BREWER, J.

In this product liability action, plaintiff appealed a judgment for defendants following a jury trial. The Court of Appeals affirmed without considering the merits of nine of plaintiff's ten assignments of instructional and evidentiary error. Purdy v. Deere and Company, 252 Or.App. 635, 287 P.3d 1281 (2012). The Court of Appeals explained that consideration of those assignments of error would be futile because, even if it determined that the trial court had erred in one or more of the ways that plaintiff asserted, the court could not reverse the trial court's judgment because of the prohibition in ORS 19.415(2) that "[n]o judgment shall be reversed or modified except for error substantially affecting the rights of a party." Id. at 639-44, 287 P.3d 1281. Plaintiff contends that, in so holding, the Court of Appeals misconstrued the standard for reversal in ORS 19.415(2) as it applies to claims of instructional and evidentiary error like those that the court declined to consider. For the reasons that follow, we reverse and remand to the Court of Appeals to address plaintiff's assignments of error.

THE FACTUAL BACKGROUND

This action arose out of an accident involving a young child, Isabelle Norton, who was seriously injured when her father, Kirk Norton, accidentally backed his riding lawnmower into her. The lawnmower, which had been manufactured by defendant Deere and Company, had been designed so that the cutting blades shut off automatically when the lawnmower was driven in reverse. However, by design, the driver could override that safety feature by pressing a button located on the lawnmower's dashboard. Norton had engaged that override feature when, unbeknownst to him, Isabelle had approached the mower from behind. As a consequence, the cutting blades were operating when Norton backed the lawnmower into his daughter.

Plaintiff, Isabelle's conservator, brought this action on the child's behalf against Deere, along with the business that sold the lawnmower to Norton, defendant Ramsey-Waite Company. The complaint — which included claims for strict liability and negligence — alleged that the mower was defective and unreasonably dangerous or had been negligently designed or marketed, in three respects: (1) it provided a mechanism for overriding the automatic shutoff feature, thus allowing the cutting blades to operate when the lawnmower was being driven in reverse; (2) the button for overriding the automatic shutoff feature had been placed on the mower's dashboard, allowing the driver to mow in reverse without turning around to ascertain whether the path is clear; and (3) it included no warnings or instructions that addressed the safe and proper operation of the lawnmower in reverse.

In their pleadings and at trial, defendants defended on the theory that the mower was not dangerously defective and that defendants had not been negligent in any of the ways that plaintiff had alleged, and that Isabelle's injuries had been caused by her father's failure to (1) use the lawnmower in the intended manner and as instructed, (2) keep a proper lookout, and (3) ensure that his daughter was properly supervised. Defendants also asserted at trial that any defect or negligence in providing an override feature did not cause Isabelle's injuries because, even if Norton had not used that feature, the blades still would have been rotating at a high rate of speed when the mower and the child came into contact.

At the close of trial, the court submitted a verdict form to the jury that asked the following three questions:

The jury answered "No" to each of the three questions, and the trial court entered judgment for defendants.

Plaintiff appealed, raising 10 assignments of error. Four of the assignments of error asserted that the trial court had erred by excluding testimony from several witnesses about instances in which other children had been injured by Deere riding mowers being driven in reverse. Plaintiff asserted that the testimony of those witnesses was relevant either to show that Deere had notice of the defect that plaintiff had alleged or to show that the mower was defective and dangerous. A fifth assignment of error challenged the trial court's refusal to admit evidence that Deere marketed small toy riding lawnmowers. In plaintiff's view, that evidence was relevant to show that Deere had advertised its riding lawnmowers as being safe around children, and had affected consumer expectations accordingly. Four more assignments of error challenged jury instructions concerning what constitutes a product defect; those instructions either were given to the jury over plaintiff's objection or were requested by plaintiff but rejected by the trial court. A tenth and final assignment of error concerned the trial court's ruling that one of defendants' witnesses — a retired Deere engineer — was qualified to testify as an expert about how long the mower blades would have continued to rotate by sheer momentum, and how far the mower would have travelled in reverse, if the automatic shutoff feature had not been overridden. That testimony was proffered to show that the alleged defect in the mower's design was not the cause of Isabelle's injuries because, even if the automatic shut off feature had not been overridden, the mower's blades still would have been rotating at a high rate of speed when they came into contact with the child.

Confronted with those assignments of error, the Court of Appeals concluded that the last one was the only assignment of error that pertained to the element of causation, as opposed to defendants' culpability. Purdy, 252 Or.App. at 648, 287 P.3d 1281. The Court of Appeals considered and rejected that assignment of error on its merits. Id. at 645-48, 287 P.3d 1281. The court declined to consider plaintiff's remaining assignments of error. The court explained that each question in the verdict form included both the issue of culpability — either the existence of a defective and unreasonably dangerous product (in the strict liability claims) or a breach of the applicable standard of care (in the negligence claims) — and the issue of causation. Id. at 642-44, 287 P.3d 1281. The court reasoned that it therefore could not tell from the verdict form whether the jury had answered "no" to each question because it found that plaintiff had failed to prove that the mower was dangerously defective or that the relevant defendant was negligent, or because it found that plaintiff had failed to prove that the relevant defendant's conduct had caused Isabelle's injuries. Id. at 644-45, 287 P.3d 1281. Because it could not tell whether the asserted errors had affected the jury's verdict, the Court of Appeals concluded that it was required to affirm. Id. at 648, 287 P.3d 1281.

In so concluding, the Court of Appeals relied on Lyons v. Walsh & Sons Trucking Co., Ltd., 337 Or. 319, 96 P.3d 1215 (2004), where this court had undertaken to apply the standard for reversal in ORS 19.415(2). Expanding on this court's construction and application of that standard in an earlier case, Shoup v. Wal-Mart Stores, Inc., 335 Or. 164, 61 P.3d 928 (2003), this court in Lyons reiterated that a party seeking reversal of a judgment bears the burden of showing from the trial court record that the asserted error substantially affected his or her rights. Lyons, 337 Or. at 326, 96 P.3d 1215. The court further held that, where a compound question on a verdict form asks whether a defendant's conduct failed to meet the relevant standard of care and, if so, whether that failure was the cause of the plaintiff's damages, a single answer of "no" does not reveal whether the jury's answer means no failure to meet the standard of care, or failure to meet the standard of care, but no causation. Id. at 325, 96 P.3d 1215. In such a situation, the court concluded, if the assignment of error challenges the jury's determination that there was no failure to meet the standard of care, the plaintiff cannot prevail because, even if the assignment is well taken, it is irrelevant (and the error, therefore, harmless) if that failure did not cause injury. Id. at 325-26, 96 P.3d 1215. Likewise, if the assignment of error challenges causation, that argument may be well taken, but it is irrelevant (and the error, therefore, harmless) if there was no failure in the first place. Id.

