UNIFORM EMPLOYEE SELECTION GUIDELINES
INTERPRETATION AND CLARIFICATION (Questions and Answers)
Equal Employment Opportunity Commission
Table of Contents
I.Purpose and Scope. 3
II.Adverse Impact, the Bottom Line and Affirmative Action. 4
III. General Questions Concerning Validity and the Use of Selection Procedures 10
IV.Technical Standards 14
A. CRITERION-RELATED VALIDITY. 16
B. CONTENT VALIDITY. 18
C. CONSTRUCT VALIDITY. 19
V. Records and Documentation. 19
The following series of questions and answers are designed to clarify and interpret, but not to modify, the uniform Guidelines on Employee Selection Procedures that were adopted on August 25, 1978, by the Equal Employment Opportunity Commission (29 CFR 1607), Office of Personnel Management (5 CFR 300), U.S. Department of Justice (28 CFR, Ch. 1, Part 50), U.S. Treasury Department (31 CFR, Ch. 1, Part 51) and the Office of Federal Contract Compliance Programs (41 CFR, Ch. 60, Part 68-3). Supplemental questions and answers (No. 91 through 93) became effective as of May 2, 1980 (45 F.R. 29350).
I. Purpose and Scope
1. Q. What is the purpose of the Guidelines?
A. The guidelines are designed to aid in the achievement of our nation's goal of equal employment opportunity without discrimination on the grounds of race, color, sex, religion or national origin. The Federal agencies have adopted the Guidelines to provide a uniform set of principles governing use of employee selection procedures which is consistent with applicable legal standards and validation standards generally accepted by the psychological profession and which the Government will apply in the discharge of its responsibilities.
2. Q. What is the basic principle of the Guidelines?
A. A selection process which has an adverse impact on the employment opportunities of members of a race, color, religion, sex, or national origin group (referred to as "race, sex, and ethnic group," as defined in Section 16P) and thus disproportionately screens them out is unlawfully discriminatory unless the process or its component procedures have been validated in accord with the Guidelines, or the user otherwise justifies them in accord with Federal law. See Sections 3 and 6. 1 This principle was adopted by the Supreme Court unanimously in Griggs v. Duke Power Co., 401 U.S. 424, and was ratified and endorsed by the Congress when it passed the Equal Employment Opportunity Act of 1972, which amended Title VII of the Civil Rights Act of 1964.
3. Q. Who is covered by the Guidelines?
A. The Guidelines apply to private and public employers, labor organizations, employment agencies, apprenticeship committees, licensing and certification boards (see Question 7), and contractors or subcontractors, who are covered by one or more of the following provisions of Federal equal employment opportunity law: Title VII of the Civil Rights Act of 1964, as amended by the Equal Employment Opportunity Act of 1972 (hereinafter Title VII); Executive Order 11246, as amended by Executive Orders 11375 and 12086 (hereinafter Executive Order 11246); the State and Local Fiscal Assistance Act of 1972, as amended; Omnibus Crime Control and Safe Streets Act of 1968, as amended; and the Intergovernmental Personnel Act of 1970, as amended. Thus, under Title VII, the Guidelines apply to the Federal Government with regard to Federal employment. Through Title VII they apply to most private employers who have 15 or more employees for 20 weeks or more a calendar year, and to most employment agencies, labor organizations and apprenticeship committees. They apply to state and local governments which employ 15 or more employees, or which receive revenue sharing funds, or which receive funds from the Law Enforcement Assistance Administration to impose and strengthen law enforcement and criminal justice, or which receive grants or other federal assistance under a program which requires maintenance of personnel standards on a merit basis. They apply through Executive Order 11246 to contractors and subcontractors of the Federal Government and to contractors and subcontractors under federally-assisted construction contracts.
4. Q. Are college placement offices and similar organizations considered to be users subject to the Guidelines?
A. Placement offices may or may not be subject to the Guidelines depending on what services they offer. If a placement office uses a selection procedure as a basis for any employment decision, it is covered under the definition of "user". Section 16. For example, if a placement office selects some students for referral to an employer but rejects others, it is covered. However, if the placement office refers all interested students to an employer, it is not covered, even though it may offer office space and provision for informing the students of job openings. The Guidelines are intended to cover all users of employee selection procedures, including employment agencies, who are subject to Federal equal employment opportunity law.
5. Q. Do the Guidelines apply only to written tests?
A. No. They apply to all selection procedures used to make employment decisions, including interviews, review of experience or education from application forms, work samples, physical requirements, and evaluations of performance. Sections 2B and 16Q, and see Question 6.
6. Q. What practices are covered by the Guidelines?
A. The Guidelines apply to employee selection procedures which are used in making employment decisions, such as hiring, retention, promotion, transfer, demotion, dismissal or referral. Section 2B. Employee selection procedures include job requirements (physical, education, experience), and evaluation of applicants or candidates on the basis of application forms, interviews, performance tests, paper and pencil tests, performance in training programs or probationary periods, and any other procedures used to make an employment decision whether administered by the employer or by an employment agency. See Section 2B.
7. Q. Do the Guidelines apply to the licensing and certification functions of state and local governments?
A. The Guidelines apply to such functions to the extent that they are covered by Federal law. Section 2B. The courts are divided on the issue of such coverage. The Government has taken the position that at least some kinds of licensing and certification which deny persons access to employment opportunity may be enjoined in an action brought pursuant to Section 707 of the Civil Rights Act of 1964, as amended.
8. Q. What is the relationship between Federal equal employment opportunity law, embodied in these Guidelines, and State and Local government merit system laws or regulations requiring rank ordering of candidates and selection from a limited number of the top candidates?
A. The Guidelines permit ranking where the evidence of validity is sufficient to support that method of use. State or local laws which compel rank ordering generally do so on the assumption that the selection procedure is valid. Thus, if there is adverse impact and the validity evidence does not adequately support that method of use, proper interpretation of such a state law would require validation prior to ranking. Accordingly, there is no necessary or inherent conflict between Federal law and State or local laws of the kind described.
Under the Supremacy Clause of the Constitution (Art. VI, Cl. 2),
however, Federal law or valid regulation overrides any contrary
provision of state or local law. Thus, if there is any conflict,
Federal equal opportunity law prevails. For example, in Rosenfeld
v. So. Pacific Co., 444 F.2d 1219 (9th Cir., 1971), the court held
invalid state protective laws which prohibited the employment of
women in jobs entailing long hours or heavy labor, because the state
laws were in conflict with Title VII. Where a State or local official
believes that there is a possible conflict, the official may wish
to consult with the State Attorney General, County or City attorney,
or other legal official to determine how to comply with the law.
II.Adverse Impact, the Bottom Line and Affirmative Action
9. Q. Do the Guidelines require that only validated selection procedures be used?
A. No. Although validation of selection procedures is desirable in personnel management, the Uniform Guidelines require users to produce evidence of validity only when the selection procedure adversely affects the opportunities of a race, sex, or ethnic group for hire, transfer, promotion, retention or other employment decision. If there is no adverse impact, there is no validation requirement under the Guidelines. Sections 1B and 3A. See also, Section 6A.
10. Q. What is adverse impact?
A. Under the Guidelines adverse impact is a substantially different rate of selection in hiring, promotion or other employment decision which works to the disadvantage of members of a race, sex or ethnic group. Sections 4D and 16B. See Questions 11 and 12.
11. Q. What is a substantially different rate of selection?
A. The agencies have adopted a rule of thumb under which they will generally consider a selection rate for any race, sex, or ethnic group which is less than four-fifths (4/5ths) or eighty percent (80%) of the selection rate for the group with the highest selection rate as a substantially different rate of selection. See Section 4D. This "4/5ths" or "80%" rule of thumb is not intended as a legal definition, but is a practical means of keeping the attention of the enforcement agencies on serious discrepancies in rates of hiring, promotion and other selection decisions.
For example, if the hiring rate for whites other than Hispanics is 60%, for American Indians 45%, for Hispanics 48%, and for Blacks 51%, and each of these groups constitutes more than 2% of the labor force in the relevant labor area (see Question 16), a comparison should be made of the selection rate for each group with that of the highest group (whites). These comparisons show the following impact ratios: American Indians 45/60 or 75%; Hispanics 48/60 or 80%; and Blacks 51/60 or 85%. Applying the 4/5ths or 80% rule of thumb, on the basis of the above information alone, adverse impact is indicated for American Indians but not for Hispanics or Blacks.
12. Q. How is adverse impact determined?
A. Adverse impact is determined by a four-step process.
(1) Calculate the rate of selection for each group (divide the number of persons selected from a group by the number of applicants from that group).
(2) Observe which group has the highest selection rate.
(3) Calculate the impact ratios, by comparing the selection rate for each group with that of the highest group (divide the selection rate for a group by the selection rate for the highest group).
