Opinion

This case presents the issue whether, under the Tort Claims Act (Gov. Code, §§ 825-825.6, 995-996.6), the County of Santa Clara (the County) must indemnify one of its deputy sheriffs and pay his costs for defending against a sexual harassment lawsuit where the evidence is undisputed that the deputy sheriff lewdly propositioned and offensively touched other deputy sheriffs working at the county jail. We conclude the answer is no. Under the Tort Claims Act, a public entity is required to pay claims and defense costs arising out of a civil lawsuit only where the employee proves that the act or omission giving rise to an injury occurred in “the scope of his or her employment as an employee of the public entity.” (Gov. Code, §§ 825, subd. (a), 825.2, subd. (b), 995; see Gov. Code, § 995.2.) Since the deliberate targeting of an individual employee by another employee for inappropriate touching and requests for sexual favors is not a risk that may fairly be regarded as typical of or broadly incidental to the operation of a county jail, such conduct must be deemed to fall outside the scope of a deputy sheriff’s employment. Consequently, the County is not obligated to indemnify the sexual harasser or his private insurer. We therefore reverse the contrary judgment of the Court of Appeal, and remand the matter with directions to vacate the judgment and to enter judgment in favor of the County.

*998I. Factual and Procedural Background

In 1980, the County promulgated a policy prohibiting sexual harassment in the workplace. At the time of the events underlying this action, the policy provided in pertinent part: “[S]exual harassment constitutes sex discrimination which is prohibited. [^Q Sexual advances, requests for sexual favors, and other verbal or physical conduct of a sexual nature constitute sexual harassment when: FID 1. Submission to such conduct is made either explicitly or implicitly a term or condition of an individual’s employment; FU 2. Submission to or rejection of such conduct by an individual is used or is threatened to be used as the basis for employment decisions affecting such individual, or; [D 3. Such conduct has the purpose or effect of interfering with any individual’s work performance or creating an intimidating, hostile, or offensive working environment.” The Santa Clara County Sheriff’s Department distributed this written policy to its employees and instructed them to study it. The policy was then discussed with employees.

In 1981, the County hired Cynthia Bates and Toni Daugherty as deputy sheriffs in the sheriff’s department.

Between April 1983 and December 1983, Bates and Deputy Sheriff Craig Nelson worked in the North County jail. While there, Nelson made lewd, suggestive and sexually offensive comments to Bates. He asked her about her sex life and made repeated comments about oral sex. Nelson also touched Bates on her legs and thighs.

Between February and June 1984, Bates and Nelson worked in the main jail together. Nelson, who was Bates’s “training officer” during this time, was responsible for evaluating Bates’s progress as a trainee and for informing his supervisors when he thought she was completely trained. At the main jail, Nelson exhibited the following conduct toward Bates: (1) he would stick out his tongue, make gestures with it and say that he “was good at eating pussy and that he knows [Bates] would enjoy it”; (2) he would come up behind Bates and whisper that he would like to take her “to the hot tubs and eat pussy and he’d love to find out what it was like if [Bates] gave him a blow job with [her] braces on”; (3) he commented that he would like to “butt fuck [another female deputy sheriff] and then pull out and come all over her face”; and (4) he told Bates: “I bet you’d like me to fuck you in the butt, I’d bet you’d love that.” Nelson also touched Bates on the back and front of her thighs three or more times. On several occasions he told Bates that in order to “get off training,” she would have to “give him head.” Nelson has admitted that he did and said these things.

*999Also in 1984, Nelson grabbed or slapped Toni Daugherty on the buttocks. Daugherty objected immediately when Nelson touched her, and he did not touch her again. Nelson called Daugherty the next day and asked about the “red marks [he] put on [her] ass.” After Daugherty reported Nelson’s behavior and the sheriff’s department began an internal investigation, she began receiving obscene phone calls at home from Nelson.

When Bates and Daugherty reported Nelson’s conduct to a captain in the sheriff’s department, he instructed them to report the incidents to the internal affairs division.

Another deputy sheriff, Zana Murphy, later reported that Nelson had made lewd and sexually suggestive comments to her as well. In particular, Nelson had discussed oral sex and sodomy with Murphy and wanted to know if she was a “swallower or a spitter.”

After interviewing witnesses, an investigator at the sheriff’s department submitted a detailed report which sustained the allegations of sexual harassment against Nelson. Based on this report, the sheriff’s department suspended Nelson without pay for 14 days. Nelson appealed the discipline pursuant to a collective bargaining agreement, and an arbitrator reduced the suspension to two days.

Additionally, the female deputies complained about alleged harassment by Sergeant David Pascual. They also charged that Lieutenants Larry Kelly and Ernie Ruch and others failed to act timely in investigating the complaints or in taking remedial action to halt the harassment. The County investigated these allegations and concluded there was insufficient evidence to support them.

In 1987, Bates, Daugherty and Murphy sued Nelson, the County and others in the federal district court in San Francisco, alleging, among other things, that Nelson had sexually harassed them in violation of title VII of the Civil Rights Act of 1964 (Title VII) (42 U.S.C. § 2000e-2) and the Fair Employment and Housing Act (FEHA) (Gov. Code, § 12940, subd. (h)). After being served in the federal action, Nelson requested the County to defend and indemnify him pursuant to sections 825 and 995 of the Government Code. The County refused, taking the position that Nelson acted outside the scope of employment when sexually harassing the three deputies.

*1000Nelson was able to obtain counsel paid for by his homeowners insurance carrier, Farmers Insurance Group (Farmers).1

Shortly before trial in the federal action, the federal district court dismissed Murphy’s claims against Nelson as time-barred. On the date set for trial, Nelson settled with Bates and Daugherty for $150,000 and was dismissed from the action. The district court, on Nelson’s motion and without objection .by the County, found that the settlement was made in good faith pursuant to Code of Civil Procedure section 877.6. The settlement amount was paid by Farmers.

The sexual harassment claims against the County and the other individual defendants proceeded to a jury trial, and the jury received instructions on the legal standards applicable to employer liability under FEHA and those applicable to constructive discharge.2 No instructions pertaining to the doctrine of respondeat superior or vicarious liability were given. The jury was directed to award damages, if any, separately against the County and the remaining individual defendants.

The jury found in favor of the female deputies. On the verdict form, the jury answered yes to a question asking if “any plaintiff’ was “a victim of sexual discrimination, sexual harassment, or retaliation in violation of California Government Code § 12,940 [szc]?” The jury also specified on the verdict form that the County had constructively discharged Bates. The jury awarded damages against the County in the amount of $400,000 to Bates, $183,000 to Daugherty and $1.6 million to Murphy.3

In June 1991, after their government claims were rejected, Farmers and Nelson filed the instant action seeking indemnity from the County and others for the amount Farmers had paid in settlement and in defense of the federal action. The parties filed cross-motions for summary judgment or summary adjudication, directed primarily to the issue of whether Nelson was acting within the scope of his employment when he sexually harassed the female deputy sheriffs. The trial court granted the County’s motion and denied that of Nelson and Farmers, finding that Nelson’s conduct was outside the scope *1001of his employment as a matter of law. The court determined that the case was distinguishable from Mary M. v. City of Los Angeles (1991) 54 Cal.3d 202 [285 Cal.Rptr. 99, 814 P.2d 1341] (Mary M.), which held that a city may be vicariously liable for the rape of a motorist committed by an on-duty police officer.

The Court of Appeal reversed. After noting that all of Nelson’s misconduct occurred on the jail premises while the deputies were in uniform and on duty, and that Nelson had authority over Bates and could give direct orders that she had to obey, the court held that the standard for scope of employment had been met. The court concluded that Nelson’s conduct was not so unusual or startling that it would be unfair to include the loss as a cost of the employer’s doing business, and that the authority of a training officer over a subordinate in the employment situation is like that of a police officer over a citizen. The Court of Appeal directed the trial court to vacate the order granting the County’s motion for summary judgment and to enter a new order granting the motion of Nelson and Farmers for summary judgment. We granted the County’s petition for review.

II. Discussion

A. Tort Claims Act

In 1963, the Tort Claims Act was enacted in order to provide a comprehensive codification of the law of governmental liability and immunity in California. (Los Angeles Police Protective League v. City of Los Angeles (1994) 27 Cal.App.4th 168, 174 [32 Cal.Rptr.2d 574].) As part of its overall statutory scheme, the Tort Claims Act provides that in the usual civil case brought against a public employee, a public entity is required to defend the action against its employee (Gov. Code, § 995 et seq.)4 and to pay any claim or judgment against the employee in favor of the third party plaintiff (§ 825 et seq.). A principal purpose of the indemnification statutes is to assure “the zealous execution of official duties by public employees.” (Johnson v. State of California (1968) 69 Cal.2d 782, 792 [73 Cal.Rptr. 240, 447 P.2d 352].)

The requirements pertaining to a public entity’s duty to defend an employee are set forth at sections 995 through 996.6. Section 995 provides in relevant part that, except as otherwise provided, upon request of an employee, “a public entity shall provide for the defense of any civil action or proceeding brought against him, in his official or individual capacity or both, *1002on account of an act or omission in the scope of his employment as an employee of the public entity.” (Italics added.) Where, as here, the public entity refuses to provide the defense as requested and the employee retains his own counsel to defend the action, the employee “is entitled to recover from the public entity such reasonable attorney’s fees, costs and expenses as are necessarily incurred . . . if the action. . . arose out of an act or omission in the scope of his employment as an employee of the public entity, but he is not entitled to such reimbursement if the public entity establishes (a) that he acted or failed to act because of actual fraud, corruption or actual malice . . . .” (§ 996.4, italics added; see also § 995.2, subd. (a) [a public entity may refuse to provide for the defense of a civil action if it determines that the employee’s act or omission was not within the scope of employment, or that the employee acted or failed to act because of actual fraud, corruption or actual malice, or that the public entity’s defense of the action would create a “specific conflict of interest” as defined therein].)

The provisions relating to a public entity’s duty to provide indemnification are addressed at sections 825 through 825.6. At all relevant times herein, section 825, subdivision (a), provided that if the public entity is requested by an employee to defend a civil action against him “for an injury arising out of an act or omission occurring within the scope of his employment as an employee of the public entity,'” and such request is made in writing not less than 10 days before the day of trial and the employee reasonably cooperates in good faith in the defense, “the public entity shall pay any judgment based thereon or any compromise or settlement of the claim or action to which the public entity has agreed.” (Stats. 1985, ch. 1373, § 1, p. 4875, italics added.) If the public entity does not conduct the defense as requested, and the employee pays the claim or judgment against him, the employee may be entitled to recover such payment from the public entity “only if he establishes that the act or omission upon which the claim or judgment is based occurred within the scope of his employment as an employee of the public entity and the public entity fails to establish that he acted or failed to act because of actual fraud, corruption or actual malice or that he willfully failed or refused to conduct the defense of the claim or action in good faith or to reasonably cooperate in good faith in the defense conducted by the public entity.” (§ 825.2, subd. (b), italics added.)

As these statutory provisions make clear, the burden rests upon the public employee to establish that the act or omission was within the scope of employment. (Los Angeles Police Protective League v. City of Los Angeles, supra, 27 Cal.App.4th at p. 176; Rivas v. City of Kerman (1992) 10 Cal.App.4th 1110, 1118-1119 [13 Cal.Rptr.2d 147].)

*1003In this case, the County is not contending that Nelson acted with “actual fraud, corruption or actual malice” within the meaning of the above statutes.5 Nor is the County arguing that Nelson failed or refused to cooperate in defense of the action, or that a specific conflict of interest existed. Rather, the County contends it had no statutory duty to defend or indemnify Nelson because his acts of sexual harassment were outside the scope of his employment as a matter of law. For the reasons below, we conclude that the Court of Appeal erred in determining this issue adversely to the County.

B. Scope of Employment

As used in the Tort Claims Act, “[t]he phrase ‘scope of his employment’ is intended to make applicable the general principles that the California courts use to determine whether the particular kind of conduct is to be considered within the scope of employment in cases involving actions by third persons against the employer for the torts of his employee.” (4 Cal. Law Revision Com. Rep. (Dec. 1963) p. 814, fti. 3.)6

In Perez v. Van Groningen & Sons, Inc. (1986) 41 Cal.3d 962 [227 Cal.Rptr. 106, 719 P.2d 676] (Perez), we explained scope of employment principles under the respondeat superior doctrine as follows: “[A]n employer is liable for risks ‘arising out of the employment.’ [Citations.] [<fl] A risk arises out of the employment when ‘in the context of the particular enterprise an employee’s conduct is not so unusual or startling that it would seem unfair to include the loss resulting from it among other costs of the employer’s business. [Citations.] In other words, where the question is one of vicarious liability, the inquiry should be whether the risk was one “that may fairly be regarded as typical of or broadly incidental” to the enterprise undertaken by the employer. [Citation.]’ [Citation.] Accordingly, the employer’s liability extends beyond his actual or possible control of the employee to include risks inherent in or created by the enterprise.” (Perez, supra, 41 Cal.3d at p. 968, italics added [employer vicariously liable for injuries sustained by plaintiff when he was knocked from a tractor driven by employee while disking employer’s orchard].) These principles were reiterated in Mary M., supra, 54 Cal.3d at page 209.

