Order

It 1s ORDERED:

The Petition for Hearing, filed on December 26, 2014 and granted on March 18, 2015, is DISMISSED as improvidently granted.

Mark D. Anderson was convicted of ten counts of sexual abuse of a minor based on evidence involving three different victims. At trial, each of the victims testified to multiple acts of abuse over a number of months and in various locations, The trial court failed to issue a factual unanimity instruction to the jury, and the defense did not request this instruction at trial.1 The jury convicted

*264Anderson on ten counts and acquitted him on one count, Count I.

Anderson challenged his convictions on Counts II, III, and VI-IX based on the trial court's faflure to issue a factual unanimity instruction.2 In Anderson I, the court of appeals agreed that the trial court's failure to issue a jury instruction regarding factual. unanimity was error, but since Anderson had not objected at trial, the court of appeals reviewed the mistake for plain error.3 The court of appeals determmed that the trial court's error was' harmless beyond a reasonable doubt.4 The court of appeals affirmed Anderson's convictions for the counts in question.5

Anderson petitioned thls court for review, and we granted Anderson's first petition,6 We remanded his case to the court of appeals

to allow the parties to provide supplemental briefing on and decide (a) whether Anderson's attorney's failure to object to the factual-unanimity instruction © was based on a tactical decision, and (b) wheth-

er the trial court's failure to provide the instruction was harmless beyond a reasonable doubt under either the Covington II 7 ) approach or the Neder[8] approach.[9]

On remand, the court 'of appeals affirmed Anderson's convictions and the analysis it used in Anderson I.10 Anderson then filed a second petition challenging the approach the court of appeals used to determine whether the factual-unanimity error-a - constitutional error-was harmless beyond a reasonable doubt under the prejudi¢e prong of the plain error analysis set forth in Adams v. State.11 We granted that petition -and requested additional briefing from both parties.12

The United States Supreme Court has used both a "guilt-based" approach and an "effect-on-the-jury" approach13 to determine whether a federal constitutional error is harmless beyond a reasonable doubt.14 The "gullt-based" approach is an outcome-oriented approach under which a court reviews the evidence not affected by the error to determine whether, in a hypothetical trial without the error, the jury would nonetheless have convicted the defendant.15 In contrast the "effect-on-the-jury" approach is a process-oriented standard under which the court asks if the error was a substantial factor in the jury's verdict in the actual trial that took

*265place.16 In the words of another commentator, "The effect.on-the-jury approach asks the historical question whether the error was a substantial factor in the jury's verdict, whereas the guilt-based approach asks the counterfactual question whether the defendant would have been convicted in a hypothetical trial absent the error."17 Under a guilt-based approach the error itself should not play a role in the court's analysis, and a reviewing court should simply weigh the evidence that was not impacted by the error.18 Under an effect-on-the-jury approach, however, a reviewing court must "evaluate the relationship of the error to the untainted evidence.19

© Both Anderson and the State agree that Alaska should--and does-use an effect-on-the-jury approach. In addition to the statements both parties made in their briefs, Anderson's counsel explained at oral argument that the parties "agree that this court has adopted the effect-on-the-jury approach to evaluating whether an error is harmless. We agree on the basic principles underlying that approach in evaluating the harm from factual concurrence errors." 'And the State conceded that both sides "seem to agree on the starting point, The State is not arguing ... [for the application of] a guilt-based approach. We are not suggesting that this court simply look at all the evidence and decide whether the jury reached the right result.... Instead, Alaska follows the effect-on—the~verdlct approac ."

We agree that the proper standard for deciding whether a constitutional error is harmless beyond a reasonable doubt under the preJudlce prong of the plain error test set forth in Adams20 is the effect-on-the -jury approach. We are not persuaded that the court of appeals erred in its decision on remand. Therefore, we DISMISS Anderson's Petition for Hearing as improvidently granted. ,

Entered at the direction of the court.