Opinion for the Court filed by Senior Circuit Judge McGOWAN.
McGOWAN, Senior Circuit Judge:At issue in this proceeding are newly promulgated Interstate Commerce Commission (ICC or Commission) regulations permitting motor common carriers to allocate higher platform handling costs to smaller-size shipments.1 Previously, the ICC had required carriers to allocate these costs in direct proportion to weight regardless of shipment size. The Commission has shifted its long-standing policy in this area so that shipping rates may reflect the greater time per pound needed to handle smaller shipments on trucking terminal platforms.
Petitioner trade associations represent the numerous corporate customers of the regulated trucking industry who regularly ship smaller-size packages. Facing several hundred million dollars in potential rate increases as a result of the Commission’s action, they raise a series of procedural and substantive challenges to the new ICC regulations. Because petitioners have not met their burden of proof with respect to these objections, we affirm.
I
In setting its rates, the regulated trucking industry must take into account the costs associated with transporting the goods it carries. See, e.g., 49 C.F.R. §§ 1139.1-.8 (1982). The ICC breaks down these costs into four general service categories including line haul, pickup and delivery, billing and collecting, and platform handling. This case concerns costs associated with platform handling, or that portion of trucking service devoted to the loading and unloading of shipper packages at carrier terminal platforms. These costs include the salaries and wages of employees who work on the platform (loading and unloading freight and processing shipping documents), the costs of operating, maintaining, and repairing platform equipment (such as hand trucks, forklifts, and other mechanical lifting devices), and other expenses directly or indirectly related to the carrier’s platform operations.
The present controversy focuses on how carriers should allocate these costs among the various-size shipments that they handle. In the past, the ICC required carriers to apportion platform costs in direct relation to weight regardless of shipment size. Thus, if a carrier moved one million pounds of traffic at a total platform handling cost of one million dollars, each pound of traffic was deemed to cost one dollar to move across the carrier’s platforms. A 100-pound shipment would therefore be charged with $100 of platform cost, while a 200-pound shipment would be charged with $200.
The problem with this cost accounting technique is that it failed to take into account the economies of scale realized by progressively larger shipments. Platform handling costs are by their nature transactional and largely represent constant labor inputs that do not vary with shipment weight. In handling a particular shipment, for example, the platform worker must locate the shipment on the platform or incoming vehicle, identify its destination from the shipping documents, determine a suitable outgoing vehicle, if necessary, and perform the required cargo movement. It is clear that this process does not take twice as long for a 200-pound shipment as for a 100-pound shipment. On the contrary, one would expect both shipments to require roughly the same time to process. Although costs do increase for substantially larger shipments (owing partly to the need for more expensive handling equipment), *340the increase in costs is not nearly proportional to the increase in weight and, in any case, tends to be offset by the productivity gains realized through mechanization.
The effect of the ICC’s long-standing policy, then, was to misallocate costs between large and small shipments, with disproportionately high charges on large shipments subsidizing artificially low charges on small shipments. The flaw in the Commission’s cost accounting procedures went unchallenged until the 1960’s, however. At that time, the railroads were introducing cost-saving measures that enabled them to gain a competitive rate advantage over trucking firms in handling larger-size shipments. To meet this competition, the trucking industry was forced to lower its rates on truckload shipments,2 but could not recoup the resulting revenue loss by increasing rates on smaller shipments because of the Commission’s existing cost allocation formula. To correct the imbalance, the industry began to press for a change in the formula. Small Shipment Rate Revision—Eastern Central Territory, 335 I.C.C. 547, 548-49 (1969), vacated and remanded sub nom. National Small Shipments Traffic Conference, Inc. v. United States, 321 F.Supp. 500 (1970) (three-judge court).
We need not detail here the checkered history of the Commission’s original attempts to accommodate this industry demand through investigatory proceedings. Suffice it to say that those attempts failed after protracted litigation before the Commission and in federal court. See National Small Shipments Traffic Conference, Inc. v. ICC, 590 F.2d 345, 346-55 (D.C.Cir.1978) (recounting the history of ICC approvals and federal court remands of two earlier cost studies). At issue in those proceedings was the validity of two earlier studies designed to show the precise handling-cost differentials between large and small shipments. After numerous rebuffs in federal court, the Commission abandoned its efforts to validate these studies and decided to undertake a third and final study of its own.3
That study began in 1976. Unlike its predecessors, it relied on a scientific random sample of platform movements nationwide. Although responsibility for the investigation rested with the Commission’s Bureau of Accounts, certain aspects of the research work were subcontracted to private firms and consultants. The purpose of the study was to correlate platform handling time (as the dependent variable) with various shipment characteristics of weight, volume, and number of pieces (as the independent variables). Errors in data collection eventually forced the Bureau of Accounts to drop the shipment-volume variable from consideration, however.
