MANNHEIMER

Affirmed.

Renee McFarland, Assistant Public Defender, and Quinlan Steiner, Public Defender, Anchorage, for the Appellant. Diane L. Wendlandt, Assistant Attorney General, Office of Special Prosecutions and Appeals, Anchorage, and Michael C. Geraghty, Attorney General, Juneau, for the Appellee.

Before: MANNHEIMER, Chief Judge, ALLARD, Judge, and HANLEY, District Court Judge.

OPINION

MANNHEIMER, Judge.

Mark D. Anderson was convicted of ten counts of second-degree sexual abuse of a minor,

based on evidence that he engaged in multiple instances of sexual contact with three different girls, each under the age of eleven. Anderson appealed his convictions on various grounds, but this Court affirmed his convictions in Anderson v. State, 289 P.3d 1 (Alaska App.2012).

The Alaska Supreme Court has now directed us to reconsider certain aspects of our decision.

The matters to be resolved all arise from the fact that the indictment against Anderson contained many counts that did not allege a specific date for the charged offense, but rather a range of dates.

As is often true in prosecutions for sexual abuse of a minor, the three girls in this case testified that Anderson engaged in sexual contact with them on numerous occasions but, for the most part, the girls were unable to identify the dates of the individual acts of sexual contact. Given the girls' testimony, Anderson argued that six of the counts in his indictment contained ranges of dates that were broad enough to potentially encompass two or more alleged acts of sexual contact—thus giving rise to the possibility that the jurors never reached unanimous agreement as to the criminal incident that formed the basis for their guilty verdicts on those six counts.

In cases like this, Alaska law requires that the jurors unanimously agree on the particular episode of criminal conduct that forms the basis for a guilty verdict.

This would not have been a problem if Anderson's jurors had been instructed on this requirement of factual unanimity—but they were not.

Anderson's trial judge neglected to instruct the jurors that, with respect to each count, they could not convict Anderson unless they unanimously agreed on the particular conduct underlying that count. Anderson's attorney did not request such a unanimity instruction, nor did he object to the judge's failure to give such an instruction. But on appeal, Anderson argued that his trial judge committed plain error by failing to give the jurors a factual unanimity instruction.

This Court agreed with Anderson that the judge's failure to give a factual unanimity instruction was obvious error,

but we concluded that this error did not rise to the level of “plain error” for two reasons.

First, we concluded that Anderson's attorney had potential tactical reasons for failing to raise this issue.

Second (and alternatively), we concluded that the lack of a factual unanimity instruction was harmless beyond a reasonable doubt—that there was no reasonable possibility that the jury's verdicts would have been different if the jurors had been properly instructed on the requirement of factual unanimity—because Anderson's defense at trial was a blanket denial of wrongdoing, coupled with the assertion that all of the girls' allegations were knowingly false, either because of ill will or as the result of adult pressure.

The supreme court has directed us to reconsider both aspects of our ruling. To aid our reconsideration of these issues, we solicited supplemental briefs from the parties.

Identifying the proper test for assessing whether the jury instruction error was harmless

Although Covington v. State

was the seminal case that established the requirement of factual unanimity in sexual abuse cases in Alaska, this Court ultimately applied the wrong test when we assessed whether the lack of a factual unanimity jury instruction was reversible error in Covington's case.

Initially, this Court reversed Covington's sexual abuse convictions because the jury was not told of the need for factual unanimity.

However, the State sought rehearing, arguing that (1) Covington did not raise the jury unanimity issue in the trial court, so Covington was required to show plain error; and (2) the jury instruction was not plainly erroneous under the facts of Covington's case.

On rehearing, this Court reinstated Covington's convictions because we agreed with the State that Covington had failed to show that the jury instruction error prejudiced the fairness of his trial. We relied primarily on the fact that Covington had not presented individual challenges to specific acts of sexual misconduct, but rather had presented a blanket defense that none of the alleged sexual abuse happened. State v. Covington ( Covington II ), 711 P.2d 1183, 1184–85 (Alaska App.1985).

However, our decision in Covington II was premised on a particular view of the doctrine of plain error—a view that we adopted in an earlier case, Van Hatten v. State, 666 P.2d 1047 (Alaska App.1983).

