Warren L. Starks appeals from a District Court ^[1] order affirming the Secretary’s denial of disability benefits and supplemental security income. We affirm.

Starks was born on November 24, 1937 and worked for most of his adult life in packinghouses as an unskilled laborer. Eventually, increasingly debilitating physical impairments caused him to stop working, and on April 3, 1985 he filed for the benefits at issue in this case. After a hearing at which Starks was represented by counsel, the AU determined that exer-tional limitations precluded Starks from returning to his former work but that he retained the residual functional capacity to perform work that is sedentary. Acknowledging that the burden of proving employa-bility had shifted to the Secretary, the AU then considered whether there were jobs classified as sedentary that Starks could perform given his “borderline” intellect, and his age, education, and work experience. Based on the evidence presented, including vocational expert testimony, and using Rule 201.18 of the Medical-Vocational Guidelines (see 20 C.F.R. Part 404, Sub-part P, App. 2, Table 1 (1988)) as a frame of reference, the AU determined that Starks could perform entry-level and unskilled sedentary jobs and found him “not disabled.”

[*189] Under Rule 201.17 of the Guidelines, a person of Starks’s residual functional capacity, age, and work experience is “disabled” and entitled to benefits if he is illiterate. The relevant Social Security Administration regulations define illiteracy as “the inability to read or write_ [S]om-eone [is] illiterate if the person cannot read or write a simple message such as instructions or inventory lists even though the person can sign his or her name. Generally, an illiterate person has had little , or no formal schooling.” 20 C.F.R. § 404.1564(b)(1) (1988). Starks claims that he is illiterate and that Rule 201.17 therefore directs a finding of “disabled.”

The primary issue before us is whether the AU’s finding that Starks is literate is supported by substantial evidence on the record as a whole. See Nettles v. Schweiker, 714 F.2d 833, 835 (8th Cir.1983). Substantial evidence is more than a mere scintilla; it is evidence that a reasonable mind might find sufficient to support a conclusion. Nettles, 714 F.2d at 835 n. 5; McMillian v. Schweiker, 697 F.2d 215, 220 (8th Cir.1983).

In Hagan v. Schweiker, 717 F.2d 1229 (8th Cir.1983), we upheld a finding of literacy. The evidence of literacy in that case consisted of the claimant’s “testimony that he has a fifth-grade education, can follow diagrams and measurements quite well, and can ‘make things out’ by picking through words.” Hagan, 717 F.2d at 1230. The evidence of literacy was also found sufficient in Glenn v. Secretary of Health & Human Services, 814 F.2d 387 (7th Cir.1987), and Elzy v. Railroad Retirement Board, 782 F.2d 1223 (5th Cir.1986). The applicant in Glenn had “completed either the fourth grade of elementary school or the sixth grade,” but could not read a newspaper, write a letter, or read the notice of hearing that the Social Security Administration had sent him. He could write “only the simplest messages” (such as “I’ll be back at 10:00”) and comprehend “only the simplest written instructions.” Glenn, 814 F.2d at 390-91. In Elzy, the applicant claimed illiteracy but school records showed he had completed the ninth grade and he admitted that if he had to he could write a letter. Elzy, 782 F.2d at 1225.

Findings of literacy were held not supported by substantial evidence in Eggleston v. Bowen, 851 F.2d 1244 (10th Cir.1988), Dollar v. Bowen, 821 F.2d 530 (10th Cir.1987), and Dixon v. Heckler, 811 F.2d 506 (10th Cir.1987). The ALJ’s finding of literacy in Eggleston was apparently based solely on an inference that the claimant’s previous employment required the ability to read and write. Rejecting the finding of literacy, the court of appeals pointed out that the inference drawn by the AU was rebutted by the claimant’s “direct testimony” that other people had done the reading and writing for him. Eggleston, 851 F.2d at 1248. In Dollar the claimant had an eighth grade education and could sign his name, but otherwise there was “absolutely no evidence” of literacy. Dollar, 821 F.2d at 535. And in Dixon the claimant had completed six or seven years of formal schooling, but she testified that she could not read a newspaper, and both she and her sister testified that she could not write. The court concluded that although “[t]here was evidence that [the claimant] could read, albeit with difficulty,” a finding of literacy was not supported by the evidence because there was “simply no indication that [the claimant] could ‘write a simple message such as instructions or inventory lists,’ ” a prerequisite, according to the court, to a finding of literacy under the regulations. Dixon, 811 F.2d at 510.

