539 F.3d 573
7th Cir.
2008

Lead Opinion

RIPPLE, Circuit Judge.

In 1997, Milwaukee County’s Juvenile Detention Center instituted a policy that required each unit of the facility to be staffed at all times by at least one officer of the same sex as the detainees housed on that unit. Because there were far more male units than female units at the facility, this policy had the effect of reducing the number of shifts available for female officers. Ersol Henry and Terri Lewis, both female officers at the facility, brought this action in the United States District Court for the Eastern District of Wisconsin, alleging sex discrimination and retaliation in violation of Title VII, 42 U.S.C. § 2000e, et seq. After a bench trial, the district court concluded that the gender-specific policy was based on a bona fide occupational qualification and that no other discrimination or retaliation had occurred; accordingly, it entered judgment in favor of the County. For the reasons set forth in this opinion, we reverse the judgment of the district court.

I

BACKGROUND

A.

At the time relevant to this appeal, Er-sol Henry and Terri Lewis were employed as Juvenile Corrections Officers (“JCOs”) at the Milwaukee County Juvenile Detention Center (“JDC”). The JDC, part of the Wisconsin State Juvenile Justice System, is a detention facility designed to house temporarily juveniles awaiting juvenile court proceedings.

The statutory mission of the Wisconsin juvenile justice system includes, inter alia, punishment, deterrence, and “the development of competency in the juvenile offender, so that he or she is more capable of living productively and responsibly in the *576community.” Wis. Stat. § 938.01(2)(c). The mission statement of the JDC, in particular, states that the facility is intended:

1. To serve the community by providing care for young people who, because they represent a danger to themselves or others, could not remain safely in the community pending the outcome of a Juvenile Court proceeding.
2. To serve the Juvenile Courts by maintaining close supervision and observation of youngsters and ensuring their appearance in court.
3. To serve the detained youth by providing the best possible care in an atmosphere of fairness and dignity so that their encounter with the Juvenile Justice System can be a positive one.

Tr. Ex. 1021.

Most juveniles at the JDC are detained there for only a short period of time. The average length of stay is ten days, although for those juveniles awaiting their initial court appearances, the average stay is approximately three weeks. Juveniles often are detained for as little as one day. Occasionally, a juvenile may be detained there for as long as one year.

In 1991, Thomas Wanta was appointed superintendent of the JDC. The facility that Mr. Wanta inherited in 1991 was a poorly-maintained, over-crowded detention facility built for indirect (also known as linear) supervision. The juveniles were housed in cells on a long hallway, and JCOs walked the halls monitoring the inmates from the outside without much interaction. Mr. Wanta believed that this indirect form of supervision was a poor way to care for and to rehabilitate juveniles; accordingly, he advanced the idea of a new facility better adapted to a method of supervision in which staff members and inmates would have closer, more direct contact.

As a result of his efforts, a new, nonlinear juvenile detention facility was completed in April 1996. The new JDC contains common rooms, classrooms and recreation rooms where the juveniles spend the majority of their daytime hours. At night, however, the juveniles are confined to their living areas, which are assigned based on their sex, age and classification.

The living areas at the new facility are organized into seven single-sex “pods.” Each can accommodate between 11 and 22 juveniles of the same sex.1 Each pod consists of a number of individual cells, a control center desk from which the staff can monitor the cells and communicate with the pod via intercom, and a common area or “day room” with tables, chairs and a television. The individual cells each contain a bed, a toilet, a desk and a small storage area. The entire cell, including the toilet, is visible from the outside through a window in the cell door.

The juveniles are monitored at all times by JCOs. During the first (morning) and second (evening) shifts, two JCOs are assigned to supervise each pod. During the third (night) shift, when the juveniles are locked inside their cells and generally are asleep, only one JCO is assigned to monitor each pod. In addition to the JCOs assigned to pods, the JDC also has a male and a female “runner” on duty at all times. The runner is a JCO who is responsible for performing the intake procedures for newly-admitted juveniles, including any necessary pat-downs and supervision of showering during the night hours.2 Another staff *577member monitors the central control center at all times.

The advent of the new facility provided an opportunity for Mr. Wanta to shift the JDC’s method of supervision from an indirect model to a direct model and to encourage JCOs to have greater interaction with the juveniles they monitored. Accordingly, in 1997, Mr. Wanta instituted a new role model/mentoring program at the JDC. The staff, including the JCOs, received basic training in mentoring, role modeling and child development in order to equip them to interact more effectively with the juveniles. In furtherance of this program, Mr. Wanta also required that a staff member of the same sex be available on each pod at all times throughout the day and night to mentor the juveniles.3

Prior to the move to the new facility, JCOs had been assigned to shifts without regard to the sex of the officer. Mr. Wan-ta’s new policy, however, required that each pod be staffed at all times by at least one JCO of the same sex as the juveniles housed on the pod. During the day shifts, when two JCOs staffed each pod, one of the two JCOs could be of the opposite sex; however, during the night shifts, when only one JCO staffed each pod, the sole JCO on duty had to be of the same sex as the juveniles in the pod. Because the JDC housed far more male juveniles than female juveniles,4 Mr. Wanta’s same-sex role model/ mentoring policy afforded male JCOs more opportunities for work than those available to female JCOs. The night shift was particularly problematic. It was perceived as the easiest shift; those officers assigned to it received premium pay; and it afforded the most opportunities for overtime.

