The opinion of the Court was delivered by
O’HERN, J.This appeal concerns appellant’s eligibility to be a candidate for appointment as a police officer in a civil service community. New Jersey’s Constitution mandates that “[ajppointments and promotions in the civil service of the State * * * shall be made according to merit and fitness to be ascertained, as far as practicable, by examination * * N.J.Const. art. VII, § 1, para. 2. The Merit System Board of the New Jersey Department of Personnel (Board), which administers the Civil Service system, ruled that appellant was not eligible for the appointment because she was unfit. However, we are unable to determine how the specific findings in the record demonstrate the unfitness to serve that is the only basis to remove her name from the list of eligible candidates. Accordingly, we remand *536the matter to the Department of Personnel to relate its findings in this case to the statutory qualifications.
I
The parties do not appear to dispute the version of the facts set forth in appellant’s brief; hence, we draw our recitation from that source.
Anastasia Vey first began working for the North Wildwood Police Department (NWPD) as a summer police officer in June 1982. She continued to work for the NWPD for the next several summers. In 1986, she decided that she wanted to work as a police officer full-time, year-round. That year, she applied as a candidate to both the New Jersey State Police and the NWPD. She passed a state-administered Civil Service examination for the position of police officer. The New Jersey Department of Personnel certified her as “eligible” for employment and placed her name on an “eligible list.” Vey also had to submit to various psychological evaluations. The State Police offered her a position, but she declined, hoping to work in her home town of North Wildwood.
The NWPD notified Vey that it was not accepting her as a member of its January 1987 class, due to a negative psychological report. In that report, which was based on a clinical interview and the results of the psychological tests, the examiner characterized plaintiff as pleasant, cooperative, and articulate; above average intellectually; suffering to a mild degree from fear of heights; well oriented, alert, and highly motivated; impulsive, anxious, and agitated; worrying more than most people; somewhat apprehensive and insecure; a likely candidate for distraction; having a low frustration tolerance; and tending to become easily bored. The evaluator concluded that “[a]lthough she appears motivated to obtain a full-time position as a police officer, there is considerable evidence at hand to suggest that her chances of experiencing success in that capacity are below average at this time.” Although the NWPD did *537not accept her for its 1987 class, it did not seek to remove her name from the eligible list at that time.
Vey sought appointment to the NWPD’s January 1988 class. She requested a new psychological evaluation. The NWPD sent her back to the same association that had performed the first evaluation. The second evaluator, who based his report in part on the first evaluation, found plaintiff to be friendly, cooperative, and verbal; average to low-average in intellectual functioning; open and frank during personality assessment; above average in interpersonal skills; positive in her outlook on life, with good reality testing; more likely than the average person to become emotional and frustrated; impulsive and suspicious; and assertive and bold, with the suggestion of likely conflict between herself and supervisory personnel. The evaluator concluded that “while Ms. Vey has many strengths, her potential for functioning well in the long term in law enforcement as a police officer are below average.”
The City of Wildwood requested that Vey’s name be removed from the eligible list. Vey appealed the proposed removal of her name. In support of her position, she submitted the report of an independent psychologist she had retained. The psychologist found Vey to be above average in leadership potential; open and cooperative; extroverted, enthusiastic, verbal, spontaneous, and uninhibited; pleasant and responsive to the needs of others; and approaching tasks and problems with an emphasis on rationality. He concluded that “Ms. Vey appears to [have] the necessary ability to be an effective police person. * * * Ms. Vey is a very attractive candidate for this position.” Vey also submitted the positive recommendations of her former supervisors and commanding officers. Those supervisors found Vey to be “an outstanding person,” “well liked,” “an asset to our squad,” and as one who “leaves nothing to be desired in all phases of her work.”
Despite those favorable records, the Medical Review Panel (Panel) found that the “weight of the documentation supports a *538conclusion that the hiring authority has met its burden of proof that [Vey] is mentally unfit to perform effectively the duties of the position,” and recommended that her name be removed from the eligible list.
