delivered the opinion of the court:
The issue in this case involves the calculation of claimant’s average weekly wage rate. Claimant was injured after working a total of four days over a three-week period and was awarded benefits by the arbitrator. On review, the Industrial Commission (Commission) modified the average weekly wage rate set by the arbitrator. The Commission inferred that claimant had worked eight hours each of the four days he was employed by respondent, which it then multiplied by his union scale wage and divided by 15 working days, the three-week period during which he was partially employed, to arrive at an average *810weekly rate of $158.40. The circuit court reversed. It calculated the wage at $594 by multiplying claimant’s union scale hourly wage by eight hours a day for five days. Respondent appeals, contending the circuit court erred in determining that the Commission’s calculation was against the manifest weight of the evidence. We agree and reverse.
The evidence presented by claimant was less than clear or comprehensive. He testified that he had been an iron worker for 14 years and was employed sporadically by 300 to 400 different companies. January 29, 1985, the date of the accident, was the fourth day, scattered over a three-week period during January 1985, that he had worked for respondent. His hourly wage was $14.85. At the time of the injury, he expected that the job would continue for another four to six weeks.
Claimant’s testimony concerning his prior employment was, to say the least, equivocal and unsubstantiated. Claimant offered that he worked at least 96 hours during 1984, although he believed he had worked more. He was not sure how much money he made during 1984, but claimed he had worked at other odd jobs in addition to those he secured as an iron worker out of the union hall. Claimant could not remember, however, for whom or when or how long he worked on any other job. Claimant was shown a W-2 form and a written communication from his prior attorney to counsel for respondent, neither of which was offered into evidence, suggesting that his actual wages for 1984 were somewhere in the $1,400 to $1,800 range. Claimant denied that this constituted his total earnings but presented no evidence to the contrary.
The arbitrator set an average weekly wage which both parties agreed had no evidentiary basis. On review, the Commission modified the average weekly wage to $158.40 after concluding that claimant proved only that he was employed for four days over a three-week period. After assuming claimant worked eight hours each of those four days, the Commission multiplied claimant’s union-scale wage by 32 hours, equalling total wages of $475.20, which it then divided by 3 weeks, or 15 days.
As we have indicated, the circuit court reversed the Commission’s decision, finding the average weekly wage was $594, an amount arrived at by multiplying the hourly rate by eight hours for five days.
Claimant bears the burden of establishing his average weekly wage under section 10 of the Workers’ Compensation Act (Act) (Ill. Rev. Stat. 1991, ch. 48, par. 138.10). (Cook v. Industrial Comm’n (1992), 231 Ill. App. 3d 729, 731, 596 N.E.2d 746, 748.) In this case, *811claimant candidly concedes that the only credible proof of his employment upon which the Commission could reasonably base its determination was the evidence concerning his limited employment at the time he was injured.
In Cook, this court recently rejected the method employed by the trial court in calculating the average weekly wage. In doing so, this court offered a hypothetical example, which is virtually identical to the underlying facts of this case, concluding that a result such as that reached by the trial court here could not be sustained because of the gross windfall to the employee.
Section 10 of the Act provides four methods for calculating average weekly wage. The Act does not, however, define the term “workweek,” or what it might reasonably constitute in any given type of employment. Because of this, it is the employee’s burden to establish what a workweek consists of in the context of the actual employment history of the claimant. This is a question of proof. There is no presumption that a workweek for every employee in the labor force is 40 hours. To accept the method used by the trial court would, in effect, shift the burden to the employer to disprove the applicability of a 40-hour week to any particular claimant’s situation. Such a result is not expressed in the statute and would be contrary to the well-established proposition that the claimant bears the burden of proving each element of his case.
Claimant argues, nevertheless, that the possibility of a windfall to the employee is no longer a relevant concern and that supreme court precedent such as Hasler v. Industrial Comm’n (1983), 97 Ill. 2d 46, 454 N.E.2d 307, is no longer viable, given legislative amendments to section 10 which have occurred since the decision in Hasler and this court’s decisions in cases such as Peoria Roofing & Sheet Metal Co. v. Industrial Comm’n (1989), 181 Ill. App. 3d 616, 537 N.E.2d 381, and Illinois-Iowa Blacktop, Inc. v. Industrial Comm’n (1989), 180 Ill. App. 3d 885, 536 N.E.2d 1008. We disagree.
As we noted in Cook, concern over a potential windfall to the employee remains a viable, although not determinative, consideration in calculating the average weekly wage rate under the current version of the statute. We stress, however, that the desirability or effect of any particular result is, of course, secondary to the primary question of whether the proof supports the calculation in the first instance.
In this case, the Commission apparently calculated the wage under the provision of section 10 of the Act which provides: “Where the employment prior to the injury extended over a period of less than 52 weeks, the method of dividing the earnings during that period by the *812number of weeks and parts thereof during which the employee actually earned wages ***.” (Ill. Rev. Stat. 1991, ch. 48, par. 138.10.) Claimant concedes that there is no reliable evidence concerning his employment history other than that he worked for four days over a three-week period in January 1985. Based on claimant’s failure to provide other credible evidence concerning prior employment, the Commission’s determination of his average weekly wage rate comports with the statutory formula.
Claimant’s argument that the Commission misinterpreted the phrase “number of weeks or parts thereof” in section 10 of the Act is without merit. Claimant’s theory is that “parts thereof” may be read in isolation. If this were true then the determination of the average weekly rate would, in virtually every case, be literally based upon the number of hours a less-than-‘ ‘full-time’ ’ employee worked multiplied upward to fill out a complete 40-hour week — even if the employee never worked a 40-hour week during his entire life. Moreover, if the legislature intended such a result, there would be no need whatsoever for the alternative methods of calculating the wage provided by section 10 of the Act.
Under the evidence presented in this case, the Commission’s determination is consistent with the statute. When the employment is noncontinuous or less than “full-time,” earnings may be divided by an entire workweek even if the employee worked only a portion of the week. (See Illinois-Iowa Blacktop, 180 Ill. App. 3d at 894, 536 N.E.2d at 1015 (McCullough, J., specially concurring).) Given the lack of proof of prior employment, the Commission’s determination is not against the manifest weight of the evidence. The circuit court is reversed.
Circuit court reversed; Commission order reinstated.
RAKOWSKI, WOODWARD, SLATER, and RARICK, JJ., concur.