The Bureau published the results of its work in 1979. Bureau of Accounts, Interstate Commerce Comm’n, 1977-1978 Motor Carrier Platform Study (1979) (ICC Statement No. 2S1-79). The study concluded that smaller-size shipments do indeed require more handling time per pound on carrier platforms, as do shipments with larger numbers of pieces. Moreover, by grouping the study data into categories by weight and number of pieces, the Bureau was able to estimate the exact handling-time differentials among these categories. Id. at 74-75 (Tables 7A-7B). These tables indicate that while shipments in the 500-999 pound category require an average of 1.4 minutes of handling time per hundredweight,4 shipments in the 1-149 pound category require 4.3 minutes, or nearly three times as long.5 Because current ICC re*341porting procedures do not provide data on number of pieces, the Bureau recommended adopting an interim formula based on weight alone that would be superseded by a more complex weight-and-pieces formula once proper data become available. Id. at 59-66.
Two months after the Bureau released its study, the ICC issued a notice of proposed rulemaking to determine whether the Commission should adopt the study recommendations. 44 Fed.Reg. 53,190 (1979) (to be codified at 49 C.F.R. §§ 1104A.1-.2) (proposed Sept. 10, 1979). Comments and replies were filed by various motor carrier rate bureaus, traffic leagues, and shipper associations, along with two divisions within the ICC acting as advocates on either side of the question. For over two years, the ICC weighed the comments it had received and analyzed the study data. In August 1982, the Commission reached a final decision to adopt the study recommendations, and petitioners thereafter filed this petition for review. See supra note 1.
II
Petitioners first raise a series of procedural challenges to the Commission’s action, including alleged improprieties in the choice of rulemaking over adjudication, separation of staff functions, ex parte contacts, and suppression of negative staff recommendations. We address each of these allegations in turn.
A. Rulemaking versus Adjudication
Petitioners contend that the controversy over platform cost allocation is essentially adversary in nature, and that the Commission therefore erred in shifting the issue from its original adjudicatory setting to the present rulemaking arena. See Brief of Petitioners at 28-30. The original adjudications to which petitioners refer are the investigatory proceedings that the Commission conducted between 1969 and 1978 to validate the two original cost studies in this area.6 Although they recognize that the present proceedings do not concern those two earlier studies, petitioners argue that the underlying controversy remains the same, and that adjudication therefore constitutes the only permissible mode of administrative action in this case.
It is hornbook administrative law that the choice of proceeding by individual adjudication or general rulemaking lies largely within the agency’s discretion. SEC v. Chenery, 332 U.S. 194, 203, 67 S.Ct. 1575, 1580-1581, 91 L.Ed. 1995 (1947). The Interstate Commerce Act grants the ICC general rulemaking powers, 49 U.S.C. § 10321(a) (Supp. V 1981), and the Commission has chosen to proceed under those powers in this case, 44 Fed.Reg. 53,190 (1979) (“This rulemaking is instituted pursuant to 49 U.S.C. 10321 and 5 U.S.C. 553, 559.”). That proposed agency action inspires adversary contention among interested parties is immaterial to whether adjudicatory or rule-making procedures must be followed. Indeed, it is the rare government action that does not provoke some type of controversy among affected individuals and corporations.
Absent specific statutory instructions, the proper focus in determining what procedures apply is on the nature of the administrative task at hand. 3 K. Davis, Administrative Law Treatise § 14:4 (2d ed. 1980). Trial-like procedures are particularly appropriate for retrospective determination of specific facts about individual parties. In certain cases, due process may even require such procedures. Compare Londoner v. Denver, 210 U.S. 373, 28 S.Ct. 708, 52 L.Ed. 1103 (1908) (individual public, works, assessments require individual hearings) with Bi-Metallic Investment Co. v. State Board of Equalization, 239 U.S. 441, 36 S.Ct. 141, 60 L.Ed. 372 (1915) (general community-wide tax increases do not). Notice- and-comment procedures, on the other *342hand, are especially suited to determining legislative facts and policy of general, prospective applicability. See, e.g., Air Line Pilots Association v. Quesada, 276 F.2d 892, 895-96 (2d Cir.1960).
