Testing to prove racial discrimination: methodology and application in Hungary

03 October 2000

Fitsum Alemu1

When discrimination occurs in Hungary today, it is often subtle and clandestine. Direct evidence of discrimination is rare, and where such direct evidence exists, corroboration is even more rare. As a result, those complainants who allege that they have been subjected to discrimination are not likely to be able to confirm the discriminatory intent of the employer, bar owner or realtor. "Testing", a technique developed primarily in the United States, can be enormously helpful in proving facial discrimination and disparate impact cases.2 Testing can be used for auditing purposes, to monitor compliance with settlements in earlier cases and to challenge discrimination in court. In the recent Patvárc case, a Hungarian court recognised for the first time testing as a valid technique for documenting discrimination.3 The court heard "testers" as witnesses and accepted the results of testing as evidence pertaining to a complaint filed by two Roma concerning allegations of discrimination in public accommodations.

What is testing?

Testing is a technique that is used to collect evidence when there is an allegation of discrimination. Testing is used mainly by civil rights organisations to uncover unlawful acts of discrimination. It is applied if a member of a protected class group suspects disparate treatment on grounds of his or her national origin, religion, gender, the colour of his or her skin, or other characteristics covered by legal prohibitions on discrimination. It is applied to gauge the existence or extent of discrimination in employment, housing, public accommodation or, indeed, any other area of social life.

There are two kinds of testing: research-oriented testing, which is used for auditing, and enforcement-oriented testing, which uses the results of testing to file a law suit or monitor compliance with injunctive relief. In enforcement-oriented testing, it is often appropriate to perform repeated tests of the same job vacancy, the apartment at issue, or the bar or club suspected of refusing to serve based on racial grounds. The goal of repeated tests is to assess the nature and extent of discrimination in anticipation of litigation, principally to determine whether the observed differences in treatment were isolated or reflect a pattern or practice of discriminatory behaviour.4

The test and the testers

In enforcement-oriented testing, first litigators should discuss the case with the complainant to draw up questions to be addressed by testing. One should also collect all materials concerning the firm or the club being tested, such as licences and earlier complaints against the firm, as well as legal provisions and case law. Then the selection and training of testers begins.

Testers are objective fact-finders who, after extensive training in both the classroom and the "field", conduct testing to uncover discrimination. A test requires two testers: a "protected tester" and a "comparison tester". A protected tester is a member of a protected group, and a comparison tester is from a non-protected group. For example, in cases alleging discrimination against a Romani person, a Romani person would serve as a protected tester, while a non-Romani person would be in the role of a comparison tester. In a case of gender discrimination against a woman, a woman would be in the protected tester status while a man would be in the role of comparison tester. In general, testers should be quite similar. The key difference should be the quality at issue in the "test", for example, the race or national origin of the tester where racial discrimination is alleged.

Training should include practice testing under close supervision, orientation about the uses of testing results to enforce civil rights laws and information regarding the nature of legal procedures in which testers may eventually be involved. During the training, paired testers should work closely with each other, get to know each other, and develop a sense of teamwork. Testers should be asked to declare explicitly that they accept the roles in the project as objective fact-finders, and to promise to maintain confidentiality.

Testers conduct their tests on the same day, posing as bona fide job or home seekers, for example. In the process of the test, testing team partners are sent at closely spaced intervals5 to seek information about a job, an apartment or the availability of a certain service. When conducting a test, testers should dress appropriately for the occasion. In testing employers, each tester should take actions that are comparable to those likely to be undertaken by his or her paired partner while still following the natural flow of each job application process. For instance, the protected class should apply first for the job at issue either by telephone or in person. Tailoring testers' conduct to the particular circumstances of each job application, maintaining a clear and complete record of the test experience, and ensuring that each tester acts in ways comparable to his or her partner is necessary to obtain evidence for litigation.

Testers record their experiences on assignment forms immediately after completion of each test. The report filed by each tester should include detailed information about job or housing availability, the application process, terms and conditions, questions asked by the tester and information volunteered by the agent or the employer. Beyond answers to the questions in the report forms, it may be worthwhile to request that testers write a detailed narrative description of their experiences during the test.