In Van Hatten, this Court held that when a defendant presents a claim of constitutional error for the first time on appeal ( i.e., when the issue is raised as a claim of plain error), an appellate court should not apply the “harmless beyond a reasonable doubt” test to determine whether the constitutional error requires reversal. Instead, this Court held that the “harmless beyond a reasonable doubt” standard applied only to cases “where errors of constitutional dimension are preserved for appeal by timely objection.”

We declared that in all other cases ( i.e., cases where the claim of constitutional error was not preserved in the trial court) the “prejudice” prong of the plain error doctrine “demand[ed] the application of a standard [less favorable to the defendant] than the harmless beyond a reasonable doubt test”.

And we identified this less favorable standard as the “appreciably affected the verdict” test—the test that the Alaska Supreme Court adopted in Love v. State, 457 P.2d 622, 630–32 (Alaska 1969), for evaluating the effect of non-constitutional errors.

In Covington II, we expressly relied on Van Hatten as the governing law on the question of how to evaluate the impact of the jury instruction error ( i.e., the lack of a factual unanimity instruction), given the fact that Covington failed to object to this error in the trial court. Thus, in Covington II, when we assessed whether the lack of a factual unanimity instruction prejudiced the fairness of Covington's trial, we did not apply the “harmless beyond a reasonable doubt” test. Instead, we applied the “appreciably affected the verdict” test—and, under this test, we concluded that the error was harmless.

But in Adams v. State, 261 P.3d 758, 773 (Alaska 2011), and again in Khan v. State, 278 P.3d 893, 901 (Alaska 2012), the Alaska Supreme Court held that even when a claim of constitutional error is raised for the first time on appeal ( i.e., when it is raised as a claim of plain error), the “harmless beyond a reasonable doubt” test continues to govern the question of whether the error (if proved) requires reversal of the defendant's conviction. Indeed, the Adams decision expressly disapproved our contrary holding in Van Hatten. Adams, 261 P.3d at 772–73.

Therefore, in resolving Anderson's case, we must apply the “harmless beyond a reasonable doubt” test set forth in Adams and Khan, rather than the “appreciably affected the verdict” test that we applied in Covington II. We are required to reverse Anderson's convictions on the six counts that he challenges for potential lack of jury unanimity—Counts II and III, and Counts VI through IX—unless the State demonstrates that the jury instruction error was harmless beyond a reasonable doubt.

What does “harmless beyond a reasonable doubt” mean in situations where the jury instructions omit or materially misdescribe an essential component of the decision the jury must make?

The jury instructions in Anderson's case informed the jurors that they were required to reach unanimous agreement as to Anderson's guilt or innocence on each count of the indictment. But the jury instructions omitted an important component of what “unanimous agreement” entailed: the instructions failed to specify that, with respect to each count, the jurors could not return a guilty verdict unless they unanimously agreed on a particular incident of sexual contact.

Although this is an error of constitutional dimension, Anderson and the State agree that the error does not require automatic reversal.

Instead, to determine whether this error requires reversal of Anderson's convictions on the six affected counts, we must assess whether the error is harmless beyond a reasonable doubt.

When the Alaska Supreme Court remanded Anderson's case to us for reconsideration, the supreme court directed us to determine “whether the trial court's failure to provide [a factual unanimity] instruction was harmless beyond a reasonable doubt under either the Covington II approach or the ... approach [adopted by the United States Supreme Court in Neder v. United States

]”.

We will discuss the Neder decision in a moment. But the decision in Covington II is irrelevant to this inquiry.

This Court's decision in Covington II offers no guidance on the question of when an error can be deemed harmless beyond a reasonable doubt—because, as we explained in the preceding section of this opinion, this Court did not employ the “harmless beyond a reasonable doubt” test in Covington II. Instead, we employed the “appreciably affected the verdict” test—the same test that applies to non-constitutional errors—because Covington failed to raise the jury instruction issue in the trial court.

It is now clear, after the supreme court's decisions in Adams and Khan, that the approach taken in Covington II was wrong—that even when a claim of constitutional error is raised for the first time on appeal, Alaska law requires an appellate court to apply the “harmless beyond a reasonable doubt” test when evaluating the effect of the error.

