (11th Cir.1995), the Eleventh Circuit Court of Appeals concluded that
a district court’s superior ability as a Bat-son factfinder stems from two advantages it has over an appellate court: the posi*601tional advantage of being there among the facts as they unfold, and of seeing and hearing the explanations as they are given; and the experiential advantage of regularly being in the business of factfinding, which an appellate court is not.
Similarly, the Eighth Circuit Court of Appeals held in United States v. Moore, 895 F.2d 484, 486 (8th Cir.1990) that
[t]he trial judge, with his experience in voir dire, is in by far the best position to make the Batson prima facie ease determination. And, because of his unique awareness of the totality of the circumstances surrounding voir dire, that determination must be treated as a finding of fact entitled to great deference on review. De novo review of the record by this court would be inappropriate because the cold record is simply not enough.
III.
The Majority seeks guidance from Title VII cases in its determination of the appropriate standard of review for step one of Batson, noting that the initial burden on the Title VII plaintiff to raise an inference of illegal discrimination is similar to the prima facie standard under Batson. Although it is clear that the two situations are closely analogous, it is unclear that Furnco Construction Corp. v. Waters, 438 U.S. 567, 576, 98 S.Ct. 2943, 57 L.Ed.2d 957 (1978), the U.S. Supreme Court case cited by the Majority, requires de novo review under Title VII.
In Fumco, the Court distinguished between a factual finding of discrimination, analogous to the third prong of the Batson test, and the initial burden of the Title VII plaintiff to come forward with evidence of discrimination sufficient to raise an inference of illegal discrimination. The court of appeals had ruled that once the trier of fact concluded that a “prima facie showing had been made out, statistics of a racially balanced workforce were totally irrelevant to the question of motive.” Fumco Constr. Corp., 438 U.S. at 579, 98 S.Ct. 2943; see also Waters v. Furnco Constr. Corp., 551 F.2d 1085 (7th Cir.1977).
The Supreme Court reversed, noting that the scope of the court’s inquiry during the prima facie stage is limited: “A ... prima facie showing is not the equivalent of a factual finding of discrimination.... Rather, it is simply proof of actions taken by the employer from which we infer discriminatory animus _” Furnco Constr. Corp., 438 U.S. at 579, 98 S.Ct. 2943 (emphasis added). In other words, once the plaintiff showed that the employer undertook certain types of suspect actions, a rebuttable presumption or “inference” that these actions were the result of discriminatory motive would obtain, and the burden of evidentiary production would shift to the defendant.4 Because the prima facie stage of a Title VII action requires only such evidence as would be necessary to raise a presumption of discriminatory intent, the trial court’s finding at that stage does not represent “the equivalent of a factual finding of discrimination.” The Court therefore permitted the defendant to introduce evidence, such as Furneo’s workforce statistics, later in the proceeding to rebut the initial presumption of discrimination. Hence, the Court’s discussion of inferences was not intended to brand the prima facie stage as a question of law, but rather was incidental to a discussion of when statistical evidence could be properly admitted in the context of a shifting eviden-tiary burden.
Subsequently, however, the words “we infer discriminatory animus” have been applied to support a de novo standard of review in Title VII cases. The Ninth Circuit, for instance, in a ease cited by the Majority, held without further discussion that “[s]ince the Supreme Court has stated that ‘a ... prima facie showing is not the equivalent of a factual finding of discrimination,’ we are content in this case to apply a de novo standard of review.” Gay v. Waiters’ & Dairy Lunchmen’s Union, Local No. 30, 694 F.2d 531, 546 (9th Cir.1982) (citation omitted).
Rather than take such an attenuated reading of Fumco, I would heed the warning of *602the Supreme Court in an earlier discrimination case:
All courts have recognized that the question facing triers of fact in discrimination cases is both sensitive and difficult.... But none of this means that trial courts or reviewing courts should treat discrimination differently from other ultimate questions of fact. Nor should they make their inquiry even more difficult by applying legal rules which were devised to govern “the allocation of burdens and order of presentation of proof,” in deciding this ultimate question.
United States Postal Serv. Bd. of Governors v. Aikens, 460 U.S. 711, 716, 103 S.Ct. 1478, 75 L.Ed.2d 403 (1982) (citation omitted).
Hence, the Title YII precedent does not dissuade me from the conclusion that a clear error standard of review at the prima facie stage of a Batson challenge is appropriate.
IV.
Using the clear error standard, I view the trial court’s decision that the defendant had not established a prima facie case to survive scrutiny.
First, I agree with the Majority that the pattern of peremptory strikes in light of the ultimate composition of the jury does not support discriminatory intent. See maj. op. at 593-594. Specifically, the final jury consisted of 33% African-American jurors, whereas the venire that was passed for cause only consisted of 26% African-American jurors. One of the three African-American jurors challenged by the prosecutor was replaced by another African-American juror— a fact that the prosecutor knew in advance because of the method of jury selection in use. Hence, the statistical evidence does not support the existence of a prima facie case.