The forms on which testers record their experience should include at least the following: the time of application; information demanded of applicants (e.g. the length of interviews, the characteristics of interviews, questions asked at interviews); the flow of information (e.g. information provided spontaneously, information that had to be requested); how applicants are treated (e.g. length of time they must wait, level of hospitality offered); manner in which jobs are described (e.g. discussion of salaries and benefits, the length of employment, and so on).6 Such well-organised testing aims at examining minute, discrete components of the hiring process and can be used to corroborate as well as dispel allegations of discrimination that have been levelled against an employer. Evidence of the ultimate disparity - that one tester was offered a job while the other was not - should be documented as carefully as possible.7

It is not the role of the tester to determine whether or not discrimination has occurred, but rather to act as an unbiased recorder of information. Only the test coordinator (the organisation or the attorney) can evaluate whether or not differential treatment has taken place. During the test, the tester should refrain from making any leading remarks about race or ethnicity in the neighbourhood, in the work place or the club they are testing; testers should be observant, meticulous record-keepers so that their experiences will be completely and accurately documented. They should record their experiences independently and should not discuss their experiences with each other until after they have been documented. Under no circumstances should a tester discuss the testing experience or the institution tested with anyone unless authorised by the test coordinator or ordered by a court. If differential treatment is established, then the organisation can file a lawsuit against the perpetrator.8

Enforcement-oriented testing may call upon testers to serve as plaintiffs and witnesses in litigation. This imposes at least three additional considerations in selecting testers. First, the personal backgrounds of testers must be free from any difficulties that might reduce their credibility as witnesses. Second, testers must be sufficiently articulate to present their experiences clearly in written witness statements and oral testimony. Third, because litigation may last for several years, testers must be willing to remain in contact with the testing program and return periodically to participate in legal proceedings over an extended period.9

Though finding capable and dedicated individuals to serve as testers is difficult, experience shows that it is possible. Some organisations hire students from local universities. Others hire research assistants to conduct tests. Most, however, identify candidates through personal referrals rather than open advertisements.

Testing in U. S. case law

Starting in the 1960s, U.S. courts decided a number of important cases that recognised testing as an effective means and, in some instances, the only method of proving discrimination. Today, testing is uniformly accepted by lawmakers, courts and administrative bodies in the U.S.


Despite the power of testing as a basis to challenge discrimination, perhaps because of that power, the use of testing to prove discrimination has come under considerable attack.10 Opponents argue that the tester has no standing. Others say testing is improper and even unethical because it involves deception. U.S. courts have dismissed most of these arguments and found that testers have standing, and that they are harmed if they are denied services or given false information about the availability of jobs or housing. For instance, in McDonnell Douglas Corporation v. Green,11 the U.S. Supreme Court found that when testers received differential treatment mirroring the allegations of discrimination, an inference of discrimination may be drawn. Some other courts ruled that testers have the same right to truthful information about the availability of possible contracts under Section 1982 of the Civil Rights Act12 as anyone else, and they have the same right to negotiate for such contracts.13

In Evers v. Dwyer14 the Supreme Court reversed a lower court's decision and held that a black man who chose to sit in the white section of a segregated bus solely to test the lawfulness of the segregation policy suffered harm protected by civil rights laws. In this case, the appellant boarded a Memphis bus on April 26, 1956, and seated himself at the front of the vehicle. The driver told him he must move to the rear, "stating that the law required it because of [his] color."15 Following the appellant's refusal to comply, two police officers boarded the bus and "ordered [the appellant] to go to the back of the bus, get off, or be arrested."16 After this warning, the appellant left the bus. The lower court accepted the defendant's argument that Mr Evers could not have standing and allege an actual controversy because he "boarded the bus for the purpose of instituting this litigation," and therefore was not "representative of a class of coloured citizens who used the buses in Memphis as a means of transportation."17 In overruling the District Court, the Supreme Court held there was an actual controversy because Mr Evers was subject to "special disabilities" and therefore had a "substantial, immediate, and real interest in the validity of the statute which imposes the disability."18

In its decision in Pierson v. Ray,19 the Court found that a black man had been harmed when he was arrested after entering the segregated section of a bus station with the purpose of testing the legality of the segregation policy. Further, in Havens Realty Corporation v. Coleman,20 the Court reached a landmark unanimous decision upholding the standing of testers and the fair housing organisation to bring suit under the Fair Housing Act.21

