The respondent, City of Fremont, Nebraska (City), has appealed from the decision of the Commission of Industrial Relations (CIR), which entered an opinion and order dated July 14, 1982, establishing wages for 33 different job classifications, beginning August 1, 1981. For reasons which we will specifically discuss, we must reverse the order of the CIR and remand the matter for further proceedings.
The International Brotherhood of Electrical Workers Local 1536 (Union) filed a petition with the CIR on August 28, 1981, alleging an industrial dispute existed because the Union and the City had reached an impasse in the negotiation of a new collective bargaining agreement.
At the pretrial conference the Union orally moved to be allowed to call James L. Jensen, the City’s expert witness, as a witness for the Union’s case in chief. The City objected, and both parties filed briefs with the CIR. The CIR ruled in favor of the *359Union and permitted it to call Mr. Jensen. At trial the City objected to the Union calling Mr. Jensen as a witness, moved to dismiss at the close of the Union’s case, and renewed its objection in the motion for new trial.
The City contends the CIR erred in (1) failing to grant the City’s motion to dismiss, since the Union failed to prove “same or similar work” or a prima facie case of noncomparability; (2) ordering wages without taking into account the differences in pay schedules in compared-to locations regarding the length of time it takes a worker to advance from the minimum to maximum pay level on the salary schedule; (3) establishing wages without sufficient data; (4) ordering wage increases in excess of the evidence regarding certain classifications of workers; (5) calculating of overall compensation, since evidence shows that the City was above the prevalent in payment of fringe benefits; (6) allowing the Union to use the City’s expert witness when competent expert witnesses were available; and (7) ordering wage and fringe benefits for nonbargaining Union employees.
In our review of orders and decisions of the CIR, we are restricted to considering whether the order of that agency is supported by substantial evidence justifying the order made, whether it acted within the scope of statutory authority, and whether its action was arbitrary, capricious, or unreasonable. IAFF Local 831 v. City of No. Platte, 215 Neb. 89, 337 N.W.2d 716 (1983); AFSCME Local 2088 v. County of Douglas, 208 Neb. 511, 304 N.W.2d 368 (1981), modified 209 Neb. 597, 309 N.W.2d 65 (1981).
The law is clear that the burden is on the moving party in a Neb. Rev. Stat. § 48-818 (Cum. Supp. 1982) case to demonstrate that existing wages are not comparable to the prevalent wage rate. General Drivers and Helpers Union v. City of West Point, 204 Neb. 238, 281 N.W.2d 772 (1979). It is necessary for the Union to establish by evidence what were “the *360prevalent wage rates paid and conditions of employment maintained for the same or similar work of workers exhibiting like or similar skills under the same or similar working conditions.” § 48-818.
In making the comparison required by statute, it is necessary to take into consideration the overall compensation presently received by the employees, having regard not only to wages for time actually worked but also to wages for time not worked, including vacations, holidays, and other excused time, and all benefits received, including insurance and pensions, and the continuity and stability of employment enjoyed by the employee.
In the present case the Union has satisfied the burden of proof of establishing a prima facie case of noncomparability. The Union’s evidence consisted of comparisons with employees of the Nebraska cities of Beatrice, Grand Island, Hastings, Norfolk, and North Platte; the public power districts of Central Nebraska Public Power and Irrigation District and Nebraska Public Power District; and the Iowa cities of Ames, Cedar Falls, and Muscatine.
The City’s evidence consisted of comparisons with employees of the Nebraska cities of Beatrice, Bellevue, Columbus, Grand Island, Hastings, Kearney, Norfolk, and North Platte; the public power districts of Norris Public Power District, Loup Public Power District, and the Nebraska Public Power District (Kramer plant); and the following cities outside Nebraska: McPherson and Manhattan, Kansas; Ames and Ankeny, Iowa; and Yankton, South Dakota.
From the arrays of both parties there were seven common employers: the Nebraska cities of Beatrice, Grand Island, Hastings, Norfolk, and North Platte; Nebraska Public Power District (Kramer plant); and Ames, Iowa. From these common employers the CIR created its array used to compute wages.
The CIR, in reaching its conclusion that the seven-member array was appropriately balanced and suf*361ficiently similar with enough like characteristics to be appropriate, used the figures presented by both parties, with minor adjustments for discrepancies. Reasons given for the discrepancies were that (1) the parties did not always agree which job class in a specific locality matched the key job class in Fremont; (2) the parties disagreed on what the maximum pay scale is on Grand Island’s pay plan; and (3) the Union provided adjusted wage figures which took into account changes in pay rate since the filing of the petition.
