Plaintiff Carrie Lavalley claims that defendant, E.B. and A.C. Whiting Company, discriminated against her on the basis of sex in violation of the Vermont Fair Employment Practices Act, 21 V.S.A. § 495(a)(1). She appeals the decision of the superior court granting defendant’s motion for summary judgment. We affirm.
Plaintiff is a full-time employee at defendant’s plant, where she earns $359.20 per week. She works on a production line as a cutter/packer, responsible for cutting long bundles of plastic fibers produced at the plant. She is required to stand throughout the shift and to lift heavy items.
In March 1991, plaintiff advised defendant that she was pregnant and unable to perform her job requirements. She presented defendant with a note from her physician stating that her condition rendered her unable to stand for extended periods of time and made it difficult for her to perform heavy lifting. Plaintiff asked defendant to accommodate her for the duration of her pregnancy either by having a co-worker assist her with heavy lifting or by temporarily assigning her to light duty. Defendant denied plaintiff’s request to be accommodated on the job, and pursuant to the company’s disability policy, classified her disability as a long-term, nonwork-related injury, and placed her on disability leave with half-pay of $160 per week. On July 5,1991, plaintiff gave birth and, after her maternity leave, returned to her former job at full pay.
Defendant’s policy divides workers into those whose disability is work-related and qualifies them for workers’ compensation benefits, and those whose disability is not work-related. The former are encouraged to accept whatever accommodations are possible, and compatible, with their ability to work. If alternative work is available, workers in this category are placed in it. Whatever the accommodation, including the placement in alternative work, the employees in this category receive full pay as long as they work.
The second category, those with nonwork-related disabilities, is further divided into two subcategories. The first subcategory is made *208up of minor and short-term disabilities that are handled by shift supervisors through accommodations that enable the employee to continue at full pay. The second subcategory is made up of those with long-term disabilities that render the worker unable to substantially perform his or her responsibilities. According to defendant’s union contract, these employees are placed on disability leave at fifty percent of salary up to a maximum of $160 per week. Plaintiff is in this subcategory.
Plaintiff brought an action alleging that defendant discriminated against her on account of her pregnancy, violating Vermont’s Fair Employment Practices Act (FEPA). See 21 V.S.A. § 495(a)(1) (“It shall be unlawful employment practice . . . [f]or any employer... to discriminate against any individual because of. . . sex. . . .”). In her complaint, she alleged that defendant violated FEPA “[b]y placing Plaintiff (who was pregnant) into a class of persons (those injured away from the job) and treating her differently from others who are disabled (those injured on the job).”
Both parties moved for summary judgment, and the superior court granted defendant’s motion. The court reasoned:
Here, the Plaintiff has not presented evidence proving discriminatory motive. In fact, the practice of paying half salary for long term non-work-related illness was motivated by a union contract, not gender. The Plaintiff’s reduction in salary was not due to her gender or pregnancy but the length of time and physical limitations created by her condition.
Even if we assume that the Plaintiff established gender as the motivating factor, we could not conclude that the Defendant violated Vermont’s Fair Employment Practices Act. Defendant’s decision to pay the plaintiff fifty percent of her salary during the time she was unable to substantially perform her employment was made and would have been absent any discriminatory motive.
Plaintiff appealed the court’s decision to this Court.
Before we address the reasoning of the trial court, we must examine the basic premise of plaintiff’s claim — that discrimination *209because of pregnancy can be a violation of FEPA.1 Defendant argues that pregnancy discrimination cannot be found to be sex discrimination because: (1) in interpreting FEPA, this Court follows the decisions of the federal courts interpreting Title VII of the Civil Rights Act of 1964, the statute after which FEPA is modeled; (2) the United States Supreme Court ruled in General Electric Co. v. Gilbert, 429 U.S. 125, 135 (1976), that pregnancy discrimination is not sex discrimination; (3) Congress overruled Gilbert by enacting the Pregnancy Discrimination Act (PDA) as an amendment to Title VII, 42 U.S.C. § 2000e(k); (4) the Vermont Legislature failed to amend FEPA to include the language of the federal amendment and left Gilbert as the proper interpretation of FEPA.