Testing to prove disparate treatment

U.S. courts have found discriminatory intent behind the "neutral" actions of individuals or local authorities.22 For example, the use of "VIP cards" and "VIP entrances" by some hotels, restaurants and clubs to mask discriminatory practices,23 and asking for proof of age from people of colour or imposing a dress code24 to delay or prevent admission of people of colour were struck down based on the discriminatory intent of the acts.25

In the U.S., testing has been shown to be one of the most effective tools for proving discrimination in such cases. For instance, in 1989 the Washington Lawyer's Committee working with the local community and Howard University experts adapted testing to prove discrimination in the provision of taxicab services in the District of Colombia. This work involved substantial research using paired teams to test for discrimination in (1) the refusal of service because of a passenger's race and (2) the refusal to transport passengers to predominantly black neighbourhoods.26 Over the course of two and a half months, the team trained and carefully matched teams of black and white testers, who ultimately conducted 292 tests. The test results showed that taxis failed to stop for black testers in 20% of the tests, while cabs passed white testers during only 3% of the tests.27 Drivers also commonly refused to bring black and white testers to predominantly black neighbourhoods in the District of Colombia.28

As a result of the testing, litigation was undertaken against three cab companies whose practices reflected particularly high levels of discrimination. On the eve of the trial, the suits were settled. The settlement included a payment of nearly $50,000 in damages and significant injunctive relief. In this case, the court ruled for the first time in U.S. history that the cab companies could be held liable for discriminatory conduct of their drivers.29 Following this litigation, the District of Columbia Department of Human Rights inaugurated an ongoing program to test the delivery of taxi services and, where necessary, to initiate charges against companies employing discriminatory activities.30

Testing was also applied to obtain evidence to prove the allegation of black secret servicemen against some chains of Denny's restaurant that they had suffered discriminatory treatment. The case was settled out of court, and the company paid $54 million to the plaintiffs for damages and $1 million to civil rights organisations to defray expenses.31

Awarding damages

The amount of damages awarded by courts has grown dramatically since the first cases in which testing was used to successfully prove discrimination. In the mid-1970s, settlements in discrimination cases involving the payment of several dollars, as token payments, were generally considered substantial victories.32 However, by 1990, the Fair Housing Council of Greater Washington (FHCGW)33 had secured more than a dozen settlements or verdicts of over $20,000 each, and recoveries in several cases have exceeded $100,000. In the majority of these cases, tester evidence was a decisive factor in the court victory.

The Supreme Court concluded in Davis v. Mansards that the fact that the plaintiff is a tester "does not affect the measure of her actual damages . . . [i]n 1984, no one should have to toughen themselves to racial discrimination - a tester has no reason to expect mistreatment at the hands of ostensibly fair minded businesspeople."34 Compensatory damages granting relief for lost employment opportunities, embarrassment and humiliation resulting from discrimination have also been awarded to testers as well as to the organisations overseeing their activities. For example, in Bryant v. Kay Brothers Builders,35 the court awarded compensatory damages to housing testers.

Today, in nearly all of the states in the U.S., there are organisations that regularly use testing to prove discrimination in employment, housing and public accommodation cases. They publish the results of the testing, use them to monitor compliance with injunctive relief and to challenge discrimination in administrative or court procedures. Tester evidence has provided a higher level of proof than previously attainable.36

Hungarian law and practive concerning testing

In today's Hungary, even members of the majority population often have difficulty fitting the criteria set by employers and finding gainful employment. Race discrimination - explicit or covert - often raises the opportunities for minorities to find work to the level of near impossibility.

