OPINION

I. INTRODUCTION

A state employee applied for occupational disability benefits, claiming that prolonged sitting at work aggravated a preexisting medical condition. The Division of Retirement and Benefits denied the claim. An administrative law judge affirmed that decision, determining that employment was not a substantial factor in causing the employee’s dis*527ability. On appeal the superior court reversed the administrative law judge’s decision. Because the administrative law judge’s decision was supported by substantial evidence, we reverse the superior court’s decision and thereby affirm the administrative law judge’s decision.

II. FACTS AND PROCEEDINGS

A. Underlying Facts And Prior Proceedings

This case comes to us for a third time.1 The underlying facts and proceedings relevant to this appeal are fully set forth in Shea II.2

In brief, Shirley Shea suffers from chronic pain and has been unable to work since 2001.3 Shea was granted non-occupational disability benefits in March 2003,4 but was denied occupational disability benefits because the Division of Retirement and Benefits’ retained expert, Dr. William Cole, concluded after reviewing Shea’s medical record that “[tjhere is not evidence from the record that the pain was caused by her occupation.” 5 In response Shea underwent a series of medical exams between August 2003 and August 2005 seeking to determine the connection, if any, between prolonged sitting at her employment and her chronic pain.6 The Division found the new information unconvincing:

In August 2005, at the Division[’s] request, Dr. William Cole reviewed all the information in Shea’s medical record, including the opinions and medical reports Shea had obtained since Dr. Cole’s opinion in March 2003. After considering this information, Dr. Cole maintained his opinion that “there is not a substantial presentation of an argument to support [Shea’s] claim that her job activities were [a] significant contributing factor to this condition, no more than the rest of the activities of daily living of her life were.” As a result, the Division affirmed its denial of Shea’s claim for occupational disability benefits.[7]

Shea appealed this decision to the Office of Administrative hearings, and a hearing was held in March 2006.8 Both Dr. Michael Smith, whom Shea had seen in 2004 for an opinion on causation,9 and Dr. Joella Beard, whom Shea had seen in 2001 for a disability impairment rating,10 testified at the hearing.11

The Administrative Law Judge (ALJ) found by a preponderance of the evidence that Shea “suffered some form of injury to her ilioinguinal nerve in the course of the 1984 procedure, resulting in long-term unresolved ilioinguinal neuralgia.” But he defined her “disabling condition [as] chronic pain syndrome, primarily resulting from the nerve injury in 1984, and referred and secondary pain related to that injury.” He also found that the initial trauma from 1984 left Shea “with a vulnerable nerve, which intermittently flared up ... as a result of the activities of everyday life, leading eventually to secondary bursitis and referred pain in a variety of areas.” He noted that Shea’s bursitis, however, “is not disabling, and the chronic pain she suffers has many sources other than her working conditions.... Her claim for disability benefits rests on whether, in light of the record as a whole, her employment was a substantial factor in a complex chronic pain syndrome.”

The ALJ found that Shea “did not prove by a preponderance of the evidence that her employment was a substantial factor in her disability” and affirmed the Division’s denial *528of Shea’s occupational disability claim;12 Shea appealed to the superioi* court, and it affirmed the ALJ’s decision.13 Shea appealed to this court, and we reversed the superior court’s decision upholding the ALJ’s decision and remanded for proceedings consistent with our explanation of the appropriate causation standard.14 ■

B. The ALJ’s Decision On Remand

The ALJ issued his decision on remand in February 2013. No new evidence was considered, but the ALJ did consider the parties’ briefs and our Shea II decision, which authorized the ALJ to “reevaluate the evidence ... as he deemfed] necessary,”15 The sole issue again was whether Shea’s employment was a substantial factor in causing her disabling pain.

The ALJ examined evidence indicating that prolonged sitting at work was a substantial factor in causing Shea’s disabling pain. Among this evidence was; “(1) Dr. Smith’s opinion [at the hearing] that-prolonged sitting aggravated a physical condition and her pain symptoms, ... (2) sitting was painful to her, (3) her job duties involved long periods of sitting, and (4) during the time she worked for the State of Alaska, her pain symptoms increased.”

The ALJ also considered evidence indicating that Shea’s employment was not a substantial factor in her disability:

(1) Ms. Shea on multiple occasions prior to becoming disabled reported that her pain was caused by a wide variety of common, every-day activities, including walking, and physical activity in general; (2) .Ms. Shea did not identify sitting as a causal factor until February, 1999, after her symptoms had become highly problematic; (3) Ms. Shea in 1998, and again in 1999, reported no significant or particular aggravating or alleviating factors; (4) Ms. Shea did not herself identify working conditions as. a causal factor until 2003, long after she had ceased working; (5) Dr. Beard’s expert medical opinion that prolonged sitting did not permanently aggravate the underlying physical condition; and (6) Dr. Beard’s observation that Ms. Shea’s pain symptoms could have been a result, in some degree, of psychological factors.

The ALJ determined that Shea had proved prolonged periods of sitting at work were a but-for cause of her disability. But the ALJ concluded that reasonable persons would not attach responsibility to the State for Shea’s disability because her employment conditions were not a sufficiently “significant and important a cause” of her disability; the ALJ therefore found Shea had not proved that prolonged sitting at work was a substantial factor in causing her disability. Shea then appealed the ALJ’s decision, to the superior court.

C. Appeal To The Superior Court

The superior court issued its decision in December 2014, reversing the ALJ’s decision after determining that it was not supported by substantial evidence, The superior court determined that “the factual findings upon which the ALJ based his conclusion that reasonable persons would not attach responsibility to Shea’s employer for her injury were not adequate to support his conclusion,” This determination was driven in part by the notion that “ft]he ALJ cannot find on one hand that Dr. Smith’s testimony was substantial evidence establishing actual cause but not substantial evidence supporting the causal component of proximate cause.”

The State appeals the superior court’s decision.

