In March 1983, Rebecca Maganuco filed suit against Leyden Community High School District 212 alleging that the school district’s maternity leave policy violated Sections 703(a)(1) and (a)(2) of Title VII of the Civil Rights Act of 1964 (“the Act”), 42 U.S.C. § 2000e-2(a)(1) and (2), as amended by the Pregnancy Discrimination Act of 1978 (“PDA”), 42 U.S.C. § 2000e(k). In November 1984, Susan Scherr filed suit against Woodland School Community Consolidated District No. 50 similarly alleging that her school district’s maternity leave policy violated Title VII. The district courts granted summary judgments in favor of the defendant school districts under both disparate treatment and disparate impact theories of liability. We reverse in part and remand.
I
Rebecca Maganuco, a tenured teacher at Leyden Community High School (“Ley-den”), became pregnant in 1981 and in June of that year informed the defendant school district that she planned to use accumulated paid sick leave to cover the period she expected to be disabled due to her pregnancy, and that as soon as she was no longer disabled, she planned to take an unpaid maternity leave for the remainder of the 1981-1982 school year. Despite her request to combine paid sick leave and an unpaid leave of absence to accommodate *976the period in which she was disabled as well as the period she chose to remain at home to care for her child, the school district’s policy forced her to choose between taking either paid sick leave for the period in which she was disabled due to her pregnancy and delivery of her child, with an immediate return to work at the termination of her disability, or unpaid leave for both the period she was disabled and the period she remained at home to provide childcare.
Susan Scherr, a tenured teacher at Woodland School Community Consolidated District No. 50 (“Woodland”), became pregnant and requested paid sick leave from April 18,1983 through the end of the 1982-1983 school year and a maternity leave for the 1983-1984 school year. Although Scherr was disabled from April 18 through the end of the 1982-1983 school year, Woodland did not allow her to take paid sick leave for this time. Instead, it required her to take an unpaid maternity leave for the entire period beginning April 18, 1983 through the 1983-1984 school year.
Each woman filed a suit alleging that her school district’s leave policies discriminated against pregnant women. In both suits, the plaintiffs argued for liability under both disparate treatment and disparate impact theories and asked for monetary damages for the days they were disabled but did not receive paid sick leave as well as permanent injunctive relief against the allegedly discriminatory maternity leave policies.
In Maganuco v. Leyden Community High School District 212, the parties filed cross-motions for summary judgment. On
October 24, 1984, Judge Paul E. Plunkett denied Maganuco’s motion for summary judgment, granted Leyden’s motion for summary judgment on Maganuco’s disparate treatment theory and denied Leyden’s motion for summary judgment on Maganu-co’s disparate impact theory. The parties filed cross-motions for reconsideration and on June 7,1985, Judge Plunkett reaffirmed his decision. The case was transferred to Judge Ann Williams and the parties proceeded with their preparations for trial on the disparate impact theory.
Meanwhile, the parties in Scherr v. Woodland Community Consolidated School District No. 50, which was also before Judge Williams, filed cross-motions for summary judgment. On December 15, 1986, Judge Williams denied Scherr’s motion for summary judgment and granted Woodland’s motion for summary judgment under both the disparate impact and disparate treatment theories. Seeking a different result on the disparate impact theory in a factually similar case also before Judge Williams, the parties in Maganuco renewed their cross-motions for summary judgment. Stating that it would be anomalous for the court to allow Maganuco to go to trial on the disparate impact claim when it had already held in Scherr that such a claim could not succeed, Judge Williams disagreed with Judge Plunkett and granted defendant Leyden’s motion for summary judgment. Both plaintiffs appealed and this Court consolidated the cases.
Both school districts provide teachers three types of leave: (1) paid sick leave— with continuing unpaid sick leave for extended illness, (2) unpaid maternity leave, and (3) an unpaid general leave of absence. In both school districts, a pregnant teacher can choose only one of the three forms of leave.
