The Effect of Extra Time on Reading Comprehension Scores for University Students With and Without Learning Disabilities M. Kay Runyan The purpose of this study was to examine the effects of extra time on the ability of university students with and without learning disabilities to complete a reading comprehension test under timed and extra-time conditions. The participants were 16 students identified as having learning disabilities and 15 normally achieving students, all attending the University of California, The Nelson-Denny Reading Comprehension and Reading Rate Test was used. Percentile ranks were obtained for reading rates on individuals, and for comprehension scores under timed and extratime conditions. The major findings of this study were that there is a significant difference between scores obtained by students with learning disabilities and by normally achieving students under timed conditions and that there are no significant differences in test performance between students with learning disabilities and normally achieving university students when students with learning disabilities are provided extra time. Normally achieving students did not perform significantly better with extra time.
D
espite the passage of Section 504 of the Rehabilitation Act of 1973, there is controversy over "reasonable accommodations" for students with learning disabilities (LD) in higher education. Some of the controversy centers around extra time on exams. There are professors who believe that the accommodation of extra exam time for students with learning disabilities is unfair to normally achieving students. Others believe that students with learning disabilities need extra time only if the test requires a large amount of reading, and that extra time is not necessary for math or other tests requiring limited reading. The extent of the effect that timing has on the test-taking performance of individuals with LD is unclear. Educational Testing Service (ETS) has recognized that standardized tests such as the Scholastic Aptitude Test (SAT) and the American College Test (ACT) do not validly measure the academic aptitude of students with learning disabilities; therefore, they allow such students extended time limits to take the tests (ACT, 1981). While this practice makes intuitive sense, it has not been based on solid empirical evidence. Much of the research involving adults 104
with learning disabilities has probed the similarities and differences between them and children with learning disabilities, since most research thus far has been on children. Cordoni, O'Donnell, Ramaniak, Kurtz, and Rosenshine (1981) reported that adults with learning disabilities did poorly on the Wechsler Adult Intelligence Scale-Revised (WAIS-R) (Wechsler, 1981) on Digit Span, Arithmetic, and Coding subtests. This is often an indication of an attention deficit, which is frequently reported as being associated with learning disabilities. Fraunheim and Heckerl (1983) tested 11 adults with learning disabilities and found deficits on the WAIS-R Arithmetic, Coding, and Digit Span subtests. Those deficiencies were interpreted as symbol-processing deficits, and the finding in this study was consistent with testing completed on the same subjects 17 years earlier. Those same adults had decoding problems, reading was worse when words were not presented in sentence contexts, the rate of oral reading was slow and laborious with rereading for comprehension, and word reversals still occurred. Speece (1987) investigated the infor-
mation-processing speed of various subtypes of individuals with learning disabilities. Her study revealed that most (76%) readers with learning disabilities exhibited a speed deficit, which supports previous research in this area (e.g., Denckla & Rudel, 1976; Stanovich, 1986). The above studies involving adults with learning disabilities offer evidence that the range of deficits (i.e., attention deficit, concentration problems, reading problems, and slow processing of printed information) demonstrated by children with learning disabilities can be generalized to adults with learning disabilities. Based on research involving adults with learning disabilities, colleges and universities are attempting to offer appropriate accommodations for students with learning disabilities. Some programs offer extra time on exams as a standard accommodation (Barbaro, 1982; Bireley & Manley, 1980; Strichart & Mangrum, 1988); however, there is controversy surrounding this issue, and many programs serving students with learning disabilities offer few services beyond tutoring (Cordoni, 1980). Cordoni attributes the inconsistency in accommodations to a lack of research on college students with learning disabilities. Studies on the effects of extra exam time for college students with learning disabilities are now beginning to emerge. These studies should be critically evaluated to determine essential accommodations for these particular students. A study conducted by Hill (1984) showed that normally achieving college students perform better than college students with learning disabilities on the ACT and Nelson-Denny Reading Test (Brown, Bennett, & Hanna, 1981) under timed conditions. College students with learning disabilities perform just as well as normally achieving college students on the ACT (on all four of the ACT subtests: English, Math, Social Studies, and Natural Science) and Nelson-Denny Reading Test under untimed testing conditions. The Nelson-Denny vocabulary subtest was the only test where the normally achieving college students performed better in both the timed and untimed situations than did college students with learning disabilities. In a more recent study, Halla (1988) concluded that there are no significant Journal of Learning Disabilities
differences in test performance between cock & Johnson, 1977), the Wide Range students with learning disabilities and Achievement Test-Revised (WRAT-R) normally achieving, upper-level under- (Jastak & Wilkinson, 1984), and other graduate, graduate, and professional achievement tests deemed appropriate by school adults on the Graduate Record the examiner. The students with learning Exam (GRE) General Test and Nelson- disabilities met the university criteria of Denny Reading Test, and that untimed 1.5 standard deviations between intellitesting conditions significantly influence gence and achievement and were receivtest performance on those two assess- ing services from the Disabled Students ments for both groups. Halla points out Program. All students with learning disthat the data may be interpreted differ- abilities had a long history of reading ently if IQs are taken into effect. The IQ problems and had developed various of the group with learning disabilities was techniques for compensating; however, a range above that of the normally all complained of slow reading rates and achieving group (1 standard deviation). having to reread material several times The