We turn, then, to the question of what “harmless beyond a reasonable doubt” means in a case like Anderson's, where the problem is that the jury instructions omitted or materially misdescribed an essential component of the jury's decision.

(a) A note regarding terminology

Before we begin a detailed examination of this question, we wish to clarify the terminology that we will be using.

The United States Supreme Court first articulated the “harmless beyond a reasonable doubt” standard in Chapman v. California, 386 U.S. 18, 87 S.Ct. 824, 17 L.Ed.2d 705 (1967). However, the Chapman opinion actually uses three phrases to describe the test for whether a constitutional error requires reversal of a criminal conviction.

Toward the end of its discussion of this point, the Supreme Court phrased the test as whether the error was “harmless beyond a reasonable doubt”.

But earlier in the same paragraph, the Supreme Court phrased the test as “whether there is a reasonable possibility that the [error] complained of might have contributed to the conviction.”

And later, still within that same paragraph, the Court phrased the test as whether the error “possibly influenced the jury adversely to [the] litigant”.

The Supreme Court cautioned its readers not to seek fine distinctions among these phrasings. In particular, the Court declared (again, in the same paragraph) that there was “little, if any, difference” between asking “whether there is a reasonable possibility that the [error] contributed to the conviction” and “requiring the beneficiary of a constitutional error to prove beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained.”

We, too, will use these phrasings interchangeably. At different points in our discussion, we may refer to the State's burden to show beyond a reasonable doubt that the verdict would have been the same, or (alternatively) we may refer to the State's burden to negate any reasonable possibility that the verdict would have been different. We mean the same thing by these phrasings. They are both intended to embody the Chapman test for assessing the effect of constitutional error.

(We note that the Montana Supreme Court has held that the “no reasonable possibility” phrasing of the test is actually preferable to the “beyond a reasonable doubt” phrasing—because “there is little, if any, difference between these two standards in terms of the burden on the State”, and because the “beyond a reasonable doubt” standard normally applies to the fact-finding of a jury, thus falsely suggesting that appellate courts “sit as a fact-finder when evaluating a constitutional violation for harmlessness.” State v. Matt, 347 Mont. 530, 199 P.3d 244, 253 (2008).

) (b) The Supreme Court's decision in Neder v. United States

In Neder v. United States, 527 U.S. 1, 119 S.Ct. 1827, 144 L.Ed.2d 35 (1999), the trial judge improperly instructed the jurors that one element of the offense—the materiality of Neder's false statements—was not to be decided by the jury, but rather was to be decided by the judge.

The jury therefore reached no decision on this element.

To resolve Neder's appeal, the Supreme Court had to decide two issues that are relevant to Anderson's case: First, does a trial judge's failure to instruct the jury on a necessary element of the offense automatically require reversal of the defendant's conviction, or is this type of error subject to harmless error analysis on appeal? And second, if this type of error can be harmless, what is the test for evaluating harmlessness in this context?

The Supreme Court first held that a judge's failure to have the jury decide every element of the offense does not require automatic reversal, but rather is the type of error that is subject to harmless error analysis.

Next, the Supreme Court took up the issue of what “harmlessness” means in this context.

The defendant in Neder argued that, even though there was overwhelming evidence that his false statement was material, it would nevertheless be improper to categorize the error in the jury instructions as harmless beyond a reasonable doubt.

Neder pointed out that his trial judge affirmatively instructed the jurors not to consider the materiality of Neder's statements. Thus, it was clear from the record that the jurors never even discussed, much less resolved, the issue of materiality. Because of this, Neder argued that it would be improper for an appellate court to declare the error harmless beyond a reasonable doubt, even if the evidence of materiality was overwhelming—because (according to Neder) the appellate court would, in effect, be deciding the element of materiality in the first instance.

The Supreme Court rejected Neder's argument, viewing it as “simply another form of the argument that a failure to instruct [the jury] on any element of the crime is not subject to harmless-error analysis.”