The school districts have similar paid sick leave policies. Teachers employed by Ley-den accumulate 17 paid days of sick leave per year and unused days accumulate from year to year up to a maximum of 180 days. Teachers who exhaust their accumulated sick leave may be granted an extended, unpaid, personal illness leave to cover the period during which they remain disabled. Pregnant teachers may elect to use their accumulated sick leave to cover the period during which they are disabled; those who do are paid for the days they are disabled and must return to work as soon as they are no longer disabled.
Teachers employed by Woodland accumulate 11 days of paid sick leave per year and may accumulate up to 180 days. Teachers who exhaust their paid sick leave *977may be eligible, after a one-day waiting period, for up to 10 additional days of paid sick leave from the sick-leave “bank.” Teachers who remain disabled after exhausting the sick-leave bank are eligible for unpaid “continuing illness” leave for the longer of 90 days or the remainder of the school year, although the Board of Education may at its discretion extend the leave for the duration of the disability. A teacher may elect to use disability leave (i.e., sick days, sick-bank days or continuing illness leave) during any period of disability related to her pregnancy and/or to the delivery of the child. Any teacher who elects disability leave must return to work immediately following the termination of the disability.
Both school districts also provide unpaid maternity leave as of right. At Leyden, any tenured teacher who becomes pregnant will be granted a maternity leave of up to three semesters, commencing on the date her doctor considers her unable to carry on her teaching duties. The teacher must give the administration no less than 60 days notice prior to the leave and at least 90 days notice prior to the end of the preceding semester of the date when she wants to start work. The administration may delay her return until the beginning of the school year following her request to return. Any teacher who opts for maternity/child-rearing leave must forgo the benefit of receiving paid sick leave for the period she is actually disabled due to the child’s birth. However, paid sick leave is available prior to the commencement of a scheduled maternity/child-rearing leave if the teacher becomes ill, whether or not the illness is related to her pregnancy.
The terms of Woodland’s unpaid maternity leave plan are provided in the Collective Bargaining Agreement between the teachers’ union and Woodland. Under the plan, teachers must schedule the leave in consultation with the Superintendent. The agreed-upon leave plan must be scheduled to take into consideration maintenance of continuity of instruction and medical factors to the maximum degree possible and must commence upon the earliest of (1) fifteen days prior to the anticipated delivery date, (2) the actual delivery date, or (3) the date on which the teacher becomes unable to perform her duties. The leave cannot exceed the balance of the school term in which it commences and one additional term, and every effort must be made to have the leave terminate prior to the start of a new school term. A teacher who elects to take a maternity leave and then becomes ill prior to the scheduled commencement of the leave, may use her sick days even if the illness is due to her pregnancy. However, the collective bargaining agreement expressly states that sick leave is not applicable during the period of the maternity leave; as a result, since maternity leave must begin at the latest date prior to the actual delivery date, a pregnant teacher who chooses maternity leave cannot take paid sick days for the period during which she is disabled. Woodland’s maternity leave policy also expressly states that nothing in the policy shall be construed as requiring any teacher to apply for a maternity leave and reasserts the independent option of using accumulated sick leave during any period of disability related to her pregnancy or the delivery of the child and returning to work immediately following the termination of the disability-
Finally, both school districts provide for an unpaid general leave of absence. Under the Leyden policy, any tenured teacher may request a leave of absence and at least one-half of those applying each year must be granted this leave. In practice, all ten requests for general leave during the time period relevant to this case were granted. General leaves of absence at Leyden must begin and end in conformance with the school year. Woodland offers, at the Board’s discretion, an unpaid leave of absence of up to one year. Such requests are normally granted and may begin at any agreed-upon time. Nothing in either of the school district’s leave policies precludes a pregnant teacher from using this leave instead of maternity leave.