group with LD could be described for maximum comprehension. Case hisas highly intelligent but, in terms of tories on 14 of the students with LD inachievement, functioning at the mean dicated inattention (short attention level of the other group. spans) and impulsivity during their early The purpose of the present study was school years. The students with learning to evaluate reading comprehension scores disabilities were interviewed by the inof adults with and without learning dis- vestigator. All students with LD reported abilities under timed and untimed con- problems with concentration when readditions. The study offers evidence that ing, attributing the lack of attention and students with learning disabilities score concentration to their slow, laborious significantly lower than their normally rate of reading. Most reported that they achieving peers under time constraints, could concentrate and read for approxbut that there is not a significant dif- imately 15 to 20 minutes, and then had ference between students with learning to take a short break. disabilities and normally achieving stuThe other 15 participants were normaldents under untimed conditions. The per- ly achieving (NA) University of Califorformance of normally achieving students nia students who had never been identiunder timed and untimed conditions is fied as having learning disabilities, speech examined, and there is not a significant problems, reading problems, or spelling difference in performance. Amount of problems, and were not on academic extra time is discussed, as evidence that probation (reported on a questionnaire). students with learning disabilities are a They were solicited randomly by making heterogeneous group and setting fixed announcements in various classes at the amounts of extra time for all students university. may not be appropriate. The two groups selected were controlled for ethnicity, gender, age, and total Scholastic Aptitude Test (SAT) (Educational Testing Service, 1986) scores. All METHOD participants in the sample were white: 16 were males and 15 were females. The Subjects mean age of the group with learning disThe participants in this study were stu- abilities was 21.3 years, ranging from 18 dents at the University of California. to 27 years. The mean age of the norThirty-one students participated in the mally achieving group was 22 years, study, 16 of whom were identified as ranging from 19 to 27. Fifteen of the having learning disabilities (LD). Prior students with learning disabilities were to entering the university, the students undergraduates and one was a graduate with learning disabilities had been tested student. Thirteen of the normally achievand identified through the use of a com- ing students were undergraduates and prehensive test battery, including the two were graduate students. Wechsler Adult Intelligence Scale-Revised The mean SAT score for students with (WAIS-R) and/or the Woodcock-Johnson learning disabilities admitted to the uniPsycho-Educational Test Battery (Wood- versity in 1987-88 was 1125, with a stanVolume 24, Number 2, February 1991
dard deviation of 185. A cutoff point was established at 1 SD above and below the mean, establishing a range from 940 to 1310. For this study, students selected had SAT scores within the range of 940 to 1310. The students with learning disabilities were selected at random, regardless of whether they took the SATs under timed or extra-time conditions. Normally achieving students met the same criteria for SAT scores. The SAT scores tended to cluster at the higher end of the range, with only one student scoring below 1000.
Instrument The Nelson-Denny Reading Test was used. This instrument has two statistically equated forms, E and F. The test contains a vocabulary section and a silent reading comprehension section. The first minute of the reading comprehension is used for obtaining a reading rate. Sentences in the first reading passage have numbers to designate words per minute. At the end of 1 minute, the numbers (designating the amount of text read) are recorded and converted to a percentile rank according to the manual conversion charts. The Reading Comprehension section and the Reading Rate sections were used. The Reading Comprehension section contains eight reading passages and a total of 36 questions. The test is designed to be completed in 20 minutes, with college students generally obtaining their maximum score within the time limit. A measure of context dependence (i.e., the degree to which questions can be answered only by reading the passage) is provided. The manual reports that both Forms E and F are based on content more than on prior knowledge. The 36 items on each form can be divided into groups of 18 interpretive items and 18 primarily literal items. Comprehension is based on the degree of abstract information that a student must extract from the reading passage to answer the questions.
Procedure The participants were individually administered the Nelson-Denny Reading Comprehension Test by the investigator, and a reading rate was obtained for each 105
participant. Form E was administered to all students except two, who had been administered the test during the previous year. Those two students were administered the alternate, Form F. All students were told that this was a reading comprehension test and that the first minute would be used to obtain a reading rate. Although this was a timed test, the students were not told how much time they had to complete the test. They were told to read normally, as they would in a test situation, and that they were to work quickly and answer questions until they were told to stop. Two measurements were obtained on each subject—one score at the end of 20 minutes and another when the test was completed. The score remained the same for both measurements if students completed the test within the 20-minute limit. At the end of 20 minutes the participants were told to mark where they were on the answer sheet (if they had not completed the test) and to continue until they were told to stop (they were allowed time to complete the test). The amount of extra time needed to complete the test was recorded. Raw scores for words per minute were converted into percentile ranks according to the Nelson-Denny conversion charts. Scores under timed conditions and untimed conditions were converted to percentile ranks and used for comparison of the two groups.
TABLE 1 Comparison off Mean Percentile Scores for Comprehension Under Timed and Extra-Time Conditions (20 minutes) Group
106
Extra time (Mean)
Paired f-value
Probability (two-tailed)
.1281 .8227
.7587 .8347
-11.73 -1.19
.000 .253
LD NA
Note. LD = learning disabilities; NA = normally achieving.
• LD./Timed
L.D. and N.A. Timed and Extra-Time • L.D. /Extra-Time •N.A./Timed
ON.A./Extra-Time
CO
LU UJ LU
0 Figure 1.
2
4
6
8 10 12 14 16 18 SUBJECT Learning disabled and normally achieving students under timed and extra-time conditions.
