Bethlehem Steel Corporation, et al. (Bethlehem), Thyssen Stahl AG, et al. (Thyssen), and Hoogovens Groep BV (Hoogovens), respectively, appeal from two decisions of the Court of International Trade, United States Steel Group v. United States, 873 F.Supp. 673 (Ct. Int’l Trade 1994); Kern-Liebers USA, Inc. v. United States, 17 I.T.R.D. (BNA) 1082, 1995 WL 33066 (Ct. Int’l Trade 1995), which affirmed the determinations of the International Trade Commission (Commission) that an industry in the United States (1) is not materially injured nor threatened with material injury by imports of hot- and cold-rolled steel from numerous countries; (2) is threatened with material injury by imports of cold-rolled steel from Germany; and (3) is threatened with material injury by imports of cold-rolled steel from the Netherlands. We affirm all aspects of the appealed decisions.
I
On June 30,1992, a group of United States steel companies filed a petition with the International Trade Commission alleging that their industry had been harmed by subsidized and “Less Than Fair Value” (LTFV)1 imports of certain flat-rolled carbon steel products from numerous countries,2 and seeking the imposition of countervailing and *1356antidumping duties against the subject imports. 57 Fed.Reg. 60,247 (1992); 58 Fed. Reg. 8974 (1993). After affirmative preliminary findings by the Department of Commerce on December 7, 1992 and February 4, 1993 that the imports in question were indeed subsidized or being sold at LTFV, the Commission commenced an investigation to determine whether the imports had caused, or threatened to cause, a material injury to an industry in the United States. See id. In conducting its investigation, the Commission divided the broad category of flat-rolled carbon steel products into four “like products:”3 (1) hot-rolled products;4 (2) cold-rolled products; (3) corrosion-resistant products;5 and (4) cut-to-length plate products. See 1 ITC Final Determination, supra, at 7, 163. The appeals in this case do not question the Commission’s categorization of products, but instead concern only the Commission’s determinations with respect to the hot-rolled and cold-rolled categories.
The Commission conducted its investigation over a three-year period, 1990-1992. During that period, the Commission collected large amounts of data concerning the subject imports including detailed information regarding the imports’ value, prices, and volume of shipments. In addition, it collected extensive data on the production, capacity utilization, and inventory levels of foreign producers, and carefully examined the patterns of domestic consumption of the subject imports. On the basis of this data, the Commission made its determinations, the majority of which are not contested on this appeal. Of those that are, the Commission determined that most of the subject hot- and cold-rolled imports had not caused, and did not threaten to cause, material injury to a domestic industry. In addition, the Commission determined that cold-rolled imports from Germany and Holland present a threat of material injury to the domestic cold-rolled industry. See 1 ITC Final Determination, supra, at 1-5.
Bethlehem et al, a group of domestic steel producers, Thyssen et al, a group of German steel producers, and Hoogovens, a Dutch steel producer, each appealed to the Court of International Trade from those aspects of the Commission’s determinations adverse to them. Bethlehem challenged the Commission’s many negative material injury determinations, while Thyssen and Hoogovens each challenged the affirmative threat findings against their home countries. In two decisions, the Court of International Trade affirmed all of the Commission’s determinations. Bethlehem, Thyssen, and Hoogovens now seek review of those decisions in this court. We have jurisdiction to review the Court of International Trade’s decisions under 28 U.S.C. § 1295(a)(5) (1994).
II
Like the Court of International Trade, we review the Commission’s findings of fact for substantial evidence and its conclusions of law de novo. Atlantic Sugar, Ltd. v. United States, 744 F.2d 1556, 1559 n. 10, 2 Fed. Cir. (T) 130, 133 n. 10 (1984); 19 U.S.C. § 1516a(b)(l)(B) (1994). Substantial evidence is “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Universal Camera Corp. v. NLRB, 340 U.S. 474, 477, 71 S.Ct. 456, 459, 95 L.Ed. 456 (1951) (quoting Consolidat*1357ed Edison Co. v. NLRB, 305 U.S. 197, 229, 59 S.Ct. 206, 216-17, 83 L.Ed. 126 (1938)). Thus, the question in this case is not whether we agree with the Commission’s decision, nor whether we would have reached the same result as the Commission had the matter come before us for decision in the first instance. By statute, Congress has allocated to the Commission the task of making these complex determinations. Ours is only to review those decisions for reasonableness.
Ill
We begin with Bethlehem’s appeal which challenges the Commission’s negative material injury determinations on three grounds. First, Bethlehem contends that conclusive evidence refutes two important subsidiary findings upon which the Commission’s determinations were, in large part, grounded. Second, it contends that the Commission improperly excluded the imports of several countries from its material injury analysis. Finally, Bethlehem contends that two of the commissioners applied an incorrect standard in determining whether the subject imports had caused a material injury. We address each argument in turn.