The Court of Appeals concluded that Lyons controls the present case because here, as in Lyons, the jury answered "no" to questions that combined two different elements — culpability and causation — of plaintiff's claims. That is, the format of the verdict form made it impossible to discern — insofar as the jury's resolution of those two elements was concerned — the rationale for the jury's verdict. Purdy, 252 Or.App. at 639-45, 287 P.3d 1281. Having rejected the single assignment of error that, in its view, related only to the element of causation, the Court of Appeals concluded that considering the remaining assignments of error would be futile, because plaintiff could not eliminate the possibility that the jury had based its answers solely on a finding that plaintiff had failed to establish causation. Id. at 645-48, 287 P.3d 1281. Accordingly, the Court of Appeals affirmed the judgment in defendants' favor.

THE PARTIES' DISPUTE

Plaintiff principally argues that the Court of Appeals' overall approach to the reversible error analysis was flawed.1 Plaintiff asserts that, in a long line of cases, this court has concluded that instructional and evidentiary errors, like the ones asserted here, "substantially affect[ed] the rights of a party," ORS 19.415(2), and therefore were grounds for reversal. Plaintiff acknowledges that a different conclusion, focusing on the format and content of the verdict form, "may be inevitable" when multiple specifications of liability — only one of which is shown to be invalid — are combined into a single question in the verdict form. Plaintiff argues, however, that the statute does not require an appellant to have requested a special verdict form to establish that an instructional or evidentiary error substantially affected the rights of the appellant. Plaintiff contends that, to the extent that this court's decision in Lyons can be understood to hold otherwise, it should be disavowed. According to plaintiff, the reasoning in Lyons is inconsistent with this court's controlling precedent, and it was effectively rejected by this court in Wallach v. Allstate Ins. Co., 344 Or. 314, 180 P.3d 19 (2008). In short, plaintiff argues that — contrary to the Court of Appeals' understanding — whether the asserted instructional and evidentiary errors constitute grounds for reversal is not controlled by Lyons.

Defendants have a different understanding of Lyons and its relevance here. According to defendants, the decision in Lyons represented a straightforward application of a broad principle announced in Shoup — a party claiming trial court error has an affirmative burden under ORS 19.415(2) to show from the trial court record that "the jury made an adverse finding regarding the issue to which [the appellant's] claims of error are directed." Defendants argue that, in Lyons, and earlier in Jensen v. Medley, 336 Or. 222, 82 P.3d 149 (2003), this court indicated that the construct that this court applied in Shoup governs all categories of trial court error, including instructional and evidentiary error. Defendants also observe that — irrespective of reservations that this court expressed in Wallach about the holding in Lyons — this court expressly declined to reconsider or overrule Lyons, leaving intact Lyons' specific holding in the context of compound verdict forms. Defendants assert, in short, that Lyons is good law and that its reasoning precludes reversal here, because it is impossible to tell whether the jury found against plaintiff on the issues to which his assignments of error are directed.

The parties thus disagree not only about the import and precedential effect of Lyons, but also about the import and effect of other decisions purporting to apply the standard for reversal set out in ORS 19.415(2) to claims of instructional or evidentiary error. To assist the reader in understanding our analysis of Lyons and other pertinent decisions, we provide the following background.

THE LEGAL BACKGROUND

The standard for reversal of a trial court judgment expressed in ORS 19.415(2) has been part of Oregon's statutory law since territorial days. See Wallach, 344 Or at 326 n. 12 (discussing earlier versions of ORS 19.415(2)). Over the years, this court has described and applied that standard in a variety of ways, depending on the particular kind of error that was asserted. Before Lyons was decided, this court had taken a generally consistent approach to instructional error. Such error was deemed to be reversible if the instruction directs the jury to apply the wrong rule, and the instructions, considered as a whole and in light of the evidence and the parties' theories of the case, permitted the jury to reach an incorrect result. Wallach 344 Or. at 322, 180 P.3d 19 (discussing earlier decisions holding that instructions permitting jury to reach legally incorrect result substantially affected a party's rights); Hernandez v. Barbo, 327 Or. 99, 107-08, 957 P.2d 147 (1998) (giving erroneous instruction substantially affected appellant's rights because it probably created erroneous impression of the law that permitted the jury to reach an incorrect result); Waterway Terminals v. P.S. Lord, 256 Or. 361, 370, 474 P.2d 309 (1970) (same).

To determine whether evidentiary error is a ground for reversal, this court generally has looked to two related standards. The first is the standard in ORS 19.415(2); the second standard is that "[e]rror may not be predicated upon a ruling which admits or excludes evidence unless a substantial right of the party is affected." OEC 103(1); see Jett v. Ford Motor Co., 335 Or. 493, 497, 72 P.3d 71 (2003) (so holding). In making those determinations, this court ordinarily has held that an error in admitting or excluding evidence does not require reversal if there is little likelihood that the error affected the verdict. See, e.g., State v. Davis, 336 Or. 19, 32, 77 P.3d 1111 (2003) (applying that construct for reversal in criminal case).

In Shoup, this court faced a novel issue that implicated ORS 19.415(2). In that case, an elderly woman who was shopping in a store was knocked to the floor when a store employee bumped into her. The woman sued the store, alleging three separate specifications of negligence. The defendant moved to strike one of the specifications of negligence from the jury's consideration, but the motion was denied. At the end of trial, the jury returned a general verdict for the plaintiff. The defendant appealed, assigning error to the trial court's refusal to withdraw the challenged specification of negligence from the jury's consideration. Shoup, 335 Or. at 166-67, 61 P.3d 928. The Court of Appeals agreed with the defendant that the challenged specification should not have been submitted to the jury, but it noted that the jury had returned a general verdict that might have been based on one or the other of the remaining specifications of negligence, neither of which had been challenged. Id. at 167, 61 P.3d 928.

The Court of Appeals concluded that it was required to reverse under ORS 19.415(2), based on the so-called "we can't tell" rule that this court had announced in Whinston v. Kaiser Foundation Hospital, 309 Or. 350, 788 P.2d 428 (1990). In Whinston, the court had held, in an appeal involving a trial court's ruling on a motion for judgment notwithstanding the verdict (JNOV) that, "[i]f the court cannot determine whether the verdict was based on a [specification] supported by the evidence or on one unsupported by the evidence, the result is a new trial." Whinston, 309 Or. at 359, 788 P.2d 428.