(4) Observe whether the selection rate for any group is substantially less (i.e., usually less then 4/5ths or 80%) than the selection rate for the highest group. If it is, adverse impact is indicated in most circumstances. See Section 4D.For example:
|Selection Rate/Percent Hired
48/80 or 60%
|12/40 or 30%
A comparison of the black selection rate (30%) with the white selection rate (60%) shows that the black rate is 30/60, or one-half (or 50%) of the white rate. Since the one-half (50%) is less than 4/5ths (80%) adverse impact is usually indicated.
The determination of adverse impact is not purely arithmetic however; and other factors may be relevant. See, Section 4D.
13. Q. Is adverse impact determined on the basis of the overall selection process or for the components in that process?
A. Adverse impact is determined first for the overall selection process for each job. If the overall selection process has an adverse impact, the adverse impact of the individual selection procedure should be analyzed. For any selection procedures in the process having an adverse impact which the user continues to use in the same manner, the user is expected to have evidence of validity satisfying the Guidelines. Sections 4C and 5D. If there is no adverse impact for the overall selection process, in most circumstances there is no obligation under the Guidelines to investigate adverse impact for the components, or to validate the selection procedures used for that job. Section 4C. But see Question 25.
14. Q. The Guidelines designate the "total selection process" as the initial basis for determining the impact of selection procedures. What is meant by the "total selection process"?
A. The "total selection process" refers to the combined effect of all selection procedures leading to the final employment decision such as hiring or promoting. For example, appraisal of candidates for administrative assistant positions in an organization might include initial screening based upon an application blank and interview, a written test, a medical examination, a background check, and a supervisor's interview. These in combination are the total selection process. Additionally, where there is more than one route to the particular kind of employment decision, the total selection process encompasses the combined results of all routes. For example, an employer may select some applicants for a particular kind of job through appropriate written and performance tests. Others may be selected through an internal upward mobility program, on the basis of successful performance in a directly related trainee type of position. In such a case, the impact of the total selection process would be the combined effect of both avenues of entry.
15. Q. What is meant by the terms "applicant" and "candidate" as they are used in the Uniform Guidelines?
A. The precise definition of the term "applicant" depends upon the user's recruitment and selection procedures. The concept of an applicant is that of a person who has indicated an interest in being considered for hiring, promotion, or other employment opportunities. This interest might be expressed by completing an application form, or might be expressed orally, depending upon the employer's practice.
The term "candidate" has been included to cover those situations where the initial step by the user involves consideration of current employees for promotion, or training, or other employment opportunities, without inviting applications. The procedure by which persons are identified as candidates is itself a selection procedure under the Guidelines.
A person who voluntarily withdraws formally or informally at any stage of the selection process is no longer an applicant or candidate for purposes of computing adverse impact. Employment standards imposed by the user which discourage disproportionately applicants of a race, sex or ethnic group may, however, require justification. Records should be kept for persons who were applicants or candidates at any stage of the process.
16. Q. Should adverse impact determinations be made for all groups regardless of their size?
A. No. Section 15A(2) calls for annual adverse impact determinations to be made for each group which constitutes either 2% or more of the total labor force in the relevant labor area, or 2% of more of the applicable workforce. Thus, impact determinations should be made for any employment decision for each group which constitutes 2% or more of the labor force in the relevant labor area. For hiring, such determination should also be made for groups which constitute more than 2% of the applicants; and for promotions, determinations should also be made for those groups which constitute at least 2% of the user's workforce. There are record keeping obligations for all groups, even those which are less than 2%. See Question 86.
17. Q. In determining adverse impact, do you compare the selection rates for males and females, and blacks and whites, or do you compare selection rates for white males, white females, black males and black females?
A. The selection rates for males and females are compared, and the selection rates for the race and ethnic groups are compared with the selection rate of the race or ethnic group with the highest selection rate. Neutral and objective selection procedures free of adverse impact against any race, sex or ethnic group are unlikely to have an impact against a subgroup. Thus there is no obligation to make comparisons for subgroups (e.g., white male, white female, black male, black female). However, there are obligations to keep records (see Question 87), and any apparent exclusion of a subgroup may suggest the presence of discrimination.
18. Q. Is it usually necessary to calculate the statistical significance of differences in selection rates when investigating the existence of adverse impact?
A. No. Adverse impact is normally indicated when one selection rate is less than 80% of the other. The federal enforcement agencies normally will use only the 80% (4/5ths) rule of thumb, except where large numbers of selections are made. See Questions 20 and 22.
19. Q. Does the 4/5ths rule of thumb mean that the Guidelines will tolerate up to 20% discrimination?
A. No. The 4/5ths rule of thumb speaks only to the question of adverse impact, and is not intended to resolve the ultimate question of unlawful discrimination. Regardless of the amount of difference in selection rates, unlawful discrimination may be present, and may be demonstrated through appropriate evidence. The 4/5ths rule merely establishes a numerical basis for drawing an initial inference and for requiring additional information.
With respect to adverse impact, the Guidelines expressly state (section 4D) that differences in selection rates of less than 20% may still amount to adverse impact where the differences are significant in both statistical and practical terms. See Question 20. In the absence of differences which are large enough to meet the 4/5ths rule of thumb or a test of statistical significance, there is no reason to assume that the differences are reliable, or that they are based upon anything other than chance.
20. Q. Why is the 4/5ths rule called a rule of thumb?
A. Because it is not intended to be controlling in all circumstances. If, for the sake of illustration, we assume that nationwide statistics show that use of an arrest record would disqualify 10% of all Hispanic persons but only 4% of all whites other than Hispanic (hereafter non-Hispanic), the selection rate for that selection procedure is 90% for Hispanics and 96% for non-Hispanics. Therefore, the 4/5 rule of thumb would not indicate the presence of adverse impact (90% is approximately 94% of 96%). But in this example, the information is based upon nationwide statistics, and the sample is large enough to yield statistically significant results, and the difference (Hispanics are 21/2 times as likely to be disqualified as non-Hispanics) is large enough to be practically significant. Thus, in this example the enforcement agencies would consider a disqualification based on an arrest record alone as having an adverse impact. Likewise, in Gregory v. Litton Industries, 472 F.2d 631 (9th Cir., 1972), the court held that the employer violated Title VII by disqualifying persons from employment solely on the basis of an arrest record, where that disqualification had an adverse impact on blacks and was not shown to be justified by business necessity.
On the other hand, a difference of more than 20% in rates of selection may not provide a basis for finding adverse impact if the number of persons selected is very small. For example, if the employer selected three males and one female from an applicant pool of 20 males and 10 females, the 4/5ths rule would indicate adverse impact (selection rate for women is 10%; for men 15%; 10/15 or 662/3% is less than 80%), yet the number of selections is too small to warrant a determination of adverse impact. In these circumstances, the enforcement agency would not require validity evidence in the absence of additional information (such as selection rates for a longer period of time) indicating adverse impact. For record keeping requirements, see Section 15A(2)(c) and Questions 84 and 85.
21. Q. Is evidence of adverse impact sufficient to warrant
a validity study or an enforcement action where the numbers involved
are so small that it is more likely than not that the difference
could have occurred by chance?
|Selection Rate/Percent Hired
Black Selection Rate ....................................... 15
15 divided by 20 = 75% (which is less than 80%).
A. No. If the numbers of persons and the difference in selection rates are so small that it is likely that the difference could have occurred by chance, the Federal agencies will not assume the existence of adverse impact, in the absence of other evidence. In this example, the difference in selection rates is too small, given the small number of black applicants, to constitute adverse impact in the absence of other information (see Section 4D). If only one more black had been hired instead of a white the selection rate for blacks (20%) would be higher than that for whites (18.7%). Generally, it is inappropriate to require validity evidence or to take enforcement action where the number of persons and the difference in selection rates are so small that the selection of one different person for one job would shift the result from adverse impact against one group to a situation in which that group has a higher selection rate than the other group.
On the other hand, if a lower selection rate continued over a period of time, so as to constitute a pattern, then the lower selection rate would constitute adverse impact, warranting the need for validity evidence.
22. Q. Is it ever necessary to calculate the statistical significance of differences in selection rates to determine whether adverse impact exists?
A. Yes. Where large numbers of selections are made, relatively small differences in selection rates may nevertheless constitute adverse impact if they are both statistically and practically significant. See Section 4D and Question 20. For that reason, if there is a small difference in selection rates (one rate is more than 80% of the other), but large numbers of selections are involved, it would be appropriate to calculate the statistical significance of the difference in selection rates.
23. Q. When the 4/5th rule of thumb shows adverse impact, is there adverse impact under the Guidelines?
A. There usually is adverse impact, except where the number of persons selected and the difference in selection rates are very small. See Section 4D and Questions 20 and 21.
24. Q. Why do the Guidelines rely primarily upon the 4/5ths rule of thumb, rather than tests of statistical significance?
A. Where the sample of persons selected is not large, even a large real difference between groups is likely not to be confirmed by a test of statistical significance (at the usual .05 level of significance). For this reason, the Guidelines do not rely primarily upon a test of statistical significance, but use the 4/5ths rule of thumb as a practical and easy-to-administer measure of whether differences in selection rates are substantial. Many decisions in day-to-day life are made without reliance upon a test of statistical significance.
