Mr. Justice House

delivered the opinion of the court:

The county collector of Sangamon County filed application for judgment in the county court against the objector railroads for 1957 taxes paid under protest. The railroads filed substantially similar objections and requested a refund of a portion of the taxes so paid. They appeal from orders overruling their respective objections and the two causes are consolidated here. The collector cross-appeals from an order overruling his motion to set aside an order denying a motion to strike the objections for want of jurisdiction and an order overruling bis objections to certain evidence offered by objectors. We have jurisdiction because the revenue is involved and constitutional questions are presented.

The objections are based upon the contention that the property of the Gulf, Mobile and Ohio Railroad Company and the property of the Chicago and North Western Railway Company in Illinois, as distributed to Sangamon County and its taxing districts by the Department of Revenue, was assessed and equalized at full, fair cash value, while the locally assessed property in the county and its taxing districts was assessed and equalized at levels of not to exceed 50% and 38.50%, respectively, of such value. They assert that the officials charged with assessment and equalization functions wilfully and deliberately assessed local property at the lower levels, and that the officials’ conduct in that regard resulted in gross discrimination against the railroads and was constructively fraudulent. The measure of recovery sought is the difference between what they were taxed and what they should have been taxed had locally assessed property been assessed and equalized at its full, fair cash value.

The collector questions the jurisdiction of the county court to entertain the objections on the theory that the order of the Department fixing the multiplier was a decision reviewable only under the Administrative Review Act, and that the remedy, if any, was by proceedings to have the valuation of locally assessed property raised. We address ourselves to the jurisdictional problem before considering the substantive issues.

Counsel for the collector categorizes as “pure dictum” two recent opinions of this court, People ex rel. Callahan v. Gulf, Mobile and Ohio Railroad Co. 8 Ill.2d 66, and Chicago, Burlington & Quincy Railroad Co. v. Department of Revenue, 17 Ill.2d 376, and thereafter ignores them.

The Burlington case was a suit against the Department contending that the valuation of the railroad’s Illinois property was excessive and was essentially aimed at relief through reduction of its valuation and not with the Department’s duty to equalize locally assessed property to full value. It was there held that the Department was not authorized to reduce the railroad’s valuation to less than full value, and it was stated (p. 391) : “But appropriate remedies are available if the plaintiff’s property is in fact bearing a disproportionate share of the tax burden in some taxing districts. Plaintiff may petition the Department for a reconsideration of the equalization multiplier * * * And such an objection may also be raised to a county collector’s application for judgment and order of sale of its real estate.” It was also pointed out that the time sequence made it unlikely that county equalization factors would have been determined at the time railroad valuation was being established. Thus while the availability of tax objections was dictum in that case, it was calculated, not inadvertent, and was indicative of the court’s view that there must be a means for taxpayers to resolve their dilemma by finding a proper tribunal before whom and a proper method by which they can present evidence to establish their charge that there has been discrimination against them.

The Gulf, Mobile and Ohio case (8 Ill.2d 66) arose by tax objections filed to the application for judgment and sale of real estate, as here, and the objections were based on the contention that the railroad’s property had been assessed at full, fair cash value but it had been excessively taxed because locally assessed property was fraudulently and intentionally undervalued. The availability of tax objections was held to be a proper remedy, but they were stricken on the sole ground that the objections did not allege sufficient facts to establish constructive fraud.

The collector asserts that the remedy of objectors is not by objections but by administrative proceedings.

The question of the exclusiveness of administrative review under the Administrative Review Act, (Ill. Rev. Stat. 1957, chap. 110, pars. 264 et seq.) and section 138 of the Revenue Act, (Ill. Rev. Stat. 1957, chap. 120, par. 619,) in tax cases such as this, has not heretofore been considered by us. However, this question is decided in People ex rel. Hillison v. Chicago, Burlington and Quincy Railroad Co. ante, p. 88, in an opinion handed down today. That case differs from the case at bar only in that section 148a of the Revenue Act (par. 629a) is not relied upon here. That case holds that it is not an exclusive remedy and is decisive of the issue.

The question of inapplicability of the administrative review remedy in discrimination cases such as this has been suggested but a determination of that issue is unnecessary in this case. The real question here is whether the tax-, objection remedy is available.

Section 146 of the Revenue Act (par. 627) empowers the Department to ascertain and determine the percentage relationship between the valuation placed on locally assessed property and its full, fair cash value, and then add to or deduct from the assessment in order to produce a ratio equivalent to 100%. Section 151 (par. 632) provides that after completion of equalization of the percentage finally determined, the Department shall certify it to the county clerks.

