OPINION

FABE, Chief Justice.

I. INTRODUCTION

In a personal injury trial resulting from a car accident, the plaintiff sought to cross-examine the defendant's medical expert about his substantial connection to the insurance industry in an effort to prove bias. But in response to defense counsel's motion in limine, the district court ruled that the plaintiff could not refer to the fact that the defendant was insured or that her insurance company and others had hired the expert witness numerous times. The trial court did permit the plaintiff to cross-examine the expert witness about his financial interest in continuing to work for "defendants" and "defense attorneys."

On appeal from the district court judgment, the superior court concluded that the district court had abused its discretion by excluding evidence of the expert's connections to the insurance industry, reasoning that the expert witness and the company which hired him had extensive dealings with the defendant's insurance company and the insurance industry more broadly and that this information was relevant to the question of bias. We agree with the superior court that the district court erred in ruling that relevant evidence of the expert witness's substantial connection to the insurance industry should be excluded. But the district court's error was harmless because at trial the plaintiff was able to elicit testimony about the witness's connection to the insurance industry. We therefore vacate the superior court's remand order and reinstate the district court's judgment.

II. FACTS AND PROCEEDINGS

A. Facts

In July 2009 Kimber Ray rear-ended an automobile in which Megan Draeger was a passenger. There was no serious physical damage to either car, and at trial the accident was described as a low speed, low impact collision.

Draeger did not complain of any injuries at the accident scene. But soon afterward, she experienced pain in her neck and shoulders, and she made an appointment to see a chiropractor six days after the accident. Draeger had a total of 24 chiropractic treatments between July and November 2009 at a total cost of $5,160. Nine months later, in August 2010, Draeger sought treatment from a physical therapist for neck pain and headaches. The physical therapist treated Draeger over nine sessions between August and September 2010.

B. Proceedings

Draeger filed suit against Ray in the district court in July 2011. Ray admitted liability for the accident, so the trial focused on the extent of Draeger's injuries and other damages related to the accident. Ray's insurer, Government Employees Insurance Company (GEICO), paid for her defense.

Ray filed a motion in limine based on Alaska Evidence Rule 4111 seeking to preclude reference at trial to the fact that Ray was covered by liability insurance with respect to Draeger's claims. Draeger partially opposed the motion, arguing that she wished to cross-examine Dr. John Ballard, an orthopedic surgeon hired by Ray's counsel to give expert testimony at trial, regarding potential bias. In particular, Draeger sought to examine Dr. Ballard about the fact that a substantial portion of his work as a medical expert is derived from referrals from insurance companies and that he had been hired many times by GEICO in particular.

District Court Judge David Wallace granted Ray's motion. Judge Wallace concluded that under the required Alaska Evidence Rule 403 balancing,2 the "minimal relevance" of the insurance evidence being offered to show bias, which is allowed under Rule 411, was outweighed by "the prejudicial nature and confusion of issues" that would more likely result if Draeger impeached Dr. Ballard by referencing GEICO or the fact that Ray was insured. The order stated that "witnesses and parties shall be instructed that no reference should be made to the fact that defendant has liability insurance" or to "the fact that persons investigating this matter may have been employed by defendant's liability insurance." However, the order noted that "this ruling does not preclude plaintiff's counsel's ability to cross examine any witness on the issue of bias that relates to the nature of their work for defendants and/or the fact [that] there may be ... financial reasons for the continuation of wanting to work for defendants."

Dr. Ballard frequently performs independent medical evaluations and medical record reviews for insurance companies. GEICO retained Dr. Ballard to evaluate medical records or conduct independent medical examinations 20 to 30 times in 2011 alone, though Dr. Ballard stated in a deposition that less than 5% of his work in Alaska comes from GEICO. Dr. Ballard does about 50 to 60 record reviews and 200 to 300 independent medical evaluations per year; 40% of those evaluations are for claims of injury arising out of automobile collisions. He also co-founded a medical evaluation company called The Independent Medical Evaluators (T.I.M.E.), which provides medical evaluations. More than 98% of T.I.M.E.'s clients are insurance companies or defense attorneys. Although Dr. Ballard was no longer a co-owner of the company when he conducted the evaluation for this case, Ray's counsel hired and paid him through T.I.M.E. for this case. Dr. Ballard made over $100,000 per year from insurance defense referrals and between $300,000 to $350,000 per year from his insurance-related medical exam work. His total annual income is between $700,000 and $800,000, which includes his private orthopedic practice. Dr. Ballard testified that his income and the prospect of future employment by Ray's counsel's office were irrelevant to the opinions he offered regarding Draeger's injuries and treatment.