As the Court of Appeal elaborated in Rodgers v. Kemper Constr. Co. (1975) 50 Cal.App.3d 608, 618-619 [124 Cal.Rptr. 143] (Rodgers): “One *1004way to determine whether a risk is inherent in, or created by, an enterprise is to ask whether the actual occurrence was a generally foreseeable consequence of the activity. However, ‘foreseeability’ in this context must be distinguished from ‘foreseeability’ as a test for negligence. In the latter sense ‘foreseeable’ means a level of probability which would lead a prudent person to take effective precautions whereas ‘foreseeability’ as a test for respondeat superior merely means that in the context of the particular enterprise an employee’s conduct is not so unusual or startling that it would seem unfair to include the loss resulting from it among other costs of the employer’s business. [Citations.]” (Italics added.) We find the Rodgers foreseeability test useful because it reflects the central justification for respondeat superior: that losses fairly attributable to an enterprise—those which foreseeably result from the conduct of the enterprise—should be allocated to the enterprise as a cost of doing business. (John R. v. Oakland Unified School Dist. (1989) 48 Cal.3d 438, 450 [256 Cal.Rptr. 766, 769 P.2d 948] (John R.); Hinman v. Westinghouse Elec. Co. (1970) 2 Cal.3d 956, 959-960 [88 Cal.Rptr. 188, 471 P.2d 988].)

In California, the scope of employment has been interpreted broadly under the respondeat superior doctrine. For example, “[t]he fact that an employee is not engaged in the ultimate object of his employment at the time of his wrongful act does not preclude attribution of liability to an employer.” (Alma W. v. Oakland Unified School Dist. (1981) 123 Cal.App.3d 133, 139 [176 Cal.Rptr. 287] (Alma W.).) Thus, acts necessary to the comfort, convenience, health, and welfare of the employee while at work, though strictly personal and not acts of service, do not take the employee outside the scope of employment. (See ibid.) Moreover, “ ‘where the employee is combining his own business with that of his employer, or attending to both at substantially the same time, no nice inquiry will be made as to which business he was actually engaged in at the time of injury, unless it clearly appears that neither directly nor indirectly could he have been serving his employer.’ [Citations.]” (John R., supra, 48 Cal.3d at p. 447.) It is also settled that an employer’s vicarious liability may extend to willful and malicious torts of an employee as well as negligence. (Mary M., supra, 54 Cal.3d at p. 209; John R., supra, 48 Cal.3d at p. 447.) Finally, an employee’s tortious act may be within the scope of employment even if it contravenes an express company rule and confers no benefit to the employer. (Mary M., supra, 54 Cal.3d at p. 209; Perez, supra, 41 Cal.3d at pp. 969-970.)

Notwithstanding the generally broad view given to scope of employment determinations, the law is clear that an employer is not strictly liable for all actions of its employees during working hours. Significantly, an *1005employer will not be held vicariously liable for an employee’s malicious or tortious conduct if the employee substantially deviates from the employment duties for personal purposes. (Alma W., supra, 123 Cal.App.3d at p. 139; see John R., supra, 48 Cal.3d at p. 447; Hinman v. Westinghouse Elec. Co., supra, 2 Cal.3d at p. 960; Jeffrey E. v. Central Baptist Church (1988) 197 Cal.App.3d 718, 721 [243 Cal.Rptr. 128] (Jeffrey E.).) Thus, if the employee “inflicts an injury out of personal malice, not engendered by the employment” (Carr v. Wm. C. Crowell Co. (1946) 28 Cal.2d 652, 656 [171 P.2d 5]) or acts out of “personal malice unconnected with the employment” (Rodgers, supra, 50 Cal.App.3d at p. 621), or if the misconduct is not an “outgrowth” of the employment (Carr v. Wm. C. Crowell Co., supra, 28 Cal.2d at p. 657), the employee is not acting within the scope of employment. Stated another way, “[i]f an employee’s tort is personal in nature, mere presence at the place of employment and attendance to occupational duties prior or subsequent to the offense will not give rise to a cause of action against the employer under the doctrine of respondeat superior.” (Alma W., supra, 123 Cal.App.3d at p. 140.) In such cases, the losses do not foreseeably result from the conduct of the employer’s enterprise and so are not fairly attributable to the employer as a cost of doing business.

To aid us in our application of these principles, we find it helpful to compare the types of situations in which the respondeat superior doctrine has and has not been applied.

Our review of the case law discloses that an employer may be subject to vicarious liability for injuries caused by an employee’s tortious actions resulting or arising from pursuit of the employer’s interests. (E.g., Perez, supra, 41 Cal.3d 962 [tractor operator carried unauthorized passenger while serving the employer’s business]; De Rosier v. Crow (1960) 184 Cal.App.2d 476 [7 Cal.Rptr. 540] [waitress employed by bowling alley/ liquor bar attempted to stop fight involving patrons and owner of bowling alley/bar]; Caldwell v. Farley (1955) 134 Cal.App.2d 84 [285 P.2d 294] [union steward struck union member who expressed opinion against strike]; Sullivan v. Matt (1955) 130 Cal.App.2d 134 [278 P.2d 499] [railroad superintendent, acting to further the interests of his company, assaulted yardman for attentions to superintendent’s secretary]; Stansell v. Safeway Stores, Inc. (1941) 44 Cal.App.2d 822 [113 P.2d 264] [assault during dispute with customer over an order]; Pritchard v. Gilbert (1951) 107 Cal.App.2d 1 [236 P.2d 412] [traveling salesman, while driving car on employer’s business, lost temper and beat motorist over near accident]; Martin v. Leatham (1937) 22 Cal.App.2d 442 [71 P.2d 336] [private detective, hired to maintain order in skating rink, engaged in altercation with patron seeking admission, and shot *1006decedent, who had intervened to stop the fight].) Vicarious liability may also be proper where the tortious conduct results or arises from a dispute over the performance of an employee’s duties, even though the conduct is not intended to benefit the employer or to further the employer’s interests. (E.g., Fields v. Sanders (1947) 29 Cal.2d 834 [180 P.2d 684, 172 A.L.R. 525] [employee truck driver beat motorist with wrench during dispute over employee’s driving on a company job]; Carr v. Wm. C. Crowell Co., supra, 28 Cal.2d 652 [employee of general contractor threw hammer at subcontractor during dispute over construction procedure].) Vicarious liability may even be appropriate for injuries caused after work hours where a dispute arises over the rights and privileges of off-duty employees. (Rodgers, supra, 50 Cal.App.3d 608 [injuries inflicted by off-duty employees of general contractor during dispute over right to use subcontractor’s equipment].) In these types of situations, the tortious actions are engendered by events or conditions relating to the employment and therefore are properly allocable to the employer.

Conversely, vicarious liability is deemed inappropriate where the misconduct does not arise from the conduct of the employer’s enterprise but instead arises out of a personal dispute (e.g., Monty v. Orlandi (1959) 169 Cal.App.2d 620, 624 [337 P.2d 861] [bar owner not vicariously liable where on-duty bartender assaulted plaintiff in the course of a personal dispute with his common law wife]), or is the result of a personal compulsion (e.g., Thorn v. City of Glendale (1994) 28 Cal.App.4th 1379, 1383 [35 Cal.Rptr.2d 1] [city not vicariously liable where fire marshal set business premises on fire during an inspection]). In such cases, the risks are engendered by events unrelated to the employment, so the mere fact that an employee has an opportunity to abuse facilities or authority necessary to the performance of his or her duties does not render the employer vicariously liable. (See Alma W., supra, 123 Cal.App.3d at p. 140.)

In a context more analogous to this case, several decisions have addressed whether an employee’s sexual misconduct directed toward a third party is within the scope of employment for respondeat superior purposes. Those cases hold that, except where sexual misconduct by on-duty police officers against members of the public is involved (e.g., Mary M., supra, 54 Cal.3d 202; White v. County of Orange (1985) 166 Cal.App.3d 566 [212 Cal.Rptr. 493]), the employer is not vicariously liable to the third party for such misconduct (e.g., Jeffrey E., supra, 197 Cal.App.3d 718 [church not liable for repeated acts of sexual assault on minor by Sunday school teacher]; Rita M. v. Roman Catholic Archbishop (1986) 187 Cal.App.3d 1453 [232 Cal.Rptr. 685] (Rita M.) [Roman Catholic archbishop not liable for seduction *1007of parishioner by priests]; Alma W., supra, 123 Cal.App.3d 133 [school district not liable for janitor’s rape of student]). In those decisions, vicarious liability was rejected as a matter of law because it could not be demonstrated that the various acts of sexual misconduct arose from the conduct of the respective enterprises. In particular, the acts had been undertaken solely for the employees’ personal gratification and had no purpose connected to the employment. Moreover, the acts had not been engendered by events or conditions relating to any employment duties or tasks; nor had they been necessary to the employees’ comfort, convenience, health, or welfare while at work. Similarly, in John. R., supra, 48 Cal.3d at page 452, we concluded that “the connection between the authority conferred on teachers to carry out their instructional duties and the abuse of that authority to indulge in personal, sexual misconduct is simply too attenuated to deem a sexual assault as falling within the range of risks allocable to a teacher’s employer.”

With the foregoing principles and case law in mind, we turn our attention to the facts of the instant case. In arguing that Nelson’s repeated acts of sexual harassment were within the scope of his employment, Farmers7 places considerable emphasis on the undisputed evidence that most of the harassment took place on the jail premises during work hours while the deputies were on duty. We are not persuaded.

Even though Farmers has shown that Nelson committed virtually all of the harassing acts during his work hours at the jail, Farmers cannot prevail on the scope of employment issue without also establishing that the acts arose out of the employment. As explained above, “[i]f an employee’s tort is personal in nature, mere presence at the place of employment and attendance to occupational duties prior or subsequent to the offense will not give rise to a cause of action against the employer under the doctrine of respondeat superior.” (Alma W., supra, 123 Cal.App.3d at p. 140; see Monty v. Orlandi, supra, 169 Cal.App.2d at p. 623; cf. John R., supra, 48 Cal.3d 438.)

On this point, Farmers does not dispute that Nelson’s repeated requests for sexual favors and his inappropriate touchings were motivated for strictly personal reasons unrelated to the guarding of inmates or the performance of any other duty of a deputy sheriff at a county jail. Furthermore, Nelson’s misconduct was not reasonably necessary to his comfort, convenience, health, and welfare while at work. Nor was it precipitated by a work-related dispute over the performance of his duties or those of his victims. Indeed, Nelson’s actions were in direct violation of the County’s policy prohibiting sexual harassment in the workplace, and the County sought to discipline him *1008for his transgressions. While the scope of employment may encompass tortious conduct that disregards the employer’s express orders (Mary M., supra, 54 Cal.3d at p. 209; Perez, supra, 41 Cal.3d at pp. 969-970), an employer will not be held vicariously liable where, as here, “ ‘it clearly appears that neither directly nor indirectly could [the employee] have been serving his employer.’ ” (John. R., supra, 48 Cal.3d at p. 447).

Farmers next argues that Nelson’s actions meet the scope of employment test because many other decisions have held employers vicariously liable for far more serious physical injuries caused by misconduct far more egregious and shocking. (E.g., Fields v. Sanders, supra, 29 Cal.2d 834; Carr v. Wm. C. Crowell Co., supra, 28 Cal.2d 652; De Rosier v. Crow, supra, 184 Cal.App.2d 476; Caldwell v. Farley, supra, 134 Cal.App.2d 84; Sullivan v. Matt, supra, 130 Cal.App.2d 134; Pritchard v. Gilbert, supra, 107 Cal.App.2d 1; Rodgers, supra, 50 Cal.App.3d 608; Stansell v. Safeway Stores, Inc., supra, 44 Cal.App.2d 822; Martin v. Leatham, supra, 22 Cal.App.2d 442.)

This argument misses the mark. Each one of those decisions involved an assault precipitated by a work-related dispute, thus clearly illustrating the principle that the tortious act must arise out of the employment. The misbehavior here, which had nothing to do with the work performed by Nelson or his victims, stands in sharp contrast to the conduct in those cases.

Farmers additionally argues that vicarious liability is proper based upon our statement in Perez, supra, 41 Cal.3d at page 968, that “[a] risk arises out of the employment when ‘in the context of the particular enterprise an employee’s conduct is not so unusual or startling that it would seem unfair to include the loss resulting from it among other costs of the employer’s business. . . .’ ” (See also Mary M., supra, 54 Cal.3d at p. 209.) Relying on this language, Fanners vigorously contends that, as a general matter, sexual harassment is foreseeable and cannot be viewed as unusual or startling in any workplace.

In support of this point, Farmers first refers us to several recent examples of highly publicized stories of harassment and to various treatises citing a number of studies and surveys indicating that on-the-job sexual harassment is pervasive. (See 1 Conte, Sexual Harassment in the Workplace (2d ed. 1994) pp. 1-2; Lindemann & Kadue, Sexual Harassment in Employment Law (1992) pp. 4-5 & fns. 12-18.) Farmers also relies upon Capitol City Foods, Inc. v. Superior Court (1992) 5 Cal.App.4th 1042 [7 Cal.Rptr.2d 418], which suggested that a legislative declaration contained in the 1984 amendment of the FEHA demonstrates the foreseeability of sexual harassment in the workplace, and Carr v. Allison Gas Turbine Div. Gen. Motors *1009(7th Cir. 1994) 32 F.3d 1007, in which Chief Judge Posner of the United States Court of Appeals for the Seventh Circuit acknowledged the propensity of male employees to sexually harass their female coworkers in a newly integrated work environment such as the one here. We are not convinced.

While it is no doubt true that sexual harassment is a pervasive problem and that many workers in many different fields of employment have experienced some form of uninvited and unwanted sexual attention, this argument stretches the respondeat superior foreseeability concept beyond its logical limits. As our decisions explain, in determining whether a risk is “unusual or startling” for respondeat superior purposes, “ ‘the inquiry should be whether the risk was one “that may fairly be regarded as typical of or broadly incidental” to the enterprise undertaken by the employer.’ ” (Perez, supra, 41 Cal.3d at p. 968, italics added; see Hinman v. Westinghouse Elec. Co., supra, 2 Cal.3d at p. 960.) Thus, it is not enough that a risk be neither unusual nor startling as a general matter; rather, the risk must be evaluated in the context of the employer’s particular enterprise. (Ibid.; Rodgers, supra, 50 Cal.App.3d at pp. 618-619.) Under the foregoing standard, we are compelled to conclude that evidence of the general prevalence of sexual harassment in workplaces and in newly integrated work environments has little, if any, probative value in determining whether lewd propositioning and offensive touchings of coworkers are typical of or broadly incidental to the particular enterprise here—a county jail.