The present dispute over the ICC’s formula for allocating platform costs ideally lends itself to notice-and-comment proceedings. No specific facts about individual parties are in issue, and the regulations adopted have general, prospective application industry-wide. In response to this observation, petitioners contend that, even if rulemaking might ordinarily be appropriate to this class of determination, this court’s remands of the earlier cost study investigations effectively limited the agency’s procedural options to continuing in an adjudicatory mode. See Brief of Petitioners at 28-30.
Even assuming that our prior remands have any bearing on the present rulemaking, we are certain that they do not limit the Commission in the manner that petitioners contend. Although our most recent remand firmly cautioned the agency not to attempt evasion of our prior mandates through procedural sleight of hand, we were careful to emphasize that the Commission otherwise remained free to choose among the various procedural options available to it. National Small Shipments, 590 F.2d at 354. It was certainly within the agency’s discretion to abandon the two earlier cost studies that we had held unsupported by record evidence and to undertake instead a third, new study designed to meet this court’s concern for methodological integrity. That the Commission chose to adopt the recommendations of that new study in a general rulemaking rather than through individual investigatory proceedings was a matter entirely within the agency’s informed discretion. We accordingly find no error in the Commission’s decision to rely on notice-and-comment procedures in this instance.
B. Separation of Staff Functions
Petitioners next contend that staff members within the Cost Development Section of the Bureau of Accounts who had primary responsibility for conducting the study improperly participated in the agency’s final decisionmaking process.7 Originally the Commission had precluded Cost Development personnel from playing a decisional role in the rulemaking. ICC Decision, Ex Parte No. MC-129 (Mar. 4, 1980), reprinted in Joint Appendix at 326-27. They were instead allowed to submit a formal reply on the record to aid the Commission in evaluating earlier submitted comments. Because employee attrition and reshuffling later reduced the number of personnel qualified to analyze the rulemaking record, the agency in late 1981 lifted its restrictions on Cost Development staff members, and they thereafter played an active role in helping the Commission reach its final decision in August 1982. ICC Decision, Ex Parte No. MC-129 (Nov. 20, 1981), reprinted in Joint Appendix at 345-46.
Petitioners’ complaint consists of two assertions: first, that due process required the ICC to maintain its separation of staff functions throughout the rulemaking and not merely until late 1981, and second, that in any case some Cost Development personnel violated the ban while it was in effect. See Brief of Petitioners at 15-16, 22, 26-27; Reply Brief of Petitioners at 8-9,11-13. Petitioners’ first assertion need not detain us long. This circuit in prior decisions has considered and rejected the argument advanced by petitioners here. Neither federal statute nor the due process clause imposes a general separation-of-functions requirement on informal rulemaking by an agency. United Steelworkers of America v. Marshall, 647 F.2d 1189, 1210-16 (D.C.Cir.1980), cert. denied, 453 U.S. 913, 101 S.Ct. 3148, 69 L.Ed.2d 997 (1981). The ICC was therefore free to lift its ban when it did, or indeed, to have imposed no ban at all.
*343Petitioners’ second assertion is more problematic. Ordinarily an agency must follow its own procedures whether they are mandated by law or not. Morton v. Ruiz, 415 U.S. 199, 235, 94 S.Ct. 1055, 1074, 39 L.Ed.2d 270 (1974). There are several reasons, however, why this principle does not require reversal of the Commission’s action in this case. First, the evidence that Cost Development personnel actually violated the ban while it was in effect is scanty at best. Compare Brief of Petitioners at 11-12, 22, 26-27, with Joint Brief of the Interstate Commerce Commission and the United States of America at 20-21. Second, none of the distinguishing factors that have moved courts in the past to overturn agency action based on aberrations from nonmandatory procedures are present in this case. More particularly, the procedural ban at issue here was not designed to protect either individual rights or wards of the federal government. See Morton v. Ruiz, 415 U.S. at 235-36, 94 S.Ct. at 1074-1075; Service v. Dulles, 354 U.S. 363, 373, 77 S.Ct. 1152, 1157-1158, 1 L.Ed.2d 1403 (1957); Oglala Sioux Tribe of Indians v. Andrus, 603 F.2d 707, 721 (8th Cir.1979); United States v. Heffner, 420 F.2d 809 (4th Cir.1969). Finally, any error introduced by Cost Development staff activity during the ban was necessarily harmless because such activity could have occurred just as well after the ban was lifted as before. Petitioners have made no showing that the timing of the alleged activity uniquely prejudiced their case before the Commission itself, and they therefore present no grounds for reversing the agency on this score. See 5 U.S.C. § 706 (1982) (rule of prejudicial error). See also Massachusetts Department of Correction v. LEAA, 605 F.2d 21, 27 (1st Cir.1979). We accordingly reject both prongs of petitioners’ separation-of-functions argument.