A
In its first assertion of error, Bethlehem argues that the Commission’s determination is premised on two subsidiary findings which Bethlehem characterizes as demonstrably wrong. Specifically, Bethlehem contests the Commission’s findings that domestic steel products are not greatly substitutable with those produced overseas, and that steel purchasers are not very price sensitive. Although Bethlehem concedes that the record developed during the investigation period contains substantial evidence which supports these findings, it argues that one piece of evidence from after the close of the investigation period ineontrovertibly refutes them.6
On December 7, 1992 and February 4, 1993, after determining that the subject imports were being subsidized or sold at LTFV, the Department of Commerce imposed preliminary tariffs pending conclusion of the Commission’s investigation and the rendering of its final material injury determination.7 After imposition of those tariffs (which effectively raised the price of imported steel), open market “spot” sales of imported cold-rolled steel declined dramatically even though demand for cold-rolled steel was rising. In addition, domestic producers were able, for the first time in several years, to increase their prices. According to Bethlehem, this proves absolutely that sales of steel products are price sensitive and that domestic purchasers will substitute domestic steel for foreign steel as the price of foreign steel rises. Bethlehem made the same argument before the Commission. The Commission, however, declined to rely heavily on that evidence primarily because it came from outside of the investigation period, and therefore could not be assessed against a comprehensive economic context. See 1 ITC Final Determination, supra, at 129 & n. 343.
It is the Commission’s task to evaluate the evidence it collects during its investigation. Certain decisions, such as the weight to be assigned a particular piece of evidence, lie at the core of that evaluative process. In this case, the Commission was well within its discretion when it discounted the probative value of Bethlehem’s evidence. During an investigation period, the Commission collects extensive economic data from which it develops a thorough understanding of extremely *1358intricate economic interactions. This thorough understanding permits the Commission to evaluate each piece of evidence in context and to reach well-supported determinations which take account of as many, aspects of a complicated economic reality as possible.
In contrast, fragmentary evidence from outside the investigation period may be difficult to interpret and is susceptible of myriad explanations. This case is no exception. Admittedly, it is possible that the rise in domestic prices and concurrent decline in imports were caused by higher import prices resulting from the preliminary tariffs. But to claim that the temporal link between these events proves that they are causally related is simply to repeat the ancient fallacy: post hoc ergo propter hoc.
Nor is there a shortage of alternative explanations for the events to which Bethlehem points. As the Commission itself noted, it is likely that the rise in prices was caused, at least to some extent, by increasing demand in a recovering steel market. Similarly, it is possible that foreign imports coincidentally declined during this period because of shifts in demand from the specialty cold-rolled products produced by foreign producers to the more commodity-like cold-rolled products dominated by domestic producers. In the absence of complete economic data for this period, especially regarding the patterns of domestic consumption, there is simply no way to know the extent to which the market effects noted by Bethlehem were caused by the imposition of preliminary tariffs. Consequently, the Commission was justified in refusing to abandon its well-supported determination on the basis of the evidence advanced by Bethlehem. We therefore decline to reverse the Commission’s determination on this ground.
B
Before turning to Bethlehem’s second assertion of error, we pause to review the concepts of cumulation and negligibility as in effect at the time of the Commission’s determination. 19 U.S.C. § 1677(7)(C)(iv)(I) (Supp. V1993) provides:
For purposes of [determining whether the subject imports have caused a material injury] the Commission shall cumulatively assess the volume and effect of imports from two or more countries of like products subject to investigation if such imports compete with each other and with like products of the domestic industry in the United States market.
As a general rule, then, in evaluating the effect of the subject imports on the domestic industry, the Commission must cumulate the subject imports from all countries, rather than separately assess the effect of each country’s imports. 19 U.S.C. § 1677(7)(C)(v) (1988), however, affords an exception to that rule. Under that section, cumulation is not required if the Commission determines that:
[Ijmports of the merchandise subject to investigation are negligible and have no discemable adverse impact on the domestic industry. For purposes of making such determination, the Commission shall evaluate all relevant economic factors regarding the imports, including, but not limited to, whether—
(I) the volume and market share of the imports are negligible,
(II) sales transactions involving the imports are isolated and sporadic, and
(III) the domestic market for the like product is price sensitive by reason of the nature of the product, so that a small quantity of imports can result in price suppression or depression.
Thus, section 1677(7)(C)(v) provides a negligibility exception to the cumulation rule of section 1677(7)(C)(iv).
In its second assertion of error, Bethlehem asserts that the Commission misapplied the negligibility exception and consequently improperly failed to cumulate the imports of certain countries.8 Focussing on only one portion of the statute, Bethlehem argues that the Commission may only exclude a country from cumulation if its im*1359ports have no “discemable adverse impact.” Here, Bethlehem contends, imports from several excluded countries did have a discerna-ble adverse impact since the Commission was able to confirm, in several specific instances, that as a result of these imports various domestic steel producers suffered confirmed lost sales. Furthermore, Bethlehem adds, the Commission’s own data indicates that the imports from each of these countries resulted in decreased revenues for domestic producers of as much as $25 million.
Bethlehem, however, misinterprets the statute. The statute requires the cumulation of imports which have a “discernable adverse impact on the domestic industry,” 19 U.S.C. § 1677(7)(C)(v) (1988) (emphasis added), not of imports which may adversely affect individual domestic producers. Thus, evidence of specific lost sales or declines in revenue, although important considerations to be taken into account by the Commission in making its negligibility determinations, do not dis-positively trigger the cumulation requirement.