III. STANDARD OF REVIEW

When the superior court acts as an intermediate court of appeal in an adminis*529trative matter, we independently review the merits of the agency’s decision.16 We review a board’s factual findings “to determine whether they are supported by substantial evidence,” which is “such relevant evidence as a reasonable mind might accept as adequate to support the board’s conclusion.”17 “[W]e view the evidence in favor of the findings,” 18 and we will not choose between competing inferences or evaluate the strength of the evidence.19 We will look only to determine if substantial evidence exists in the record, taking into account evidence in the record detracting from the supporting evidence’s weight.20 “The conclusion that a work-related injury or hazard is not a substantial factor in causing an employee’s disability must be supported by substantial evidence. It is a legal question whether the quantum of evidence is substantial enough to support such a conclusion in the contemplation of a reasonable mind.”21

IV. DISCUSSION

To qualify for occupational disability benefits, Shea had the burden of proving22 that her employment was “terminated because of a total and apparently permanent occupational disability.”23 An “occupational disability” is

a physical or mental condition that, in the judgment of the administrator, presumably permanently prevents an employee from satisfactorily performing the employee’s usual duties for an employer or the duties of another comparable position or job that an employer makes available and for which the employee is qualified by training or education; however, the proximate cause of the condition must be a bodily injury sustained, or a hazard undergone, while in the performance and within the scope of the employee’s duties and not the proximate result of the wilful negligence of the employee.[24]

The presumption of compensability from the workers’ compensation context does not apply in the occupational disability benefits context, and the employee must prove by a “preponderance of the evidence that the disability was proximately caused by an injury which occurred in the course of employment.”25 An employee’s underlying injury does not need to have been caused by the employment; instead, an employee can qualify for occupational disability benefits by showing that the employment aggravated a preexisting condition by causing increased pain or other symptoms.26 This aggravation must be shown to be a substantial factor in contributing to the employee’s disability.27

The ALJ concluded that Shea had established sitting contributed to her disabling pain, but that she did not prove her employment was a substantial factor in causing her disability. This conclusion rested on the ALJ’s determination that “some reasonable persons would consider prolonged sitting at work so significant and important a cause as to attach legal responsibility for it and others would not” and that “the latter group is larger.” Five reasons supported this eonclu*530sion: (1) Dr, Beard’s testimony that Shea’s ilioinguinal neuralgia28 had resolved weakened Dr. Smith’s opinion that prolonged sitting at work was a substantial factor of Shea’s disability; (2) Shea did not report prolonged sitting at work as a pain trigger to any of her physicians prior to her disability claim’s denial; (3) ordinary daily activity aggravated Shea’s pain; (4) Dr. Smith’s testimony limited the aggravation of Shea’s chronic pain symptoms from prolonged sitting at 5% to 10%; and (5) expert testimony indicated psychological factors may have contributed to Shea’s chronic pain.

A. The Factual Findings Underpinning The ALJ’s Proximate Cause Decision Are Supported By Substantial Evidence.

Shea challenges the factual findings underpinning the ALJ’s proximate cause decision as unsupported by substantial evidence. On appeal we do not reweigh the evidence or choose between competing inferences, but we will look to the record’s entirety to ensure that substantial evidence supports the ALJ’s factual findings.29

1. Dispute between Dr. Beard’s testimony and Dr. Smith’s testimony

Dr. Beard explained that in her experience the nerve damage caused by Shea’s 1984 procedure would likely not continue for as long as Shea’s pain had continued. Dr. Beard also noted that ilioinguinal neuralgia is a condition aggravated by heavy lifting, extreme positioning, and sometimes standing or walking for long periods of time. When questioned about prolonged sitting’s effect on il-ioinguinal neuralgia, Dr. Beard stated that it would not “cause a permanent worsening or flare of [the] condition” if the person had the ability to adjust her position. This testimony directly contradicted Dr. Smith’s testimony that prolonged sitting would affect Shea’s condition, which was based on the belief that her ilioinguinal neuralgia had not resolved and remained the primary problem.

The dissent asserts that the ALJ erroneously concluded that Shea’s neuralgia had resolved. But the ALJ made no such finding. The ALJ found there was dispute between the doctors concerning the underlying cause of Shea’s disability and whether her ilioingui-nal neuralgia had resolved, and that this dispute weakened Shea’s case. Portions of Dr. Beard’s testimony provide substantial evidence for the ALJ’s finding that the ilioin-guinal neuralgia diagnosis was disputed: “I do not believe she had chronic pelvic pain,” and “even back then, no, I would not have diagnosed her or given her that label of ‘chronic pelvic pain.’ ”

The dissent endorses the superior court’s assertion that Dr. Beard’s testimony “does not necessarily show that Shea’s injury had abated; it shows either that (1) Shea’s condition was atypical, or (2) something else was causing her pain. Neither is proof that her ilioinguinal neuralgia resolved.” But that formulation shows that there were two competing factual inferences the ALJ could have drawn from Dr. Beard’s testimony, with evidence supporting either. Although the ALJ did not state he was drawing the latter inference — again, this was the superior court’s formulation — he did find that Dr. Beard’s testimony raised a dispute concerning the etiology of Shea’s current symptoms and that her pain could be caused by something other than ilioinguinal neuralgia substantially aggravated by work requirements.30

The dissent waves away this evident dispute by asserting that “the factual issue regarding the etiology of Shea’s pain was either uncontested or had been resolved in Shea’s favor,” and maintaining that the possibility of any medical dispute “is flatly contra-*531dieted by the ALJ’s earlier finding ... that Shea’s 1984 procedure ‘resultfed] in long-term unresolved ilioinguinal neuralgia.’ ” But on remand we authorized the ALJ to “reevaluate the evidence ... as he deemfed] necessary,” 31 and he was entitled to rely more heavily this time around on Dr. Beard’s testimony to discount the persuasiveness of Dr. Smith’s opinion concerning the role work requirements played in the progression of Shea’s disability.