C. Ex Parte Contacts
During the ICC’s two-year internal staff review of the rulemaking record, the Bureau of Accounts hired a private consultant to evaluate the various methodological criticisms that had been levelled against the platform study up to that point. The consultant, Mr. Daniel Boger, an expert in statistics and econometrics, reanalyzed the record data to assess the impact of any measurement error on the study’s final conclusions. His report to the Bureau, which found the weight-only formula to be “quite robust,” was influential in quelling mounting staff concern about final adoption of the platform study recommendations. See Memorandum from Dan Boger to Ronald Young, Director of ICC Bureau of Accounts (Nov. 9, 1981), reprinted in Joint Appendix at 472-90.
Petitioners contend that these oral and written communications between Mr. Boger and the ICC staff constituted improper ex parte contacts which invalidate the Commission’s entire rulemaking proceeding. In United Steelworkers of America v. Marshall, we held that in the informal rulemaking context private technical consultants may assist the agency in analyzing record data without running afoul of the prohibition on ex parte contacts. 647 F.2d at 1217-20. Because such consultants operate as the functional equivalent of regular staff, they constitute agency insiders. As a result, no improper contact between administrative personnel and outside parties ever arises.8
Because the ICC has statutory authority to retain private consultants,9 we would normally have no cause to press our inquiry further. Petitioners seek to distinguish the fact pattern in this case from that in Steelworkers, however, by pointing out that the ICC never formally approved a consultant *344contract for Mr. Boger’s services. We agree with petitioners that the bureaucratic paperwork necessary for Mr. Boger’s formal employment lie in a regrettable state of disarray. See Reply Brief of Petitioners at 13-16. Nevertheless, the evidence indicates that at the very least an implied-in-fact contract existed between Mr. Boger and the agency for his work on the platform study record data. See Fincke v. United States, 675 F.2d 289 (Ct.Cl.1982); Hagan v. United States, 671 F.2d 1302 (Ct.Cl.1982). This relationship was sufficient to constitute him an agency insider for purposes of the ex parte contacts rule. We therefore find no error in the Commission’s utilization of Mr. Boger to assist in evaluating the rulemaking record in this instance.
D. Suppression of Negative Staff Recommendations
Petitioners’ final procedural complaint alleges that, after several ICC offices and sections had recommended rejecting the platform study, the Bureau of Accounts unfairly seized control of the decisionmaking process to push through a Commission decision adopting the study findings in toto. In so doing, petitioners contend, the Bureau suppressed earlier staff evaluations critical of the study, lest those evaluations reach Commission members and dissuade them from following the Bureau’s lead. See Brief of Petitioners at 28, 30. The evidence shows that four ICC offices or sections at one time or another did express serious reservations in writing about final adoption of the platform study recommendations. See Joint Appendix at 239-62 (Office of Special Counsel), 371-438 (former Investigations and Rulemaking Section of the Bureau of Accounts, and Mathematics and Statistics Section of the former Office of Policy and Analysis), 502-03 (Office of Transportation Analysis). The questions are whether the Bureau of Accounts improperly concealed these reservations from the Commission and, if so, what legal consequences flow from such conduct.
Under existing law, an agency decisionmaking body such as the Commission may delegate detailed consideration of the administrative record to its subordinates while retaining the final power of decision for itself.10 United States v. Morgan, 313 U.S. 409, 421-22, 61 S.Ct. 999, 1004-1005, 85 L.Ed. 1429 (1941). Rather than wade through the entire record personally, then, members of the body are free to rely on summaries prepared by agency staff. Braniff Airways, Inc. v. CAB, 379 F.2d 453, 461 (D.C.Cir.1967). Because of the strong presumption of regularity in administrative proceedings, reviewing courts will not normally entertain procedural challenges that members of the body inadequately considered the issues before reaching a final decision, id. at 460, or that staff reports on which the body relied imperfectly summarized the record under review, cf. Montrose Chemical Corp. v. Train, 491 F.2d 63, 71 (D.C.Cir.1974).11
At some point, however, staff-prepared synopses may so distort the record *345that an agency decisionmaking body can no longer rely on them in meeting its obligations under the law.12 More particularly, in informal rulemaking employing notice-and-comment procedures, dependence on severely skewed staff summaries may breach the decisionmaker’s statutory duty to accord “consideration” to relevant comments submitted for the record by interested parties. See 5 U.S.C. § 553(c) (1982). Certainly, if subordinates systematically eliminated from their reports all mention of record comments adverse to the agency’s final action, the consideration requirement would not be satisfied unless the decisionmakers took independent steps to familiarize themselves with withheld portions of the record.