TABLE 2 Comparison of Mean Percentile Scores for Normally Achieving Groups (20 minutes) Groups Variable Reading Rate Comprehension (timed) Comprehension (extra time)
RESULTS The effects of extra time for the two groups are given in Table 1 and Figure 1. The paired t tests indicate a significantly higher mean percentile rank comprehension score for subjects with learning disabilities when given extra time (/?= .000). For normally achieving subjects, there was no significant effect of extra time (p= .253). Table 2 shows the comparison of mean percentiles for the normally achieving group and the group with LD for each of the variables measured. Reading rate was significantly higher for normally achieving students compared to that of students with learning disabilities (p = .000). The same result was obtained for the Nelson-Denny Comprehension score under timed conditions (p = .000). How-
Timed (Mean)
NA Mean SD
LD Mean SD
.5953
.247
.1069
.8227
.115
.8347
.124
i value8
Probability (two-tailed)
.116
6.98
.000
.1281
.106
17.46
.000
.7587
.203
1.26
.218
Note. NA = normally achieving; LD = learning disabilities. Separate variance estimate.
a
ever, under extra-time conditions, there was no significant difference for the comprehension score between the normally achieving students and students with learning disabilities (/?= .218). When scores of the normally achieving subjects under timed conditions were compared to the scores of subjects with learning disabilities under extra time, again there was no significant difference between the two groups for the comprehension score 0? = .288). The tests were administered individual-
ly, allowing the investigator to observe the students. The students with learning disabilities did not complete the reading sections within the 20-minute time limit. The questions to which they responded were generally correct; however, they had responded to fewer questions under the time limit. Thirteen of the normally achieving students completed the comprehension sections within the 20 minutes. Two subjects needed extra time, completing the test with an extra 3 and 4 minutes, respectively. Journal of Learning Disabilities
DISCUSSION In this study, the following hypotheses were tested: (a) Students with learning disabilities will produce significantly lower scores than normally achieving students as measured by the Nelson-Denny Reading Comprehension Test under timed conditions; (b) both groups will increase their reading comprehension scores when tested under extra-time conditions, but the groups will not differ significantly in the untimed testing condition; and (c) the scores for the students with learning disabilities, when given extra time, will not differ significantly from those of the normally achieving students under timed conditions. Results of the comparisons between university students with learning disabilities and normally achieving university students support the hypothesis that students with learning disabilities score significantly lower than normally achieving individuals under timed conditions on reading comprehension. The study supports the hypothesis that students with learning disabilities perform as well as their normally achieving peers when given extra time, and that normally achieving students did not perform significantly better when given extra time. All but two of the normally achieving students finished in the 20 minutes. Their scores did not significantly increase the performance of the group, and none scored a perfect score. The extra time allowed the students with learning disabilities to respond to all questions. Of importance was the amount of extra time required for each individual student. The extra time required by students with learning disabilities varied a good deal, ranging from 4 minutes to 29 minutes. Students were not allowed to go back and change answers in the extra time; they were allowed only to finish the test. This lends support to the idea that students with learning disabilities have varied rates of processing printed information, and therefore a fixed amount of extra time for all students with learning disabilities on standardized tests may not be appropriate. The presentfindingscan be interpreted in the following ways: Normally achieving students often perform up to their capability under timed testing conditions Volume 24, Number 2, February 1991
and have little room for improvement. Students with learning disabilities often cannot perform up to their capabilities under timed conditions, and show considerable improvement when allowed extra time. Hill (1984), using a large sample and additional tests (ACT and Nelson-Denny Reading and Vocabulary Tests), interpreted his findings in the same way. All students with learning disabilities reported having to reread the passages several times to gain maximum comprehension. This suggests that these students may experience problems with attention span and concentration while reading. LeBerge and Samuels (1974) developed a theory on automaticity of information processing in which the primary component is attention. For the disabled reader, information at the visual and phonological levels is not processed automatically. A major portion of attention is needed for decoding and processing, leaving little to focus at the semantics level, thus reducing understanding and necessitating rereading a text for comprehension. Koppell (1979) suggested that attentional deficits may not necessarily be causes of reading disability but may result from a poor reader's previous failure experiences or anxiety about past performance. Senf (1976) agrees that attention deficits may not be primary disorders but may be secondary to previous problems in organizing and processing information. While the relationship between learning disabilities and attention disorder remains under investigation, it is clear that many students have both problems; often, however, one or the other remains undiagnosed. Standardized (timed) tests are "speed tests"—test in which individual differences in performance depend on speed of performance. The time limit on these tests is such that some individuals do not complete all the items, especially those persons with learning disabilities. "Power tests," on the other hand, have time limits long enough to permit most people to attempt all items on the test. Speed of performance is not an important factor on power tests (Anastasi, 1982). This is a critical issue, because previous research supports the concept that students with learning disabilities do not process information at a normal rate (Speece, 1987;
Rudel, 1981), and this slower processing may account for the difference in their performance on timed and untimed tests.