Dr. Ballard testified that, in his opinion, Draeger's chiropractic treatments in 2009 were reasonable and related to her cervical, thoracic, and lumbar strains resulting from the accident. However, he testified that Draeger's later physical therapy in 2010 was not related to the accident and that, in his opinion, Draeger had recovered from the accident no later than November 2009.

During cross-examination, Dr. Ballard confirmed that he was hired for evaluation in this case through T.I.M.E., and when asked by Draeger's counsel whether "more than 98% of [his evaluations through T.I.M.E.] are for insurance companies or defense attorneys," Dr. Ballard answered, "Correct." Ray's attorney objected. At a bench conference Judge Wallace reiterated his order regarding testimony about insurance: "I thought I made it clear in my decision not to mention insurance. You can say `defense attorneys,' you can say `defense.'" Neither party requested a curative instruction. There were no further references to insurance during trial.3

The jury awarded Draeger $5,160 for past economic loss, equal to the full amount of her chiropractic treatments from July to November 2009. It also awarded her $775 for past non-economic damages. The jury did not award Draeger any damages for future economic loss or future non-economic loss.

Draeger appealed to the superior court. She argued that it was improper for the district court to preclude her from questioning Dr. Ballard about the extent to which he was paid for his work as an expert witness by insurance companies. Judge John Suddock issued an opinion reversing the district court's order limiting Draeger's cross-examination of Dr. Ballard and remanding the case to the district court for a new trial. The superior court concluded that the district court had erred in preventing Draeger from cross-examining Dr. Ballard regarding his "relationship with the insurance industry in general and GEICO in particular," despite the fact that the district court's ruling allowed Draeger to impeach Dr. Ballard regarding his near-exclusive work for "defendants" and "defense lawyers." Ray filed a petition for review, which we granted.

III. STANDARD OF REVIEW

"In an appeal from a judgment of a superior court acting as an intermediate court of appeal, we independently review the judgment of the district court."4 "We review a trial court's decision to admit or exclude evidence for an abuse of discretion."5 We review a trial court's Rule 403 ruling for abuse of discretion by "balanc[ing] the danger of unfair prejudice against the probative value of the evidence to determine whether the potential danger predominated so greatly as to leave us firmly convinced that admitting the challenged evidence amounted to a clear abuse of discretion."6 We reverse only if "the error affected the substantial rights of a party."7

IV. DISCUSSION

A. The District Court Abused Its Discretion By Not Admitting Evidence Of Dr. Ballard's Substantial Connection To The Insurance Industry.

1. Evidence of a witness's connection to the insurance industry is admissible to show bias if its probative value outweighs the danger of unfair prejudice.

Alaska Evidence Rule 411 provides that "[e]vidence that a person was or was not insured against liability is not admissible" to prove "whether the person acted negligently or otherwise wrongfully."8 But the rule allows courts to admit this evidence when offered for another purpose, such as to show "bias or prejudice of a witness."9 When Rule 411 does not bar evidence, it may still be excluded under Alaska Evidence Rule 403.10 Rule 403 provides that relevant evidence "may be excluded if its probative value is outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury," among other factors. Alaska Evidence Rules 411 and 403 thus work in conjunction: If the trial court decides that evidence of liability insurance can be admitted despite Rule 411, the court must then perform a Rule 403 balancing analysis to determine whether the evidence's probative value outweighs the danger of unfair prejudice.11 "Under Evidence Rule 403, the trial court bears primary responsibility for determining admissibility of evidence."12 As we have held previously, trial courts generally "have broad discretion in applying [the Evidence Rule 403] balancing test."13