Moreover, we cannot agree that Capitol City Foods, Inc. v. Superior Court, supra, 5 Cal.App.4th 1042, stands for the proposition that the Legislature, in declaring a policy against sexual harassment in the workplace in its 1984 amendment of the FEHA, has indicated such conduct ordinarily is foreseeable in the respondeat superior sense. Initially we note that the Court of Appeal’s remarks to that effect were based in part upon a concession by the employer. (See 5 Cal.App.4th at p. 1048.) More importantly, the proposition finds no support in the one legislative declaration cited in the opinion (ibid., citing Stats. 1984, ch. 1754, pp. 6403-6404) or in the actual statutory provisions of the FEHA.

The comment gleaned from Carr v. Allison Gas Turbine Div. Gen. Motors, supra, 32 F.3d at page 1012, also fails to persuade us otherwise. In that case, Chief Judge Posner declared: “General Motors was astonishingly unprepared to deal with problems of sexual harassment, foreseeable though they are when a woman is introduced into a formerly all-male workplace.” (32 F.3d at p. 1012, italics added.) But Chief Judge Posner made that reference to foreseeability in the context of analyzing the employer’s negligence in *1010failing to take appropriate responsive action in the face of sexual harassment complaints by the plaintiff. (Ibid.) Because it is abundantly clear that foreseeability in the respondeat superior context is distinct from the negligence test for foreseeability (Rodgers, supra, 50 Cal.App.3d at pp. 618-619; see also Rita M., supra, 187 Cal.App.3d at p. 1461; Alma W., supra, 123 Cal.App.3d at p. 142), Chief Judge Posner’s statements prove unhelpful to Farmers’s position.

Likewise, the analysis offered by Justice Mosk fails to convince us. To justify the conclusion that lewd propositioning and offensive touching fall within the scope of employment at a county jail as a matter of law, Justice Mosk relies in part upon a number of cases reflecting sexual harassment lawsuits in police and jail settings. These cases are cited, Justice Mosk says, “not for their law but for their facts, i.e., to show the frequency with which women police and correctional officers complain of sexual harassment by fellow officers or superiors.” (Dis. opn. of Mosk, J., post, at p. 1032.)

Unlike Justice Mosk, we do not believe it appropriate to rely upon a survey of published cases to establish that sexual harassment in police and jail settings is an inherent risk for respondeat superior purposes. First and foremost, Justice Mosk’s approach ignores a basic tenet of respondeat superior law, i.e., that for the doctrine to apply, the tortious act must arise out of the employment.8 Second, Justice Mosk cites no authority to support his unusual approach, and the approach offers no principled basis for determining when a particular type of occurrence may be deemed to constitute an inherent risk. (See conc. opn. of Werdegar, J., post, at pp. 1023-1024.) Third, applying the same type of approach in other contexts, we would be forced to conclude, in direct conflict with John R., supra, 48 Cal.3d 438, and numerous other California and out-of-state decisions, that sexual molestation by teachers and clergy is an inherent risk of schools and religious institutions simply because of the frequency with which such misconduct is claimed.

Furthermore, while Justice Mosk prefers not to discuss the legal analyses of his cases (virtually all of which involved claims under Title VII and other federal acts), a review discloses that none of them stands for the proposition that sexual harassment is within the scope of a police or correctional officer’s employment under the common law. Indeed, at least two of the decisions appear to reject that notion. (See Hirschfeld v. New Mexico Corrections Dept. (10th Cir. 1990) 916 F.2d 572, 576-577; Ball v. City of Cheyenne, Wyo. (D.Wyo. 1993) 845 F.Supp. 803, 809.) For all of the *1011foregoing reasons, we regard Justice Mosk’s analysis and conclusions as unsound.

Both Farmers and Justice Mosk additionally rely upon particular evidence in the record to argue that sexual harassment of female deputy sheriffs was foreseeable because profanity and sexually explicit language and banter were common at this particular county jail, especially in 1983 and 1984 when the workforce was first integrated. Specifically, both Nelson and a lieutenant named Armand Tiano submitted declarations stating that profanity, sexually explicit language, banter and horseplay were extremely common among coworkers and peers at the county jail. Additionally, Sergeant Pascual testified that jails are “vulgar places where there is talk of everything,” and that the vulgar and suggestive language used by Nelson was “not uncommon” and “just regular jail talk.” This argument is without merit.

Even if the evidence shows that the use of profanity and sexually explicit language was not uncommon at this particular county jail, it still falls far short of establishing that serious misconduct such as asking individual employees for sexual favors and targeting those individuals for inappropriate touching is either typical of or broadly incidental to the operation of a county jail or to the duties and tasks of deputy sheriffs at such a jail. (See Perez, supra, 41 Cal.3d at p. 968; Hinman v. Westinghouse Elec. Co., supra, 2 Cal.3d at p. 960.)

Moreover, factors that might be relevant to whether the County itself acted negligently are not relevant to whether the County should be vicariously liable for an employee’s misconduct regardless of its own fault. (John R., supra, 48 Cal.3d at p. 450, fn. 9; see also 48 Cal.3d at p. 451, fn. 10; Hinman v. Westinghouse Elec. Co., supra, 2 Cal.3d at p. 960 [in making respondeat superior determination, “ ‘we are not looking for that which can and should reasonably be avoided, but [for] the more or less inevitable toll of a lawful enterprise’ ”].) Accordingly, even assuming arguendo that the usage of profanity and crude language at the jail should have put the County on notice that Nelson’s actions were “foreseeable” in a negligence sense despite the absence of a causal link between the acts of sexual harassment and Nelson’s work as a deputy sheriff, that is a matter lacking relevance in scope of employment analysis.

Alternatively, Farmers contends that, at least with respect to Bates, Nelson’s misconduct occurred in large part because he was her training officer, and was able to abuse the supervisory authority conferred upon him by the *1012County.9 By virtue of his position, Fanners argues, Nelson had authority or apparent authority to commit quid pro quo harassment.10 (This evidently refers to when Nelson told Bates he would not allow her to “get off training” unless she consented to “give him head.”) In Farmers’s view, this case is comparable to Mary M., supra, 54 Cal.3d 202, which held that a city may be vicariously liable where an on-duty police officer abused his official authority and raped a woman after detaining her for a traffic stop.

The attempted analogy to Mary M. fails. As Farmers concedes, Nelson did not act as a “supervisor” for two of his three victims (Daugherty and Murphy). As for Bates, the undisputed evidence shows that Nelson was her training officer from March 1984 to June 1984 and that during this time Nelson told her several times that she would have to “give him head” in order to get off training. But the undisputed evidence additionally reflects that Nelson had harassed and propositioned Bates previously, between April 1983 and December 1983, when he had no supervisory authority over Bates but was merely her coworker.

Even if we focus solely on the period of time when Nelson was Bates’s supervisor, the work-related authority of a supervisor over a trainee employee in a county sheriff’s department is in no way comparable to the extraordinary power and authority that police officers exercise over members of the public. As emphasized in Mary M., police officers occupy a unique position of trust in our society. They are given the authority to detain, to arrest and to use deadly force if necessary. When officers abuse their authority by committing crimes against members of the community, they violate the public trust and may erode the community’s confidence in the integrity of its police force. (54 Cal.3d at pp. 206-207.) Plainly there is no parallel between the supervisory authority in the instant case and the formidable, official authority at issue in Mary M.

Moreover, Mary M., supra, 54 Cal.3d 202, did not suggest that an employer may be vicariously liable for an employee’s misconduct whenever there is an abuse of a job-created, hierarchical relationship in which the *1013employee is afforded a high degree of authority over the victim. Stressing quite clearly that its conclusion “flows from the unique authority vested in police officers” (54 Cal.3d at p. 218, fn. 11), Mary M. distinguished but did not purport to overrule previous cases rejecting the application of respondeat superior for misconduct occurring in relationships of a hierarchical nature where, at least in the eyes of the victim, the wrongdoer’s authority might be considered very great. (E.g., John R., supra, 48 Cal.3d at p. 452 [connection between the extensive authority conferred on teachers to carry out their instructional duties and the abuse of that authority to indulge in personal, sexual misconduct is simply too attenuated to deem a sexual assault as falling within the range of risks allocable to a teacher’s employer]; Jeffrey E., supra, 197 Cal.App.3d at p. 723 [Sunday school teacher’s sexual molestation of minor not equivalent to police officer’s abuse of official authority over members of the public]; Rita M., supra, 187 Cal.App.3d 1453 [seduction of parishioner by seven priests not foreseeable in the sense required for respondeat superior]; see also Virginia G. v. ABC Unified School Dist. (1993) 15 Cal.App.4th 1848, 1854-1855 [19 Cal.Rptr.2d 671] [conduct of teachers who sexually molest students under their supervision will not be imputed to school district, though districts may be liable for their own negligence in hiring and supervising teachers].) Even though the above authorities did not involve misconduct by supervisors against trainee employees, they nonetheless support the conclusion that, for purposes of respondeat superior, employees do not act within the scope of employment when they abuse job-created authority over others for purely personal reasons.

Finally, we consider whether imposing vicarious liability would further the three policy justifications for the respondeat superior doctrine: (1) to prevent recurrence of the tortious conduct; (2) to give greater assurance of compensation for the victim; and (3) to ensure that the victim’s losses will be equitably borne by those who benefit from the enterprise that gave rise to the injury. (Mary M., supra, 54 Cal.3d at p. 209; John R., supra, 48 Cal.3d at pp. 450-452.) For the reasons that follow, we find that, on balance, the underlying purposes of the doctrine do not support its application in this case.

Of the three policy justifications, the first recognizes that “imposing liability on the employer may prevent recurrence of the tortious conduct, because it ‘creates a strong incentive for vigilance by those in a position “to guard substantially against the evil to be prevented.” ’ ” (Mary M., supra, 54 Cal.3d at p. 214.) In John R., supra, we first recognized that prevention and *1014deterrence “play[] little role in the allocation of responsibility for the sexual misconduct of employees generally . . . .” (48 Cal.3d at p. 451.) We then determined that the imposition of vicarious liability for sexual misconduct of a student by a teacher “would be far too likely to deter [school] districts from encouraging, or even authorizing, extracurricular and/or one-on-one contacts between teachers and students or to induce districts to impose such rigorous controls on activities of this nature that the educational process would be negatively affected.” (Ibid., fn. omitted.) Conversely, in Mary M., supra, it was concluded that the goal of prevention would be furthered in a case of police rape , because “[t]here is little or no risk that preventive measures would significantly interfere with the ability of police departments to enforce the law and to protect society from criminal acts.” (54 Cal.3d at p. 215.)

In this case, we find it significant that public entities such as the County are already required by the FEHA to “take all reasonable steps to prevent harassment from occurring.”11 (§ 12940, subd. (h)(1).) The FEHA makes it unlawful for public entities and any persons acting as their agents to sexually harass any employee or applicant, and for public entities to fail to take immediate and appropriate corrective action if they know or should know of sexual harassment by employees other than agents or supervisors. (§ 12940, subd. (h)(1).) Public entities may be directly liable to sexually harassed employees for compensatory damages in civil actions under the FEHA.12 (See Commodore Home Systems, Inc. v. Superior Court (1982) 32 Cal.3d 211, 221 [185 Cal.Rptr. 270, 649 P.2d 912].) Title VII also prohibits *1015sexual harassment in the workplace. (Meritor Savings Bank v. Vinson (1986) All U.S. 57 [91 L.Ed.2d 49, 106 S.Ct. 2399].)13

Because this potential for direct liability already furnishes powerful motivation for the County to establish and maintain programs and procedures designed to eliminate sexual harassment of employees at the jail, the imposition of vicarious liability is not essential to “ ‘create[] a strong incentive for vigilance by those in a position “to guard substantially against the evil to be prevented.” ’ ” (Mary M., supra, 54 Cal.3d at p. 214.) Indeed, since the FEHA already requires the County to “take all reasonable steps to prevent harassment from occurring” (§ 12940, subd. (h)(1), italics added), the imposition of liability under respondeat superior might, in effect, simply create an incentive to adopt any and all possible preventive measures, irrespective of reasonableness or probable effectiveness. In our view, the goal of deterring sexual harassment would not and should not be advanced by encouraging precautionary measures that are unreasonable or of marginal benefit.

In considering this policy justification, we recognize that the instant case, which concerns a public employee’s right to statutory indemnification, comes to us in a posture different from the usual case in which an injured third party seeks to hold a public entity vicariously liable for an employee’s tort. There is nothing about the particular context of this case, however, that detracts from the above analysis. If anything, deterrence objectives are better served by denying sexual harassers the right to indemnity than by insulating them from financial responsibility for their own misconduct.

In sum, no one would dispute that the prevention of sexual harassment in the workplace is of utmost importance. But subjecting the County to vicarious liability simply is unnecessary as an incentive “to guard substantially against the evil to be prevented” (Mary M., supra, 54 Cal.3d at p. 214), and is just as likely to accomplish more harm than good. We therefore conclude that the concerns of prevention and deterrence do not support a finding that Nelson’s misconduct was within the scope of his employment.

