In November, 1979 the Crane Company closed the foundry it had operated since 1942 at Indian Orchard, Massachusetts. The foundry had produced steel castings used to make steam and water valves. Believing that the closing was due to import competition, representatives of the foundry workers applied to the Secretary of *27Labor for a certification of group eligibility to apply for benefits under the trade adjustment assistance provisions of the Trade Act of 1974, 19 U.S.C. §§ 2271-2322. The Secretary determined, after investigation, that increased imports had not “contributed importantly” to the closing of the foundry and therefore denied certification. The foundry workers, through their collective bargaining representative, Local 167, International Molders and Allied Workers’ Union, AFL-CIO, petition for review of that decision, claiming that it was based on faulty methodology and questionable data. We conclude that the Secretary’s decision is supported by substantial evidence and therefore affirm.
Under the worker adjustment assistance program, workers injured by import competition may receive benefits such as unemployment compensation, job search and relocation allowances, training, and other employment services. See generally Fortin v. Marshall, 608 F.2d 525, 525-26 (1st Cir. 1979); Usery v. Whitin Machine Works, Inc., 554 F.2d 498, 500 (1st Cir. 1977). To apply for benefits, a group of workers must first obtain a certification of group eligibility from the Secretary of Labor. The Secretary is required to certify a group of workers if he determines that three statutory criteria are met. We are concerned here only with the Secretary’s negative determination of the third criterion:1
“that increases of imports of articles like or directly competitive with articles produced by such workers’ firm or an appropriate subdivision thereof contributed importantly to such total or partial separation, or threat thereof, and to such decline in sales or production.” 19 U.S.C. § 2272(3).
The key phrase “contributed importantly” is defined in § 2272 to mean “a cause which is important but not necessarily more important than any other cause.”
Local 167 attacks both the methodology and the data relied upon by the Secretary to evaluate whether increasing imports “contributed importantly” to the foundry closing. First, petitioner objects to the Secretary’s focus upon Crane customers who decreased their purchases of Crane products and increased their purchases of imports. Petitioner argues that this approach fails to detect more subtle forms of import substitution, as would occur if a customer decreased both imports and domestic purchases, while increasing the import share of its reduced total purchase requirements. Petitioner also contends that the data relied upon by the Secretary was incomplete because it was derived from a survey of fewer than all of Crane’s customers, and unreliable because it was based, in part, upon the unverified representations of the Crane Company. Before considering these points, we first describe the Secretary’s investigation and its results.
The OTAA Investigation
The Office of Trade Adjustment Assistance [OTAA] is the body within the Department of Labor charged with responsibility for investigating petitions for certification of group eligibility for worker adjustment assistance. 29 C.F.R. § 90.12 (1980). The investigation in this case occurred in two distinct phases. After receiving the foundry workers’ petition, the OTAA published a Notice of Investigation, 44 Fed.Reg. 62378 (1979) , inviting requests by interested parties for a public hearing. When no hearing request was received, the OTAA proceeded to evaluate the petition based upon information supplied by the Crane Company and available trade data. In this first phase of its investigation the OTAA limited itself to the market for steel castings — the product made at the Indian Orchard foundry. The OTAA concluded that steel castings imports had not “contributed importantly” to the closing of the foundry and, on January 4, 1980, published a Notice of Negative Determination to that effect. 45 Fed.Reg. 1164 (1980) .
*28Several months later, counsel for Local 167 sought administrative reconsideration of this decision. The OTAA deemed the request for reconsideration untimely, but agreed to reopen its original investigation to evaluate import competition in the market for valves, since the first phase of the investigation had revealed that most of the steel castings produced at the Indian Orchard foundry were used to manufacture valves at the Crane Company’s Indian Orchard machine shop. The second phase of the investigation treated the foundry and machine shop as an integrated unit, and thus the “appropriate subdivision” of the firm for certification purposes. 19 U.S.C. § 2272. After examining valve import competition, the OTAA concluded that neither valve nor casting imports had contributed importantly to the foundry closing. A final Notice of Negative Determination was published in the Federal Register on May 23, 1980, 45 Fed.Reg. 35043 (1980);
In both phases of its investigation the OTAA employed a two-stage analysis, looking at the impact of import competition on the industry as a whole and on the individual firm. The object of the first stage is to determine possible indirect effects of import competition.'on the firm by examining the extent of import penetration industry-wide, as well as the recent trend in import purchases, both absolutely, in comparison to import levels in past years, and relatively, in comparison to domestic production and consumption. The second stage of the investigation is designed to determine whether customers of the firm in question have reduced their purchases of domestic products while increasing their purchases of imports. The OTAA gathers this information by submitting a questionnaire to a sample of firm customers.
