Download Gray Oral Reading Test-4 (GORT-4) and more Lecture notes Learning disability in PDF only on Docsity!
Test Review: Gray Oral Reading Test-4 (GORT-4)
Name of Test: Gray Oral Reading Test-4 (GORT-4) Author(s): Wiederholt, J. Lee and Bryant, Brian R. Publisher/Year: Pro-Ed 1967, 1986, 1992, 2001 Forms: Form A and Form B Age Range: 6 years, 0 months to 18 years, 11 months Norming Sample:
GORT-4 used a test administrator recruitment procedure common to other Pro-Ed tests in which the customer database was used to invite potential examiners to participate. At the four major test sites (one in each geographic region), site coordinators supervised the administration of tests to students. At these sites, the authors state that “several hundred children in Grades 1 through 12” (Wiederholt & Bryant, 2001, p. 49) participated in Fall, 1999 and Fall, 2000.
Authors report that students with disabilities were those who attended regular rather than special education classes. The sample was stratified by ages and region, gender, race, ethnicity, parental income, and parental education. This data was presented in table form (Wiederholt & Bryant, 2001, Table 4.2, pp. 53-55). The percentage of the sample compares favourably to the national school age percentages taken from the 1997 Statistical Abstract of the United States: The National Data Book by the U.S. Bureau of Census.
Total Number: 1 677 students Number and Age: 6 years (n=87) to 18 years (n-72) with 12 month intervals for groups Location: 28 states (and the District of Columbia in four major geographic regions) Demographics: Age, race (White, Black, Other), ethnicity (Native American, African American, Hispanic, Asian, European American and Other), gender, urban vs. rural, geographic region (NW, Midwest, West, and South), family income (under 15,000 to over 75, 000) and educational attainment of parents (less than bachelor’s degree, bachelor’s degree, and master’s/professional/doctoral degrees) and disability status (no disability, learning disability, speech/language disorder, mental retardation, and other) were reported. The authors stratified the sample on these characteristics. Rural/Urban: yes SES: by family income Other: no Summary Prepared By: Eleanor Stewart 18 October 2007
Test Description/Overview:
The test kit consists of the examiner’s manual, the student workbook containing both Forms A and B, and Form A and Form B record forms. The kit is contained in a sturdy cardboard box. The examiner needs a stopwatch.
The preface provides an interesting history of Dr. William Gray and his career in reading assessment and research. His influence is acknowledged with the test name, Gray Oral Reading Test, a testament to his original work in developing the present day edition. Each edition of the GORT is described in order for the reader to appreciate the progression toward the revisions made for the fourth edition. Comment from the Buros reviewer: “The GORT-3 was well received, but criticism of the technical aspects of the test was made. These areas of criticism involved documentation of the equivalency of Forms A and B, need for more extensive validity data for assessing student progress in reading, socioeconomic description of students in the standardization sample, and the normative data being 15 years old” (Crumpton, & Miller Whitehead, 2003, p. 418).
Improvements included in the 2001 edition of the GORT-4, are listed on page x (Wiederholt & Bryant, 2001). Changes include: new normative data comparable to the U.S. population, normative sample stratification, bias is addressed, additional reliability is calculated for subsamples, validity studies addressing new normative sample as well as subgroups are included, and clear evidence of form equivalence.
In the introductory chapter, the authors describe the nature of reading and the purpose of the GORT-4. They begin with the statement, “The nature of reading is complex and not completely understood” (Wiederholt & Bryant, 2001, p. 1). They outline a brief history of how reading has been conceptualized first as “thought getting” by Huey in 1908 followed by Gray’s proposition that “reading assumes not only fluent, accurate recognition of words, but in addition, a clear grasp of the meanings implied by the symbols used” (Quote in text, p. 1). The authors state that theorists hold three broad views of how reading develops. These views are described as: bottom-up, top-down, and interactive theories. They briefly outline the central tenets of each theory. They then introduce the view that certain skills may be more important at certain stages of reading development with a brief comment on phonemic skills. Finally, they re-state Gray’s view as follows, “Regardless of how one learns to read, the ultimate goal of reading instruction is the same-that is, the fluent recognition of words and a clear grasp of the meaning implied by the symbols used. The GORT-4 is a measure of this ultimate goal” (p. 2).
Comment: I was surprised at how short this introduction was at only 1.5 pages. Also, the brief statement above is the only reference
proficiency, to identify strengths and weaknesses, to document progress in intervention, and to use in research on reading. Each of these purposes is briefly described in the manual (Wiederholt & Bryant, 2001, pp. 4-5).
Areas Tested: Reading Comprehension Other: fluency, accuracy, rate
Who can Administer: A range of individuals such as teachers, special educators, and speech pathologists are listed as potential testers. The authors state that whoever uses the GORT-4 should have formal training in test administration and interpretation and suggest that such training is available through colleges and “local agencies or private consultants” (Wiederholt & Bryant, 2001, p. 7) however, they note that testers should consult schools and professional organizations if the purpose of testing involves the diagnosis of disabilities or qualification for special programs. The authors provide five general rules to make test administration more reliable. These are familiar injunctions about familiarizing oneself with the test prior to administering it, creating an appropriate test environment, etc.