II
The defendant school districts, aided by the United States as amicus curiae, *978raise an argument that questions the theoretical foundations of the PDA. They contend that the PDA is a sweeping mandate of equal treatment, but does not contemplate disparate impact claims.1 They find support for this contention in both the plain language and legislative history of the PDA. From the language of the statute they emphasize the portion that states, “women affected by pregnancy, childbirth, or related medical conditions shall be treated the same for all employment-related purposes.” 42 U.S.C. § 2000e(k) (emphasis added). From the legislative history they argue that “[t]he primary purpose of the PDA was to eliminate the need for employees to rely on disparate impact analysis to support their Title VII claims based on pregnancy classifications ‘[b]y making clear that distinctions based on pregnancy are per se violations of Title VII.’ ” (H.R. Rep. No. 95-948, 95th Cong., 2d Sess. 3 (1978) U.S.Code Cong. & Admin.News 1978, pp. 4749, 4751). By the same token, Congress intended that disparate impact analysis would play no role in the pregnancy context after passage of the PDA. (Br. of the United States at 19.)
In their arguments, the defendants err by neglecting the context of the PDA and treating it as an independent statutory enactment. It is not. The PDA is an amendment to a highly developed statutory scheme added in response to the Supreme Court’s decision in General Electric v. Gilbert, 429 U.S. 125, 97 S.Ct. 401, 50 L.Ed.2d 343 (1976).
In Gilbert, the Court held that General Electric’s exclusion of pregnancy coverage from its otherwise comprehensive disability plan did not constitute discrimination based on sex. While acknowledging that pregnancy is, of course, confined to women, the Court nevertheless reasoned that the pregnancy classification was not sex-based because it divided potential beneficiaries into two groups — “pregnant women and non-pregnant persons.” Id. at 135-36, 97 S.Ct. at 407-08.
Congress “ ‘unambiguously expressed its disapproval of both the holding and the reasoning of the Court in the Gilbert decision’ ” by amending Title VII. California Federal Savings and Loan Association v. Guerra, 479 U.S. 272, 107 S.Ct. 683, 691, 93 L.Ed.2d 613 (1987) (quoting Newport News Shipbuilding & Dry Dock Co. v. EEOC, 462 U.S. 669, 678, 103 S.Ct. 2622, 2628, 77 L.Ed.2d 89 (1983)). The amendment contained in the PDA provides in relevant part that:
(k) The terms “because of sex” or “on the basis of sex” include, but are not limited to, because of or on the basis of pregnancy, childbirth or related medical conditions; and women affected by pregnancy, childbirth, or related medical conditions shall be treated the same for all employment-related purposes, including receipt of benefits under fringe benefit programs, as other persons not so affected but similar in their ability or inability to work....
42 U.S.C. § 2000e(k). Congress inserted the PDA as a new subsection to the definitions section which is applicable “[f]or the purposes of this subehapter.” 42 U.S.C. § 2000e. Placed where it is within the Act as section 701(k), the PDA supplements the meaning of “because of sex” and “on the basis of sex” within Title VII as a whole. Newport News, 462 U.S. at 678, 103 S.Ct. at 2628.
As a definitional amendment, the PDA provides no substantive rule to govern pregnancy discrimination. Instead, as part of Title VII, the PDA finds force through the substantive sections of the Act, sections 703(a)(1) and (2).2 Title VII prohibits two types of employment discrimination. *979First, it prohibits disparate treatment-intentional discriminatory treatment of employees based on impermissible criteria. International Brotherhood of Teamsters v. United States, 431 U.S. 324, 335-36 n. 15, 97 S.Ct. 1843, 1854-55 n. 15, 52 L.Ed.2d 396 (1977). Second, it prohibits disparate impact — facially neutral practices that nevertheless have a discriminatory impact and are not justified by business necessity. Id. Because the PDA is part of Title VII and derives its substance and procedures from the Act as a whole, a claim of pregnancy discrimination, like any other claim of discrimination under Title VII, may be based either on a theory of disparate treatment or a theory of disparate impact.3 See, e.g., Chambers v. Omaha Girls Club, Inc., 834 F.2d 697, 700-01 (8th Cir.1987); King v. Trans World Airlines, Inc., 738 F.2d 255, 258-59 n. 2 (8th Cir.1984); Hayes v. Shelby Memorial Hospital, 726 F.2d 1543, 1547 (11th Cir.1984); Zuniga v. Kleberg County Hospital, 692 F.2d 986, 989 (5th Cir.1982); Abraham v. Graphic Arts International Union, 660 F.2d 811, 819 (D.C.Cir.1981) (all using disparate impact under the PDA).