*1016The second policy justification for the application of respondeat superior is to give greater assurance of compensation to the victim. (Mary M., supra, 54 Cal.3d at p. 209.) As pointed out above, the County may be directly liable to sexual harassment victims in civil actions under the FEHA and Title VII. Thus, regardless of the outcome here, deputy sheriffs who have been sexually harassed may presently secure compensation directly from the County in many types of situations.14 Even though the imposition of vicarious liability might additionally assure compensation to victims of sexual harassment where the FEHA would not—i.e., where the harassment is by a nonsupervising coworker and the County, its agents and supervisors do not act unlawfully or unreasonably—it might also, under the reasoning of John R., supra, 48 Cal.3d at page 451, “tend to make insurance, already a scarce resource, even harder to obtain, and could lead to the diversion of needed funds from [the oversight of inmates and jail security] to cover claims.” Moreover, the particular context of this case fails to aid Farmers in our policy analysis since indemnification of Nelson under sections 825.2 and 995 is unnecessary to assure compensation to the victims. On balance, we find this second factor to be, at best, neutral in our determination.

The third policy justification to consider is whether the application of respondeat superior would ensure that the victim’s losses will be equitably borne by those who benefit from the enterprise that gave rise to the injury. (Mary M., supra, 54 Cal.3d at p. 216.) In Mary M., supra, where the awesome and dangerous power delegated to police officers was involved, it was concluded that “[t]he cost resulting from misuse of that power should be borne by the community, because of the substantial benefits that the community derives from the lawful exercise of police power.” (Id. at p. 217, italics added.)

*1017Here, the third justification for imposing vicarious liability is discemibly weaker. Significantly, Nelson exercised no job-conferred authority over Daugherty or Murphy at the time he harassed them, and had no authority over Bates when he initially harassed her in 1983. And though Nelson continued to harass Bates after becoming her training officer in 1984, his supervisory authority did not include the extraordinary power to detain, arrest or jail his trainee. Thus, even if we were to consider only the misconduct Nelson committed as a supervisor, this case does not implicate “the considerable authority and control inherent in the responsibilities of an officer in enforcing the law” or the “substantial benefits” that the community derives therefrom.15 (Mary M., supra, 54 Cal.3d at p. 218.) We therefore conclude that the connection between Nelson’s duties and his deliberate targeting of the three women for sexual harassment was “simply too attenuated” to be deemed as falling within the range of risks allocable to the community in this case. (See John R., supra, 48 Cal.3d at p. 452 [drawing a similar conclusion regarding sexual abuse and the authority of a teacher over a student]; Jeffrey E., supra, 197 Cal.App.3d at p. 723 [same, regarding authority of a Sunday school teacher over a minor].)

Accordingly, consideration of the three respondeat superior policy justifications reinforces our determination that Nelson’s lewd propositioning and offensive touching of his trainee and coworkers were not within the scope of his employment at the county jail.16

Despite what Justice Mosk suggests, our conclusion—that the sexual harassment here was not within the scope of employment even though it occurred during work hours in a workplace that may be characterized as traditionally male-dominated—finds overwhelming support in the decisions of other jurisdictions. (E.g., Tumminello v. City of New York (1995) 212 A.D.2d 434 [622 N.Y.S.2d 714] [New York law]; Smith v. American Exp. Travel Services (1994) 179 Ariz. 131 [876 P.2d 1166] [Arizona law]; Phelps *1018v. Vassey (1993) 113 N.C.App. 132 [437 S.E.2d 692] [North Carolina law]; Carrv. U.S. West Direct Co. (1989) 98 Or.App. 30 [779 P.2d 154] [Oregon law]; Dockter v. Rudolf Wolff Futures, Inc. (N.D.Ill. 1988) 684 F.Supp. 532 [Illinois law]; Valdez v. Church’s Fried Chicken, Inc. (W.D.Tex. 1988) 683 F.Supp. 596 [Texas law]; Favors v. Aleo Mfg. Co. (1988) 186 Ga.App. 480 [367 S.E.2d 328] [Georgia law]; Bennett v. Corroon and Black Corp. (La.Ct.App. 1987) 517 So.2d 1245 [Louisiana law]; Davis v. United States Steel Corp. (4th Cir. 1985) 779 F.2d 209 [South Carolina law]; but see Carlson v. Crater Lake Lumber Co. (1991) 105 Or.App. 314 [804 P.2d 511] [Oregon law]; Davis v. Black (1991) 70 Ohio App.3d 359 [591 N.E.2d 11] [Ohio law].) While these out-of-state authorities are not controlling in our determinations, they nonetheless demonstrate that Justice Mosk’s contrary conclusion is not in sync with the national trend.

On a last note, we observe that the Court of Appeal, in determining that Nelson’s acts were within the scope of employment, attributed significant weight to the fact that the County had elected to defend other employees also accused in the same lawsuit of sexual harassment. For the reasons that follow, we disagree with the view that a public entity’s decision to defend certain employees accused of sexual harassment is an appropriate factor for determining scope of employment.

First, we have been presented with no sound legal or policy reason to support the.conclusion that, under sections 995 and 995.2, a public entity may not properly defend a public employee in a civil suit alleging sexual harassment if the entity determines that the harassment charges are not well founded. Indeed, providing a defense in such circumstances is consistent with one of the purposes of governmental defense statutes, to provide public employees acting in the scope of employment with a measure of protection from the harassment of vexatious lawsuits. (39 Ops.Cal.Atty.Gen. 71, 73 (1962), citing Huffaker v. Decker (1946) 77 Cal.App.2d 383, 388 [175 P.2d 254] [both discussing earlier statutes].) We therefore decline to hold that a public entity’s agreement to undertake representation of certain employees for such reasons will defeat the entity’s right to refuse the defense of other employees whose acts of sexual harassment are undisputed.

Second, we see no inconsistency in the County’s apparent determination that employees accused of failing to properly investigate and respond to sexual harassment charges were acting within the scope of employment. It is reasonable to assume that employees who investigate such charges and attempt remedial measures are engaged in conduct required of their employment, and that shortcomings in fulfilling such duties are either incidental to or reasonably foreseeable as a direct consequence of such duties.

*1019In closing, we observe that both the County and Farmers take the position that the scope of employment question may be resolved as a matter of law in this case. We agree. “Ordinarily, the determination whether an employee has acted within the scope of employment presents a question of fact; it becomes a question of law, however, when ‘the facts are undisputed and no conflicting inferences are possible.’ ”17 (Mary M., supra, 54 Cal.3d at p. 213.) Mindful of this standard, we conclude that the facts, even when viewed in the light most favorable to Farmers, present no basis for finding that Nelson acted within the scope of his employment.18 Consequently, Farmers cannot prevail on its claims against the County for indemnification and defense costs under sections 825.2 and 996.4.

III. Conclusion

We conclude that, because Nelson’s repeated acts of sexual harassment fell outside the scope of his employment as a deputy sheriff, public funds may not be used to indemnify him for liabilities arising out of his own despicable conduct. In so concluding, however, we wish to emphasize that our holding will not, as various amici curiae seem to fear, eliminate the incentive for employers to prevent or respond to sexual harassment in the *1020workplace. Nor will our holding leave sexual harassment victims without adequate means to recover compensation for their injuries.

Even though, under our analysis, the respondeat superior doctrine would not subject an employer to vicarious liability for sexual harassment exceeding the scope of employment, employers remain directly liable to sexually harassed workers for violations of the FEHA (§ 12940, subd. (h)) and Title VII (42 U.S.C. § 2000e-2).19 The availability of damages and other relief for employer misconduct and statutory violations provide strong incentives to private and public employers alike to eliminate sexual harassment from their workplaces. Finally, the goal of eradicating harassment from the public sector is more effectively advanced by denying sexual harassers the right to indemnity than by insulating them from financial responsibility for their own misconduct.

The judgment of the Court of Appeal is reversed, and the matter is remanded to that court with directions to vacate the judgment in favor of Farmers and to enter judgment in favor of the County.

Lucas, C. J., Arabian, J., George, J., and Werdegar, J., concurred.

I agree with, and have signed, the majority opinion. I write separately because, in addition to distinguishing the decision in Mary M. v. City of Los Angeles (1991) 54 Cal. 3d 202 [285 Cal.Rptr. 99, 814 P.2d 1341] (Mary M.) (as does the majority), I would go further and overrule Mary M., because I believe that case was wrongly decided. By declining to overrule that decision at this time, we run the risk that lower courts in future cases will feel constrained to follow the aberrant holding of that decision.

*1021In Mary M. this court restated the principles governing whether an employee was acting within the scope of employment for purposes of respondeat superior, formulated by the court in cases such as Perez v. Van Groningen & Sons, Inc. (1986) 41 Cal.3d 962 [227 Cal.Rptr. 106, 719 P.2d 676] and John R. v. Oakland Unified School Dist. (1989) 48 Cal.3d 438 [256 Cal.Rptr. 766, 769 P.2d 948]. Mary M. did not alter these established rules; it only purported to apply these rules in holding that a police officer acts within the scope of his employment when he rapes a woman he has detained. This startling result was based upon the proposition that the “potential for abuse” is inherent in the “extraordinary power and authority” granted to police officers. (Mary M. v. City of Los Angeles, supra, 54 Cal.3d 202, 216-217.) This court concluded that “[t]he cost resulting from misuse of that power should be borne by the community, because of the substantial benefits that the community derives from the lawful exercise of police power.” (Id. at p. 217.)

The majority in the present case distinguishes the decision in Mary M. by noting that the holding in that case applies only to “sexual misconduct by on-duty police officers against members of the public.” (Maj. opn., ante, at p. 1006.) The present case is different, because it involves acts of sexual harassment by Deputy Sheriff Craig Nelson against female deputy sheriffs working at a county jail. Nelson was the training officer of one of his victims and told her repeatedly that she would have to perform sexual acts in order successfully to complete her training. As the majority correctly concludes, the decision in Mary M. is not controlling in the present case, even as to the acts of harassment committed by Nelson in his role as training officer, because “the work-related authority of a supervisor over a trainee employee in a county sheriff’s department is in no way comparable to the extraordinary power and authority that police officers exercise over members of the public.” (Maj. opn., ante, at p. 1012; Thorn v. City of Glendale (1994) 28 Cal.App.4th 1379, 1384 [35 Cal.Rptr.2d 1] [in holding that arson committed by an on-duty fire marshal was outside the scope of his employment, the Court of Appeal observed that Mary M. “appears to have established a special rule for the independent wrongful acts of police officers based upon their unique position of both trust and power in our society”].) I would go further, however, and recognize that the holding in Mary M. is an aberration that should be overruled.

The majority in the present case correctly concludes that Nelson’s acts of sexual harassment were outside the scope of his employment, because they were undertaken solely for his personal gratification and had no purpose connected to his employment. The same was true in Mary M. The rape committed by the police officer in Mary M. was undertaken solely for his *1022personal gratification and had no purpose connected to his employment. (See Thorn v. City of Glendale, supra, 28 Cal.App.4th 1379, 1383 [arson by an on-duty fire marshal was “the result. . . of a personal compulsion”].)

The circumstance that the police officer in Mary M. abused the authority vested in him in committing the rape does not bring his act within the scope of his employment if the crime was “so unusual or startling” that it cannot fairly be said to have arisen from the employment. As we stated in Perez v. Van Groningen & Sons, Inc., supra, 41 Cal.3d 962: “A risk arises out of the employment when ‘in the context of the particular enterprise an employee’s conduct is not so unusual or startling that it would seem unfair to include the loss resulting from it among other costs of the employer’s business. [Citations.] In other words, where the question is one of vicarious liability, the inquiry should be whether the risk was one “that may fairly be regarded as typical of or broadly incidental” to the enterprise undertaken by the employer. [Citation.]’ ” (Id. at p. 968.)

Thankfully, it is both unusual and startling for an on-duty police officer to rape a woman whom he has detained. It seems unfair to include the loss resulting from such a heinous and shocking crime among the losses to be expected from the operation of a police force by a public entity. Why should the public bear the financial burden imposed as a result of such misconduct, in situations where there has been no showing that the public entity was negligent either in hiring or supervising its employee?

The decision in Mary M. reached a contrary conclusion, reasoning that “[i]n view of the considerable power and authority that police officers possess, it is neither startling nor unexpected that on occasion an officer will misuse that authority by engaging in assaultive conduct.” (Mary M. v. City of Los Angeles, supra, 54 Cal.3d 202, 217.) But it is one thing to say that a public entity must expect that some police officers will abuse their authority by, for example, using excessive force in effectuating an arrest or detention (id. at p. 215) and quite another to conclude that a public entity must expect that some officers will rape women they have detained. Must a public entity similarly expect that some officers will misuse their authority to commit theft or murder while on duty? (See, e.g., People v. Von Villas (1992) 11 Cal.App.4th 175, 196-209 [15 Cal.Rptr.2d 112].) Is the public entity responsible for such crimes by off-duty police officers if the officers, in perpetrating the offenses, misuse their authority? I do not believe that such crimes, committed solely for personal reasons, fall within the scope of a police officer’s employment. (Mary M. v. City of Los Angeles, supra, 54 Cal.3d 202, 242 (conc. opn. of Baxter, J.).) Of course, if a public entity negligently hires or retains an officer who it knows, or reasonably should know, poses a danger of committing such misconduct, the entity may be held directly liable for the resulting injury.

*1023Police officers should be governed by the same standard employed in determining whether the misconduct of other employees falls within the scope of employment. Police officers occupy a position of trust and authority in our society, but the same is true of other public employees, such as teachers. As Justice Baxter observed in his concurring opinion in Mary M.: “A schoolteacher alone at his home with an impressionable child has as much power and opportunity to commit a sexual assault against the child, especially one of tender years, as a police officer has to commit an assault against a citizen.” (Mary M. v. City of Los Angeles, supra, 54 Cal.3d 202, 234-235 (conc. opn. of Baxter, J.).)