Hungarian law allows the use of testing to obtain evidence, gives testers standing as plaintiffs, permits testers to be called as witnesses, and allows the tester's report to be attached as documentary evidence. First of all, Article 164(1) of the Hungarian Code of Civil Procedure orders that, "the facts required for the adjudication of a case must be proved usually by the party who is interested in their acceptance by the court." The same law also demands that "unless otherwise provided by the act, a court is not bound to formal rules of evidence in a civil action, and is free to use presentations of the case by the parties as well as all other evidences fit for detecting the facts of a case."37 Under the principle of "the right to evaluate evidence freely," in Hungary, the court is not bound to any specific type of evidence. Individual testimonies pertaining to how facts in a case were obtained are considered, under Hungarian law, valid as evidence.38 Hungarian law also provides that either in theory or in practice, "documentary evidence is considered to be evidence since it carries the thoughts of a human being."39 The law also provides that, "the plaintiff could file an action by his name or - under authorization - on behalf of others."40

Using testing in employment discrimination cases in Hungary

The Legal Defence Bureau for National and Ethnic Minorities (NEKI), the organisation for which I work, in cooperation with the ERRC, applied testing in 1999 in the Lajos B. case. Mr. Lajos B., a Romani university student, read an advertisement in February 1999 that there was an opening at T. Ltd. to distribute leaflets. He called the company and was told that there was a vacancy, and that he could come for an interview. After he filled out an application form, he was told that they would call him. After a few days he learned from the company that there was no longer a job vacancy. In his complaint he alleged that he had not been hired for the job because of his Romani origin.41

To prove disparate treatment in this case, we decided to test the company's job application process. The testers applied for the job in the same manner as the complainant had done before, i.e. they called first, then went to inquire about the job in person. The first testers applied for the job on April 29, 1999. Mr Béla C., a Romani man, called the firm, followed by Mr Péter B., a non-Romani man, half an hour later. Both then went to the company, filled in the application forms and were told to call back later. Mr József R. and Mr Roland T., Romani and non-Romani testers respectively, applied on May 4, 1999. All four applicants called the company and discovered that the non-Roma had been recruited, whereas the Roma had not. The recruited testers signed contracts and were given leaflets for distribution. To further document the discrimination, we asked the testers not to carry out the job but to return the leaflets, saying that distribution was too difficult. The same day we sent another Romani tester, knowing there was certainly work to be done. He also filled in the application forms, and he too was later turned down by the company by telephone.42

Because of a lack of evidence, it was difficult to prove the specific complaint of Mr Lajos B. and did not file on his behalf. Instead, we filed an action on behalf of Mr József R., who was a tester, as well as a bona fide job seeker. In the submission it was argued that the employer had violated Article 5 of the Hungarian Labour Law43 and many international laws to which Hungary is party. The submission asked the court to declare the occurrence of the infringement, order the company to employ Mr József R., have the infringement discontinued and the perpetrator restrained from further infringement, and order the company to pay a fine to be used for public purposes,44 since some other Roma had also allegedly been the victims of the company's discriminatory behaviour. In addition to the written submission and evidence, the Labour Court also heard the witness testimony of the plaintiff. In its decision, the court found that there had not been an employment relationship, but rather an independent contractual relationship, an issue which is not covered by Hungarian Labour Law. The court, therefore, ordered the remittal of the case to a civil court. The plaintiff filed an appeal with a second instance labour court of Hungary. The case was pending as of October 11, 2000.

Using testing in public accommodation cases

The reports of civil rights organisations and the Office of the Minority Ombudsman show that immigrants and Roma also suffer discrimination in public accommodation in Hungary. However, the lack of anti-discrimination provisions specifically prohibiting discrimination in public accommodation, a lack of case law, and passivity on the part of public administrators such as consumer protection officers or town clerks make it difficult to obtain remedies in such cases.

Despite these difficulties, NEKI has conducted more than twenty tests to prove discrimination in public accommodation. In Patvárc and in Szikcó (a town where Roma were reportedly not allowed to enter a disco), NEKI sent pairs of Romani and non-Romani testers to discos, pubs and restaurants against which allegations had been made.

In NEKI's tests for discrimination against Roma in public accommodation, forms requested that testers provide the following information: the time of the testing, details on the behaviour of the security personnel and the owners of the bars such as the information or documents they had requested from the tester; reasons for denying services; and any use of racial epithets or other derogatory remarks about the tester. Concerning services, NEKI requested that testers provide a written record of whether the bartender took the tester's order, and if so, how long the tester had to wait to be served. Concerning the bar's complaint book, NEKI requested that testers record whether the book was accessible, and whether the owner or the bartender provided the book for use when the tester wanted to record his complaint. If the bar did not provide the complaint book, the NEKI forms asked testers to provide the reason for the denial.45