The context in which the PDA was enacted indicates that, contrary to defendants’ argument, Congress did not intend the same treatment of pregnancy and other disabilities to mean identical treatment. As discussed infra, Congress passed the PDA in reaction to Supreme Court decisions that permitted employers to exclude pregnancy from insurance coverage because of its unique status. When it eon-sidered the PDA, Congress had before it extensive evidence of discrimination against pregnancy, particularly in disability and health insurance programs like those challenged in Gilbert and Nashville Gas Co. v. Satty, 434 U.S. 136, 98 S.Ct. 347, 54 L.Ed.2d 356 (1977), and opposition to the PDA came from those concerned with the cost of including pregnancy in health and disability benefit plans, not from those who favored special accommodation of pregnancy. Cal.Fed., 107 S.Ct. at 692. In this context, it is clear that the statutory requirement that pregnancy receive the “same” treatment as other disabilities was not intended as an ultimate end, but as a means of assuring that pregnancy would not be excluded, as it was in Gilbert, from any list of compensable disabilities. See Note, Sexual Equality Under the Pregnancy Discrimination Act, 83 Colum.L. Rev. 690, 719.4 Moreover, interpreting “same” to mean “identical” would go against “[t]he entire thrust ... behind this legislation [which] is to guarantee women the basic right to participate fully and equally in the workforce, without denying them the fundamental right to full participation in family life.” 123 Cong.Rec. 29658 (1977) (quoted with approval in Cal.Fed., 107 S.Ct. at 693-94).
In their argument that Congress intended the PDA to preclude a disparate impact approach, the defendants ignore not only the broad context in which the PDA was passed, but also the remainder of the sen*980tence they quote for support. See infra p. 980. As to the broad context, under Gilbert, pregnancy classifications were gender neutral. Therefore when the Supreme Court considered a pregnancy discrimination claim in Satty, it used the theory applicable to facially neutral practices — disparate impact — and subsequently found that the employer’s policy of denying accumulated seniority to female employees returning from pregnancy leave violated Title VII. Congress recognized that the result of Gilbert and Satty was to leave employers and employees in the untenable position of guessing whether their policies violated Title VII. H.R.Rep. No. 95-948, 95th Cong., 2d Sess. 3 (1977). Through the PDA, Congress eliminated a great deal of this confusion by reversing the premise that pregnancy-based classifications were gender neutral and expressly recognizing pregnancy-based classifications as sex-based. As proponents of the bill repeatedly emphasized, the “amending legislation was necessary to re-establish the principles of Title VII law as they had been understood prior to the Gilbert decision.” Newport News, 462 U.S. at 679, 103 S.Ct. at 2628. Accordingly, the PDA did not abolish “the impact approach.” It restored it to its original purpose under Title VII as a means of challenging facially neutral employment practices that nevertheless discriminate. See Note, Employment Equality Under the Pregnancy Discrimination Act of 1978, 94 Yale L.J. 929, 936 (1985).
As to the narrow context, the remainder of the quotation relied on by the defendants makes clear the error in their conclusion that Congress intended disparate impact analysis to play no role in pregnancy discrimination cases. The full quotation reads as follows: “By making it clear that distinctions based on pregnancy are per se violations of Title VII, the bill would eliminate the need in most instances to rely on the impact approach, and thus would obviate the difficulties in applying the distinctions created in Satty.” Id. As the full quotation reveals, Congress anticipated that the PDA would eliminate the need for a disparate impact theory “in most instances,” thereby recognizing that in some instances plaintiffs would still need to rely on Title VII’s disparate impact approach to rectify discrimination against them.