The decision in Mary M. created special rules, purportedly applicable only to on-duty police officers, for determining whether sexual misconduct falls within the scope of employment for purposes of respondeat superior. I believe we should recognize that the general rules governing the doctrine of respondeat superior—and, in particular, the issue whether particular misconduct comes within the scope of employment—apply in all cases, including those involving sexual misconduct committed by on-duty police officers. Under these general rules, rape of a detainee by a police officer falls outside the scope of employment, because such misconduct is committed solely for the officer’s personal gratification and is “so unusual or startling” that it cannot fairly be said to have arisen from the employment. Accordingly, rather than simply distinguish the present case from Mary M., I believe we should overrule the decision in that case. By explicitly recognizing at this time the flaw in the Mary M. decision, we would assure that, in the future, all cases will be governed by the general rules of respondeat superior ably set forth and applied in the majority opinion.

Lucas, C. J., concurred.

I fully concur in the majority opinion. I write separately to highlight what I perceive to be an analytical flaw in Justice Mosk’s dissent.

What this case asks is whether a public-entity employer, here Santa Clara County, that is directly liable for its own negligence in connection with coemployee sexual harassment, and automatically liable for harassment by a supervisor (Gov. Code, § 12940, subd. (h)), should, in addition, be required under principles of respondeat superior to indemnify the proven wrongdoer (see Gov. Code, §§ 825, 995) merely because his wrongful acts occurred at the worksite, during work hours. As the majority aptly demonstrates, law, logic, common sense and public policy dictate the answer be “no.”

In determining whether an employer is strictly liable for its employee’s tort under the doctrine of respondeat superior, Justice Mosk, like the majority, acknowledges the test we should apply is “whether the employee’s *1024conduct was reasonably foreseeable ‘in the context of the particular enterprise’ in which it took place.” (Dis. opn. of Mosk, J., post, at p. 1029.) Purporting then to apply this test, Justice Mosk asserts that the relevant question therefore “is the incidence of sexual harassment by coworkers in traditionally male workplaces, and specifically in county jails, that have recently been integrated by sex.” (Ibid., italics in original.) Having thus framed the “relevant” question, Justice Mosk concludes Deputy Nelson’s conduct was foreseeable for respondeat superior purposes, because it occurred at a time when Deputy Bates was one of the few women to integrate a formerly and traditionally all-male workplace—a big-city jail—and studies and case law “both show that harassment by coworkers is pervasive in traditionally male workplaces that have recently been integrated by sex . . . .” (Ibid.)

In thus relying on the asserted frequency of sexual harassment in male-dominated workplaces, Justice Mosk, in my view, has confused the fact of male resentment of female encroachment on what previously may have been viewed as exclusively male “territory,” with the concept of the nature of the duties and tasks the employees of the enterprise, male or female, are required to perform. That sexual harassment might be a foreseeable consequence of integrating a workforce due to male resentment of the female presence, that it may even have been predictable, is not, however, to say such harassment is “reasonably foreseeable ‘in the context of the particular enterprise’ ” as that concept applies to respondeat superior liability. Reasonable foreseeability in the latter sense embodies legal and policy judgments about the degree to which it is fair to impose liability on the employer for an employee’s conduct. The dissent oversimplifies these difficult judgments by treating “foreseeability” primarily as a matter of statistics. The cited cases do not, in my view, support such an approach.

Justice Mosk relies on Carr v. Wm. C. Crowell Co. (1946) 28 Cal.2d 652 [171 P.2d 5] for the broad principle that because “[m]en [or women] do not discard their personal qualities when they go to work” (id. at p. 656), an employer is liable for injuries inflicted by one employee on another caused by such human qualities as animosity or emotional flare-up (dis. opn. of Mosk, J., post, at pp. 1033-1034). But Carr involved a dispute between two employees, one a contractor and the other a subcontractor, over the performance of their tasks. Rejecting the argument that when the defendant threw his carpenter’s hammer at the plaintiff he was not acting in the scope of his employment, this court held that for liability to apply “[i]t is sufficient. . . if the injury resulted from a dispute arising out of the employment” (28 Cal.2d at p. 654)—i.e., out of the performance of the employees’ duties (id. at p. 657).

*1025The opinion in Carr v. Wm. C. Crowell Co., supra, 28 Cal.2d 652, rather than supporting Justice Mosk’s view, instead highlights the difficulty of articulating a plausible link between sexual harassment and Deputy Nelson’s duties as a deputy sheriff. That sexual harassment in the workplace occurs frequently, or in some enterprises is pervasive, is deplorable. That it does thus occur is not, however, determinative of the issue before the court. As the majority correctly observes, “evidence of the general prevalence of sexual harassment in . . . newly integrated work environments has little, if any, probative value in determining whether” sexual harassment of a female deputy sheriff is a risk “broadly incidental” to the operation of a county jail and, thus, one for which the law imposes liability on the employer. (Maj. opn., ante, at p. 1009.)

Justice Kennard’s views in this respect are similar to Justice Mosk’s. While she believes the relevant facts are disputed, Justice Kennard nevertheless finds potentially dispositive the factual question whether Deputy Nelson’s “behavior was typical of the activities of employees at the jail . . . .” (Dis. opn. of Kennard, J., post, p. 1043.) Does this mean that identical acts of sexual harassment in the state’s many county jails are, or are not, within the scope of employment depending upon the frequency of harassment in each? Is sexual misconduct broadly incidental to the operation of jails in some counties but not in others? To my mind, the implausiblity of such a conclusion illustrates the error in giving overly much weight to statistical considerations.

Because Deputy Nelson’s sexual harassment of Deputy Bates was not broadly incidental to his duties as a deputy sheriff, the County of Santa Clara cannot be held strictly liable for his tortious actions under the doctrine of respondeat superior.

Although I authored the majority opinion, I write separately to reiterate my disagreement with Mary M. v. City of Los Angeles (1991) 54 Cal.3d 202 [285 Cal.Rptr. 99, 814 P.2d 1341] (Mary M.) and to clarify that I adhere to the views set forth in my concurrence to that opinion and to those expressed by Justice George’s concurrence in the case at bar.

I also write separately because, while I agree with Justice George that Mary M. was wrongly decided and should be overruled, I do not believe this case presents the proper vehicle because the facts here are amply distinguishable.

I dissent.

Although Deputy Sheriff Nelson’s conduct was undoubtedly deplorable, it nevertheless fell “within the scope of his employment” for purposes of the indemnification statute (Gov. Code, § 825.2, subd. (b)) because it was “reasonably foreseeable” in the broad sense in which that term is used in the law of respondeat superior, i.e., because sexual harassment of a woman deputy sheriff working in a big-city jail a decade ago was not “so unusual or startling” that it would be unfair to include the resulting expense in the county’s costs of doing business.

The applicable rules of law are not in dispute.

First, as the majority observe (maj. opn., ante, at p. 1003), the term “scope of employment” as used in the Tort Claims Act has the same meaning that it *1026has “in cases involving actions by third persons against the employer for the torts of his employee” (4 Cal. Law Revision Com. Rep. (Dec. 1963) p. 814, fn. 3), i.e., the same meaning that it has in the common law doctrine of respondeat superior.

Second, in California the term “scope of employment” has been given a broad meaning for respondeat superior purposes. As the majority acknowledge (maj. opn., ante, at p. 1004), in this state an employee’s conduct may be within the scope of employment for respondeat superior purposes even if (1) it constitutes a willful and malicious tort, or (2) violates an express rule or policy of the employer, or (3) confers no benefit whatever on the employer; and (4) if the employee is combining his own business with that of his employer, “no nice inquiry will be made” into which activity he was actually engaged in when the injury occurred. (See generally, Perez v. Van Groningen & Sons, Inc. (1986) 41 Cal.3d 962, 968-970 [227 Cal.Rptr. 106, 719 P.2d 676] (Perez).) These rules are obviously applicable here.

Third, the test for determining when an employee’s conduct is within the scope of employment for respondeat superior purposes is set forth in our cases. The leading modem decision of this court on the doctrine of respondeat superior is Hinman v. Westinghouse Elec. Co. (1970) 2 Cal.3d 956 [88 Cal.Rptr. 188, 471 P.2d 988] (Hinman). Our unanimous opinion in that case recited (id. at p. 960) that “California cases have long recognized that the employer’s responsibility for the torts of his employee extends beyond his actual or possible control of the servant to injuries which are ‘risks of the enterprise.’ (Carr v. Wm. C. Crowell Co., 28 Cal.2d 652, 655-656 [171 P.2d 5]; George v. Bekins Van & Storage Co., 33 Cal.2d 834, 843 [205 P.2d 1037]; Fields v. Sanders, 29 Cal.2d 834, 841 [180 P.2d 684, 172 A.L.R. 525].) Chief Justice Traynor has pointed out: ‘The principal justification for the application of the doctrine of respondeat superior in any case is the fact that the employer may spread the risk through insurance and carry the cost thereof as part of his costs of doing business.’ (Johnston v. Long, 30 Cal.2d 54, 64 [181 P.2d 645].) Thus, it must be deemed settled in California that in accordance with the principal justification for the doctrine, the employer’s liability extends to the risks inherent in or created by the enterprise.” (Italics added and original italics deleted.)

The emphasized test, of course, is necessarily somewhat general in its terms. The question is how to determine whether a risk is “inherent in or created by” the enterprise on the facts of a specific case. The best answer so far to that question was given by the scholars Harper and James in the first edition of their treatise and adopted into the law of California in the much *1027cited case of Rodgers v. Kemper Constr. Co. (1975) 50 Cal.App.3d 608 [124 Cal.Rptr. 143] (Rodgers). The opinion in Rodgers makes it clear that “reasonable foreseeability” is not an alternative to the “inherent or incidental risk” test of the scope of employment, but a practical method of applying that test to specific facts. Thus the Rodgers opinion explains that “One way to determine whether a risk is inherent in, or created by, an enterprise is to ask whether the actual occurrence was a generally foreseeable consequence of the activity. However, ‘foreseeability’ in this context must be distinguished from ‘foreseeability’ as a test for negligence. In the latter sense ‘foreseeable’ means a level of probability which would lead a prudent person to take effective precautions whereas ‘foreseeability’ as a test for respondeat superior merely means that in the context of the particular enterprise an employee’s conduct is not so unusual or startling that it would seem unfair to include the loss resulting from it among other costs of the employer’s business. (2 Harper & James, The Law of Torts, pp. 1377-1378. . . .)” (50 Cal.App.3d at pp. 618-619, italics added and original italics deleted.)

This court has repeatedly quoted with approval the test of foreseeability adopted in Rodgers (see Perez, supra, 41 Cal.3d at p. 968; John R. v. Oakland Unified School Dist. (1989) 48 Cal.3d 438, 450, fn. 9 [256 Cal.Rptr. 766, 769 P.2d 948] (lead opn. of Arguelles, J.), 464-465 (conc. & dis. opn. of Kaufman, J.) (John R.); Mary M. v. City of Los Angeles (1991) 54 Cal.3d 202, 209 [285 Cal.Rptr. 99, 814 P.2d 1341]), and the majority do so again (maj. opn., ante, at pp. 1003-1004). In addition, the foreseeability test of Rodgers “has been widely followed by the courts of appeal. . . .” (Childers v. Shasta Livestock Auction Yard, Inc. (1987) 190 Cal.App.3d 792, 803-804 [235 Cal.Rptr. 641] [citing cases and calling Rodgers “Clearly the leading case in this area”]; accord, Debbie Reynolds Prof. Rehearsal Studios v. Superior Court (1994) 25 Cal.App.4th 222, 227-228 [30 Cal.Rptr.2d 514]; State Farm Mut. Auto. Ins. Co. v. Haight (1988) 205 Cal.App.3d 223, 242 [252 Cal.Rptr. 162] [calling Rodgers “The classic analysis of the question of foreseeability in this context”].)1

Thus the correct test for determining when an employee’s conduct is within the scope of employment for respondeat superior purposes is simply the test we articulated in Hinman, i.e., whether the conduct is a risk “inherent *1028in or created by the enterprise,”2 and the best way to determine whether a risk is inherent in or created by an enterprise is to ask, with Rodgers and Harper and James, whether the employee’s conduct was “so unusual or startling” in the context of that enterprise that it would be unfair to include the resulting loss in the employer’s costs of doing business. I shall apply that test to the facts alleged.

First, as is often true, it will be helpful to identify what this case is not about. It is not about a group of priests who seduced a 16-year-old parishioner in the confessional and elsewhere. (Rita M. v. Roman Catholic Archbishop (1986) 187 Cal.App.3d 1453 [232 Cal.Rptr. 685].) It is not about a Sunday school teacher who repeatedly molested a second grader entrusted to his care. (Jeffrey E. v. Central Baptist Church (1988) 197 Cal.App.3d 718 [243 Cal.Rptr. 128].) It is not about a school janitor who molested and sexually assaulted an 11-year-old student in the janitor’s office. (Alma W. v. Oakland Unified School Dist. (1981) 123 Cal.App.3d 133 [176 Cal.Rptr. 287].) It is not about a junior high school mathematics teacher who molested a 14-year-old student engaged in correcting other students’ papers at the teacher’s apartment (John R., supra, 48 Cal.3d 438), or an elementary school teacher who molested a 5-year-old pupil in the classroom (Kimberly M. v. Los Angeles Unified School Dist. (1989) 215 Cal.App.3d 545 [263 Cal.Rptr. 612]). And it is not about a dance instructor who sexually abused a 15-year-old dance student in a trailer adjacent to the rehearsal studio. (Debbie Reynolds Prof. Rehearsal Studios v. Superior Court, supra, 25 Cal.App.4th 222.) In each of those cases the church or school had no reason to believe that its employee would betray his trust so far as to sexually molest a minor in his charge. In each case, therefore, under the foreseeability test of respondeat superior the employee’s conduct was “so unusual or startling” in the context of the particular enterprise that it would have been unfair to include the loss in the employer’s costs of doing business.