LIMITATIONS AND IMPLICATIONS The present study has several limitations, one of which is the small number of subjects involved in the research. Another limitation is that all the students attended the University of California on one campus, and therefore results may not be representative of all college students with learning disabilities. Growing numbers of students with learning disabilities are seeking admittance to colleges and universities (Strichart & Mangrum, 1988), but there is a paucity of research on the impact of timed exams on the performance of students with such disabilities. With the current controversy on providing extra exam time for students with learning disabilities in the postsecondary setting, there is a need for additional research.
ABOUT THE
AUTHOR
M. Kay Runyan is a doctoral candidate in the Joint Doctoral Program in Special Education at the University of California at Berkeley and San Francisco State University. She is coordinator of services to students with learning disabilities at the University of California at Berkeley. Address M. Kay Runyan, Disabled Students Program, University of California at Berkeley, Berkeley, CA 94720.
AUTHOR'S
NOTES
1. This research is a pilot study for a doctoral dissertation and was first presented at The Next Step Symposium, Harvard College, November 1989. 2. The research was supported in part by the University of California, Berkeley, in the form of an A&PS Development award. I gratefully acknowledge the assistance ofNori Ison as a reviewer of the statistics.
REFERENCES ACT. (1981). Assessment supervisors' manual of instruction for special testing. Iowa City, I A: American College Testing Program. Anatasi, A. (1982). Psychological testing (5th ed.). New York: Macmillan. Barbaro, F. (1982). The learning disabled college student: Some considerations in settling objectives. Journal of Learning Disabilities, 10, 599-603. Bireley, M., & Manley, E. (1980). The learning 107
disabled student in a college environment: A report of Wright State University's Program. Journal of Learning Disabilities, 13, 12-15. Brown, J.I., Bennett, J.M., & Hanna, G. (1981). Nelson-Denny reading test manual (Forms E and F). Boston: Riverside. Cordoni, B.K. (1980). College programs for learning disabled students. Perceptions, 3, 1-5. Cordoni, B.K., O'Donnell, J.P., Ramaniak, N.V., Kurtz, T., & Rosenshine, K. (1981). Wechsler Adult Intelligence Score patterns for learning disabled young adults. Journal of Learning Disabilities, 14, 404-407. Denckla, M.B., & Rudel, R. (1976). Rapid "Automatized" Naming (R.A.N.): Dyslexia differentiated from other learning disabilities. Neuropsychologia, 14, 471-479. Fraunheim, J.G., & Heckerl, J.B. (1983). A longitudinal study of psychological and achievement test performance in severe dyslexic adults. Journal of Learning Disabilities, 16, 339-346. Halla, J. (1988). GRE test scores of LD and non-
LD adults. Unpublished doctoral dissertation, ing disabilities. In S.R. Forness, K.A. Kavale, & Texas Technical University, Lubbock. M. Bender (Eds.), Handbook of learning disabilHill, G. (1984). Learning disabled college students: ities, Vol. 1, dimensions and diagnosis (p. 142). Austin, TX: PRO-ED. Assessment of academic aptitude. Unpublished doctoral dissertation, Texas Technical University, Speece, D.L. (1987). Information processing subtypes of learning disabled readers. Training DisLubbock. abilities Research 2(2), 91-102. Jastak, S., & Wilkinson, G. (1984). Wide range achievement test-Revised. Los Angeles: Western Stanovich, K.E. (1986). Cognitive processes and the Psychological Services. reading problems of learning-disabled children: Koppell, S. (1979). Testing the attentional deficit Evaluating the assumption of specificity. In J.K. notion. Journal of Learning Disabilities, 12, Torgesen & B.Y.L. Wong (Eds.), Psychological 43-48. and educational perspectives on learning disLeBerge, D., & Samuels, S.J. (1974). Toward a abilities (pp. 87-131). New York: Academic Press. theory of automatic information processing in Strichart, S., & Mangrum, C. (1988). College and reading. In S.R. Forness, K.A. Kavale, & M. the learning disabled student. Philadelphia: Grune Bender (Eds.), Handbook of learning disabilities, & Stratton. Vol. 1, dimensions and diagnosis (p. 140). Austin, Wechsler, D. (1981). Wechsler adult intelligence TX: PRO-ED. scale-Revised. Cleveland, OH: Psychological Rudel, R.G. (1981). Residual effects of childhood Corp. reading disabilities. Bulletin of the Orton Society, Woodcock, R., & Johnson, M.B. (1977). Wood31, 89-102. cock-Johnson psycho-educational test battery. Senf, G.M. (1976). Future research needs in learnAllen, TX: DLM Teaching Resource.
Tools to help treat Attention Deficit Disorder These PRO-ED products offer the indepth information you need to help students with ADD, their parents, and their teachers.
Attention Deficit Disorder This book and videotape by Dale R. Jordan will help you distinguish the four types of ADD, identify ADD behavior patterns, and foster the development of appropriate behaviors. #1457 Book $11.00 #1458 Videotape (VHS) $98.00
Attention Deficit Disorder and Hyperactivity, Second Edition This new revision, written by Ronald J. Friedman and Guy T. Doyal for parents and teachers, presents current research findings and clinical experiences with nearly a thousand children with ADD and their families. #3547 $11.00
108
Identification and Treatment of Attention Deficit Disorder This new handbook by Nancy Nussbaum and Erin Bigler contains practical information based on research for parents, teachers, and professionals. #1506 $8.00
Help! This Kid's Driving Me Crazy! The videotape by Jane Judy, Lynne Adkins, and Janis Cady (supported by a 64-page booklet) will help parents and professionals work more effectively with young children with ADD. #3548 Book #3550 Videotape (VHS)
$ 5.00 $89.00
ORDER NOW! For more information or a FREE catalog, write or fax today!