At Vey’s request, the Board remanded the matter to the Panel for further consideration. The Panel again recommended that Vey be removed from the eligible list. At that time, the Board accepted the Panel’s findings and concluded that “the appointing authority has met its burden of proof that Ms. Vey is mentally unfit to perform effectively the duties of a Police Officer and therefore [we order] that her name be removed from the eligible list.”
On Vey’s appeal, the Appellate Division found that the record disclosed Vey to be “a very intelligent, decent and well-motivated person,” and noted that “[w]ere we the fact-finder we might well differ with the personnel decision of the appointing authority.” However, a majority of the court concluded that it was required to accord deference to the agency in matters of “scientific specialization.” A dissenting member of the Appellate Division concluded that
[t]he mere fact that a psychological evaluation of [Vey] based on testing resulted in a characterization of “the applicant’s potential for functioning well as a police officer [as] below average ” cannot support the conclusion of the expert or the Medical Review Board that “the hiring authority has met its burden of proof that [Vey] is mentally unfit to perform effectively the duties of the position.”
Vey appealed to us as of right. R. 2:2-1(a).
II
As noted, the New Jersey Constitution reaffirms our longstanding commitment that appointments and promotions in the Civil Service shall be made “according to merit and fitness.” N.J. Const. art. VII, § 1, para. 2. The 1986 reform of the Civil Service Act (Act) recognizes that commitment. The Act created a Department of Personnel to replace the Civil Service Commission. The Department of Personnel consists of a Merit System *539Board, a Commissioner of Personnel, and other subdivisions. N.J.S.A. 11A:2-1. For the selection and appointment of civil-service employees, the Commissioner must administer examinations that “test fairly the knowledge, skills and abilities” required to perform the task of a given job classification. N.J. S.A.. 11A:4-I(a). Those examinations “may include, but are not limited to, written, oral, performance [tests] and evaluation of education and experience.” Ibid. On the basis of such tests, the Commissioner establishes an “eligible list.” Once the Commissioner has certified an eligible list, an appointing authority must make its appointments from the list. N.J.S.A. HA:4-5. The Board has the authority to hear and decide appeals concerning various civil-service employment matters. N.J.S.A. HA:2-6. The Board may submit psychological appeals to the Panel for a recommendation. N.J.A.C. 4A:4-6.5(f).
The Civil Service Act delegates broad authority to the Board to adopt and enforce rules and regulations to carry out the Act. N.J.S.A. 11A:2-1(a).1 One regulation provides for the removal for cause of a candidate’s name from an eligible list. N.J.A.C. 4:1-8.14(a)(5) (current version at N.J.A.C. 4A:4-6.2(a)(5)). Cause includes when the candidate “[i]s physically or psychologically unfit to perform effectively the duties of the title.” N.J.A.C. 4:1-8.14(b)(3) (current version at N.J.A.C. 4A:4-6.-1(a)(3)).
A candidate has a right to appeal a decision to remove his or her name. N.J.S.A. 11A:2-14. The Act vests the Board with the final administrative determination over such an appeal. On appeal, the appointing authority must file with the Board all psychological and psychiatric reports that were the basis for the requested removal. N.J.A.C. 4:l-8.26(d) (current version at *540N.J.A.C. 4A:4-6.5(c)). Those reports must include a “diagnosis or specific, detailed statement” showing
a behavioral pattern that clearly indicates that the appellant is unable to effectively perform the duties of the title; or
specific reasons for the requested removal on a psychiatric or psychological basis.
[N.J.A.C. 4:1-8.26(d)(3) (current version at N.J.A.C. 4A:4-6.5(e)(3)) (numeration and structure added).]
The appointing authority bears the burden of proving that the candidate meets the regulatory description and is subject to removal for psychological unfitness. N.J.A.C. 4:1-8.26(e) (current version at N.J.A.C. 4A:4-6.3(b)).