The combined OTAA investigation reached the following results with respect to steel castings and valves.
A. Steel Castings
Based on the available data, the OTAA study showed that steel casting imports increased steadily in recent years both absolutely and relative to domestic shipments. Measured in absolute terms, imports increased in value 279.6 percent from 1975 to 1979. Part of this increase was due to inflation, but imports had also captured an increasing share of the American market; the ratio of imports to domestic shipments grew 155.5 percent over the same period. In the years immediately preceding the foundry closing, the import ratio decreased slightly from 1977 to 1978,2 but then almost doubled from 1978 to 1979. The OTAA found these increases significant, and perhaps even understated by the data, but discounted their potential impact since the level of import penetration was minimal, never amounting to more than 2.3 percent of domestic shipments.
The second phase of the investigation, concentrating on the Indian Orchard foundry, was simplified considerably by the OTAA’s discovery that 90 percent of the foundry’s output had gone to other facilities operated by the Crane Company. The OTAA accepted the statement of a representative of the Crane Company that it would replace the production of the Indian Orchard foundry from domestic sources, with the exception of one facility whose purchases amounted to less than one percent of the Indian Orchard foundry’s annual output. Based on this information, the OTAA concluded, without conducting a customer survey, that there was no significant substitution of imported castings among foundry customers.
B. Valves
The trend in purchases of imported valves found by the OTAA roughly tracked that of castings: a small but increasing degree of import penetration. The study showed that valve imports increased in value 112.5 percent from 1975-79, while the ratio of imports to domestic shipments increased 41.6 percent over the same period; yet the ratio of imports to domestic shipments never exceeded 5.1 percent. Other *29factors indicated that the valve industry/ had not suffered from import competition, The OTAA study showed that, unlike the situation in the casting industry, employment in the valve industry had increased annually since 1975. The study also revealed that the United States had enjoyed a substantial balance of trade in valves each year since 1976, exporting approximately twice the amount by value that it imported.
To determine if increasing imports of valves had contributed to reduced demand for castings at the Indian Orchard machine shop, the OTAA surveyed a sample of four purchasers of valves from the machine shop. Together these customers accounted for 49 percent of valve sales in 1978 and 68 percent in 1979. The OTAA examined the customers’ purchases over two sets of periods, comparing 1979 to 1978 and the first two months of 1980 to the same period in 1979. Over the first period it found that three of the four customers had increased valve purchases from Indian Orchard in 1979, compared to 1978, while maintaining or decreasing their purchases of imports. The fourth customer decreased purchases from Indian Orchard from 1978 to 1979, but, so the Secretary informs us, also decreased its purchases of imports to nearly the same extent. During this period, total Indian Orchard valve sales decreased 7.3 percent, of which the survey customers accounted for 38 percent.
Over the second comparison period, one customer increased purchases from Indian Orchard, while decreasing imports. The other three decreased purchases from Indian Orchard, but only one of these also reported increasing purchases of imports. That customer reduced its Indian Orchard purchases by $50,000, while increasing import purchases by $10,000, but at the same time increased purchases from other domestic sources by $700,000. During the second comparison period total Indian Orchard valve sales decreased 21.8 percent, of which the survey customers accounted for 70 percent.
Thus, the survey showed that over both comparison periods, only one customer had directly substituted imports for Indian Orchard valves, and that customer simultaneously had substantially increased its reliance on other domestic sources.. Based on these data, the OTAA determined that customer substitution of imported valves had not figured importantly in the decline of machine shop sales, and consequently could not have contributed to the closing of the foundry.