In sharp contrast, we are not dealing here with priests or schoolteachers or even ordinary office personnel, and the workplace where these events occurred was not a church or a school or an ordinary office. Rather, it was a big-city jail, and all the participants were adults and coworkers—indeed, all were deputy sheriffs, in uniform and on duty, doing the difficult and often stressful work of guarding or transporting accused or convicted criminals. Traditionally, of course, this work was done exclusively by men, just as the *1029vast majority of their charges—the jail inmates—were men. Indeed, there are few more male-dominated environments than jails and prisons. It was not until relatively recently that law enforcement authorities began to hire women as deputy sheriffs and correctional officers, and women are still a small minority of those so employed. In the case at bar the harassment occurred over a decade ago—in 1983 and 1984—only a few years after the Santa Clara Sheriff’s Department first employed women as deputy sheriffs.

These facts are relevant because I do not rely on the studies cited by the majority showing the widespread incidence of sexual harassment in the American workplace in general. (See, e.g., 1 Conte, Sexual Harassment in the Workplace (2d ed. 1994) pp. 1-2; Lindemann & Kadue, Sexual Harassment in Employment Law (1992) pp. 4-7.) Rather, as the majority emphasize, the test we should apply is whether the employee’s conduct was reasonably foreseeable “in the context of the particular enterprise” in which it took place. The question, therefore, is the incidence of sexual harassment by coworkers in traditionally male workplaces, and specifically in county jails, that have recently been integrated by sex.

The answer is clear: studies and case law both show that harassment by coworkers is pervasive in traditionally male workplaces that have recently been integrated by sex, and especially so in military-style institutions like law enforcement. “Co-worker harassment tends to occur most often in situations where women have entered jobs or workplaces traditionally occupied by male incumbents. Thus, plaintiffs in many co-worker harassment cases are women who have entered traditionally ‘male’ jobs such as police officer, firefighter, plumber, electrician, truck driver, engineer, car salesperson, pilot, air traffic controller, securities trader, surgeon, miner, automobile mechanic, airline mechanic, and railroad engineer. Other plaintiffs are women who work in traditionally ‘male’ work environments such as fire stations, warehouses, assembly line manufacturing operations, prisons, oil refining companies, paper mills, construction sites, steel plants, and the military.” (Lindemann & Kadue, Sexual Harassment in Employment Law, supra, pp. 234-235, fns. omitted, italics added, citing cases; accord, 1 Conte, Sexual; Harassment in the Workplace, supra, pp. 98-100, citing cases.)

Thus when a woman was hired by General Motors as a tinsmith apprentice in one of its factories, “She was the first woman to work in the tinsmith shop, and her male coworkers were unhappy about working with a woman.” (Carr v. Allison Gas Turbine Div. Gen. Motors (7th Cir. 1994) 32 F.3d 1007, 1009.) The male coworkers expressed their displeasure by sexually harassing the woman in a variety of ways, and she filed an employment discrimination *1030action in federal district court alleging a hostile work environment. Reversing a defense judgment, the circuit court directed judgment for the plaintiff on the question of liability. The court rejected inter alia a trial court finding that the plaintiff did not prove General Motors knew or should have known of the harassment and failed to stop it; instead the court reasoned, in an opinion by Chief Judge Richard Posner, that “General Motors was astonishingly unprepared to deal with problems of sexual harassment, foreseeable though they are when a woman is introduced into a formerly all-male workplace.” (Id. at p. 1012, italics added.)3

Although various explanations for coworker harassment of members of a female minority have been proposed, there is much support for the view that men employed in a traditionally male setting may perceive women entering their workplace as threatening either their job security or their self-esteem, and may use sexual harassment as a means to resist the intrusion. (Lindemann & Kadue, Sexual Harassment in Employment Law, supra, p. 235.) “For women in male-dominated jobs, harassment is less likely to take the form of supervisors’ demands for sexual favors and more likely to take the form of sexual taunts and other actions by co-workers that are part of a larger pattern of hostility intended to drive the women away.” (Schultz, Telling Stories About Women and Work: Judicial Interpretations of Sex Segregation in the Workplace in Title VII Cases Raising the Lack of Interest Argument (1990) 103 Harv. L.Rev. 1749, 1832-1833, fn. 321.) Other scholars agree: “ ‘By making insulting comments and touching women sexually, some men may try to “make life miserable” for women in the [nontraditional] jobs, encouraging them to leave. The relatively high turnover rate among women in [these jobs] suggests that this is a successful strategy to force women out.’ ” (Id. at p. 1834, fn. 328, quoting Gutek, Sex and the Workplace (1985) p. 119.) For example, when the Nevada Department of Prisons hired its first group of women correctional officers in the mid-1970’s, their integration into the prison system proved to be a “laborious and difficult process”: after establishing that women could in fact perform the duties of correctional officers, it was “necessary to train and indoctrinate the all-male staff which was then in place. Some of these staff members had expressed beliefs such as, ‘prisons are no place for women.’ Other officers were reluctant to recognize the status of the newly-hired women as full-fledged correctional officers .... [S]ome staff members manifested their opposition to the employment of women officers by engaging in sexual harassment *1031and sex discrimination.” (Snow v. Nevada Dept, of Prisons (D.Nev. 1984) 582 F.Supp. 53, 55.)

Whatever the reason, it is clear that the incidence of sexual harassment in traditionally male workplaces is high. In 1986, for example, “[a] study of women in the traditionally male fields of engineering, science, and management revealed that 75 percent of the respondents had experienced one or more types of harassment.” (1 Conte, Sexual Harassment in the Workplace, supra, p. 2, in. omitted, citing Lafontaine & Tredeau, The Frequency, Sources, and Correlates of Sexual Harassment Among Women in Traditional Male Occupations (1986) 15 Sex Roles 433, 436.) The situation in traditional military service is comparable. Thus in a recent study of 333 former servicewomen who sought Veterans’ Administration hospital services in 1992 and 1993, 90 percent of the subjects younger than 50 reported they had been sexually harassed while in the military. (Sex Abuse of Military Women, S.F. Chronicle (May 12, 1995) p. A4, cols. 4-6.) Even in a broader study of 10,750 servicewomen on active duty conducted in 1988 by the Department of Defense, fully 64 percent reported they had been sexually harassed. (Ibid.)

Turning to the particular context of the case at bar, we find that sexual harassment is also all too common in local police forces. One group of cases adjudicates complaints by women employees of police departments charging that they were sexually harassed by police officers or superiors. (See, e.g., Lankford v. City of Hobart (10th Cir. 1994) 27 F.3d 477, 478 [police dispatchers]; Henson v. City of Dundee (11th Cir. 1982) 682 F.2d 897, 899 [same]; Dirksen v. City of Springfield (C.D.111. 1994) 842 F.Supp. 1117, 1119 [police secretary]; Ball v. City of Cheyenne (D.Wyo. 1993) 845 F.Supp. 803, 806-807 [police dispatcher]; Froyd v. Cook (E.D.Cal. 1988) 681 F.Supp. 669, 671 [same, applying California law].)

Another group of cases addresses complaints by women police officers charging that they were sexually harassed by fellow officers or superiors. (See, e.g., Accardi v. Superior Court (1993) 17 Cal.App.4th 341, 346 [21 Cal.Rptr.2d 292]; Andrews v. City of Philadelphia (3d Cir. 1990) 895 F.2d 1469, 1474-1475, 1479; Barcume v. City of Flint (E.D.Mich. 1993) 819 F.Supp. 631; Poulsen v. City of North Tonawanda, N.Y. (W.D.N.Y. 1993) 811 F.Supp. 884, 888-889; Watts v. New York City Police Dept. (S.D.N.Y. 1989) 724 F.Supp. 99, 101-102; Haehn v. City of Hoisington (D.Kan. 1988) 702 F.Supp. 1526, 1529; Arnold v. City of Seminole (E.D. Okla. 1985) 614 F.Supp. 853, 858-859, 862-863.)

More specifically, the case law also illustrates the prevalence of sexual harassment of female law enforcement personnel working in prisons and *1032jails. One group of cases adjudicates complaints by women correctional officers in state prisons charging that they were sexually harassed by fellow officers or superiors. (See, e.g., Morgan v. Ford (11th Cir. 1993) 6 F.3d 750, 752, 756; Minteer v. Auger (8th Cir. 1988) 844 F.2d 569, 571; cf. Hirschfeld v. New Mexico Corrections Dept. (10th Cir. 1990) 916 F.2d 572, 574 [sexual harassment of state prison employee by correctional officer]; Cuesta v. Texas Dept, of Criminal Justice (W.D.Tex. 1991) 805 F.Supp. 451, 456-457 [sexual harassment of parole caseworker by supervisor].)

Still more numerous are complaints like those made in the case at bar, i.e., complaints by women deputy sheriffs or other women correctional officers in county jails charging that they were sexually harassed by fellow officers or superiors. (See, e.g., Crighton v. Schuylkill County (Mar. 14, 1995, E.D.Pa. Civ. A. No. 94-5658); Anthony v. County of Sacramento, Sheriffs Dept. (E.D.Cal. 1994) 845 F.Supp. 1396, 1399 [applying California law]; Sherod v. Wahl (Mar. 19, 1993, N.D.Ill. No. 91 C 7953); Sims v. Montgomery County Com’n (M.D.Ala. 1990) 766 F.Supp. 1052, 1070-1074; Bennett v. New York City Dept, of Corrections (S.D.N.Y. 1989) 705 F.Supp. 979, 984-985; cf. Handley v. Phillips (M.D.Pa. 1989) 715 F.Supp. 657, 674 [sexual harassment of county jail matron by warden].) For example, in Sims v. Montgomery County Com’n, supra, 766 F.Supp. 1052, 1070, the federal district court concluded that “sexual harassment in the Montgomery County Sheriff’s Department has permeated all ranks, from the lowest level corrections officers and deputy sheriffs to the sheriff himself, and is so pervasive and severe as to render the working conditions in the department psychologically intolerable for female officers.”

The majority seek to distinguish these cases on several grounds, but none refutes the simple reason why I cite them: I cite these cases not for their law but for their facts, i.e., to show the frequency with which women police and correctional officers complain of sexual harassment by fellow officers or superiors. The sheer number of such complaints revealed by these cases is certainly relevant to the issue whether sexual harassment of women deputy sheriffs by fellow officers is “unusual or startling.” And these complaints represent but the tip of the iceberg: because “Women often remain silent when confronted with sexual harassment” (Lindemann & Kadue, Sexual Harassment in Employment Law, supra, at p. 6, fn. omitted), it is to be expected that few women police and correctional officers will even file a formal administrative complaint charging misconduct by their fellow officers, and fewer still will “make a federal case out of it” by litigating a Title VII action (42 U.S.C. § 2000e et seq.) against their own employer. The few who do, therefore, speak for the many who experience such sexual harassment.

*1033The majority cite cases from other jurisdictions that it claims support its conclusion that the sexual harassment in this case was not within the scope of employment even though it occurred during work hours in a traditionally male-dominated workplace. But in only one of the cited cases (Tumminello v. City of New York (1995) 212 A.D.2d 434 [622 N.Y.S.2d 714]) did the workplace have anything to do with law enforcement (there the parties were both detectives); all the other cases involved such sex-neutral workplaces as an insurance office, a fast-food restaurant, a commodity brokerage firm, a city public works department, a travel agency, and a manufacturing company.

More important, in none of the cases cited by the majority did the court apply the California test for determining when an employee’s conduct is within the scope of employment for respondeat superior purposes. Instead each inquired, solely or primarily, whether the conduct furthered the employer’s business or was included in the employee’s duties. (E.g., Tumminello v. City of New York, supra, 622 N.Y.S.2d at p. 715; Phelps v. Vassey (1993) 132 N.C.App. 132 [437 S.E.2d 692, 695]; Dockter v. Rudolf Wolff Futures, Inc. (N.D.Ill. 1988) 684 F.Supp. 532, 536 [applying Illinois law].) But in California the employer’s business and the employee’s duties are not the determinants of scope of employment: applicable here are the many compensation cases holding that “quarrels, assaults, or horseplay among employees ‘may reasonably be regarded as an incident of the employment,’ even though they are in no way intended to further the employer’s business, if they are engendered by the associations or conditions of employment . . . .” (5 Harper et al., The Law of Torts, supra, § 26.8, p. 43, fn. 19, italics added.) This is because “Such associations ‘include the faults and derelictions of human beings as well as their virtues and obediences. Men do not discard their personal qualities when they go to work. Into the job they carry their intelligence, skill, habits of care and rectitude. Just as inevitably they take along also their tendencies to carelessness and camaraderie, as well as emotional makeup. In bringing men [and women] together, work brings these qualities together, causes frictions between them, creates occasions for lapses into carelessness, and for fun-making and emotional flareup. . . . These expressions of human nature are incidents inseparable from working together. They involve risks of injury and these risks are inherent in the working environment.’ ” (Carr v. Wm. C. Crowell Co. (1946) 28 Cal.2d 652, 656 [171 P.2d 5].) Here, when women deputy sheriffs were thrust into the traditionally male workplace of the county jail, sexual harassment became a risk “inherent in or created by the enterprise” and hence within the scope of employment for respondeat superior purposes.