pro-ed
8700 Shoal Creek Boulevard Austin, Texas 78758 Office: 512/451-3246 Fax: 512/451-8542
Journal of Learning Disabilities
D
espite the passage of Section 504 of the Rehabilitation Act of 1973, there is controversy over "reasonable accommodations" for students with learning disabilities (LD) in higher education. Some of the controversy centers around extra time on exams. There are professors who believe that the accommodation of extra exam time for students with learning disabilities is unfair to normally achieving students. Others believe that students with learning disabilities need extra time only if the test requires a large amount of reading, and that extra time is not necessary for math or other tests requiring limited reading. The extent of the effect that timing has on the test-taking performance of individuals with LD is unclear. Educational Testing Service (ETS) has recognized that standardized tests such as the Scholastic Aptitude Test (SAT) and the American College Test (ACT) do not validly measure the academic aptitude of students with learning disabilities; therefore, they allow such students extended time limits to take the tests (ACT, 1981). While this practice makes intuitive sense, it has not been based on solid empirical evidence. Much of the research involving adults 104
with learning disabilities has probed the similarities and differences between them and children with learning disabilities, since most research thus far has been on children. Cordoni, O'Donnell, Ramaniak, Kurtz, and Rosenshine (1981) reported that adults with learning disabilities did poorly on the Wechsler Adult Intelligence Scale-Revised (WAIS-R) (Wechsler, 1981) on Digit Span, Arithmetic, and Coding subtests. This is often an indication of an attention deficit, which is frequently reported as being associated with learning disabilities. Fraunheim and Heckerl (1983) tested 11 adults with learning disabilities and found deficits on the WAIS-R Arithmetic, Coding, and Digit Span subtests. Those deficiencies were interpreted as symbol-processing deficits, and the finding in this study was consistent with testing completed on the same subjects 17 years earlier. Those same adults had decoding problems, reading was worse when words were not presented in sentence contexts, the rate of oral reading was slow and laborious with rereading for comprehension, and word reversals still occurred. Speece (1987) investigated the infor-
mation-processing speed of various subtypes of individuals with learning disabilities. Her study revealed that most (76%) readers with learning disabilities exhibited a speed deficit, which supports previous research in this area (e.g., Denckla & Rudel, 1976; Stanovich, 1986). The above studies involving adults with learning disabilities offer evidence that the range of deficits (i.e., attention deficit, concentration problems, reading problems, and slow processing of printed information) demonstrated by children with learning disabilities can be generalized to adults with learning disabilities. Based on research involving adults with learning disabilities, colleges and universities are attempting to offer appropriate accommodations for students with learning disabilities. Some programs offer extra time on exams as a standard accommodation (Barbaro, 1982; Bireley & Manley, 1980; Strichart & Mangrum, 1988); however, there is controversy surrounding this issue, and many programs serving students with learning disabilities offer few services beyond tutoring (Cordoni, 1980). Cordoni attributes the inconsistency in accommodations to a lack of research on college students with learning disabilities. Studies on the effects of extra exam time for college students with learning disabilities are now beginning to emerge. These studies should be critically evaluated to determine essential accommodations for these particular students. A study conducted by Hill (1984) showed that normally achieving college students perform better than college students with learning disabilities on the ACT and Nelson-Denny Reading Test (Brown, Bennett, & Hanna, 1981) under timed conditions. College students with learning disabilities perform just as well as normally achieving college students on the ACT (on all four of the ACT subtests: English, Math, Social Studies, and Natural Science) and Nelson-Denny Reading Test under untimed testing conditions. The Nelson-Denny vocabulary subtest was the only test where the normally achieving college students performed better in both the timed and untimed situations than did college students with learning disabilities. In a more recent study, Halla (1988) concluded that there are no significant Journal of Learning Disabilities
differences in test performance between cock & Johnson, 1977), the Wide Range students with learning disabilities and Achievement Test-Revised (WRAT-R) normally achieving, upper-level under- (Jastak & Wilkinson, 1984), and other graduate, graduate, and professional achievement tests deemed appropriate by school adults on the Graduate Record the examiner. The students with learning Exam (GRE) General Test and Nelson- disabilities met the university criteria of Denny Reading Test, and that untimed 1.5 standard deviations between intellitesting conditions significantly influence gence and achievement and were receivtest performance on those two assess- ing services from the Disabled Students ments for both groups. Halla points out Program. All students with learning disthat the data may be interpreted differ- abilities had a long history of reading ently if IQs are taken into effect. The IQ problems and had developed various of the group with learning disabilities was techniques for compensating; however, a range above that of the normally all complained of slow reading rates and achieving group (1 standard deviation). having to reread material several times The group with LD could be described for maximum comprehension. Case hisas highly intelligent but, in terms of tories on 14 of the students with LD inachievement, functioning at the mean dicated inattention (short attention level of the other group. spans) and impulsivity during their early The purpose of the present study was school years. The students with learning to evaluate reading comprehension scores disabilities were interviewed by the inof adults with and without learning dis- vestigator. All students with LD reported abilities under