Defendant is correct that FEPA is patterned after Title VII of the CM Rights Act, Graff v. Eaton, 157 Vt. 321, 323, 598 A.2d 1383, 1384 (1991), and makes it unlawful for any employer to discriminate against any individual because of his or her sex. 21 V.S.A. § 495(a)(1). We find, however, significant weaknesses in its argument.
We start with two preliminary points. First, this is not a case where the Legislature has adopted a statute patterned on the statute of another state or the United States after a determinative court decision interpreting the model statute. Where the court decision precedes our adoption, “the presumption is that the Legislature also adopted the construction given the statute by the courts of the other [jurisdiction].” Hartnett v. Union Mut. Fire Ins. Co., 153 Vt. 152, 154, 569 A.2d 486, 487 (1989). The section of FEPA that sets out the regulatory standard was adopted in 1963.1963, No. 196, § 1. It was amended to include sex discrimination in 1971,1971, No. 9, § 1, and was again amended in 1976 to its current form.2 1975, No. 198 (Adj. Sess.), § 1. All of the legislative action preceded the date of the Gilbert decision. Thus, there is no presumption that the Legislature intended to adopt the construction of the statute in Gilbert. See State v. Wilcox, 160 Vt. 271, 273 n.1, 628 A.2d 924, 925 n.1 (1993).
Second, although Vermont has patterned FEPA on Title VII, we are not bound by federal court interpretations of Title VII in *210construing FEPA. “[W]e look to federal case law for guidance in construing” identical provisions. Hodgdon v. Mt. Mansfield Co., 160 Vt. 150, 165, 624 A.2d 1122, 1130 (1992). In other words, federal decisions represent persuasive authority on the proper interpretation of FEPA. They are not, however, the only .sources of persuasive authority. Many states have enacted employment discrimination laws patterned in whole or in part on Title VII. Decisions from the courts of those states are also sources of persuasive authority
We make these points to emphasize that we will not adopt an interpretation of FEPA solely because the federal courts, including the United States Supreme Court, have so interpreted Title VII. Nor do we believe that the Vermont Legislature must react to every federal decision interpreting Title VII or risk that its inaction will be interpreted as an endorsement of the federal decision.
In this case, we are more persuaded by the decisions of the courts of our sister states, which have overwhelmingly found in interpreting similar or identical statutes that pregnancy discrimination can be sex discrimination. See Colorado Civil Rights Comm’n v. Travelers Ins. Co., 759 P.2d 1358, 1365 (Colo. 1988); Massachusetts Elec. Co. v. Massachusetts Comm’n Against Discrimination, 375 N.E.2d 1192, 1199 (Mass. 1978); Minnesota Mining & Mfg. Co. v. State, 289 N.W.2d 396, 398-99 (Minn. 1979); Bankers Life & Casualty Co. v. Peterson, 866 P.2d 241, 244 (Mont. 1993); Castellano v. Linden Bd. of Educ., 386 A.2d 396, 402 (N.J. Super. Ct. App. Div. 1978), rev’d in part on other grounds, 400 A.2d 1182, 1183 (N.J. 1979); Brooklyn Union Gas Co. v. New York State Human Rights Appeal Bd., 359 N.E.2d 393, 397 (N.Y. 1976); Anderson v. Upper Bucks County Area Vocational Technical Sch., 373 A.2d 126, 130 (Pa. Commw. Ct. 1977); Frank’s Shoe Store v. West Virginia Human Rights Comm’n, 365 S.E.2d 251, 257 (W Va. 1986); Kimberly-Clark Corp. v. Labor & Indus. Review Comm’n, 291 N.W.2d 584, 586 (Wis. 1980). We agree with the reasoning of the Massachusetts Supreme Judicial Court that “[p]regnancy is a condition unique to women, and the ability to become pregnant is a primary characteristic of the female sex. Thus any classification which relies on pregnancy as the determinative criterion is a distinction based on sex.” Massachusetts Elec. Co., 375 N.E.2d at 1198.