The procedure of tax objection has been available since the 1845 Revenue Act and has evolved with little change through five successive act's culminating in sections 194 and 235 of the Revenue Act of 1939. (Ill. Rev. Stat. 1957, chap. 120, pars. 675 and 716.) There seems to be no doubt but that constructive fraud was available as a defense in such a proceeding at all times until 1947. In that year section 138 (par. 619) was amended to make the provisions' of the Administrative Review Act (previously enacted in 1945) applicable to all proceedings for judicial review of final administrative decisions of the Department of Revenue. These sections have remained in force in the same form since 1947, except for an amendment added in 1957 which was primarily for the purpose of requiring payment of 100% of the amount of taxes .protested rather than 75%' as theretofore provided. This amendment tends to indicate' a legislative intent that the issue of excessive taxation re-, suiting- from- constructive fraud may continue to be used-as a defense in objections .'to application for judgment; since the General Assembly-must have been aware of our holding in the Gulf, Mobile and Ohio case. As the statutes now read, administrative review procedure is more appropriately applicable to original assessments, such as railroads and capital stock, than the equalization of assessments of local officials by the application of multipliers.

We are fully aware that relief by way of tax objection is a cumbersome and ponderous process. Nevertheless, it is a legal remedy available to objectors and they have seen fit to pursue it. We now turn to the question of whether they have sustained the burden of proving gross discrimination amounting to constructive fraud.

In order to establish their case objectors must show the level at which their property and that of locally assessed property was assessed, and that the difference was grossly discriminatory and hence constructively fraudulent. We will first consider the asessment of the railroads’ properties.

It is the duty of the Department of Revenue to assess all railroad property, other than noncarrier real estate, at its fair, cash value, (Ill. Rev. Stat. 1957, chap. 120 par. 561,) and the several factors used by the Department to determine such value have been approved by this court. (Chicago and North Western Railway Co. v. Department of Revenue, 6 Ill.2d 278; Chicago, Burlington & Quincy Railroad Co. v. Department of Revenue, 17 Ill.2d 376.) The official reports of the Department state that objectors’ properties were assessed at full, fair cash value, as do letters certifying the assessed values for the period of 1945 to 1957. There is a presumption of statutory compliance, and the testimony of Raymond S. Danis, railroad assessor of the Department, verifies the full valuation. There is no evidence in the record to overcome the presumption and proof, and we must conclude that objectors’ property was assessed at 100%.

The evidence of the ratio of locally assessed property to full, fair cash value consisted of exhibits, sales-assessment ratio studies, appraisals and appraisal studies, a survey by the Sangamon County board of review, and the testimony of experts.

The Department of Revenue determines the ratio of assessments of locally assessed property to full, fair cash value. The assessed value is obtained from abstracts of the assessment books by each county clerk after such books have been certified by the boards of review. Full value is obtained from edited samples of actual sales in each county. These samples have been secured and edited in the same manner for each year since 1951. The Department considers the sales ratios for the first three j^ears of the four-year period next preceding the subject year, since ratio studies of those three years will have been completed. The ratio of assessed to full value is first determined in each transaction, and the ratios for all eligible transactions are tabulated to determine median ratios. Separate medians in Sangamon County are determined for rural and urban property, and such median ratios are weighted to represent the proportions which each class of property is of all the property in the taxing district for which the ratio is being determined, and finally the weighted ratios are then averaged to determine the average weighted sales ratio for the county.

The sales-assessment ratios in Sangamon County thus obtained show a steady decline since 1951. For 1957 they were found to be 9.72 before equalization and after applying a multiplier of 4.7619 the equalized ratio was 46.29% of full, fair cash value. The studies and results were verified by E. L. Maynard, supervisor of the property tax division of the Department. He testified that there has been at no time a diminishing market value of property in the county from 1951 to date, and that there has been a slight annual increase over the period. He further testified that in his opinion the transactions provided adequate samples to furnish a high degree of accuracy.

Dr. Rolf A. Weil, Dean of the College of Administration of Roosevelt University expressed the opinion that the foregoing method was accurate and substantially that recommended by the National Association of Tax Administrators. He made studies of the assessment-sales ratios of the townships in which the objectors’ railroads passed and found them to be 48.05% for the Gulf, Mobile and Ohio townships and 38.90% for the Chicago and North Western townships, both after equalization. Representative appraisals also were tabulated by him (647 and 170 respectively,) and the assessment-appraisal ratios in the Gulf, Mobile and Ohio townships were 49.13% and in the Chicago and North Western townships 31.43%.

Dr. Frederick A. Ekeblad of Northwestern University testified that in his opinion the procedure followed by the Department of Revenue is well known and accepted, and affords a statistical basis for arriving at the ratio of assessment to full value with a minimum number of sample sales. He made calculations to test the reliability of ratio studies, and testified that if the full value of all property in the county were known, the probability of the assessment value being as high as 48.60% of full value was one chance in a thousand and for being as high as 49.80% was one chance in ten million. His expert opinion, as well as those of Reeves and Dr. Weil, was that the assessment-sales ratio demonstrated a level of equalized value of locally assessed property to full, fair cash value in 1957 of less than 50%.

H. Clyde Reeves, former Kentucky Commissioner of Revenue, gave an opinion that there were adequate samples to assure accuracy in the method used by the Department and agreed with the conclusions reached by Maynard, Dr. Ekeblad, and Dr. Weil.