This analysis suggests that petitioners do have a legal right that their comments reach Commission members in at least summary form, and that those comments be considered before final action is taken.13 Neither the APA nor the due process clause, however, accords similar treatment to staff evaluations that move beyond a mere summary of record comments to express the independent judgments of subordinate agency personnel. An agency is free to structure its internal policy debate in any manner it deems appropriate. Mid-level managers may therefore filter out the evaluations of lower-level personnel if they so choose, so long as relevant record comments are not eliminated in the process as well.
The evidence in this case establishes that Commission members were made fully aware of all comments in the rulemaking record, particularly those criticizing the agency’s proposed action. In September 1981, the Commission published the following pertinent notice in the Federal Register:
Since the receipt of the comments in this proceeding, the Commission has been re-evaluating the [platform] study in light of the issues raised by the comments. While a few of the comments supported the study, the majority criticized both the study and its proposals. Many of the comments questioned the completeness of the data collected, the validity of the statistical analysis used, and the truth of some of the underlying assumptions on which the study recommendations were based.
Thoughtful consideration of the comments received in this proceeding has taken a significant amount of time. We understand the importance of the issues involved to both shippers and motor common carriers and we recognize the need for prompt implementation of a reasonable method of cost allocation. By the same token, the Commission obviously cannot simply ignore the serious questions raised as to the validity of the 1977-1978 platform study. The Commission is handling this proceeding as expeditiously as possible and we expect to issue a final decision in this proceeding in the near future.
46 Fed.Reg. 47,801, 47,802 (1981). This statement leaves no doubt that Commission members were thoroughly familiar with adverse comments in the rulemaking record. Moreover, in its final decision, the Commission addressed and responded to each of the concerns raised in the various opposing comments. See 132 M.C.C. at 855-62.
Even regarding the handling of negative staff recommendations by the Bureau of Accounts, petitioners have mischaracterized the internal contention that surrounded final adoption of the platform study recommendations. The Bureau did not impose its will on other division heads nor stifle their expressions of legitimate concern about the study’s validity. Rather, what appears to have occurred is a lively debate in which other ICC offices seriously questioned the soundness of various aspects of the study methodology, and in which the Bureau re*346sponded with refined statistical analyses to support its conclusions. The Bureau of Accounts ultimately prevailed, not because it unfairly suppressed legitimate internal dissent, but because it convinced other division heads that its weight-only formula was not materially flawed by admitted shortcomings in the study’s design and execution.14 Indeed, petitioners concede as much in their final brief. See Reply Brief of Petitioners at 18-19. Because they have adduced no evidence suggesting bad faith or improper behavior on the part of ICC personnel, petitioners have failed to establish their final procedural challenge. We accordingly find no error in the manner in which agency staff personnel handled the various tasks delegated to them by the full Commission.
Ill
Petitioners also raise two substantive challenges to the Commission’s decision.15 The first is that the platform study incorrectly estimated the total number of shipments in each weight category and, in particular, failed to take proper account of the phenomenon of “cluster” shipments. The second is that the Commission arbitrarily applied the study results to allocate nonstudied as well as studied items of expense. We examine each of these arguments in turn.
A. Shipment Count Estimates and Cluster Shipments
The ICC undertook the present study to measure direct platform handling time in relation to individual shipment characteristics. Rather than collecting data on entire shipments, however, the study recorded only partial data on shipment segments. These segments consisted of individual movements from truck to platform, platform to truck, or truck to truck. Due to the nature of the sampling procedure, it was impossible to aggregate the movements of any one shipment and thus to derive unsegmented data reflecting the shipment’s total processing across the terminal platform.16 To counteract this deficiency, the Bureau of Accounts was forced to estimate the total number of shipments in each weight category and then to divide the total handling time for that category by the estimated shipment count. The Bureau derived this estimate by taking one-half the sum of all truck-to-platform and platform-to-truck movements, and then adding to that number the total number of truck-to-truck movements. Bureau of Accounts, Interstate Commerce Comm’n, 1977-1978 Motor Carrier Platform Study 30-31 (1979) (ICC Statement No. 2S1-79).17
Underlying this calculation is the assumption that every shipment that does not move directly from truck to truck receives two handlings, one off of an arriving truck *347onto the platform and another back off of the platform onto a departing truck. Id. at 31. Numerous comments submitted for the record demonstrated, however, that this assumption does not necessarily hold.18 In particular, petitioners point out that many small shipments receive fewer than two full handlings because of shipping practices known as “assembly” and “distribution.” Assembly shipments are a handling method in which numerous small packages reaching the platform from distinct origins are consolidated into one large shipment for forwarding to a single destination. Distribution shipments involve the reverse handling method in which one large shipment arriving from a single origin is broken down on the platform into numerous small packages for forwarding to disparate destinations. Both types of shipments are known generically as “cluster” shipments.