Ill
The use of psychological tests to predict or evaluate employee job performance is a recognized part of the American workplace. Psychological tests, like intelligence or agility tests, are only as good as their correlation to actual job performance. The most significant discussion of such employment tests has arisen in connection with the fair-employment claims. Title VII of the Civil Rights Act of 1964 permits the use of employment tests, 42 U.S.C.A. § 2000e-2(h), and “those tests may have a differential impact so long as they are valid for [employee] groups.” Comment, Psychological Aptitude Tests and the Duty to Supply Information: NLRB v. Detroit Edison Co., 91 Harv.L.Rev. 869, 873 n. 30 (1978). “Validity is predictive power: the more accurately a test predicts an employee’s job performance, the more valid it is. Empirical validity * * * is established by a direct statistical correlation between test scores and performance ratings and is not necessarily obvious from inspection of the test questions.” Id. at 873 n. 31.
Particularly in the case of police and fire fighters, courts have had to examine the validity of various tests, including intelligence, agility, and psychological tests. In Craig v. County of Los Angeles, 626 F.2d 659, 662 (9th Cir.1980), cert. denied, 450 U.S. 919, 101 S.Ct. 1364, 67 L.Ed.2d 345 (1981), the court explained a validation process that involves three distinct steps. *541First, the employer must specify the trait or characteristic that the selection device is being used to identify or measure. Next, the employer must establish that the particular trait or characteristic is an important element of work behavior. Finally, the employer must “demonstrate by ‘professionally acceptable methods’ that the selection device is ‘predictive of or significantly correlated’ with the element of work behavior identified in the second step.” Ibid, (quoting Albemarle Paper Co. v. Moody, 422 U.S. 405, 431, 95 S.Ct. 2362, 45 L.Ed.2d 280, 304 (1975)); see also Washington v. Davis, 426 U.S. 229, 250-51, 96 S.Ct. 2040, 2052-53, 48 L.Ed.2d 597, 613-14 (1976) (despite disproportionate impact of intelligence testing, test was validated for purpose of predicting success in police training program); Blake v. City of Los Angeles, 595 F.2d 1367, 1378-79 (9th Cir.1979) (test for police officer), cert. denied, 446 U.S. 928, 100 S.Ct. 1865, 64 L.Ed.2d 281 (1980); Williams v. City & County of San Francisco, 483 F.Supp. 335, 342 (N.D.Cal.1979) (test for juvenile probation officer), rev’d, 685 F.2d 450 (9th Cir.1982).
The reason for validation is that frequently the tests themselves incorporate, perhaps inadvertently, non-job-related criteria. For example, a test used by Sears, Roebuck & Co. in hiring for commission-sales positions asked such questions as: “ ‘Do you have a low pitched voice?’ ‘Do you swear often?’ ‘Have you participated in wrestling?’ ‘Have you played on a football team?’ ” Note, Getting Women Work That Isn’t Women’s Work: Challenging Gender Biases in the Workplace Under Title VII, 97 Yale L.J. 1397, 1406 (1988) (quoting brief submitted in Equal Employment Opportunity Comm’n v. Sears, Roebuck & Co., 628 F.Supp. 1264 (N.D.Ill.1986)), aff'd 839 F.2d 302 (7th Cir.1988). The question is whether the hiring criteria were simply modeled on traits associated with men or whether such tests do predict effectively who will be a good commission salesperson.
Although the case before us is not an employment-discrimination case and does not raise questions of disproportionate impact of the testing criteria, the difficulty here is that a gap *542exists between the evidence of the identified personality traits and the condition of unfitness that has to be met under the regulation. The use of the specific personality traits ascribed to this candidate as predictors of job performance has not been validated. It may be that such validation would not be required if the test disclosed a recognized mental disease or defect. Such a person might meet the criterion of the regulation that speaks of disqualification because the candidate is “psychologically unfit.” The record does not disclose whether that language in the regulation refers to commonly-recognized categories of mental disorder. See State v. Pitts, 116 N.J. 580, 608, 562 A.2d 1320 (1989) (noting mental disorders recognized by the American Psychiatric Association in DSM III-R: Diagnostic and Statistical Manual of Psychiatric Disorders 226-28 and 251 (3d ed. Rev.1987)). In this case, however, the record describes a variety of seemingly unremarkable personality traits with a conclusion that they demonstrate a below-average potential on the part of appellant. In psychological reviews of applicants, evaluators may use language containing euphemisms that are well understood by others in the field to convey a diagnosis of mental dysfunction. If that is so, then the diagnosis must be clearly stated because courts are simply unable to review agency action if there is a coded way of communicating significant medical findings.
