Geraldine Laverne Mayes, a black female 43 years of age, filed a complaint with the Mayor’s Commission on Human Rights and Community Relations of the City of Springfield, Missouri, alleging that she had been diseriminatorily discharged from employment because of her race. The Commission held a hearing, found racial discrimination and ordered French, among other things, to provide Mrs. Mayes with backpay and to place her on its callback list.
French thereupon filed a petition for review in the Circuit Court of Greene County pursuant to the provisions of Chapter 536 and Rule 100.1 Upon consideration of the whole record, the trial court concluded that the Commission’s order was not supported *721by competent and substantial evidence and reversed its finding and order.
At the outset, we must note that the substantive ordinance involved, which is codified as Springfield City Code Section 18A-3(a), and which was proved and offered on trial, is a rescript of that part of Title VII of the Civil Rights Act of 1964, as amended, now codified as 42 U.S.C.A. § 2000e-2(a) and our § 296.020(1). So, assuming as we do that Chapter 296 was not intended to be preemptive, see Ruggeri v. City of St. Louis, 441 S.W.2d 361, 365[6] (Mo.1969), we have before us a Title VII employment discrimination case. Given the paucity of precedents construing our fair employment practices act, we shall regard the federal decisions and the regulations implementing and interpreting the federal statutes as persuasive. Kansas City v. Missouri Commission on Human Rights, 632 S.W.2d 488, 490[3] (Mo. banc 1982).
The Commission heard a good deal of oral testimony; it is summarized and set forth in the trial court’s 13-page, 48-paragraph memorandum, and in the view we take of this appeal, an extensive summary of the subjective evidence received is wholly unnecessary. Presumably the Commission is aware, as was the trial court, that the federal courts have developed at least two theories for approaching and analyzing discrimination claims: a disparate treatment theory and a disparate impact theory. Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971); McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973). Put differently, Title VII violations were originally perceived as specific acts of discrimination against an individual person, actionable under the provisions of 42 U.S.C.A. § 2000e-5(f). Since 1972, the federal statute, and presumably its imitative rescripts, have viewed employment discrimination primarily in terms of systems and practices which impose a substantially disproportionate impact on a protected group. Discriminatory employment systems and practices are actionable under the provisions of 42 U.S.C.A. § 2000e-6(e).2
The Commission chose to regard this claim as combining a complaint of disparate treatment and disparate impact. Because the Commission’s argument in this court was rigidly put in terms of “burdens” and the “shifting” thereof, we shall follow that mechanistic approach in dealing with the claim of disparate treatment, noting one crucial limitation the Commission apparently overlooked.
Because this is a case in which the claimant relied on the disparate treatment theory to establish a case of discriminatory discharge, it may be said that it had three phases or “shifts of burden”: 1) plaintiff’s establishment of a prima facie case; 2) the employer’s rebuttal of the prima facie case by articulation of some legitimate non-discriminatory reason for the discharge, and 3) plaintiff’s proof that the reason given was pretextual. Johnson v. Bunny Bread Co., 646 F.2d 1250, 1253[1] (8th Cir.1981); McCosh v. City of Grand Forks, 628 F.2d 1058, 1062[2] (8th Cir.1980). In this connection, however, we must bear in mind that the E.E.O. Act does not prohibit arbitrary discharge; it is directed only to specific impermissible bases of discrimination. Garcia v. Gloor, 618 F.2d 264, 269[7] (5th Cir.1980). Therefore in a discriminatory discharge case, assuming the third “phase” is reached, “proof of a discriminatory motive is critical.” Johnson v. Bunny Bread Co., supra, 646 F.2d at 1254, quoting International Brotherhood of Teamsters v. United States, 431 U.S. at 335, n. 15, 97 S.Ct. 1854.