The realities of this situation are best understood, moreover, not by the members of this court in our ivory tower far removed from the scene, but by *1034the correctional officers who actually work in our jails and prisons. Among the filings by amici curiae in this case is a brief by the California Correctional Peace Officers Association, which advises us that as a matter of fact “Sexual harassment is commonplace in correctional facilities.” The brief goes further: endeavoring to assist us by using the language of our test, the correctional officers’ association explains that “Sexual harassment in a correctional facility is indeed foreseeable and is neither unusual nor startling. Unfortunately, sexual harassment in a newly gender-integrated correctional facility is typical [of] and broadly incidental to the enterprise being undertaken by the employer.” It would be presumptuous for us to believe we know the conditions in the trenches better than the frontline troops themselves.

Finally, we need not speculate on the particular working conditions at the time and place of the events herein, i.e., in the Santa Clara County jail in 1983 and 1984: the record shows they were typical. At the arbitration hearing in this case Sergeant Pascual of the Santa Clara County Sheriff’s Department testified in relevant part as follows:

“Q. You have been around for a long time. Is it unusual for deputies working North County Jail to discuss things of sexual nature? A. No, police work, up until a few years ago, was primarily a male-oriented work. It is going to take us a little bit of time before we realize we now have ladies amongst us. Insofar as discussing sex, I don’t know what you mean by sex. Is cussing sex? Is talking about your home life sex? Is talking about when you were in the army sex? You talk about everything. Sometimes you use quote unquote the F-word. . . . It is a vulgar place to begin with. There is talk of everything.

“Q. Would it be uncommon in the jail to hear deputies talking about giving or getting head? A. Not uncommon.”

Sergeant Pascual went on to relate a specific exchange of a sexual nature between Deputy Bates and Deputy Nelson, and characterized it as follows:

“A. . . . That is just regular jail talk.

“Q. Regular jail talk? A. Amongst the deputies. It is just a way to get rid of the stress of the job.

“Q. When you said that this is jail talk, I want to just clarify that. Does that mean that you have heard that sort of expression before this occasion? A. You mean that type of talk?

*1035“Q. Yes. A. Yes, that is very common amongst deputies.

“Q. Male and female alike? A. Yes, yes.”

The views of Sergeant Pascual were confirmed by an even more experienced officer. Lieutenant Tiano had been a member of the Santa Clara County Sheriff’s Department since 1971. In support of plaintiffs’ motion for summary judgment, Lieutenant Tiano filed a sworn declaration in which he stated as follows:

“2. I have read Sergeant Pascual’s testimony as given in the arbitration hearing of Deputy Nelson. What Sergeant Pascual says is true, and has been true for as long as I have been employed in the county jails.

“3. The jails have always been a place where strong language has been used, including profanity and conversation laced with sexual innuendo. This conversation takes place often in a joking manner between jail personnel, both male and female. In general, profanity, sexually explicit language, banter, and horseplay are extremely common among co-workers and peers.

“4. This type of interaction is and was so common at county jails that it would be fair to say that every employee of the jails or [the county department of corrections] either knew of such interactions or should have known of such interactions. While the above description is true today, it is especially true of the period of Deputy Nelson’s alleged sexual harassment of Cynthia Bates, Toni Daugherty, and Zana Murphy. This occurred in roughly 1983-84 when the Department had recently begun employing women as deputy sheriffs. As Sergeant Pascual testified, the behavior of many male deputy sheriffs did not change immediately upon the arrival of the female deputy sheriffs.”

In short, as the California Correctional Peace Officers Association observes in its brief, “To claim that such behavior was unforeseeable under the set of circumstances presented by this case is ludicrous.”

The majority dismiss the foregoing evidence on the reasoning that “even assuming arguendo that the usage of profanity and crude language at the jail should have put the County on notice that Nelson’s actions were ‘foreseeable’ in a negligence sense despite the absence of a causal [nexus] between the acts of sexual harassment and Nelson’s work as a deputy sheriff, that is a matter lacking relevance in scope of employment analysis.” (Maj. opn., ante, at p. 1011, italics added.) This reasoning does not withstand scrutiny. It attempts to draw a distinction between foreseeability “in a negligence sense” *1036and foreseeability for purposes of respondeat superior, asserting that the former is irrelevant to the latter. But this would be true only if the difference between the two were a difference in kind; it is not—it is simply a difference in degree. There is only one kind of foreseeability, but foreseeability requires a higher level of probability for negligence purposes than for respondeat superior purposes. As the majority elsewhere correctly observe, foreseeability as a test for negligence “ ‘means a level of probability which would lead a prudent person to take effective precautions whereas “foreseeability” as a test for respondeat superior merely means that in the context of the particular enterprise an employee’s conduct is not so unusual or startling’ ” that the loss should not be borne by the employer. (Maj. opn., ante, at p. 1004, italics added and original italics deleted.) The evidence that “sexually explicit. . . horseplay” was common among coworkers in the Santa Clara County jail in 1983 and 1984 is obviously relevant to the inquiry whether such conduct was foreseeable in the respondeat superior sense.

The majority compound their error by declaring a general rule to the effect that “factors that might be relevant to whether the County itself acted negligently are not relevant to whether the County should be vicariously liable for an employee’s misconduct regardless of its own fault.” (Maj. opn., ante, at p. 1011, italics added.) This rule is inconsistent with the majority’s earlier and correct explanation that the difference between foreseeability for negligence purposes and for respondeat superior purposes is merely the required “level of probability.” And the majority’s rule is unsupported by authority. First, the majority rely on two footnotes in John R., supra (48 Cal.3d at pp. 450, fn. 9, and 451, fn. 10). But that opinion was signed by only one other justice of this court; two justices dissented on the issue, and the remaining three justices (at p. 455) “concur[red] in the majority’s [¿ic] holding,” not in its opinion. In turn, the only authority cited in those two footnotes was the second case on which the majority now rely, Hinman, supra, 2 Cal.3d at page 960. But the majority quote Hinman out of context: at the page in question, Hinman was simply explaining that the “modem and proper basis” for the doctrine of respondeat superior is no longer the employer’s “control or fault” but “the risks incident to his enterprise.” Nothing in Hinman supports the majority’s new rule that evidence of foreseeability in the negligence sense is “irrelevant” to the question of foreseeability in the respondeat superior sense.

I do not, of course, condone the offensive remarks and acts complained of in this case. But neither do I agree with the majority’s seemingly naive view of that conduct: like Captain Renault in the classic film Casablanca, the majority profess to be “shocked, shocked to find that [sexual harassment *1037was] going on”4 a decade ago in the Santa Clara County jail. I believe, rather, that in the rough-and-tumble locker-room atmosphere of that traditionally male-dominated and recently sexually integrated workplace such conduct cannot plausibly be deemed “unusual,” less still “startling.”

To hold otherwise is simply to deny reality. As Justice Kaufman has wisely observed, “Sadly, however, we have learned that sexual harassment and assaults—in the home as well as the workplace—are not uncommon occurrences. This is a hard truth to accept. But putting our collective heads in the sand will not make it go away. And clinging to a less ‘pessimistic’ view of human nature . . . will not compensate the victims of such outrages. On the contrary, indulging such illusions merely deepens and perpetuates the injustice.” (John R., supra, 48 Cal.3d 438, 464 (conc. & dis. opn. of Kaufman, J.).)

For the foregoing reasons I conclude that Nelson’s conduct at least crossed the low threshold of general foreseeability that suffices to trigger respondeat superior liability in California. It was therefore within the scope of his employment for purposes of the indemnification statute, and the trial court erred in denying plaintiffs’ motion for summary judgment.5

I would affirm the judgment of the Court of Appeal.

Under California law, a government entity must indemnify any employee sued by a third party for acts arising within the scope of employment by reimbursing the employee for the costs incurred in litigating and settling the lawsuit. In this case, three women employed as deputy sheriffs at a county jail sued a male colleague, contending that he had sexually harassed them. After lengthy litigation, the parties settled the matter, and the male deputy and his insurance company (which had paid for the deputy’s lawyer and most of the cost of settlement) sought indemnification from the county for their litigation and settlement expenses.

The trial court granted summary judgment for the county, concluding that the deputy’s acts of sexual harassment were outside the scope of employment, and that therefore the county was not obligated to reimburse the deputy or his insurer for their litigation and settlement expenses. The Court of Appeal reversed, holding that the deputy’s acts were within the scope of employment, thus entitling the deputy and his insurer to indemnification by the county.

The majority agrees with the trial court that, as a matter of law, the harassing deputy’s conduct was outside the scope of his employment. Justice Mosk, on the other hand, as expressed in his dissenting opinion, agrees with the Court of Appeal that, as a matter of law, the conduct fell within the scope of employment. I find neither of these extreme views persuasive. As I see it, whether the deputy’s acts of harassment were within or outside the scope of employment is a question of fact that, on the record in this case, may not be resolved on summary judgment but must be determined by the trier of fact.

I

In 1983, Deputy Sheriff Cynthia Bates was one of the first women deputies assigned to work at Santa Clara County’s North County jail. Also working at the jail was Deputy Craig Nelson. In February 1984, Deputy Bates was assigned to the main jail, and Deputy Nelson became her training officer. In June 1984, Bates complained to her superiors that at both jails Nelson had made offensive, sexually explicit comments to her and had touched her on her legs and thighs. Another deputy, Toni Daugherty, alleged that in January or February of 1984 Nelson had grabbed her on the buttocks.

Santa Clara County (hereafter the County) conducted an internal investigation. Deputy Nelson acknowledged making the comments in question to *1039Deputy Bates, but denied that they were offensive; he asserted that the comments were made in a joking manner and were not intended to harass. He admitted “brushing” Bates on the thigh, but denied doing it repeatedly. He denied grabbing Deputy Daugherty’s buttocks, but admitted that he had touched them in an unsuccessful attempt to put double-stick tape on them as a practical joke.

Following its investigation, the County concluded that Deputy Nelson had sexually harassed the two women deputies, and suspended Nelson without pay for fourteen days. On administrative appeal, an arbitrator reduced the suspension to two days. Thereafter, Deputies Bates and Daugherty sued Deputy Nelson and the County in federal court, asserting claims of sexual harassment and differential treatment in violation of both title VII of the Civil Rights Act of 1964 (42 U.S.C. § 2000e-2) and California’s Fair Employment and Housing Act (Gov. Code, § 12940, subd. (h)). Joining in the complaint was a third deputy, Zana Murphy, who alleged, among other things, that Nelson had made offensive, sexually explicit comments to her also. The County refused to defend Nelson. He then obtained representation from Farmers Insurance Group, under a personal liability provision contained in his homeowner’s insurance policy. Nelson successfully moved for summary judgment on all of Murphy’s claims on the ground that they were untimely. On the eve of trial, Nelson settled with Bates and Daugherty, paying them a total of $150,000.1

Farmers Insurance Group and Nelson (hereafter jointly referred to as Farmers) then filed this action, asserting that the County was obligated to indemnify them for the amount of the settlement and the cost of defending the action. Both sides moved for summary judgment. The motions were based on declarations, the report of the arbitrator who had conducted the hearing on Deputy Nelson’s administrative appeal of the discipline imposed *1040against him,2 partial transcripts of the testimony in the disciplinary proceeding, and partial transcripts of the trial of the lawsuit brought by Deputies Bates, Daugherty, and Murphy against the County. These materials produced a significant area of disagreement relating to the conditions at the two jails at which the conduct complained of had occurred.

Farmers asserted that Deputy Nelson’s behavior, particularly his sexually oriented comments, was intended to be humorous, and was typical behavior at the jail.3 As described in the arbitrator’s report, Deputy Nelson told an internal affairs investigator: “Everything said or done was done in a joking manner. If at any point [Deputy Bates] would have said she was offended . . . that would have been the end of it. ... I definitely would not have continued on if I had known it was upsetting her.” In a declaration supporting Farmers’ motion for summary judgment, Nelson asserted that “profanity and sexually explicit language, banter, and horseplay are extremely common among the co-workers and peers in the jail.” Farmers also submitted a declaration from Lieutenant Armand Tiano, who had worked at the jail for many years, stating: “The jails have always been a place where strong language has been used, including profanity and conversation laced with sexual innuendo. This conversation takes place often in a joking manner between jail personnel, both male and female. In general, profanity, sexually explicit language, banter, and horseplay are extremely common among co-workers and peers.” There was also testimony by Sergeant David Pascual in the County’s disciplinary proceeding against Deputy Nelson that sexual banter was “very common” at the jail, “just a way to get rid of the stress of the job”; he described the jail as “a vulgar place to begin with. There is talk of everything.”

In its opposition to Farmers’ motion for summary judgment, the County asserted that the question whether sexual jokes and innuendo were common at the jail was a disputed issue of fact. The County relied on a statement by *1041Deputy Steven Cutright, quoted in the arbitrator’s ruling, that although there was “occasional joking around” among the employees at the jail, the joking was “not of a sexual nature.” The County also relied on disciplinary hearing testimony by Deputy Gagnon that “lewd, sexual comments” by deputies were not common at the jail.

The trial court granted the County’s motion for summary judgment, finding as a matter of law that Deputy Nelson’s acts of harassment4 were outside the scope of his employment. The Court of Appeal disagreed, holding that Nelson’s conduct as a matter of law fell within the scope of employment, and ordering the trial court to grant Farmers’ motion for summary judgment.