timed and untimed con- problems with concentration when readditions. The study offers evidence that ing, attributing the lack of attention and students with learning disabilities score concentration to their slow, laborious significantly lower than their normally rate of reading. Most reported that they achieving peers under time constraints, could concentrate and read for approxbut that there is not a significant dif- imately 15 to 20 minutes, and then had ference between students with learning to take a short break. disabilities and normally achieving stuThe other 15 participants were normaldents under untimed conditions. The per- ly achieving (NA) University of Califorformance of normally achieving students nia students who had never been identiunder timed and untimed conditions is fied as having learning disabilities, speech examined, and there is not a significant problems, reading problems, or spelling difference in performance. Amount of problems, and were not on academic extra time is discussed, as evidence that probation (reported on a questionnaire). students with learning disabilities are a They were solicited randomly by making heterogeneous group and setting fixed announcements in various classes at the amounts of extra time for all students university. may not be appropriate. The two groups selected were controlled for ethnicity, gender, age, and total Scholastic Aptitude Test (SAT) (Educational Testing Service, 1986) scores. All METHOD participants in the sample were white: 16 were males and 15 were females. The Subjects mean age of the group with learning disThe participants in this study were stu- abilities was 21.3 years, ranging from 18 dents at the University of California. to 27 years. The mean age of the norThirty-one students participated in the mally achieving group was 22 years, study, 16 of whom were identified as ranging from 19 to 27. Fifteen of the having learning disabilities (LD). Prior students with learning disabilities were to entering the university, the students undergraduates and one was a graduate with learning disabilities had been tested student. Thirteen of the normally achievand identified through the use of a com- ing students were undergraduates and prehensive test battery, including the two were graduate students. Wechsler Adult Intelligence Scale-Revised The mean SAT score for students with (WAIS-R) and/or the Woodcock-Johnson learning disabilities admitted to the uniPsycho-Educational Test Battery (Wood- versity in 1987-88 was 1125, with a stanVolume 24, Number 2, February 1991
dard deviation of 185. A cutoff point was established at 1 SD above and below the mean, establishing a range from 940 to 1310. For this study, students selected had SAT scores within the range of 940 to 1310. The students with learning disabilities were selected at random, regardless of whether they took the SATs under timed or extra-time conditions. Normally achieving students met the same criteria for SAT scores. The SAT scores tended to cluster at the higher end of the range, with only one student scoring below 1000.
Instrument The Nelson-Denny Reading Test was used. This instrument has two statistically equated forms, E and F. The test contains a vocabulary section and a silent reading comprehension section. The first minute of the reading comprehension is used for obtaining a reading rate. Sentences in the first reading passage have numbers to designate words per minute. At the end of 1 minute, the numbers (designating the amount of text read) are recorded and converted to a percentile rank according to the manual conversion charts. The Reading Comprehension section and the Reading Rate sections were used. The Reading Comprehension section contains eight reading passages and a total of 36 questions. The test is designed to be completed in 20 minutes, with college students generally obtaining their maximum score within the time limit. A measure of context dependence (i.e., the degree to which questions can be answered only by reading the passage) is provided. The manual reports that both Forms E and F are based on content more than on prior knowledge. The 36 items on each form can be divided into groups of 18 interpretive items and 18 primarily literal items. Comprehension is based on the degree of abstract information that a student must extract from the reading passage to answer the questions.
Procedure The participants were individually administered the Nelson-Denny Reading Comprehension Test by the investigator, and a reading rate was obtained for each 105
participant. Form E was administered to all students except two, who had been administered the test during the previous year. Those two students were administered the alternate, Form F. All students were told that this was a reading comprehension test and that the first minute would be used to obtain a reading rate. Although this was a timed test, the students were not told how much time they had to complete the test. They were told to read normally, as they would in a test situation, and that they were to work quickly and answer questions until they were told to stop. Two measurements were obtained on each subject—one score at the end of 20 minutes and another when the test was completed. The score remained the same for both measurements if students completed the test within the 20-minute limit. At the end of 20 minutes the participants were told to mark where they were on the answer sheet (if they had not completed the test) and to continue until they were told to stop (they were allowed time to complete the test). The amount of extra time needed to complete the test was recorded. Raw scores for words per minute were converted into percentile ranks according to the Nelson-Denny conversion charts. Scores under timed conditions and untimed conditions were converted to percentile ranks and used for comparison of the two groups.
TABLE 1 Comparison off Mean Percentile Scores for Comprehension Under Timed and Extra-Time Conditions (20 minutes) Group
106
Extra time (Mean)
Paired f-value
Probability (two-tailed)
.1281 .8227
.7587 .8347
-11.73 -1.19
.000 .253
LD NA
Note. LD = learning disabilities; NA = normally achieving.
• LD./Timed
L.D. and N.A. Timed and Extra-Time • L.D. /Extra-Time •N.A./Timed
ON.A./Extra-Time
CO
LU UJ LU
0 Figure 1.
2
4
6
8 10 12 14 16 18 SUBJECT Learning disabled and normally achieving students under timed and extra-time conditions.
