dissenting.
I disagree with the majority’s conclusions that the trial court acted within its discretion by excluding some of defendant’s proffered opinion evidence about the character for truthfulness of certain state witnesses and that the trial court’s error in excluding other portions of that evidence was harmless. The majority’s conclusion that the trial court had the discretion to exclude testimony by several of defendant’s opinion witnesses is based on its failure to distinguish between the foundational requirements for reputation and opinion evidence. Its conclusion that the exclusion of testimony by other opinion witnesses was harmless is based on its flawed assumption that, as long as impeachment testimony from at least one witness is permitted, similar testimony from other witnesses can properly be viewed as cumulative. I disagree with both conclusions and, therefore, respectfully dissent.
Defendant assigns error to the trial court’s exclusion of testimony by some of his witnesses for lack of a proper foundation. The majority concludes that the trial court’s exclusion of several of the witnesses was proper. In reaching that conclusion, it ignores the well-settled distinction between the foundational requirements for reputation and opinion evidence. Because I would join the vast majority of other courts in concluding that a trial court has considerably less discretion to exclude opinion witnesses for lack of foundation than it does to exclude reputation witnesses, I would hold that the trial court erred as a matter of law in this case in excluding several of defendant’s opinion witnesses. Moreover, I would hold that the trial court committed at least one *161reversible error with respect to each count of which defendant was convicted. Accordingly, I would reverse defendant’s sodomy convictions.
OEC 608 governs the admissibility of reputation and opinion evidence. In most respects, the language of OEC 608 is substantially the same as that of the federal rule, FRE 608.1 Because Oregon courts have not yet addressed the precise foundation required for opinion evidence under OEC 608.1 would look to interpretations of FRE 608 and similar state rules in construing Oregon’s rule. Unlike the majority, I would rely on that authority to conclude that the foundational requirements for opinion evidence differ from the requirements for the admission of reputation evidence.
OEC 608 provides that “[t]he credibility of a witness may be attacked or supported by evidence in the form of opinion or reputation, but * * * [t]he evidence may refer only to character for truthfulness or untruthfulness[.]” OEC 608(l)(a). The trial court correctly concluded that foundations must be established for both reputation and opinion evidence about a witness’s character for truthfulness. See, e.g., State v. Caffee, 116 Or App 23, 27, 840 P2d 720 (1992), rev den 315 Or 312 (1993); Charles Alan Wright and Victor James Gold, 28 Federal Practice and Procedure § 6114, at 53-54, 60-62 (1993). However, both the trial court and the majority failed to address whether the foundational requirements for reputation and opinion evidence differ. Because opinion evidence tends to be more reliable than reputation evidence and because its defects are more easily exposed on cross-examination,2 I would follow the time-honored distinction between the foundational requirements recognized in other jurisdictions for the two types of character evidence. *162Accordingly, I would hold, as a matter of law, that the foundational requirements were satisfied for testimony by several of the excluded opinion witnesses.
As several federal and state courts have acknowledged,
“ ‘no prerequisite conditioned upon long acquaintance or recent information about the witness [is required for the admission of opinion evidence]; cross-examination can be expected to expose defects of lack of familiarity and to reveal reliance on isolated or irrelevant instances of misconduct or the existence of feelings of personal hostility toward the principal witness.’ ”
*163United States v. Watson, 669 F2d 1374, 1382 (11th Cir 1982) (quoting United States v. Lollar, 606 F2d 587, 589 (5th Cir 1979) (citations omitted)). See also John B. Weinstein, 4 Weinstein’s Federal Evidence § 608.13(2), at 608-25 (2d ed 1997) (stating that the only foundation required for lay opinions as to character is that of personal knowledge); State v. Carsner, 126 Idaho 911, 894 P2d 144, 150-51 (Ct App), rev den (Idaho 1995).
