[¶ 1] The employee, Barbara Bossie, appeals from a decision of the Workers’ Compensation Board granting her petition for award. Concluding that neither of the first three methods for computing the average weekly wage in 39-A M.R.S.A. § 102(4)(A), (B) & (C) (Supp.1997), could be fairly applied, the Board purported to apply subsection 102(4)(D) by dividing her yearly earnings by 52. Bossie contends that, because there was no evidence of the earnings of similarly situated employees pursuant to subsection D, the Board should have applied subsection B to determine her average weekly wage and divided her total earnings by the number of weeks she worked in the preceding year. St. Pierre v. St Regis Paper Co., 386 A.2d 714, 719-20 (Me.1978). We agree and vacate the decision of the Board.
*579I.
[¶2] Bossie was employed at School Administrative District No. 24 (S.A.D.# 24) as a cook for twenty-four years. She worked during the school year for 36 weeks from August to June. The parties stipulated that, at the time of her injury, Bossie earned an hourly wage of $9.75 for a regular 30 hour week, with total earnings, including holiday pay, of $10,530 a year. The parties also agreed that an additional $188.19 a month, or $2,258.28 a year, in employer contributions to Bossie’s health insurance should be included in her average weekly wage. Bossie left work due to a gradual work-related injury in January 1993. The Board granted her petition for award in 1996 and awarded short-term total and continuing partial incapacity benefits pursuant to 39-A M.R.S.A. § 102(4)(D) by dividing Bossie’s total yearly earnings by 52 to arrive at an average weekly wage of $246.26. We granted Bossie’s petition for appellate review pursuant to 39-A M.R.S.A. § 322 (Supp.1997).
II.
[¶ 3] The selection of the proper method for determining the “avérage weekly wage” proceeds sequentially through four alternative provisions outlined in 39-A M.R.S.A. § 102(4)(A), (B), (C) & (D) (Supp. 1997). Frank v. Manpower Temp. Servs., 687 A.2d 623, 625 (Me.1996). By its terms, the first method, subsection A, does not apply because Bossie had not been employed for at least 200 full working days during the immediately preceding year prior to her injury. 39-A M.R.S.A § 102(4)(A). The second method, applicable to employments that do not continue for more than 200 days in the immediately preceding year, is therefore applicable by its terms:
B. When the employment or occupation did not continue pursuant to paragraph A for 200 full working days, “average weekly wages, earnings or salary” is determined by dividing the entire amount of wages or salary earned by the injured employee during the immediately preceding year by the total number of weeks, any part of which the employee worked during the same period. The week in which employment began, if it began during the year immediately preceding the injury, and the week in which the injury occurred, together with the amounts earned in those weeks, may not be considered in computations under this paragraph if their inclusion would reduce the average weekly wages, earnings or salary.
39-A M.R.S.A. § 102(4)(B).1 Although Bos-sie bore the burden of proof to establish the average weekly wage on her petition for award, Frank, 687 A.2d at 625, she met that burden by presenting sufficient evidence to support a calculation pursuant to subsection B.
[¶ 4] S.A.D. #24 contends, however, that because Bossie ordinarily worked 36 weeks out of the year, application of subsection B would lead to an inflated average weekly wage, and therefore it was appropriate for the Board to apply the so-called “fail-back” provision, subsection D. Subsection D provides:
D. When the methods set out in paragraph A, B or C of arriving at the average weekly wages, earnings or salary of the injured employee can not reasonably and fairly be applied, “average weekly wages” means the sum, having regard to the previous wages, earnings or salary of the injured employee and of other employees of the same or most similar class working in the same or most similar employment in the same or a neighboring locality, that reasonably represents the weekly earning capacity of the injured employee in the employment in which the employee at the time of the injury was working.
39-A M.R.S.A. § 102(4)(D).
