OPINION

COOPER, P. J.

SUMMARY

Television writers filed class action lawsuits against studios, networks, production companies and talent agencies, asserting an industrywide pattern and practice of age discrimination. The writers served subpoenas on third parties, including the Writers Guild of America, seeking data on Writers Guild members from which they could prepare a statistical analysis to support their claims of age discrimination. A privacy notice was sent to 47,000 Writers Guild members, advising them of their right to object to disclosure of personal information on privacy grounds. Some 7,700 individuals filed objections. The writers moved to overrule the objections. The trial court sustained the objections in their entirety. The writers sought a writ directing the trial court to vacate its order and allow access to certain of the requested information, arguing the information was critical to proving their claims and privacy concerns were minimal. We grant the writ petition.

FACTUAL AND PROCEDURAL BACKGROUND

This is the latest, but not the last, chapter in a story that began in the year 2000, when a number of television writers filed a class action lawsuit in federal court against various networks, studios and talent agencies, alleging an industrywide pattern or practice of age discrimination. It is unnecessary to recite the history of the litigation or the details of the current lawsuits, all of which are described in Alch v. Superior Court (2004) 122 Cal.App.4th 339 [ 19 Cal.Rptr.3d 29] ( Alch). Suffice it to say that the present litigation involves 23 separate class action lawsuits filed by hundreds of writers (writers or petitioners) against 12 different groups of related television networks, studios and production companies (the employers) and 11 talent agencies. The writers challenge the hiring practices of the employers, and the allegedly discriminatory representation and referral practices of the talent agencies. In addition to claiming disparate treatment on the basis of age, the writers allege the employers use facially neutral practices that have a disparate impact on older writers.

At the time the writ petition now before us was filed, 13 of the lawsuits were assigned to Judge Wendell Mortimer, Jr., nine were assigned to Judge Emilie Elias, and one was assigned to Judge Anthony Mohr. The order now under review was issued by Judge Mortimer and governs the 13 cases in his court. However, the same questions are pending before Judge Elias and Judge Mohr, both of whom await this court's decision on Judge Mortimer's ruling. In the interim, Judge Mortimer has retired, and the cases in his court have been assigned to Judge Elias.

The parties are in the discovery phase of the lawsuits. The writers (and several of the employers or agencies) served subpoenas upon a number of third parties, including the Writers Guild of America, West, Inc.; the Writers Guild of America, East, Inc.; the Producer-Writers Guild of America Pension Plan; the Writers' Guild-Industry Health Fund (collectively, Writers Guild); and four payroll companies used by some of the employers. The subpoenas sought documents including information the writers assert is necessary for statistical analyses of the hiring and representation practices of the employers and talent agencies. The writers intend to compare the age composition of writers actually employed by the networks and studios with the age composition of writers deemed qualified, available and interested in television writing positions, "as best as the latter can be approximated." This information cannot be obtained from applicant records of the employers, both because the employers and agencies did not systematically maintain such records and because the writers claim that older writers were deterred from seeking employment or representation. The writers claim the best approximation of the pool of qualified, available and interested writers is based on Writers Guild membership, refined to eliminate categories of writers likely not to be qualified, interested or available for television writing positions.

The subpoenas served by the writers sought personal information about Writers Guild members, including demographic information (date of birth, race, gender); employment and agency representation records (employment contracts, internal memoranda, correspondence, and e-mail); earnings records; employment application records (resumes, scripts, and other pitch materials); writing qualifications ("script coverage" and other critiques of work submitted); and health and disability records regarding employability. Because the requested information necessarily implicated the privacy rights of nonparty writers, the parties negotiated (and the three judges eventually approved) an order governing notice to all persons whose records were encompassed within the subpoenaed information. The order included a form of privacy notice, an objection form on which recipients could object to the disclosure of all or specific categories of information, and a list of frequently asked questions and corresponding answers. The notice advised recipients that a motion could be filed to overrule any objection. It also advised that a court order would restrict use of and access to the requested records, which would be made available only in connection with the litigation. (The parties had previously agreed to, and the three judges approved, a protective order applicable to all discovery in all the cases.)

The privacy notice was sent to approximately 47,000 individuals whose information was contained in the databases of the subpoenaed third parties. Of this number, some 7,700 recipients objected to the disclosure of some or all of the requested information. The objectors were then sent a second notice, telling them that the writers intended to move to overrule their objections, and advising them of their rights to respond to the writers' motion in writing and at a hearing.

On June 29, 2007, the writers filed a motion "to Modify Protective Order and Partially Overrule Objections to Disclosure of Personal Information." In this motion, the writers conceded the court should uphold the objections as to health and medical records, but argued the court should otherwise overrule the objections. The writers also suggested further enhancements to the existing protective order, including a requirement that all information about identifiable writers (other than petitioners) be treated as "attorneys' eyes only" (the highest level of protection). About 75 of the objectors submitted written oppositions to the motion, as did various employers and agencies.