Determinations made by the CIR in accepting or rejecting claimed comparables are within the field of expertise of the CIR and should be given due deference, and, by necessity, in attempting to arrive at comparables, require granting some discretion to the CIR. AFSCME Local 2088 v. County of Douglas, supra.
The second assignment of error and primary contention of the City is that the CIR failed to take into account the differences in pay schedules in compared-to locations regarding the length of time it takes a worker to advance from the minimum to the maximum pay level. The CIR, in denying the motion for new trial and reconsideration, found that this issue had been considered by the CIR when it rendered its original decision.
Fremont had a salary schedule of 2y2 years, compared to salary schedules of 10 years for Grand Island, 8 years for Hastings, 5 to 8 years for North Platte, 7 years for Norfolk, and 4% years for Ames, Iowa. The City argues the CIR should have adopted offered evidence of what the pay would have been for the other cities, determined at the 21/¿>-year mark, in order to restructure Fremont’s salary schedule.
The Union argues the salary schedule is not at issue, and since the CIR set wages by plugging in high and low salary figures to set the salary structure, the CIR should not consider midrange salary compu*362tations. The Union also argues the evidence offered at trial by the City was not to establish a prevalent wage, but to show that there was an inappropriate classification comparison used for the “key job classifications.”
In the pretrial order the structure of the salary schedule was not among the items listed by the Union as being at issue. Nor was the issue of salary schedule raised in the City’s answer or amended answer. The City did not request the CIR to amend the pay schedule, because it was satisfied with its salary schedule, but thought the other cities’ pay periods should be adjusted to be comparable to the short term used in Fremont.
The City contends adjustments should have been made in other employers’ compared-to schedules to reflect the differences in salary schedules, or, at the very least, the City should now be permitted to adjust its salary schedule to the comparable employers. We do not agree with the Union’s assertion that the salary schedule was not at issue.
The Union further argues an adjustment in the salary schedule to reflect a 2%-year midrange adjustment would be inappropriate because the method of adjustment is one used for determining compensation in teacher pay cases and may not necessarily be applied to nonteacher cases.
With regard to a determination of conditions of employment under § 48-818, the CIR has the following alternatives: If, in the compared-to cities, the exact content of a condition of employment is clear from the total evidence, the CIR may enter an order either as to that condition of employment or find that the subject city’s condition in this regard is lesser or greater than the prevalent, and adjust the “overall compensation” accordingly.
Section 48-818 states in part: “The findings and order or orders may establish or alter the scale of wages, hours of labor, or conditions of employment, or any one or more of the same.” We believe the *363timing difference in the salary schedules — which, it seems to us, is another way of defining experience and longevity in employment — is a condition of employment in this situation which must be adjusted to reach a comparability determination.
“The salary step schedule merely determined that employees with less than 1 year’s experience, employees with 1 year’s experience, and employees with more than 2 years’ experience, all within the same job classification, should be paid wages differently depending upon their experience. It was a recognition that experience and longevity in employment may provide á basis for an increase in compensation over the starting salary fixed for a particular job classification.” Plattsmouth Pol. Dept. Coll. Barg. Comm. v. City of Plattsmouth, 205 Neb. 567, 569, 288 N.W.2d 729, 731 (1980).
Evidence offered by the City, but apparently disregarded in part by the CIR, was a series of exhibits, Nos. 2 to 12, prepared by James L. Jensen. These exhibits consisted of minimum, midrange, and maximum salary ranges for a variety of occupations in the various cities surveyed. He explained the purpose of using midrange figures as follows: “We use the mid-range salary for a number of reasons. I want, first of all, to determine that they have a range of salary and how they’re generally structured for any given class of work. And so that we know that after a person’s been on the job for two, two and a half years, normally think of that person as being able to function at a journeyman level for whatever that job might be. So I’d like to know what the salary is for — at that stage of a person’s development of the job. Also we need to know the maximum salaries. . . . The mid-range salary I use is, if we can determine it is — and usually we can — is what the person would normally be paid in that job after being on the job for two to two and a half years.”
If we understand What the CIR did, it computed a *364minimum and maximum salary for Fremont, using those comparable figures from the cities in its array. As a result — and again we must add the limiting word “apparently” — Fremont will now be paying maximum salaries after 2y2 years, whereas Grand Island does not reach that point until 10 years; North Platte, sometime between 5 to 8 years; Norfolk, after 7 years; and Ames, Iowa, for 4% years.