"When reviewing the exclusion of evidence under Evidence Rule 403 as unfairly prejudicial, we first consider the relevance of the [excluded evidence] and then determine whether ... [the superior court's exclusion of it] constitutes a clear abuse of discretion."14 "The credibility of witnesses is always a material issue, so the only question of materiality or relevance when evidence is offered to impeach for bias is whether the evidence tends in reason to demonstrate the existence of some fact, state of mind or condition that a reasonable person would take into account in assessing the credibility of the witness under attack."15

Although evidence of bias is relevant and probative, courts sometimes view evidence of liability insurance as prejudicial. Rule 411 "is designed to prevent a jury from deciding a close case on an improper basis — i.e., whether or not a party is insured."16 Courts can exclude evidence of a defendant's insurance "to avoid prejudice in the verdict, which might result from the jury's knowledge that insurance, and not the defendant, would be responsible for paying any resulting award of damages."17 However, we have yet to consider whether the danger of unfair prejudice can outweigh the probative value of evidence offered to show an expert's bias in favor of the insurance industry in an automobile injury case.18 We turn next to that question.

2. The probative value of evidence of a witness's connection to the insurance industry is likely to outweigh the danger of unfair prejudice if the connection is substantial.

We have not previously articulated a steadfast rule to guide the trial court's determination of whether the probative value of a witness's connection to the insurance industry offered to show bias outweighs the danger of unfair prejudice under Rule 403. The majority of jurisdictions that have considered this question allow insurance evidence to show bias only in cases where a "substantial connection" exists between a witness and the insurance industry.19 "The substantial connection analysis looks to whether a witness has a sufficient degree of connection with the liability insurance carrier to justify allowing proof of this relationship as a means of attacking the credibility of the witness."20

Some jurisdictions define what constitutes a substantial connection between a witness and an insurer in terms of an ownership, agency, or employment relationship — that is, a relationship in which a witness has a "direct interest in the outcome of the litigation."21 These jurisdictions often limit when a witness's pecuniary interest can be used to establish a substantial connection by requiring something "[b]eyond mere payment in exchange for testimony [at] trial."22 Courts have found a substantial connection where an expert witness was employed by and consulted for an insurance company and 10-20% of the expert's practice consisted of reviewing records for insurance companies,23 and where an expert was employed by a consulting firm that derived roughly 30% of its income from insurance companies.24

Importantly, courts have recognized a substantial connection in the absence of a formal employment relationship between an expert witness and an insurer. For example, in Lombard v. Rohrbaugh, the Virginia Supreme Court upheld a trial court's decision to admit evidence of a medical expert's relationship with an insurance company in an automobile accident case because the medical expert had received over $100,000 annually from the insurance company for at least two years.25 The defendant argued that opposing counsel should not have been permitted to mention insurance because the witness was not an employee of the insurance company.26 The Lombard court held that "[a] witness'[s] status as an employee of an insurance company providing coverage to a party is evidence of potential bias, but the absence of an employer-employee relationship does not define the limits of cross-examination. At issue is the potential for bias because of the witness'[s] interests in the case, not artificial labels."27 We agree. This is particularly true given the modern corporate structure where employment and consulting relationships are often created ad hoc or through an intermediary and do not conform to traditional direct employment relationships.

The Virginia Supreme Court has addressed a fact pattern that is similar to the one in this case. In Henning v. Thomas, a plaintiff in a medical malpractice case hired an expert witness through a company that "identifies other physicians throughout the country who are willing to review medical records and provide medical testimony."28 The witness signed an agreement with the company to assess cases and was to receive a set fee for reviewing depositions or transcripts and testifying; the company kept a portion of the overall fee paid to it by the plaintiff.29 The defendant sought to cross-examine the witness as to who retained him to testify, but the trial court did not permit "any question other than whether [the witness] was being paid to come to court to give his testimony."30 The Virginia Supreme Court reversed, reasoning:

We similarly do not think a party's cross-examination should be limited solely to questions regarding whether a witness is being paid to testify.