During the time of their employment as JCOs, Ms. Henry and Ms. Lewis primarily worked one of the day shifts. Prior to 1997, however, they each had earned a substantial amount of additional income from voluntary overtime, predominantly by working the night shift. According to a collective bargaining agreement, voluntary overtime at the JDC traditionally had been apportioned according to seniority. Employees with the most seniority could “put in” for overtime, and they would receive the first opportunities to work their preferred shifts. Ms. Lewis and Ms. Henry were relatively senior employees, and they often were able to work overtime at the old JDC.

After Mr. Wanta instituted the same-sex pod policy, however, far fewer women were allowed to work the third shift because there were far fewer female pods than male pods at the facility. As a result of the same-sex role model/mentoring program, most of the available night shifts with premium pay were reserved for male employees. Female officers like Ms. Henry and Ms. Lewis no longer were able to get the same number of overtime hours as they previously had received. Instead, male employees with less seniority were allowed to work these shifts. Consequently, Ms. Henry and Ms. Lewis received significantly less compensation than they *578had received prior to the institution of the same-sex role model/mentoring program.5

B.

Prior to the institution of this policy, Ms. Henry and Ms. Lewis each had filed an unrelated EEOC complaint alleging harassment and discrimination in the workplace. In October of 1997, they filed additional grievances regarding the sex-specific requirements of the third shift. The plaintiffs allege that they subsequently were treated even more negatively than they had been treated before they filed these complaints.

Ms. Henry testified at trial that her Head of Shift, Bobby Bell, became aggressive and threatening toward her after she began objecting to the JDC’s sex-based shift assignments and to other instances of alleged discrimination. She also stated that she once was removed from the voluntary overtime list for being tardy by just one minute and that she was criticized informally by management for minor things such as eating a sandwich and wearing a sweater at work. She testified that other employees were not disciplined or criticized for such minor infractions. She also noted that her timecard had been written on in red ink, which she found embarrassing, and that occasionally her timecards had been missing altogether— an unusual occurrence.

Ms. Lewis testified that, after filing her EEOC charges, her doctor’s notes were no longer accepted without question for sick leave purposes. Both plaintiffs stated that their managers called them very early in the mornings on weekends and on vacation days when the managers knew that they would be unable to work. They also asserted that, despite repeated requests, they no longer were assigned to work together or to work at the “easier” JCO positions which they preferred. Finally, they testified that they were referred to as “troublemakers” and other derogatory terms by their managers and fellow JCOs.

Ultimately, Ms. Lewis and Ms. Henry decided to take positions in different departments because of this alleged harassment. Although their hourly pay rate was greater in these positions than their hourly rate had been while working as JCOs, their total income was less because they were unable to get overtime pay in these positions.

C.

Ms. Henry and Ms. Lewis brought this action in the district court. They alleged that, in violation of Title VII, they had been denied overtime assignments on the third shift at the JDC because of their sex. They also alleged that they were subjected to a variety of other indignities in the workplace on account of their sex and that the defendants had retaliated against them for filing complaints of discrimination.

Before the district court, the County denied that Ms. Henry and Ms. Lewis had experienced any harassment on account of their sex or in retaliation for their EEOC charges. The County admitted that it had assigned positions on the third shift based on the employees’ sex; however, it asserted as a defense that its sex-based classifications were a bona fide occupational qualification (“BFOQ”) permissible under 42 U.S.C. § 2000e-2(e). It based this defense largely on the testimony of the JDC’s superintendent, Mr. Wanta, who testified *579that, in his professional judgment, a same-sex role model/ mentoring program would best facilitate the rehabilitative goals of the JDC.

After a three-day bench trial, the district court concluded that the same-sex staffing policy on the third shift was a BFOQ. It found that “[t]he essence of the JDC’s business is to ensure and promote the care, rehabilitation, safety and security of the juveniles entrusted to its care.” R.60 at 11. Based on Mr. Wanta’s testimony, it then concluded that “same gender role modeling furthers the twin goals of rehabilitation and security in the juvenile detention setting,” id. at 4, and that “[s]ame gender shift assignments serve to protect the privacy interests of the juvenile detainees,” id. at 5. Finally, relying on general studies that heterosexual assaults are statistically more likely than homosexual assaults, the court found that “[s]ame gender shift assignments also serve the goals of risk management and security.” Id. In the district court’s view, the “alternative to protecting against this risk ... is to hire an additional staff member for each pod on third shift at an approximate cost of $750,000 per year.” Id. Accordingly, it found that gender was a BFOQ for JCOs working the third shift; therefore, the sex-based shift assignments did not violate the anti-discrimination provisions of Title VII.