Referring to the City’s exhibits 2, 6, and 9, as well as table 23 of the CIR’s opinion and order, we can perhaps demonstrate the objection raised by the City. The following is a listing of the minimum, midrange, and maximum hourly wages paid by comparable cities surveyed for three randomly selected occupations:
Water Serviceman II
Grand Island $5.43 $6.29 $6.86
North Platte 6.16 6.82 7.50
Norfolk 5.43 6.00 7.23
Ames, Iowa 6.33 7.56 8.45
Fremont * 5.93 7.90 7.90
Lineman First Class
Grand Island 7.68 8.99 9.86
North Platte 8.24 • 9.04 11.37
Norfolk N/A
Ames, Iowa 9.74 11.04(?) 11.04
Fremont * 8.63 11.87 11.87
Meter Reader
Grand Island 5.10 5.94 6.49
North Platte 5.87 6.48 7.16
Norfolk 5.16 5.70 6.91
Ames, Iowa 5.56 6.30 7.41
Fremont * 5.22 6.96 6.96
*As a result of the order of the CIR.
We may accept from the record the fact that the relationship between the minimum and maximum salaries of the surveyed cities and Fremont seems to be reasonable. However, the disparity appears when attempting to correlate the midrange salaries, *365realizing that Fremont must pay that salary all during the time that the other cities do not reach their maximum salary levels, Grand Island, for example, for up to another 7y2 years.
There may be a logical explanation for what the CIR did in this regard, but it does not appear in its order. For that reason the order of the CIR must be reversed and the cause remanded with directions to adjust the wage schedules for timing differences, or to justify its failure to do so.
The City also complains that the CIR erred in establishing wages without sufficient data to utilize the key class method. In establishing the wages for the 33 job classes, the CIR surveyed 13 key job classes to set wages within each progression. Specifically, the City argues that (1) the CIR did not meet two of the nine criteria established by the CIR to promote use of the key class method; (2) the CIR established wages for Lineman First Class in excess of the evidence for that class of workers; and (3) the CIR did not provide the City with an offset for more favorable fringe benefits or adjust the fringe benefits according to what was prevalent.
The City contends the Union failed to satisfy guidelines (5) and (6), which, in summary form, are: (5) “Key classifications” together with related lines of progression must permit direct or computed establishment of at least 85 percent of the classes involved; and (6) The “key classifications” or wage rates tendered should have at least one in each regular line of progression, or lines of progression should be established as comparable, or their relative market value established or apparent to the commission from its previous experience.
The City argued, that Utility Technician II, Tree Trimmer IV, and Electric Serviceman II classifications should be eliminated because there were four or fewer comparisons within the seven-employer array and that such a reduction would leave only 25 classes, or 75.75 percent of the 33 classes of work. *366Furthermore, the City also argued that the nonbargaining classes of Electrical Technician II and Engineering Technician II should be eliminated to reduce the figure to 74.19 percent.
In AFSCME Local 2088 v. County of Douglas, 208 Neb. 511, 304 N.W.2d 368 (1981), the CIR guideline (5) was interpreted. We allowed the CIR to add several nonsurvey positions to fit within one of the clerical lines of progression to reach the required percentage of job classifications. In the present case we find no error in the CIR including nonbargaining classes to meet the required percentage of job classifications. The technical argument on the specific number of comparisons is a more difficult question.
The City contends job classifications based upon three and four comparisons should be rejected because they fall short of the CIR’s expressed preference, suggesting that it has rejected the use of two employers and three employers where there were other comparable employments available. However, we believe the CIR should be allowed to make a statistically significant determination with three comparisons in a seven-employer array.
We believe that where the employer has a large number of job classifications, but established lines of progression and established relationships between lines of progression, “key job classifications” may be utilized to establish wages for all classes, provided reasonable requirements of prevalence and relevance are met. We think that to set down a specific mathematical rule in § 48-818 cases would limit the discretion allowed the CIR to make comparability determinations.
Also, guidelines used by the CIR are not statutory requirements, and the failure of the evidence to strictly comply with the guidelines does not require us to find the'action of the CIR in developing “key job classifications” was arbitrary and capricious. Thus, despite the CIR’s past expressed preference for six to eight comparisons, the actual number of *367comparisons needed is a question of reasonableness for the CIR.
The City next argues guideline (6) is not met because (1) there is not a key class in each line of progression, (2) there is no evidence to conclude the lines of progression are comparable, and (3) there is no proof of the relative market value. In AFSCME Local 2088 v. County of Douglas, supra, we found the requirements of guideline (6) could be satisfied by all the evidence taken as a whole, and the requirement concerning the lines of progression is only one of three options of that guideline.
In the present case a comparison of the job classifications in the bargaining unit with the Union wage rate exhibits proves the Union submitted key classification data in every line of progression, and computation of wage rates based on the key classification data is possible for 100 percent of the classes involved in the bargaining unit. The City’s contention is unsupported by the evidence.