Finally, defendants argue that even if plaintiffs generally can use the disparate impact theory under the PDA, the plaintiffs in these cases cannot because their claims are under section 703(a)(1) for discrimination “with respect to ... compensation, terms, conditions or privileges of employment” and the disparate impact theory is inapplicable to violations of this section of Title VII. The disparate impact theory developed in cases alleging violations of section 703(a)(2) (e.g., Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971); Dothard v. Rawlinson, 433 U.S. 321, 97 S.Ct. 2720, 53 L.Ed.2d 786 (1977)), and the Supreme Court has not yet expressly decided whether subsection (a)(1), like (a)(2), allows a suit based on disparate impact. Other courts, including this one, have followed the Supreme Court’s lead by not addressing the issue directly and have applied a disparate impact analysis to section 703(a)(1) claims without considering whether the claims were being brought under section 703(a)(1) or (a)(2). See, e.g., Liberies v. County of Cook, 709 F.2d 1122, 1130 (7th Cir.1983); EEOC v. Ball Corp., 661 F.2d 531, 540 (6th Cir.1981). However, since oral argument in this case, we addressed the issue directly and held that disparate impact theory applies under section 703(a)(1) as well. Colby v. J.C. Penney Co., 811 F.2d 1119, 1127 (7th Cir.1987). Two other Circuits have addressed the issue directly and both have held that disparate impact claims can be brought under section 703(a)(1). EEOC v. J.C. Penney Co., 843 F.2d 249, 252 (6th Cir.1988) (expressly agreeing with our reasoning in Colby); Wambheim v. J.C. Penney Co., 705 F.2d 1492, 1494 (9th Cir.1983).
Having resolved the more general legal question of whether plaintiffs Scherr and Maganuco can attempt to establish pregnancy discrimination on a disparate impact theory, we turn to the specific inquiries of whether issues of material fact exist that preclude the grants of summary judgment *981in favor of the defendant school districts in both cases on both theories.
Ill
Disparate Treatment
A. Scherr v. Woodland
In October 1984, Judge Plunkett granted the defendant school district Ley-den’s motion for summary judgment on the disparate treatment theory because he found that the school district’s “policies, at least on their face, ‘discriminate’ between forseeable and unforseeable disabilities, not between pregnancy-related disabilities and other disabilities.” Maganuco v. Leyden Community High School District 212, No. 83 C 1975, slip op. at 8 (N.D.Ill.1984). Thus, in a manner reminiscent of the since rejected theory of Gilbert, he held that the policies were not facially discriminatory on the basis of sex.
One month later, Susan Scherr filed her virtually identical suit against Woodland and Woodland defended on the theory that its leave policies differentiated between forseeable and unforseeable disabilities. Judge Williams correctly recognized this argument for what it was — “a post-hoc attempt to fit its [Woodland’s] leave policies within Judge Plunkett’s scheme.” Scherr v. Woodland School Community Consolidated District No. 50, No. 84 C 10185, slip op. at 8 n. 2 (N.D.Ill.1986) [1986 WL 15115], Judge Williams nevertheless granted summary judgment in the defendant’s favor on the disparate treatment theory because she found that no group was treated more favorably than pregnant teachers in the application of defendant’s sick leave and non-disability leave policies.