[¶ 5] S.A.D. #24 argues that although subsection B is applicable on its face, subsection D would be the best method for calculating the average weekly wage of an employee, like Bossie, with a long-term history of employment for substantially less than the normal full working year. As Professor Larson states in his treatise, the average weekly *580wage determination is not based solely on what that employee is theoretically capable of éarning, but on the employee’s actual work-history, e.g., the employee’s willingness to work fiill-time and the availability of full-time employment in the competitive labor market. 2 A. Larson, The Law of Workmen’s Compensation, §§ 60.21(c), 60.22(a) (1993). Professor Larson is critical of jurisdictions that determine the earnings of long term part-time employees based on what those employees might earn in hypothetical full-time employment:
The flaw in this reasoning is that the purpose of the wage calculation is not to arrive at some theoretical concept of loss of earning capacity; rather it is to make a realistic judgment on what the claimant’s future loss is in the light of all the factors that are known. One of these factors is the established fact of claimant’s choice of a part-time relation to the labor market. If this is clear, and above all there is no reason to suppose it will change in the future period into which the disability extends, then it is unrealistic to turn a part-time able-bodied worker into a full-time disabled worker.
Id. at § 60.21(c). In discussing the anala-gous situation of a school teacher, Professor Larson states:
If a school teacher, for example, is paid $2,000 a month for nine months of the year, there is no reason to calculate earning capacity on the unrealistic basis of $2,000 a month for twelve months; if the wage statute says that monthly wages shall be multiplied by twelve, this (and any comparable situation) is an appropriate occasion for application of the residual clause [our subsection D], which exists for the express purpose of taking care of just such nonstandard wage relations.
Id. at § 60.22(a).
[¶ 6] Although we agree with S.A.D. #24 that subsection D might have been the best method of determining the average weekly wage in this case, as the party asserting the application of subsection D, S.A.D. # 24 bore the burden of providing evidence to support a determination pursuant to that subsection. See Fernald v. Dexter Shoe Co., 670 A.2d 1382, 1385 (Me.1996) (employer asserting that employee’s earnings varied from week to week bears burden of proof on employee’s petition to show variability). Subsection D does not require rigid adherence to an exact mathematical formula. Roberts v. Smith, 415 A.2d 1089, 1090 (Me.1980). Nevertheless, as Bossie contends, method D expressly requires a consideration of the earnings of “comparable employees.” Harrigan v. Maine Veterans Home, 1997 ME 224, ¶ 7, 704 A.2d 1215, n. 4; St. Pierre, 386 A.2d at 719-20. Indeed, we held in St. Pierre that the “comparable employee” requirement of former 39 M.R.S.A. § 2(2)(C) (1989), repealed by the Maine Workers’ Compensation Act of 1992, P.L.1991, ch. 885, § A-7, the predecessor statute to subsection D, could not be satisfied by the consideration of the earnings of only one “comparable employee.” 386 A.2d at 719-20. We stated, in pertinent part, that the fall-back provision
plainly mandates that consideration be given to the earnings of similarly situated employees, in the plural. The picking of a single ‘comparable employee’ and the averaging of that single employee’s wages during the preceding year does not comply with the statute’s requirement. Except by the rarest of coincidences, choice of even one ‘comparable employee’ in addition to [the one chosen] would have changed, either up or down, the resulting, arithmetic average. Even that, however, would not satisfy the statute. While the technique of designating a single ‘comparable’ worker has the advantage of administrative convenience to recommend it, that method of computation is unduly restrictive and, as the facts of the present case illustrate, may well disadvantage the employee by ascribing to him lower ‘average weekly wages’ than he is by law entitled to. Manifestly, both St. Regis and the commission erred in failing to give due regard to all the factors enumerated in paragraph C_ Whatever greater burden is imposed upon the parties and the commission by our reading of paragraph C to require more than merely averaging the wages of a single ‘comparable employee’ is required, we believe, by the clear intendment of the statute. The legislature alone *581has the authority to authorize an administrative shortcut or any other rule of convenience that those administering' the Workmen’s Compensation Act may devise.
Id. (citations omitted) (emphasis added). In this case, S.A.D. #24 did not present evidence of the earnings of comparable employees and, therefore, did not provide a basis for the application of subsection D.