The City also claimed the CIR established wages for Lineman First Class in excess of the evidence for that classification of workers. The basis of the City’s argument is that the selection of Lineman Apprentice as the “key job classification” was inappropriate because the duties of the class were in dispute. Again, the City’s contention is unsupported. The Union submitted evidentiary exhibits and testimony relevant to this dispute. The CIR considered this evidence along with the City’s evidence and resolved this dispute. We find this portion of the decision to be supported by substantial evidence.
The utilization of percentages of key classifications may distort the value of jobs below and above the key classification by valuing a class below the key class somewhat lower than what might be established by the market, while at the same time valuing a class above the key class somewhat higher than what might be established by the market. However, we believe this risk of somewhat higher valuation is *368acceptable in order to make the task of analyzing large numbers of job classifications manageable.
We affirm the use of the “key job classification” by the CIR in arriving at comparable salaries for public employees having a large number of classifications.
In addition, the City contends there should have been an offset or adjustment because almost every one of the fringe benefits was above the prevalent. This contention that the CIR failed to consider fringe benefits is unsupported by the record. The CIR’s order specifically states: “In considering ‘overall compensation’ the Commission finds that fringe benefits and other working conditions in dispute shall remain unchanged.” Furthermore, tables 14 through 22, attached to the CIR order, are sufficient substantial evidence to support the CIR’s decision, and we conclude the CIR decision is not arbitrary or unreasonable in this regard.
The sixth assignment of error argued by the City is the CIR’s allowing the Union to call the City’s expert witness, James L. Jensen, to testify. In IAFF Local 831 v. City of No. Platte, 215 Neb. 89, 337 N.W.2d 716 (1983), where the same issue was presented, we allowed the union to call the city’s expert witness as a witness in its case in chief when the expert was merely called upon to express an opinion he had already formed prior to trial. In that case we said: “The trial performs a truth-seeking function in settling controversies between the parties. It runs contrary to that function to suggest that unfavorable evidence in the form of an expert’s opinion may be kept from either party in the interest of justice and to prevent prejudice.” Id. at 98, 337 N.W.2d at 722.
In the case before us Mr. Jensen was not called upon to make any additional investigation, but was called upon to express opinions he had already formed prior to trial. We believe the CIR was correct to allow Mr. Jensen to testify.
*369The City further argues that although Mr. Jensen is a qualified expert witness in this matter, his expert testimony should have been excluded because he had insufficient foundation upon which to base his opinion. An expert witness should not be allowed to express an opinion where the evidence reveals there is no adequate basis for the opinion. Plattsmouth Pol. Dept. Coll. Barg. Comm. v. City of Plattsmouth, 205 Neb. 567, 288 N.W.2d 729 (1980).
The City relies on our holding in Plattsmouth Pol. Dept., supra, to argue the expert’s testimony did not include a verification of actual work performed by those classified under various job positions.
In the present case Mr. Jensen is not without knowledge of the working conditions and content of individual job classifications in the compared-to utilities offered by both sides. First, the Union presented comparability evidence on each of the utility employers in its array, separate from the testimony of Mr. Jensen and of other witnesses. Second, Mr. Jensen has conducted wage and fringe benefit surveys for various municipalities for 16 years.
Prior to his consulting work, Mr. Jensen was employed as an administrative assistant by the city of Ames, Iowa, one of the compared-to utilities in the arrays of both parties. He had conducted the annual wage survey for Fremont for the last 5 years. These surveys included wage and fringe benefit comparisons for electrical, water, and sewage employees in Fremont and several Nebraska cities. The other cities in Mr. Jensen’s prior surveys included Grand Island, Hastings, Norfolk, North Platte, and Beatrice. As a consultant, Mr. Jensen had previously participated in drafting Grand Island’s job descriptions, drafted those of Beatrice and Fremont, developed a job classification plan for Hastings, and was in the process of developing a job classification plan for North Platte.
The third reason to find sufficient foundation is that Mr. Jensen personally visited the surveyed utili*370ties, and upon arriving at a given site checked with and visited with supervisors, department heads, or employees in verifying the job descriptions for comparability. Mr. Jensen did not always contact the actual personnel filling positions for every job in the survey, but, instead, used his judgment whether it was practical or feasible to track down and observe crews of personnel in action. We find that sufficient foundation existed to form an adequate basis to express an opinion.
The last assignment of error is that the CIR erred in ordering. wages and fringe benefits for nonbargaining units Electrical Technician II and Engineering Technician II. The Union agrees with the City that these two classifications are not within the bargaining unit and should be excluded. This court agrees.
The action of the CIR in this case is reversed, and the cause is remanded for further proceedings in accordance with this opinion.
Reversed and remanded with directions.