Other aspects of the statute resist Bethlehem’s reading. For example, in making a negligibility finding, the Commission is instructed to consider, among other things, whether the volume and market share of the imports are negligible and whether sales transactions involving the imports are isolated and sporadic. 19 U.S.C. § 1677(7)(C)(v)(I,II) (1988). Thus, the statute contemplates that a country may import small amounts of a product to the United States and still be deemed negligible, although those imports will presumably have some adverse impact on the particular domestic producers who would otherwise supply the demand they fill. Therefore, we disagree with Bethlehem that the Commission improperly excluded several countries from its cumulation and decline to reverse the Commission on this ground.9
C
Bethlehem’s third assertion of error questions the manner in which the Commission, and in this case, Commissioners Brunsdale and Crawford in particular, determine whether an industry in the United States is materially injured or threatened with material injury by reason of subsidized and/or LTFV imports. A review of the relevant statutes puts Bethlehem’s argument in context.
Countervailing and/or antidumping duties may not be imposed on merchandise found to be subsidized and/or dumped unless:
the Commission determines that—
(A) an industry in the United States—
(i) is materially injured, or
(ii) is threatened with material injury, ... by reason of imports of that merchandise.
19 U.S.C. § 1671(a)(2) (1988) (for subsidized merchandise); 19 U.S.C. § 1673(2) (1988) (for LTFV merchandise).
*1360With respect to both subsidized and LTFV imports, “material injury” is defined as “harm which is not inconsequential, immaterial, or unimportant.” 19 U.S.C. § 1677(7)(A) (1988).
The Commission makes its determinations by tallying the votes of the six individual commissioners, each of whom is obligated to determine whether particular imports cause or threaten to cause the requisite harm. Congress has not left the commissioners at sea in the performance of their individual, and ultimately collective, duties. Instead, Congress has supplied the analytical tools, in the form of statutory tests, some of which must be applied and some of which may be applied before the Commission, acting through its commissioners, arrives at its measurement of the causal effects of particular imports. The statutory tests are many, and specific; in each case, the Commission:
(i) shall consider—
(I) the volume of imports of the merchandise which is the subject of the investigation,
(II) the effect of imports of that merchandise on prices in the United State for like products, and
(III) the impact of imports of such merchandise on domestic producers of like products, but only in the context of production operations within the United States; and
(ii) may consider such other economic factors as are relevant to the determination regarding whether there is material injury by reason of imports.
19 U.S.C. § 1677(7)(B) (1988).
These three general tests are further refined explicitly in the statute. With regard to volume of imports, “the Commission shall consider whether the volume of imports of the merchandise, or any increase in that volume, either in absolute terms or relative to production of consumption in the United States, is significant.” 19 U.S.C. § 1677(7)(C)(i) (1988). Concerning the effect of imports on price, the Commission must consider whether:
(I) there has been significant price underselling by the imported merchandise as compared with the price of like products of the United States, and
(II) the effects of imports of such merchandise otherwise depresses prices to a significant degree or prevents price increases, which otherwise would have occurred, to a significant degree.
19 U.S.C. § 1677(7X000 (1988).
When considering the impact of LTFV or subsidized imports on the state of the domestic industry, the specific commands made on the Commission by Congress are even more detailed:
In examining the impact required to be considered under subparagraph (B)(iii),10 the Commission shall evaluate all relevant economic factors which have a bearing on the state of the industry in the United States, including, but not limited to—
(I) actual and potential decline in output, sales, market share, profits, productivity, return on investments, and utilization of capacity,
(II) factors affecting domestic prices,
(III) actual and potential negative effects on cash flow, inventories, employment, wages, growth, ability to raise capital, and investment, and
(IV) actual and potential negative effects on the existing development and production efforts of the domestic industry, including efforts to develop a derivative or more advanced version of the like product.
The Commission shall evaluate all relevant economic factors described in this clause within the context of the business cycle and conditions of competition that are distinctive to the affected industry.
19 U.S.C. § 1677(7)(C)(iii) (1988).
Furthermore, as we have noted above, each commissioner is free to determine whether or not to exclude from the material injury analysis imports from countries deemed negligible. We note this additional factor simply to underscore the complicated network of statutory tests that must or may *1361be employed by each commissioner in deciding whether the causal effects of particular imports support imposition of countervailing or antidumping duties.
Over the course of years, differing commissioners have employed differing methodologies to reach their conclusions on the extent to which LTFV and/or subsidized imports have harmed, or threatened to harm, domestic industries. For purposes of this case, the differing methodologies are described as the “one step” and the “two step” analyses. Under the one-step analysis, a commissioner assesses in a unitary process both the current state of the domestic industry and whether that state is materially injured by reason of LTFV or subsidized imports. Under the two-step analysis, a commissioner first assesses the state of the relevant domestic industry. If the assessment produces a conclusion that the industry is materially injured, then the analysis proceeds to its second step, which is the separate inquiry asking whether the pertinent imports contribute in a non-de minimis way to such material injury. Over time, some commissioners have opted for one of these methodologies, and others have preferred the second analytical tool. And in some instances, as with Commissioner Watson in this ease, the commissioner simply recites the statutory language, without specifying which of the one- or two-step analyses, or some other analytical construct, has been used to assist the commissioner in fulfilling the statutory requirements.