And the ALJ’s findings prior to remand concerning the cause of Shea’s disability are not as clear as the dissent now asserts. The preponderance of evidence may have indicated that Shea had long-term unresolved ilioin-guinal neuralgia, but Shea’s disabling condition was found to be “chronic pain syndrome, primarily resulting from the nerve injury in 1984, and referred and secondary pain related to that injury.” The causes of that referred and secondary pain are and always have been in dispute: the ALJ found prior to remand that “the chronic pain she suffers has many sources other than her working conditions”; and in its brief on appeal the State argues that evidence of how other “ordinary daily life activities also aggravated Shea’s pain is relevant and material evidence of alternative causation.”

The dissent correctly notes that the “material, contested question” in this case is “whether sitting at work was so important a cause in Shea’s pain that reasonable persons would regard it as a cause and attach responsibility to it.” That being the inquiry, it is not clear how the possibility of an alternative cause of Shea’s pain was “not legally relevant” after remand — when Shea bore the burden of demonstrating the relationship between work requirements and her “complex chronic pain syndrome” — particularly when Shea’s own expert witness attributed no more than 10% of her symptoms to those work requirements.

Substantial evidence supported the ALJ’s interpretation of Dr. Beard’s testimony. And as the superior court noted, “fi]n other areas of law, the Alaska Supreme Court has held that where there are two or more conflicting medical opinions — each of which constitutes substantial evidence — the reviewing court will affirm the decision of the agency below.” 32 This principle supports upholding the ALJ’s finding that Dr. Beard’s testimony weakened Dr. Smith’s opinion that prolonged sitting at work was a substantial factor in Shea’s disability.

2. Shea’s prior failure to report to a physician that sitting at work increased her pain

The ALJ determined that prior to Shea’s disability claim denial she did not report prolonged sitting at work as a pain trigger to any physician. Shea did not report until 1999 that sitting in general increased her pain. And she did not report until 2003 thát prolonged sitting at work increased her pain, after she filed her occupational disability claim. Substantial evidence in the record supports the ALJ’s finding.

Shea asserts that the ALJ impermissibly disregarded testimony that prior to 1999 she had reported to her employer that sitting aggravated her pain. The ALJ’s fact finding on this issue relates only to when Shea first reported the link to a physician. Though Shea is correct that the record contains evidence she reported pain from sitting to her employer, this evidence does not controvert the ALJ’s physician-specific finding.

3. Ordinary activities and Shea’s pain

The record supports the finding that ordinary daily activities aggravated Shea’s pain. Between 1989 and 2001 Shea told various physicians that her pain was aggravated by a range of routine daily activities including working in her yard, any physical activity, using stairs, sitting, standing, walking, bending forward or backward, and lifting. Shea’s husband testified that Shea ceased doing routine activities like cooking, gardening, and walking her dog because those activities increased her pain.

4. Dr. Smith’s testimony on the upper bound of prolonged sitting as a source of Shea’s pain

Dr. Smith stated that prolonged sitting increased Shea’s pain by “maybe 5 or 10 *532percent, at the most.” There is no dispute about this testimony.

5. Dr. Beard’s testimony regarding psychological factors

Dr. Beard’s testimony clearly supports the ALJ’s observation that psychological factors may have contributed to Shea’s disabling pain. Dr. Beard testified: “I don’t believe she was disabled as much as she felt she was disabled. Her perception of her disability ... exceeded ... what would be ... mostly medically reasonable”; in situations like Shea’s the pain’s original source can resolve and the pain “itself becomes their driving force. And it’s a big emotional, psychological, psychosocial dilemma as to what is really driving this pain. So in her case, I would say ... it was her perception that she was not able to return to work, not necessarily ... that that was accurate.” Dr. Beard noted that without further psychological evaluation she could not definitively say whether Shea’s pain had a psychological, rather than physical, cause, but that based on the evidence at hand she saw no grounds to rule it out.

The ALJ did not find that psychological factors did cause Shea’s disability. He rioted only that they may have done so. Substantial evidence supports this observation.

B. The ALJ’s Ultimate Factual Determination That Prolonged Sitting At Work Was Not The Proximate Cause Of Shea’s Disability Is Supported By Substantial Evidence.

The causation question disputed here is whether Shea’s employment was a proximate cause of her disability. A proximate cause is a cause that is “so important in bringing about the injury that reasonable [persons] would regard it as a cause and attach responsibility to it.”33 The appropriate inquiry is “whether the conduct has been so significant and important a cause that the defendant should be legally responsible.”34 The ALJ determined that Shea’s employment was not so significant and important a cause that the State should be legally responsible.

Because reasonable minds could disagree whether Shea’s employment was a substantial factor in causing her disability, we review the ALJ’s determination for substantial evidence.35 We give deference to an ALJ’s decision if it is supported by evidence “substantial enough to support [the] conclusion in the contemplation of a reasonable mind.”36

Dr. Beard’s testimony supports the ALJ’s conclusion. It casts doubt on whether Shea suffered from ilioinguinal neuralgia during her State employment, breaking the causal chain between Shea’s prolonged sitting and the aggravation of a preexisting condition. The dissent suggests that Dr. Beard’s testimony should be discounted because Dr. Beard applied an incorrect legal standard when offering her opinion on whether Shea’s disability was work related.37 But as the ALJ noted, Dr. Beard’s mistake about the appropriate legal standard was relevant only to her “opinion as to causation.” The dissent offers no reason Dr. Beard's diagnosis of Shea’s condition and potential aggravating factors should likewise be dis*533counted or why it should not cast doubt on the premise of Dr. Smith’s opinion, which was that “ilioinguinal neuralgia is the primary problem.’’