The parties in all the lawsuits agreed to a joint hearing before the three judges in which objectors could present oral opposition to the writers' motion. Five lawyers appeared at the hearing on September 17, 2007, on behalf of various objectors, and several objectors appeared representing themselves. Each judge then heard separate argument from counsel the following day. On September 19, 2007, Judge Mortimer denied the writers' motion to overrule the objections, sustaining the objections in their entirety. The court stated, inter alia:

— A major issue existed "as to the probative value and usefulness of the desired information." The court referred to the employers' argument that "the end result of the information gathering will not prove anything, and that the information will be flawed," and posed questions as to how those who have no current desire to write for television could be eliminated from the data, and how those who want to write for television but have no credits to date (and are thus not Writers Guild members) could be included. The court referred to the writers' proposal to use a third party neutral expert to eliminate nontelevision writers from the databases before production, but stated this "may be an impossible task and merely allows another opportunity for the information to be leaked." However, the court said it would "not decide at this time whether or not the desired discovery is calculated to lead to the discovery of admissible evidence," instead noting the issues just described.

— While there was "a strong protective order already in place to protect the information sought," no protective order could "assure that all produced information will remain outside the public domain." The court further observed that "[b]oth sides agree that the individuals' names would need to be disclosed and that there is no practical way to redact the names from all the documents."

— Social Security numbers might be necessary in some cases to verify validity of information, and "[t]his, of course, raises another concern, that of identify theft." And, if access to the requested information were allowed, both parties might want to investigate further, including talking to the individuals and others; "[a]n inevitable result is that they could potentially be seen as cooperating with or aligning themselves with the plaintiffs against the defendants," the latter of whom "are their actual or potential agents and past or future employers"; this "could affect their viability in the marketplace."

— "The over 7,700 objectors do not ask for, or want, any part of this lawsuit. They merely want to be left alone."

— The objectors' privacy rights outweighed the public interest in pursuing the litigation. "[N]ot only do we have many personal privacy concerns protected by the Constitution, [but also] the state has an interest in protecting its citizens against fraud and identify theft." And, the 175 individual plaintiffs had "other avenues of proof open to them, as "plaintiffs have stated that they may still be able to put together a meaningful statistical study based upon information from non-objectors."

The court also observed that it would need further briefing and a hearing "[b]efore `private' information is obtained concerning those who did not respond to the privacy notice. . . ."

The writers moved for "clarification and/or reconsideration" of Judge Mortimer's order, to the extent the order protected "four discrete categories of information" from disclosure. Those categories included (1) date of birth and other basic demographic data, and (2) employment data such as a writer's employer, job title, credits and dates of employment; in addition, the writers sought (3) "a link for the multiple databases" containing the demographic and employment history information, as well as (4) two types of anecdotal evidence of discrimination against class members: documents containing words or phrases indicating age was a consideration in hiring and documents containing lists of or references to preferred writers. This information, the writers contended, was the bare minimum necessary to litigate their claims of systemic practices of age discrimination. They argued these discrete categories of information were either publicly available, albeit at great cost in time, or not especially sensitive private information. Judge Mortimer denied the motion.

The writers petitioned for a writ of mandate directing the court to vacate its order and issue a new order permitting access to specified information concerning the objectors. We issued an alternative writ, and the matter was set for hearing on our order to show cause. Two returns were filed on behalf of various employers and talent agencies (collectively, real parties in interest), and others filed joinders to those returns. We permitted the filing of a brief on behalf of 13 of the nonparty objectors (the Wolf objectors). We also allowed an amicus curiae brief in support of real parties in interest from NBC Universal, Inc., and other defendants in one of the cases pending before Judge Elias, as well as an amicus curiae brief in support of the writers by The Impact Fund and several other entities involved in civil rights and public interest advocacy. We now grant the writ.

DISCUSSION

A. Summary of conclusions.

We are well aware that a reviewing court may not substitute its opinion for that of the trial court if there is a basis, supported by the evidence, for the trial court's ruling, and that we may set aside an order denying discovery only if there was no legal justification for the order. ( Tien v. Superior Court (2006) 139 Cal.App.4th 528, 535 [ 43 Cal.Rptr.3d 121].) We also recognize that the trial court was faced, to some extent, with a moving target: the information initially subpoenaed was more comprehensive — and considerably more sensitive on the privacy scale — than the information the writers requested in their motion to overrule the objections, and the latter, too, was more inclusive than the information ultimately sought when the writers asked for reconsideration. These differences, however, highlight the error in the trial court's analysis. It used a broad brush to deny the writers access to all data about the objectors out of hand, and wholly failed to consider whether a more nuanced approach to the different categories of data would satisfy the balance that must be struck between privacy interests and a litigant's need for discovery. (See Valley Bank of Nevada v. Superior Court (1975) 15 Cal.3d 652, 658 [ 125 Cal.Rptr. 553, 542 P.2d 977] ( Valley Bank) [considerations which will affect the exercise of the trial court's discretion in evaluating privacy claims include the "`ability of the court to make an alternative order which may grant partial disclosure'"; where possible, "`courts should impose partial limitations rather than outright denial of discovery'"].)

In short, while the trial court purported to weigh the objectors' privacy rights against the public interest in pursuing the litigation, it failed to follow the dictates of Valley Bank in doing so. In addition to failing to analyze the different categories of data requested, the court gave short shrift to "the public interest in pursuing [the] litigation." Indeed, it erroneously stated that the writers, in their brief, had indicated "that they may still be able to put together a meaningful statistical study based upon information from non-objectors." On the contrary, the writers submitted evidence that no meaningful statistical study could take place if data from the objectors were omitted from it. Under these circumstances, we can reach no other conclusion than that the trial court's orders denying access to any and all data from the objectors were without legal justification.