For the purpose of determining whether the error at issue in Shoup required reversal, Whinston was not directly on point, because that case had not involved any explicit consideration of the standard set out in ORS 19.415(2). Still, it had obvious implications for that standard, at least in cases involving a specific pattern of circumstances — that is, where multiple specifications of negligence, one of which is invalid, are submitted to the jury in a single, general verdict question. To state the obvious: In such circumstances, a reviewing court cannot tell whether the error in submitting the invalid specification affected the verdict, because the jury might have based its verdict on one of the valid specifications. The Court of Appeals understandably applied that rule in Shoup to hold that, because it could not tell whether the jury had based its verdict on one or more of the valid specifications of negligence, or on the single invalid specification, it must remand for a new trial.

On review in Shoup, this court stated that, when it adopted the "we can't tell" rule in Whinston, it had failed to consider whether that rule was compatible with the statutory standard for reversal set out in ORS 19.415(2).2 Shoup, 335 Or. at 170, 61 P.3d 928. With that question squarely before it, the court began its analysis by rejecting the defendant's defense of the rule — that, in any case posing a "we can't tell" problem, the mere possibility that the jury based its verdict on an invalid specification of liability establishes that the appellant's rights were "substantially affected" within the meaning of ORS 19.415(2). The court explained that the "outcome might have been different" construct appeared to be more lenient than the standard set out in ORS 19.415(2). The court concluded that

335 Or. at 174, 61 P.3d 928 (emphasis added). Applying that holding to the case before it, the court 19 concluded that the record — in particular, the verdict form — did not show that the jury had based its verdict on the invalid specification of negligence. Therefore, the court reasoned, the defendant had not shown that the trial court's error in submitting that specification to the jury had substantially affected the defendant's rights. Id. at 178-79, 61 P.3d 928. In that context, the court suggested, the problem could have been avoided by using a special verdict form: "A special verdict would have allowed defendant to show the claims or specifications upon which the judgment was based and thus to provide us with a record that would allow us to determine whether the trial court error was prejudicial." Id. at 178, 61 P.3d 928.

The court in Shoup made particular mention of two cases that bear on our discussion here, and which we now describe briefly. The first was Baker v. English, 324 Or. 585, 932 P.2d 57 (1997), where this court applied the statutory standard for reversal in the context of an erroneous discovery ruling; the second was Hernandez, where this court applied the statutory standard to instructional error.

The court in Baker held that the erroneous denial of discovery in that case did not require reversal because the requesting party "already knew or had possession of qualitatively the same information as that contained in the denied discovery." 324 Or. at 593, 932 P.2d 57. As pertinent here, the court acknowledged that it often had determined whether an error required reversal by inquiring whether the error permitted the jury to reach an erroneous verdict. The court explained, however, that although that inquiry was helpful in some cases, "[the court's] focus in this and all similar cases is the statutory test set forth in ORS 19.415(2): `No judgment shall be reversed or modified except for error substantially affecting the rights of a party.'" Id. at 592-93, 932 P.2d 57. The court concluded that the discovery error in that case did not satisfy the statutory standard:

Id. at 593, 932 P.2d 57. Thus, in Baker, the denial of discovery could not have had the statutorily required effect on the right that the defendant posited, that is, to formulate and execute a trial theory and strategy. Accordingly, the trial court's erroneous ruling could not furnish a basis for reversing the judgment.

In the other important case discussed in Shoup, Hernandez, a woodworker who had been injured by a saw brought a product liability action against its manufacturer. The defendant asserted an affirmative defense of comparative fault that alleged ten separate specifications of the plaintiff's negligence that had contributed to the accident. The plaintiff requested an instruction that limited the kind of conduct that qualified for consideration by the jury as comparative negligence in the circumstances. The trial court declined to give the instruction. After the jury returned a verdict for the defendant, the plaintiff appealed, assigning error to the trial court's refusal to give the requested instruction. Hernandez, 327 Or. at 102-05, 957 P.2d 147. On review, this court concluded that the instruction was proper and that the trial court's refusal to give it could have affected the jury's verdict on one of the ten specifications of comparative negligence. Id. at 112, 957 P.2d 147. The court further concluded that the error required reversal under the construct described in Baker, 324 Or. at 590, 932 P.2d 57, and Waterway Terminals, 256 Or. at 370, 474 P.2d 309, because the jury "may have based its comparative fault assessment on a misperception of the evidence resulting directly from the court's failure to give plaintiff's requested jury instruction." Id. Thus, in Hernandez, this court held that instructional error had "substantially affect[ed]" the appellant's rights within the meaning of ORS 19.415(2), where, based on the record in that case, the trial court's refusal to give a valid instruction permitted the jury to reach an erroneous verdict.

In Shoup, this court commented on the decisions in Hernandez and Baker:

335 Or. at 172 n. 2, 61 P.3d 928. Viewed in the context of the overall discussion in Shoup, that comment is most logically understood to mean that the generally-applied construct for determining whether instructional error is a basis for reversal under ORS 19.415(2) is not a universal proxy for the statutory standard, but that its application in Hernandez was correct. That is how this court interpreted it in Wallach. 344 Or. at 323-24, 180 P.3d 19.

Shortly after it decided Shoup, this court applied ORS 19.415(2) in State v. Pine, 336 Or. 194, 82 P.3d 130 (2003). In Pine, the defendant appealed his conviction for third-degree assault, arguing that the trial court had erred by instructing the jury that, to be convicted of that offense, the defendant himself need not have caused physical injury to the victim. The state countered that, under Shoup, any error in giving the instruction was not a ground for reversal, because the verdict was supported by evidence that the defendant himself had caused physical injury to the victim. On review, this court acknowledged that Shoup might be distinguishable on the ground that it was a civil, not a criminal, case, but the court instead chose to "rely on a more fundamental reason to distinguish Shoup from th[e] case [before it]. * * * [I]f the jury had believed defendant's version of the facts, it nonetheless could have convicted him under the challenged instruction. If * * * that instruction incorrectly stated the law, then the jury's guilty verdict effectively would have convicted defendant of a crime that the legislature did not enact." 336 Or. at 200, 82 P.3d 130 (emphasis added).

That brings us, finally, to Lyons, the case that the Court of Appeals deemed to be controlling here. In Lyons, the plaintiffs' decedent, a state trooper, was killed, along with his colleague, Rector, when the police vehicle that Rector was driving and in which the decedent was a passenger was struck by a tractor trailer. The plaintiffs brought a wrongful death action against the company that owned the truck and employed its driver, alleging that the accident was caused by the negligence of the defendant's employee. Lyons, 337 Or. at 321-22, 96 P.3d 1215. The plaintiffs asserted — and the defendant did not dispute — that, if Rector had been negligent, he and his employer, the Oregon State Police, were immune from liability to the plaintiffs under the exclusive remedy provision of the workers' compensation statutes. Id. at 322, 96 P.3d 1215.