In this case, Commissioners Brunsdale and Crawford used the one-step analysis in reaching their determinations on material injury. According to Bethlehem, their use of the one-step analysis is forbidden by the relevant statutes, and the two-step method used by the other commissioners in this case must be determined, by this court, to be the only statutorily permitted tool of analysis for rendering material injury determinations. Bethlehem begins with the premise that the statutes require a determination of material injury whenever imports contribute in a non-de minimis way to further injury to a domestic industry. Since some commissioners have characterized the effect of the two-step analysis in this manner, Bethlehem urges us to read the two-step analysis into the statute as the only permissible way to assess material injury. Bethlehem argues that the one-step method of analysis violates the statute because it necessarily requires a greater quantum of impact from imports on a domestic industry, in order to find material injury or threatened material injury, than would use of the two-step test. Bethlehem argues that the outcome of this case could change if we were to impose the two-step test on Commissioners Brunsdale and Crawford since it is possible that under the two-step test, they might assess the material injury aspect of the imports differently. Given the votes of other commissioners in this case, additional votes of material injury could tip the scale in Bethlehem’s favor.
Bethlehem advanced the same argument in the Court of International Trade, seeking to characterize the difference of modes of analysis of various commissioners as fundamentally different statutory interpretations. The Court of International Trade rejected that characterization, and instead held that the two methods of analysis are simply that, and neither fails to comport with the statutory commands setting forth the manner in which the Commission is to dice a record to reach a material injury determination. Judge Resta-ni, for the Court of International Trade, correctly noted that the two-step method has the virtue of some decisionmaking efficiencies, but that the statute by its terms does not mandate it as the only method of analysis. She also, rightly, noted that “[t]he statutory language fits very well with a one-step mode of analysis.” United States Steel Group, 873 F.Supp. at 695. This is so, in part at least, because of the awkwardness of reading the statute to say that material injury must be determined in every case by asking, as Judge Restani put it, “if such imports were a non-de minimis contributing cause of the state of the ‘materially injured industry.’ ” Id. at 695.
At bottom, Bethlehem seeks a ruling from this court that there should be a single methodology, applicable to each of the commissioners, for determining whether a domestic industry is injured, or threatened with inju*1362ry, by reason of subsidized and/or LTFV imports. The statute on its face compels no such uniform methodology, and we are not persuaded that we should create one, even were we so empowered.
Congress has crafted an intricate statute, and committed its enforcement to the Department of Commerce and the International Trade Commission. Congress has populated the Commission with six independent commissioners, each confirmed to office by the United States Senate. As Bethlehem candidly and commendably notes in its brief, commissioners are free to attach different weight to the various statutory tests which they are required to employ when evaluating the presence or threat of injury. Also, Bethlehem notes and does not challenge the indisputable proposition that each commissioner is free to attach different weight to factual information bearing on, and determinate of, the many statutory tests; and that commissioners may ultimately reach different factual conclusions on the same record. In the end, of course, the factual conclusions of each commissioner will drive the legal conclusion he or she reaches, namely, whether the requisite injury has been shown. The invitation to employ such diversity in methodologies is inherent in the statutes themselves, given the variety of the considerations to be undertaken and the lack of any Congressionally mandated procedure or methodology for assessment of the statutory tests.
This court has no independent authority to tell the Commission how to do its job. We can only direct the Commission to follow the dictates of its statutory mandate. So long as the Commission’s analysis does not violate any statute and is not otherwise arbitrary and capricious, the Commission may perform its duties in the way it believes most suitable. Because we do not believe that the statute compels the commissioners to employ either the one-step or two-step approaches, we disagree with Bethlehem that Commissioners Brunsdale and Crawford employed a causation standard which was contrary to law, and therefore decline to reverse the Commission on this ground.
Having considered and rejected each of Bethlehem’s three assertions of error, we affirm the decision of the Court of International Trade with respect to all aspects of Bethlehem’s appeal.
IV
We next address Thyssen’s appeal. By a vote of four to two, the Commission determined that German cold-rolled imports threaten a domestic industry with material injury. 1 ITC Final Determination, swpra, at 1-5, 131. Two of the four commissioners in the majority, Commissioners Watson and Rohr, wrote jointly to explain the basis of their decision, while the other two, Commissioners Newquist and Nuzum, each wrote separately. Id. at 133, 300, 351. Thyssen attacks the decisions of each of these Commissioners as being unsupported by substantial evidence.
Thyssen’s strongest arguments focus on two of the subsidiary findings which formed, in part, the basis of Commissioners Watson and Rohr’s ultimate decision. First, Thyssen attacks a finding by those Commissioners that German producers might shift production from corrosion-resistant steel to cold-rolled steel in order to avoid United States tariffs on German corrosion-resistant imports. Second, Thyssen attacks a finding by those Commissioners that German import prices declined significantly over the investigation period.
A
In another of its determinations,11 the Commission concluded that German imports of corrosion-resistant steel had caused a material injury to the domestic corrosion-resistant steel industry, and that tariffs should therefore be imposed on those products. Id. at 1-5. In light of this, Commissioners Watson and Rohr found that German producers might shift cold-rolled feedstock from the *1363production of corrosion-resistant products to cold-rolled sales in an effort to avoid the corrosion-resistant tariff. This subsidiary-finding was grounded, in turn, on several underlying factual assertions. Specifically, the Commissioners noted that:
The four German producers of cold-rolled products also produce corrosion-resistant products. The increase in German 'production of corrosion-resistant products between 1990 to 1992 was accounted for almost exclusively by increases in sales in the U.S. market. In addition, German exports of corrosion-resistant steel are projected to decline in 1993.... [T]he facts outlined above suggest that there is the possibility that some captive cold-rolled products used in corrosion-resistant steel would be diverted for sale as cold-rolled products [in the United States].