After repeated testing and evaluation of the assignment forms, NEKI initiated administrative procedures in the Consumer Production Office, and at the same time filed an action in courts against the owner of the restaurants. Though in both cases, the discriminative action of the bar owners targeted the Romani community, there is no provision in Hungarian law to allow the possibility of a class action lawsuit. For this reason, NEKI always asks the court to order the perpetrator to pay a fine to be used for public purposes, as well as to rule that the infringement be discontinued and the perpetrator restrained from further infringement. In the future, NEKI would like to see changes to the Hungarian legal regime, such that damages or fines will be used directly in the community affected by the discriminatory act or acts.


Although the Hungarian legal system is not a precedent-based one, the decision in Patvárc is path breaking and encourages civil rights organisations around Europe to apply testing to prove discrimination. Before organisations adopt this method, however, it is necessary to study its history and effectiveness in the U.S., as well as in other countries where testing has been successfully used. Research into domestic civil law, and civil procedure law as well as laws on evidence is also necessary to determine effective strategies in cases using testing. In order to ensure that testing remains an effective method for proving discrimination, the power of this tool should not be abused: it should not be used to harass entrepreneurs.

Beyond a particular case, the results derived from testing can inform a variety of public deliberations. Enforcement agencies can use testing results to measure the effectiveness of their efforts and to direct resources to those segments of the community where disparities in treatment are most pronounced. Similarly, legislators can rely upon testing results to assess the effectiveness of prohibitions on discrimination and to assess whether the allocation of resources is sufficient and well targeted to redress the discrimination at issue.46 Governmental agencies can use testing as a component of their systemic investigation, or fund organisations which apply testing. Finally, companies can conduct self-testing in cooperation with civil rights organisations to avoid embarrassing and costly litigation.