II

As a general rule, under the California Tort Claims Act a public entity must, upon request, provide for the defense of any employee sued “on account of any act or omission in the scope of his employment. . . .” (Gov. Code, § 995, italics added.)5 If the public entity refuses to do so, the employee may recover from the public entity “such reasonable attorney’s fees, costs and expenses as are necessarily incurred ... in defending the action . . .” (§ 996.4), as well as the cost of any judgment or claim paid by the employee (§ 825.2).6

Whether in this case the County must reimburse Farmers and Deputy Nelson for their litigation and settlement costs turns, therefore, on whether Nelson’s conduct towards Deputies Bates, Daugherty, and Murphy was in the scope of his employment. The Legislature intended the phrase “scope of . . . employment,” as used in the California Tort Claims Act, to have the meaning that our courts have given it in decisions “involving actions by third *1042persons against the employer for the torts of his employee.” (4 Cal. Law Revision Com. Rep. (Dec. 1963) p. 814, fn. 3.) Those cases provide that an employer may be held vicariously liable for an employee’s acts committed in the “scope of employment.” (Mary M. v. City of Los Angeles (1991) 54 Cal.3d 202, 208 [285 Cal.Rptr. 99, 814 P.2d 1341]; Perez v. Van Groningen & Sons, Inc. (1986) 41 Cal.3d 962, 967 [227 Cal.Rptr. 106, 719 P.2d 676]; Carr v. Wm. C. Crowell Co. (1946) 28 Cal.2d 652, 654-655 [171 P.2d 5].)

This court recently summarized the principles governing scope of employment: “ ‘A risk arises out of the scope of employment when “in the context of the particular enterprise an employee’s conduct is not so unusual or startling that it would seem unfair to include the loss resulting from it among other costs of the employer’s business. [Citations.] In other words, where the question is one of vicarious liability, the inquiry should be whether the risk was one ‘that may fairly be regarded as typical of or broadly incidental’ to the enterprise undertaken by the employer. [Citation.]” ’ ” (Mary M. v. City of Los Angeles, supra, 54 Cal.3d at p. 209, citing Perez v. Van Groningen & Sons, Inc., supra, 41 Cal.3d 962, 968, and Rodgers v. Kemper Constr. Co. (1975) 50 Cal.App.3d 608, 619 [124 Cal.Rptr. 143], brackets in Mary M.)

Acts that do not benefit the employer may nonetheless fall within the scope of employment; so may acts that are willful or malicious, and those that violate the employer’s express orders or policies. (Mary M. v. City of Los Angeles, supra, 54 Cal.3d at p. 209.)

In deciding whether in a particular case an employee’s conduct falls within the general rule of liability or within one of the exceptions to that rule, California long ago abandoned the “motive” test for determining scope of employment; under that test an act is within the scope of employment only if motivated by a desire to benefit the employer. We discarded this test nearly 50 years ago. (Fields v. Sanders (1947) 29 Cal.2d 834, 838-839 [180 P.2d 684, 172 A.L.R. 525].) As explained in Rodgers v. Kemper Constr. Co., supra, 50 Cal.App.3d at page 621: “Traditionally, before an employer could be held vicariously liable for an employee’s assault, proof was required that the employee intended to benefit or further the interest of the employer. (See 2 Harper & James, The Law of Torts [(1956)], p. 1392.) However, the ‘motive test,’ though still the ‘majority rule,’ has been abandoned in California (Fields v. Sanders, supra, 29 Cal.2d 834, 838-839; Carr v. Wm. Crowell Co., supra, 28 Cal.2d 652, 654; Note, 35 Cal.L.Rev. 126-128) and by federal courts applying federal tort law (Ira S. Bushey & Sons, Inc. v. United States [(1968)] 398 F.2d 167, 170-171).”

In applying the principles discussed above to the facts of this case, the pertinent inquiry is this: were Deputy Nelson’s acts of harassment so unusual or startling as to fall outside the scope of his employment? The *1043majority holds that, as a matter of law, Nelson acted outside the scope of his employment and that therefore this case may be resolved on summary judgment. As I shall explain, such a resolution is impossible on the record before this court.

A motion for summary judgment may be granted only when “there is no triable issue as to any material fact and ... the moving party is entitled to a judgment as a matter of law.” (Code Civ. Proc., § 437c, subd. (c).) Whether an employee has acted within the scope of employment is generally a question of fact, not of law. (Mary M. v. City of Los Angeles, supra, 54 Cal.3d at p. 213; Perez v. Van Groningen & Sons, Inc., supra, 41 Cal.3d at p. 968; Ducey v. Argo Sales Co. (1979) 25 Cal.3d 707, 722 [159 Cal.Rptr. 835, 602 P.2d 755].) It becomes a question of law that may be resolved on a motion for summary judgment only when “the facts are undisputed and no conflicting inferences are possible.” (Perez v. Van Groningen & Sons, Inc., supra, 41 Cal.3d at p. 968.) In this case, because the facts are in dispute, and because conflicting inferences are possible, the case may not be resolved on summary judgment.

True, certain facts—the details of the harassing acts committed by Deputy Nelson—are, for the most part, undisputed. But the facts relating to whether that conduct arose in the scope of employment are hotly disputed. Farmers asserts that Deputy Nelson’s actions were nothing more than sexually oriented joking and “horseplay,” that this behavior was typical of the jailhouse environment, and that Nelson simply did not realize that the women deputies found his conduct offensive. The County, by contrast, denies that Nelson’s conduct was common jailhouse behavior.

If, as Farmers asserts, such behavior was typical of the activities of employees at the jail, then Deputy Nelson’s acts of sexual harassment, although actionable, could not be considered so “unusual” or “startling” that it would be unfair to hold the County vicariously liable for his conduct.7 (Perez v. Van Groningen & Sons, Inc., supra, 41 Cal.3d at p. 968.) But if, as *1044the County contends, Deputy Nelson’s conduct was not typical of the jailhouse environment, then his acts of harassment were outside the scope of his employment.8

To conclude that, as a matter of law, Deputy Nelson’s actions were outside the scope of employment, the majority appears to accept Farmers’ contention that sexually oriented joking and horseplay were typical of the jailhouse environment. The majority asserts, however, that Nelson’s behavior was not horseplay but a serious attempt to sexually solicit and/or assault the women deputies. The majority states the evidence is “undisputed” that Nelson “lewdly propositioned” the women deputies, and that he engaged in “the deliberate targeting of an individual employee by another employee for inappropriate touching and requests for sexual favors” (maj. opn., ante, at p. 997); it reasons that even if sexually oriented joking and horseplay were commonplace at the jail, the evidence “falls far short of establishing that serious misconduct such as asking individual employees for sexual favors and targeting those individuals for inappropriate touching is either typical of or broadly incidental to the operation of a county jail or to the duties and tasks of deputy sheriffs at such a jail” (maj. opn., ante, at p. 1011).

It is by no means certain, however, that Deputy Nelson engaged in “serious misconduct such as asking individual employees for sexual favors *1045and targeting those individuals for inappropriate touching” as claimed by the majority. In its opposition to the County’s motion for summary judgment, Farmers alleged that Deputy Nelson’s conduct was done in a joking manner. As previously noted, the record contains evidence to support Farmers’ allegation; it also contains evidence to the contrary. Thus, whether Nelson was engaged in joking and horseplay is a disputed issue of fact.

As I have noted earlier, a motion for summary judgment may not be granted unless there are no disputed issues of material fact. (Code Civ. Proc., § 437c, subd. (c).) Accordingly, the majority may not base its conclusion that Deputy Nelson acted outside the scope of employment on the “fact” that he targeted the women deputies for sexual favors and inappropriate touching when that “fact” is disputed by Farmers. It is, of course, possible that Deputy Nelson deliberately targeted the women deputies for sexual favors and inappropriate touching, but for purposes of evaluating the County’s motion for summary judgment we must assume that Farmers’ description of Deputy Nelson’s actions is correct. By adopting the County’s characterization of Nelson’s actions, the majority has resolved a disputed issue of fact in the County’s favor. In so doing, the majority errs: disputed issues of fact must be resolved not by an appellate court, whose review is limited to the record of the proceeding in the trial court, but by the trier of fact, whose task it is to weigh the evidence and evaluate the credibility of witnesses at the trial proceedings.

The majority also asserts that Deputy Nelson’s acts of sexual harassment fell outside the scope of employment because they “were motivated for strictly personal reasons unrelated to the guarding of inmates or the performance of any other duty of a deputy sheriff at a county jail.” (Maj. opn., ante, at p. 1007.) The majority, quoting part of a sentence from the lead opinion in John R. v. Oakland Unified School Dist. (1989) 48 Cal.3d 438, 447 [256 Cal.Rptr. 766, 769 P.2d 948], states that an employer is not vicariously liable for the acts of an employee when “ ‘ “it clearly appears that neither directly nor indirectly could [the employee] have been serving his employer.” ’ ” The majority, however, ignores the rest of that sentence, which says, “ ‘where the employee is combining his own business with that of his employer, or attending to both at substantially the same time, no nice inquiry will be made as to which business he was actually engaged in at the time of injury . . . .’” Here, when Deputy Nelson was harassing the female deputies, he was on duty, engaged in his job of guarding the inmates at the jail and “ ‘serving his employer.’ ” The fact that Nelson’s acts of harassment may have benefited himself rather than his employer does not foreclose the possibility that he was acting within the scope of his employment. As this *1046court pointed out in Perez v. Van Groningen & Sons, Inc., supra, 41 Cal.3d at page 969: “There is no requirement that an employee’s act benefit an employer for respondeat superior to apply.” (See also Mary M. v. City of Los Angeles, supra, 54 Cal.3d at pp. 218-219 [To determine scope of employment, “it is necessary to examine the employee’s conduct as a whole, not simply the tortious act itself.”].)

To bolster its conclusion that Deputy Nelson’s conduct was outside the scope of employment, the majority cites case authority from nine states*’ holding acts of sexual harassment to be outside the scope of employment. (Maj. opn., ante, at pp. 1017-1018.) These cases, however, are of little persuasive value here, because almost all of them were decided in jurisdictions that apply the “motive” test in determining whether an employee’s act is within the scope of employment. This test, as mentioned earlier, was discarded in California half a century ago. (Rodgers v. Kemper Constr. Co., supra, 50 Cal.App.3d at p. 621.) Although the motive of an employee is relevant in determining whether the relationship between the actionable conduct and the employment is sufficiently close to arise within the scope of employment, it is not dispositive. Rather, as I have pointed out previously, under California law the relevant inquiry is whether, in light of all the circumstances, the employee’s conduct “is not so unusual or startling that it would seem unfair to include the loss resulting from it among other costs of the employer’s business.” (Perez v. Van Groningen & Sons, Inc., supra, 41 Cal.3d at p. 968.)

Here, the evidence that Farmers presented in support of its motion for summary judgment, and in opposition to the County’s motion for summary judgment, gives rise to a triable issue of fact with regard to whether Deputy Nelson acted within the scope of employment when he sexually harassed Deputies Bates, Daugherty, and Murphy. I therefore disagree with the majority’s holding that, as a matter of law, Deputy Nelson’s conduct was outside the scope of his employment.

III

I now turn to Justice Mosk’s dissenting view, with which I disagree, that as a matter of law Deputy Nelson’s conduct fell within the scope of his employment.

Justice Mosk cites general surveys showing that sexual harassment is common in “traditionally male workplaces, and specifically in county jails, that have recently been integrated by sex.“ (Dis. opn. of Justice Mosk, ante, at p. 1029, italics omitted.) Justice Mosk then observes: “[M]en employed in *1047a traditionally male setting may perceive women entering their workplace as threatening either their job security or their self-esteem and may use sexual harassment as a means to resist the intrusion.” (Id. at p. 1030.) I have no quarrel with that observation. If sexual harassment resulting from male resentment of women performing jobs previously held by men is neither unusual nor startling in a particular enterprise, then it would indeed be proper to conclude that sexual harassment of this nature arises, as a matter of law, within the scope of employment. I question, however, Justice Mosk’s reliance on the surveys in question to conclude that in this case Deputy Nelson’s conduct fell within the scope of his employment. I see two problems with that approach.

First, Farmers does not assert that Deputy Nelson’s conduct arose from resentment of the fact that women deputies had been given a work assignment previously held only by male deputies. Rather, Farmers argues that Nelson’s sexually oriented joking and “horseplay” constituted common behavior at the jail. Absent a claim by Farmers that Deputy Nelson’s behavior stemmed from resentment of the integration of the workforce at the jail, general surveys or studies showing that such resentment causes some male employees to engage in acts of harassment are of little relevance in this case.

Second, in determining whether the trial court in this case properly granted the County’s motion for summary judgment, this court should not rely on general surveys or studies that were not considered by the trial court and are not a part of the appellate record; instead, our task is to examine the record to determine the existence of triable issues of fact. As I noted earlier, a motion for summary judgment may be granted only when “there is no triable issue as to one or more material facts.” (Code Civ. Proc., § 437c, subd. (c).) Here, Farmers asserts that Deputy Nelson’s behavior was typical of the jailhouse environment. The County, however, denies that. Although the evidence may support a finding by the trier of fact that Nelson’s conduct arose in the scope of employment, it would be improper for this court to conclude as a matter of law that the County cannot prevail on its assertion that he was not engaged in typical jailhouse behavior.

IV

It cannot be said that sexual harassment is part of an employee’s job description, or that it advances the employer’s interests. It does not follow from this observation, however, that such conduct automatically falls outside the scope of employment. Nor does the prevalence of sexual harassment in certain work environments automatically establish that such conduct is within the scope of employment.

*1048Instead, whether an employee’s acts of harassment fall within the scope of employment should be decided on a case-by-case basis, by closely examining the nature of the conduct and its relationship to the employer’s enterprise. Under California law, the pertinent inquiry is whether, in the context of a particular enterprise, the conduct is so unusual or startling that it would be unfair to burden the employer with payment of damages resulting from the employee’s conduct. Often, as in this case, the evidence relevant to that determination (which includes evidence not only of the employee’s conduct but also of the workplace context in which it occurred) will be in dispute. In those situations, as in this one, the resolution of those factual disputes, and the ultimate decision as to whether the acts of harassment are linked closely enough to the enterprise so as to fall within the scope of employment, must be left to the trier of fact that hears the evidence.

For this reason, I would reverse the judgment of the Court of Appeal and direct that court to order the trial court to deny the motions for summary judgment made by both Farmers and the County.