B. The legal principles.

The principles governing the intersection of the right to privacy under the California Constitution and a civil litigant's right to discovery have been discussed in several cases. But none of the precedents involves circumstances comparable to those in this case. A number of cases involve situations in which the privacy rights of a party to litigation conflict with the other party's discovery rights. And, the Supreme Court has concluded that third party contact information — name, address, and phone number — may appropriately be given to civil litigants when the third parties are given an opportunity to object to the disclosure of their contact information. ( Pioneer Electronics (USA), Inc. v. Superior Court (2007) 40 Cal.4th 360 [ 53 Cal.Rptr.3d 513, 150 P.3d 198] ( Pioneer).) This case, however, requires us to take the next step, and to determine whether it is appropriate to require the disclosure of personal information — beyond contact information — to civil litigants over the objections of the individuals whose information is sought. Our decision is guided by the principles summarized in Pioneer and announced in Hill v. National Collegiate Athletic Assn. (1994) 7 Cal.4th 1 [ 26 Cal.Rptr.2d 834, 865 P.2d 633] ( Hill), which delineate the elements of a cause of action for invasion of the state constitutional right to privacy, and Valley Bank, supra, 15 Cal.3d at pp. 657-658, which noted that, "`[i]n order to facilitate the ascertainment of truth and the just resolution of legal claims, the state clearly exerts a justifiable interest in requiring a businessman to disclose communications, confidential or otherwise, relevant to pending litigation.'" We review the pertinent principles.

— As subsequent cases have confirmed, discovery orders implicating privacy rights are evaluated under the framework established in Hill, and reiterated in Pioneer, supra, 40 Cal.4th at pages 370-371. First, the privacy claimant must possess a legally protected privacy interest, of which there are two general types, autonomy privacy (the interest in making intimate personal decisions or conducting personal activities without observation, intrusion or interference) and informational privacy. Informational privacy — the form at issue in this case — is the interest "in precluding the dissemination or misuse of sensitive and confidential information." ( Hill, supra, 7 Cal.4th at p. 35.) Information in this class is deemed private "when well-established social norms recognize the need to maximize individual control over its dissemination and use to prevent unjustified embarrassment or indignity." ( Ibid.) Second, the privacy claimant must have a reasonable expectation of privacy under the specific circumstances, including "customs, practices, and physical settings surrounding particular activities [which] may create or inhibit reasonable expectations of privacy." ( Id. at p. 36.) Third, actionable invasions of privacy "must be sufficiently serious in their nature, scope, and actual or potential impact to constitute an egregious breach of the social norms underlying the privacy right." ( Id. at p. 37.) Finally, if the three criteria for invasion of a privacy interest exist — a legally protected privacy interest, a reasonable expectation of privacy under the particular circumstances, and a serious invasion of the interest — then the privacy interest "must be measured against other competing or countervailing interests in a `"balancing test."'" ( Pioneer, supra, 40 Cal.4th at p. 371, quoting Hill, supra, 7 Cal.4th at p. 37.) Pioneer explained: "`Conduct alleged to be an invasion of privacy is to be evaluated based on the extent to which it furthers legitimate and important competing interests.' [Citation.] Protective measures, safeguards and other alternatives may minimize the privacy intrusion. `For example, if intrusion is limited and confidential information is carefully shielded from disclosure except to those who have a legitimate need to know, privacy concerns are assuaged.' [Citation.]" ( Pioneer, supra, 40 Cal.4th at p. 371, quoting Hill, supra, 7 Cal.4th at p. 38.)

— In Pioneer, the Supreme Court considered the extent to which California's constitutional right to privacy protected purchasers of possibly defective DVD players, who had already complained to the seller, "from having their identifying information disclosed to plaintiff during civil discovery proceedings in a consumers' rights class action against the seller." ( Pioneer, supra, 40 Cal.4th at p. 363.) The court concluded that the trial court's order — which required written notice of the proposed disclosure to all complaining customers, and gave them the opportunity to object to the release of their own identifying information — involved no serious invasion of privacy. The court observed that the disclosure "involves no revelation of personal or business secrets, intimate activities, or similar private information, and threatens no undue intrusion into one's personal life, such as mass-marketing efforts or unsolicited sales pitches." ( Id. at p. 373.) The court observed that its conclusion there was no serious invasion of privacy "could end our inquiry" without any balancing of interests, but it proceeded to examine the respective interests involved, reinforcing its conclusion that the trial court's order was not an abuse of discretion. ( Ibid.)

— In Valley Bank, a case predating Hill, the Supreme Court engaged in "a careful balancing of the right of civil litigants to discover relevant facts, on the one hand, with the right of bank customers to maintain reasonable privacy regarding their financial affairs, on the other." ( Valley Bank, supra, 15 Cal.3d at p. 657.) The court, "[s]triking a balance between the competing considerations," concluded that before a bank could disclose private information, it was required to notify the customer of the proceedings and give him a fair opportunity to assert his interests "by objecting to disclosure, by seeking an appropriate protective order, or by instituting other legal proceedings to limit the scope or nature of the matters sought to be discovered." ( Id. at p. 658.) In the course of its discussion, the court observed that "we readily acknowledge that relevant bank customer information should not be wholly privileged and insulated from scrutiny by civil litigants." ( Id. at p. 657.)