The Court held that in these circumstances— i.e., instances where an element of the offense is wrongly removed from the jury's consideration—the question is whether, given the evidence at trial and the way the case was litigated, the appellate court can “conclude beyond a reasonable doubt that the jury verdict would have been the same absent the error” ( i.e., if the jury had been asked to decide the omitted element of the offense). 527 U.S. at 18–19, 119 S.Ct. at 1838.

In the next paragraph of the Neder decision, the Court restated this test—but this time using the converse formulation: the Court declared that the defendant's conviction should be reversed “[if] the record contains evidence that could rationally lead [the jury] to a contrary finding with respect to the omitted element”. 527 U.S. at 19, 119 S.Ct. at 1839.

(c) Our discussion of this point of law in our earlier decision in Anderson's case, and why we re-affirm our conclusions

When Anderson's appeal was first presented to us, the parties agreed that Anderson's trial judge committed obvious error by failing to instruct the jury on the requirement of factual unanimity. This raised a question of whether the error was harmless.

With respect to six of the counts in the indictment, Anderson argued that this lack of a factual unanimity instruction could not be deemed harmless. Anderson pointed out that, according to the testimony of the three girls, many acts of sexual abuse occurred during the time spans covered by the six counts. Thus, he argued, there was at least a reasonable possibility (if not an outright probability) that the jurors did not reach unanimous agreement as to the particular incident(s) of sexual contact underlying each of those six guilty verdicts.

In our earlier decision in this case, we acknowledged that there was reason to suspect that the jurors who decided Anderson's case never unanimously identified a particular incident of sexual contact for each of those six counts.

But we held that this fact was not determinative of whether the error (the lack of the jury instruction) was harmless beyond a reasonable doubt.

As shown by the Supreme Court's decision in Neder, the question is not whether the record demonstrates that, despite the error in the jury instructions, the jury nevertheless reached unanimous agreement on every essential element of the offense charged.

If that were the rule, then the Supreme Court would unquestionably have reversed Neder's conviction—because Neder's trial judge did not simply fail to tell the jurors about the element of materiality; instead, he affirmatively instructed the jurors not to consider this element.

Thus, it was almost certain that Neder's jurors failed to reach a decision (unanimous or otherwise) concerning the element of materiality.

But under the Neder test, the question is not whether the State can show beyond a reasonable doubt that, despite the error in the jury instructions, the jurors somehow figured out that they had to reach unanimous agreement on the component of their verdict that was omitted from the jury instructions. Rather, the question is whether the State can show beyond a reasonable doubt that, if the jury had been properly instructed, they would have returned the same verdict.

(Or, phrased another way, the State must show that there is no reasonable possibility that the jury, if properly instructed, would have returned a different verdict.)

In Anderson's supplemental brief to this Court, he criticizes this approach and declares that it is inconsistent with existing Alaska law.

More specifically, Anderson argues that the Neder approach is flawed because it “asks the counter-factual question [of] whether the defendant would have been convicted in a hypothetical trial absent the error.” Anderson declares that the inquiry should instead focus on “whether the error was a substantial factor in the jury's verdict”.

We agree with Anderson that the question is whether the error affected the jury's verdict—or, more precisely, whether there is a reasonable possibility that the error affected the jury's verdict. But to answer that question, a court must necessarily ask another: Might the jury have returned a different verdict if they had been properly instructed?

Almost certainly, the course of the jury's discussions in Anderson's case ( i.e., the specific content of their deliberations) would have been different if the jurors had known that they were required to unanimously agree on one or more particular incidents of sexual contact before returning a guilty verdict on any count.

But the issue is not whether the course of the jury's deliberations might have been different. Instead, the issue is whether the outcome of their deliberations might have been different. We must decide whether the State has shown, beyond a reasonable doubt, that the outcome of the jury's deliberations would have been the same even if the error had not occurred— i.e., even if the jurors had been apprised of the requirement of factual unanimity.

Anderson objects that this process leads us into “hypothetical” and “counter-factual” inquiries. But if we are to adhere to the principle that jury instruction errors do not automatically require reversal, and that these errors can potentially be harmless, this is the only practical way to perform the harmless error analysis.

Anderson raises one other objection to the way we approached the question of harmless error in our earlier decision.

Anderson contends that court decisions in this area of law can be divided into two camps, each adopting a distinct approach to the question of harmless error. The first of these approaches focuses on the effect that the error had on the jury, while the second approach focuses on whether the evidence at trial clearly demonstrated the defendant's guilt.