Petitioner’s Contentions
In our review of the denial of certification, we must accept the Secretary’s findings of fact if supported by substantial evidence. 19 U.S.C. § 2322(b). There can be little question that, if the findings are supportable, the Secretary’s determination must be affirmed. Far from an “important” factor, the OTAA study showed that import substitution of castings and valves could have contributed only slightly to the closing of the foundry. The question before us, therefore, is whether, as petitioner contends, the method adopted by the OTAA was so flawed as to fatally taint the findings relied upon for the Secretary’s decision.
A. Relative Import Substitution
Local 167 attacks the method chosen by the OTAA to analyze the customer survey results as embodying a simplistic notion of the phenomenon of import substitution. In stating the basis for its recommendation, the OTAA analyzed the survey results solely in terms of a dual test measuring direct import substitution: an increase of purchases of imports paired with a decrease of purchases from Indian Orchard. This approach, petitioner correctly observes, would permit relative import substitution to go unnoticed if a customer had increased or decreased purchases of both imports and domestic products, thereby escaping one part of the OTAA’s dual test, while increasing the import share of its total purchases.3 *30The Secretary does not challenge this characterization of the OTAA methodology, but defends it as an adequate, if approximate, means of assessing the impact of import competition.
A case for limiting the inquiry to direct substitution could be made by interpreting the phrase “increases of imports” in § 2272(3) to mean only absolute increases. The Secretary has closed off that route, however, by defining “increases of imports” to mean absolute or relative increases.4 29 C.F.R. § 90.16(3) (1980); see also id. § 90.2 (“ ‘increased imports’ means that imports have increased either absolutely or relatively”). In the first phase of its investigation, the OTAA evaluates relative substitution by computing the industry-wide ratios of imports to domestic production and consumption, and the Secretary relies on the relative degree of import penetration revealed by those ratios to support his determination in this case. No reason has been offered us why relative substitution is any more difficult to gauge or any less significant a factor to consider at the individual firm level.
It is easy to imagine a case where, in a period of generally declining demand, management — considering if a firm should try to weather the storm — would base its decision in important part on whether the firm maintained its market share of the reduced demand, or lost ground to import competition. In such a case, petitioner’s challenge to the OTAA’s dual test would be persuasive. The evidence for that scenario, however, is almost nonexistent in this case. Only the two customers who, over the second survey period, decreased their purchases from Indian Orchard but did not increase their import purchases could have engaged in relative import substitution.5 All others, excluding the one customer who showed direct substitution, either increased purchases from Indian Orchard while maintaining or decreasing their purchases of imports, or decreased Indian Orchard purchases and imports proportionally. It is not possible to tell from the OTAA summary of the survey results whether those two customers decreased their import purchases by a proportionate amount. The summary does show, however, that the aggregate purchases of imports for all four customers, including the one who increased imports, declined as a percentage of total demand from 6.7 percent in the first two months of 1979 to 4.5 percent in the same period of 1980. Thus, contrary to the scenario postulated by petitioner, over the relevant period the import share of total demand declined.
We agree with petitioner and Judge Leventhal writing in United Glass & Ceramic Workers v. Marshall, 584 F.2d 398, 405 (D.C. Cir.1978), that “looking to whether, a customer had directly shifted to imports is not a very sophisticated test.” Indeed, the Secretary’s own regulations appear to demand a more discerning look. But, at least where, as here, the data count so strongly against the existence of the phenomenon on which petitioner’s argument is premised, we cannot say that the method adopted by the OTAA is not reasonably calculated to detect an important causal relationship between *31increased reliance on imports and the loss of the petitioning workers’ jobs. We must therefore defer, in this case, to the Secretary’s choice of analytic technique.
B. Partial Survey-
Local 167 also objects to the OTAA’s failure to survey all Indian Orchard customers. The OTAA admittedly conducts only a partial customer survey, the necessary consequence of which is that some customers who may have switched to imports are not polled. We are informed that the OTAA normally attempts to survey several customers of different sizes and geographic locations, accounting for 25-30 percent of the subject firm’s sales. The Indian Orchard valve customers surveyed here substantially exceeded those percentages, accounting for 49 percent and 68 percent of sales during the two survey periods. Because the Crane Company absorbed nearly all of the foundry’s production of castings, the only casting “customer” surveyed accounted for 90 percent of sales.