Regarding the plaintiffs’ other claims, the district court found that the allegedly harassing incidents did not affect the terms and conditions of their employment because they were “trivial inconveniences” that did “not rise to the level of an adverse employment action.” Id. at 7. It also found that the alleged retaliatory actions were not materially adverse actions that would dissuade a reasonable employee from engaging in protected activity. Id. at 8. Finally, the district court found that the plaintiffs had failed to demonstrate any causal connection between their gender or their EEOC charges and the alleged harassment. Id. at 9-10. Accordingly, it denied their claims of discrimination and retaliation under Title VII.

The district court therefore entered judgment in favor of the defendants on May 31, 2007. Ms. Henry and Ms. Lewis timely appealed.

II

DISCUSSION

A.

We begin by setting forth the general principles that must govern our decision. Title VII makes it unlawful for an employer to “limit, segregate, or classify his employees” based on their sex in a way that would adversely affect their employment. 42 U.S.C. § 2000e-2(a)(2). An exception exists, however, when sex is a bona fide occupational qualification (“BFOQ”) that is “reasonably necessary to the normal operation of that particular business or enterprise.” Id. § 2000e-2(e); see also Int’l Union, United Auto. Aerospace & Agric. Implement Workers v. Johnson Controls, Inc., 499 U.S. 187, 201, 111 S.Ct. 1196, 113 L.Ed.2d 158 (1991); Torres v. Wis. Dep’t of Health & Social Servs., 859 F.2d 1523, 1527 (7th Cir.1988) (en banc).

The Supreme Court has cautioned that the BFOQ defense is written narrowly and is to be read narrowly. Int’l Union, 499 U.S. at 201, 111 S.Ct. 1196. It has made clear that the defense is “meant to be an extremely narrow exception to the general prohibition of discrimination on the basis of sex.” Dothard v. Rawlinson, 433 U.S. 321, 334, 97 S.Ct. 2720, 53 L.Ed.2d 786 (1977). In Dothard, the Court quoted approvingly the Fifth Circuit’s formulation in Diaz v. Pan American World Airways, 442 F.2d 385, 388 (5th Cir.1971), that “dis*580crimination based on sex is valid only when the essence of the business operation would be undermined.” Dothard, 433 U.S. at 333, 97 S.Ct. 2720; see also Torres, 859 F.2d at 1527.

Employers bear the burden of establishing the affirmative defense that a particular qualification is a BFOQ. See Meacham v. Knolls Atomic Power Lab., -U.S.-, 128 S.Ct. 2395, 2400-01, 171 L.Ed.2d 283 (2008); Everson v. Mich. Dep’t of Corr., 391 F.3d 737, 748 (6th Cir. 2004). Employers also bear the burden of proving that they could not rearrange job responsibilities or otherwise eliminate the clash between the business necessities and the employment opportunities of female officers. Torres v. Wis. Dept. of Health & Social Servs., 838 F.2d 944, 953 n. 6 (7th Cir.1988) (reheard and reversed on other grounds in Torres, 859 F.2d 1523).6

B.

Before the district court, Milwaukee County relied on a series of cases from this and other circuits that examined whether sex could be a BFOQ for officers in adult female correctional facilities. See Torres, 859 F.2d at 1532; Everson, 391 F.3d at 747-60; Robino v. Iranon, 145 F.3d 1109, 1110-11 (9th Cir.1998); Tharp v. Iowa Dep’t of Corr., 68 F.3d 223, 226 (8th Cir.1995). In each of these cases, the court held that the goals of security, safety, privacy and rehabilitation could, in some circumstances, justify sex-based assignments in female prisons. Milwaukee County submits that, like the sex-specific shift assignments in adult female correctional facilities, the sex-specific assignments at issue here should be upheld because they are necessary to protect the juveniles’ safety and privacy and to further the facility’s rehabilitative goals.

In Torres, we determined that the unique circumstances of the female prison at issue required prison administrators to “innovate and experiment.” 859 F.2d at 1529. In the course of our decision, we noted that, because prison administrators in general face unusually difficult challenges in dealing with the “perplexing sociological problems of how best to achieve the goals of the penal function in the criminal justice system: to punish justly, to deter future crime, and to return imprisoned persons to society with an improved chance of being useful, law-abiding citizens,” prison administrators’ decisions must receive some degree of deference. Id. Accordingly, although the decisions of prison officials are not accorded as much deference in Title YII cases as they are in constitutional cases, their decisions “are entitled to substantial weight when they are the product of a reasoned decision-making process, based on available information and experience.” Id. at 1532 (emphasis added); see also Robino, 145 F.3d at 1110; Everson, 391 F.3d at 750. Milwaukee County contends that, like in Torres, the administrators of the JDC were entitled to substantial deference in their decision to implement a sex-specific policy regarding shift assignments. In its view, the goals and circumstances of the juvenile detention context, when compared to the female corrections context, are equally complex and challenging.

We agree that the administrators of juvenile detention facilities, like the administrators of female correctional facilities, are entitled to substantial deference when fashioning policies to further the goals of the facility. We do not agree, *581however, that the discretion accorded to these individuals in either context is effectively unlimited. A defendant ultimately must introduce sufficient evidence to prove that the administrator’s judgment — that a particular sex classification is reasonably necessary to the normal operation of the institution — is “the product of a reasoned decision — making process, based on available information and experience.” See Torres, 859 F.2d at 1532.