As explained in Federal Practice and Procedure, opinion testimony admitted under FRE 608 must be both rationally based and helpful to the trier of fact. Wright and Gold, 28 Federal Practice and Procedure § 6114, at 53-54 (describing the requirements of FRE 701). The requirement that the opinion be rationally based includes two components: that the “opinion * * * be based on data that is logically connected to the truthfulness or untruthfulness of the other witness * * * [and that] the opinion of the character witness * * * be based on sufficient perception of the other witness * * Id. § 6114, at 53. Although an opinion witness’s perception of the substantive witness need not be contemporaneous with the trial, see, e.g., Weinstein, 4 Weinstein’s Federal Evidence § 608.13[2], at 608-25 (noting that recent information about the substantive witness is not a formal prerequisite to admission of opinion testimony), if the perception on which the opinion is based is remote from the crime and the trial, the opinion testimony may be excluded. See, e.g., Caffee, 116 Or App at 27 (in a case in which the only recent contact consisted of letter writing, the court excluded the opinion evidence because of the lack of a current basis for the testimony). As recently explained by a federal district court, “[t]he most relevant time period [in determining the admissibility of opinion evidence] is from the date of the crime to the date of the witness testifying.” United States v. Crumby, 895 F Supp 1354, 1364 (D Ariz 1995).3
*164Here, defendant sought to introduce the testimony of several witnesses about the character for truthfulness of Jennifer and Kelly, the mothers of the two alleged victims. He also sought to introduce opinion evidence that one of the child victims, K, was untruthful. Most of the character witnesses had had contact with the state witnesses that was roughly contemporaneous with the alleged crimes. The trial court committed legal error in applying the standard for admission of reputation evidence to defendant’s proffered opinion evidence.4
One example of the trial court’s use of the wrong foundational standard is its exclusion of the opinion testimony of Jennifer’s former boss, Southerton, on the ground that her knowledge of Jennifer was not “community-based” or “characteristic.”5 The majority does not address the fact that the trial court clearly used the wrong standard in excluding the former boss’s testimony. It merely holds that the trial court properly held that an opinion about Jennifer’s character for truthfulness could not be based on a single instance of untruthfulness. In so holding, it ignores both the *165efforts of other jurisdictions that have struggled with the issue and reached opposite conclusions, as well as the practical reality of how opinions are formed.
Southerton had worked with Jennifer for three weeks in 1996, the year in which the crime allegedly had occurred and been reported by Jennifer, and she had seen Jennifer on and off throughout the day during those three weeks. She fired Jennifer after discovering that she had lied in her criminal background check about whether she had any criminal convictions. The issue was important for Southerton because the business for which she employed Jennifer was a day care center for children. Further, the criminal background form that Jennifer falsely completed stated that it was important to be truthful in completing it.
As explained in federal case law and treatises, the trial court’s reasons for excluding the testimony do not provide a proper basis for the exclusion of opinion evidence. Instead, as long as the opinion is rationally based on the witness’s personal knowledge and would be helpful to the trier of fact, her opinion should usually be admitted and any defects in its bases be explored on cross-examination. Wright and Gold, 28 Federal Practice and Procedure § 6114, at 53-54; Weinstein, 4 Weinstein’s Federal Evidence § 608.13[2], at 608-25; Watson, 669 F2d at 1382.
There is no indication in the record that the trial court concluded or could have properly concluded that the proffered evidence did not meet those requirements. It is clear that the opinion was rationally based on the witness’s perception. First, it is beyond question that telling a lie is probative of one’s truthfulness, so the opinion was rational in that sense. Second, the opinion was based on sufficient perception. Although Southerton knew Jennifer for only a brief period, she interacted with her throughout each day of that brief period, and she caught her in one overt lie. Other appellate courts have concluded that it is an error of law to exclude opinion witnesses who had similar opportunities to perceive the substantive witnesses. See, e.g., Watson, 669 F2d at 1382 n 6; Wright and Gold, 28 Federal Practice and Procedure § 6114, at 54 n 13 (citing Honey v. People, 713 P2d 1300, 1303 (Colo 1986); State v. Morrison, 351 SE2d 810 (NC App 1987)).
*166For example, in Watson the court held that it was an error of law to exclude several opinion witnesses based on insufficient perception. One of them was an employee of a public agency to which the substantive witness had made a single false complaint. Another had supervised the substantive witness for a period of two weeks. A third opinion witness had lived near the substantive witness for several years, but the opinion witness’s only contact with the substantive witness apparently consisted of their working together on two or three occasions for less than an hour each. Watson, 669 F2d at 1382 n 6. Jennifer’s former boss had more contact with the substantive witness than did the three opinion witnesses in Watson; it is clear that the degree of interaction between the opinion witness and Jennifer satisfies the minimal constraints of the personal knowledge requirement and that the context of the lie provided an ample basis for Southerton to form an opinion about Jennifer’s character for truthfulness. Moreover, it is clear that the testimony sought to be introduced would have been helpful to the trier of fact.
The majority nonetheless concludes that the exclusion of Southerton’s testimony was proper apparently because it believes that it is irrational to base an opinion as to truthfulness on one lie and because it believes that admitting the testimony would blur the distinction between opinion evidence and evidence of specific bad acts. On the contrary, there is nothing irrational about basing an opinion on a single act, especially one as probative of character for truthfulness as an overt lie made after being told the importance of truthfully reporting one’s criminal background. The practical reality is that we must often make our decisions about others based on very limited information. It is unusual to catch someone in an overt lie, because lies are designed to be hidden. Therefore, catching someone in one overt lie in a short period of time logically supports an inference that the person has an untruthful character. Because the excluded evidence was rationally based and helpful to the trier of fact, the trial court did not have the authority to exclude it for lack of foundation. The proper place to explore weaknesses in the basis for the opinion, such as the length of time that Southern knew Jennifer, would have been on cross-examination.