— The parties in this case agree that, in the discovery context — and assuming a serious invasion of a privacy interest has been established — the writers must show they have a "compelling need" for the data requested, and they may do so by showing the information is "directly relevant" and "essential to the fair resolution" of the lawsuit. ( Britt v. Superior Court (1978) 20 Cal.3d 844, 848, 859 [ 143 Cal.Rptr. 695, 574 P.2d 766] ( Britt) [defendants sought discovery of "extensive and intimate details of both [plaintiffs'] own and others' activities in various local political associations"; a trial court could properly compel disclosure only if "such associational activities are directly relevant to the plaintiffs claim, and disclosure of the plaintiffs affiliations is essential to the fair resolution of the lawsuit . . ."]; Save Open Space Santa Monica Mountains v. Superior Court (2000) 84 Cal.App.4th 235, 252 [ 100 Cal.Rptr.2d 725] ( Save Open Space) [case involving associational privacy rights of third parties; right of associational privacy is not absolute, but party seeking discovery of private matters "must do more than satisfy the relevancy standard"; party is required to demonstrate a compelling need for the discovery, and the "`compelling need must be so strong as to outweigh the privacy right when these two compelling interests are carefully balanced'"].)

— In evaluating privacy claims, "considerations which, among others, will affect the exercise of the trial court's discretion" include "`the purpose of the information sought, the effect that disclosure will have on the parties and on the trial, the nature of the objections urged by the party resisting disclosure, and ability of the court to make an alternative order which may grant partial disclosure, disclosure in another form, or disclosure only in the event that the party seeking the information undertakes certain specified burdens which appear just under the circumstances.'" ( Valley Bank, supra, 15 Cal.3d at p. 658; see also Schnabel v. Superior Court (1993) 5 Cal.4th 704, 714 [ 21 Cal.Rptr.2d 200, 854 P.2d 1117] ( Schnabel) [in weighing the privacy interests of a third party in a marital dissolution proceeding, "the court should consider the nature of the information sought, its inherent intrusiveness, and any specific showing of a need for privacy, including any specific harm that disclosure of the information might cause"]; Puerto v. Superior Court (2008) 158 Cal.App.4th 1242, 1251 [ 70 Cal.Rptr.3d 701] ( Puerto) [if there is a reasonable expectation of privacy and the invasion of privacy is serious, "then the court must balance the privacy interest at stake against other competing or countervailing interests, which include the interest of the requesting party, fairness to the litigants in conducting the litigation, and the consequences of granting or restricting access to the information"]; Save Open Space, supra, 84 Cal.App.4th at p. 253 ["`strength of an individual's interest in keeping personal information private depends in large part on the consequences of disclosure'"].)

C. The issues presented in this case.

We emphasize that our decision addresses the information identified in the writ petition. With respect to those information requests — described in the writ petition and in the writers' motion to the trial court for reconsideration — the trial court erred in refusing to overrule the objections on privacy grounds. To be clear, we delineate in the margin the data categories at issue, which consist of demographic and work history information. The data requested does not include medical and disability records, and does not include financial earnings records.

Our overall conclusion as to the data requested is that a serious invasion of a reasonable expectation of privacy was established under the Hill and Pioneer criteria, requiring the privacy invasion to "`be evaluated based on the extent to which it furthers legitimate and important competing interests.'" ( Pioneer, supra, 40 Cal.4th at p. 371, quoting Hill, supra, 7 Cal.4th at p. 38.) We further conclude, however, that the writers demonstrated that the information requested is "directly relevant" to their claims and "essential to the fair resolution" of their lawsuit. ( Britt, supra, 20 Cal.3d at p. 859, italics omitted.) And finally, the trial court erred in its purported balancing of the privacy rights of the objectors and the countervailing interests of the litigants and the state, failing to analyze the types of information requested, failing to consider the state's interest in preventing discrimination, and erroneously concluding the writers could proceed with a statistical analysis without information from the objectors. We discuss each of these points in turn.

1. The criteria for invasion of a privacy interest were established.

As a preliminary matter, the writers claim their data requests, as now formulated, do not require the court to balance the privacy interests of the objectors against the discovery interests of the writers, because there is no serious invasion of a privacy interest. ( Pioneer, supra, 40 Cal.4th at p. 371; Hill, supra, 7 Cal.4th at p. 37.) We agree that the data requests no longer include the more sensitive items, and we are cognizant that in Hill's words, the invasion of privacy must be "sufficiently serious in [its] nature, scope, and actual or potential impact to constitute an egregious breach of the social norms underlying the privacy right." ( Hill, supra, 7 Cal.4th at p. 37.) Nonetheless, even with the omission of the patently sensitive items such as medical and financial records, the data requests include 14 or 15 different items of information for each Writers Guild member. This agglomeration of information, including as it does name and work history information, goes well beyond the information disclosures that courts have found do not amount to serious invasions of privacy. (See Pioneer, supra, 40 Cal.4th at pp. 372-373; Puerto, supra, 158 Cal.App.4th at p. 1253.) Indeed, the writers' own reliance on the existing protective order to assuage privacy concerns demonstrates that those concerns exist. A protective order is a safeguard for the protection of information admittedly subject to privacy concerns. Accordingly, while we think the degree of sensitivity of the information to be disclosed may properly be considered when the court balances the competing interests, we conclude that the data requests are an invasion of privacy which must "`be evaluated based on the extent to which it furthers legitimate and important competing interests.'" ( Pioneer, supra, 40 Cal.4th at p. 371, quoting Hill, supra, 7 Cal.4th at p. 38.)