Judge Williams’ rejection of the forseea-bility-unforseeability distinction left Woodland without an explanation for differences in its policies. Rather than formulate another explanation for the differences, on appeal Woodland argues that no differences exist. It asserts that under its leave policies, no teacher, man or woman, pregnant or not pregnant, may combine paid and unpaid leaves to run consecutively. Both Woodland and Scherr agree that such a “no combination” policy exists for maternity leave; this policy is stated in Article VI(D)(4) of the Collective Bargaining Agreement. However, the parties dispute the existence of a “no combination” policy for non-pregnant teachers. Both agree that no such restriction is expressly stated in defendant’s written nondisability leave of absence policy, Section 4152.8 of the Policy Manual, but Woodland asserts the existence of an informal, unwritten policy.5 Woodland supports this assertion with the following statement in the affidavit of Superintendent Arden Luce: “8. Paid Sick leave or any other disability related leave is not available during the term of any non-disability related leave of absence identified in paragraphs 6 [maternity leave] — 7 [general leave of absence] above.”
In determining whether a genuine issue of material fact exists, the district court must construe the facts alleged in the light most favorable to the party opposing the motion for summary judgment. Richardson v. Penfold, 839 F.2d 392, 394 (7th Cir.1988). Upon review, we must consider the entire record and all reasonable inferences drawn from the record in the same light. Id.; Jakubiec v. Cities Services Co., 844 F.2d 470, 471 (7th Cir.1988). There is substantial evidence to support Scherr’s contention that Woodland did not have a legitimate “no combination” policy applicable to non-pregnant teachers prior to her lawsuit. The purported policy was never formally adopted, never made the subject of collective bargaining, never communicated to those to whom it applied, and never actually applied. Whether Woodland has a no combination policy applicable to non-pregnant teachers is a genuine issue of material fact that goes to the heart of Scherr’s ability to make a prima facie case of dispár*982ate treatment.6
In his dissenting opinion, Judge Manion contends that the disability leave plans of the school districts, in isolation, do not result in a disparate impact on pregnant teachers since the pregnant teachers are not forced into forgoing their paid sick days, but instead presumably choose to forgo them in favor of the longer maternity leave. Indeed, the dissent regards the availability of unpaid maternity leave as a gratuitous “extra option,” since maternity leave is not available to non-pregnant teachers. This fails to recognize that although maternity leave is not available to non-pregnant teachers, a similar general leave of absence is available to all teachers.
Plaintiffs simply seek to be allowed to combine paid sick leave with unpaid maternity leave if the school districts allow or have failed to establish a legitimate policy against combination of paid sick leave and a general leave of absence. The dissent must then assume that no material question of fact exists regarding the existence of a policy prohibiting the combination of paid sick leave with any unpaid leave. We, on the other hand, hold that there is sufficient evidence to raise a question of fact as to whether Woodland had a legitimate “no-combination” policy in place prior to this lawsuit for both pregnant and non-pregnant teachers since no such policy was ever formally adopted or communicated to the teachers prior to this suit. Accordingly, we reverse and remand for such a determination.
B. Maganuco v. Leyden
As discussed above, in Maganuco Judge Plunkett granted the defendant Ley-den’s motion for summary judgment on the disparate treatment theory because he found that the policies discriminated on the basis of the forseeability of the disability, not on the basis of sex. Although rejecting this rationale, we affirm the grant of summary judgment.
Like Woodland, Leyden asserts that it treats all teachers alike, that no Leyden teacher may take disability leave days during the period of a non-disability leave or schedule disability and non-disability leave to run consecutively, so that there is no facial discrimination to support Maganuco’s disparate treatment theory. Also like Woodland, Leyden has a written policy prohibiting a teacher from combining paid sick leave for the period of disability due to pregnancy and childbirth with unpaid maternity leave for the following period in which she is not disabled (Board/Union Contract, Article X, Section 4), but no written policy prohibiting the combination of paid sick leave with a general leave of absence.7 And like Woodland, Leyden has never granted a request to combine sick *983leave with a general leave of absence (no teacher has ever made such a request) and there is no evidence that it ever articulated such a policy prior to this litigation.