In Johnson v. Bunny Bread Co., supra, 646 F.2d at 1253, the court set out the elements of a prima facie case of discriminatory discharge as follows: A claimant *722must show that 1) he was a member of a protected class; 2) he was capable of performing the job, and 3) he was discharged from the job. Here, the complainant certainly established that she was a member of a protected class and that she was discharged. We are in doubt that she established she was capable of performing the job. Rather, her own testimony consisted of a series of reasons why she had not performed satisfactorily on the three-employee production line at which she worked. One or another of the other employees adjusted the conveyor belt to such speed that she was unable- to keep pace; she was unable to remain at the production line because she had a difficult menstrual period, which for some reason required her to change clothes frequently during a single shift, etc.
Assuming arguendo that the claimant made out a prima facie case, the burden then shifted to the employer to articulate some legitimate non-discriminatory reason for the discharge. McDonnell Douglas Corp. v. Green, supra, 411 U.S. at 802, 93 S.Ct. at 1824, 36 L.Ed. at 678. The employer’s burden is not however, a burden of persuasion; the burden of articulation only requires an employer to present admissible evidence sufficient to raise a genuine issue of fact as to whether the complainant was discriminated against. Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 253-256, 101 S.Ct. 1089, 1093-1095, 67 L.Ed.2d 207, 215-217 (1981); Johnson v. Bunny Bread Co., supra, 646 F.2d at 1254. Placing this burden of production on the defendant serves simultaneously to meet the plaintiff’s prima facie case and to frame a legitimate reason for the action with sufficient clarity so that the complainant will have a full and fair opportunity to demonstrate pretext. The sufficiency of the articulation is to be judged by the degree to which it fulfills this function. The employer is not required to persuade the trier of fact that it was actually motivated by the proffered reasons. Texas Dept, of Community Affairs v. Burdine, supra, 450 U.S. at 254-256, 101 S.Ct. at 1094-1095, 67 L.Ed. at 216-217; Johnson v. Bunny Bread Co., supra, 646 F.2d at 1254.
One Culbertson, French’s personnel director, testified that Mrs. Mayes was discharged because of “poor attitude.” She had unsatisfactory reports from both of her supervisors, including such conduct as too much talking, lack of hustle and enthusiasm and too many restroom trips. Complainant’s own testimony indicates as much. Mrs. Mayes had personal problems with another employee, Doris, and a supervisor, Linda Tucker. She knew she was to be evaluated on her cooperation with her peers and supervisors. French does not counsel probationary employees concerning bad working habits because it believes such counsel will temporarily correct the undesirable habit but eventually the specific behavior trait will return. Such “articulation” was sufficient to show a legitimate nondiscriminatory reason for complainant’s discharge and it framed the factual issue with sufficient clarity to provide the plaintiff with a full and fair opportunity to demonstrate pretext. Miller v. Jones Truck Lines, Inc., 476 F.Supp. 1048 (E.D.Ark.1979). See also, Texas Department of Community Affairs v. Burdine, supra, 450 U.S. at 251, 101 S.Ct. at 1092, 67 L.Ed.2d at 214; General Motors Corporation v. Fair Employment Practices Division of the Council on Human Relations of the City of St. Louis, 574 S.W.2d 394, 398[4, 5] (Mo. banc 1978).
The complainant then had the opportunity, and the burden, of proving the articulated reasons were pretextual. Furnco Construction Corp. v. Waters, 438 U.S. 567, 578, 98 S.Ct. 2943, 2950, 57 L.Ed.2d 957, 968 (1978). As noted, some proof of discriminatory motive is crucial. And, in a discriminatory discharge case, a complainant must show a similarity between his conduct and that of white employees not discharged. Johnson v. Bunny Bread Co., supra, 646 F.2d at 1255.