Dr. Beard’s testimony substantially undercuts Dr. Smith’s, and, even if it did not, Dr. Smith’s opinion that Shea’s employment contributed to the increase in her pain by no more than 10% would not necessarily make it a substantial factor “if reasonable persons would [not] regard the injury as a cause of the disability and [not] attach responsibility to it.”38 Dr. Smith testified it was more likely than not that Shea’s work environment aggravated her condition by 5% to 10%. Dr. Smith himself characterized this percentage as “small.” Even if the ALJ relied exclusively on Dr. Smith’s testimony, he could reasonably conclude that a 5% to 10% increase in Shea’s symptoms resulting from work was not “so significant and important a cause as to attach legal responsibility to her employer for her disability.” ' Our Shea II decision leaves room for such a finding.39 And although the dissent certainly is correct that Shea II “permits the opposite conclusion as well,” the ALJ is the fact finder here, not this court.

That Dr. Smith’s testimony in itself could have been sufficient to support the ALJ’s proximate cause finding is also relevant to the dissent’s arguments about the ALJ’s allegedly “contradictory findings,” The dissent argues that the ALJ made “contradictory findings” because Dr. Smith’s testimony was sufficient to establish but-for cause yet insufficient to establish proximate cause. We conclude instead that the ALJ correctly required a higher showing to establish that prolonged sitting was a “substantial factor” in Shea’s disability rather than simply “one among a multitude of aggravating factors, no one of which stood out as of particular significance.” 40 “Substantial factor” necessarily requires a higher showing than but-for cause in this context: “but-for cause” requires that Shea’s work was a factor, in causing her disability; “legal cause” demands in addition that the factor be substantial. The ALJ could have relied exclusively on Dr. Smith’s testimony throughout and found without contradiction that it sufficed to establish but-for cause but not proximate cause.41

Shea’s failure until 2003 to report to a physician that prolonged sitting at work caused her disabling pain further supports the conclusion that a reasonable person would not attach legal responsibility to her employer. Her failure to report to a physician that work aggravated her pain indicates that her employment was not, in her mind, so significant a cause of her increased pain that she thought it particularly relevant to medical diagnosis and treatment.

The dissent asserts that the ALJ’s finding here provides no persuasive evidence concerning the legal causes of Shea’s disability because “Shea’s reports to her supervisor *534indicate ‘that she knew [sitting] was a factor in what caused her pain,’” and because “Shea had no need to report her work conditions to her doctor” as her supervisor was working with her to mitigate her pain. This argument misapprehends the finding’s relevance. It is undisputed that Shea’s pain manifested at work. But the relevant inquiry is not whether sitting for prolonged periods aggravated the contemporaneous expression of symptoms. Rather the relevant inquiry is whether prolonged sitting in some way precipitated or worsened the symptoms or the underlying disease process on a “presumably permanent[ ]” basis.42

The dissent acknowledges this distinction when it discusses how Dr. Smith “testified that prolonged sitting at work aggravated Shea’s pain in two ways: by increasing it during work, after which the pain would return to baseline, and by increasing her baseline level of pain.” We are concerned only with the latter. Shea discussing prolonged sitting at work with her supervisors is evidence of the former, but it does not address the relevant inquiry.43

And although for medical diagnosis and treatment purposes, as the dissent states, “prolonged sitting at work is not reasonably distinguishable from prolonged sitting in general,” in the ALJ’s opinion sitting also is not distinguishable in significance from other activities Shea reported to physicians, including “standing, walking, exercise, bending forward or backwards, cold, and stairs.” Shea failed to identify to physicians that work requirements were a likely contributing factor to her disability but addressed work requirements with her supervisors; these actions support the ALJ’s finding prior to remand that “prolonged sitting was one among a multitude of aggravating factors, no one of which stood out as of particular significance.” And this lack of distinction is relevant to determining whether work requirements should be considered the legal cause of Shea’s disability.

*535The dissent correctly notes that the ALJ’s observations about psychological factors were “tepid” and based on mere speculation because Shea did not undergo sufficient psychological evaluation. But Dr. Beard’s very firm opinion that “I do not believe [Shea] had chronic pelvic pain” was not mere speculation; Dr. Beard based her opinion on personal examination and experience treating patients with similar conditions. Dr. Beard’s opinion undercut the persuasiveness of Shea’s arguments. And because Shea bore the burden of proof, the State’s “theory of the case” did not have to include an alternative diagnosis as the dissent seems to suggest. Rather, existing plausible alternative bases for Shea’s pain support the ALJ’s determination that she failed to show work requirements were the proximate cause of her disability.

Other daily activities increased Shea’s pain and other factors, some perhaps psychological, may have played a role in her symptoms. This evidence supports a reasonable conclusion that Shea’s employment was not so significant or important a cause as to justify holding the State legally responsible. Because, in the contemplation of a reasonable mind, the evidence relied on is substantial enough to support the ALJ’s conclusion, we affirm the ALJ’s decision.

Y. CONCLUSION

We REVERSE the superior court’s decision reversing the ALJ’s decision, thereby AFFIRMING the ALJ’s decision sustaining the administrator’s denial of Shea’s occupational disability benefits.

dissenting.

I disagree with the court’s determination that substantial evidence in the record supports the denial of Shirley Shea’s occupational disability claim. Superior Court Judge Philip R. Volland correctly concluded that key facts in the administrative decision are either irrelevant to the ultimate question of proximate cause or simply not supported by evidence in the record. I would therefore affirm Judge Volland’s decision reversing the administrative denial of benefits.

The only question decided adversely to Shea by the administrative law judge was proximate cause, and the ALJ limited his adverse finding to one part of proximate cause, which he called the “ Attach Responsibility’ requirement.” The ALJ agreed with Shea that she had carried her burden of proving that, but for her work at the State, she “would not have experienced disabling chronic pain in 2001,” the year she became unable to work. But the ALJ rejected Shea’s claim that her work was a proximate cause based on the ALJ’s view that the group of people who would not “consider prolonged sitting at work so significant and important a catise as to attach legal responsibility for” the disability was “larger” than the group who would. The ALJ evidently counted himself among the former group and discussed the evidence he relied on to reject Shea’s claim. But not all of the evidence relied on by the ALJ was substantial; Some was speculative or nonexistent. And much evidence relied on by the court in today’s opinion either does not reflect the correct legal standard or is not related to the issues disputed in the case. This evidence cannot serve as substantial evidence to support the finding about proximate cause.