Because of the prevalence of cluster shipment handling, petitioners contend that the study’s assumption that all platformed shipments receive two handlings has the effect of substantially underestimating the total number of small shipments and thus of overestimating the average handling time in lower weight brackets. See Brief of Petitioners at 39; Joint Appendix at 294-95 (Statement of Dr. Gary Andrew, Statistical Consultant to National Small Shipments Traffic Conference, Inc.). If petitioners are correct, the result would be to allocate inaccurately high costs to small shipments under the study’s final weight-only formula. Consequently, they contend that the Commission’s decision to adopt the platform study recommendations in the face of this alleged defect amounts to arbitrary and capricious agency action.
Virtually all participants in the rulemaking process agreed that the failure to track entire shipments as opposed to mere shipment segments amounted to a serious flaw in the study methodology. See, e.g., Joint Appendix at 374 (Report of Dr. S. West, ICC Statistical Consultant), 448-52 (Memorandum of Dan C. Boger, ICC Statistical Consultant). Indeed, the Commission itself had noted the same shortcoming in its first platform study in 1975,19 but evidently took no steps to prevent a recurrence of the .problem in its later investigations. Nonetheless, the issue before the court is whether the shipment count estimates actually employed by Commission staff were adequate to compensate for this shortcoming or instead, as petitioners contend, introduced unacceptable bias into the final study results.
In its reply statement on the record, the Cost Development Section of the Bureau of Accounts responded to petitioners’ criticism by demonstrating that two other methods exist for estimating shipment counts and that both methods yield the same relative handling times as those found in the study. See Verified Statement of James C. Lundgren, ICC Mathematical Statistician (Mar. 24, 1980). The key observation to be made is that neither of these alternative methods relies on the assumption challenged by petitioners that every platformed shipment receives two handlings.20 If there were any systematic bias either for or against small shipments in any one of the three methods, the bias should show up as a difference in the computed ratios of mean handling times among the methods. In calculations that the Cost Development Section presented for the record, however, no such difference surfaced.
*348This convergence of all three methods around a single set of relative handling times suggests that the original method that the Bureau of Accounts employed to estimate shipping counts produces reliable results and does not run afoul of the phenomenon of cluster shipments. It was, of course, open to petitioners to argue that all three methods are uniformly biased against small shipments, but they have proffered no such argument to this court. They instead limit themselves to quibbling with the Commission’s reliance on relative as opposed to absolute handling times in approving the disputed methodology. Compare 132 M.C.C. at 857 with Reply Brief of Petitioners at 23-24. The Commission might have avoided even this objection had it employed suitable sampling procedures and thereby obviated the need to rely on estimates of total shipment counts. Nonetheless, we find this remaining argument of petitioners to be unavailing. The whole purpose for conducting the platform study was to derive ratios of average handling time among various-size shipments so that platform expenses can be properly allocated. The precise determination of absolute values of shipment counts among the various estimation methods is immaterial to this task, so long as confidence exists in the soundness of the resulting proportional relationships. We are satisfied that the Commission acted rationally in adopting the ratios which the platform study recommended and which were confirmed in the rulemaking record by two alternative modes of estimation unchallenged by petitioners on review.
B. Nonstudied Expenses
The platform study measured only direct cargo handling time on the platform in relation to various shipment characteristics, but did not measure other components of platform expense such as those for administrative and clerical salaries, building rent and depreciation, office equipment, cargo loss and damage insurance, and utilities. These latter, nonstudied expenses account for approximately one-half of all variable platform costs at many terminals.21 See Joint Appendix at 277-79 (Appendix A to Statement of Gerald W. Fauth, Jr., Transportation Cost Consultant to National Small Shipments Traffic Conference, Inc.). Petitioners contend that these expenses have no relation to direct handling time, and that the Commission therefore erred in allocating them on the basis of the study results.
In its decision, the Commission rejected petitioners’ complaint as outside the scope of the present rulemaking because, in its words,
We do not believe that the question of allocating indirect expenses bears directly on the motor carrier platform study, but is a criticism of the present method of distributing expenses in motor carrier cost formulas such as Highway Form A. Highway Form A is designed to distribute service expenses on the basis of special studies primarily based on direct operational functions within these services. Since the subject platform study is a special study, the results should be implemented on the same basis as other service expenses. Changing the methodology for distributing service expenses lies beyond the scope of this proceeding.