25. Q. Are there any circumstances in which the employer should evaluate components of a selection process, even though the overall selection process results in no adverse impact?
A. Yes, there are such circumstances: (1) Where the selection procedure is a significant factor in the continuation of patterns of assignments of incumbent employees caused by prior discriminatory employment practices. Assume, for example, an employer who traditionally hired blacks as employees for the "laborer" department in a manufacturing plant, and traditionally hired only whites as skilled craftsmen. Assume further that the employer in 1962 began to use a written examination not supported by a validity study to screen incumbent employees who sought to enter the apprenticeship program for skilled craft jobs. The employer stopped making racial assignments in 1972. Assume further that for the last four years, there have been special recruitment efforts aimed at recent black high school graduates and that the selection process, which includes the written examination, has resulted in the selection of black applicants for apprenticeship in approximately the same rates as white applicants.
In those circumstances, if the written examination had an adverse impact, its use would tend to keep incumbent black employees in the laborer department, and deny them entry to apprenticeship programs. For that reason, the enforcement agencies would expect the user to evaluate the impact of the written examination, and to have validity evidence for the use of the written examination if it has an adverse impact.
(2) Where the weight of court decisions or administrative interpretations holds that a specific selection procedure is not job related in similar circumstances.
For example, courts have held that because an arrest is not a determination of guilt, an applicant's arrest record by itself does not indicate inability to perform a job consistent with the trustworthy and efficient operation of a business. Yet a no arrest record requirement has a nationwide adverse impact on some minority groups. Thus, an employer who refuses to hire applicants solely on the basis of an arrest record is on notice that this policy may be found to be discriminatory. Gregory v. Litton Industries, 472 F.2d 631 (9th Cir., 1972) (excluding persons from employment solely on the basis of arrests, which has an adverse impact, held to violate Title VII). Similarly, a minimum height requirement disproportionately disqualifies women and some national origin groups, and has been held not to be job related in a number of cases. For example, in Dothard v. Rawlinson, 433 U.S. 321 (1977), the Court held that height and weight requirements not shown to be job related were violative of Title VII. Thus an employer using a minimum height requirement should have evidence of its validity.
(3) In addition, there may be other circumstances in which an enforcement agency may decide to request an employer to evaluate components of a selection process, but such circumstances would clearly be unusual. Any such decision will be made only at a high level in the agency. Investigators and compliance officers are not authorized to make this decision.
26. Q. Does the bottom line concept of Section 4C apply to the administrative processing of charges of discrimination filed with an issuing agency, alleging that a specific selection procedure is discriminatory?
A. No. The bottom line concept applies only to enforcement actions as defined in Section 16 of the Guidelines. Enforcement actions include only court enforcement actions and other similar proceedings as defined in Section 16I. The EEOC administrative processing of charges of discrimination (investigation, finding of reasonable cause/no cause, and conciliation) required by Section 706(b) of Title VII are specificallly exempted from the bottom line concept by the definition of an enforcement action. The bottom line concept is a result of a decision by the various enforcement agencies that, as a matter of prosecutorial discretion, they will devote their limited enforcement resources to the most serious offenders of equal employment opportunity laws. Since the concept is not a rule of law, it does not affect the discharge by the EEOC of its statutory responsibilities to investigate charges of discrimination, render an administrative finding on its investigation, and engage in voluntary conciliation efforts. Similarly, with respect to the other issuing agencies, the bottom line concept applies not to the processing of individual charges, but to the initiation of enforcement action.
27. Q. An employer uses one test or other selection procedure to select persons for a number of different jobs. Applicants are given the test, and the successful applicants are then referred to different departments and positions on the basis of openings available and their interests. The Guidelines appear to require assessment of adverse impact on a job-by-job basis (Section 15A[a]). Is there some way to show that the test as a whole does not have adverse impact even though the proportions of members of each race, sex or ethnic group assigned to different jobs may vary?
A. Yes, in some circumstances. The Guidelines require evidence of validity only for those selection procedures which have an adverse impact, and which are part of a selection process which has an adverse impact. If the test is administered and used in the same fashion for a variety of jobs, the impact of that test can be assessed in the aggregate. The records showing the results of the test, and the total number of persons selected, generally would be sufficient to show the impact of the test. If the test has no adverse impact, it need not be validated.
But the absence of adverse impact of the test in the aggregate does not end the inquiry. For there may be discrimination or adverse impact in the assignment of individuals to, or in the selection of persons for, particular jobs. The Guidelines call for records to be kept and determinations of adverse impact to be made of the overall selection process on a job by job basis. Thus, if there is adverse impact in the assignment or selection procedures for a job even though there is no adverse impact from the test, the user should eliminate the adverse impact from the assignment procedure or justify the assignment procedure.
28. Q. The Uniform Guidelines apply to the requirements of Federal law prohibiting employment practices which discriminate on the grounds of race, color, religion, sex or national origin. However, records are required to be kept only by sex and by specified race and ethnic groups. How can adverse impact be determined for religious groups and for national origin groups other than those specified in Section 4B of the Guidelines?
A. The groups for which records are required to be maintained are the groups for which there is extensive evidence of continuing discriminatory practices. This limitation is designed in part to minimize the burden on employers for record keeping which may not be needed.
For groups for which records are not required, the person(s) complaining may obtain information from the employer or others (voluntarily or through legal process) to show that adverse impact has taken place. When that has been done, the various provisions of the Uniform Guidelines are fully applicable.
Whether or not there is adverse impact, Federal equal employment opportunity law prohibits any deliberate discrimination or disparate treatment on grounds of religion or national origin, as well as on grounds of sex, color, or race.
Whenever "ethnic" is used in the Guidelines or in these Questions and Answers, it is intended to include national origin and religion, as set forth in the statutes, executive orders, and regulations prohibiting discrimination. See Section 16P.
29. Q. What is the relationship between affirmative action and the requirements of the Uniform Guidelines?
A. The two subjects are different, although related. Compliance with the Guidelines does not relieve users of their affirmative action obligations, including those of Federal contractors and subcontractors under Executive Order 11246. Section 13.
The Guidelines encourage the development and effective implementation of affirmative action plans or programs in two ways. First, in determining whether to institute action against a user on the basis of a selection procedure which has adverse impact and which has not been validated, the enforcement agency will take into account the general equal employment opportunity posture of the user with respect to the job classifications for which the procedure is used and the progress which has been made in carrying out any affirmative action program, Section 4E. If the user has demonstrated over a substantial period of time that it is in fact appropriately utilizing in the job or group of jobs in question the available race, sex or ethnic groups in the relevant labor force, the enforcement agency will generally exercise its discretion by not initiating enforcement proceedings based on adverse impact in relation to the applicant flow. Second, nothing in the Guidelines is intended to preclude the use of selection procedures, consistent with Federal law, which assist in the achievement of affirmative action objectives. Section 13A. See also, Questions 30 and 31.
30. Q. When may a user be race, sex or ethnic-conscious?
A. The Guidelines recognize that affirmative action programs may be race, sex or ethnic conscious in appropriate circumstances, (See Sections 4E and 13; See also Section 7, Appendix). In addition to obligatory affirmative action programs (See Question 29), the Guidelines encourage the adoption of voluntary affirmative action programs. Users choosing to engage in voluntary affirmative action are referred to EEOC's Guidelines on Affirmative Action (44 FR 4422, January 19, 1979). A user may justifiably be race, sex or ethnic-conscious in circumstances where it has reason to believe that qualified persons of specified race, sex or ethnicity have been or may be subject to the exclusionary effects of its selection procedures or other employment practices in its work force or particular jobs therein. In establishing long and short range goals, the employer may use the race, sex, or ethnic classification as the basis for such goals (Section 17[a]).
In establishing a recruiting program, the employer may direct its recruiting activities to locations or institutions which have a high proportion of the race, sex, or ethnic group which has been excluded or underutilized (section 17[b]). In establishing the pool of qualified persons from which final selections are to be made, the employer may take reasonable steps to assure that members of the excluded or underutilized race, sex, or ethnic group are included in the pool (Section 17[e]).
Similarly, the employer may be race, sex or ethnic-conscious in determining what changes should be implemented if the objectives of the programs are not being met (Section 17[g]).
Even apart from affirmative action programs a user may be race, sex or ethnic-conscious in taking appropriate and lawful measures to eliminate adverse impact from selection procedures (Section 6A).
31. Q. Section 6A authorizes the use of alternative selection procedures to eliminate adverse impact, but does not appear to address the issue of validity. Thus, the use of alternative selection procedures without adverse impact seems to be presented as an option in lieu of validation. Is that its intent?
A. Yes. Under Federal equal employment opportunity law the use of any selection procedure which has an adverse impact on any race, sex or ethnic group is discriminatory unless the procedure has been properly validated, or the use of the procedure is otherwise justified under Federal law. Griggs v. Duke Power Co., 401 U.S. 424 (1971); Section 3A. If a selection procedure has an adverse impact, therefore, Federal equal employment opportunity law authorizes the user to choose lawful alternative procedures which eliminate the adverse impact rather than demonstrating the validity of the original selection procedure.