There was much more evidence of other tests, studies and criteria bearing upon the value placed on locally assessed property which substantiates the evidence of its valuation at just under 50%. We see no purpose, however, in further comment, since it would merely be cumulative. It is worthy of note, however, that a survey conducted by the Sangamon County board of review, an important cog in tax assessment officialdom, verified the opinion evidence of the experts. The survey revealed the equalized ratio of assessed to full value of 40.81% for the county, 42.66% for the Gulf, Mobile and Ohio townships and 33.14% for the Chicago and North Western townships.

The evidence is uncontradicted and shows conclusively that there has been gross discrimination against these objecting taxpayers. The discrimination has been recognized by the Department of Revenue and by the local assessment officials as well, and amounts to constructive fraud.

The measure of recovery is the difference between the amount of taxes extended against the objectors and the amount which would have been extended had the locally assessed property been equalized at 100% of full, fair cash value. This leaves the question of whether the State, county or township ratio of underassessment is to be applied in determining the amount of refunds to which objectors are entitled.

The collector cites a number of cases under his points and authorities, without discussion, and infers that the ratio of equalized locally assessed property and its full value must be ascertained from the State as a whole. Prior to 1946 a State-wdde equalization ratio of less than 100% was fixed by the Tax Commission for debasement of railroad property. The Commission certified the assessed value of railroad properties to the counties on the basis of the ratio prevailing in the several counties. It was held in Mobile and Ohio Railroad Co. v. Tax Commission, 374 Ill. 75, that the State-wide ratio should be used and that case was followed in People ex rel. Hempen v. Baltimore and Ohio Railroad Co. 379 Ill. 543; People ex rel. Ross v. Chicago, Milwaukee, St. Paul and Pacific Railroad Co. 381 Ill. 58 and People ex rel. Ross v. Chicago, Burlington and Quincy Railroad Co. 381 Ill. 374. These cases arose before the 1945 revision of the 1939 Revenue Act when full valuation was substituted, and are of no assistance in the situation here presented. In those cases the question w^as whether the method of debasing railroad property, on other than a Statewide basis, was a proper method of equalization. Under the method of equalization now used there is no lack of uniformity in the valuation of railroad property from county to county, which is all that the State-wide ratio produced under the debasement method of equalization.

The method now used will produce uniformity if properly applied, but the evidence in this case shows that the objective of equalization at full value has not been achieved by the Department. To use an average State-wide ratio in determining the degree of discrimination against railroad property would penalize the taxpayers in those counties in which the goal of full-value assessment has most nearly been achieved, and we are therefore of the opinion that the State-wide ratio is not appropriate.

The question then is whether discrimination should be ascertained at the township or county level. The collector contends that the evidence is insufficient to establish the township level of assessment. The amount of the railroads’ assessment allocated to the several townships by the Department of Revenue was in the record, and, in our opinion, the proof was sufficient to show the ratio of underassessment in the railroad townships. He also asserts that the county level is inappropriate because of the difference in ratios between the townships.

In deciding the level to be used it must be borne in mind that this is not a case of equalization but is the seeking of a formula to provide a refund of discriminatory taxes. While assessments are originally made at the township level, the board of review has the responsibility of equalizing between townships to the end that the townships within the county be assessed uniformly. In order to establish Statewide uniformity it is the duty of the Department of Revenue to assign a multiplier to each county, and the county, not the township, is the unit recognized for application of the multiplier. If the taxing officials at each stage followed the statute literally there would be uniformity at all levels, but that has not been attained. There is not complete equalization between the townships, nor in fact between individual properties within a township.

We are of the opinion, nevertheless, that the countywide weighted average should be used in this case. The proof does not indicate a gross difference between application of the county ratio and the township ratio. Since the use of either ratio would produce substantially the same result, considerations of practicality dictate the use of the county ratio. Although at least one of the objectors will have a lesser recovery by use of the county ratio, both objectors urged its use and substantial justice will not be denied by its use.

Both objectors have contended that the county wide ratio of assessed to full, fair cash value is not to exceed 50% and, although the proof established that it was just below that percentage, the)r apparently concede the fairness of the round figure of 50%.

The Chicago and North Western protested on the basis of a ratio of locally assessed property of 38.50% to full, fair cash value, but its recovery is limited by use of the county ratio fixed at 50%. The Gulf, Mobile and Ohio is in a different position. Its protest is predicated upon a ratio of locally assessed property of 55% to full, fair cash value, and of course its recovery is limited to the amount of its protest. With these limitations, objectors are entitled to a refund based upon the measure of recovery above stated.

This decision does not solve the problem in other counties, save only as to the availability of the remedy and the measure of recovery to be used. Similar proceedings with adequate proof would have to be made in each county where discrimination is charged in order to recover, because the figures used in determining the multiplier to be applied in each county are different. The fact that an erroneous multiplier is used in one county does not establish that the multipliers in the other counties are erroneous; and even if such other multipliers can be shown to be erroneous, they would not necessarily produce the same discrepancy, percentage-wise, as here alleged.

The judgment of the county court of Sangamon County is reversed, and the cause is remanded, with directions to order refunds according to the measure of recovery, as limited, provided herein.

Reversed and remanded, with directions.