In accordance with their view of the effect of that immunity, the plaintiffs sought various instructions admonishing the jury not to "weigh or consider" Rector's conduct unless it found that it was the "sole and exclusive" cause of the accident. Id. at 323, 96 P.3d 1215. The trial court declined to give the requested instructions and, instead, instructed the jury not to compare Rector's fault with the truck driver's fault. At the close of trial, the court submitted a special verdict form to the jury that combined the issues of negligence and causation into a single, compound question: Whether the defendant was negligent in any of the ways alleged and "and, if so, was such negligence a cause of damage to the plaintiffs." Id. at 323, 96 P.3d 1215 (emphasis added). The jury answered "no" to that question, and thus the defendant prevailed.

The plaintiffs unsuccessfully appealed to the Court of Appeals and then sought review before this court, on the grounds that the trial court had erred in giving the comparative fault instruction and in refusing to give the requested "sole and exclusive cause" instruction. Although this court initially allowed review to consider those issues on their merits, it eventually concluded that it could not reach them "because of the nature of the verdict that the jury rendered." Id. at 324, 96 P.3d 1215. The court observed that the jury had rendered its verdict on a compound question that made it impossible to tell whether the answer rested (1) on a determination that the defendant was not negligent, or, alternatively, (2) on a determination that the defendant was negligent but that its negligence had not caused the plaintiff's damages. Id. at 325, 96 P.3d 1215.

The court explained:

337 Or. at 325, 96 P.3d 1215. This court referred to the holding in Shoup that a party seeking reversal of a judgment bears the burden of making a record that demonstrates prejudicial error, and then held:

Id. at 326, 96 P.3d 1215.

In at least two respects, the court's explanation of its holding was somewhat opaque. First, the court stated that the plaintiff had not "advanced a single factual theory of liability that the form of jury verdict reflected." Id. That statement was literally accurate; the plaintiff in Lyons, like the plaintiff in Shoup, had pleaded several factually distinct specifications of negligence. See Lyons, 337 Or. at 322, 96 P.3d 1215 (describing the plaintiff's multiple theories of negligence). However, the "two different rationales that the jury verdict form identified" in Lyons — culpability and causation — had no obvious connection to the circumstance that the plaintiff had pleaded multiple theories of liability. Rather, those different "rationales" were common elements (as opposed to "theories of liability") in each of the specifications of negligence. Second, the court's explanation suggested that the case did not involve "other kinds of asserted trial error, such as a faulty jury instruction, that may call for a different analysis of whether the error `substantially affect[s] the rights of a party,'" even though the plaintiff's fundamental position on appeal was that the trial court had improperly instructed the jury.

In Wallach, this court addressed some of the uncertainty that arose out of Lyons. In Wallach, an insured who had been injured in an automobile accident brought an uninsured motorist (UM) action against his insurer. The plaintiff thereafter was involved in two more accidents and, in the UM action, sought to hold the insurer liable for aggravation of the injuries sustained in the first accident that resulted from the second and third accidents. The plaintiff proffered an instruction that was consistent with his aggravation theory, and the trial court gave the instruction despite the insurer's objection that the instruction erroneously stated the law. The jury returned a verdict awarding significant money damages to the plaintiff. The insurer appealed, asserting that the trial court had erred in giving the "aggravation" instruction. The Court of Appeals agreed with the insurer that the instruction was erroneous and reversed and remanded the case to the trial court. Wallach v. Allstate Ins. Co., 206 Or.App. 137, 145, 135 P.3d 404 (2006).

The plaintiff sought review, arguing that the instruction was correct and that, even if it was not, the Court of Appeals was bound by Lyons to affirm, because the defendant had failed to show that the asserted error actually affected the verdict. The plaintiff argued that the defendant had failed to make the required showing because the challenged instruction was relevant only to the claimed aggravation damages sustained in the second and third accidents, and the jury's award may have been based solely on the injuries that resulted from the first accident.

This court in Wallach distinguished Lyons on the ground that, in Lyons, this court had described that case as one involving a "narrow problem" pertaining to the form of the verdict, not a faulty jury instruction. Wallach, 344 Or. at 328, 180 P.3d 19. The court noted that, in an unbroken line of cases preceding Lyons, this court had determined whether instructional error had "substantially affected" the rights of a party under ORS 19.415(2) by asking whether, based on the instructions read as a whole, the erroneous instruction told the jury to apply the wrong legal rules to a claim or defense and, if so, whether, in light of the trial court record, the error permitted the jury to reach a legally erroneous outcome. Wallach, 344 Or. at 326, 180 P.3d 19. The court further observed that it had applied the same construct of the statutory standard in two instructional error cases, Hernandez and Pine, where verdicts could have been based on alternative theories of liability or guilt that 15 would have been unaffected by the instructional error. Id. at 323-26, 180 P.3d 19.3 In the absence of an indication that the court in Lyons had meant to overrule Hernandez and Pine, this court in Wallach applied the construct of the statutory standard that the court had applied in those two cases. Id. at 328, 330, 180 P.3d 19.

The court's decision in Wallach was not unanimous. Noting that the majority had attempted (1) to distinguish Lyons as a case involving an error in the form of the verdict and (2) to limit the holding in Lyons accordingly, the dissent protested that the distinction did not withstand scrutiny. Wallach, 344 Or. at 341-42, 180 P.3d 19 (Durham, J., dissenting). The dissent opined that the underlying problem in Lyons was an erroneous jury instruction and that, because the court had applied the "we can't tell" standard for reversal against the appellant in that case, it was bound to apply the same standard to the instructional error in Wallach Id.

Wallach, 344 Or. at 329, 180 P.3d 19.

THE PROBLEM OF LYONS

To sum up thus far, in Wallach, this court cabined the holding in Lyons to circumstances involving a particular kind of verdict form. Without reconsidering the holding in Lyons, this court in Wallach concluded that, because the case before it did not involve a similar verdict form, the holding in Lyons was not controlling. This court's holding in Wallach therefore permitted the bench and bar to infer that Lyons does govern the analysis in instructional error cases that do involve such a verdict form. That is precisely the conclusion that the Court of Appeals reached in this case when it decided that, in the absence of a special verdict form that eliminated the possibility that the jury's verdict was based solely on the issue of causation, it could not consider plaintiffs' claims of instructional and evidentiary error, all of which were concerned with elements of culpability. Purdy, 252 Or.App. at (542 n. 2, 287 P.3d 1281. As a consequence, the ongoing vitality of Lyons looms prominently in our review.