Many users, while wishing to validate all of their selection procedures, are not able to conduct the validity studies immediately. Such users have the option of choosing alternative techniques which eliminate adverse impact, with a view to providing a basis for determining subsequently which selection procedures are valid and have as little adverse impact as possible.
Apart from Federal equal employment opportunity law, employers
have economic incentives to use properly validated selection procedures.
Nothing in Section 6A should be interpreted as discouraging the
use of properly validated selection procedures; but Federal equal
employment opportunity law does not require validity studies to
be conducted unless there is adverse impact. See Section 2C.
III. General Questions Concerning Validity and the Use of Selection Procedures
32. Q. What is "validation" according to the Uniform Guidelines?
A. Validation is the demonstration of the job relatedness of a selection procedure. The Uniform Guidelines recognize the same three validity strategies recognized by the American Psychological Association:
(1) Criterion-related validity--a statistical demonstration of a relationship between scores on a selection procedure and job performance of a sample of workers.
(2) Content validity--a demonstration that the content of a selection procedure is representative of important aspects of performance on the job.
(3) Construct validity--a demonstration that (a) a selection procedure measures a construct (something believed to be an underlying human trait or characteristic, such as honesty) and (b) the construct is important for successful job performance.
33. Q. What is the typical process by which validity studies are reviewed by an enforcement agency?
A. The validity study is normally requested by an enforcement officer during the course of a review. The officer will first determine whether the user's data show that the overall selection process has an adverse impact, and if so, which component selection procedures have an adverse impact. See Section 15A(3). The officer will then ask for the evidence of validity for each procedure which has an adverse impact. See Sections 15B, C, and D. This validity evidence will be referred to appropriate personnel for review. Agency findings will then be communicated to the user.
34. Q. Can a user send its validity evidence to an enforcement agency before a review, so as to assure its validity?
A. No. Enforcement agencies will not review validity reports except in the context of investigations or reviews. Even in those circumstances, validity evidence will not be reviewed without evidence of how the selection procedure is used and what impact its use has on various race, sex, and ethnic groups.
35. Q. May reports of validity prepared by publishers of commercial tests and printed in test manuals or other literature be helpful in meeting the Guidelines?
A. They may be. However, it is the user's responsibility to determine that the validity evidence is adequate to meet the Guidelines. See Section 7, and Questions 43 and 66. Users should not use selection procedures which are likely to have an adverse impact without reviewing the evidence of validity to make sure that the standards of the Guidelines are met.
The following questions and answers (36-81) assume that a selection procedure has an adverse impact and is part of a selection process that has an adverse impact.
36. Q. How can users justify continued use of a procedure on a basis other than validity?
A. Normally, the method of justifying selection procedures with an adverse impact and the method to which the Guidelines are primarily addressed, is validation. The method of justification of a procedure by means other than validity is one to which the Guidelines are not addressed. See Section 6B. In Griggs v. Duke Power Co., 401 U.S. 424, the Supreme Court indicated that the burden on the user was a heavy one, but that the selection procedure could be used if there was a "business necessity" for its continued use; therefore, the Federal agencies will consider evidence that a selection procedure is necessary for the safe and efficient operation of a business to justify continued use of a selection procedure.
37. Q. Is the demonstration of a rational relationship (as that term is used in constitutional law) between a selection procedure and the job sufficient to meet the validation requirements of the Guidelines?
A. No. The Supreme Court in Washington v. Davis, 426 U.S. 229 (1976) stated that different standards would be applied to employment discrimination allegations arising under the Constitution than would be applied to employment discrimination allegations arising under Title VII. The Davis case arose under the Constitution, and no Title VII violation was alleged. The Court applied a traditional constitutional law standard of "rational relationship" and said that it would defer to the "seemingly reasonable acts of administrators and executives." However, it went on to point out that under Title VII, the appropriate standard would still be an affirmative demonstration of the relationship between the selection procedure and measures of job performance by means of accepted procedures of validation and it would be an "insufficient response to demonstrate some rational basis" for a selection procedure having an adverse impact. Thus, the mere demonstration of a rational relationship between a selection procedure and the job does not meet the requirement of Title VII of the Civil Rights Act of 1964, or of Executive Order 11246, or the State and Local Fiscal Assistance Act of 1972, as amended (the revenue sharing act) or the Omnibus Crime Control and Safe Streets Act of 1968, as amended, and will not meet the requirements of these Guidelines for a validity study. The three validity strategies called for by these Guidelines all require evidence that the selection procedure is related to successful performance on the job. That evidence may be obtained through local validation or through validity studies done elsewhere.
38. Q. Can a user rely upon written or oral assertions of validity instead of evidence of validity?
A. No. If a user's selection procedures have an adverse impact, the user is expected to produce evidence of the validity of the procedures as they are used. Thus, the unsupported assertion by anyone, including representatives of the Federal government or State Employment Services, that a test battery or other selection procedure has been validated is not sufficient to satisfy the Guidelines.
39. Q. Are there any formal requirements imposed by these Guidelines as to who is allowed to perform a validity study?
A. No. A validity study is judged on its own merits, and may be performed by any person competent to apply the principles of validity research, including a member of the user's staff or a consultant. However, it is the user's responsibility to see that the study meets validity provisions of the Guidelines, which are based upon professionally accepted standards. See Question 42.
40. Q. What is the relationship between the validation provisions of the Guidelines and other statements of psychological principles, such as the Standards for Educational and Psychological Tests, published by the American Psychological Association (Wash., D.C., 1974) (hereinafter "American Psychological Association Standards")?
A. The validation provisions of the Guidelines are designed to be consistent with the generally accepted standards of the psychological profession. These Guidelines also interpret Federal equal employment opportunity law, and embody some policy determinations of an administrative nature. To the extent that there may be differences between particular provisions of the Guidelines and expressions of validation principles found elsewhere, the Guidelines will be given precedence by the enforcement agencies.
41. Q. When should a validity study be carried out?
A. When a selection procedure has adverse impact on any race, sex or ethnic group, the Guidelines generally call for a validity study or the elimination of adverse impact. See Sections 3A and 6, and Questions 9, 31, and 36. If a selection procedure has adverse impact, its use in making employment decisions without adequate evidence of validity would be inconsistent with the Guidelines. Users who choose to continue the use of a selection procedure with an adverse impact until the procedure is challenged increase the risk that they will be found to be engaged in discriminatory practices and will be liable for back pay awards, plaintiffs' attorneys' fees, loss of Federal contracts, subcontracts or grants, and the like. Validation studies begun on the eve of litigation have seldom been found to be adequate. Users who choose to validate selection procedures should consider the potential benefit from having a validation study completed or well underway before the procedures are administered for use in employment decisions.
42. Q. Where can a user obtain professional advice concerning validation of selection procedures?
A. Many industrial and personnel psychologists validate selection procedures, review published evidence of validity and make recommendations with respect to the use of selection procedures. Many of these individuals are members or fellows of Division 14 (Industrial and Organizational Psychology) or Division 5 (Evaluation and Measurement) of the American Psychological Association. They can be identified in the membership directory of that organization. A high level of qualification is represented by a diploma in Industrial Psychology awarded by the American Board of Professional Psychology.
Individuals with the necessary competence may come from a variety of backgrounds. The primary qualification is pertinent training and experience in the conduct of validation research.
Industrial psychologists and other persons competent in the field may be found as faculty members in colleges and universities (normally in the departments of psychology or business administration) or working as individual consultants or as members of a consulting organization.
Not all psychologists have the necessary expertise. States have boards which license and certify psychologists, but not generally in a specialty such as industrial psychology. However, State psychological associations may be a source of information as to individuals qualified to conduct validation studies. Addresses of State psychological associations or other sources of information may be obtained from the American Psychological Association, 1200 Seventeenth Street, NW., Washington, D.C. 20036.
43. Q. Can a selection procedure be a valid predictor of performance on a job in a certain location and be invalid for predicting success on a different job or the same job in a different location?
A. Yes. Because of differences in work behaviors, criterion measures, study samples or other factors, a selection procedure found to have validity in one situation does not necessarily have validity in different circumstances. Conversely, a selection procedure not found to have validity in one situation may have validity in different circumstances. For these reasons, the Guidelines require that certain standards be satisfied before a user may rely upon findings of validity in another situation. Section 7 and Section 14D. See also, Question 66. Cooperative and multi-unit studies are however encouraged, and, when those standards of the Guidelines are satisfied, validity evidence specific to each location is not required. See Section 7C and Section 8.
44. Q. Is the user of a selection procedure required to develop the procedure?
A. No. A selection procedure developed elsewhere may be used. However, the user has the obligation to show that its use for the particular job is consistent with the Guidelines. See Section 7.
45. Q. Do the Guidelines permit users to engage in cooperative efforts to meet the Guidelines?
A. Yes. The Guidelines not only permit but encourage such efforts. Where users have participated in a cooperative study which meets the validation standards of these Guidelines and proper account has been taken of variables which might affect the applicability of the study to specific users, validity evidence specific to each user will not be required. Section 8.