However, Leyden prevails where Woodland lost because of a critical factual difference in the leave policies of the two school districts — general leaves of absence at Ley-den must begin and end in conformance with the school year. Because a general leave of absence must begin at the beginning of the school year, no teacher could ever take paid sick leave immediately prior to the beginning of his or her general unpaid leave of absence. Prior to that time, school would be in summer recess and the teacher would not be working, and thus not entitled to paid sick leave. From the terms of its leave of absence policy, it is clear that Leyden would not allow any teacher to use paid sick leave immediately prior to taking a leave of absence. Because no teacher may combine the two types of leave, there is no disparate treatment. We affirm the grant of summary judgment in favor of Leyden on the disparate treatment theory.
IV
Disparate Impact
In both Scherr and Maganuco, Judge Williams granted the defendant school districts’ motions for summary judgment on the plaintiffs’ disparate impact claims. We reverse and remand both cases.
A. Scherr v. Woodland
In Scherr, the court stated that it “believe[d] that neither plaintiff’s nor defendant’s evidence sheds any light on the impact of defendant’s leave policies on pregnant teachers.” The court nevertheless granted summary judgment for the defendants because it found that Scherr “has pointed to no teacher or group of teachers upon whom defendant’s leave policies had an impact different from the impact of those policies on pregnant teachers.” The district court erred in two ways. First, finding neither party’s evidence dispositive, it construed the facts alleged in the light most favorable to the party making the motion. The facts must be construed to favor the party opposing the motion. See Jakubiec, 844 F.2d at 471.
Second, the district court required Scherr to present evidence of hypothetical non-pregnant teachers who would want to combine paid and unpaid leave. This is the type of evidence that would be appropriate to require in a disparate treatment case, but is inappropriate to require in a disparate impact case. “The distinguishing features of the factual issues that typically dominate in disparate impact cases do not imply that the legal issue is different than in cases where disparate treatment analysis is used,” Watson v. Fort Worth Bank and Trust, — U.S. -, 108 S.Ct. 2777, 2785, 101 L.Ed.2d 827 (1988) (O’Connor, J., plurality opinion), but the “factual issues and the character of the evidence are inevitably somewhat different when the plaintiff is exempted from the need to prove intentional discrimination_ The evidence in these ‘disparate impact’ cases usually focuses on statistical disparities, rather than specific incidents, and on competing explanations for those disparities.” Id. 108 S.Ct. at 2784. Therefore rather than insisting on proof of the hypothetical non-pregnant teacher who would combine paid sick leave with unpaid general leave, the district court should look to the proof of the needs of pregnant teachers and compare that to the actual coverage. See California Federal Savings and Loan Association v. Guerra, 758 F.2d 390, 396 (9th Cir.1985), affirmed, 479 U.S. 272, 107 S.Ct. 683, 93 L.Ed.2d 613 (1987) (equality in the disability context compares coverage to actual need, not coverage to hypothetical identical needs). Consequently we reverse and remand.
B. Maganuco v. Leyden
In Maganuco, Judge Williams disagreed with Judge Plunkett’s prior ruling and granted Leyden’s motion for summary judgment. Her rationale was that “it would be anomalous for this court to allow plaintiff to go to trial on the disparate impact claim when the court has already *984held in Scherr that such a claim cannot succeed.” In Scherr, the court held that Scherr’s disparate impact claim could not succeed; it did not hold that no disparate impact claim could succeed under the PDA. Therefore, it would not have been anomalous for Maganuco to proceed to trial on her claim if issues of material fact in her case remained unresolved, even though Scherr was precluded from proceeding to trial in her case. We agree with Judge Plunkett’s original determination that whether Leyden’s leave policies have a disparate impact on pregnant teachers is an issue that, on the basis of the current record, cannot be resolved on cross-motions for summary judgment.
Affirmed in part, reversed in part and remanded, costs to be borne by the respective parties.
. This argument was made by the amicus curiae and adopted by the defendant school districts.