According to Anderson, the courts that follow this second approach will overlook constitutional errors in the lower court proceedings, or will declare these errors to be harmless, as long as the evidence of the defendant's guilt is “overwhelming”. And he criticizes this second approach as flawed—as fundamentally inconsistent with the principle that all criminal defendants are entitled to procedural fairness, no matter how guilty they may be.

We agree with Anderson in this matter. (And we question whether any appellate courts are truly, or at least wittingly, in the second camp he describes.)

But Anderson makes a further assertion: he contends that, in this Court's earlier decision in his case, we endorsed the second camp's approach. Nor is Anderson alone in this assertion: the State agrees with Anderson that, in our earlier decision, this Court adopted an “overwhelming evidence of guilt” approach to the question of harmless error.

The parties' characterization of our earlier decision reminds us of how difficult it can be to describe these legal concepts in a manner free from imprecision and ambiguity. It appears that our earlier opinion in Anderson's case was wanting in this regard.

In our earlier opinion, we did not rely on the theory that the evidence against Anderson was so overwhelming, and that Anderson's guilt was so evident, that no error—not even a constitutional error—could possibly require reversal of his convictions.

Rather, when we explained why we concluded that the jury instruction error in Anderson's case was harmless beyond a reasonable doubt, we focused on the arguments that Anderson raised in his defense—arguments that the accusations against him were totally false, and that the three girls had made these accusations either as the result of ill-will or emotional turmoil, or as the result of pressure from overly suspicious adults. Anderson v. State, 289 P.3d at 7–8. We concluded that, given the evidence in Anderson's case, and given the tenor of Anderson's defense, there was no reasonable possibility that the jury's verdicts would have been different even if the jurors had been properly instructed on the need for factual unanimity. Id. at 8.

This is not to say that the strength of the State's evidence is irrelevant to the assessment of whether a constitutional error is harmless. But if the strength of the State's evidence is to be relevant, it must be relevant to the particular issue or issues affected by the constitutional error.

For instance, in cases where an element of the offense is either unwittingly omitted from the jury instructions or, as in Neder, expressly removed from the jury's consideration, the strength of the State's evidence on that element would be relevant to the assessment of whether there was a reasonable possibility that the jury, if properly instructed, would have reached a different decision. Thus, in Neder, the Supreme Court considered the fact that no one disputed the materiality of Neder's false statements, and that the evidence of the statements' materiality was overwhelming.

This is not the same as saying that the error should be ignored because, overall, the evidence of Neder's guilt was compelling. Rather, the strength of the evidence that Neder's false statements were material was relevant to the Supreme Court's assessment of whether the outcome of Neder's trial was potentially affected by the constitutional error—by the fact that the trial judge mistakenly directed the jurors not to consider whether Neder's false statements were material.

In Anderson's case, and in Covington, the error in the jury instructions was of a different kind, and the strength of the State's evidence had only a lesser relevance to this error. Instead, in both cases, the more important factors were the general tenor of the victims' testimony and the details of the defense that was offered at trial. These were the factors that allowed this Court to meaningfully assess whether the jury instruction error was harmless— i.e., whether the failure to inform the jury of the requirement of factual unanimity might have made a difference to the outcome of the trial.

(As we explained earlier, this Court used the wrong test in Covington II when we assessed whether the jury instruction error might have made a difference to the outcome of the trial: we mistakenly used the “appreciably affected the verdict” test that applies to non-constitutional error, rather than the “harmless beyond a reasonable doubt” test that applies to constitutional error. But the principle here is the same, regardless of whether the error is constitutional or non-constitutional.)

Why we again conclude that the jury instruction error in Anderson's case was harmless beyond a reasonable doubt

We now must determine whether the State has shown, beyond a reasonable doubt, that the jury instruction error in Anderson's case was harmless. Phrased another way (but intended to mean the same thing), we must decide whether there is a reasonable possibility that the verdicts reached by Anderson's jury would have been different if the jurors had received a proper instruction on factual unanimity.