2. The writers demonstrated that the requested information is directly relevant to their claims and essential to the fair resolution of these lawsuits.

We entertain no doubt that the writers have demonstrated that the information requested is "directly relevant" to their claims and "essential to the fair resolution" of their lawsuit. ( Britt, supra, 20 Cal.3d at p. 859, italics omitted.) The writers have alleged both claims of disparate treatment (a pattern and practice of intentional discrimination), and "disparate impact" claims based on employment practices that are facially neutral but have a disparate impact on older writers. Statistical proof is indispensable in a disparate impact case: "`The plaintiff must begin by identifying the specific employment practice that is challenged.'" "`Once the employment practice at issue has been identified, causation must be proved; that is, the plaintiff must offer statistical evidence of a kind and degree sufficient to show that the practice in question has caused the exclusion of applicants for jobs or promotions because of their membership in a protected group.'" ( Carter v. CB Richard Ellis, Inc. (2004) 122 Cal.App.4th 1313, 1323-1324 [ 19 Cal.Rptr.3d 519], quoting Watson v. Fort Worth Bank Trust (1988) 487 U.S. 977, 994 [ 101 L.Ed.2d 827, 108 S.Ct. 2777].) The writers cannot prove their disparate impact claims without access to evidence from which they can perform a statistical analysis. And while, at least in theory, a pattern and practice of discrimination (the disparate treatment claim) may be established without statistical evidence, plaintiffs "normally seek to establish a pattern or practice of discriminatory intent by combining statistical and nonstatistical evidence, the latter most commonly consisting of anecdotal evidence of individual instances of discriminatory treatment." (1 Lindemann Grossman, Employment Discrimination Law (3d ed. 1996) p. 45.) Consequently, we find the conclusion is virtually inescapable that the writers have a "compelling need" for the information they have sought, which is clearly "directly relevant" to their claims and "essential to the fair resolution" of their lawsuit.

The employers and agencies resist this conclusion with one principal argument. They contend that in order to establish a compelling need for the data, the writers must first prove that, if access is provided, they will be able to "turn the massive amount of requested data into admissible evidence." Absent such proof, they say, the writers are merely speculating that the data can be transformed into admissible evidence. Ergo, the writers have not shown the data is "directly relevant" and "essential" to the fair resolution of the lawsuit. Specifically, the real parties in interest's argument proceeds this way:

(1) A valid statistical analysis must compare the ages of persons actually hired or represented, and the ages of individuals who were qualified, interested, and available to be hired or represented.

(2) Writers seek to create a so-called "proxy pool" of qualified, interested and available writers based on the subpoenaed data of Writers Guild members.

(3) The data requested is "inherently flawed" because it is "massively over and underinclusive" of writers actually qualified, interested, and available to write for television. It is overinclusive because it will include not just television writers, but writers in other genres (feature film, radio, theater, and so on) who have no interest in or qualifications for writing for television, and it is underinclusive because it will omit writers who are interested and qualified but who are not Writers Guild members.

(4) Writers "never credibly explained or presented any evidence" as to how these flaws would be surmounted; the writers' contentions they could surmount the problems "are simply not evidence."

Real parties in interest's argument is, in effect, a claim that, because privacy interests are involved, the writers must prove that the data they seek will prove their case before they may have access to the data. But there is no support in law, or in logic, for this claim. First, we are aware of no precedent, and the employers and agencies cite none, for the proposition that a statistical study must be proved valid in advance of its performance simply because the underlying data is subject to privacy claims. Indeed, we know of no principle requiring subpoenaed information to be proved "admissible" in advance of its production. Second, such a rule would be wholly impractical and unreasonable in the context of class action litigation requiring complex statistical analysis. At this stage of the discovery process, uncertainty is inevitable about the "kind and character" of data contained in the multiple databases that have been subpoenaed from multiple third parties. Some information in the databases doubtless will be, in the end, irrelevant or unusable for any number of reasons, including the subject's lack of interest or availability for television writing. But that does not mean that the overall body of information subpoenaed — demographic and work history information of Writers Guild members — is not directly relevant and essential to the writers' case.

The writers acknowledge that they cannot, at this stage, identify precisely how their expert will attempt to distinguish potential applicants for employment from persons in the data who are not potential applicants. They cannot do so until after data discovery, and discovery on the employers' hiring policies, has occurred. The writers "can state only that their expert will consider the data, policy statements and labor economic theories." As the writers' expert Madden explained: "I cannot determine the specific characteristics or qualifications that should be included in the analyses of whether age affects hire until I have reviewed the data and the hiring policies and practices." Madden did not have a complete list of all items in the relevant databases, "let alone information on which of these items contain reliable information for the majority or for all of the writers. Therefore, I cannot even know which characteristics the databases will allow me to consider."