We think a representative partial survey is an acceptable technique for assessing “important” causation, without placing the burden on the Secretary to demonstrate in each case that a total survey is not feasible. The statute does not impose such a duty, nor are we, as a reviewing court, able to say that a partial survey will not uncover evidence of a significant causal relationship. Petitioner does not argue that the sample taken was unrepresentative and, in fact, the decline in machine shop business attributable to the survey customers was roughly commensurate with their share of machine shop sales. In sum, we are not convinced that a representative sample is objectionable either generally, or under the circumstances of this case.
C. Lack of Verification
Lastly, Local 167 finds fault with the OTAA’s acceptance, without verification, of the Crane Company’s statement that the output of the Indian Orchard foundry would be replaced almost entirely with castings from other domestic sources. This statement formed the crux of the Secretary’s negative determination with respect to the foundry. Petitioner contends that the OTAA abused its discretion by simply taking the word of the Crane Company on so important a question.
Petitioner’s objection is, in part, that the statement of the Crane Company was a mere prediction, which, as far as the OTAA knows, never came true. The relevance of the prediction, however, depends less on its accuracy as a forecast of the future than on its veracity as an expression of the belief of the Crane Company at the time it was made. If unforeseen events later produced a shift to imports, that would have no bearing on what caused the foundry to close when it did. The principal complaint, then, is that the OTAA erred in crediting the Crane Company statement.
We recognize that the certification process does not allow petitioner the opportunity to test credibility that it would have in an adversary proceeding. The Secretary is charged with investigating the petition and the petitioning workers are relegated to presenting evidence at a public hearing, if they so choose. To insure willing cooperation with the customer surveys, the Secretary guarantees survey respondents that confidential business information will not be disclosed. 29 C.F.R. § 90.33 (1980). This shield of confidentiality was extended to the Crane Company statement, now questioned by petitioner, up until the time for briefing this appeal. Thus, even if the foundry workers had sought a hearing in this case, their lack of access to the contrary evidence would have precluded effective rebuttal.
Because of the ex parte nature of the certification process, and the remedial purpose of the trade adjustment assistance program, the Secretary is obliged to conduct his investigation with the utmost regard for the interest of the petitioning workers. We cannot say, however, that the Secretary has shirked his duty to investigate by accepting as credible an authoritative company response to an official query *32given under a guarantee of confidentiality. The duty of verification proposed by petitioner, which is applicable to customer surveys generally, would greatly increase the investigative burden of the Department of Labor. We note that there are not objective circumstances in this case suggesting that the company gave a less than truthful response, nor does it appear that the company would have financially benefitted from the denial of certification. It was for the Secretary to decide, under these circumstances, whether to credit the Crane Company statement without further checking.
Affirmed.
. The other two requirements, that a significant number of workers have lost employment and that sales or production of the firm have decreased absolutely, are not contested. See 19 U.S.C. § 2272(1), (2).
. The employees were first informed of the closing, on a temporary basis, during 1978.
. If only direct import substitution is taken into account, a shift in a customer’s total purchases might obscure a relative increase in imports, as the following example illustrates: In year one, *30a customer buys 50 units of imports and 50 units of domestic products; in year two, it buys 40 units of imports and 10 units of domestic products. Comparing year two to year one, the customer has decreased its purchases of both imports and domestic products, but, because it has halved consumption, it has increased the import share of total purchases from 50% to 80%. In the same manner, a customer might increase the import share of total purchases while increasing both imports and domestic products in absolute terms.
. A Senate amendment adding the requirement that the Secretary find an absolute increase of imports was deleted by the Conference Committee in favor of the present language of § 2272(3). H.R.Conf.Rep.No.93-1644, 93rd Cong., 2d Sess. (1974), reprinted in [1974] U.S. Code Cong. & Ad.News 7186, 7367, 7381.
. No customer increased both domestic purchases and imports. Thus, we are not called . upon to decide if, in such a case, inquiry into relative import substitution is foreclosed by the requirement of an absolute decrease in sales or production of domestic products, 19 U.S.C. § 2272(2); see United Glass & Ceramic Workers v. Marshall, 584 F.2d 398, 406 & n. 29 (D.C.Cir.1978), or if that requirement is satisfied merely by the decline in sales that inevitably follows a plant closing.