Legal framework

The court sets out the statutory definition of “occupational disability” and cites our case law holding that an employee can qualify for occupational disability benefits when an occupational hazard causes an increase in symptoms, even if it does not cause a worsening of the underlying condition itself. Under the statute and our case law on aggravation, Shea had to show that it was more likely true than not1 that “a hazard undergone, while in the performance and within the scope of [her] duties” was a proximate cause of a physical condition that “presumably permanently prevents [her] from satisfactorily per*536forming [her] usual duties”2 The statute uses the phrase “presumably permanently” only to modify the duration of an employee’s inability to perform her job functions, not to measure the impact on a condition or symptom.3 The State does not contest that Shea is permanently precluded from performing the usual duties of her job.

The “hazard” for. Shea was sitting, which she alleged aggravated her pain from the ilioinguinal neuralgia diagnosed by her doctors. In the first administrative decision, the ALJ wrote that “both parties agree that the primary cause of [Shea’s] disability is a neurological injury incurred in 1984 that became disabling during the time Ms. Shea was working for the State of Alaska.” The ALJ also wrote in that decision that “the preponderance of the evidence is that [Shea] suffered some form of injury to her ilioinguinal nerve in the course of the 1984 procedure, resulting in long-term unresolved ilioinguinal neuralgia.” That finding was not appealed, and in the State’s brief in the current appeal, it asserts that “[i]n Dr. Beard’s expert medical opinion, Shea’s pain was not work-related, but was caused by her 1984 ... procedure.” The ALJ referred to the neurological injury only as ilioinguinal neuralgia, and the ALJ also noted that the State “did not attempt to prove that Ms. Shea’s pain was a psychological response to the original non-work-related ilioinguinal neuralgia and as such was not proximately caused by her employment.” The ALJ’s words do not suggest that the parties “raised a dispute about the etiology of Shea’s current symptoms.” To the contrary, the ALJ’s decision demonstrates that the question of the etiology of Shea’s symptoms was not an issue on remand.

Neither party disputed that Shea was disabled from working. As the ALJ said in the first decision, the only dispute was “whether the prolonged periods of sitting were a substantial factor in her disability.”4 To show that prolonged sitting was a proximate cause of her disability, Shea only needed to show that prolonged sitting was a proximate cause of her increased pain.

We have previously observed that “increased pain or other symptoms can be as disabling as deterioration of the underlying disease itself,” and we have rejected a distinction between worsening of an underlying condition and worsening of the symptoms.5 Elaborating on the question of increased symptoms in the context of workers’ compensation, we recognized that “when a job worsens an employee’s symptoms such that she can no longer perform her job functions, that constitutes an ‘aggravation’ — even when the job does not actually worsen the underlying condition.”6 The question is not, as the court puts it in its opinion today, “whether prolonged sitting in some way precipitated or worsened the symptoms or the underlying disease process on a ‘presumably perma-nente ]’ basis.” Instead the question is whether the hazard Shea underwent at work worsened her symptoms to such an extent that she was permanently precluded from performing her job duties.7 She could demonstrate this not only by showing that sitting at work permanently caused an increase in her baseline pain, but also by showing that sitting at work increased her pain to such an extent that she could not perform her work and that this circumstance was permanent— that it would not improve.8 Shea presented evidence supporting both propositions.

*537Administrative decision on appeal in this case

The ALJ decided on remand that Shea had proved but-for causation but had not shown that sitting at work was so significant a factor in her disability that reasonable persons would consider it a cause and attach responsibility to it. In reaching that decision, the ALJ acknowledged that Shea had presented substantial evidence that her work was a substantial factor in causing her disability, and he specifically identified the following evidence to support this statement: Dr. Smith’s opinion “that prolonged- sitting aggravated a physical condition and her pain symptoms,” coupled with the increase in Shea’s pain symptoms while she worked for the State and the evidence that Shea “sat for prolonged periods of time while working a,nd found it painful.”

The ALJ then listed the six pieces of “substantial evidence” that countered Shea’s proof; that list is set out in the court’s opinion. The ALJ discussed separately the evidence that persuaded him Shea had not shown that work was so significant a cause in her disability that he would attach legal responsibility to it. The ALJ’s discussion focused on the manner in which Dr. Beard’s opinion differed from Dr. Smith’s and the timing of Shea’s reports regarding her pain, although he mentioned other factors as well.

The ALJ examined testimony of the two doctors whose opinions he had given the most weight, Dr. Smith and Dr. Beard. The ALJ acknowledged that Dr. Beard’s opinion about aggravation was based on an incorrect legal standard and “did not directly rebut Dr. Smith’s opinion as to aggravation of chronic pain,”9 The ALJ stated that Dr. Beard’s “testimony, in substance, was to the effect that Ms. Shea’s ilioinguinal neuralgia had resolved.” The ALJ reasoned that this aspect of Dr. Beard’s opinion “weaken[ed] the basis for [Dr. Smith’s] opinion, which was that ... [the] [ ]ilioinguinal neuralgia[] had not resolved.” (Emphasis in original.) Notably, the ALJ did not infer from Dr. Beard’s testimony that something other than the ilioinguinal neuralgia was the underlying cause of Shea’s chronic pain. Citing Smith v. University of Alaska, Fairbanks,10 the ALJ commented that Dr, Beard’s testimony “has some persuasive weight even in the absence of a definitive statement applying the correct legal standard.” But nothing in the ALJ’s discussion even hints at a finding “that Dr. Beard’s testimony raised a dispute concerning the etiology of Shea’s current symptoms.” Yet, the court insists that the ALJ found “that Dr. Beard’s testimony raised a dispute concerning the etiology of Shea’s current symptoms” and “found there was dispute between the doctors concerning the underlying cause of Shea’s disability,” but the administrative decisions simply do not support this assertion. In discussing the “ ‘Attach Responsibility requirement,” the ALJ wrote;