*723Statistical evidence may be used in a discriminatory discharge case, Harper v. Trans World Airlines, Inc., 525 F.2d 409, 412—413[4], 34 A.L.R.Fed. 639, 644-645 (8th Cir.1975), but to have evidentiary value, statistics must have statistical relevance. As the trial court correctly concluded and noted, small samples are usually held insufficient to support any conclusion as to discriminatory effect or discriminatory intent. Harper v. Trans World Airlines, Inc., supra, 525 F.2d at 412[4], 34 A.L.R.Fed. at 644-645; Robinson v. City of Dallas, 514 F.2d 1271, 1273 (5th Cir.1975). The reason given by statisticians for the rejection of small samples is that they probably do not truly reflect the employer’s policy. D. Baldus & J. Cole, Statistical Proof of Discrimination, § 9.12, p. 301 (1980). The only statistical evidence before the Commission having any bearing upon discriminatory motive or having any remote tendency to indicate a similarity between the complainant’s conduct and that of white employees not discharged was Commission’s Exhibits 9 and 10. Exhibit 10, as a sample, includes 3 employees; Exhibit 9 undertakes to compare 14 employees. We might develop the matter further, but the samples are too small to permit an inference of discriminatory intent. See Harper v. Trans World Airlines, Inc., supra, 525 F.2d at 412, 34 A.L.R.Fed. at 644-645. Taking the most sanguine view of the complainant’s proof, it might be said she proved that she was arbitrarily discharged; it cannot be said she established that she was discriminatorily discharged.
As noted, the Commission treated this case as combining a claim of disparate treatment and disparate classwide impact. We need not, for our present purposes, analyze the powers invested in the Commission by the substantive ordinance in evidence. It is sufficient to say it appears to have been the intent of the City of Springfield, as a constitutional charter city, to create the Mayor’s Commission on Human Rights and Community Relations in the image of the E.E.O.C. itself. The Commission has, in any case, purported to find that French’s employment policies have a disparate class-wide impact; its jurisdiction to entertain disparate impact claims is not challenged, and as noted, we have assumed Chapter 296 was not intended to be preemptive.
The general purpose of the disparate impact principle is to eliminate even those facially neutral or benign employment practices which operate to impose a substantially disproportionate burden on a protected group. Griggs v. Duke Power Co., supra, 401 U.S. at 430-431, 91 S.Ct. at 853, 28 L.Ed.2d at 163-164; Wright v. National Archives & Records Service, 609 F.2d 702, 711 (4th Cir.1979); Robinson v. City of Dallas, supra, 514 F.2d at 1272. An intent to discriminate is not an element of the disparate impact principle.
To establish a prima facie case under the disparate impact theory, it must be shown that a particular job requirement or selection or promotion process selects applicants for hire or promotion in a racial pattern significantly different from that of the relevant pool of applicants. Albemarle Paper Co. v. Moody, 422 U.S. 405, 425, 95 S.Ct. 2362, 2375, 45 L.Ed.2d 280, 301 (1975); Bauer v. Bailar, 647 F.2d 1037, 1042 (10th Cir.1981); Wright v. National Archives & Records Service, supra, 609 F.2d at 711. Establishment of a prima facie case requires proof that the discrimination is more than an occurrence of isolated, accidental or sporadic discriminatory acts. The complainant must “establish by a preponderance of the evidence that racial discrimination [is] the [employer’s] standard operating procedure — the regular rather than the unusual practice.” International Brotherhood of Teamsters v. United States, supra, 431 U.S. 324, 336, 97 S.Ct. 1843, 1855, 52 L.Ed.2d 396, 416 (1977) (footnote omitted). Statistical evidence may be used and may be sufficient of itself to establish a prima facie case. Robinson v. City of Dallas, supra, 514 F.2d at 1273.