The defendant advances two additional arguments in support of a prima facie showing, neither of which was presented to the trial judge. He points to the similarity in answers on questionnaires filled out by a white juror who was not excused and an African-American juror who was excused. I agree with the Majority that it was incumbent upon the defendant to raise this argument initially in making his Batson challenge.
The other argument that the defendant now makes is that the prosecutor’s remarks to the jury prior to the commencement of voir dire reveal discriminatory intent. The Majority notes that the remarks were part of the totality of the circumstances to be considered by the trial judge in making a prima facie determination and then holds that the prosecutor’s statements, when combined with his pattern of peremptory strikes, establish a prima facie case under Batson. Yet, the Majority recognizes that “the prosecutor’s opening statements may very well have been an attempt to prevent possible racial bias of the potential members of the jury, rather than an indication of his own racial bias.” Maj. op. at 596.
Here, the standard of review adopted by the court plays a major role. The recognition that the opening remarks could be interpreted two different ways acknowledges the critical role of the trial judge. Indeed, under the clear error standard that I would adopt, “[wjhere there are two permissible views of the evidence, the fact finder’s choice between them cannot be clearly erroneous.” Anderson v. Bessemer City, 470 U.S. 564, 574, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985). The decision of the lower court will only be overturned if, upon review of the entire record, the reviewing court is “left with the definite and firm conviction that a mistake has been made.” United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 92 L.Ed. 746 (1948).
The statistical make-up of the venire, the number and effect of the challenges, and the ultimate composition of the panel do not add up to a prima facie case. The element that, in the Majority’s view, tips the balance to a prima facie ease is an element that the Majority acknowledges can be seen two different ways. The trial judge was in the courtroom. She witnessed and heard the voir dire. She concluded that there was no prima facie case.
I find no clear error in that decision.
*603V.
Lastly, I am not troubled by the fact that the trial judge proceeded beyond the prima facie stage of the Batson analysis even after having found no prima facie case. Indeed, the procedure she used was curious and not optimal. However,' the prima facie determination is neither moot5 nor clouded by the subsequent completion of the record.
It is good practice for the trial judge to articulate the reasons for a finding of no prima facie case on the record, with reference to the underlying facts supporting that conclusion. See Moore, 895 F.2d at 486. I do not view this trial judge as having done anything more than just that. She supported her own conclusion and then permitted the prosecutor to supplement the record for appellate purposes. She did not, in my view, stray impermissibly into the second or third stages of the Batson analysis. She found the showing by the defendant to be inadequate even to shift to the prosecution the burden of coming forward with race-neutral reasons. The record supports that conclusion and I would end the inquiry there.
VI.
I agree with the reasoning and the result reached by the court of appeals in this case. I therefore respectfully dissent from the Majority opinion.
I am authorized to state that Justice HOBBS and Justice RICE join in this dissent.
. See Holl v. Commissioner, 54 F.3d 648, 650 (10th Cir.1995).
. Subsequent case law has clarified that not only does the defendant have the right to challenge any prosecutorial efforts to select the jury on racial grounds, but the prosecutor has the same right to object to the defendant’s challenges. See Georgia v. McCollum, 505 U.S. 42, 59, 112 S.Ct. 2348, 120 L.Ed.2d 33 (1992). Further, jurors themselves have a right to serve without being removed for discriminatory reasons. Id. at 48, 112 S.Ct. 2348.
. This portion of the opinion has been read as a pronouncement that the standard of review at the second stage of Batson is de novo. Although I would not leap to that conclusion, I do not take issue with it here. The second stage of Batson is a narrow, circumscribed point in the proceedings. If the trial court finds that the prosecutor’s explanation of the challenge at issue is racially motivated, such a determination would be viewed as an ultimate finding of fact — whether it technically occurred at step two or after an opportunity for defense comment. If, on the other hand, the trial court finds that the explanation is race-neutral, only the statement itself is at issue and an appellate court can read the text of the explanation and revisit the trial court's conclusion. Most probably, in the event that the trial court finds the explanation to be race-neutral, it will proceed to step three and once again reach an ultimate determination, reviewable under the clear error standard. Hence, the standard imposed at stage two is of little consequence in practice.
. See United States Postal Serv. Bd. of Governors v. Aikens, 460 U.S. 711, 714, 103 S.Ct. 1478, 75 L.Ed.2d 403 (1983) (noting that by establishing a prima facie case, a plaintiff in a Title VII action creates a rebuttable presumption of unlawful discrimination).
. I agree completely with the Majority opinion on this issue. See maj. op at section III.