As discussed, plaintiff contends that Lyons does not control here. Although plaintiff suggests other reasons why that may be so,4 his primary thesis is that Lyons was wrongly decided. Plaintiff asserts that, before Lyons, this court consistently had held that instructional error required reversal of a trial court judgment if the error permitted the jury to reach a legally erroneous result.5 Plaintiff further observes that Shoup did not involve instructional or evidentiary error and that this court in Shoup took pains to distinguish instructional error from the kind of error presented in that case. Shoup, 335 Or. at 172 n. 2, 61 P.3d 928. Finally, plaintiff argues that, although the court in Wallach expressly declined to overrule Lyons, it did, in effect, limit the holding in Lyons to its own procedural posture. Plaintiff nevertheless asserts that Lyons was wrongly decided, even in the supposedly narrow circumstances in which it arose.

This court does not overrule its prior decisions lightly. As we stated in Farmers Ins. Co. v. Mowry, 350 Or. 686, 698, 261 P.3d 1 (2011), "[s]tability and predictability are important values in the law." Because of the importance of those values, we will not overrule prior decisions "simply because the personal policy preferences of the members of the court may differ from those of our predecessors who decided the earlier case." Id. (internal quotation marks omitted). We set out the governing considerations in Mowry:

Id. at 698, 261 P.3d 1.

Lyons concerned the meaning and application of a statute, ORS 19.415(2), and our determination whether to reconsider that case here places particular emphasis on the court's use of the appropriate interpretive framework, which, broadly conceived, would include adequate consideration of the relevant context, including case law interpreting or applying the statute. Mowry, 350 Or. at 698, 261 P.3d 1. Thus, the criticism that plaintiff has leveled against this court's decision in Lyons — that, on the questionable premise that the case did not involve a faulty jury instruction, the court in Lyons failed to consider a body of pertinent case law — could be an adequate basis for reconsidering that case.

Notably, this court accepted review in Lyons to address the merits of the plaintiff's assignments of instructional error but, without consulting the parties in that case, decided that it could not reach those assignments of error because of the inherent ambiguity of the verdict that the jury had rendered. The court's analysis of the issue was sparse; it consisted of (1) a brief quotation setting out the primary holding of Shoup, unaccompanied by any explanation of the procedural circumstances in which that holding arose; (2) a conclusion that Shoup applied equally to "the narrow problem that the form of jury verdict used in the present case poses" — presumably a reference to the facts that the verdict form asked the jury to answer a two-part question and that the court could not tell which part of the question had driven the jury's answer; and (3) a statement that the case did not involve "other kinds of asserted trial error, such as faulty jury instructions, that may call for a different analysis." Lyons, 337 Or. at 326, 96 P.3d 1215. As discussed, because the particular error that the appellant had asserted was that the jury instructions were faulty, the court's statement was questionable. Also as discussed, the court did not directly consider the line of cases that set out and applied a different construct of the standard for reversal in the context of asserted instructional error. Those concerns provide sufficient grounds for reconsidering the analysis applied in that case. See Mowry, 350 Or. at 693, 261 P.3d 1 (stating pertinent considerations for decision whether to reexamine precedent). We now turn to that task.

We begin by reiterating that the task before the court in Lyons was to determine the meaning of, and apply, a statute. The standard currently codified in ORS 19.415(2) has been a part of this state's civil and criminal law since the Deady Code. See General Laws of Oregon, Civ. Code, ch. VI, § 533, p. 284 (Deady 1845-1864) (providing that, in civil appeals, the judgment "shall only be reversed or modified for errors substantially affecting the rights of the appellant"); General Laws of Oregon, Crim. Code, ch. XXIII, § 246, pp. 482-83 (Deady 1845-1864) (providing that, in criminal appeals, the court "must give judgment, without regard to * * * technical errors, defects or exceptions which do not affect the substantial rights of the parties").

As noted, this court construed ORS 19.415(2) in Shoup, and it reached two conclusions that are pertinent here. First, to require reversal under the statute, an error must — in an important or essential manner — have materially or detrimentally influenced a party's rights; it is insufficient to speculate that the error might have changed the outcome in the case. In that regard, the court said:

Shoup, 335 Or. at 173, 61 P.3d 928. Second, the statute imposes a burden on the party asserting an error to demonstrate that the error had the required prejudicial effect. In that regard, the court said:

Id. at 173-74, 61 P.3d 928.

As noted, defendants argue that, to satisfy those requirements, an appellant must show that the jury actually made an adverse finding on the issue to which the appellant's claims of error are directed. Plaintiff responds that the effects of the asserted instructional and evidentiary errors on his rights in this case would not necessarily be discernible from the content of a special verdict and, therefore, the generally applicable constructs of the standard set out in ORS 19.415(2) for those types of errors ought to govern here.

A careful reading of the words of the statute in light of this court's decision in Shoup supports plaintiff's view. Just as the legislature could have conditioned reversal on the mere possibility that an error affected the jury's decision if it had so intended, it also could have conditioned reversal on a showing that the jury found against the appellant on the particular element of a claim or defense to which the asserted error pertained, if it had meant to impose that requirement. This court rejected the former approach in Shoup, and we reject the latter here, because, it too, fails to correspond to the statutory standard.

That standard asks, instead, whether — in an important or essential manner — the error had a detrimental influence on a party's rights. Shoup, 335 Or. at 172-73, 61 P.3d 928. It does not pretend to measure mathematical probabilities; rather, it assesses the extent to which an error skewed the odds against a legally correct result. This court's previous decisions that have applied the standard to instances of instructional and evidentiary error generally indicate that little likelihood is not enough, but more — that is, "some" or a "significant" likelihood that the error influenced the result — will suffice for reversal. See State v. Lopez-Minjarez, 350 Or. 576, 587, 260 P.3d 439 (2011) (because "erroneous instruction had no significant likelihood of affecting the jury's verdict" on a charge, it did not substantially affect defendant's rights); see also Davis, 336 Or. at 29 n. 7, 77 P.3d 1111 ("under [ORS 19.415(2)] the analysis whether an appellate court must affirm a judgment despite trial error is similar to the analysis that Article VII (Amended), section 3, requires.").6

In Hernandez, this court referred to "the general rule [that] the parties in a civil action are entitled to jury instructions on their theory of the case if their requested instructions correctly state the law, are based on the current pleadings in the case, and are supported by evidence." Hernandez, 327 Or. at 106, 957 P.2d 147. The court in Hernandez concluded that, because the trial court refused to give the plaintiff's requested instruction, the jury applied an incomplete and inaccurate legal rule to the facts, which, in the context of the record as a whole, permitted the jury to reach a legally incorrect outcome; accordingly, the error substantially affected the appellant's rights. Id. at 112, 957 P.2d 147; see also Wallach 344 Or. at 329, 180 P.3d 19 (erroneous instruction that permitted jury to find defendant liable for damages for which it was not liable, as well as damages for which it was liable, substantially 15 affected defendant's rights); Pine, 336 Or. at 209-10, 82 P.3d 130 (instruction that permitted jury to convict defendant based on facts contrary to statute defining crime substantially affected defendant's rights).