If a complainant establishes a pri-ma facie case the burden shifts to the em*724ployer to prove that the questioned requirements or selection or promotion processes (“criteria”) are “job related,” i.e., that they have a manifest relation to the employment in question. Albemarle Paper Company v. Moody, supra, 422 U.S. at 425, 95 S.Ct. at 2375, 45 L.Ed.2d at 301; Bauer v. Bailar, supra, 647 F.2d at 1042. If the employer proves a manifest relation between the criterion and the job performance the burden again shifts to the complainant to demonstrate other methods which would serve the employer’s legitimate, interests without resulting in disproportionate exclusion of the complainant’s protected group. Albermarle Paper Co. v. Moody, supra, 422 U.S. at 425, 95 S.Ct. at 2375, 45 L.Ed.2d at 301.
One of the objectives of the disparate impact theory is to restrict the use of subjective judgments in hiring and promotion which result over an extended period in disproportionate exclusion of the members of a protected group. Bauer v. Bailar, supra, 647 F.2d at 1042. French’s criterion may be somewhat subjective, but it is not on its face discriminatory; the criterion is essentially based on performance in training, i.e., performance during the “probationary period.” The appropriate E.E.O.C. guideline, 29 C.F.R. § 1607.14 B(3), provides, in pertinent part:
“... Where performance in training is used as a criterion, success in training should be properly measured and the relevance of the training [may] be shown ... through a demonstration of the relationship between measures of performance in training and measures of job performance. Measures of relative success in training include but are not limited to instructor evaluations, performance samples, or tests.... ”
Reduced to its essence, French’s criterion for the retention of probationary employees came within this definition.
The Commission inferentially found that French’s criterion produced a disparate classwide impact. This finding or inference is based wholly on statistical proof, which the trial court found insufficient. We also find the statistical evidence to be without probative value, for much the same reason as the trial court, as reflected in paragraphs 30 to 37, inclusive, of the trial court’s findings.
In the first place, the Commission chose to use raw percentage data based upon small samples. Admitting that judicial approaches to the persuasiveness of statistical proof vary, depending on the specific type of discrimination involved, the judicial tendency is to regard invalidated statistical proof as insufficient to prove either disparate treatment or disparate impact. One finds a typical rejection of such data in the majority opinion in Eubanks v. Pickens-Bond Const. Co., 635 F.2d 1341, 1349-1350 (8th Cir.1980): (our emphasis)
“... The district court merely stated the actual number and percentage of black foremen for the three years in question and the number and percentage of black cement finishers on the First National Bank job during that period. To infer a prima facie case of discrimination from these numbers and percentages, the court needed to go beyond presentation of the raw data to application of an appropriate and recognized test for determining the statistical significance of any observed disparity. See Mayor v. Educational Equality League, 415 U.S. 605, 619-20, 94 S.Ct. 1323, 1332-1333, 39 L.Ed.2d 630 (1974) (rejecting as unreliable ‘simplistic percentage comparisons’ of the racial composition of a thirteen-member nominating panel and the population of Philadelphia). Cf. Hazelwood School District v. United States, supra, 433 U.S. [299] at 308-09, n. 14, 97 S.Ct. [2736] at 2741-2742, n. 14 [53 L.Ed.2d 768 (1977)] . (standard deviation analysis). See also D. Baldus & J. Cole, supra, at 322-28 (selected methods for calculating significance) .... ”
We wish to be specific in our holding in this case. Our ruling should not be taken *725as giving a judicial stamp of approval to any particular method of statistical proof of discrimination, nor as a pretense of statistical expertise. We rule only on the facts before us and undertake only such analysis of the principles involved as is essential to a resolution of this appeal. We do hold:
1. That the sample size in Exhibits 9 and 10 is too small to establish a prima facie case of disparate treatment through statistical analysis, at least without the application of some recognized method of statistical validation. Eubanks v. PickensBond Const. Co., supra, 635 F.2d at 1350[9, 10]; Harper v. Trans World Airlines, Inc., supra, 525 F.2d at 412; D. Baldus & J. Cole, supra, at 297-303.
2. Because Exhibits 9 and 10 do not reflect the pool of actual applicants during the period covered, it is impossible to construct a model reflecting anything but the “actual” selection rate. See D. Baldus & J. Cole, supra at 78-79. Although a gross disparity in the actual selection rate is indicated, the proof is not probative on the issue of disparate impact because the sample is small and is not statistically validated.