The district court concluded that the JDC’s policy of assigning shifts according to an employee’s sex was based on its administrator’s reasonable belief that the policy would “promote” the goals of rehabilitation, security and privacy. R.60 at 11. All of these functions are, as the district court concluded, essential to the mission of the JDC. However, Title VII’s standard is not satisfied simply because a policy promotes an essential function of an institution. Although sex-based assignments might be helpful in pursuing these goals, in order to satisfy the anti-discrimination strictures of Title VII, Milwaukee County must show that the contested sex classifications are “reasonably necessary.” See Torres, 859 F.2d at 1527 (emphasis in original); see also Kimel v. Fla. Bd. of Regents, 528 U.S. 62, 86, 120 S.Ct. 631, 145 L.Ed.2d 522 (2000) (noting that the “reasonably necessary” standard of the BFOQ defense is “a far cry from the rational basis standard”).

We must conclude that Milwaukee County’s contention that sex-based assignments are reasonably necessary to achieve these goals, at least on the third shift, is not supported by the record before us. The employer, Milwaukee County, has the burden to demonstrate that it could not rearrange job responsibilities to eliminate or minimize the conflict between the inmates’ privacy, security and rehabilitation interests and the employees’ rights under Title VII. See Torres, 838 F.2d at 952 n. 6; Chambers v. Omaha Girls Club, Inc., 834 F.2d 697, 704 (8th Cir.1987). Although reducing the number of opposite-sex staff on the pods may help to promote security, efficient risk management and privacy, Milwaukee County has failed to establish that its policy was reasonably necessary for these goals. We address each proffered justification in turn.

The evidence in the record does not support the conclusion that the juveniles’ safety or security, or the institution’s ability to manage risk effectively, was at all in jeopardy because of the presence of opposite-sex JCOs on the third shift. The district court correctly asserted that heterosexual assaults and misconduct are statistically more likely than homosexual attacks. The record establishes, however, that there has not been a single instance of staff-on-inmate sexual assault at the JDC, on any shift, by either sex; nor has there been a significant problem with false accusations against the staff. Furthermore, other safety precautions, such as door alarms and the presence of supervisors, runners and video cameras, currently are working to prevent actual and alleged security breaches. Although Milwaukee County contends that a staff member may be able to circumvent the alarm system in order to enter a juvenile’s cell at night, the record contains no evidence that this contingency has occurred or was likely to occur at the JDC.

More fundamentally, Milwaukee County offered no reasons why the numerous alternatives to same-sex staffing suggested by the plaintiffs at trial, such as improving the alarm system, installing additional cameras, leaving the doors open between the pods at night or increasing the frequency of supervisor patrols, would not have mitigated any concern. Notably, Mr. Wanta testified at trial that he had not *582investigated the cost of any of these options. He did assert that adding an additional staff member on each of the pods during the third shift would have been prohibitively costly; however, he did not provide any data on this point, and such an augmentation in personnel certainly was not the only option available to minimize the already-minimal risk of staff-on-juvenile sexual assault. The BFOQ defense extends only to those policies that are “reasonably necessary to the normal operation” of the institution, 42 U.S.C. 2000e-2(e)(1). It does not excuse investigation of and reliance upon alternatives that involve minor additional costs or inconveniences.7

Milwaukee County’s proffered privacy justification is even more difficult to justify on the record before us. The record affirmatively shows that the JDC allowed JCOs of the opposite sex to monitor the pods during both of the daytime shifts. It is undisputed that the vast majority of the time that the juveniles were unclothed occurred during these daytime shifts. Showering generally took place during the second shift, when members of the opposite sex were permitted to staff the pods. The only showering that occurred on the third shift was monitored by one of the runners who performed the intake procedures. The juveniles were provided with pajamas, which they were required to wear at night. They changed into this attire on the second shift, and they changed out of it on the first shift — again, while JCOs of the opposite sex were permitted to view them. Although Milwaukee County presented testimony that third-shift JCOs occasionally viewed juveniles using the toilet, masturbating or otherwise acting out sexually, it is undisputed that this situation occurred on the first and second shifts as well.

This situation is therefore very different from cases such as Torres, 859 F.2d at 1530, and Everson, 391 F.3d at 753, which involved the presence of male guards in the housing units of all-female prisons. In Torres, we held that the superintendent reasonably made a “professional judgment that giving women prisoners a living environment free from the presence of males in a position of authority was necessary to foster the goal of rehabilitation.” 859 F.2d at 1530. In Everson, the court determined that, “given the endemic problem of sexual abuse in Michigan’s female facilities” and other unique circumstances, 391 F.3d at 761, the presence of male guards, at any time, in the housing units was a threat to “the security of the prison, the safety of inmates and the protection of the privacy rights of inmates,” id. at 753. In Everson and Torres, prison officials permitted no officers of the opposite sex to guard the living units at any time because the specific needs of the institutions and the prisoners housed in those institutions reasonably required such a policy. The same-sex policy was aimed at a specific condition and was tailored to address that specific condition. The plans were quite clearly “the product of a reasoned decision-making process, based on available information and experience.” Torres, 859 F.2d at 1532. Here, by contrast, we are faced with the fact that the JDC allowed JCOs of the opposite sex to guard the juveniles during those times when the privacy interests of the juveniles were most in jeopardy. Under these circumstances, we cannot say that, with respect to privacy concerns, the same-sex policy is the product of the same sort of comprehensive professional decision-making as exhibited in Torres and Everson. The inconsistencies in imple*583mentation cast a significant doubt on whether the policy is reasonably necessary to achieve the institution’s goal of protecting the privacy interests of the juveniles. Therefore, on this factor as well, the record reveals a failure of proof on the part of Milwaukee County.