*167In addition, although the majority states otherwise, admitting Southerton’s testimony would in no way conflate opinion evidence with evidence of specific untruthful acts. Specific acts evidence is prohibited in most cases because of its potential prejudicial effect. The perceived danger is that the factfinder will judge the substantive witness, often the criminal defendant, based on the witness’s prior acts rather than making a decision based on the evidence in the case. The drafters of the rules of evidence, both in Oregon and elsewhere, have simply decided that opinion evidence does not carry the same risk of prejudice as does evidence of prior bad acts. The majority is without power to refuse to enforce that decision, and its contention that some opinion evidence is as prejudicial as some bad acts evidence is a non sequitur.
I also take issue with the majority’s analysis of the trial court’s exclusion of the opinion witnesses who would have testified as to Kelly’s character for truthfulness. The trial court excluded the opinion testimony of all four of the character witnesses whom defendant wished to use to impeach Kelly, the mother of complaining witness B. The opinion testimony of three of the witnesses was excluded because their interaction with Kelly had occurred about three years before trial, during the year that the crime allegedly had occurred and been reported by Kelly.6 The other witness’s opinion testimony was excluded based simply on a lack of foundation, without further explanation. Although remoteness from the time of trial is a basis for excluding reputation evidence, it is usually not a basis for excluding opinion evidence. See Watson, 669 F2d at 1381-82 & 1382 n 5; Weinstein, 4 Weinstein’s Federal Evidence § 608.13[2], at 608-25; Carsner, 894 P2d at 151. The majority errs in relying on Caffee to conclude otherwise.
While it is possible that some interaction could be too remote from both the substantive witness’s report of the *168alleged crime and the trial to permit a finding of personal knowledge, that is not the case here. Here, the character witnesses’s interaction with Kelly occurred almost contemporaneously with her reports of the alleged crime and only about three years before trial.7 The trial court apparently assumed that the only relevant reference point was the time of trial. As previously stated, the relevant period for opinion evidence extends from the time of the crime to the time of the trial. Crumby, 895 F Supp at 1364. Because Kelly’s contact with the character witnesses occurred at about the same time as the crime and as the statements by Kelly about the crime that were identified or admitted at trial, I would conclude that the character witness had personal knowledge of Kelly’s character during a relevant time period.8 In excluding the opinion testimony of three of the witnesses because of its remoteness, the trial court relied on an incorrect understanding of the foundation requirements for opinion testimony. It therefore committed an error of law.9
Finally, I address whether exclusion of the character witnesses in this case requires reversal. To warrant reversal, an evidential error must have substantially affected the rights of a party. OEC 103(1); ORS 19.415(2). “[A]n error by the trial court substantially affects the rights of an aggrieved *169party if the outcome of the case either would have or may have been different if the error had not occurred. Where such prejudice exists, the error requires reversal.” Hernandez v. Barbo Machinery Co., 327 Or 99, 112, 957 P2d 147 (1998) (citations omitted). As in many sexual assault cases, the state’s case against defendant turned on the credibility of its witnesses. See, e.g., State v. Reynolds, 324 Or 550, 559 n 5, 931 P2d 94 (1997); State v. Bairnsfather, 591 So 2d 686, 688 (La 1991). The state offered no physical evidence of the alleged abuse. In that circumstance, the erroneous exclusion of evidence that bears on the credibility of the state’s witnesses could well have affected the result in the case. The majority concludes otherwise by labeling as cumulative the evidence that it believes was erroneously excluded.
I begin with the neighbor who would have testified as to the character of K, one of the alleged child victims in the case. That witness is the only witness who would have testified as to K’s character, and the majority agrees that the trial court committed error in excluding her testimony. It concludes, however, that the error was harmless. The testimony at issue concerned the credibility of one of the state’s most important witnesses: one of the two alleged child victims. No other witness testified as to the child’s character for truthfulness, and the entire case hinged on the credibility of witnesses; no physical evidence was offered. It follows that the exclusion of that evidence could have affected the result in the case and, therefore, that it was not harmless.