In this case, Dr. Smith was of the opinion that prolonged sitting at work aggravated Ms. Shea’s chronic pain. Dr, Beard’s opinion was couched in terms of the underlying physical condition, and thus did not directly rebut Dr. Smith’s opinion. Nonetheless, Dr. Beard’s testimony provides support for the view that prolonged sitting at work was not a substantial factor in Ms. Shea’s disability, because her testimony, in substance, was to the effect that Ms. Shea’s ilioinguinal neuralgia had resolved. Thus, while Dr., Beard did not directly rebut Dr. Smith’s opinion as to aggravation of chronic pain, her testimony weakens the basis for his opinion, which was that Ms. Shea’s underlying physical condition (ilioin-guinal neuralgia) had not resolved. In light of Dr. Beard’s testimony, Dr. Smith’s opinion that Ms, Shea’s chronic pain was work-related is only marginally peesuasive.[11]

The ALJ did not attempt to reconcile his interpretation of Dr. Beard’s testimony with his statement in the initial decision that “the *538preponderance of the evidence is that [Shea] suffered some form of injury to her ilioingui-nal nerve in the course of the 1984 procedure, resulting in long-term unresolved ilioin-guinal neuralgia.”

The ALJ also discussed at length his reasoning related to Shea’s reports about the impact of sitting at work. A cracial aspect of the ALJ’s analysis was his concern that Shea had not reported to her physicians that prolonged sitting at work was causing her pain to increase. According to the ALJ, “[a] claimant’s reported perception of pain and the activities that generate it, made for the purpose of assisting in the diagnosis and treatment of that pain, is important evidence of the cause of that pain.” The ALJ then wrote that “reasonable persons will consider whether the claimant reported a work injury as a source of pain during the period of employment, and, if not, whether there is a reasonable explanation for the failure to do so.” (Emphasis added.) The ALJ then found the following (as later revised):

In this case, Ms. Shea did not identify sitting as a causal factor until 1999, and she did not identify sitting at work as a causal factor until two years after she retired, in 2003.... [T]he alleged relationship between her working conditions and the disability is one that Ms. Shea might reasonably be expected to have identified through direct personal experience, and to tell her treating physician: “It hurts when I sit for a prolonged period of time” Under these circumstances, that Ms, Shea did not report to a physician prolonged sitting at work as contributing to her chronic pain while she was still employed suggests that prolonged sitting at work was not so significant and important a cause as to attach legal responsibility to her employer.[12]

The ALJ cited the fact that Shea had “not reported] to a physician working conditions as contributing to her pain until after she had stopped working” and “evidence that the ilioinguinal neuralgia had resolved and that psychological factors may have contributed to her disability” as supporting his conclusion that Shea had not shown that reasonable persons would attach legal responsibility to her employer for her disability. Because the ALJ focused on these two reasons to reject Shea’s claim, and because the ALJ acknowledged that reasonable people could also conclude that Shea’s work was a legal cause of her disability, the factual underpinnings of his main reasons must be examined in detail.

Reporting work-related sitting to a physician

With regard to Shea’s report of sitting as a cause of her pain, the record shows that Shea did just what the ALJ said she was “reasonably expected to do.” She reported to her treating physicians, beginning in 1999, that sitting was causing her increased pain,13 The ALJ notably failed to discuss “the reasonable explanation” for her failure to report employment-specific prolonged sitting to her doctors: She had discussed the problem with her supervisor at work, and the employer was taking steps to address the problem. From the point of view of medical diagnosis and treatment, prolonged sitting at work is not reasonably distinguishable from prolonged sitting in general except that the worker may have less control over varying her activity or work conditions, such as the type of chair she uses. If an employer is attempting to address those conditions, as Shea’s supervisor was, a physician does not really need to know for purposes of diagnosis and treatment whether the prolonged sitting is happening at work or elsewhere. Because Shea’s supervisor was working with her to address her increased pain at work, Shea had no need to report her work conditions to her doctor, particularly in light of her earlier report to her physicians that sitting increased her pain. As the superi- or court observed, Shea’s reports to her supervisor indicate “that she knew [sitting] was *539a factor in what caused her pain.” Thus, I agree with the superior court that Shea’s failure to report to a physician that prolonged sitting at work was causing increased pain was “not particularly convincing evidence” that could serve to undermine the evidence supporting Shea.14

Dr. Beard’s testimony

The ALJ cited Dr. Beard’s testimony “to the effect that Ms. Shea’s ilioinguinal neuralgia had resolved” to support his ultimate conclusion and to counter Dr. Smith’s opinion that prolonged sitting at work probably did increase Shea’s pain.15 The superior court carefully examined the portion of Dr. Beard’s testimony that the ALJ relied on and concluded, correctly in my view, that this testimony did not support a finding that the ilioinguinal neuralgia had resolved. Dr. Beard testified that Shea “had just recently been evaluated ... and diagnosed with that condition” at the time of Dr. Beard’s examination. Dr. Beard then testified that Shea’s disability is “really chronic pain,” where a patient stops seeking treatment for the underlying condition and instead seeks treatment for the pain. This testimony does not support a finding that the ilioinguinal neuralgia had resolved. And, as the ALJ pointed out, the State “did not attempt to prove that Ms. Shea’s pain was a psychological response to the original nonwork-related ilioinguinal neuralgia and as such was not proximately caused by her employment.”