We turn now to the particular facts in this case. Starks has been formally educated in public schools through the tenth grade. When asked by the AU if he could read and write, Starks responded, “Not really.” Administrative Transcript (Tr.) at 66. He testified that he does not read a newspaper and that if he were to he would “stumble over the words,” knowing “some” but not knowing a “lot.” Tr. at 66-67. He further testified that he would have trouble reading such things as lists of products and inventories. He also testified that he took an oral examination to obtain his driver’s license and that he does not write checks (but contrary to what is ar[*190] gued on appeal, he did not testify that he had to take an oral examination or that he cannot write a check). Two long-time acquaintances of Starks provided testimony and a letter indicating that Starks could not read well enough to understand newspaper articles and important papers such as job applications and unemployment reports. Tr. at 91, 250. Another acquaintance described Starks’s reading ability as “not very good.” Tr. at 249. After the hearing before the AU, Starks was given a Wide Range Achievement Test on which he obtained “a standard score of 55 on the reading section, a 59 on the spelling section and a 63 standard score on the arithmetic portion.” Tr. at 216. The psychologist reporting Starks’s scores described them as “below a third grade equivalent.” Tr. at 216. Other evidence in the record bearing on the question of Starks’s literacy was a Reconsideration Disability Report indicating that the interviewer did not perceive Starks to have difficulty reading or writing, Tr. at 166, a Report of Contact stating that Starks “reports that he is able to read and write but has difficulty with some of the big words,” Tr. at 168, and a comment by a psychologist that on an achievement test given in 1953 Starks apparently scored “a 5th grade equivalent.” Tr. at 216.

Viewing the above evidence in light of the aforementioned cases, we believe the AU’s finding that Starks is literate is supported by substantial evidence. Starks may be unable despite his years of formal schooling to read well enough to understand newspaper articles, but as the court in Glenn pointed out, such proficiency is not required by the regulations. Rather, a finding of literacy is warranted even when the claimant is able to read and write only very simple messages. Starks’s acknowledgment that he would know “some” but not a “lot” of the words in newspaper articles, his achievement test scores (“below a third grade equivalent”), and the Report of Contact (“able to read and write but has difficulty with some of the big words”) suggest such an ability. That evidence, in conjunction with his years of formal education (through a level substantially higher than that of claimants found literate in Hagan and Glenn), the Reconsideration Disability Report (no perceived difficulty reading or writing), and the psychologist’s comment (on achievement test given in 1953 Starks apparently scored “a 5th grade equivalent”), we think puts this case more in line with Hagan, Glenn, and Elzy, than with Eggleston, Dollar, and Dixon. We reiterate: in Eggle-ston the finding of literacy was based solely on an inference drawn by the ALJ, an inference rebutted by the claimant’s testimony; in Dollar there was “absolutely no evidence” of literacy except for the claimant’s eighth grade education and ability to sign his name; and in Dixon the court found “no indication” that the claimant could write a simple message.

Starks presents arguments challenging the reliability of some of the evidence adduced at the hearing before the AU. Starks points out, for example, that he testified he could not understand simple written instructions, and that the psychologist reporting Starks’s achievement test scores as “below a third grade equivalent” did not indicate how far below. He also argues that the Report of Contact is suspect because it erroneously reports that Starks understood a letter sent him by the Social Security Administration, and that a claimant’s years of formal schooling may be considered conclusive evidence of literacy only if there is no contradictory evidence. To each of these arguments, however, there is a counter-argument. First, the credibility of testimony at the hearing is a question for the AU. See Smith v. Heckler, 760 F.2d 184, 187 (8th Cir.1985). Given the AU’s finding that Starks’s testimony concerning “foot pain, shortness of breath, etc., [was] not credible” and the evidence suggesting literacy, the AU was entitled to believe that Starks has an ability to read and write greater than that which he claimed. Second, it is reasonable to interpret “below a third grade equivalent” to mean that Starks’s achievement test scores fell between the averages for second and third graders. Third, although the Re[*191] port of Contact may have contained an error in one respect it does not follow that the other information in the report was erroneous. And fourth, while the level of Starks’s formal education is not conclusive evidence of his literacy it is at least some evidence thereof, and that is how the AU treated it.

For the purpose of “identifying a class of workers with negligible employment opportunities, the standard for literacy has been pitched quite low.” Glenn, 814 F.2d at 391. Even so, this is a close case and had the issue of Starks’s literacy been presented to us in the first instance we might have decided it differently. Our task, however, is not to try the issue de novo or reweigh the evidence but only to determine whether the Secretary’s determination was supported by substantial evidence. We conclude that although Starks may read and write very poorly, substantial evidence supports the Secretary’s finding that he is literate within the meaning of the regulations. ^[2]

The order of the District Court upholding the decision of the Secretary is affirmed.