That approach to analyzing the effect of error under ORS 19.415(2) is consistent with a presumption to which this court has adhered for many years; that is, "[w]e presume that a jury follows a trial court's instructions." Wallach 344 Or. at 326, 180 P.3d 19. It does not mean, however, that all or most instructional error substantially affects a party's rights for the purposes of ORS 19.415(2). "A jury instruction does not constitute reversible error unless it prejudiced the defendant when the instructions are considered as a whole." State v. Bowen, 340 Or. 487, 516, 135 P.3d 272 (2006), cert. den., 549 U.S. 1214, 127 S.Ct. 1258, 167 L.Ed.2d 89 (2007). Moreover, in making that determination, the court not only considers the instructions as a whole, it considers the instructions in the context of the evidence at trial and the parties' theories of the case with respect to the various charges, claims, and defenses at issue. Lopez-Minjarez, 350 Or. at 578, 260 P.3d 439 (2011) (to assess whether error in instructing jury substantially affected appellant's rights on various charges involved, "it is important to describe both sides' respective evidence and theories of the case"); see also State v. Phillips, 354 Or. 598, 613, 317 P.3d 236 (2013) (erroneous failure to give jury concurrence instruction did not substantially affect defendant's rights because, "on the facts in this case, the factual findings necessary to find defendant liable on one theory either subsumed or were the same as the factual findings on the other theory"); U.S. National Bank v. Boge, 311 Or. 550, 566, 814 P.2d 1082 (1991) (giving erroneous instruction regarding standard of conduct substantially affected plaintiff's rights because "[t]he standard for [plaintiff's] conduct was the central issue to be decided," and error permitted jury to reach incorrect result).7

That kind of record-based review is consistent with this court's statement in Shoup that ORS 19.415(2) "places the burden to make a record that demonstrates prejudicial error on whichever party * * * seeks reversal," and that that rule applies "in every case." 335 Or. at 173-74, 61 P.3d 928. As noted, in Shoup, this court concluded that the defendant was "unable to show that the trial court's error [in allowing the invalid specification to go to the jury] substantially affected [its] rights." Id. at 179, 61 P.3d 928. In so holding, the court relied on two factors in the record. First, the court noted that identical evidence applied to all three specifications of negligence in that case, two valid and one invalid:

Id. at 178, 61 P.3d 928. In sum, the record showed that there was little likelihood that the jury had found for the plaintiff on the invalid specification without also finding for the plaintiff on one or both of the valid specifications.

Id. at 178-79, 61 P.3d 928. Thus, it was significant that the defendant had affirmatively eschewed — by objecting to a theory-specific verdict form — the opportunity to create a record that would have established whether the jury had found for the plaintiff on only the invalid theory of liability. In that circumstance, the court held, the defendant had "not identified anything in the record to demonstrate that the jury based its verdict on [the invalid specification of negligence]" and that the defendant, therefore "[was] unable to show that the trial court's error substantially affected [the] defendant's rights." Id. at 178-79, 61 P.3d 928.

That holding constituted a particular application of the standard in ORS 19.415(2) based on the record before the court. The error did not substantially affect the defendant's rights, because the same evidence applied to all three theories of liability, and there was little likelihood that the jury had based its verdict on the invalid theory alone. And, to punctuate the analysis, the record showed that the defendant had actively prevented the use of a verdict form that would have shown whether the jury had based its verdict on the invalid theory of liability. Accordingly, the defendant was in no position to complain about the error.

Here, the verdict form that the trial court submitted to the jury was not as specific as the type of verdict form to which this court referred to in Shoup. That is, the form in this case required the jury to enter separate verdicts on plaintiff's claims for strict liability and negligence, but it combined the various specifications of liability for each claim into a single verdict on that claim. However, for two reasons, a theory-specific verdict form like the one that the court referred to in Shoup would not necessarily have revealed whether the posited instructional and evidentiary errors in this case adversely influenced the result that the jury reached. First, although plaintiff asserted several theories of liability on each claim, culpability and causation were common elements in each of those theories. In that circumstance, a separate verdict on each theory of liability would not have shown the jury's findings as to those elements.

Second, for good reason, this court in Shoup was careful to distinguish the problem in that case from circumstances involving instructional error. Where multiple theories of liability are advanced and a challenged jury instruction or evidence applies to one or more (but not all) theories, a verdict showing the jury's decision on a theory-specific basis might indicate (if, for example, the error pertained only to a theory of liability on which the jury found for the appellant) that the error did not adversely influence the jury's decision. However, a theory-specific verdict would not necessarily demonstrate whether the converse is true. In the case of instructional error, to show more about the likelihood that the error adversely influenced the jury's decision, a verdict form would have to create a decision tree setting out the jury's deliberative path in considerably greater detail. Because jury instructions often apply to single elements of a claim or defense, the verdict form would need to include separate findings with respect to each triable issue of fact to which a disputed instruction might apply. And even that level of detail would not necessarily reveal that (or the extent to which) any particular error substantially affected the appellant's rights, because, as discussed, that determination — for both instructional and evidentiary error — generally turns on broader, record-based, considerations that are not reflected in the content of a verdict form. See, e.g., Lopez-Minjarez, 350 Or. at 584-91, 260 P.3d 439 (determining effect of instructional error on appellant's rights by examining evidence and parties' theories as to each charge); Jett, 335 Or. at 500-01, 72 P.3d 71 (evidentiary error was harmless under ORS 19.415(2) and OEC 103(1) because "ample evidence in the record made the same point that" the appellant sought to make by introducing the excluded evidence).