132 M.C.C. at 861. Thus, the Commission interpreted petitioners’ objection as questioning not the platform study itself, but the agency’s general policy of allocating indirect support expenses on the basis of direct service inputs, in this case, platform handling time.
We agree with the Commission that reconsideration of its motor carrier cost formulas is irrelevant to the present proceeding. In an earlier rulemaking, the agency specifically refused to adopt these formulas as presumptively more valid than any other reasonable cost finding method. Rules to Govern the Assembling and Presenting of Cost Evidence, 337 I.C.C. 298, 302-03 (1970). The existence of the formulas therefore *349does not bar petitioners from asserting in future agency proceedings that some other method of distributing indirect costs would be more appropriate. Moreover, at oral argument, counsel for the Commission emphasized the agency’s commitment to provide a forum in which petitioners’ concerns would be granted full consideration. Based on that commitment, and on the current legal effect of the motor carrier cost formulas,22 we find no error in the ICC’s refusal to open up the present rulemaking to full exploration of the indirect cost allocation issue. We accordingly affirm the Commission’s disposition of petitioners’ final substantive complaint.
IV
Because a presumption of procedural regularity and substantive rationality attaches to final agency action, aggrieved parties bear the burden of demonstrating to a reviewing court that challenged agency action merits reversal. Although numerous objections of both a procedural and a substantive nature have been raised to the newly promulgated ICC regulations at issue in this case, we find that petitioners have failed to show that the agency has acted either improperly or arbitrarily. We accordingly deny their petition to reverse the Commission decision under challenge in this proceeding.
It is so ordered.
. The ICC adopted the regulations under review in 1977-1978 Platform Study of Class I and Class II Motor Common Carriers of General Freight Subject to Accounting Instruction 27, 132 M.C.C. 851 (1982). The regulations were duly published in 47 Fed.Reg. 36,184 (1982) (to be codified at 49 C.F.R. §§ 1104A.1-.2). Petitioners seek review under 28 U.S.C.A. §§ 2321(a), 2342(5), 2344 (West 1978 & Supp. 1983).
. In industry parlance, truckload shipments are distinguished from Iess-than-truckload and any-quantity shipments.
. The first study had been conducted by the industry itself and relied in part on a “through basis” analysis of costs and revenues. See Small Shipment Rate Revision—Eastern Central Territory, 350 I.C.C. 586, 592-602 (1975), vacated and remanded sub nom. National Small Shipments Traffic Conf., Inc. v. ICC, No. 76-1114 (D.C.Cir. Sept. 23, 1977) (unpublished opinion). The second study, conducted by the ICC’s Bureau of Accounts, was a survey of actual platform handling times based on a nonscientific judgment sample. See id. at 606-18.
. One hundredweight equals 100 pounds.
. These figures may be obtained by dividing the number of man-minutes per shipment in Table 7A by the average number of pounds per shipment, corrected to hundredweight, for each *341weight bracket. See Bureau of Accounts, supra, at 63, 74.
. See supra note 3 and accompanying text. ICC investigatory proceedings are not strictly speaking adjudications, but rather hybrid informal rulemakings. See National Small Shipments Traffic Conf., Inc. v. ICC, 590 F.2d at 350 (citing 49 U.S.C. § 316(g) (1970) (current version at 49 U.S.C. § 10708 (Supp. V 1981))).
. In petitioners’ words, this allowed the Section to play “the triple role of (1) investigator of platform costs, (2) advocate of its own study (by its filed comments) and (3) advisor to decision-maker on the merits of the study.” See Brief of Petitioners at 16.
. For a discussion of how contacts with outside parties might be analyzed once they do occur in an informal rulemaking not involving competing private claims to a valuable privilege, see generally Sierra Club v. Costie, 657 F.2d 298, 400-10 (D.C.Cir.1981).
. See 5 U.S.C. § 3109(b) (1982) (an “agency” may retain private consultants); id § 5721(1)(A) (definition of “agency” includes Executive agencies); id. § 105 (definition of “Executive agency” includes independent establishments); 49 U.S.C. § 10301(a) (Supp. V 1981) (the ICC is an “independent establishment”).
. For purposes of the above analysis, it is immaterial whether the agency decisionmaker is a single individual or a body composed of several individual members.