46. Q. Must the same method for validation be used for all parts of a selection process?
A. No. For example, where a selection process includes both a physical performance test and an interview, the physical test might be supported on the basis of content validity, and the interview on the basis of a criterion-related study.
47. Q. Is a showing of validity sufficient to assure the lawfulness of the use of a selection procedure?
A. No. The use of the selection procedure must be consistent with the validity evidence. For example, if a research study shows only that, at a given passing score the test satisfactorily screens out probable failures, the study would not justify the use of substantially different passing scores, or of ranked lists of those who passed. See Section 5G. Similarly, if the research shows that a battery is valid when a particular set of weights is used, the weights actually used must conform to those that were established by the research.
48. Q. Do the Guidelines call for a user to consider and investigate alternative selection procedures when conducting a validity study?
A. Yes. The Guidelines call for a user, when conducting a validity study, to make a reasonable effort to become aware of suitable alternative selection procedures and methods of use which have as little adverse impact as possible, and to investigate those which are suitable. Section 3B.
An alternative procedure may not previously have been used by the user for the job in question and may not have been extensively used elsewhere. Accordingly, the preliminary determination of the suitability of the alternative selection procedure for the user and job in question may have to be made on the basis of incomplete information. If on the basis of the evidence available, the user determines that the alternative selection procedure is likely to meet its legitimate needs, and is likely to have less adverse impact than the existing selection procedure, the alternative should be investigated further as a part of the validity study. The extent of the investigation should be reasonable. Thus, the investigation should continue until the user has reasonably concluded that the alternative is not useful or not suitable, or until a study of its validity has been completed. Once the full validity study has been completed, including the evidence concerning the alternative procedure, the user should evaluate the results of the study to determine which procedure should be used. See Section 3B and Question 50.
49. Q. Do the Guidelines call for a user continually to investigate "suitable alternative selection procedures and suitable alternative methods of using the selection procedure which have as little adverse impact as possible?"
A. No. There is no requirement for continual investigation. A reasonable investigation of alternatives is called for by the Guidelines as a part of any validity study. Once the study is complete and validity has been found, however, there is generally no obligation to conduct further investigations, until such time as a new study is called for. See, Sections 3B and 5K. If a government agency, complainant, civil rights organization or other person having a legitimate interest shows such a user an alternative procedure with less adverse impact and with substantial evidence of validity for the same job in similar circumstances, the user is obliged to investigate only the particular procedure which has been presented. Section 3B.
50. Q. In what circumstances do the Guidelines call for the use of an alternative selection procedure or an alternative method of using the procedure?
A. The alternative selection procedure (or method of use) should be used when it has less adverse impact and when the evidence shows that its validity is substantially the same or greater for the same job in similar circumstances. Thus, if under the original selection procedure the selection rate for black applicants was only one half (50 percent) that of the selection rate for white applicants, whereas under the alternative selection procedure the selection rate for blacks is two-thirds (67 percent) that of white applicants, the new alternative selection procedure should be used when the evidence shows substantially the same or greater validity for the alternative than for the original procedure. The same principles apply to a new user who is deciding what selection procedure to institute.
51. Q. What are the factors to be considered in determining whether the validity for one procedure is substantially the same as or greater than that of another procedure?
A. In the case of a criterion-related validity study, the factors include the importance of the criteria for which significant relationships are found, the magnitude of the relationship between selection procedure scores and criterion measures, and the size and composition of the samples used. For content validity, the strength of validity evidence would depend upon the proportion of critical and/or important job behaviors measured, and the extent to which the selection procedure resembles actual work samples or work behaviors. Where selection procedures have been validated by different strategies, or by construct validity, the determination should be made on a case-by-case basis.
52. Q. The Guidelines require consideration of alternative procedures and alternative methods of use, in light of the evidence of validity and utility and the degree of adverse impact of the procedure. How can a user know that any selection procedure with an adverse impact is lawful?
A. The Uniform Guidelines (Section 5G) expressly permit the use of a procedure in a manner supported by the evidence of validity and utility, even if another method of use has a lesser adverse impact. With respect to consideration of alternative selection procedures, if the user made a reasonable effort to become aware of alternative procedures, has considered them and investigated those which appear suitable as a part of the validity study, and has shown validity for a procedure, the user has complied with the Uniform Guidelines. The burden is then on the person challenging the procedure to show that there is another procedure with better or substantially equal validity which will accomplish the same legitimate business purposes with less adverse impact. Section 3B. See also, Albemarle Paper Co. v. Moody, 422 U.S. 405.
53. Q. Are the Guidelines consistent with the decision of the Supreme Court in Furnco Construction Corp. v. Waters, -- U.S. --, 98 S.Ct. 2943 (1978) where the Court stated: "Title VII *** does not impose a duty to adopt a hiring procedure that maximizes hiring of minority employees."
A. Yes. The quoted statement in Furnco v. Waters was made on a record where there was no adverse impact in the hiring process, no different treatment, no intentional discrimination, and no contractual obligations under E.O. 11246. Section 3B of the Guidelines is predicated upon a finding of adverse impact. Section 3B indicates that, when two or more selection procedures are available which serve a legitimate business purpose with substantially equal validity, the user should use the one which has been demonstrated to have the lesser adverse impact. Part V of the Overview of the Uniform Guidelines, in elaborating on this principle, states: "Federal equal employment opportunity law has added a requirement to the process of validation. In conducting a validation study, the employer should consider available alternatives which will achieve its legitimate purpose with lesser adverse impact."
Section 3B of the Guidelines is based on the principle enunciated in the Supreme Court decision in Albermarle Paper Co. v. Moody, 422 U.S. 405 (1975) that, even where job relatedness has been proven, the availability of other tests or selection devices which would also serve the employer's legitimate interest in "efficient and trustworthy workmanship" without a similarly undesirable racial effect would be evidence that the employer was using its tests merely as a pretext for discrimination.
Where adverse impact still exists, even though the selection procedure
has been validated, there continues to be an obligation to consider
alternative procedures which reduce or remove that adverse impact
if an opportunity presents itself to do so without sacrificing validity.
Where there is no adverse impact, the Furnco principle rather than
the Albermarle principle is applicable.
IV. Technical Standards
54. Q. How does a user choose which validation strategy to use?
A. A user should select a validation strategy or strategies which are (1) appropriate for the type of selection procedure, the job, and the employment situation, and (2) technically and administratively feasible. Whatever method of validation is used, the basic logic is one of prediction; that is, the presumption that level of performance on the selection procedure will, on the average, be indicative of level of performance on the job after selection. Thus, a criterion-related study, particularly a predictive one, is often regarded as the closet to such an ideal. See American Psychological Association Standards, pp. 26-27.
Key conditions for a criterion-related study are a substantial number of individuals for inclusion in the study, and a considerable range of performance on the selection and criterion measures. In addition, reliable and valid measures of job performance should be available, or capable of being developed. Section 14B(1). Where such circumstances exist, a user should consider use of the criterion-related strategy.
Content validity is appropriate where it is technically and administratively feasible to develop work samples or measures of operationally defined skills, knowledges, or abilities which, are a necessary prerequisite to observable work behaviors. Content validity is not appropriate for demonstrating the validity of tests of mental processes or aptitudes or characteristics; and is not appropriate for knowledges, skills or abilities which an employee will be expected to learn on the job. Section 14C(1).
The application of a construct validity strategy to support employee selection procedures is newer and less developed than criterion-related or content validity strategies. Continuing research may result in construct validity becoming more widely used. Because construct validity represents a generalization of findings, one situation in which construct validity might hold particular promise is that where it is desirable to use the same selection procedures for a variety of jobs. An overriding consideration in whether or not to consider construct validation is the availability of an individual with a high level of expertise in this field.
In some situations only one kind of validation study is likely to be appropriate. More than one strategy may be possible in other circumstances, in which case administrative considerations such as time and expense may be decisive. A combination of approaches may be feasible and desirable.
55. Q. Why do the Guidelines recognize only content, construct, and criterion-related validity?
A. These three validation strategies are recognized in the Guidelines since they represent the current professional consensus. If the professional community recognizes new strategies or substantial modifications of existing strategies, they will be considered and, if necessary, changes will be made in the Guidelines. Section 5A.
56. Q. Why don't the Uniform Guidelines state a preference for criterion-related validity over content or construct validity?
A. Generally accepted principles of the psychological profession support the use of criterion-related, content or construct validity strategies as appropriate. American Psychological Association Standards, E, pp. 25-26. This use was recognized by the supreme Court in Washington v. Davis, 426 U.S. 229, 247, fn. 13. Because the Guidelines describe the conditions under which each validity strategy is inappropriate, there is no reason to state a general preference for any one validity strategy.
57. Q. Are the Guidelines intended to restrict the development of new testing strategies, psychological theories, methods of job analysis or statistical techniques?
A. No. The Guidelines are concerned with the validity and fairness selection procedures used in making employment decisions, and are not intended to limit research and new developments. See Question 55.