Thus, it comes as no surprise that, apart from the import of its holding in Lyons, this court has never suggested that an appellant must have proposed an element-by-element or factual issue-by-issue questionnaire format for a jury verdict to demonstrate the existence of reversible instructional or evidentiary error under ORS 19.415(2).8 In particular, this court in Shoup did not suggest that, in every circumstance, the effect of an error on an appellant's rights must (or even could) be shown through the use of a special verdict. Nor did this court in Shoup purport to disavow previous decisions in which it had applied ORS 19.415(2) to assignments of instructional or evidentiary error.

If the court had taken those decisions into account in Lyons, the answer that the court gave in that case would — and should — have been different. Lyons involved an instructional error problem, and the proper application of the statutory standard for reversal in that case was the familiar one that this court had earlier used in Pine, and later used in Wallach and Lopez-Minjarez. Generally speaking, if a trial court incorrectly instructs the jury on an element of a claim or defense, and — when the instructions are considered as a whole in light of the evidence and the parties' theories of the case at trial — there is some likelihood that the jury reached a legally erroneous result, a party has established that the instructional error substantially affected its rights within the meaning of ORS 19.415(2). Lopez-Minjarez, 350 Or. at 584-91, 260 P.3d 439; Wallach, 344 Or. at 329, 180 P.3d 19; Pine, 336 Or. at 200, 210, 82 P.3d 130.9 That is so, logic dictates, even if it cannot be definitively shown that the jury did base its verdict on the erroneous instruction or if a more elaborate, element-by-element verdict form might have shown that the jury did not base its verdict on the erroneous instruction. Thus, the fact that the verdict form in Lyons did not pose separate questions for the jury to answer on particular factual elements of plaintiff's multiple theories of liability did not require a different construct of the statutory standard in that case. We therefore disavow and overrule our contrary holding in Lyons.10

APPLICATION

With that understanding, we return to this case. As discussed, the Court of Appeals concluded that it could not consider nine of plaintiff's ten assignments of error because, even if established, none of those errors would constitute a ground for reversal. That was so, in the court's estimation, because, in the absence of a special verdict that separated the issues of culpability and causation, it could not tell whether, after it had rejected the sole claimed error regarding causation, any of the remaining claimed instructional or evidentiary errors (all of which pertained only to the issue of culpability) 15 had affected the jury's decision. The court concluded that it was bound by this court's holding in Lyons to apply that construct to all categories of trial court error, including the instructional and evidentiary errors at issue here. 252 Or.App. at 642, 287 P.3d 1281.

Because we have overruled our holding in Lyons, we reverse and remand to the Court of Appeals to consider plaintiff's remaining assignments of error. If that court concludes that one or more of plaintiff's assignments of error are meritorious, it can then determine whether the error substantially affected plaintiff's rights.

The decision of the Court of Appeals is reversed and the case is remanded to that court for further proceedings.

BALMER, C.J., concurred and filed and opinion.

BALMER, C.J., concurring.

This case raises the ubiquitous and difficult issue of what an appellate court should do in the face of trial court error: When should error lead to reversal and when, notwithstanding error of some kind, should the judgment below nevertheless be affirmed? The question has vexed appellate court review of criminal and civil cases for centuries and touches on such fundamental concerns as fairness, the protection of constitutional rights, the role of juries, the legitimacy of trial and appellate court processes, and the prudent use of judicial resources. See generally Roger J. Traynor, The Riddle of Harmless Error (1970). Courts have struggled to articulate the appropriate test and, once articulated, to undertake the more nuanced task of applying it to specific cases.

The majority opinion is a useful addition to the ongoing effort of the Oregon appellate courts to articulate and apply the proper test for reversing a trial court judgment when the appellant demonstrates error in the trial court proceedings. I agree with the majority's analysis and disposition of this case and with its discussion of Shoup v. Wal-Mart Stores, Inc., 335 Or. 164, 61 P.3d 928 (2003) and ORS 19.415(2). I write separately to emphasize several aspects of the task at hand and to suggest at least some ways that a party seeking to reverse a trial court judgment can provide assistance to the appellate court.

Few legal proceedings are flawless, of course, and to reverse every trial court judgment because there was error of some kind in the proceeding would undermine critical goals of justice, finality, and efficiency. In fact, courts at various times have done just that — reversing convictions because, for example, the indictment charged a defendant with entering a building with the intent to commit "larcey," and the statute criminalized only entry with intent to commit "larceny," or when the indictment stated that the offense was "against the peace of the State," instead of "against the peace and dignity of the State." Traynor, The Riddle of Harmless Error at 3-4, 85 n. 3. Yet to require an appellant to demonstrate that the trial court judgment was "clearly wrong," as some courts have, see id. at 17-18, 89 n. 45, would pose an often insurmountable hurdle for the party seeking reversal and would countenance serious legal error at the trial level.

The Oregon statute providing for appellate review of trial court judgments sets a standard for reversal between those two extremes. It requires a party seeking reversal to demonstrate that trial court error "substantially affect[ed]" the party's rights, but does not require a showing that the error necessarily led to an incorrect judgment. ORS 19.415(2) puts it this way: "No judgment shall be reversed * * * except for error substantially affecting the rights of a party." In Shoup, we analyzed and applied that provision, noting that "reversal of a judgment is the exception, not the rule" and that, while the standard for reversal is neutral as between plaintiffs and defendants, "it places the burden to make a record that demonstrates prejudicial error on whichever party loses in the trial court and then seeks reversal or modification of the judgment on appeal." 335 Or. at 173-74, 61 P.3d 928.11 Moreover, when a party seeks reversal of a jury verdict, the Oregon Constitution imposes a further barrier by barring "re-examin[ation]" of any fact tried to a jury, "unless the court can affirmatively say there is no evidence to support the verdict." Or. Const., Art. VII (Amended), § 3. The obvious and intended (and salutary) effect of those legal principles is to place a thumb on the scale in favor of the trial court judgment.

The more difficult questions are how to articulate the legal standard — other than simply repeating the words of ORS 19.415(2) — and how to apply it to specific cases. The majority correctly points out that it is not enough for the appellant to argue that the error "possibly" affected the outcome of the case. 355 Or. at 226, 324 P.3d at 467. Except for the most trivial of errors, it almost always can be argued that an error at trial — in evidence admitted or excluded, in an instruction given or not given, in a comment by a judge, attorney, or witness that should not have been made — "possibly" affected the outcome. Nor, as the majority states, must the appellant show that the jury actually found against the appellant on the claim or defense to which the error pertained. Id. Lacking access to the jury room or the individual and group decision making processes that occur there, parties and courts rarely will be able to determine whether the error actually had an effect on the verdict. To set the bar that high would be inconsistent with ORS 19.415(2), which does not require a party to demonstrate that it would have prevailed, absent the error, but only that the error "substantially affect[ed]" its rights.