Id. at 134-35 (emphasis added) (footnotes omitted).
The Commission now concedes that most of the additional corrosion-resistant steel produced in Germany between 1990 and 1992 was not sold in the United States. As a result, Thyssen argues, Commissioners Watson and Rohr’s subsidiary finding that German producers might divert production from corrosion-resistant to cold-rolled products, which was expressly grounded on the facts outlined above, is not supported by substantial evidence.
Although we consider the issue close, we agree with Thyssen that this subsidiary finding was not supported by substantial evidence. To advance their argument from premise (i.e., that German corrosion-resistant steel was subject to United States duty) to conclusion (i.e., that German producers would produce more cold-rolled steel at the expense of corrosion-resistant steel), the Commissioners needed, as an analytical matter, to assert that some significant amount of German corrosion-resistant steel was intended for sale in the United States, and would thus be affected by the tariff. In large part, the Commissioners filled that analytical gap with the assertion, now conceded to be inaccurate, that “[ t]he increase in German production of corrosion-resistant products between 1990 to 1992 was accounted for almost exclusively by increases in sales in the U.S. market.” Id. at 134. Other than that assertion, Commissioners Watson and Rohr relied on no significant evidence that German corrosion-resistant steel would be affected by the United States tariffs. We therefore agree with Thyssen that this subsidiary finding is not supported by substantial evidence. Consequently, we exclude this subsidiary finding from our substantial evidence calculus, and examine the remainder of Commissioners Watson and Rohr’s opinion to determine whether substantial evidence supports their ultimate affirmative threat determination.
B
In its second persuasive argument, Thys-sen contends that Commissioners Watson and Rohr improperly inferred from steadily declining German Average Unit Values (AUVs) that there was a “probability that imports of cold-rolled steel from Germany will enter the United States in the immediate future at prices that will have a suppressing or depressing effect on U.S. prices.” Id. According to Thyssen, however, the apparent decline in German AUVs was not due to falling German prices, but was rather due to a shift in demand from expensive German steel products to inexpensive German steel products. As Thyssen convincingly demonstrates, when these shifts in demand are taken into account, the true decrease in German unit values was a mere 1.8% rather than the 14% which emerges from the Commission’s calculations.12
Computing an average is arguably the most basic of all statistical techniques. It permits compression of large quantities of data into a single representative figure capable of easy comprehension and assimilation. In that respect, it is undoubtedly a valuable tool. Averaging, however, is not without its dangers. Average values speak only to the attributes of the composite figures in the aggregate. They say nothing about the corn-*1364posite figures individually. Therefore, to properly identify whether a particular composite factor is the primary influence driving the change in an average value over time, it is, at times, necessary to disaggregate the composite data, and examine each portion of the data discretely.
Like any other average, AUVs have the advantages and dangers described above. AUVs are computed by multiplying, for each product, the price of the product times the quantity sold, summing these results, and then dividing the total by the total number of products sold. Thus, falling AUVs may be caused either by declining prices or by a redistribution of sales from expensive to inexpensive products. The Commission’s reb-anee on AUVs as an indication of falling prices, therefore, implicitly assumes that the distribution of product sales remains constant.
We recognize that the Commission must assimilate and interpret large quantities of data in making these determinations. Therefore, we do not hold, as a general rule, that the Commission may not rely on AUV trends as indicative of corresponding changes in price. In fact, we believe that the Commission’s implicit “constant product distribution” presumption is wholly appropriate in most instances. Like most presumptions, however, this one is rebuttable. When a party can show that the AUVs reflect shifts in demand rather than pricing trends, the probative value of the AUVs is severely undermined. See Chr. Bjelland Seafoods A/S v. United States, 17 I.T.R.D. (BNA) 1020, 1024, 1995 WL 25327 (Ct. Int’l Trade 1995).
This is such a case. Here, Thyssen showed, using rehable data, that the declining AUVs relied on by the Commission were
1990
Column 1 2 3 4
Tariff Quantity Quantity Value AUV
Category
790921 2 786 381 1
720922 132 50825 385 106
790923 142 58127 410 110
Grp 37 25 37808 1523 H UJ
Other 22 12063 554 16
Total 322 159609 495 247
not, for the most part, due to a decrease in the price of German steel. Therefore, in considering these AUVs, Commissioners Watson and Rohr should have discounted their probative value and should not have relied on them significantly in reaching an affirmative threat determination. For that reason, in reviewing Commissioners Watson and Rohr’s decision for substantial evidence, we do not give weight to the portion of their analysis which relied on the decline in German import AUVs.
We note, however, that in addition to falling AUVs, Commissioners Watson and Rohr relied on several other pieces of evidence to support their finding that German cold-rolled imports threatened to suppress prices in the United States. Specifically, the Commissioners also relied on the fact that German products undersold domestic steel products 71 times while overselling domestic products in only 64 instances, and that German cold-rolled products were concentrated, on the whole, in the types of fungible commodity products more likely to compete with domestic production. See Kern-Liebers USA, Inc. v. United States, 17 I.T.R.D. (BNA) 1082, 1105-06, 1995 WL 33066 (Ct. Int’l Trade 1995). Taken as a whole, therefore, it is likely that there is substantial evidence to support Commissioners Watson and Rohr’s subsidiary finding that German cold-rolled imports threaten to suppress domestic prices.