The County also contends that same-sex staffing on the third shift is necessary to further the JDC’s mission of rehabilitation. Contrary to the submission of the plaintiffs, we have no doubt that the County is correct in stating that this goal is a very important goal of the institution.8

The record contains substantial testimony from Mr. Wanta on this factor. He described the basis for his determination that a role model/mentoring policy was necessary to the JDC’s rehabilitative efforts. Tr. at 607-19, 634-35, 658-60. He noted that, in formulating the policy, he had relied upon his own experience with juvenile corrections, information learned from his attendance at various seminars and committee meetings, his interviews with the juveniles and staff at the JDC, consultations with experts in the field and professional literature supporting such programs.9 Each of these sources suggested that the direct role modeling/mentoring form of supervision, rather than the indirect form of supervision, was the best available method of providing care and rehabilitation to juveniles in detention facilities. Mr. Wanta, in his professional judgment, concluded that institution of the direct role model/mentoring form of supervision was necessary to achieve the JDC’s mission of rehabilitation. The foundation for his belief is well established in the record.

The conclusion that the effectiveness of these role model/mentoring programs requires the presence of at least one staff member of the same sex as the juveniles in each pod at all times, however, does not find the same strong foundation in the record. Mr. Wanta expressed his belief that same-sex role model/mentoring was more effective than cross-gender programs. He testified that, in formulating this view, he relied upon his own personal experiences, as well as literature on mentoring programs, which “indicate that gender mentoring improves the chances of child behavior changes being positive.” Tr. at 635. He continued: “[A]ll the statistics and the research that I have seen *584indicates that a male mentoring a male, and a female mentoring a female, exponentially increases the chance of success than would cross-gender ... mentoring.” Id.

In support of his belief, Mr. Wanta referenced a report from the Office of Juvenile Justice and Delinquency Prevention (“OJJDP”), Juvenile Mentoring Programs: A Progress Review (Sept.2000), R.48, Ex. 1022. Although this publication was not available at the time of his decision, he suggested that it supported his view that same-sex role modeling was necessary for rehabilitation. This report, however, is specific to one-on-one mentor matching programs for at-risk youth. Moreover, its findings regarding the effect of sex-specific assignments, even in this circumstance, are inconclusive at best. See id. at 5. Mr. Wanta did not explain how anecdotal evidence from these after-school mentoring programs is relevant to the vastly different juvenile detention setting. More importantly, it does not speak to the issue of whether, to achieve success in mentoring, same-sex supervision on the third shift, when the juveniles are isolated and likely asleep in their cells, is reasonably necessary. The other publications admitted into evidence by the County appear to add little to this precise question.10

We are well aware that the professionals who have the great responsibility for running penal institutions need to innovate and experiment if they are to succeed in resolving the crisis in this important area of governance. Indeed, we already have recognized that the inability to proffer solid empirical evidence in support of a particular policy certainly is not fatal to a BFOQ defense. See Torres, 859 F.2d at 1532. Even if we defer to Mr. Wanta’s judgment that a mentoring program is important to the success of juvenile institutions such as the one he manages, and even if we defer to his judgment with respect to the need for same-sex mentoring of juveniles in such an environment, we still must be satisfied in the present litigation that these professional judgments require a rigid rule that such a same-sex mentoring program reasonably necessitates the presence of a JCO of the same sex at all times. Milwaukee County had the responsibility to introduce sufficient evidence in the record to support the conclusion that such same-sex presence at all times was reasonably necessary to meet the institution’s essential goals. Here, the record, although perhaps demonstrating the worth of a mentoring program and the usefulness of mentors of the same sex, does not present a sufficiently strong case with respect to the need for the presence of those mentors seven days a week and twenty-four hours a day.

At trial, Mr. Wanta explained how the JCOs were to act as role models and mentors at the JDC:

On a practical sense, I just wanted a staff member to be a good, positive person. To display positive attitude. To display respect towards the juvenile. Respect towards your peers. Respect toward any individual that came across their area. To model how to interact. Whether interacting with that juvenile, so the other juveniles who observe can *585see that interaction going on and the positive outcomes to that type of interaction. Interaction with your peers in a positive way.... [T]he biggest focus was ... for staff to demonstrate proper behavior.

Tr. at 609. JCOs were not trained or expected to act as counselors; if any serious counseling was necessary, the JCOs were instructed to refer the juvenile to an on-site nurse trained in mental health. Instead, the JCOs carried out their role model/mentoring responsibilities by providing the juveniles with a constant model of proper behavior in their interactions with others.