If, on the other hand, the evaluator simply notes various personality traits, then there ought to be some validation that the described personality traits do in fact correlate to job performance. The agency’s own standard disqualifies a candidate for being “unfit,” not for failing to be the best qualified for the job. For example, an evaluator may find that an applicant is impulsive. If that characteristic is shown to pose a danger of loss of control or of excessive use of force by officers, then it would certainly be a personality trait that could make one unfit to serve as a police officer. Similarly, if a personality trait such as excessive timidity or unwillingness to *543use force inhibits the effective performance of police duties, that too might constitute grounds for disqualification.
In Vey’s case, evaluators found her to have a below-average potential in part because they believed her “boldness” and “suspiciousness” could pose potential problems between Vey and her supervisors. At first glance, it just does not seem to us to be a sign of unfitness that a police officer be bold or suspicious. As Vey’s attorney noted, “[i]t is difficult to envision an effective police officer who is neither bold nor suspicious.” In fact, Vey’s summer employment record contained several positive evaluations and recommendations for rehire from her supervisors. The point is that the record is simply silent concerning whether and how the use of the identified personality traits has been validated as an accurate predictor of performance as a police officer. Even though this is not an employment-discrimination case, we believe that the agency “should reasonably bear the burden of establishing the job-validity of its * * * psychological tests.” Maine Human Rights Comm’n v. City of Auburn, 425 A.2d 990, 997 (Me.1981).
We do not in any sense suggest or encourage the appointment of below-average candidates for police service. In many instances, the police chief makes the final hiring decision. The police chief should always look for superior candidates, not just average candidates. In interviews, the police chief seeks to identify those applicants who, in addition to performing well on the Civil Service examinations, measure up to the police chief’s judgment. The final test, consequently, is “inevitably subjective.” Ibid. That does not mean that it is inappropriate, particularly in the case of relatively small departments. No candidate can ask for more than this: a fair chance to impress that final interviewer with those qualities of leadership, loyalty, and determination that will make a good police officer.
IV
Although administrative agencies are entitled to discretion in making decisions, that discretion is not unbounded *544and must be exercised in a manner that will facilitate judicial review. Administrative agencies must “articulate the standards and principles that govern their discretionary decisions in as much detail as possible.” Van Holten Group v. Elizabethtown Water Co., 121 N.J. 48, 67, 577 A.2d 829 (1990). When the absence of particular findings hinders or detracts from effective appellate review, the court may remand the matter to the agency for a clearer statement of findings and later reconsideration. Application of Howard Sav. Inst, 32 N.J. 29, 53, 159 A.2d 113 (1960).
This case contains a gap in the administrative record. The record does not disclose the basis for equating the described personality traits with the projected job performance. Nor does the record disclose to us a finding of a recognized mental disorder. The findings recite the presence of personality traits, but there is no evidence of a correlation between such nonpathological test results and actual job performance. If a trait is to be deemed disqualifying to the extent that it demonstrates psychological unfitness, there must be evidence in the record of both the trait itself and the trait’s correlation with the standard of psychological unfitness. That correlation may be demonstrated in any of the familiar ways for establishing the validity of such psychological tests. Because the record does not disclose the basis for the agency’s equation of the described traits with the condition of unfitness, the matter must be remanded to the agency.
The judgment of the Appellate Division is reversed and the cause is remanded to the Department of Personnel for further proceedings in accordance with this opinion.
At the inception of this case, N.J.A.C. 4:1-8.1 to -16.13 contained the relevant Civil Service Rules. Those Rules have since been repealed and recoded in N.J.A.C. 4A:4-1.1 to -7.10. No one has suggested that those changes substantively affect this case.