Even if the Commissioners’ subsidiary price-suppression finding was not supported by substantial evidence, however, we find
1991 1992
6 7 8 9 5
Value AUV Quantity Value AUV
431 361 ^ IQ 20097 372
39089 367 139 49586 356
42807 389 106 38547 362
23208 1741 13 23947 1908
7916 477 24 11212 471
113451 458 336 143389 427
*1365that the other evidence relied on by Commissioners Watson and Rohr, taken as a whole, was sufficient to support their ultimate conclusion. Specifically, Commissioners Watson and Rohr found that Germany has increasing inventories and much unused capacity, as well as decreasing markets at home and abroad to sell its steel. 1 ITC Final Determination, supra, at 133-84. In addition, the Commissioners found that market penetration of German imports increased over the period of investigation, albeit not greatly. Id. In view of these facts, it was reasonable for the Commissioners to conclude that unless the other markets for German cold-rolled steel recovered significantly, an uncertain possibility, German producers would turn to the United States to sell their excess steel. This very real prospect supports Commissioners Watson and Rohr’s finding that German imports of cold-rolled steel present a threat of material injury to the domestic cold-rolled industry.
We have carefully considered Thyssen’s other attacks on the evidence relied upon by Commissioners Watson and Rohr, as well as Thyssen’s arguments with respect to the decisions of Commissioners Newquist and Nu-zum, but find them insufficient on the facts of this ease to weaken our conclusion that substantial evidence supports the determinations of each of these Commissioners.13 Consequently, we affirm the Commission’s affirmative determination that imports of German cold-rolled steel products threaten the United States cold-rolled industry with material injury.
V
We turn finally to Hoogovens’ appeal. By a vote of three to three, the Commission determined that cold-rolled imports from the Netherlands present a threat of material injury. Id. at 3-4.14 Like Thyssen, Hoogo-vens argues that the determinations of the three commissioners who voted in the affirmative, Commissioners Newquist, Rohr and Nuzum, were not supported by substantial evidence. Hoogovens directs its arguments at two pieces of evidence in particular: sample pricing data tending to show underselling by Dutch imports, and non-weighted AUVs tending to show declining import prices.
A
The pricing data relied on by the Commission to show underselling of Dutch imports consisted of sample data which represented less than 4% of the total quantity of Dutch cold-rolled steel imported to the United States. At oral argument, Hoogovens abandoned the argument in its brief that a 4% sampling rate is per se inadequate. Instead, Hoogovens argues that on the facts of this case the small quantity of data collected by the Commission does not fairly represent the price of imported Dutch cold-rolled steel generally. Had the Commission examined a wider sample of Dutch steel, Hoogovens argues, the Commission would have recognized that most cold-rolled steel imported from Holland consists of high-quality specialty products for which United States consumers pay a price premium.
In their briefs to this court, Hoogovens and the Commission vigorously dispute whose fault it is that the Commission collected the allegedly unrepresentative data. In particular, the Commission points out the *1366difficulties it faces in conducting an investigation and the need for cooperation from all parties in collecting sufficient and accurate data. According to the Commission, Hoogo-vens raised its concerns with the sampled data too late in the proceedings for the Commission to take corrective action, even assuming such action was appropriate. In response, Hoogovens argues that the burden of conducting an adequate investigation lies squarely with the Commission, and that the Commission cannot avoid responsibility for any deficiencies in its investigation by blaming Hoogovens for its failure to point out inadequacies in the Commission’s data.
We need not decide, however, whether Hoogovens or the Commission has the better of this debate because Hoogovens has failed to make a threshold showing that the sample data relied on by the Commission was, in fact, not representative. Sampling, like averaging, is a necessary statistical technique without which the Commission’s task would be virtually impossible. Of course, sampling has its own dangers including the danger that the sample data may not be representative of the data as a whole. We are certain, however, that the Commission recognizes this danger and takes appropriate steps to avoid it when necessary. Consequently, Hoogovens has the burden to show that the sample relied on by the Commission was not representative. General allegations that the cold-rolled steel market is not homogeneous and that small samples consequently yield skewed results are insufficient to meet this burden.
Hoogovens points us to no quantitative evidence to indicate that the sampled data relied on by the Commission was not representative.15 Instead, it presents anecdotal evidence in the form of affidavits from two of its United States customers which state that they pay a premium for Dutch cold-rolled products because of their high quality. As the Court of International Trade noted, however, anecdotal evidence of this type is simply insufficient to show that the data relied on by the Commission was not representative. See Kern-Liebers USA, Inc. v. United States, 17 I.T.R.D. (BNA) 1082, 1102, 1995 WL 33066 (Ct. Int’l Trade 1995). Although the affidavits make passing references to quantities of steel purchased and the size of premiums paid, they do not provide the type of specific quantitative information necessary to weigh the affidavits against the other evidence of record in order to determine whether the sample data was not representative. Consequently, we reject Hoogovens’ first argument.
B
Hoogovens’s second argument, like Thys-sen’s, focuses on what it considers to be the tenuous link between AUVs and prices. According to Hoogovens, because the cold-rolled industry comprises so many different products, it is impossible for trends in aggregate unit values to accurately reflect pricing trends. As a result, Hoogovens argues, Commissioner Newquist’s determination, which relied in part on declining import AUVs as evidence that “domestic prices must [continue to decline] to compete with the unfair imports,” is not supported by substantial evidence. 1 ITC Final Determination, supra, at 299.
