help ensure technically sound and nondiscriminatory usage of selection procedures, it is not a substitute for documented evidence of validity.

c. A thorough job analysis and careful development of a selection procedure to measure or predict job performance will enhance the probability that the selection procedure will be shown valid for the job, and documentation thereof will be given appropriate weight by the enforcement agencies.

$8 Employment agencies and employment services

a. An employment agency, including private employment agencies and State employment agencies which agrees to a request by an employer or labor organization to devise and utilize a selection program should follow the standards for determining adverse impact and (if necessary) showing validity as set forth in these guidelines. An employment agency is not relieved of its obligation herein because the user did not request such validation or has requested the use of some lesser standard of validation than is provided in these guidelines. The use of an employment agency does not relieve an employer or labor organization of its responsibilities under Federal law to provide equal employment opportunity or its obligations as a user under these guidelines.

b. Where an employment agency or service is requested to administer a selection program which has been devised elsewhere and to make referrals pursuant to the results, the employment agency or service should obtain evidence of the absence of adverse impact, or of validity, as described in these guidelines, before it administers the selection program and makes referarls pursuant to the results. The employment agency must furnish on request such evidence of validity. An employment agency or service will be expected to refuse to make referrals based on the selection procedures where the employer or labor organization does not supply satisfactory evidence of validity.

$9 Restrictions on Adopting More Stringent Standards

Where a user has followed discriminatory employment practices, which are demonstrated by facts other than past failure to conform with requirements for validation of selection procedures, the user should allow members of the group discriminated against to compete for current and future vacancies on the basis of qualifications and standards no more severe than those utilized by the user in the past unless the user demonstrates that the increased standards are required by business necessity. Nothing in this paragraph is intended to prohibit a user who has not previously followed merit standards from adopting merit standards; nor does it preclude a user who has previously used invalid selection procedures from developing and using procedures which are validated in accord with these guidelines.

$10 Retesting

Users should provide a reasonable opportunity for retesting and reconsideration. The user may however take reasonable steps to preserve the security of its procedures. Where examinations are administered periodically with public notice, such reasonable opportunity exists, unless persons who have previously been tested are precluded from retesting.

$11 Affirmative Action

The use of selection procedures which have been validated pursuant to these guidelines does not relieve users of their obligations to undertake affirmative action to assure equal employment opportunity.

PART II-TECHNICAL STANDARDS

$12 Technical Standards for Validity Studies

The following minimum standards, as applicable, should be met in conducting a validity study. Nothing in these guidelines is intended to preclude the develepment and use of other professionally acceptable techniques with respect to validation of selection procedures.

a. Any validity study should be based upon a review of information about the job for which the selection procedure is to be used. The review should consist of a job analysis except as provided in § 12(c) (3) below with respect to criterion related validity. Any method of job analysis may be used if it provides the information required for the specific validation strategy used. Generally, the job analysis should result in either (1) a determination of the important duties performed on the job, and should set forth on what bases duties were determined

to be important or critical, such as the proportion of time spent on the respective duties, their levels of difficulty, their frequency of performance the consequence of error, or other appropriate factors; or (2) a direct identification by an analysis of the work process of the knowledges, skills, abilities (as defined in Appendix I(1) below) or other characteristics affecting work behavior which are necessary or important for successful job performance.

b. Measures of job knowledges, skills or abilities which are normally acquired by an applicant through subsequent work experience, formal training programs or on the job training should not be included in the selection procedure, except as a basis for selecting among qualified candidates if significant reduction in costs or improvement in job performance can be shown.

c. Criterion-related validity.-(1) Users choosing to validate a selection procedure by a criterion-related validity strategy should determine whether it is technically feasible to conduct such a study in the particular employment context. These guidelines do not require a user to hire or promote persons for the purpose of making it possible to conduct a criterion-related study; and do not require such a study on a sample of less than thirty (30) persons.

(2) There should be a review of information as described in subsection 12a above, to determine measures of work behaviors or performance that are relevant to the job in question. These measures or criteria are relevant to the extent that they represent critical or important job duties or work behaviors as developed from the job analysis. The possibility of bias should be considered both in selection of the measures and their application.