. In their brief, respondents imply that the Montrose Chemical decision grants agency staff carte blanche to distort and mischaracterize the record in summaries prepared for agency decisionmakers. See Joint Brief of the Interstate Commerce Comm’n and the United States at 17-18. This implication is unjustified. The issue before the Montrose court was whether staff summaries relied on by the EPA Administrator in cancelling a chemical registration were shielded from disclosure under exemption 5 of the Freedom of Information Act as predecisional deliberative materials. See 491 F.2d at 64-67 (construing 5 U.S.C. § 552(b)(5) (1970)). The court held that, as part of the deliberative process, the summaries were exempt, even though they portrayed the record inaccurately and such inaccuraries introduced error into the underlying adjudicatory process. Id. at 71. The court’s earlier discussion of the Morgan cases merely served to discern the limits of the deliberative process, id. at 67-70, and was not meant to question the well-established principle that evidence of bad faith or other improper behavior may justify judicial inquiry into the thoroughness with which agency decisionmakers considered the record under review, Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 420, 91 S.Ct. 814, 825, 28 L.Ed.2d 136 (1971).
. Such distortion could conceivably result from negligence or intentional bias on the part of agency staff.
. This right remains subject, however, to the long-standing rule that courts will not probe the mental processes of administrative decisionmakers absent strong evidence of bad faith or other misconduct. Overton Park, 401 U.S. at 420, 91 S.Ct. at 825.
. See, e.g., Memorandum from William Southard, Director of Office of Transportation Analysis, to Ronald Young, Director of Bureau of Accounts (Feb. 18, 1982) (outlining flaws in platform study and recommending its rejection), reprinted in Joint Appendix at 502-03; id. (Feb. 25, 1982) (retracting earlier negative recommendation based on intervening discussions with Bureau personnel in which it was agreed that flaws were not as severe as originally envisioned), reprinted in Joint Appendix at 504.
. In the proceedings before the Commission, petitioners and intervenors originally raised seven substantive arguments against adopting the platform study. See 132 M.C.C. at 855-62. They have formally briefed only two of these in their petition for review before this court. See Brief of Petitioners at 33 & n.*. We have nonetheless examined their five remaining arguments and are not persuaded that the Commission’s responses to them lack a rational basis. Those arguments therefore present no grounds for reversing the Commission’s decision.
. This was true for all shipment segments except truck-to-truck movements which by definition involve complete platform handlings (i.e., such shipments are moved directly from an arriving truck to a departing truck without an intervening stop on the platform).
. The actual formula was slightly more complicated than described above since it also included a weighting factor for each shipment. See Bureau of Accounts, supra, at 30. Besides the three types of platform movements mentioned above, the sample also included a number of platform-to-platform movements, but these are irrelevant for purposes of estimating shipping counts.
. See, e.g., Joint Appendix at 254-57 (Supplemental Comments of ICC Office of Special Counsel), 259-60 (Supplemental Verified Statement of Robert Hasek, Transportation Analyst, ICC Bureau of Accounts), 290-96 (Statement of Dr. Gary Andrew, Statistical Consultant to National Small Shipments Traffic Conference, Inc.), 317-20 (Verified Statement of A.N. Glick, American Home Products Corp.).
. Small Shipment Rate Revision, 350 I.C.C. at 608 (“Ideally, the [first platform] study would have followed each shipment beginning with its unloading from a line-haul or pickup and delivery until it was stowed in another vehicle, but this was not possible.”).
.The first method assumes instead that each such shipment receives only one handling (from truck to platform), while the second method also assumes a single handling but in the opposite direction (from platform to truck). See Verified Statement of James C. Lundgren, supra, at 3-5.
. The costs at issue here are indirect variable costs. See Bureau of Accounts, Interstate Commerce Comm’n, 1977-1978 Motor Carrier Platform Study 61 (1979) (ICC Statement No. 2S1-79). They are not constant costs, as petitioners suggest in one portion of their reply brief. See Reply Brief of Petitioners at 22 n.*.
. Under the 1970 ICC decision cited above, the motor carrier cost formulas are best characterized as general statements of policy because they repose in agency decisionmakers a discretionary power to accept alternative cost finding methodologies. See 337 I.C.C. at 300 (“[P]arties [are still] free to employ other methods of estimating costs.”); id. at 303 (“[Preferred status for costs developed under the specific formulas . .. might discourage further independent research and development in the area of transportation costing.”). See generally Guardian Fed. Sav. & Loan Ass’n v. FSLIC, 589 F.2d 658, 666-68 (D.C.Cir.1978) (finding general statement of policy where discretionary power is retained to vary from recommended standards).