So, the bar for the appellant is somewhere above "possibly affected" the result, but below "necessarily affected" the result. Elaborating on Shoup, the majority articulates the statutory standard to be "whether — in an important or essential manner — the error had a detrimental influence on a party's rights"; the appellate court must "assess[] the extent to which an error skewed the odds against a legally correct result." 355 Or. at 226, 324 P.3d at 467. That is not like a preponderance of the evidence test to determine whether a plaintiff in a civil case has met its burden of showing that a particular assertion or fact was "more likely true than not." "Some" likelihood — more than "a little" — that the error influenced the result is required, id. but "how much" more will depend on factual and legal issues in the case as determined from the trial court record. The majority, correctly in my view, eschews a more precise quantification of the probability that the error affected the result. Of course, the critical task is applying the statutory standard to the record in a particular case.

Fortunately, there are other aspects of appellate review that this court routinely considers that make the problem less daunting than it might appear in the abstract. First, we have long recognized that some trial errors are substantial and more likely to have affected the result, while others are less likely to have done so. Evidentiary error, for example, is "not presumed to be prejudicial," OEC 103(1), and we have relied on that presumption in affirming judgments notwithstanding the erroneous exclusion or admission of evidence. See State v. Gibson, 338 Or. 560, 575-77, 113 P.3d 423, cert. den., 546 U.S. 1044, 126 S.Ct. 760, 163 L.Ed.2d 591 (2005) (citing OEC 103(1) and concluding that improperly admitted evidence was not prejudicial). Similarly, when considering a claim of instructional error, we do not look at the challenged instruction in isolation, but rather examine the instructions as a whole to determine whether they accurately state the law. State v. Oatney, 335 Or. 276, 290, 66 P.3d 475 (2003), cert. den., 540 U.S. 1151, 124 S.Ct. 1148, 157 L.Ed.2d 1045 (2004). If the instructions as a whole provide a complete and accurate statement of the law sufficient for the jury to properly decide the issues before it, an appellate court is unlikely to find that a claimed instructional error was prejudicial and to reverse a judgment on that ground. See State v. Bowen, 340 Or. 487, 516-17, 135 P.3d 272 (2006), cert. den., 549 U.S. 1214, 127 S.Ct. 1258, 167 L.Ed.2d 89 (2007) (concluding that, because instructions considered as a whole included complete and correct statements of the law necessary to decide the charges, fact that instructions perhaps should have been given in different sequence and that lesser-included offense instruction should have been given in connection with aggravated murder charge, rather than intentional murder charge, was not prejudicial).

Moreover, we do not look at trial court errors in the abstract — rather, we examine those errors in the context of the trial record as a whole, including the parties' claims and defenses, the evidence admitted and excluded, the parties' theories of the case, and the instructions. This court was able to reach the result it did in Shoup only after considering those factors. Shoup, 335 Or. at 178-79, 61 P.3d 928. As the majority notes, in both civil and criminal cases, this court has generally engaged in that kind of contextual, record-based review to assess whether error is prejudicial. 355 Or. at 227-28, 230-31, 324 P.3d at 467-68, 469-70 (discussing cases).

The approaches just discussed regarding evidentiary and instructional error, and the appellate courts' commitment to reviewing the trial court record, provide some structure to the ORS 19.415(2) inquiry. That difficult inquiry, however, inevitably requires the exercise of some amount of discretion by the appellate court. See Traynor, The Riddle of Harmless Error at 15-17 (discussing discretion in harmless error review).

But litigants need not leave themselves at the mercy of an appellate court trying to determine from the record whether trial court error "affected" the judgment "a little," "some," or "a lot." Special verdict forms, such as the form proposed by the plaintiff (but objected to by the defendant) in Shoup that would have asked the jury to state separately whether it found the defendant negligent based on its own negligence or that of its employee, can greatly assist an appellate court in determining whether error was prejudicial. Verdict forms separating liability, causation, and damages can clearly demonstrate that an error did — or did not — affect the verdict. See ORCP 61 B (authorizing use of special verdict). Interrogatories can separate jury determinations as to the liability of one party among multiple defendants or one claim among multiple claims and can even address issues related to specific evidence, the admissibility of which might become a critical issue on appeal. See ORCP 61 C (authorizing use of interrogatories).

I fully agree with the majority that an appellant need not "prove," by means of a verdict form or an interrogatory, that the jury based its verdict on improperly admitted evidence or an instruction that turned out to be erroneous. And the majority is correct that attempting to come up with a verdict form or interrogatories that would allow an appellate court to determine whether any particular jury instruction or evidentiary ruling — later asserted to be erroneous — necessarily affected the judgment could, in a complicated case, quickly lead to a complex "decision tree" for the jury's deliberative path that might create more problems for jurors, parties, and the court than it would solve. See 355 Or. at 230, 324 P.3d at 469 (discussing limitations of verdict forms).

However, lawyers aware of the difficulty that appellate courts face in determining whether error is prejudicial or harmless — and of the amount of discretion appellate courts inevitably exercise in that area — should consider whether they and their clients would be better off on appeal if they used verdict forms or interrogatories that provided greater insight into the basis for the jury's verdict. Here, the verdict form consisted of separate questions about products liability and negligence as to Deere and Company and negligence as to Ramsey-Waite Company. However, each question actually was a compound question asking both about a theory of liability (products liability or negligence) and about the very different issue of causation. Because causation was vigorously disputed at trial, it would have made sense for one or all parties to have requested a verdict form that separately addressed that issue.

Similarly, in Lyons v. Walsh & Sons Trucking Co., Ltd., 337 Or. 319, 96 P.3d 1215 (2004), the court was confronted with a verdict form that asked the compound question of whether the defendant was negligent and, if so, whether its negligence was a cause of the plaintiff's damages. The jury answered "no." 337 Or. at 323, 96 P.3d 1215. Causation was a key issue at trial and the subject of assignments of error related to instructions and a requested interrogatory. Yet the compound question posed to the jury prevented the court from determining whether the jury had concluded that the defendant was not negligent or that the defendant, although negligent, did not cause the plaintiff's damages. Again, simply separating the question of liability from the question of causation on the verdict form would have provided substantial assistance to the appellate courts in seeking to determine whether any trial court errors had "substantially affect[ed]" the judgment. In appropriate cases, lawyers can assist their clients, and the fair and efficient administration of justice, by crafting verdict forms and interrogatories that will help appellate courts determine when trial court errors should result in reversal. See Traynor, The Riddle of Harmless Error at 23 (advocating use of special verdicts and interrogatories).