We also believe that Gilbert is a weak precedent upon which to rely, given the later congressional action. The Gilbert decision has been roundly criticized. See, e.g., California Fed. Sav. & Loan Ass’n v. Guerra, 479 U.S. 272, 284-85 (1987). In enacting the Pregnancy *211Discrimination Act in 1978, Congress “unambiguously expressed its disapproval of both the holding and the reasoning of the Court in the Gilbert decision.” Newport News Shipbuilding & Dry Dock Co. v. EEOC, 462 U.S. 669, 678 (1983). Congress found that the Court “erroneously interpreted congressional intent and that amending legislation was necessary to reestablish the principles of Title VII law as they had been understood prior to the Gilbert decision.” Id. at 679. It is clear that Congress viewed the Pregnancy Discrimination Act as a reenactment of the proper meaning of Title VII, rather than the creation of expanded civil rights protection for pregnant workers. By not following Gilbert, we are taking proper guidance from the federal law.
Having determined that plaintiff’s claim does not fail because discrimination against pregnant women can never be sex discrimination, we must determine whether FEPA was violated in this case. Plaintiff argues this as a case of disparate treatment and claims that the trial court erred in applying the burden-shifting rules that apply to such claims.
We have held that the standards and burdens of proof under FEPA are identical to those existing under Title VII of the federal Civil Rights Act of 1964. Hodgdon, 160 Vt. at 161, 624 A.2d at 1128; Graff, 157 Vt. at 323, 598 A.2d at 1384. In a series of decisions, we have applied the two main United States Supreme Court precedents that define these standards and burdens in a disparate-treatment case, McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1973), and Price Waterhouse v. Hopkins, 490 U.S. 228 (1989) (superseded in part by Civil Rights Act of 1991, Pub. L. No. 102-166, § 107, 105 Stat. 1071, 1075 (1991) (codified at 42 U.S.C. § 2000e-2)). See In re McCort, 162 Vt. 481, 491-93, 650 A.2d 504, 510-11 (1994); Hodgdon, 160 Vt. at 159-62, 624 A.2d at 1127-29; Graff, 157 Vt. at 324-27, 598 A.2d at 1384-86; State v. Whitingham Sch. Bd., 138 Vt. 15, 18-22, 410 A.2d 996, 998-1000 (1979). Plaintiff argues that the trial court erred in applying the standards and burdens of McDonnell Douglas and Price Waterhouse because it failed to find that she had established a prima facie case under Price Waterhouse and failed to apply the three-step test of McDonnell Douglas. We conclude that the proper analysis of this case does not depend on the burden allocations of those cases.3
*212It is important to emphasize some points about this case. As specified in the complaint, this is a challenge to an employment policy of defendant, with no claim that plaintiff has been treated unfairly except in conformity with the policy. Although plaintiff’s affidavit in opposition to summary judgment stated that “there were jobs available for me if the Defendant had wanted to employ me full time,” this fact is fully consistent with defendant’s policy. There was no allegation below that the employer treated nonpregnant employees with nonwork-related injuries more favorably by providing them with job accommodation. Cf. Adams v. Nolan, 962 F.2d 791, 796 (8th Cir. 1992) (pregnancy discrimination found where employer accommodated employee with nonwork-related disability for one month but failed to accommodate pregnant employee who sought accommodation for one month). On the contrary, she appears to have been accorded exactly the same rights as were available to all other employees, whether male or female. Nor was there any allegation that defendant departed from its policy in a discriminatory manner, or that the employer implemented the policy to intentionally discriminate against pregnant women. See Lambert v. Genesee Hosp., 10 F.3d 46, 58 (2d Cir. 1993) (no finding of disparate treatment where plaintiffs failed to offer any evidence that employer “departed from [leave of absence] policy in a discriminatory manner, or that the policy itself was designed to discriminate against pregnant women”), cert. denied, 511 U.S. 1052 (1994).
Unlike the employment policy in Gilbert and virtually all of the pregnancy discrimination cases, the policy does not specifically condition any employment or benefit rule on pregnancy. Indeed, the policy is facially neutral and affects pregnant women because their temporary disability, and those of others both male and female, is not work related. See Adams, 962 F.2d at 794-95 (policy that determines availability of light-duty assignments to those with “non-work related injury or illness” is not discriminatory if applicable to all such persons without intent to discriminate against pregnant workers); Ulloa v. American Express Travel Related Servs. Co., 822 F. Supp. 1566, 1571 (S.D. Fla. 1993) (no disparate treatment where pregnant employee who was terminated for taking leave of absence exceeding twelve *213weeks failed to show she was treated less favorably than nonpregnant employees who took leave of absence exceeding twelve weeks); Atwood v. City of Des Moines, 485 N.W.2d 657, 660 (Iowa 1992) (police department’s change in light-duty policy was not pregnancy discrimination where new policies accorded pregnant police officer same rights as were available prospectively to all other employees); see generally Wimberly v. Labor & Indus. Relations Comm’n, 479 U.S. 511, 517 (1987) (state unemployment compensation rule that denies benefits to those who leave jobs for nonwork-related reasons is facially neutral and does not discriminate against pregnant worker).