58. Q. Is a full job analysis necessary for all validity studies?
A. It is required for all content and construct studies, but not for all criterion-related studies. See Sections 14A and 14B(2). Measures of the results or outcomes of work behaviors such as production rate or error rate may be used without a full job analysis where a review of information about the job shows that these criteria are important to the employment situation of the user. Similarly, measures such as absenteeism, tardiness or turnover may be used without a full job analysis if these behaviors are shown by a review of information about the job to be important in the specific situation. A rating of overall job performance may be used without a full job analysis only if the user can demonstrate its appropriateness for the specific job and employment situation through a study of the job. The Supreme Court held in Albemarle Paper Co. v. Moody, 422 U.S. 405 (1975), that measures of overall job performance should be carefully developed and their use should be standardized and controlled.
59. Q. Section 5J on interim use requires the user to have available substantial evidence of validity. What does this mean?
A. For purposes of compliance with 5J, "substantial evidence" means evidence which may not meet all the validation requirements of the Guidelines but which raises a strong inference that validity pursuant to these standards will soon be shown. Section 5J is based on the proposition that it would not be an appropriate allocation of Federal resources to bring enforcement proceedings against a user who would soon be able to satisfy fully the standards of the Guidelines. For example, a criterion-related study may have produced evidence which meets almost all of the requirements of the Guidelines with the exception that the gathering of the data of test fairness is still in progress and the fairness study has not yet produced results. If the correlation coefficient for the group as a whole permits the strong inference that the selection procedure is valid, then the selection procedure may be used on an interim basis pending the completion of the fairness study.
60. Q. What are the potential consequences to a user when a selection procedure is used on an interim basis?
A. The fact that the Guidelines permit interim use of a selection procedure under some conditions does not immunize the user from liability for back pay, attorney fees and the like, should use of the selection procedure later be found to be in violation of the Guidelines. Section 5J. For this reason, users should take steps to come into full compliance with the Guidelines as soon as possible. It is also appropriate for users to consider ways of minimizing adverse impact during the period of interim use.
61. Q. Must provisions for retesting be allowed for job-knowledge tests, where knowledge of the test content would assist in scoring well on it the second time?
A. The primary intent of the provision for retesting is that an applicant who was not selected should be given another chance. Particularly in the case of job-knowledge tests, security precautions may preclude retesting with the same test after a short time. However, the opportunity for retesting should be provided for the same job at a later time, when the applicant may have acquired more of the relevant job knowledges.
62. Q. Under what circumstances may a selection procedure be used for ranking?
A. Criterion-related and construct validity strategies are essentially empirical, statistical processes showing a relationship between performance on the selection procedure and performance on the job. To justify ranking under such validity strategies, therefore, the user need show mathematical support for the proposition that persons who receive higher scores on the procedure are likely to perform better on the job.
Content validity, on the other hand, is primarily a judgmental process concerned with the adequacy of the selection procedure as a sample of the work behaviors. Use of a selection procedure on a ranking basis may be supported by content validity if there is evidence from job analysis or other empirical data that what is measured by the selection procedure is associated with differences in levels of job performance. Section 14C(9); see also Section 5G.
Any conclusion that a content validated procedure is appropriate for ranking must rest on an inference that higher scores on the procedure are related to better job performance. The more closely and completely the selection procedure approximates the important work behaviors, the easier it is to make such an inference. Evidence that better performance on the procedure is related to greater productivity or to performance of behaviors of greater difficulty may also support such an inference.
Where the content and context of the selection procedure are unlike those of the job, as, for example, in many paper-and-pencil job knowledge tests, it is difficult to infer an association between levels of performance on the procedure and on the job. To support a test of job knowledge on a content validity basis, there must be evidence of a specific tie-in between each item of knowledge tested and one or more work behaviors. See Question 79. To justify use of such a test for ranking, it would also have to be demonstrated from empirical evidence either that mastery of more difficult work behaviors, or that mastery of a greater scope of knowledge corresponds to a greater scope of important work behaviors.
For example, for a particular warehouse worker job, the job analysis may show that lifting a 50-pound object is essential, but the job analysis does not show that lifting heavier objects is essential or would result in significantly better job performance. In this case a test of ability to lift 50 pounds could be justified on a content validity basis for a pass/fail determination. However, ranking of candidates based on relative amount of weight that can be lifted would be inappropriate.
In another instance, a job analysis may reflect that, for the job of machine operator, reading of simple instructions is not a major part of the job but is essential. Thus, reading would be a critical behavior under the Guidelines. See Section 14C(8). Since the job analysis in this example did not also show that the ability to read such instructions more quickly or to understand more complex materials would be likely to result in better job performance, a reading test supported by content validity alone should be used on a pass/fail rather than a ranking basis. In such circumstances, use of the test for ranking would have to be supported by evidence from a criterion-related (or construct) validity study.
On the other hand, in the case of a person to be hired for a typing pool, the job analysis may show that the job consists almost entirely of typing from manuscript, and that productivity can be measured directly in terms of finished typed copy. For such a job, typing constitutes not only a critical behavior, but it constitutes most of the job. A higher score on a test which measured words per minute typed, with adjustments for errors, would therefore be likely to predict better job performance than a significantly lower score. Ranking or grouping based on such a typing test would therefore be appropriate under the Guidelines.
63. Q. If selection procedures are administered by an employment agency or a consultant for an employer, is the employer relieved of responsibilities under the Guidelines?
A. No. The employer remains responsible. It is therefore expected
that the employer will have sufficient information available to
show: (a) What selection procedures are being used on its behalf;
(b) the total number of applicants for referral by race, sex and
ethnic group; (c) the number of persons, by race, sex and ethnic
group, referred to the employer; and (d) the impact of the selection
procedures and evidence of the validity of any such procedure having
an adverse impact as determined above.
A. CRITERION-RELATED VALIDITY
64. Q. Under what circumstances may success in training be used as a criterion in criterion-related validity studies?
A. Success in training is an appropriate criterion when it is (1) necessary for successful job performance or has been shown to be related to degree of proficiency on the job and (2) properly measured. Section 14B(3). The measure of success in training should be carefully developed to ensure that factors which are not job related do not influence the measure of training success. Section 14B(3).
65. Q. When may concurrent validity be used?
A. A concurrent validity strategy assumes that the findings from a criterion-related validity study of current employees can be applied to applicants for the same job. Therefore, if concurrent validity is to be used, differences between the applicant and employee groups which might affect validity should be taken into account. The user should be particularly concerned with those differences between the applicant group and current employees used in the research sample which are caused by work experience or other work related events or by prior selection of employees and selection of the sample. See Section 14B(4).
66. Q. Under what circumstances can a selection procedure be supported (on other than an interim basis) by a criterion-related validity study done elsewhere?
A. A validity study done elsewhere may provide sufficient evidence if four conditions are met (Sec. 7B):
1. The evidence from the other studies clearly demonstrates that the procedure was valid in its use elsewhere.
2. The job(s) for which the selection procedure will be used closely matches the job(s) in the original study as shown by a comparison of major work behaviors as shown by the job analyses in both contexts.
3. Evidence of fairness from the other studies is considered for those groups constituting a significant factor in the user's labor market. Section 7B(3). Where the evidence is not available the user should conduct an internal study of test fairness, if technically feasible. Section 7B(3).
4. Proper account is taken of variables which might affect the applicability of the study in the new setting, such as performance standards, work methods, representative ness of the sample in terms of experience or other relevant factors, and the currency of the study.
67. Q. What does "unfairness of selection procedure" mean?
A. When a specific score on a selection procedure has a different meaning in terms of expected job performance for members of one race, sex or ethnic group than the same score does for members of another group, the use of that selection procedure may be unfair for members of one of the groups. See section 16V. For example, if members of one group have an average score of 40 on the selection procedure, but perform on the job as well as another group which has an average score of 50, then some uses of the selection procedure would be unfair to the members of the lower scoring group. See Question 70.
68. Q. When should the user investigate the question of fairness?
A. Fairness should be investigated generally at the same time that a criterion-related validity study is conducted, or as soon thereafter as feasible. Section 14B(8).
69. Q. Why do the Guidelines require that users look for evidence of unfairness?
A. The consequences of using unfair selection procedures are severe in terms of discriminating against applicants on the basis of race, sex or ethnic group membership. Accordingly, these studies should be performed routinely where technically feasible and appropriate, whether or not the probability of finding unfairness is small. Thus, the Supreme Court indicated in Albermarle Paper Co. v. Moody, 422 U.S. 405, that a validation study was "materially deficient" because, among other reasons, it failed to investigate fairness where it was not shown to be unfeasible to do so. Moreover, the American Psychological Association Standards published in 1974 call for the investigation of test fairness in criterion-related studies wherever feasible (pp. 43-44).
70. Q. What should be done if a selection procedure is unfair for one or more groups in the relevant labor market?
A. The Guidelines discuss three options. See Section 14B(8)(d). First, the selection instrument may be replaced by another validated instrument which is fair to all groups. Second, the selection instrument may be revised to eliminate the sources of unfairness. For example, certain items may be found to be the only ones which cause the unfairness to a particular group, and these items may be deleted or replaced by others. Finally, revisions may be made on the method of use of the selection procedure to ensure that the probability of being selected is compatible with the probability of successful job performance.