TABLE 2 Comparison of Mean Percentile Scores for Normally Achieving Groups (20 minutes) Groups Variable Reading Rate Comprehension (timed) Comprehension (extra time)
RESULTS The effects of extra time for the two groups are given in Table 1 and Figure 1. The paired t tests indicate a significantly higher mean percentile rank comprehension score for subjects with learning disabilities when given extra time (/?= .000). For normally achieving subjects, there was no significant effect of extra time (p= .253). Table 2 shows the comparison of mean percentiles for the normally achieving group and the group with LD for each of the variables measured. Reading rate was significantly higher for normally achieving students compared to that of students with learning disabilities (p = .000). The same result was obtained for the Nelson-Denny Comprehension score under timed conditions (p = .000). How-
Timed (Mean)
NA Mean SD
LD Mean SD
.5953
.247
.1069
.8227
.115
.8347
.124
i value8
Probability (two-tailed)
.116
6.98
.000
.1281
.106
17.46
.000
.7587
.203
1.26
.218
Note. NA = normally achieving; LD = learning disabilities. Separate variance estimate.
a
ever, under extra-time conditions, there was no significant difference for the comprehension score between the normally achieving students and students with learning disabilities (/?= .218). When scores of the normally achieving subjects under timed conditions were compared to the scores of subjects with learning disabilities under extra time, again there was no significant difference between the two groups for the comprehension score 0? = .288). The tests were administered individual-
ly, allowing the investigator to observe the students. The students with learning disabilities did not complete the reading sections within the 20-minute time limit. The questions to which they responded were generally correct; however, they had responded to fewer questions under the time limit. Thirteen of the normally achieving students completed the comprehension sections within the 20 minutes. Two subjects needed extra time, completing the test with an extra 3 and 4 minutes, respectively. Journal of Learning Disabilities
DISCUSSION In this study, the following hypotheses were tested: (a) Students with learning disabilities will produce significantly lower scores than normally achieving students as measured by the Nelson-Denny Reading Comprehension Test under timed conditions; (b) both groups will increase their reading comprehension scores when tested under extra-time conditions, but the groups will not differ significantly in the untimed testing condition; and (c) the scores for the students with learning disabilities, when given extra time, will not differ significantly from those of the normally achieving students under timed conditions. Results of the comparisons between university students with learning disabilities and normally achieving university students support the hypothesis that students with learning disabilities score significantly lower than normally achieving individuals under timed conditions on reading comprehension. The study supports the hypothesis that students with learning disabilities perform as well as their normally achieving peers when given extra time, and that normally achieving students did not perform significantly better when given extra time. All but two of the normally achieving students finished in the 20 minutes. Their scores did not significantly increase the performance of the group, and none scored a perfect score. The extra time allowed the students with learning disabilities to respond to all questions. Of importance was the amount of extra time required for each individual student. The extra time required by students with learning disabilities varied a good deal, ranging from 4 minutes to 29 minutes. Students were not allowed to go back and change answers in the extra time; they were allowed only to finish the test. This lends support to the idea that students with learning disabilities have varied rates of processing printed information, and therefore a fixed amount of extra time for all students with learning disabilities on standardized tests may not be appropriate. The presentfindingscan be interpreted in the following ways: Normally achieving students often perform up to their capability under timed testing conditions Volume 24, Number 2, February 1991
and have little room for improvement. Students with learning disabilities often cannot perform up to their capabilities under timed conditions, and show considerable improvement when allowed extra time. Hill (1984), using a large sample and additional tests (ACT and Nelson-Denny Reading and Vocabulary Tests), interpreted his findings in the same way. All students with learning disabilities reported having to reread the passages several times to gain maximum comprehension. This suggests that these students may experience problems with attention span and concentration while reading. LeBerge and Samuels (1974) developed a theory on automaticity of information processing in which the primary component is attention. For the disabled reader, information at the visual and phonological levels is not processed automatically. A major portion of attention is needed for decoding and processing, leaving little to focus at the semantics level, thus reducing understanding and necessitating rereading a text for comprehension. Koppell (1979) suggested that attentional deficits may not necessarily be causes of reading disability but may result from a poor reader's previous failure experiences or anxiety about past performance. Senf (1976) agrees that attention deficits may not be primary disorders but may be secondary to previous problems in organizing and processing information. While the relationship between learning disabilities and attention disorder remains under investigation, it is clear that many students have both problems; often, however, one or the other remains undiagnosed. Standardized (timed) tests are "speed tests"—test in which individual differences in performance depend on speed of performance. The time limit on these tests is such that some individuals do not complete all the items, especially those persons with learning disabilities. "Power tests," on the other hand, have time limits long enough to permit most people to attempt all items on the test. Speed of performance is not an important factor on power tests (Anastasi, 1982). This is a critical issue, because previous research supports the concept that students with learning disabilities do not process information at a normal rate (Speece, 1987;
Rudel, 1981), and this slower processing may account for the difference in their performance on timed and untimed tests.
LIMITATIONS AND IMPLICATIONS The present study has several limitations, one of which is the small number of subjects involved in the research. Another limitation is that all the students attended the University of California on one campus, and therefore results may not be representative of all college students with learning disabilities. Growing numbers of students with learning disabilities are seeking admittance to colleges and universities (Strichart & Mangrum, 1988), but there is a paucity of research on the impact of timed exams on the performance of students with such disabilities. With the current controversy on providing extra exam time for students with learning disabilities in the postsecondary setting, there is a need for additional research.