As we noted earlier, the charges against Anderson involved three different young girls. Two of the girls (K.M. and A.K.) were sisters living in the same household, but the third girl (G.B.) lived in a separate household, and she was not acquainted with the other two girls.

The sexual abuse involving G.B. was reported first. Several months later, when the police were nearing completion of their investigation into the allegations involving G.B., the police received a separate, independent report that Anderson had sexually abused K.M. and A.K.

The only obvious link between the case involving G.B. and the case involving K.M. and A.K. was that each girl had a parent who was friends with Anderson through work. (The mother of G.B., and the father of K.M. and A.K., both worked at the same car dealership where Anderson was employed.)

At trial, Anderson's attorney argued that all of the accusations against Anderson were false. In his summation to the jury, the defense attorney pointed out that Anderson had never wavered in his denial of these charges, from the time he was first interviewed by the police. The defense attorney also pointed out that there was no medical evidence (or other physical evidence) to corroborate the girls' testimony, nor had any adult ever observed Anderson acting inappropriately with the girls.

With respect to the charges involving G.B., the defense attorney suggested that G.B. was a troubled young girl who was manifesting behavioral problems (her parents were separated at the time), and the defense attorney contended that G.B.'s accusation of sexual abuse was the product of a suggestion planted in her mind by an overly suspicious day-care provider.

According to the trial testimony, this day-care provider observed G.B. playing in a manner that, to him, contained an inappropriate sexual aspect. The day-care provider's observation prompted him to “ask [G.B.] several times what was wrong”. When G.B. did not immediately respond, the day-care provider told G.B., “You need to tell me who's touched you inappropriately, so I can fix it.” At that point, G.B. told him that “Mr. Andy [ i.e., Anderson] had done some things.”

Anderson's attorney argued that G.B.'s answer to the day-care provider was simply the product of suggestion—and that, as more and more adults became involved, G.B. felt that she could no longer retract what she had said.

To explain the apparently independent accusations that Anderson had sexually abused K.M. and A.K. (two sisters from a different family), the defense attorney noted that, after the investigation began into the sexual abuse of G.B., G.B.'s mother informed the managers of the car dealership where she worked—the same dealership where the father of K.M. and A.K. worked. The defense attorney suggested that K.M.'s and A.K.'s father overheard people talking about these allegations—and that, from there, “[it was] a small leap, at that point, to [K.M.'s] ears.”

K.M. was the older of the two sisters, and the defense attorney argued that she decided to fabricate her own accusations against Anderson simply because she “[wanted] to get some man in trouble.”

Defense Attorney: Our theory of the case [is that] it was a game [for K.M.]. She was having fun, [playing] a game [whose object was] getting a guy in trouble.
The defense attorney acknowledged that K.M.'s younger sister, A.K., also reported being sexually abused by Anderson, but the attorney suggested that A.K. simply went along with her sister's lie because K.M. was older and was “the leader”.

We recognize that the charges against Anderson were based on different acts of sexual contact that were factually distinct. The State's evidence concerning these acts varied somewhat in content and probative strength. Thus, it was logically possible that the jurors might believe that the State had proved some of these acts of sexual contact but not others. But even though this outcome may have been a logical possibility, it was not a reasonable possibility, given the way Anderson's case was litigated.

In their summations to the jury, Anderson's attorney and the prosecutor each presented one theory of the case. Anderson's attorney offered one blanket defense to all the charges against Anderson: the charges were false, the three girls knew that the charges were false, and Anderson was factually innocent of any wrong-doing. Likewise, the State's theory of the case was consistent as to each act of sexual contact. The prosecutor relied heavily on the fact that K.M.'s and A.K.'s descriptions of the abuse (and the time frame during which it occurred) dovetailed with the description of abuse given by G.B. The prosecutor argued that this was strong evidence of Anderson's guilt, because K.M. and A.K. were not acquainted with G.B. The prosecutor also argued that the three girls were too young to invent such charges, and that they had no motive to lie—but that Anderson did.

Given the parties' theories of the case, and given the evidence presented at trial, we conclude that even if Anderson's jury had been instructed on the need for factual unanimity, there is no reasonable possibility that the jury would have reached different verdicts. Accordingly, we again conclude (as we did in our previous decision) that the trial judge's error in failing to give a factual unanimity instruction was harmless beyond a reasonable doubt.