The Federal enforcement agencies recognize that there is serious debate in the psychological profession on the question of test fairness, and that information on that concept is developing. Accordingly, the enforcement agencies will consider developments in this field in evaluating actions occasioned by a finding of test unfairness.
71. Q. How is test unfairness related to differential validity and to differential prediction?
A. Test unfairness refers to use of selection procedures based on scores when members of one group characteristically obtain lower scores than members of another group, and the differences are not reflected in measures of job performance. See Sections 16V and 14B(8)(a), and Question 67.
Differential validity and test unfairness are conceptually distinct. Differential validity is defined as a situation in which a given instrument has significantly different validity coefficients for different race, sex or ethnic groups. Use of a test may be unfair to some groups even when differential validity is not found.
Differential prediction is a central concept for one definition of test unfairness. Differential prediction occurs when the use of the same set of scores systematically over predicts or under predicts job performance for members of one group as compared to members of another group.
Other definitions of test unfairness which do not relate to differential prediction may, however, also be appropriately applied to employment decisions. Thus these Guidelines are not intended to choose between fairness models as long as the model selected is appropriate to the manner in which the selection procedure is used.
72. Q. What options does a user have if a criterion-related study is appropriate but is not feasible because there are not enough persons in the job?
A. There are a number of options the user should consider, depending upon the particular facts and circumstances such as:
1. Change the procedure so as to eliminate adverse impact (see Section 6A);
2. Validate a procedure through a content validity strategy, if appropriate (see Section 14C and Questions 54 and 74);
3. Use a selection procedure validated elsewhere in conformity with the Guidelines (see Sections 7-8 and Question 66);
4. Engage in a cooperative study with other facilities or users (in cooperation with such users either bilaterally or through industry or trade associations or governmental groups), or participate in research studies conducted by the state employment security system. Where different locations are combined, care is needed to insure that the jobs studied are in fact the same and that the study is adequate and in conformity with the Guidelines (see Sections 8 and 14 and Question 45).
5. Combine essentially similar jobs into a single study sample.
See Section 14B(1).
B. CONTENT VALIDITY
73. Q. Must a selection procedure supported by content validity be an actual "on the job" sample of work behaviors?
A. No. The Guidelines emphasize the importance of a close approximation between the content of the selection procedure and the observable behaviors or products of the job, so as to minimize the inferential leap between performance on the selection procedure and job performance. However, the Guidelines also permit justification on the basis of content validity of selection procedures measuring knowledges, skills, or abilities which are not necessarily samples of work behaviors if: (1) The knowledge, skill, or ability being measured is operationally defined in accord with Section 14C(4); and (2) that knowledge, skill, or ability is a prerequisite for critical or important work behaviors. In addition users may justify a requirement for training, or for experience obtained from prior employment or volunteer work, on the basis of content validity, even though the prior training or experience does not duplicate the job. See Section 14B(6).
74. Q. Is the use of a content validity strategy appropriate for a procedure measuring skills or knowledges which are taught in training after initial employment?
A. Usually not. The Guidelines state (Section 14C) that content validity is not appropriate where the selection procedure involves knowledges, skills, or abilities which the employee will be expected to learn "on the job". The phrase "on the job" is intended to apply to training which occurs after hiring, promotion or transfer. However, if an ability, such as speaking and understanding a language, takes a substantial length of time to learn, is required for successful job performance, and is not taught to those initial hires who possess it in advance, a test for that ability may be supported on a content validity basis.
75. Q. Can a measure of a trait or construct be validated on the basis of content validity?
A. No. Traits or constructs are by definition underlying characteristics which are intangible and are not directly observable. They are therefore not appropriate for the sampling approach of content validity. Some selection procedures, while labeled as construct measures, may actually be samples of observable work behaviors. Whatever the label, if the operational definitions are in fact based upon observable work behaviors, a selection procedure measuring those behaviors may be appropriately supported by a content validity strategy. For example, while a measure of the construct "dependability" should not be supported on the basis of content validity, promptness and regularity of attendance in a prior work record are frequently inquired into as a part of a selection procedure, and such measures may be supported on the basis of content validity.
76. Q. May a test which measures what the employee has learned in a training program be justified for use in employment decisions on the basis of content validity?
A. Yes. While the Guidelines (Section 14C) note that content validity is not an appropriate strategy for knowledges, skills or abilities which an employee "will be expected to learn on the job," nothing in the Guidelines suggests that a test supported by content validity is not appropriate for determining what the employee has learned on the job, or in a training program. If the content of the test is relevant to the job, it may be used for employment decisions such as retention or assignment. See Section 14C(7).
77. Q. Is a task analysis necessary to support a selection procedure based on content validity?
A. A description of all tasks is not required by the Guidelines. However, the job analysis should describe all important work behaviors and their relative importance and their level of difficulty. Sections 14C(2) and 15C(3). The job analysis should focus on observable work behaviors and, to the extent appropriate, observable work products, and the tasks associated with the important observable work behaviors and/or work products. The job analysis should identify how the critical or important work behaviors are used in the job, and should support the content of the selection procedure.
78. Q. What is required to show the content validity of a paper-and-pencil test that is intended to approximate work behaviors?
A. Where a test is intended to replicate a work behavior, content validity is established by a demonstration of the similarities between the test and the job with respect to behaviors, products, and the surrounding environmental conditions. Section 14B(4).
Paper-and-pencil tests which are intended to replicate a work behavior are most likely to be appropriate where work behaviors are performed in paper and pencil form (e.g., editing and bookkeeping). Paper-and-pencil test of effectiveness in interpersonal relations (e.g., sales or supervision), or of physical activities (e.g., automobile repair) or ability to function properly under danger (e.g., firefighters) generally are not close enough approximations of work behaviors to show content validity.
The appropriateness of tests of job knowledge, whether or not in pencil and paper form, is addressed in Question 79.
79. Q. What is required to show the content validity of a test of a job knowledge?
A. There must be a defined, well recognized body of information, and knowledge of the information must be prerequisite to performance of the required work behaviors. The work behavior(s) to which each knowledge is related should be identified on an item-by-item basis. The test should fairly sample the information that is actually used by the employee on the job, so that the level of difficulty of the test items should correspond to the level of difficulty of the knowledge as used in the work behavior. See Section 14C(1) and (4).
80. Q. Under content validity, may a selection procedure for entry into a job be justified on the grounds that the knowledges, skills or abilities measured by the selection procedure are prerequisites to successful performance in a training program?
A. Yes, but only if the training material and the training program closely approximate the content and level of difficulty of the job and if the knowledges, skills or abilities are not those taught in the training program. For example, if training materials are at a level of reading difficulty substantially in excess of the reading difficulty of materials used on the job, the Guidelines would not permit justification on a content validity basis of a reading test based on those training materials for entry into the job.
Under the Guidelines a training program itself is a selection procedure if passing it is a prerequisite to retention or advancement. See Section 2C and 14C(17). As such, the content of the training program may only be justified by the relationship between the program and critical or important behaviors of the job itself, or through a demonstration of the relationship between measures of performance in training and measures of job performance.
Under the example given above, therefore, where the requirements
in the training materials exceed those on the job, the training
program itself could not be validated on a content validity basis
if passing it is a basis for retention or promotion.
C. CONSTRUCT VALIDITY
81. Q. In Section 5, "General Standards for Validity Studies," construct validity is identified as no less acceptable than criterion-related and content validity. However, the specific requirements for construct validity, in Section 14D, seem to limit the generalizability of construct validity to the rules governing criterion-related validity. Can this apparent inconsistency be reconciled?
A. Yes. In view of the developing nature of construct validation for employment selection procedures, the approach taken concerning the generalizability of construct validity (section 14D) is intended to be a cautious one. However, construct validity may be generalized in circumstances where transportability of tests supported on the basis of criterion-related validity would not be appropriate. In establishing transportability of criterion-related validity, the jobs should have substantially the same major work behaviors. Section 7B(2). Construct validity, on the other hand, allows for situations where only some of the important work behaviors are the same. Thus, well-established measures of the construct which underlie particular work behaviors and which have been shown to be valid for some jobs may be generalized to other jobs which have some of the same work behaviors but which are different with respect to other work behaviors. Section 14D(4).
As further research and professional guidance on construct validity
in employment situations emerge, additional extensions of construct
validity for employee selection may become generally accepted in
the profession. The agencies encourage further research and professional
guidance with respect to the appropriate use of construct validity.
V. Records and Documentation
82. Q. Do the Guidelines have simplified record keeping for small users (employers who employ one hundred or fewer employees and other users not required to file EEO-1, et seq. reports)?
A. Yes. Although small users are fully covered by Federal equal employment opportunity law, the Guidelines have reduced their record-keeping burden. See option in Section 15A(1). Thus, small users need not make adverse impact determinations nor are they required to keep applicant data on a job-by-job basis. The agencies also recognize that a small user may find that some or all validation strategies are not feasible. See Question 54. If a small user has reason to believe that its selection procedures have adverse impact and validation is not feasible, it should consider other options. See Sections 7A and 8 and Questions 31, 36, 45, 66, and 72.