[¶ 7] S.A.D. # 24 contends that this case is distinguishable from St. Pierre because, in that case, the employee had only worked for the employer a short time prior to his injury. 386 A.2d at 716. By contrast, Bossie’s 24 year work history with her employer provided ample evidence of her earning capacity, and, therefore, the Board was permitted to forego an examination of the earnings of comparable employees pursuant to subsection D. We disagree. The Legislature has carefully outlined four alternative methods to best arrive at a fair estimate of the employee’s earning capacity. As we have stated, “these methods are not to be applied in the alternative as a matter of choice, but are to be applied in the order stated, to the facts as they exist in the particular ease, upon the principle of resorting to the best evidence obtainable in determining the employee’s average wage.” Thibeault’s Case, 119 Me. 336, 338, 111 A. 491, 492 (1920). The Legislature has reasonably concluded that in cases when the methods of subsections A, B, or C cannot be fairly applied, the next most reliable approach to arriving at a fair estimate of the employee’s earning capacity must include a consideration of the earnings of comparable employees. We therefore do not interpret the express language requiring consideration of comparable employees as mere surplus-age. St. Pierre, 386 A.2d at 719-20. Moreover, as Bossie contends, the Board has, in effect, applied the methodology of subsection C, the seasonal employment provision, to an employment that regularly continues for more than 26 weeks in a calendar year and, therefore, does not fall within the terms of that provision. 39-A M.R.S.A. § 102(4)(C)(1). Such an interpretation could potentially lead to the application of the seasonal employment provision to all employees regularly employed for more than 26, but fewer than 52, weeks in a calendar year, and would therefore defeat the purpose of the 26 week limitation in subsection (C)(1). We see no support for this approach in the statutory scheme.
[¶ 8] S.A.D. # 24 next contends that, while former 39 M.R.S.A. § 2(2)(C), at issue in St. Pierre, may have required consideration of comparable employees in all cases, such a result is not required pursuant to our current subsection D. Former subsection C is remarkably similar to subsection D:
C. In cases where the foregoing methods of arriving at the “average weekly wages, earnings or salary” of the injured employee cannot reasonably and fairly be applied, said “average weekly wages” shall be taken at such sum as, having regard to the previous wages, earnings or salary of the injured employee and of other employees of the same or most similar class, working in the same or most similar employment in the same or a neighboring locality, shall reasonably represent the weekly earning capacity of the injured employee at the time of the injury in the employment in which he was working at such time.
39 M.R.S.A. § 2(2)(C) (1989) (emphasis added), repealed by the Maine Workers’ Compensation Act of 1992, P.L.1991, ch. 885, § A-7. Current subsection 102(4)(D), by contrast, provides that “ ‘average weekly wages’ means the sum, having regard to the previous wages, earnings or salary of the injured employee and of other employees of the same or most similar class working in the same or most similar employment in the same or a neighboring locality_” 39-A M.R.S.A. § 102(4)(D) (emphasis added). S.A.D. #24 contends that the replacement of the language “shall be taken” in former subsection C, with the word “means” in current subsection D, suggests that the consideration of comparable employees is now a matter of Board discretion. We disagree. As we have stated, “‘when statutory language has acquired a consistent and entrenched meaning through prior judicial decisions, we will not abandon our traditional interpretation of that language unless there is express, statutory language plainly showing a legislative intent to abrogate those prior decisions.’ ” Bureau *582v. Staffing Network, Inc., 678 A.2d 583, 588 (Me.1996) (quoting Tripp v. Philips Elmet Corp., 676 A.2d 927, 930-81 (Me.1996)). The use of the word “means” in the context of the current provision is no less mandatory than the phrase “shall be taken” in former subsection (2)(C).
The entry is:
The decision of the Workers’ Compensation Board vacated. Remanded to the Workers’ Compensation Board for further proceedings consistent with the decision herein.
. Both parties concede that because Bossie’s employment continued longer than 26 weeks, she is not a seasonal employee pursuant to 39-A M.R.S.A. § 102(4)(C)(1) (Supp.1997).