Nevertheless, disability schemes that are facially neutral with respect to pregnancy may still be subject to a disparate-impact claim. See, e.g., Scherr v. Woodland Sch. Community Consol. Dist. No. 50, 867 F.2d 974, 979 (7th Cir. 1988) (facially neutral practices that have disparate impact on pregnant women can constitute pregnancy discrimination under Title VII). Under a disparate-impact theory, the plaintiff is relieved of the burden of having to show that the employer acted with discriminatory intent; rather, the plaintiff need only show that the employment practice has a discriminatory impact on a protected class and is not justified by business necessity. See, e.g., EEOC v. Warshawsky & Co., 768 F. Supp. 647, 651 (N.D. Ill. 1991).
Plaintiff, however, failed to present either to the superior court or to this Court a claim of disparate impact. A theory of disparate impact should be pled specifically in the complaint. See Johnson v. Allyn & Bacon, Inc., 731 F.2d 64, 73 (1st Cir.) (disparate-impact theory not presented to trial court cannot be raised for first time on appeal), cert. denied, 469 U.S. 1018 (1984); 2 L. Larson, Employment Discrimination § 20.03, at 20-13 (2d ed. 1994) (“[A]t-tempts to parlay a disparate treatment case into a disparate impact case at a later stage of the litigation will usually not succeed if the pleadings did not allege facts sufficient to provide the requisite fair notice to the defendant. . . .”). Even if plaintiff has made a disparate-impact claim, we have no record of the impact of defendant’s policy, except on plaintiff. We do not know whether the policy has a disproportionate impact on women employees generally.4 Nor do we *214know that the entire policy has a negative impact on pregnant women generally.5
*215As a result of plaintiff’s failure to make a disparate-impact claim, plaintiff’s claim is reduced to one that defendant’s policy is sex discrimination per se because it excludes pregnant women from doing some job that they might be capable of performing. In other words, plaintiff’s claim is that FEPA prevents an employer from making disability-related employment decisions on the basis of whether the disability is work-related. We disagree that the policy violates FEPA on its face. Cf. Wimberly, 479 U.S. at 517 (neutral unemployment compensation disqualification policy, based on whether an employee leaving is work related, does not violate pregnancy discrimination prohibition of Federal Unemployment Tax Act). Plaintiff would have been treated no differently had she suffered a nonwork disability unrelated to pregnancy. Defendant was not required to grant to plaintiff benefits that were otherwise not available to other employees with nonwork-related disabilities.
The policy distinction that plaintiff challenges is fundamentally rooted in the workers’ compensation laws, which provide that a worker who “receives a personal injury by accident arising out of and in the course of his employment” is entitled to compensation as provided by law. 21 V.S.A. § 618. If the worker “is unable to perform work for which the employee has previous training or experience, the employee shall be entitled to vocational rehabilitation services, including retraining and job placement, as may be reasonably necessary to restore the employee to suitable employment.” Id. § 641. In cases of partial temporary disability, the amount of compensation is based on the difference between the “average weekly wage before the injury and the average weekly wage which he or she is able to earn thereafter.” Id. § 646.
In view of the separate statutory scheme for workers who are injured in the course of their employment, we cannot conclude that the Legislature intended FEPA to reach the gender consequences of that statutory scheme. We do not accept the dissent’s argument that there need be no consistency between an employer’s disability benefits policy and its disability employment policy. As noted above, the workers’ compensation laws provide rights to both replacement income and retraining and job placement to restore the worker to suitable employment. If we accept plaintiff’s argument here, we are necessarily holding that the governmental foundation for the distinctions defendant’s policy draws are also discriminatory. We conclude that defendant did not discriminate against plaintiff on account of sex in violation of 21 V.S.A. § 495(a)(1).