(3) The criteria should be fully described, including the rationale for their selection. Proper safeguards should be taken to insure that scores on selection procedures do not enter into any judgments of employer adequacy that are to be used as criterion measures. Criterion measures may include measures other than work proficiency such as length of service, regularity of attendance, training time or properly measured success in job relevant training. Whatever criteria are used should represent major or critical work behaviors. Job behaviors such as accident rate, tardiness, absenteeism and turnover, may be used as criteria without a full job analysis if the user can show the importance of the criterion to the particular employment context. A standardized rating of overall work performance may be utilized without a job analysis where the review of information about the job shows that it is an appropriate criterion.

(4) The sample subjects should insofar as feasible be representative of the candidates normally available in the relevant labor market for the job or jobs in question, and should insofar as feasible include the racial, ethnic and sex groups normally available in the relevant job market.

(5) The degree of relationship between selection procedure scores and criterion measures should be computed, using professionally acceptable statistical procedures. Generally, a selection procedure is considered related to the criterion, for the purposes of these guidelines, when the relationship between performance on the procedure and performance on the criterion measure is statistically significant at the .05 level of significance, which means that it is sufficiently high as to have a probability of no more than one (1) in twenty (20) to have occurred by chance. (6) Fairness of the selection procedure.-i. When one racial, ethnic or sex group characteristically obtains lower scores on a seeletion instrument than another group without corresponding differences in job performance, use of the selection instrument may unfairly deny employment opportunities to the group that obtains the lower scores.

ii. Where a selection procedure results in a significant adverse impact on a racial, ethnic or sex group identified in accordance with the classifications set forth in § 4 above and that group is a significant factor in the relevant labor market, the user generally should investigate the possible existence of test unfairness for that group if it is technically feasible to do so. The greater the severity of the adverse impact on a group, the greater the need to demonstrate the fairness of the procedure for the group. Separate studies of test fairness need not be undertaken where the weight of evidence from other studies shows that selection procedure or a substantially similar procedure is a fair predictor for the group in question and for the same or similar jobs; however, where the severity of adverse impact on a group is significantly greater than in the other studies referred to, a user should perform a separate study of test fairness.

iii. Users conducting a study of test fairness should review the APA Standards regarding investigation of possible bias in testing. Any professionally accepted model of test fairness may be selected which is appropriate to the particular em

ployment context. Investigation of test fairness in samples where the range of scores on selection procedures or criterion measures is severely restricted for any subgroup sample (as compared to other subgroup samples) may produce misleading evidence of test unfairness. That factor should accordingly be taken into account in conducting such studies and before reliance is placed upon the results. iv. If unfairness is demonstrated the user may either revise or replace the selection instrument in accordance with these guidelines, or may continue to use the selection instrument operationally with appropriate revision in its use to assure compatability between successful job performance and the probability of being selected.

v. In addition to the general conditions needed for technical feasibility for the conduct of a criterion-related study (see Appendix I(8), below) an investigation of test fairness requires the following:

(1) a sufficient number of persons in each group for findings of statistical and practical significance. These guidelines do not require a user to hire or promote persons on the basis of group classifications for the purpose of making it possible to conduct a study of test fairness; and do not require a user to conduct a study of test fairness on a sample of less than thirty (30) persons for each group involved in the study.

(2) The samples for each group should be comparable in terms of the actual job they perform, length of time on the job where time on the job is likely to affect performance, and other relevant factors likely to affect validity differences. vi. If a study of test fairness is otherwise called for but is technically infeasible, the use of a selection procedure which has otherwise met the validity standards of these guidelines will be considered in accord with these guidelines unless the technical infeasibility resulted from discriminatory employment practices which are demonstrated by facts other than past failure to conform with requirements for validation of selection procedures. However, when it becomes technically feasible for the user to perform a study of test fairness and such a study is otherwise called for, the user should conduct the study of test fairness.