. Section 703 of Title VII provides that:
(a) It shall be an unlawful employment practice for an employer—
(1) to fail or refuse to hire or to discharge any individual or otherwise to discriminate against any individual with respect to his compensation, terms, conditions, or privileges of employment, because of such individual’s race, color, religion, sex, or national origin; or
(2) to limit, segregate, or classify his employees or applicants for employment in any way which would deprive or tend to deprive any individual of employment opportunities.
. The Supreme Court explained the difference between the two theories as follows:
'Disparate treatment’ ... is the most easily understood type of discrimination. The employer simply treats some people less favorably than others because of their race, color, religion, sex or national origin. Proof of discriminatory motive is critical, although it can in some situations be inferred from the mere fact of differences in treatment....
Claims of disparate treatment may be distinguished from claims that stress ‘disparate impact.' The latter involve employment practices that are facially neutral in their treatment of different groups but that in fact fall more harshly on one group than another and cannot be justified by business necessity.... Proof of discriminatory motive, we have held, is not required under a disparate-impact theo-ry_ Either theory may, of course, be applied to a particular set of facts.
International Brotherhood of Teamsters v. United States, 431 U.S. 324, 335-36 n. 15, 97 S.Ct. 1843, 1854-55 n. 15.
. For additional discussion of whether the “same” treatment requires identical treatment and the choice between equal treatment and special treatment, see Dowd, Maternity Leave: Taking Sex Differences into Account, LIV Fordham L.Rev. 699 (1986); Minnow, Justice Engendered, 101 Harv.L.Rev. 10 (1987); Note, The Conflict Between State Guaranteed Pregnancy Benefits and the Pregnancy Discrimination Act: A Statutory Analysis, 74 Geo.L.J. 1743 (1986); Note, Employment Equality Under the Pregnancy Discrimination Act of 1978, 94 Yale L.J. 929 (1985).
. Consideration of an unwritten, informal plan is appropriate. The EEOC regulations promulgated pursuant to the PDA expressly apply to both written and unwritten employment poli-eles and practices as well as formal and informal sick leave plans. 29 C.F.R. § 1604.10(a) and (b).
. The elements of a prima facie case of sex discrimination when the plaintiff alleges that employees of one sex are discriminately denied a benefit that is extended to employees of the opposite sex is an unsettled issue in this Circuit. In Washington County v. Gunther, 452 U.S. 161, 166 n. 8, 101 S.Ct. 2242, 2246 n. 8, 68 L.Ed.2d 751 (1981), the Supreme Court did not expressly state the necessary elements. Resolution of the cases before us have not required resolution of the issue and in the absence of further guidance from the Supreme Court we have not squarely addressed it. Beard v. Whitley County REMC, 840 F.2d 405, 411 (7th Cir.1988); Jennings v. Tinley Park Community Consolidated School District No. 146, 796 F.2d 962, 965 n. 2 (7th Cir.1986). Instead, in both cases we assumed, without so holding, "that to establish a prima facie case plaintiffs need only show a difference in treatment between two major groups of employees of different sex.” Beard, 840 F.2d at 411; Jennings, 796 F.2d at 965 n. 2; see also Boyd v. Madison County Mutual Insurance Co., 653 F.2d 1173, 1178 (7th Cir.1981), certiorari denied, 454 U.S. 1146, 102 S.Ct. 1008, 71 L.Ed.2d 299 (1982) (affirming district court’s determination that prima facie case existed without formulating a standard). Resolution of the Woodland case does not demand resolution of the issue, so that again we make the same legal assumption and remand to the district court for the predicate factual detemination.
. Article X, Section 4 of the Leyden Board/Union Contract provides in part that a pregnant teacher will be “granted a maternity leave to commence on the date her doctor considers her unable to carry on her teaching duties.” As a result, a pregnant teacher is foreclosed from using her paid sick leave at the beginning of her pregnancy-related disability followed by a maternity leave, since the maternity leave must begin on the date the teacher becomes disabled due to her pregnancy, rather than on the later date when her paid sick leave is exhausted.