In short, we cannot accept the notion that information is not "directly relevant" or "essential" unless proof is first produced as to how anticipated (and acknowledged) problems in the data will be controlled for or remedied. Statistical analysis in complex cases cannot be perfect; distinctions and adjustments are made as the data may allow. Certainly it is possible that, in the end, the statistical analysis presented by the writers and their expert witnesses will be flawed, or will not show what they hope it will show. But the contrary is also possible, and the answer cannot be known until both the data and information on specific hiring practices is available for analysis. While the employers and agencies pose a laundry list of questions about how the writers will account for various qualifications and restrictions they say must be factored into a proposed proxy pool, this is not the time for those questions. The time for those questions is after the writers have assessed the data available and have constructed their statistical analysis, which they cannot do absent the data they seek. We note that the real parties in interest have adduced no evidence that a viable statistical analysis cannot be done, instead relying solely on their claim that the writers must prove it can be done before they may have access to the data. But this court is in no position to conclude — nor was the trial court (nor did it do so) — that the writers will be unable to construct a statistical showing that is both admissible and persuasive. The idea that they must prove they can do so, before they have a complete picture of the data available for the analysis, is a proposition unsupported by either precedent or common sense.

3. The trial court abused its discretion when it concluded the writers' need for data to prove their case was outweighed by the objectors' privacy interests.

As we have seen, the objectors have serious privacy concerns, but at the same time, the writers have shown that the data they seek is essential to the fair resolution of their lawsuit. Accordingly, as Hill and Pioneer make clear, the trial court was required to evaluate the invasion of the objectors' privacy "based on the extent to which it furthers legitimate and important competing interests." ( Hill, supra, 7 Cal.4th at p. 38.) Valley Bank listed a number of considerations that would affect the exercise of the trial court's discretion in evaluating privacy claims. ( Valley Bank, supra, 15 Cal.3d at p. 658.)

In this case, the trial court refused to require the production of any of the objectors' information, without regard to the strength of the privacy interest in the particular category of data. This failure ignores the teaching of Valley Bank, which requires the court to consider its "`ability . . . to make an alternative order which may grant partial disclosure,'" and which expressly stated that, "[w]here it is possible to do so, `. . . the courts should impose partial limitations rather than outright denial of discovery.'" ( Valley Bank, supra, 15 Cal.3d at p. 658.) The court's wholesale denial of access to the objectors' information was thus necessarily an abuse of discretion. The court also failed to consider the effect that denial of access to the data would have on the writers' ability to prove their claims, indeed, erroneously stating the writers had suggested they can make a statistical case without the objectors' information. It did not consider the "`nature of the objections urged by the party resisting disclosure'" ( ibid.), or the consequences to the objectors of disclosure of the different categories of information ( Save Open Space, supra, 84 Cal.App.4th at p. 253), except to raise the specter of identity theft. And, it failed to give any real consideration to the effect of the protective order in assuaging the privacy concerns. We address the major categories of information sought by the writers, along with the errors in the trial court's analysis.

a. Work history information.

The most sensitive category of data the writers now request is the work history information. Even here, however, the data sought is not of the kind involved in cases refusing to permit the disclosure of third party personnel files, and, in contrast to those cases, here the writers have shown a compelling need for the information. For example, in Harding Lawson Associates v. Superior Court (1992) 10 Cal.App.4th 7 [ 12 Cal.Rptr.2d 538] ( Harding Lawson), the court concluded that the trial court's discovery order was overbroad and unjustified, "[t]o the extent the court's discovery order required production of confidential information contained in personnel files of employees other than [plaintiff]." ( Id. at p. 10.) The court observed that the plaintiff "has not shown a compelling need for particular confidential documents in third party personnel files or that the information could not be obtained through depositions or from nonconfidential sources." ( Ibid.; see also Board of Trustees v. Superior Court (1981) 119 Cal.App.3d 516, 526 [ 174 Cal.Rptr. 160] ( Board of Trustees) [university's personnel, tenure and promotion files relating to a defendant faculty member had no direct relevance to plaintiffs defamation action; plaintiff "merely argues that such disclosure might lead to the required proof of malice of one or more of the several defendants"].)

This case is entirely different from either of Harding Lawson or Board of Trustees. First, the data sought includes the identity of the person's employer and talent agency representative, job title, period of employment, the productions on which he or she worked, his or her credits and awards, and so on. (See fn. 8, ante.) The writers no longer seek other sensitive information ordinarily found in personnel files, such as evaluation of the person's work (script coverages and other critiques), income information, employment contracts and the like. Moreover, as the Supreme Court observed in Hill, "customs, practices, and physical settings surrounding particular activities may create or inhibit reasonable expectations of privacy." ( Hill, supra, 7 Cal.4th at p. 36.) In the context of screenwriting activity, it is difficult to see how or why a writer might expect the identity of a studio or production company employing him or her to remain confidential, and the same is true of other work history information such as the productions on which a writer has worked, credits and awards and the like. Indeed, some of the information can be retrieved through Internet searches, albeit with an impractical investment of time when tens of thousands of persons are involved. Consequently, we do not see why these work history records, which do not include evaluative materials, would be considered especially sensitive, under the particular circumstances. And, the trial court identified no untoward consequences of the disclosure of work history information of the type now sought. The information is essential to the resolution of the writers' claims of industry-wide discrimination; it will be subject to a comprehensive protective order which all the parties agree is, and the trial court described as, "strong"; and it involves "no revelation of personal or business secrets, intimate activities, or similar private information, and threatens no undue intrusion into one's personal life. . . ." ( Pioneer, supra, 40 Cal.4th at p. 373.) Under these circumstances, the trial court abused its discretion in not permitting access to work history data.

b. Demographic information.