(7) If the relationship between a selection procedure and a criterion measure is significant, and the user wishes to utilize a selection procedure for rank ordering, the score distribution of the selection procedure and of the criteria for job performance should be studied to determine whether differences in scores on the procedure reliably predict only the difference between successful and unsuccessful performance, or reliably predict different levels of performance.

d. Content validity.—(1) There should be a job analysis, as defined in subsection 12a above, which results in a definition of the performance domain with respect to the job in question. Definitions of the performance domain include the critical or important job duties, knowledges, skills, abilities (as defined in Appendix I(1) below) or other characteristics affecting work behavior required for successful performance of the job or a substantial part of it. A selection procedure based on inferences about psychological processes cannot be supported by content validity alone. Thus content validity by itself is not an appropriate validation strategy for intelligence, aptitude, personality or interest tests.

(2) When a selection procedure is used to rank job candidates, the definition of the performance domain should include the knowledges, skills, abilities (as defined in Appendix I(1) below) or other characteristics affecting work behavior that differentiate between barely acceptable job perofrmance and higher levels of performance.

(3) A selection procedure which is a representative sample of the performance domain of the job as defined in accordance with subsection (1) above, is a content valid procedure for the purpose of these guidelines. The reliability of selec tion procedures justified on the basis of content validity should be a matter of concern to the user. Whenever it is feasible to do so, appropriate statistical estimates should be made of the reliability of the selection procedures.

(4) The relationship between the content of the selection procedure and the performance domain of the job should be described. The closer the content of the procedure is to actual work samples (i.e., job tasks or job behaviors), the clearer the relationship. Therefore, the requirement for empirical evidence supporting the relationship increases as the content of the procedure less resembles work samples.

(5) A requirement for specific prior training or for work experience in a particular job classification, including the level or amount of training or experience, should be justified on the basis of the relationship between the content

of the training or experience and the performance domain of the job for which the training or experience is to be required.

(6) If a higher score on a content valid selection procedure test can be expected to result in better job performance, the results may be used to rank persons who score above minimum levels.

e. Construct validity.-Construct validity is a more complex strategy than either criterion-related or content validity. Accordingly, users choosing to validate a selection procedure by use of this strategy should be careful to follow professionally accepted standards, such as those contained in the APA Standards and the standard text books and journals in the field of personnel selection.

(1) There should be a job analysis as defined in §12a above. This job analysis should result in a determination of the constructs that represent successful performance of the important or critical duties of the job.

(2) A selection procedure should be selected or developed which is believed to measure the construct(s) identified in subparagraph (1) above.

(3) The selection procedure may be used operationally if the standards of subparagraphs (1) and (2) are met and there is adequate empirical research evidence showing that the procedure is validly and usefully related to performance of critical job duties. Normally, the empirical research evidence would take the form of one or more criterion related validity studies meeting the requirements of §12c. See also second sentence of $12.

(4) Where a selection procedure satisfies the standards of subsections (1), (2) and (3) above, it may be used operationally for other jobs which are shown by an appropriate job analysis to include the same construct(s) as an essential element in job performance.

f. Practical usefulness.-Users should evaluate selection procedures to assure that each is useful, in a practical sense, in the employment setting.

(1) The selection procedure should be related to successful job performance of at least one critical job duty or other significantly important aspect of job performance. The use of a single selection instrument which is related to only one job duty or a small proportion of job performance will be subject to close review.

(2) If a selection procedure is justified on the basis of content validity alone, it will be considered practically useful if it represents a substantial proportion of the performance domain.

(3) There are many factors involved in determining test usefulness for procedures justified on the basis of criterion-related validity. Generally, if other factors remain the same, the greater the magnitude of the relationship (e.g., correlation coefficient) between performance on a selection procedure and performance on the job, the more likely it is that the procedure will be useful. Other factors relevant in determining practical usefulness are the percentage of persons who can be expected to be selected out of those who are candidates for selection, and the percentage of individuals in the present workforce who are considered satisfactory or superior and who have not been selected on the basis of the selection procedure, or other simialr selection procedures. These guidelines recognize that there are no absolutes in this area, and that the concepts of test usefulness is best evaluated in each particular situation, and that the considerations may extend beyond those set forth herein.