Given this description of the role of JCOs in the mentoring program, we must conclude that the record does not respond to the question of why a JCO of the same sex is reasonably necessary during the night shift. According to the record, the opportunity for the JCO to interact with the juveniles on the third shift is very minimal. The third shift begins after the juveniles are locked down for the night, and the JCOs on third shift were instructed to encourage them to sleep. Although there was testimony at trial that staff occasionally spoke with juveniles through their doors at night, particularly when they were ill or acting out and it was necessary to calm them down, this interaction was kept to a minimum. There is no evidence that the JCOs ever spoke with the juveniles about confidential counseling matters at night. Indeed, because the other juveniles in the pod would have been able to hear any conversations that occurred, the JCOs testified at trial that they were encouraged to avoid these types of discussions on the third shift.

The County provided no reasons why opposite-sex JCOs were incapable of appropriately interacting with these juveniles to the extent necessary to provide a good behavioral role model on the third shift. Mr. Wanta’s assertion that “consistency” in the same-sex mentoring program requires the presence of someone of the same sex within each pod at all times is simply not justified by the record. Furthermore, the County failed to provide evidence that the many non-discriminatory alternatives proffered by the plaintiffs would have been intolerable here. The plaintiffs questioned Mr. Wanta about numerous other possibilities for encouraging rehabilitation through the availability of same-sex staff, such as hiring more JCOs for the third shift, leaving the doors between pods open or increasing the frequency of supervisor rotations. They also noted that numerous other mechanisms to encourage rehabilitation already were in place at the facility, including educational programs, counselors, guest speakers, community mentors and other programs available to facilitate the JDC’s mission. The County failed to explain why the presence of a same-sex JCO within each pod during the hours that the juveniles were sleeping was reasonably necessary to its rehabilitative efforts.

Accordingly, we must conclude that the County failed to meet its burden to prove that the sex-based classification at issue here was reasonably necessary for the rehabilitation, security or privacy functions of the JDC. Therefore, Milwaukee’s BFOQ defense must fail. The JDC’s third shift policy adversely affected the plaintiffs’ employment. It is undisputed that overtime pay had been a significant and expected component of the plaintiffs’ compensation prior to the institution of the séx-based policy. Not only did the majority of overtime work available occur on the third shift, but the third shift also offered a fifty cent per hour pay premium. Accordingly, the dramatic reduction in the opportunity for women to work on the *586third shift constituted an adverse employment action. See Lewis v. City of Chicago, 496 F.3d 645, 653-54 (7th Cir.2007) (holding that the denial of the opportunity for overtime pay, when that pay is a significant and recurring part of an employee’s total earnings, can constitute an adverse employment action). Because the JDC’s third-shift policy adversely affected the plaintiffs’ employment opportunities, we must conclude that it is in violation of Title VII.

C.

Finally, we turn to the plaintiffs’ workplace harassment and retaliation claims. Both plaintiffs testified regarding a number of incidents that occurred, both before and after they had filed their EEOC charges, which they believed were forms of harassment and retaliation. The district court, however, held a bench trial and determined that these incidents were merely trivial inconveniences and isolated acts that did not rise to the level of harassment necessary to establish a claim under either the discrimination or retaliation provisions of Title VII. The court also found that the plaintiffs did not present any evidence of a causal connection between their sex or their complaints and the alleged harassment. We review these findings of fact for clear error. Fed.R.Civ.P. 52(a); Cerros, 288 F.3d at 1044.

Title VII prohibits sex discrimination in the terms and conditions of employment. Jackson v. County of Racine, 474 F.3d 493, 499 (7th Cir.2007). In order to establish a Title VII claim based on workplace harassment, then, the plaintiffs must prove that a reasonable person would find the alleged harassment to be so severe or pervasive as to create a hostile work environment, thus affecting the terms and conditions of employment. Id.; Kriescher v. Fox Hills Golf Resort & Conf. Ctr. FHR, Inc., 384 F.3d 912, 915 (7th Cir.2004). Furthermore, they must show that there is a “link between this treatment and [their] sex.” Jackson, 474 F.3d at 499.

The analysis regarding the plaintiffs’ retaliation claims is similar. Title VII prohibits employers from discriminating against employees who report or oppose practices made unlawful under Title VII. 42 U.S.C. § 2000e-3(a). The range of conduct prohibited under the anti-retaliation provision is broader than its anti-discrimination provision, however, because “the discriminatory acts proscribed by Title VU’s anti-retaliation provision are not limited to those that affect the terms and conditions of one’s employment.” Lewis, 496 F.3d at 655 (internal quotation marks and citation omitted). Nevertheless, in order to establish a claim for retaliation under Title VII, the plaintiffs must prove that the alleged employment actions were materially adverse such that they would dissuade a reasonable person from engaging in protected activity. Id. Additionally, they must show a causal link between these alleged adverse actions and their protected activity. Basith v. Cook County, 241 F.3d 919, 933 (7th Cir.2001).