The superior court considered other parts of Dr. Beard’s testimony, to the effect that the duration of Shea’s symptoms “would not be typical” if the neuralgia was caused by the earlier surgery because ilioinguinal neuralgia eases usually “resolve relatively quickly.” But, as the superior court correctly determined, this testimony does not show that Shea’s neuralgia had resolved: Rather, “it shows either that (1) Shea’s condition was atypical, or (2) something else was causing her pain.” Because the ALJ incorrectly inferred that Shea’s neuralgia had resolved in order to discredit Dr. Smith’s opinion about causation, the superior court correctly discounted this part of the ALJ’s decision.

The court takes the position that the ALJ made the latter inference and “did find that Dr. Beard’s testimony raised a dispute concerning the etiology of Shea’s current symptoms and that her pain could be caused by something other than ilioinguinal neuralgia substantially aggravated by work requirements.” But nothing in the ALJ’s opinion supports the court’s assertion that “the etiology of Shea’s current symptoms” was in question; this assertion is flatly contradicted by the ALJ’s earlier finding, based on the preponderance of the evidence, that Shea’s 1984 procedure “result[ed] in long-term unresolved ilioinguinal neuralgia.” The relevant material and contested issue was not the etiology of Shea’s symptoms; if that issue was contested, the ALJ resolved it in the first decision. The contested question was narrower: By the time the ALJ discussed this portion of Dr. Beard’s testimony, the only unresolved question was whether sitting at work was so important a cause in Shea’s pain that reasonable persons would regard it as a cause and attach responsibility to it. The ALJ had already determined that sitting was a but-for cause — that “but for periods of prolonged sitting at work, Ms. Shea would not have been disabled in 2001.”

Rather than focusing on this narrow question, the court claims that the ALJ “did find that Dr. Beard’s testimony raised a dispute *540about the etiology of Shea’s current symptoms.” But as set out above, the ALJ described Dr. Beard’s testimony as testimony that the “ilioinguinal neuralgia had resolved.” The court points to no evidence or testimony to show that the inference the ALJ actually made is supported by substantial evidence, and as a result, this “evidence” cannot be a basis for the court or the ALJ to discount Dr. Smith’s opinion. And the ALJ said Dr. Beard’s testimony undercut Dr. Smith’s because “her testimony weakens the basis for his opinion, which was that Ms. Shea’s underlying physical condition (ilioinguinal neuralgia) had not resolved,” (emphasis in original) not because the testimony pointed to an alternative cause for the pain.

Furthermore, if Dr. Beard’s testimony posited another medical condition that might have been the underlying cause of Shea’s pain and inability to work, that testimony was not relevant. According to the State “[i]n Dr.- Beard’s expert medical opinion, Shea’s pain was not work-related, but was caused by her 1984 ... procedure.” And in the context of his discussion of but-for causation, the ALJ wrote that the State “did not establish an alternative explanation, supported by an expert medical opinion, as to why [Shea] began to experience disabling pain during her employment.” He acknowledged “evidence that psychological factors contributed to Ms. Shea’s disability” but also wrote that the State “did not attempt to prove that Ms. Shea’s pain was a psychological response to the original non-work-related ilioinguinal neuralgia and as such was not proximately caused by her employment.” Because the factual issue regarding the etiology of Shea’s pain was either uncontested or had been resolved in Shea’s favor, if Dr. Beard’s testimony supported an inference that something other than .ilioinguinal neuralgia was causing the pain, that factual inference was not legally relevant.16 Because substantial evidence must be relevant evidence,17 the inference the court today attributes to the ALJ — that “Dr. Beard’s testimony raised a dispute concerning the etiology of Shea’s current symptoms” — is not substantial evidence because it did not make more or less likely the existence of a material fact. In discussing legal cause, the ALJ made no inferences as to other possible causes of Shea’s chronic pain because the possibility of other possible causes was not contested.

The other part of Dr. Beard’s testimony that the ALJ relied on was her opinion that psychological factors may have contributed to Shea’s disability. While the superior court agreed that substantial evidence supported the ALJ’s “tepid” finding that psychological factors “may have contributed to Shea’s disability” (emphasis in original), the superior court also highlighted Dr. Beard’s testimony that there was no psychological evaluation of Shea, making impossible a determination of the degree to which psychological factors were involved.18 I agree with the superior court’s conclusion that this finding has no bearing on the ultimate question of causation because there was no evidence beyond the speculative possibility that psychological factors were at play. We have held in a variety of contexts that speculation is not substantial evidence.19 The State agreed, and continues *541to agree, that the underlying basis of Shea’s pain was a neurological injury in 1984, which the ALJ repeatedly characterized as ilioin-guinal neuralgia. The only contested issue was causation; if the State’s theory of the case was that something other than pain related to the ilioinguinal neuralgia was Shea’s underlying medical condition, that theory of the case was not apparent to the ALJ. And it cannot be an alternative theory the court can use to affirm the ALJ’s decision because it would require additional fact finding.20 While the court contends that “the State’s ‘theory of the ease’ did not have to include an alternative diagnosis” in order for the ALJ to find that Shea had not met her burden of proof, the court cannot manufacture a dispute about the evidence and make alternative findings that the fact-finder did not make in order to affirm.21 Nothing in the ALJ’s decisions supports the court’s statements that a dispute regarding the existence of ilioinguinal neuralgia was the reason the ALJ found against Shea on the question of legal cause.

In sum, Dr. Beard’s testimony did not contain the “evidence” the ALJ considered so substantial as to undercut what the ALJ conceded was adequate proof of causation by Shea. And her speculation that psychological factors were a factor is not substantial evidence.

Other factors

The other two factors identified by the ALJ as important to his decision about legal responsibility were (1) Dr. Smith’s testimony that “the aggravation of [Shea’s] chronic pain symptoms due to prolonged sitting was limited to 5-10%” and (2) evidence that “ordinary daily activities also aggravated her chronic pain.” But evidence about the impact of ordinary daily activities is not persuasive evidence to undermine Shea’s claim, particularly in light of the ALJ’s finding that her daily activities did not change during the time she became increasingly disabled and, eventually, unable to perform her work. I agree with the superior court that these two findings are contradictory.