(4) In determining whether a user may continue to utilize a selection procedure operationally which has not been clearly shown under subparagraphs (1) and (2) above, either to have usefulness or to lack usefulness, the following additional factors as appropriate should also be taken into account: the availability of other selection instruments of greater or substantially equal validity; the degree to which a procedure has an adverse impact; and the need of an employer, required by law or regulation to follow merit principles, to have an objective system of selection.

APPENDIX I-DEFINITIONS

The following definitions shall apply throughout these guidelines: (1) Ability: The present power to perform a function.

(2) Adverse Impact: Defined in § 4 of these guidelines.

(3) Employer: Any employer subject to the provisions of the Civil Rights Act of 1964, as amended, including state or local government and any Federal agency subject to the provisions of Sec. 717 of the Civil Rights Act of 1964, as amended,

and any Federal contractor or subcontractor or federally assisted construction contractor or subcontractor covered by Executive Order 11246, as amended.

(4) Employment agency: Any employment agency subject to the provisions of the Civil Rights Act of 1964, as amended.

(5) Labor organization: Any labor organization subject to the provisions of the Civil Rights Act of 1964, as amended, and any committee controlling apprenticeship or other training.

(6) Racial, ethnic or sex group: Any group of persons identifiable on the grounds of race, color, religion, sex or national origin.

(7) Selection procedure: Any measure, combination of measures, or procedure used as a basis for any employment decision. Selection procedures include the full range of assessment techniques from traditional paper and pencil tests, performance tests, and physical, educational and work experience requirements through informal or casual interviews and unscored application forms.

(8) Technical feasibility: The existence of conditions permitting the conduct of meaningful criterion related validity studies. These conditions include: (a) having or being able to obtain a sufficient number of persons available for the study to achieve findings of statistical and practical significance; (b) having or being able to obtain a sufficient range of scores on the selection procedure and job performance measures to produce validity results which can be expected to be representative of the results if the ranges normally expected were utilized; and (c) having or being able to devise unbiased, reliable and relevant measures of job performance or other criteria of employee adequacy. [Generally, if the number of persons available in a job classification or group of classifications is thirty (30) or more, the agencies issuing these guidelines would consider that there are a sufficient number of persons available.] With respect to investigation of possible unfairness, the same considerations are applicable to each group for which the study is made. See § 12c (6).

(9) Unfairness of Selection Procedure (differential prediction): A condition in which members of group characteristically obtains lower scores on a selection instrument than members of another group, without corresponding differences in job performance. See § 12c (6).

(10) User: Any employer, agency, labor organization, employment agency, licensing or certification board, or accrediting association which uses a selection procedure as a basis for any employment decision. Whenever an employer, labor organization, or employment agency is required by law to restrict recruitment for any occupation to those applicants who have met licensing or certification requirements, the licensing of certifying authority will be considered the user with respect to those licensing or certification requirements. Whenever a state employment agency or service does no more than administer or monitor a procedure as permitted by Department of Labor regulations, and does so without making referrals or taking any other action on the basis of the results, the state employment agency will not be deemed to be a user.

APPENDIX II-DOCUMENTATION OF VALIDITY EVIDENCE

Documentation standards will be issued as a part of guidelines when published. Mr. HAWKINS. Maybe we can get back to the EEOC and find out what their explanation is.

Mr. HAMPTON. Mr. Chairman, the fact there is a disagreement on how you validate tests between the Civil Service Commisison and the EEOC is absolutely true. If they would say we were holding out, it would be because we didn't agree with them; and if we would say they are holding out, it would be because they didn't agree with us. The Justice and Labor Departments have not made their decisions, but the question is basically there are a multitude of methodologies that are totally acceptable for evaluating tests and not just one methodology. This is where the difference lies. We think there are a multiplicity of acceptable methods for validating tests that are practical and less onerous.

We are anxious to get on with this business, because I think what we are interested in is progress, and I don't think we are going to have it as long as two agencies are differing on things. Now, the policy level