3. The only data before the Commission which reflects the effect of French’s selection policy is Commission’s Exhibit 8, which shows that of the total number of French employees in 1979, 95.59 percent were white and 2.27 percent were black; in 1980, 95.53 percent were white and 3.19 percent were black. This exhibit indicates compliance with the “four-fifths” rule guideline set out for the evaluation of selection rates for disparate impact by the E.E.O.C. This guideline is set out as 29 C.F.R. § 1607.4 D, and in pertinent part reads:
“A selection rate for any race, sex, or ethnic group which is less than four-fifths (⅜⅞) (or eighty percent) of the rate for the group with the highest rate will generally be regarded by the Federal enforcement agencies as evidence of adverse impact, while a greater than four-fifths rate will generally not be regarded by Federal enforcement agencies as evidence of adverse impact....”
There are other statistical flaws in the Commission’s analysis, discernible by resort to standard texts, but we need not lengthen the opinion by discussing them.
Finally, we must address the Commission’s insistent point that the trial court erred in reversing its decision because it substituted its view of the evidence for that of the Commission by finding there was no competent and substantial evidence upon the whole record to support the Commission’s findings. In addressing this contention, we do not agree that in every ease, the review of the trial court and that of this court is necessarily limited by § 536.140.1. See: Drcy v. State Tax Commission, 323 S.W.2d 719, 722[2] (Mo.1959). We do agree that if we review under § 536.140.1, then both we and the circuit court review the decision of the agency; both this court and the circuit court defer to the administrative decision and we, like the circuit court, must affirm the administrative decision unless it appears that the determination does not rest on competent and substantial evidence or is otherwise invalid. Phipps v. School Disk of Kansas City, 645 S.W.2d 91, 95-96 (Mo.App.1982).3 And, in this case, we need go no further than to consider the scope of appellate review under § 536.140.1. The term “substantial evidence” means evidence which has probative force upon the issues. State ex rel. Rice v. Public Service Commission, 359 Mo. 109, 114, 220 S.W.2d 61, 64[3] (Mo. banc 1949). The Commission based its *726findings on statistics and inferences which had no probative force, and the trial court so held. Further to be noted is that the cause came to this court by way of appeal just as any other civil case, § 536.140.6, and the trial court heard no new or additional evidence. Therefore, although we review the record made before the Commission, Fleming Foods of Missouri, Inc. v. Runyan, 634 S.W.2d 183, 184[1] (Mo. banc 1982), as the disposition of that case shows, it is the judgment of the trial court to which our mandate is directed. Accordingly and for the reasons stated, the judgment of the trial court is affirmed.
PREWITT, P.J., and TITUS, Alternate Judge, concur. MAUS, J., took no part in the consideration or the decision of this cause.. References to statutes and rules are to RSMo 1978 and Missouri Rules of Court (13th ed. 1982).
. See H.R.Rep. No. 238, 92d Cong., 1st Sess. 8 (1972), reprinted in [1972] U.S.Code Cong. & Ad.News 2137, 2143-2144; Jackson & Mathe-son, The Continuing Violation Theory and the Concept of Jurisdiction in Title VII Suits, 67 Geo.L.J. 811, 823-824, nn. 71, 74 (1979).
. We welcome the opportunity to dispel any notion that Newcomb v. Patton, 608 S.W.2d 145 (Mo.App.1980), should be taken as expressing any view of the law in conflict with Phipps. In Newcomb, this court had a hybrid record before it; most of the record was produced in the trial court and we should have held ourselves at liberty to redetermine the facts and redeclare the law of the case, as did the court in Phipps, supra, 645 S.W.2d at 100. As to the scope of appellate review of a noncontested case, we most cordially agree with the rationale articulated in the Phipps case.