The majority goes to great lengths to portray this case as a credibility contest between defendant and Jennifer, K’s mother, and to downplay the importance of K’s credibility to the verdict against defendant. If the jury had not believed K, it could not have convicted defendant on the count involving her. Only one witness sought to testify as to K’s character. By no stretch of the imagination could that testimony be cumulative. The defense may have proceeded on the theory that Jennifer had lied and coached K to lie. The fact that defendant chose to emphasize the possibility that Jennifer was lying and coaching K, rather than emphasizing the possibility that K was lying to Jennifer and everyone else, does not change the fact that the state had the burden to prove defendant’s guilt beyond a reasonable doubt. Moreover, it *170would be easier for the jury to believe that Jennifer had coached K to lie if the jury believed that if was someone who would lie. In this case, admissible evidence that could have materially advanced defendant’s case was excluded. I would reverse defendant’s conviction as to the count involving K.10
I now turn to the witnesses who would have testified about Kelly, the mother of B, the second alleged child victim in this case. The excluded testimony bears on count three, which named B as the victim. The crime against B allegedly occurred between January 1, 1994, and September 1, 1994. Only one witness was permitted to testify with respect to Kelly’s character for truthfulness. That testimony concerned her reputation rather than the witness’s opinion of Kelly’s character, and, because it was admitted with very little foundation, it had only slight probative value. The court excluded the opinion testimony of four witnesses, including that of the witness who testified as to Kelly’s reputation.
I focus on the excluded opinion testimony of one of the witnesses, Debra Coonradt. She would have testified that her contact with Kelly began in late May 1994 when Kelly and her family moved the bus in which they lived onto Coonradt’s property, and that it continued until Kelly and her family moved away in early July. During that period, Coonradt interacted daily with Kelly and got to know her very well. The trial court excluded her opinion testimony as too remote, and the majority concludes that the court acted within its discretion in doing so. As discussed above, both analyses are based on errors of law with respect to the foundational requirements for admission of opinion evidence. Based on the short time lapse between the contact and Kelly’s first reports of her daughter’s accusations, and the frequency of contact between Kelly and Coonradt, I would conclude that the evidence was sufficiently probative that it could have affected the jury’s decision with respect to the *171count involving B. The lack of other viable character testimony about Kelly reinforces my conclusion that the admission of Coonradt’s testimony could have affected the jury’s decision to convict. Not only was the character evidence that was admitted with respect to Kelly of very slight probative value, it also was of a different category from the evidence at issue here. Because no similar testimony was admitted, Coonradt’s testimony could not have been cumulative. Hence, I would reverse the count involving B.
Because I believe that the trial court improperly conflated the foundational requirements for reputation and opinion evidence and that its error in that regard was not harmless, I would reverse both sodomy counts. Accordingly, I respectfully dissent.
The crucial difference between the two rules is that the federal rule allows cross-examination about specific acts of untruthful conduct while the Oregon rule does not permit such an inquiry. Compare OEC 608(2) with FRE 608(b).
Although the federal rationale for the relatively low foundational requirements for opinion evidence is not entirely applicable to the state rule because the federal rule, FRE 608(b), allows cross-examination into the specific acts that form the basis for the character witness’s opinion, while OEC 608(2) does not, I am persuaded that much of the federal rationale for the foundational requirements for opinion evidence is applicable to the Oregon rule and that the inherent differences between reputation and opinion evidence compel the different foundational requirements for them.
*162First, in spite of the fact that no inquiry into specific acts is permitted in state court, cross-examination can still be used to reveal most defects in the bases for opinion testimony, such as personal hostility toward the witness who is the subject of the opinion testimony or a limited relationship with the witness. Moreover, if the trial court determines that the character witness’s opinion is based on irrelevant instances of misconduct, it is free to exclude the testimony.
There also are inherent differences between reputation and opinion evidence that lead to their different foundational requirements. Reputation evidence is necessarily based on hearsay, a form of evidence that the law has traditionally viewed as suspect. Because of the potential unreliability of reputation evidence, it is appropriate to require those testifying about a person’s reputation to show long acquaintance with the person to be impeached, the community in which she lives, and the circles within which she moves. Without those safeguards, the risk is great that the reputation evidence will be completely unreliable because it will be based on unrepresentative opinions of a few out-of-court declarants that may themselves be based on irrelevant acts of the witness to be impeached. See, e.g., Charles Alan Wright and Kenneth W. Graham, Jr., 22 Federal Practice and Procedure § 5265, at 583 (1978) (stating that opinion evidence generally tends to be more reliable than reputation evidence). In contrast, when a character witness testifies as to her opinion, the rules require that the opinion be based on personal knowledge. See id. § 5265, at 584. The personal knowledge requirement obviates the need for the additional foundational requirements of reputation evidence. Moreover, because the opinion witness testifies from her personal knowledge, cross-examination can be expected to reveal defects in the bases of her testimony that could not be as readily revealed, or revealed at all, in the cross-examination of a reputation witness. See John B. Weinstein, 4 Weinstein’s Federal Evidence § 608.13121, at 608-25 (2d ed 1997). Finally, because a reputation witness claims to speak for the community, whereas an opinion witness merely offers a personal view, a jury is likely to give more weight to the testimony of the reputation witness. The greater persuasive force of reputation testimony justifies the higher foundational requirements. Thus, although the Oregon Legislature has determined under OEC 608 that inquiry into specific acts will not be allowed on cross-examination, the inherent differences between reputation and opinion evidence and the fact that effective cross-examination will reveal most defects in a witness’s opinion persuade me that the best course is to adhere to the widely accepted distinction between the foundation required for reputation and opinion evidence.