After a review of the record, we conclude that the district court did not err in its determination that the alleged incidents did not rise to the level of harassment or retaliation. The Supreme Court has emphasized the necessity of separating “significant from trivial harms.” Burlington N. & Santa Fe Ry. Co. v. White, 548 U.S. 53, 68, 126 S.Ct. 2405, 165 L.Ed.2d 345 (2006). It further cautioned that “[a]n employee’s decision to report discriminatory behavior cannot immunize that employee from those petty slights or minor annoyances that often take place at work.” Id. Title VII’s anti-retaliation provision *587prohibits only those employer actions that are likely to deter victims of discrimination from invoking the Act’s remedial mechanisms. Id. We agree with the district court that the alleged incidents, such as being told not to wear sweaters or eat in front of the juveniles, unspecified “intimidation” and door slamming by the head of shift, missing or marked up time-cards, occasional early morning phone calls, and not being assigned to work together on the same shift or in easier pods are the type of petty slights and minor annoyances that generally will not create such deterrence. Additionally, because these incidents do not rise to the level of an adverse action under the anti-retaliation provision of Title VII, they certainly do not constitute workplace harassment sufficient to establish a discrimination claim. See id. at 67, 126 S.Ct. 2405.

Furthermore, the district court did not err when it concluded that the plaintiffs had failed to prove a causal link between their discrimination complaints and the alleged harassment. The plaintiffs did not even attempt to show that any of the alleged harassment was tied to their sex. As for retaliation, the plaintiffs point only to “suspicious timing” and an allegation that they were referred to as “troublemakers” to suggest that the incidents of which they complain were a result of retaliation for their protected activity. The district court here found that the timing of the incidents in question was not suspicious; the difficulties the plaintiffs had been having with their managers had been ongoing well before they engaged in any protected activity.11 The district court further found that the managers’ alleged use of the term “troublemakers” was not in reference to any of the plaintiff’s protected activities. R.60 at 10. Instead, it found that the plaintiffs in fact had been difficult employees, constantly lodging trivial grievances and generally complaining whenever superiors tried to correct their behavior; the term “troublemaker” had been used in reference to the plaintiffs’ general behavior, not their EEOC complaints. Id. We cannot say that this finding was clearly erroneous.

Conclusion

Accordingly, the judgment of the district court is affirmed in part and reversed in part. The case is remanded to the district court for proceedings consistent with this opinion. The parties shall bear their own costs in this appeal.

Affikmed in part; Reversed and RemaNded in part.

Notes

. One pod at the JDC was reserved for females. The other six pods housed male juveniles.

. Other than these newly-admitted juveniles, all regular showering takes place on the second shift.

. Although the juveniles generally are confined to their cells at night, cell occupants can, and occasionally do, talk with JCOs through their doors. Staff members are permitted to speak with the juveniles through their doors at night; however, they are not permitted to enter the individual cells at night except in an emergency. If they do enter the cells at night, an alarm will sound.

. The ratio of male detainees to female detainees at the JDC was between 4:1 and 6:1 during the relevant time period. Six of the pods at the new JDC were reserved for males; one was reserved for females.

. At the time of trial, despite the efforts of Mr. Wanta, the same-sex overtime policy was not being followed strictly at the JDC. On the advice of counsel, the JDC began allowing females to work in male pods during the third shift, so long as they kept the doors between the pods open. Tr. at 659.

. See also Everson v. Mich. Dept. of Corr., 391 F.3d 737, 749 (6th Cir.2004); Reed v. County of Casey, 184 F.3d 597, 600 (6th Cir.1999); Chambers v. Omaha Girls Club, Inc., 834 F.2d 697, 704 (8th Cir.1987); Edwards v. Dep't of Corr., 615 F.Supp. 804, 809 (M.D.Ala.1985).

. See, e.g., Chambers, 834 F.2d at 701 (noting that "[t]he employer must demonstrate that there is a compelling need to maintain that practice, and the practice cannot be justified by routine business considerations” (internal quotation marks omitted)).

. Ms. Henry and Ms. Lewis take issue with the district court's determination that rehabilitation is part of the "essence of the business” of the JDC. They note that neither the Wisconsin regulations nor the JDC's own mission statement mentions the word "rehabilitation.” They also assert that the juveniles are in the JDC for very brief periods of time, and, unlike in a penal institution, rehabilitation in a short-term detention facility is a futile goal.

Based on its interpretation of the juvenile justice statute and the JDC’s mission statement, as well as on testimony from Mr. Wanta and others, the district court concluded that rehabilitation indeed was part of the essence of the JDC's business. Wisconsin’s juvenile justice code requires the juvenile justice system to take action directed at providing the best possible care for the juveniles and at preventing future delinquent behavior by developing their ability to lead productive and responsible lives. Wis. Stat. § 938.01(2)(c), (f). Wisconsin courts also have recognized the rehabilitative goals of the juvenile justice system. See In re Hezzie R., 219 Wis.2d 848, 580 N.W.2d 660, 668 & n. 4 (1998). Mr. Wanta and other administrators testified that the juveniles' time at the JDC, generally a result of a crisis situation, is often a valuable opportunity for intervention and rehabilitation. Tr. at 623-24. This evidence supports the district court's finding that rehabilitation is part of the "essence” of the JDC.