83. Q. Is the requirement in the Guidelines that users maintain records of the race, national origin, and sex of employees and applicants constitutional?
A. Yes. For example, the United States Court of Appeals for the First Circuit rejected a challenge on constitutional and other grounds to the Equal Employment Opportunity Commission regulations requiring State and local governmental units to furnish information as to race, national origin and sex of employees. United States v. New Hampshire, 539 F.2d 277 (1st Cir. 1976), cert. denied, sub nom. New Hampshire v. United States, 429 U.S. 1023. The Court held that the record keeping and reporting requirements promulgated under Title VII of the Civil Rights Act of 1964, as amended, were reasonably necessary for the Federal agency to determine whether the state was in compliance with Title VII and thus were authorized and constitutional. The same legal principles apply to record keeping with respect to applicants.
Under the Supremacy Clause of the Constitution, the Federal law requiring maintenance of records identifying race, sex and national origin overrides any contrary provision of State law. See Question 8.
The agencies recognize, however, that such laws have been enacted to prevent misuse of this information. Thus, employers should take appropriate steps to ensure proper use of all data. See Question 88.
84. Q. Is the user obliged to keep records which show whether its selection processes have an adverse impact on race, sex, or ethnic groups?
A. Yes. Under the Guidelines users are obliged to maintain evidence indicating the impact which their selection processes have on identifiable race, sex or ethnic groups. Sections 4 A and B. If the selection process for a job does have an adverse impact on one or more such groups, the user is expected to maintain records showing the impact for the individual procedures. Section 15A(2).
85. Q. What are the record keeping obligations of a user who cannot determine whether a selection process for a job has adverse impact because it makes an insufficient number of selections for that job in a year?
A. In such circumstances the user should collect, maintain, and have available information on the impact of the selection process and the component procedures until it can determine that adverse impact does not exist for the overall process or until the job has changed substantially. Section 15A(2)(c).
86. Q. Should applicant and selection information be maintained for race or ethnic groups constituting less than 2% of the labor force and the applicants?
A. Small employers and other small users are not obliged to keep such records. Section 15A(1). Employers with more than 100 employees and other users required to file EEO-1 et seq. reports should maintain records and other information upon which impact determinations could be made, because section 15A2 requires the maintenance of such information for "any of the groups for which records are called for by section 4B above." See also, Section 4A.
No user, regardless of size, is required to make adverse impact determinations for race or ethnic groups constituting less than 2% of the labor force and the applicants. See Question 16.
87. Q. Should information be maintained which identifies applicants and persons selected both by sex and by race or ethnic group?
A. Yes. Although the Federal agencies have decided not to require computations of adverse impact by subgroups (white males, black males, white females, black females--see Question 17), the Guidelines call for record keeping which allows identification of persons by sex, combined with race or ethnic group, so as to permit the identification of discriminatory practices on any such basis. Section 4A and 4B.
88. Q. How should a user collect data on race, sex or ethnic classifications for purposes of determining the impact of selection procedures?
A. The Guidelines have not specified any particular procedure, and the enforcement agencies will accept different procedures that capture the necessary information. Where applications are made in person, a user may maintain a log or applicant flow chart based upon visual observation, identifying the number of persons expressing an interest, by sex and by race or national origin; may in some circumstances rely upon personal knowledge of the user; or may rely upon self-identification. Where applications are not made in person and the applicants are not personally known to the employer, self-identification may be appropriate. Wherever a self-identification form is used, the employer should advise the applicant that identification by race, sex and national origin is sought, not for employment decisions, but for record-keeping in compliance with Federal law. Such self-identification forms should be kept separately from the application, and should not be a basis for employment decisions; and the applicants should be so advised. See Section 4B.
89. Q. What information should be included in documenting a validity study for purposes of these Guidelines?
A. Generally, reports of validity studies should contain all the information necessary to permit an enforcement agency to conclude whether a selection procedure has been validated. Information that is critical to this determination is denoted in Section 15 of the Guidelines by the word "(essential)."
Any reports completed after September 25, 1978, (the effective date of the Guidelines) which do not contain this information will be considered incomplete by the agencies unless there is good reason for not including the information. Users should therefore prepare validation reports according to the format of Section 15 of the Guidelines, and should carefully document the reasons if any of the information labeled "(essential)" is missing.
The major elements for all types of validation studies include the following:
When and where the study was conducted.
A description of the selection procedure, how it is used, and the results by race, sex, and ethnic group.
How the job was analyzed or reviewed and what information was obtained from this job analysis or review.
The evidence demonstrating that the selection procedure is related to the job. The nature of this evidence varies, depending upon the strategy used.
What alternative selection procedures and alternative methods of using the selection procedure were studied and the results of this study.
The name, address and telephone number of a contact person who can provide further information about the study.
The documentation requirements for each validation strategy are set forth in detail in Section 15 B, C, D, E, F, and G. Among the requirements for each validity strategy are the following:
A description of the criterion measures of job performance, how and why they were selected, and how they were used to evaluate employees.
A description of the sample used in the study, how it was selected, and the size of each race, sex, or ethnic group in it.
A description of the statistical methods used to determine whether scores on the selection procedure are related to scores on the criterion measures of job performance, and the results of these statistical calculations.
The content of the job, as identified from the job analysis.
The content of the selection procedure.
The evidence demonstrating that the content of the selection procedure is a representative sample of the content of the job.
A definition of the construct and how it relates to other constructs in the psychological literature.
The evidence that the selection procedure measures the construct.
The evidence showing that the measure of the construct is related to work behaviors which involve the construct.
90. Q. Although the records called for under "Source Data", Section 15B(11) and section 15D(11), are not listed as "Essential", the Guidelines state that each user should maintain such records, and have them available upon request of a compliance agency. Are these records necessary? Does the absence of complete records preclude the further use of research data compiled prior to the issuance of the Guidelines?
A. The Guidelines require the maintenance of these records in some form "as a necessary part of the study." Section 15A(3)(c). However, such records need not be compiled or maintained in any specific format. The term "Essential" as used in the Guidelines refers to information considered essential to the validity report. Section 15A(3)(b). The Source Data records need not be included with reports of validation or other formal reports until and unless they are specifically requested by a compliance agency. The absence of complete records does not preclude use of research data based on those records that are available. Validation studies submitted to comply with the requirements of the Guidelines may be considered inadequate to the extent that important data are missing or there is evidence that the collected data are inaccurate.
91. Q. What constitutes a "reasonable investigation of alternatives" as that phrase is used in the Answer to Question 49?
A. The Uniform Guidelines call for a reasonable investigation of alternatives for a proposed selection procedure as a part of any validity study. See Section 3B and Questions 48 and 49. A reasonable investigation of alternatives would begin with a search of the published literature (test manuals and journal articles) to develop a list of currently available selection procedures that have in the past been found to be valid for the job in question or for similar jobs. A further review would then be required of all selection procedures at least as valid as the proposed procedure to determine if any offer the probability of lesser adverse impact. Where the information on the proposed selection procedure indicates a low degree of validity and high adverse impact, and where the published literature does not suggest a better alternative, investigation of other sources (for example, professionally-available, unpublished research studies) may also be necessary before continuing use of the proposed procedure can be justified. In any event, a survey of the enforcement agencies alone does not constitute a reasonable investigation of alternatives. Professional reporting of studies of validity and adverse impact is encouraged within the constraints of practicality.
92. Q. Do significant differences between races, sexes, or ethnic groups on criterion measures mean that the criterion measures are biased?
A. Not necessarily. However, criterion instruments should be carefully constructed and data collection procedures should be carefully controlled to minimize the possibility of bias. See Section 14B(2). All steps taken to ensure that criterion measures are free from factors which would unfairly alter the scores of members of any group should be described in the validation report, as required by Section 15B(5) of the Guidelines.
93. Q. Can the use of a selection procedure which has been shown to be significantly related to only one or two job duties be justified under the Guidelines?
A. Yes. For example, where one or two work behaviors are the only
critical or important ones, the sole use of a selection procedure
which is related only to these behaviors may be appropriate. For
example, a truck driver has the major duty of driving; and in addition
handles customer accounts. Use of a selection procedure related
only to truck driving might be acceptable, even if it showed no
relationship to the handling of customer accounts. However, one
or two significant relationships may occur by chance when many relationships
are examined. In addition, in most practical situations, there are
many critical and/or important work behaviors or work outcomes.
For these reasons, reliance upon one or two significant relationships
will be subject to close review, particularly where they are not
the only important or critical ones.
[Questions and Answers 91-93 read as added, effective May 2, 1980.]94. Proposed. Q. Do federal employment nondiscrimination laws apply to employers and other UGESP-covered entities when they use the Internet and related electronic data processing technologies for recruitment and selection?
[Questions and Answers 94-98 are proposed guidelines.]Back | Uniform Guidelines | Home