4Focusing on the time period from the crime to the trial is logical whenever the credibility of the witness sought to be impeached concerns the criminal conduct. The fact that the case law does not require an opinion witness’s personal knowledge of the substantive witness to be based on interaction very close to the time of trial indicates that, at least with opinion testimony, the focus is not solely on whether the witness is an honest person at the time of trial. Instead, because a witness’s honesty on the stand is based on the reliability of her perception of the underlying events, a witness’s credibility comprises two components: (1) whether *164the witness is honest in her perception of the events at issue and (2) whether the witness tells the truth about that perception at trial. In other words, if a person’s perception of the events at issue is colored by dishonesty, even if the person has somehow become an honest person by the time of trial, the portrayal of the events at the time of the crime will still be tainted by the dishonest perception. Thus, an opinion witness may testify so long as the knowledge of the substantive witness is not too remote from either the time of trial or the period during which the disputed events allegedly occurred. The point is even more obvious where, as here, the evidence at trial includes statements made by the witness at the time of the crime as well as at trial.
Ordinarily, the trial court’s decision to admit or exclude character witnesses under OEC 608 is reviewed for abuse of discretion. Caffee, 116 Or App at 27. However, because the issue of whether the trial court based its decision to exclude the evidence on the proper foundational standard is a question of law, that issue should be reviewed for errors of law. United States v. Watson, 669 F2d 1374, 1382-83 (11th Cir 1982). See also State v. Reynolds, 324 Or 550, 558, 931 P2d 94 (1997) (holding that the trial court erred as a matter of law in refusing to allow defendant to offer surrebuttal testimony as to his character for truthfulness after the state had attacked it and that that error required reversal); United States v. Dotson, 799 F2d 189, 193 (5th Cir 1986) (holding that the record must contain some indication that the trial court evaluated the opinion evidence according to the proper foundational standards).
If opinion evidence had to be community based and characteristic in order to be admissible, it would be indistinguishable from reputation evidence, and there would have been no need to draft the evidence rules to allow the admission of opinion evidence.
The one character witness against Kelly who testified was allowed to testify only as to Kelly’s reputation for truthfulness, although defendant sought to have him testify about his opinion of her character for truthfulness as well. The trial court’s basis for excluding the opinion testimony was improper foundation. (The witness’s reputation testimony was apparently admitted only because the state failed to object.)
Because it is difficult to pin down the exact time lapse between the alleged crime and the witnesses’ interaction with Kelly, I focus on the time lapse between Kelly’s reports of the crime and her interaction with the opinion witnesses. In the indictment, the state alleged simply that the crimes against B had occurred between January 1, 1994, and September 1, 1994. Kelly’s testimony at trial indicated that she and her family had moved from Oregon to Washington in August 1994 and that B had had no contact with defendant after the move. The opinion witnesses who would have testified about Kelly indicated that their interaction with her had ended when she moved. Moreover, it appears from the trial transcript that B told Kelly about defendant’s alleged crimes against her at the end of September or the beginning of October 1994, and that Kelly first began to report the events to authorities toward the end of October 1994.
I need not decide whether, if Kelly’s testimony on the stand were her only statements that were subject to impeachment, the trial court could have excluded the character witness’s opinion testimony based on its determination that a time lapse of roughly three years between the trial testimony and the contact did not meet the personal knowledge requirement.
Caffee is the case on which the majority relies for the proposition that the foundational requirements for reputation and opinion evidence are the same, and it is the case from which Kirkpatrick extrapolates in describing the foundational requirements for opinion evidence. In fact, Caffee did not purport to decide the issue that we are faced with here.
Because I would reverse as to the count involving K based on the exclusion of the witness who would have testified as to her character, there is no need for me to address whether it was harmless error to exclude the testimony of witnesses who would have to testified as to Jennifer’s character for truthfulness.