. See, e.g., Office of Juvenile Justice and Delinquency Prevention ("OJJDP”), Delinquency Prevention: Desktop Guide to Good Juvenile Detention Practice (Roush ed., 1996), R.48, Ex. 1013 (praising the virtues of direct supervision over indirect supervision).

. See OJJDP, Psychiatric Disorders of Youth in Detention (April 2006), R.48, Ex. 1017 (discussing the prevalence of juveniles with mental disorders in juvenile detention facilities); U.S. Dep't of Justice, Bureau of Justice Statistics, Prison Rape Elimination Act of 2003: Sexual Violence Reported by Coirectional Authorities 2004 (July 2005), R.48, Ex. 1015 (discussing nationwide statistics on sexual assault in correctional facilities); OJJDP, Juvenile Mentoring Programs: 1998 Report to Congress (Dec.1998), R.48, Ex. 1023 (discussing mentor matching programs outside the detention context, and noting that most programs use gender as a matching criteria).

. Furthermore, suspicious timing alone is not enough to establish causation. See, e.g., Sauzek v. Exxon Coal USA, Inc., 202 F.3d 913, 918 (7th Cir.2000) (“Speculation based on suspicious timing alone, however, does not support a reasonable inference of retaliation; instead, plaintiffs must produce facts which somehow tie the adverse decision to the plaintiffs’ protected actions.”); Culver v. Gorman & Co., 416 F.3d 540, 546 (7th Cir.2005) (noting that timing “will rarely be sufficient in and of itself to create a triable issue”).



Concurrence Opinion

EASTERBROOK, Chief Judge,

concurring.

The court’s opinion applies Torres v. Wisconsin Department of Health & Social Services, 859 F.2d 1523 (7th Cir.1988) (en banc). I dissented from that decision, because I thought that the majority’s approach could allow an employer to establish a bona fide occupational qualification on the basis of wishful thinking rather than proof. 859 F.2d at 1535-38. After Torres the Supreme Court emphasized that the BFOQ defense requires an employer to prove that “objective, verifiable requirements [that only one sex can fulfil] ... concern job-related skills and aptitudes.” Automobile Workers v. Johnson *588Controls, Inc., 499 U.S. 187, 201, 111 S.Ct. 1196, 113 L.Ed.2d 158 (1991). See also Dothard v. Rawlinson, 433 U.S. 321, 97 S.Ct. 2720, 53 L.Ed.2d 786 (1977). Cf. Johnson v. Phelan, 69 F.3d 144 (7th Cir. 1995). But we need not decide whether Torres requires a fresh look, because the court’s opinion (which I join) shows that the employer’s justifications flunk the Torres standard.

Milwaukee County decided that at least one guard of the same sex as the prisoners is necessary not only for the prison as a whole (to perform body-cavity searches, for example), but also for each pod in the prison. A pod in Milwaukee’s Juvenile Detention Center may contain as few as 11 inmates, and because women occupy only one pod the County’s policy reduced women’s opportunities for employment as guards. The County expresses concern about sexual assaults but admits that no guard has ever sexually assaulted any prisoner in its care — and data from prisons elsewhere do not show that female guards are likely to assault male prisoners sexually. That left the County with stereotypes, such as the proposition that guards of the same sex serve as mentors or role models for the prisoners.

I call this a stereotype because it is based on folk wisdom. It could, in principle, be based on facts, such as proof that recidivism rates fall (or legitimate income after release rises) when a prison has more guards of the inmates’ sex. But the County conceded that it lacked such data when it adopted this policy, and neither expert witnesses nor any published studies supplied an empirical foundation for the policy at trial.

Employers frequently assert that inmates (or students) respond more favorably to guards (or teachers) of their own sex or race. If this sort of justification had been advanced for matching the race of the inmates and the guards (or students and their teachers), courts would not go along. See Wygant v. Jackson Board of Education, 476 U.S. 267, 106 S.Ct. 1842, 90 L.Ed.2d 260 (1986). Cf. United States v. Virginia, 518 U.S. 515, 116 S.Ct. 2264, 135 L.Ed.2d 735 (1996). (Whether short boot-camp programs are exceptional for this purpose, see Wittmer v. Peters, 87 F.3d 916 (7th Cir.1996), is not important to a case that involves the long-term operation of regular prisons.) Why then should courts accept the same sort of “justification” for sex discrimination?

The majority in Torres concluded that Title VII should not be used to block experiments that might lead to the sort of data that would establish a BFOQ. Twenty years have passed since Torres, and Wisconsin’s prisons (like those in other states) have had ample opportunity to try different policies. Other states (and other prisons in Wisconsin) allow people to guard the opposite sex both day and night; data should not be hard to come by. But instead of producing data, the defendants in this case reiterated sexual stereotypes. A court that permits a state (or for that matter a federal agency) to make decisions influenced by intuitions about what the data ultimately will show must insist that the state (or agency) find out whether those intuitions are sound or simply superstitions. See Bechtel v. FCC, 10 F.3d 875 (D.C.Cir.1993). Wisconsin has not made anything of the 20 years’ breathing space provided by Torres, and the time has passed when cheap talk and unverifiable assertions of “professional experience” may substitute for the kind of evidence that Johnson Controls and Dothard contemplate.