The other major category of data requested was demographic information, including (a) name, (b) date of birth, (c) date of death (if applicable), (d) gender, (e) race, and (f) residential ZIP code. Again, the court failed to consider any of the factors it should have considered in assessing the strength of the objectors' privacy concerns, and in comparing those concerns with the consequences to the litigants of refusing access to the data. ( Valley Bank, supra, 15 Cal.3d at p. 658.) It did not consider the "`nature of the objections urged by the party resisting disclosure'" ( Valley Bank, supra, 15 Cal.3d at p. 658), and it did not consider "the consequences of granting or restricting access to the information." ( Puerto, supra, 158 Cal.App.4th at p. 1251.) Instead, the court merely observed that some of the objectors "felt so strongly about objecting" that they hired attorneys or personally appeared; some were well-known individuals; and "[a]lthough their reasons and their concerns differ," their objections were all grounded in their right to privacy. But, had the trial court assessed the nature of the objections in relation to the categories of information initially sought, it would have found, for example, that the "well-known individuals" — presumably the Wolf objectors — were concerned about financial, medical, and evaluative information, none of which is sought in this writ petition, and none of which was sought in the writers' motion to the trial court for reconsideration. Indeed, in their brief to this court, the Wolf objectors define "private information" to exclude demographic information.

In short, while the demographic data is personal information, it is hardly sensitive information. (See Puerto, supra, 158 Cal.App.4th at p. 1253 [contact information "while personal, is not particularly sensitive, as it is . . . not medical or financial details, political affiliations, sexual relationships, or personnel information"].) And, on the other hand, demographic information is obviously critical to any statistical analysis. Dates of birth, for example, are patently essential to an age discrimination case. All the factors we have identified as weighing in favor of disclosure with respect to the work history information apply with respect to the demographic information: the consequences of nondisclosure are dire for the writers' case, while the consequences of disclosure for the objectors are relatively inconsequential, particularly in view of the existence of a strong protective order limiting access to the information. The only point cited by the trial court that marginally relates to the consequences of disclosure of demographic information is the interest in preventing identity theft, and that was raised in connection with Social Security numbers (which are not sought by the writers and are discussed post). So, we again think it plain that the trial court abused its discretion in sustaining the objections to disclosure of demographic information.

We note a few comments concerning the remaining items the writers seek.

— Petitioners seek to use a "unique identifier" for each Writers Guild member, so that the multiple databases containing different items of the requested information can be linked by their experts. Petitioners concede that Social Security numbers are sensitive private information. However, when the privacy notice was sent to all Writers Guild members, the databases from eight different entities — with names, Social Security numbers and contact information — were given to a neutral contractor, who replaced the Social Security number with the same random number each time the Social Security number appeared. The writers proposed using the same system, so that only the neutral contractor would have access to Social Security numbers. Nonetheless, the trial court stated that the writers had "candidly admit[ted] that Social Security numbers may be necessary in some cases to verify validity of information," and that "[t]his, of course, raises another concern, that of identity theft." Again, the trial court appears simply to have ignored completely the existence of protective measures and safeguards commonly used to assuage privacy concerns (see Pioneer, supra, 40 Cal.4th at p. 371 ["[p]rotective measures, safeguards and other alternatives may minimize the privacy intrusion"]), including a known administrative method of avoiding production of Social Security numbers to the litigants.

— In addition to the work history and demographic data, the writers' motion to the trial court for reconsideration identified two types of anecdotal information they say is critical to their lawsuits: (a) documents containing words or phrases that might indicate that age was a consideration in hiring decisions; and (b) documents consisting of or containing lists of or references to preferred writers. The trial court did not directly address this document request. However, the writers state that, by treating the objectors' names as private information, the trial court's order has the effect of requiring real parties in interest to withhold from production documents in the two requested categories — those with age-based comments and references to preferred writers — if they contain any names (except names of petitioners or of persons known not to be objectors). Similarly, petitioners would be unable to ask deposition questions concerning incidents involving anyone other than themselves "for fear it would disclose information about an objector." The real parties in interest do not address this aspect of the writers' information requests and, given our ruling that it was error to deny the writers access to the objectors' names along with the other demographic information, we need not linger long on the point. It is clear, however, that to establish a pattern and practice claim, the writers have a compelling need for precisely this kind of document, which would be directly relevant to proving intentional discrimination. Moreover, beyond names, documents containing age-based comments and lists of preferred writers will not necessarily contain any private information about the subjects of those documents. To the extent such documents do contain information implicating privacy concerns, redaction procedures and other remedies may be employed to resolve those concerns on a case-by-case basis.

c. Other considerations in the balancing process.