The question of whether Anderson's attorney made a tactical decision not to object when the trial judge failed to include a jury instruction on factual unanimity

When the supreme court remanded Anderson's case to this Court, the supreme court directed us to re-evaluate one additional issue: whether Anderson's claim of plain error should fail because Anderson's attorney might have made a tactical decision not to object to the omission in the jury instructions.

In our earlier decision, we concluded that Anderson's claim of plain error failed because the record supported an inference that Anderson's attorney had tactical reasons for not insisting on a jury unanimity instruction. Anderson, 289 P.3d at 5. But after considering the arguments presented in the parties' supplemental briefs, we conclude that our earlier decision may have been based on a mistaken legal approach to the question of “tactical decision”.

First of all, when a claim of error comes to an appellate court as a claim of “plain error”, the record of the trial court proceedings will rarely contain direct evidence of an attorney's tactical decision-making or reasoning. (In those few cases where the record does contain direct evidence of an attorney's desire to let an error occur, any later claim of error will often fall into the category of “invited error” rather than “plain error”.)

Moreover, appellate courts do not receive evidence and do not engage in fact-finding.

Thus, when an appellate court speaks of an attorney making a “tactical decision” not to object to an apparent error, the appellate court is not making a finding of historical fact about the attorney's state of mind. Rather, the appellate court is speaking of, and categorizing, the inferences that might reasonably be drawn from the record.

The next question is to identify the sorts of inferences that justify the label “tactical decision”.

In Anderson's supplemental brief to this Court, he argues that the label “tactical decision” should apply only when the record of the lower court proceedings affirmatively shows that the attorney (1) was consciously aware of the error, (2) deliberately refrained from objecting, and (3) was seeking an identifiable, tangible benefit by withholding an objection.

The State, on the other hand, argues that this proposed definition of “tactical decision” is too narrow, and that it conflicts with the presumption of attorney competence. More specifically, the State argues that, because of the presumption of attorney competence, an appellate court should presume that trial attorneys are aware of errors that are occurring in their presence, unless the record affirmatively demonstrates otherwise. Additionally, because of this same presumption of competence, the State argues that the label “tactical decision” should apply not only to situations where the attorney was seeking an identifiable, tangible benefit by failing to object, but also to situations where the attorney simply concluded that it was not worthwhile to object.

Although this debate may seem like the kind of topic that only appellate judges and lawyers would care about, the definition of “tactical decision” has significant consequences for our criminal justice system.

An overly broad definition of “tactical decision”— i.e., one that encompasses too many failures to object—means, in practice, that appellate courts will be enjoined to overlook serious, prejudicial errors in the criminal justice process. One of an appellate court's primary tasks is to ensure the procedural fairness of the justice system, and the success of that endeavor would be jeopardized by a rule that required appellate courts to “look the other way”—to ignore obvious error in the lower court—based on the speculative possibility that the defendant's trial attorney had some undisclosed reason for withholding an objection.

On the other hand, an overly narrow definition of “tactical decision”—one that requires too much affirmative proof that the trial attorney consciously withheld an objection—means that there will be cases where a defense attorney can “save up” the errors that occur in the trial process, holding them in reserve in the event that the trial ends badly for the defendant.

An overly narrow definition of “tactical decision” also potentially conflicts with the presumption of attorney competence: creating the possibility that appellate courts will reverse criminal convictions based on an appellate attorney's theory of why the error was crucially detrimental to the defense case—even though, if this matter were investigated at an evidentiary hearing (for example, in a post-conviction relief proceeding), the trial attorney might credibly explain that, under the attorney's chosen litigation strategy, the error appeared to be inconsequential.

Given that so much is at stake, given that the issue is so debatable, given the fact that this Court still lacks a third permanent member, and given the fact that we have already concluded that the jury instruction error was harmless beyond a reasonable doubt in Anderson's case, we have decided not to re-visit the question of whether Anderson's attorney might have made a tactical decision not to object to the jury instruction error.

Conclusion

For the reason that the error in the jury instructions was harmless beyond a reasonable doubt, we again AFFIRM the judgement of the superior court.