The trial court's substantial connection analysis should look primarily to "whether a witness has a sufficient degree of connection with [a] liability insurance carrier to justify allowing proof of this relationship as a means of attacking the credibility of the witness."32 Where an expert witness has significant ties to the insurance industry as indicated by receiving a sizable portion of his or her income from insurance work, being hired by a firm that derives a large portion of its income from insurance companies, or facts that otherwise suggest an interest in the outcome of the litigation, the probative value of that substantial connection is likely to outweigh the danger of unfair prejudice, and is thus likely admissible to show bias under Rule 411 and Rule 403.

3. To establish a substantial connection, a party may present evidence about a witness's ties to the insurance industry without necessarily eliciting evidence about the other party's individual insurance coverage.

There are two types of testimony regarding insurance that may be at issue in this type of case. The first is whether the defendant was insured against liability, and if so, by which company. The second is whether an expert witness hired by the defense may have a substantial connection to the insurance industry and possibly a small number of companies in particular, which may be evidence of bias. At trial Draeger focused primarily on this second category: Dr. Ballard's connection to the insurance industry. Her opposition to Ray's motion in limine argued that "[s]tating merely that these experts do most of their work for the defense, or for defense attorneys, only exposes the tip of the iceberg. Working for a particular entity with a singular incentive to minimize the harm resulting from auto collisions implies a stronger incentive for bias than working for different defense attorneys or `the defense.'" Because Draeger's primary interest was to explore Dr. Ballard's connection to the insurance industry, admission of such evidence as how often Dr. Ballard worked for insurance companies or firms that serve the insurance industry and what percentage of his income was from his insurance work33 — without revealing details about Ray's coverage — would have satisfied Draeger's goal of attempting to demonstrate Dr. Ballard's potential bias.

The district court, however, only allowed Draeger to probe the expert's potential bias with a limited set of terms revolving around "the defense": Its order stated that "this ruling does not preclude plaintiff's counsel's ability to cross examine any witness on the issue of bias that relates to the nature of their work for defendants and/or the fact [that] there may be ... financial reasons for the continuation of wanting to work for defendants." This ruling excluded the use of the word "insurance" in any regard. But restricting cross-examination to the use of the terms "defendants," "defense counsel," or "defense expert" may not convey the witness's potential bias to a juror who may assume that attorneys represent both personal injury plaintiffs and defendants. The South Carolina Supreme Court addressed this specific question in Yoho v. Thompson, when the trial court had "informed [the plaintiff] that she could discuss [the expert witness's] bias by using generic terms such as `defense,' `defendants,' and `defense lawyer,' but that she could not discuss his possible bias by using the word `insurance.'"34 The Yoho court rejected this approach, reasoning that references to insurance work are "qualitatively different from showing [that the witness] works for `the defense' generally, and [are] much more indicative of possible bias."35

Parties should not have to couch a witness's relationship with the insurance industry in code by referencing only "defense attorneys" or "defense work" when presenting evidence of an expert witness's bias. Jurors may not understand this reference. And the point is that an expert witness with a substantial connection to insurance companies is working for the side with an interest in minimizing claims, and describing the witness's relationship with insurers in clear terms is the most direct way to relay the entirety of that message to the jury.36 Though a jury may infer from this evidence that a party is insured, this inference presents a low risk of unfair prejudice because jurors will be aware that Alaska law requires drivers to carry automobile insurance.37 The opposing attorney should be able to elicit evidence about the witness's connection to an institutional set of insurance clients and to argue that the expert witness may wish to ensure repeat business.

But a party may not need to specify any insurance company by name in order to establish a substantial connection to insurance clients. In many cases, evidence of a witness's frequency of work for the insurance industry more broadly and the percentage of the witness's income derived from insurance work sufficiently demonstrates the potential bias. Thus, in many cases, the trial court will not need to admit details of a party's individual insurance coverage or information about any insurance company by name. The opposing party can often adequately expose the potential bias of an expert witness by showing that he has a close connection to the insurance industry.38

4. Dr. Ballard has a substantial connection to the insurance industry and the district court abused its discretion by not admitting evidence of that connection.