A few more observations about the trial court's analysis in this case may be useful. First, as we have concluded, the court's analysis was fundamentally flawed because it failed to consider the significant differences in privacy concerns attaching to different categories of information, and consequently failed to consider the possibility of requiring partial disclosure rather than denying discovery outright. The court's error may have been generated in part by its apparent failure to give any consideration to the significance of the state's interest in preventing invidious forms of discrimination. The court's decision, as it should, describes the constitutional right to privacy in considerable detail, and then states that those rights outweigh the public interest in pursuing this litigation. But the court does not describe "the public interest in pursuing this litigation." Instead, it returns to the "many personal privacy concerns protected by the Constitution" and points to the state's interest "in protecting its citizens against fraud and identity theft." The omission of any reference to the nature of the public interest in this case is, we think, telling. These are putative class actions alleging an industrywide pattern and practice of age discrimination. The state has an interest in the ascertainment of truth in all legal proceedings in its courts. This interest is accentuated in cases of discrimination, as California unquestionably "has a legitimate and compelling state interest generally in the battle against discrimination on the basis of race, gender, age, national origin, or other invidious categories of discrimination." ( Pacific-Union Club v. Superior Court (1991) 232 Cal.App.3d 60, 79 [ 283 Cal.Rptr. 287]; see Department of Fair Employment Housing v. Superior Court (2002) 99 Cal.App.4th 896, 905 [ 121 Cal.Rptr.2d 615] [state's compelling interest in protecting against discrimination by marital status outweighs intrusion on the privacy of third party tenants and applicants that might result from disclosure of rental application records].) The state likewise has an interest in ensuring that undue burdens are not imposed on the class action device in discrimination cases. (See Keating v. Superior Court (1982) 31 Cal.3d 584, 609 [ 183 Cal.Rptr. 360, 645 P.2d 1192] [Supreme Court has "repeatedly emphasized the importance of the class action device for vindicating rights asserted by large groups of persons"].)

Second, it appears the trial court's analysis may also have been skewed by the sheer number of persons who objected to the disclosure of their personal data. However, the number of objectors has no apparent relevance to the balancing process. While 7,700 people objected, at least four times that number of people did not object. And, if a privacy invasion is indeed serious, we cannot see how it may be rendered more — or less — serious based on the mere number of persons subjected to the privacy invasion. (See Puerto, supra, 158 Cal.App.4th at p. 1255 ["[i]t appears that the large number of witnesses identified by [defendant], rather than the actual significance of the privacy invasion with respect to each witness, may have impacted the [trial] court's analysis"; Court of Appeal saw "no manner in which the mere numerosity of witnesses alters the underlying analysis of the seriousness of the intrusion on the witnesses' privacy rights"].)

Third, it is not clear how much weight the court gave to identity theft, but it mentioned the state's interest in protecting its citizens from fraud and identity theft, while failing to mention the state's interest in preventing invidious discrimination. Certainly many of the objectors mentioned identity theft as a concern, and real parties in interest assert — without citation of authority — that the danger of identity theft "increases exponentially" when information is "assembled in one location." Nonetheless, we do not think the theoretical possibility of identity theft is entitled to any significant weight when a court is balancing privacy interests against the interests of civil litigants in pursuing discrimination claims. Identity theft is theoretically possible in scores of ways from scores of sources in our daily lives. There is nothing to suggest that this litigation presents an opportunity for identity theft any greater than exists in all civil litigation, and in all aspects of ordinary life in the information age. To suggest that litigation of civil rights or any other matters should not proceed because of a remote and undifferentiated possibility that private information may be stolen is simply not an acceptable solution to general concerns about identity theft.

Fourth, the trial court appeared to give no weight at all to the admittedly "strong protective order already in place to protect the information sought," observing that "no protective order can assure that all produced information will remain outside the public domain." The latter observation is unquestionably true; no guarantee can be made that information will not be leaked or stolen. But a guarantee "that all produced information will remain outside the public domain" is not and cannot be the standard for assessing a protective order; if it were, the production of private information would never be compelled. The court's failure to address the significance of a strong protective order in reducing privacy concerns, and to articulate why the order in this case did not suffice with respect to any of the information requests, again disregards the teaching of Pioneer. "`[I]f intrusion is limited and confidential information is carefully shielded from disclosure except to those who have a legitimate need to know, privacy concerns are assuaged.'" ( Pioneer, supra, 40 Cal.4th at p. 371.)

d. Other considerations urged by real parties in interest.

The employers and talent agencies argue that, if this court permits access to the requested data, we must "be prepared to allow the production of all of it," including income, health and disability status information. Otherwise, they say, their "due process right to all information relevant to such persons' qualifications, interest and availability" will be violated. But that is not a question presented by this writ petition, and the record in this case provides no basis for an answer in any event. Real parties in interest did not move to overrule objections to disclosure of medical and financial information, and have no basis for requesting production of additional data in this forum. And no due process rights are implicated in any event; the employers and talent agencies will have ample opportunity in the trial court to assert that any statistical analysis the writers present is defective, whether because of the absence of medical or financial data or for any other reason.

DISPOSITION

The writ petition is granted. Let a peremptory writ of mandate issue directing the trial court to vacate its orders of September 19, 2007, and November 16, 2007, denying the writers' motion to overrule objections and denying their motion for reconsideration, and to enter a new and different order granting the writers access to the requested information to the extent consistent with the views expressed in this opinion. Costs are awarded to petitioners.

Chavez, J., concurred.