Unlike Thyssen, however, Hoogovens fails to provide specific data that this decline in average unit values is not representative of a corresponding decline in import prices. As explained above in part IV, the Commission is generally entitled to rely on falling AUVs as evidence of a corresponding decline in prices, when it feels that such an inference is appropriate. Consequently, the burden here is on Hoogovens to show, by hard evidence, that in this case falling AUVs are not representative of falling prices, but rather are due *1367to some other factor, such as a redistribution in domestic consumption. Hoogovens general allegations that the breadth of products encompassed within the cold-rolled universe precludes reliance on AUVs are not sufficient to meet this burden. We therefore decline to reverse Commissioner Newquist’s determination on this ground.
Finally, we note that Hoogovens does not address any of the other evidence relied on by the Commission in reaching its determination. Specifically, in addition to the pricing data and AUVs challenged by Hoogovens, Commissioner Newquist’s determination relied on other substantial evidence including, in part, his findings that the domestic cold-rolled industry was in a vulnerable condition; that imports of cold-rolled steel from the countries he had cumulated16 increased during the investigation period; and that domestic production decreased during the investigation period. Id. at 300-02. Similarly, Commissioner Rohr’s determination was grounded on his findings that Dutch utilization levels remained at relatively low levels; that home market shipments were low; that third country shipments had declined steadily over the investigation period; as well as on other significant evidence. Finally, Commissioner Nuzum’s determination was grounded, in part, on her findings that imports from the countries she had cumulated17 had risen substantially from 1991-1992 and that producers in those countries had substantial unused capacity. Id. at 387-88. Thus, even were we to agree with Hoogovens’ arguments with respect to the sample pricing data and AUVs, it is less than certain that we would, on the record as a whole, find that the Commission’s determination was not supported by substantial evidence.
VI
Having concluded our analysis with respect to each of the appeals before us, we close by again stressing our limited role in these determinations. We are here primarily to decide whether, on the basis of the record before us, reasonable decisionmakers could have concluded, on the record in this case, that German and Dutch cold-rolled imports threaten to cause material injury while hot- and cold-rolled imports from other countries have not caused, and do not threaten to cause, such injury. Because we agree with the Court of International Trade that the Commission could so conclude, the decision of that court is affirmed in all respects.
AFFIRMED.
. Two classes of imports are at issue in this case. Subsidized imports are imported products whose manufacture or exportation is subsidized by a government, corporation or citizen of particular foreign countries. 19 U.S.C. § 1671(a)(1) (1988). Duties imposed to offset such subsidies are called ''countervailing” duties. In contrast, "Less than Fair Value” (LTFV) imports are imports which are sold in the United States at less than their fair market value. 19 U.S.C. § 1673(1) (1988). The act of selling imports at less than fair market value is called “dumping.” As a result, duties imposed on LTFV imports are called "antidump-ing” duties.
By statute, the Department of Commerce must impose countervailing and antidumping duties on subsidized and LTFV imports when the conditions of 19 U.S.C. §§ 1671, 1673 (1988) are met. In relevant part, those conditions are that: (1) the Department of Commerce determines that the subject imports are indeed being subsidized or sold at LTFV, 19 U.S.C. §§ 1671(a)(1); 1673(1) (1988); and (2) the Commission determines that: "an industry in the United States: (i) is materially injured, or (ii) is threatened with material injury ... by reason of [the subject imports],” 19 U.S.C. §§ 1671(a)(2)(A), 1673(2)(A) (1988).
. The petition alleged subsidized imports from Austria, Belgium, Brazil, France, Germany, Italy, Korea, Mexico, New Zealand, Spain, Sweden, and the United Kingdom. 57 Fed.Reg. 60,247, 60,248 (1992). It also alleged LTFV imports from Argentina, Australia, Austria, Belgium, Brazil, Canada, Finland, France, Germany, Italy, Japan, Korea, Mexico, the Netherlands, Poland, Romania, Spain, Sweden, and the United Kingdom. 58 Fed.Reg. 8974, 8974 (1993).
. In determining whether an industry in the United States is materially injured by particular imports, the Commission must first delineate the relevant industry. By statute, the relevant domestic industry is defined as: “the domestic producers as a whole of a like product ...19 U.S.C. § 1677(4)(A) (1988) (emphasis added). The statute defines "like product” as: "a product which is like, or in the absence of like, most similar in characteristics and uses with, the article subject to an investigation.” 19 U.S.C. § 1677(10) (1988). Thus, in making material injury determinations, the Commission first identifies classes of imported "like products.” Having identified those classes of products, the Commission then determines whether imports of products falling within the identified classes have caused material injury to the domestic industry which produces those products.
. For a description of the products included in each of these categories see 2 ITC Final Determination, supra, at 1-17 to 1-37.
. In its final determination, the Commission divided this category into two like products: corrosion resistant products other than clad plate, and corrosion resistant clad plate. 1 ITC Final Determination, supra, at 7.
. In Bethlehem’s words: "Certainly there is information on the record that both supports and detracts from these findings. Obviously, Petitioners believe that the overwhelming hulk of the evidence detracts from these findings, and so argued below. However, Petitioners recognize that the substantial evidence test is quite deferential to agency decisionmakers and would not be persisting in this appeal in the absence of undisputed record evidence that renders the ITC determination factually impossible.”