Affirmed.
Following initial briefing in this case, we requested that the Attorney General submit a brief on this point. The Attorney General has done so, arguing that discrimination on the basis of pregnancy is sex discrimination prohibited by FEPA. The brief of the Attorney General has been helpful, and, as set out in the text, we agree with the Attorney General’s position.
There have been further amendments to this section, but they are unrelated to the dispute before us.
The dissent argues that plaintiff is entitled to reach the jury on a claim that defendant’s “explanation . . . is a pretext for discrimination against pregnant women.” 166 Vt. at 220, 692 A.2d at 376. Even if we assume that plaintiff made out a prima facie *212case of discrimination, she has never made a claim of pretext, arguing instead that the policy itself is discriminatory irrespective of defendant’s reasons for implementing it. As a result, the trial court concluded that “plaintiff has not presented evidence proving discriminatory motive.” The conclusion is clearly correct on this record.
The only evidence before us is contained in an exhibit attached to defendant’s memorandum in support of summary judgment and sworn to by defendant’s controller. It shows the gender, period, and reason of all persons who were placed in temporary disability status and received the same benefits as plaintiff. In 1989, seventeen persons were placed in that status and all were male. In 1990, thirteen persons were placed in *214that status and five were female. The disability for one of the females was pregnancy related. In 1991, twenty-one persons were placed in that status and three were female. Again, the disability of one of the females was pregnancy related. This information is of limited value, however, because we do not know the gender makeup of the workforce.
For example, those with work-related disabilities receive workers’ compensation benefits of unknown amounts. The amount would, of course, affect the desirability of this alternative if no other job were available.
We do not even know whether defendant’s policy generally affects pregnant workers in the way it affected plaintiff. If the nature of a pregnant worker’s job is such that it can be performed up to a short period before giving birth, with reasonable workplace accommodations, the availability of other jobs may be irrelevant. Thus, we do not accept the dissent’s argument that “defendant has employed a classification that always excludes pregnant women from job opportunities.” 166 Vt. at 218, 692 A.2d at 374. We can determine whether the impact of the defendant’s policy is disparate, under any theory, only if there is evidence of how it has affected pregnant and nonpregnant employees.
Assuming that a disability scheme treats employees who suffer nonwork-related injuries less favorably than employees who suffer work-related injuries, a question arises as to the applicability of the disparate-impact analysis to a claim that a facially neutral disability scheme has a discriminatory effect on the basis of pregnancy. On this point, the precedents under the PDA are split.
Some courts have held that, under the PDA, the proper comparison in disparate-treatment or disparate-impact cases is between pregnant women and nonpregnant persons. See EEOC v. Ackerman, Hood & McQueen, Inc., 956 F.2d 944, 948 (10th Cir.) (comparison in pregnancy discrimination claim is between pregnant women and nonpregnant workers), cert. denied, 506 U.S. 817 (1992); Maganuco v. Leyden Community High Sch. Dist. 212, 939 F.2d 440, 444 (7th Cir. 1991) (to establish claim that sick leave policy had disparate impact on pregnant women, comparison is between women disabled due to pregnancy and male coworkers or women who have not experienced pregnancy-related disability). The Tenth Circuit Court in Ackerman reasoned:
The [PDA] requires courts to inquire whether the employer treats pregnancy or pregnancy-related conditions differently than other medical conditions. . . .
. . . The clear language of the PDA requires the court to compare treatment between pregnant persons and “other persons not so affected but similar in their ability or inability to work.” The comparison is, therefore, between pregnant and nonpregnant workers, not between men and women.
956 F.2d at 948 (quoting 42 U.S.C. § 2000e(k)).
Others hold that, because the PDA was a definitional amendment, it did not add a new cause of action for pregnancy discrimination. Warshawsky, 768 F. Supp. at 653. The Warshawsky court concluded, therefore, that “the statistical analysis under a disparate impact theory is the traditional ‘sex discrimination’ analysis.” Id. at 654 (footnote omitted). In other words, the comparison is between female employees and male employees. See id. at 655 (finding disparate impact where first-year female employees were more likely to be fired than first-year male employees).
Because of the failure of plaintiff to produce evidence of the impact of defendant’s policy, we do not have to determine how we will resolve this question under FEPA.