ABOUT THE
AUTHOR
M. Kay Runyan is a doctoral candidate in the Joint Doctoral Program in Special Education at the University of California at Berkeley and San Francisco State University. She is coordinator of services to students with learning disabilities at the University of California at Berkeley. Address M. Kay Runyan, Disabled Students Program, University of California at Berkeley, Berkeley, CA 94720.
AUTHOR'S
NOTES
1. This research is a pilot study for a doctoral dissertation and was first presented at The Next Step Symposium, Harvard College, November 1989. 2. The research was supported in part by the University of California, Berkeley, in the form of an A&PS Development award. I gratefully acknowledge the assistance ofNori Ison as a reviewer of the statistics.
REFERENCES ACT. (1981). Assessment supervisors' manual of instruction for special testing. Iowa City, I A: American College Testing Program. Anatasi, A. (1982). Psychological testing (5th ed.). New York: Macmillan. Barbaro, F. (1982). The learning disabled college student: Some considerations in settling objectives. Journal of Learning Disabilities, 10, 599-603. Bireley, M., & Manley, E. (1980). The learning 107
disabled student in a college environment: A report of Wright State University's Program. Journal of Learning Disabilities, 13, 12-15. Brown, J.I., Bennett, J.M., & Hanna, G. (1981). Nelson-Denny reading test manual (Forms E and F). Boston: Riverside. Cordoni, B.K. (1980). College programs for learning disabled students. Perceptions, 3, 1-5. Cordoni, B.K., O'Donnell, J.P., Ramaniak, N.V., Kurtz, T., & Rosenshine, K. (1981). Wechsler Adult Intelligence Score patterns for learning disabled young adults. Journal of Learning Disabilities, 14, 404-407. Denckla, M.B., & Rudel, R. (1976). Rapid "Automatized" Naming (R.A.N.): Dyslexia differentiated from other learning disabilities. Neuropsychologia, 14, 471-479. Fraunheim, J.G., & Heckerl, J.B. (1983). A longitudinal study of psychological and achievement test performance in severe dyslexic adults. Journal of Learning Disabilities, 16, 339-346. Halla, J. (1988). GRE test scores of LD and non-
LD adults. Unpublished doctoral dissertation, ing disabilities. In S.R. Forness, K.A. Kavale, & Texas Technical University, Lubbock. M. Bender (Eds.), Handbook of learning disabilHill, G. (1984). Learning disabled college students: ities, Vol. 1, dimensions and diagnosis (p. 142). Austin, TX: PRO-ED. Assessment of academic aptitude. Unpublished doctoral dissertation, Texas Technical University, Speece, D.L. (1987). Information processing subtypes of learning disabled readers. Training DisLubbock. abilities Research 2(2), 91-102. Jastak, S., & Wilkinson, G. (1984). Wide range achievement test-Revised. Los Angeles: Western Stanovich, K.E. (1986). Cognitive processes and the Psychological Services. reading problems of learning-disabled children: Koppell, S. (1979). Testing the attentional deficit Evaluating the assumption of specificity. In J.K. notion. Journal of Learning Disabilities, 12, Torgesen & B.Y.L. Wong (Eds.), Psychological 43-48. and educational perspectives on learning disLeBerge, D., & Samuels, S.J. (1974). Toward a abilities (pp. 87-131). New York: Academic Press. theory of automatic information processing in Strichart, S., & Mangrum, C. (1988). College and reading. In S.R. Forness, K.A. Kavale, & M. the learning disabled student. Philadelphia: Grune Bender (Eds.), Handbook of learning disabilities, & Stratton. Vol. 1, dimensions and diagnosis (p. 140). Austin, Wechsler, D. (1981). Wechsler adult intelligence TX: PRO-ED. scale-Revised. Cleveland, OH: Psychological Rudel, R.G. (1981). Residual effects of childhood Corp. reading disabilities. Bulletin of the Orton Society, Woodcock, R., & Johnson, M.B. (1977). Wood31, 89-102. cock-Johnson psycho-educational test battery. Senf, G.M. (1976). Future research needs in learnAllen, TX: DLM Teaching Resource.
Tools to help treat Attention Deficit Disorder These PRO-ED products offer the indepth information you need to help students with ADD, their parents, and their teachers.
Attention Deficit Disorder This book and videotape by Dale R. Jordan will help you distinguish the four types of ADD, identify ADD behavior patterns, and foster the development of appropriate behaviors. #1457 Book $11.00 #1458 Videotape (VHS) $98.00
Attention Deficit Disorder and Hyperactivity, Second Edition This new revision, written by Ronald J. Friedman and Guy T. Doyal for parents and teachers, presents current research findings and clinical experiences with nearly a thousand children with ADD and their families. #3547 $11.00
108
Identification and Treatment of Attention Deficit Disorder This new handbook by Nancy Nussbaum and Erin Bigler contains practical information based on research for parents, teachers, and professionals. #1506 $8.00
Help! This Kid's Driving Me Crazy! The videotape by Jane Judy, Lynne Adkins, and Janis Cady (supported by a 64-page booklet) will help parents and professionals work more effectively with young children with ADD. #3548 Book #3550 Videotape (VHS)
$ 5.00 $89.00
ORDER NOW! For more information or a FREE catalog, write or fax today!
pro-ed
8700 Shoal Creek Boulevard Austin, Texas 78758 Office: 512/451-3246 Fax: 512/451-8542
Journal of Learning Disabilities