. See 19 U.S.C. §§ 1671b(d)(2), 1673b(d)(2) (1988) which provides:
If the preliminaiy determination of the [Department of Commerce] ... is affirmative, the [Department of Commerce] ... shall order the posting of a cash deposit, bond, or other security, as it deems appropriate, for each entry of the merchandise concerned equal to [the estimated amount of countervailing or antidump-ing duty that would be due assuming an affirmative final material injury determination]....
. Bethlehem contests the exclusion of imports of cold-rolled steel from Argentina, Belgium, France, Italy, and Spain from the Commission’s cold-rolled cumulation, and the exclusion of hot-rolled steel from Brazil, Germany, and Japan from the Commission’s hot-rolled cumulation.
. Bethlehem also argues that the Commission’s negligibility determinations in this case should be subject to the same burdens analysis developed by this court in Creswell Trading Co. v. United States, 15 F.3d 1054, 1059-62 (Fed.Cir.1994). We disagree. The analysis in Cresweil was adopted to meet "the context of an Item (d) investigation [where] Commerce necessarily requires information that is within the knowledge and control of the government of the exporting country, its exporters, or both, which information Commerce cannot obtain independently.” Creswell, 15 F.3d at 1060. Cresweil thus presented a situation in which the "party” bearing the burden of persuasion on an issue was not in a position to gather evidence on that issue. There, the Department of Commerce bore the burden of showing, to a preponderance, that certain im-. ported products were subsidized by the Indian government. Noting that Commerce had no way of collecting the evidence necessaiy to support such a contention, we created a legal presumption which (when triggered by an initial modest showing by Commerce) shifted the burden of going forward onto the Indian government to put forth evidence that the imports in question were not subsidized. Rebuttal of that presumption, of course, shifted the burden of going forward back to Commerce. We decline Bethlehem’s invitation to extend the analysis and rule of Cresweil to the circumstances of this case. Unlike the subsidy determination in Cresweil, negligibility determinations call upon the Commission to collect and evaluate evidence to which it has access. Therefore, the appropriate way for us to review the Commission’s negligibility findings is to examine the evidence of record as a whole, and determine whether substantial evidence supports those findings.
. This appears to be a typographical error. It probably should read "subparagraph (B)(i)(III).”
. As noted above, the Commission investigated and made determinations with respect to two additional like-products, corrosion resistant products and cut-to-length plate products, which are not at issue on this appeal. Thyssen's first argument derives, circuitously, from those other determinations.
. In its brief, Thyssen provides the following table, based on data from the record below, to illustrate its contention that the decline in German AUVs was not a result of falling German import prices:
*1364As the chart demonstrates, the apparent decline in German unit values was 14% ($495 minus $427 divided by $495). The real decline in German unit values, however, was much smaller. Applying the 1992 unit values from column 9 to the 1990 product mix in column 1, for example, yields an average unit value of approximately $488, only 1.4% lower than 1990's AUV of $495.
. Specifically, we have considered Thyssen's argument that, given a unanimous negative material injury finding, the Commission may not find a threat of material injury absent some showing that existing patterns of trade will likely change in the imminent future. We find this argument to be without merit. First, the cases cited by Thyssen on this point, not binding on this court in any event, do not support Thyssen's proposition. Moreover, purely as a logical matter, Thys-sen's proposition is incorrect. Even Thyssen must concede that at the moment in time when an injurious trade practice begins, it frequently will not cause an instantaneous material injury to the domestic industry. Consequently, between commencement of the injurious trade practice and the ripening of its materially injurious effect, the Commission may detect a threat of material injury (even absent a finding that patterns of trade are likely to further change), although a finding of present material injury would be premature. Any such determination, of course, would depend upon the particular facts of a given case.
. By statute, a vote of three to three constitutes an affirmative determination of a threat of material injury. See 19 U.S.C. § 1677(11) (1988).
. At one point in its brief, Hoogovens does assert that the AUV of its largest volume specialty product (accounting for [13% ] of Hoogovens’ imports) was consistently higher than that for the domestic product. Hoogovens fails, however, to provide any record support for that assertion. Furthermore, it does not appear that Hoogovens presented this evidence to the Court of Intema-tional Trade. See Kern-Liebers USA, Inc. v. United States, 17 I.T.R.D. (BNA) 1082, 1102, 1995 WL 33066 (Ct. Int’l Trade 1995) ("Furthermore, the only data Hoogovens submitted, to support its claim that the pricing data, though accurate, was unrepresentative of its total imports, were the generalized affidavits by the company presidents of two of Hoogovens’ major purchasers.").
. Exercising the discretion granted him by statute, Commissioner Newquist cumulated imports from various countries for purposes of his threat determinations. See 19 U.S.C. § 1677(7)(F)(iv)(I) (1988 & Supp V 1993). Specifically, Commissioner Newquist cumulatively assessed the effects of imports from Belgium, Brazil, Canada, France, Germany, Japan, Korea, and the Netherlands. 1 ITC Final Determination, supra, at 269.
. Commissioner Nuzum cumulated the imports of Belgium, Brazil, France, Germany, Italy, Korea, the Netherlands, and Spain in making her determination. 1 ITC Final Determination, supra, at 385-87.