  1. Fitsum Alemu has been a staff attorney at the Budapest-based non-governmental organisation Legal Defence Bureau for National and Ethnic Minorities (NEKI). He has been involved in a joint project of ERRC and NEKI aimed at developing strategic anti-discrimination litigation.
  2. See Boggs, Roderic V.O., Joseph M. Sellers, and Marc Benedick (Ph.D.), “The Use of Testing in Civil Rights Enforcement” (hereinafter: “The Use of Testing”), September 1991, p.26 (unpublished document). Facial discrimination: a kind of discrimination that occurs when a legal regulation, official measure, or an individual’s intentional act directly aims at discrimination in any form against protected classes. Disparate treatment: a form of discrimination which occurs when individuals or state agencies treat protected classes less favourably because of their race, national origin, sex, religion or their social status, but the perpetrator denies the act. To prove this kind of discrimination the plaintiff generally must demonstrate that, for example, in cases of employment discrimination: (i) he is a member of a protected class; (ii) he applied for the job and was qualified for a job for which the employer was seeking applicants; (iii) despite his qualification, he was rejected; (iv) and after his application was rejected, the position remained open and the employer continued to seek applications from persons whose qualifications were similar to those of the complainants, or later the position was filled by a member of a non-protected class with the same qualifications as the complainant. See Alemu, Fitsum, “Definition of Discrimination” in White Booklet 1999, p.76. For the purposes of this paper, “protected class group” will be used to describe persons whom laws shelter from discrimination, such as Roma, blacks, Latinos, persons with certain disabilities, members of families with children and so on.
  3. See the decision of the Balassagyarmati City Court, 8.P.20.761/1999/18 számú ítélete, p.2. On February 17, 2000, the Court held that the refusal of a pub owner in the town of Patvárc to serve Roma had been discriminatory. Also see “Hungarian Court Finds Pub Owner Guily of Discrimination”, “Snapshots from around Europe” Roma Rights 1/2000.
  4. See “The Use of Testing,” p.20.
  5. First the protected class and then the non-protected class.
  6. See NEKI, Assignment Sheet to Test Employment Discrimination, 1999; “The Use of Testing,” pp.17-18; Essed, P. Understanding Everyday Racism, Newberry Park, CA: Sage Publications, 1991.
  7. Ibid. p.17.
  8. See Alemu, Fitsum. “Information for Testers,” 1999 (unpublished document); and “Contract Between NEKI and Testers,” NEKI, 1999 (unpublished document).
  9. See “The Use of Testing,” pp.10-14, and “Contract Between NEKI and Testers,” Op. cit.
  10. See “The Use of Testing,” pp.38-44.
  11. 411 U.S. 793 (1973).
  12. Section 1982 — Property Rights of Citizens: “All citizens of the United States shall have the same right, in every State and Territory, as is enjoyed by white citizens thereof to inherit, purchase, lease, sell, hold, and convey real and personal property.”
  13. See Coel v. Rose Tree Manor Apartments, Inc., No. 84-1521, U.S. Dist. LEXIS 801 (E.D.Pa. 1988).
  14. 358 U.S. 202 (1958).
  15. Evers, p.203-204.
  16. Evers, p.204.
  17. Evers, p.203.
  18. Evers, p.204.
  19. 386 U.S. 547 (1967).
  20. 455 U.S. 363 (1982).
  21. This case was initiated by Housing Opportunities Made Equal, a Richmond, Virginia-based human rights organisation that advocates fair housing.
  22. See Katzenbach v. McClung, 379 U.S. 294 (1964), Tillman v. Wheaton-Haven Recreation Association, Inc., 451 F.2d 1211 (1971), United States v. Glass Menagerie, Inc., 702 F.Supp 139 (E.D.Ky. 1988), New York v. Peter & John’s Pump House, 914 F.Supp 809 (N.D.N.Y. 1996) and United States v. Cantrell, 307 F.Supp 259 (E.D.La. 1969).
  23. United States v. Glass Menagerie, Inc., 317 F.Supp 139 (E.D.Ky. 1988).
  24. New York v. Peter & John’s Pump House, 914 F.Supp 809 (N.D.N.Y. 1996).
  25. See also U.S. v. Cantrell in which a local ordinance prohibited bars and cocktail lounges from admitting “any military personnel in Uniform, as well as from selling any such person any food or beverages” to discriminate against black persons who frequented the bar. The act was found to be discriminatory.
  26. See “The Use of Testing,” p.6.
  27. See Ridley, Bayton and Outtz, “Taxi Service in the District of Columbia: Is It Influenced by the Patrons Race and Destination?”, p.17 (June 1989).
  28. Ibid, pp.21, 27.
  29. See Floyd-Mayers v. American Cab Co., 732 F.Supp 243 (D.D.C. 1990).
  30. See Wheeler, “Undercover Riders Gauge Prejudice of D.C. cabbies,” in The Washington Post, Metro Section, July 17, 1990.
  31. Sniffen, Michael J., Associated Press, “Denny’s Agrees to Retrain Managers,” April 10, 2000; National Fair Housing Advocate, February 1997; Fair Housing Council, Spring 1997 Report.
  32. See Fitsum, Alemu, “Fair Housing in the USA and Hungary,” in Fundamentum, 1998/1-2, pp.141-153.
  33. A civil rights organisation that advocates fair housing issues. The author of this article worked with this organisation for three months as an intern.
  34. 597 F.Supp 334, p.347 (N.D. Ind. 1984).
  35. C.A. No. B-86-354 (D.Md. February 18, 1988).
  36. See “The Use of Testing,” p.4.
  37. Article 6 of the Hungarian Code of Civil Procedure.
  38. Articles 166 and 173 of the Hungarian Code of Civil Procedure.
  39. Kengyel Miklós, Magyar Polgári Eljárás jog, Osiris, 1998, p.269.
  40. Article 2 of the Hungarian Code of Civil Procedure; Article 10 of Council of Ministers Decree No. 106/1952; No. 199 Civil Council Decision of the Supreme Court.
  41. NEKI, White Booklet 1999, pp.34-36.
  42. Ibid.
  43. Article 5(1) of the Hungarian Labour Code bans discrimination in employment.
  44. According to Article 78 of the Hungarian Civil Code and Article 100 of the Hungarian Labour Law.
  45. NEKI, “Testers Assignment Form in Public Accommodation Discrimination,” 1999 (unpublished document).
  46. See “The Use of Testing,” p.21.


Challenge discrimination, promote equality


Receive our public announcements Receive our Roma Rights Journal


The latest Roma Rights news and content online

join us

Find out how you can join or support our activities