There is a substantial connection between Dr. Ballard and the insurance industry. Like the expert in Lombard v. Rohrbaugh,39 Dr. Ballard is highly compensated by the industry; he receives between $300,000 to $350,000 a year for his insurance reviews. This represents a large percentage of his total yearly income of up to $800,000, which includes his private orthopedic practice. And like the expert witness in Mitchell v. Glimm who was hired by a consulting firm that derives a sizable portion of its income from insurance companies,40 Dr. Ballard was hired for this case by a company that does 98% of its work for insurance companies or defense attorneys. The financial entanglements of both Dr. Ballard and the consultancy through which he was hired create a substantial connection to the insurance industry.

Here, the district court excluded all reference to insurance as more prejudicial than probative, ordering that "witnesses and parties shall be instructed that no reference should be made to the fact that persons investigating this matter may have been employed by defendant's liability insurance." On appeal the superior court reversed the district court's exclusion of insurance evidence and suggested that evidence of a substantial connection is invariably more probative than prejudicial. The trial court has broad discretion under Rule 403 balancing to determine whether the probative value of evidence outweighs its potential to create unfair prejudice.41 But if evidence of an expert witness's substantial connection to the insurance industry is available, the weight of factors that the trial court must balance will generally be static because the potential for unfair prejudice will probably not vary and thus should tilt in favor of admission, absent unusual factual circumstances. Because evidence of Dr. Ballard's substantial connection to the insurance industry was available and was probative of bias, and there were no unusual factual circumstances in this case to suggest that the risk of unfair prejudice outweighs its probative value, the district court abused its discretion by excluding direct evidence of the connection.

B. Because The Jury Heard Evidence Of Dr. Ballard's Substantial Connection To The Insurance Industry, The District Court's Error Was Harmless.

Although it was an abuse of discretion for the district court to attempt to exclude direct evidence of Dr. Ballard's substantial connection to the insurance industry, the error was harmless. "When the trial court has erroneously excluded evidence, a party must show that the error was harmful or prejudicial before we will reverse the trial court. The test for determining whether an error was harmless is whether on the whole record the error would have had a substantial influence on the verdict of a jury of reasonable lay people."42 Thus, "[a]n erroneous ruling requires reversal only if our review of the record convinces us that it probably had a substantial effect on the jury."43

Though the district court aimed to exclude all evidence about insurance through its order granting Ray's motion in limine, the jury ultimately did hear evidence about Dr. Ballard's substantial connection to the insurance industry. During cross-examination Draeger's counsel established that Dr. Ballard was hired for evaluation in this case through T.I.M.E. and that he performs more than 100 medical examinations per year through T.I.M.E. in auto injury cases. When Draeger's lawyer asked Dr. Ballard to confirm that nearly all "evaluations through T.I.M.E.[,] like what [he] did in this case, more than 98% of [evaluations] are for insurance companies or defense attorneys," Dr. Ballard answered, "Correct." Though Ray's attorney objected and during a bench conference the district court instructed the parties not to mention insurance again, the trial court did not strike the testimony or give any curative or limiting instruction to the jury after Dr. Ballard's testimony regarding T.I.M.E.'s primarily insurance-based clientele. Thus, the jury did hear some of the evidence that Draeger's attorney advocates for here. Because the district court's attempted exclusion of all insurance evidence was unsuccessful, it could not have had a substantial effect on the jury. And because Dr. Ballard provided an answer to a question that should have been allowed, the district court's error was harmless.

V. CONCLUSION

Although it was error to fail to admit evidence of the expert witness's substantial connection to the insurance industry, which was not unfairly prejudicial, this error was harmless because the jury ultimately did hear some testimony about that connection. We thus AFFIRM in part the superior court's reasoning but REVERSE and VACATE the superior court's order remanding for new trial and REINSTATE the district court's judgment.