The amount of aggravation attributed by Dr. Smith to an “over time” worsening of Shea’s chronic pain was five to ten percent; he indicted that Shea could experience increased pain while sitting, after which her pain would return to baseline. Quoting part of the statute out of context, the court contends that “the relevant inquiry is whether prolonged sitting in some way precipitated or worsened the symptoms or the underlying disease process on a ‘presumably perma-nente ]’ basis.” But that is not what the statute says: AS 39.35.680(27) refers to “a physical or mental condition, that in the judgment of the administrator, presumably permanently prevente an employee from satisfactorily performing the employee’s usual duties.” An employee’s inability to perform her work is what must be permanently precluded. Requiring a permanent worsening of baseline pain is no different from requiring a permanent worsening of an employee’s condition. The court’s misstatement of the statute is essentially a reformulation of a distinction we rejected.22 And Dr, Smith’s testimony showed that sitting at work both increased Shea’s pain while she was working and furthered her permanent disabling pain by increasing her baseline level of pain over time. Dr. Smith also testified that there was no cure for Shea’s pain. In spite of efforts by Shea’s employer to ameliorate the working *542conditions that increased her pain, by the time Shea’s employment with the State ended, her baseline pain had increased to the point where that pain made completing her work impossible.

The court also argues that Dr. Smith’s opinion, standing alone, was sufficient to defeat Shea’s claim because he testified that prolonged sitting would permanently increase Shea’s pain by five to ten percent, and Shea II “leaves room for such a finding.” Shea II permits the opposite conclusion as well,23 and the ALJ did not rely exclusively on Dr. Smith’s opinion to find against Shea, instead listing it as one of many pieces of evidence he weighed. Indeed, the reason the ALJ gave for considering factors such as Shea’s reports to her physicians was “the close balance in the expert medical testimony.” In other words, the ALJ did not find Dr. Smith’s opinion persuasive on the point as the court suggests.

A detailed examination of the reasons given by the ALJ for rejecting Shea’s claim demonstrates that substantial evidence in the record does not support his conclusions.

Causation overall

In his summary about the overall question of causation, the ALJ wrote that Dr. Smith’s opinion alone was insufficient to “establish that prolonged sitting was a substantial factor in [Shea’s] disability.” But the ALJ concluded that Dr. Smith’s opinion in combination with two other factors was “substantial evidence that prolonged sitting at work was a substantial factor in her disability.” The ALJ also listed evidence that, in his view, undercut Shea’s case, including evidence that Shea did not “identify sitting as a causal factor until February, 1999, after her symptoms had become highly problematic.” But this evidence is consistent with Dr. Smith’s opinion that “over time” prolonged sitting would cause Shea’s baseline pain to increase by five to ten percent. Shea’s job with the State, which required prolonged sitting, began in 1993, and it was following the six year’s from 1993 to 1999 that Shea identified prolonged sitting as causing increased pain.

The ALJ listed “Dr. Beard’s expert medical opinion that prolonged sitting did not permanently aggravate the underlying physical condition” as “substantial evidence that prolonged sitting at work was not a substantial factor in Ms. Shea’s disability,” even though the ALJ later acknowledged that this opinion appeared not to apply the correct legal standard. Relying on our opinion in Smith v. University of Alaska, Fairbanks,24 the ALJ wrote that Dr. Beard’s “testimony has some persuasive weight even in the absence of a definitive statement applying the correct legal standard.” But Smith has no relationship to Dr. Beard’s opinion about aggravation. There is no indication that the doctor’s testimony in Smith used an incorrect legal standard; rather, the testimony did not, in the Board’s view, use an adequately specific probability formula.25 In contrast, Dr. Beard gave a definite statement using an incorrect legal definition of aggravation, testifying that she meant “to cause a permanent worsening or flare of [the] condition,” which she identified as ilioinguinal neuralgia. Smith does not license reliance on an expert opinion that has the wrong legal standard as its basis. And if the ALJ intended to refer to the quotation in Smith from Larson’s treatise— to the effect that a fact finder should consider “the real substance of what the witness intended to convey” rather than engage in “a game of ‘say the magic word’ ”26 — that quote does not endorse use of an incorrect legal standard. Moreover, when the ALJ discussed the substance of Dr. Beard’s opinion, he focused on the inference that Shea’s ilioingui-nal neuralgia had resolved. As set out earlier, that inference is not supported by the record.

*543In his final analysis, Judge Volland correctly concluded that the ALJ made contradictory findings in accepting Dr. Smith’s opinion to find “but for” cause but discrediting the same opinion when analyzing proximate cause. The ALJ found that, after Shea began to work for the State, her “level of pain was substantially greater, even though her everyday activities did not change (except in response to pain).” This fact, in conjunction with Dr. Smith’s expert opinion that it was at least 51% likely that prolonged sitting aggravated Shea’s injury, led the ALJ to find that Shea had established that, “but for” her State employment, her condition would not have worsened. As the superior court wrote, “If Shea’s daily activities remained unchanged and sitting was a 51% cause, it cannot be said that the daily activities defeat proximate cause.” The ALJ credited Dr. Smith’s testimony to find “but for” cause, and I agree with the superior court that “[t]he ALJ cannot find on one hand that Dr. Smith’s testimony was substantial evidence establishing actual cause but not substantial evidence supporting the causal component of proximate cause.”

The evidence relied on by the ALJ to deny Shea’s claim was either incomplete or irrelevant to the question of proximate cause. And the court’s current theoiy of the case is significantly different from the decision on review or the State’s litigation position and does not apply the legal standard we have previously adopted for aggravation claims. I therefore respectfully dissent from the court’s opinion and would affirm Judge Vol-land’s decision reversing the ALJ’s decision.