Administration Time: The manual indicates 15 to 45 minutes (Wiederholt & Bryant, 2001, p. 10). Though testing can be completed in one session, two sessions are also allowed if, for example, the student becomes fatigued.
Test Administration (General and Subtests):
Chapter 2 presents the administration and scoring information. Entry points are identified by grade and starting point story with the caveat that testing should begin at the child’s grade level unless the examiner has prior knowledge that the child may be a poor reader in which case testing may begin at a lower level. For the Comprehension subtests, basal and ceiling rules are identified with three examples provided to assist the examiner. If a basal is not initially established, the examiner is instructed to proceed to ceiling and then return to establish five correct responses to comprehension questions. Form A and B use the same basal and ceiling rules. The examiner scores the student’s response as correct (1 point) or incorrect (0 point) and sums the correct responses for each story which is transferred to Section II on the record form. A completed example of the form is presented in Figure 2.2 (Wiederholt & Bryant, 2001, p. 12). For the Fluency score, the examiner is concerned with timing the reader as the passage is read to completion. Then the time taken (in seconds) is recorded in the record booklet. Time intervals are converted to values on a 0 to 5 point Rate score. Similarly, accuracy is recorded as the number of deviations from the print and converted to a 0 to 5 point Accuracy Score. The Fluency score is the sum of the Rate and Accuracy Scores. The Fluency score determines the basal as a score of 9 or 10 for a story
and ceiling as a score of 2 or less. Figure 2.3 illustrates the procedure for establishing basal and ceiling scores in this manner (Wiederholt & Bryant, 2001, p. 14).
Chapter 2 continues with specific administration and scoring instructions (Wiederholt & Bryant, 2001, pp. 15-26) in which prompting and timing, noting deviations, determining scores, and identifying and calculating miscues is discussed. The examiner’s instructions to the student are printed in red type in this section. Examples of each marking system are provided in Figures 2.5 and 2.6 (i.e., slash marking and miscue marking). The authors explain that they chose to present a simple marking system though they are aware that some examiners may choose to separate graphic and phonemic miscues. They state that the examiner must analyze at least two different stories with a minimum of 25 miscues across the two stories. The student’s miscues are classified according to the outline in Section V of the record booklet. An example is illustrated in Figure 2.7 (p. 22). Five categories of miscues are used as follows: meaning similarity, function similarity, graphic/phonemic similarity, multiple sources, and self-correction (each is described in Chapter 1; and specific examples are provided in this chapter for purpose of assisting the reader to learn the system). The student’s miscues per category are then totaled and a percentage is calculated for all miscue categories. The authors allow that other types of miscues may be produced. These miscues should be noted along with general impressions and observations of the student’s behaviour and performance during the reading tasks. Space is provided on the record booklet for this purpose.
Comment: The miscue schema is easy to follow and use. I think that it may be easier for someone who works in linguistics or related area. Novice examiners may have some difficulty with this test. Also, I wonder if subjective judgment enters into scoring, especially in assessing meaning elements in miscuing. I think that examiners should be proficient English language users to use this system accurately and fairly.
Test Interpretation:
Interpreting the results of the standardized scores is explained beginning with the Total Scores which the authors state have no clinical value in themselves but are used as the basis for calculating the standard scores, percentiles, and age and grade equivalents. The subtest standard scores are based on 100 +/-3. The Oral Reading Quotient has a base of 100 with standard deviation of 15. Percentiles are explained and follow the norm for this type of score. Age and grade equivalents are based on the average performance of the normative group. The authors state that these scores were developed “through interpolation, extrapolation, and smoothing…” (Wiederholt & Bryant, 2001, p. 36). They acknowledge the well-known APA concerns but note that examiners will have to align their results with employer requirements and legislative requirements. Therefore, the GORT-4 authors deliver these scores “reluctantly”. They encourage examiners to read the cautions regarding these scores and cite references to Aiken (2000), Anastasi
chapter on interpreting results of the GORT-4, the authors discuss additional assessment and instruction.
Standardization: Age equivalent scores Grade equivalent scores Percentiles Standard scores (mean 10 , SD 3) Stanines Other: Composite (from standard scores) and Conversion Scores (for rate and accuracy) These conversions are made when the time taken to read the passage and the number of deviations are converted into number values from 0 to 5 that then are added to lend themselves to the overall Fluency Score which ranges from 0 to 10. The Fluency Score is said to equally emphasize both rate and accuracy aspects. The procedure for developing the conversion scores from a sample of the normative data is presented on page 57. Basically, the authors divided the performance into quarters with the top 25% representing the best readers receiving a rating of 5, followed by the next group and so on down to the bottom values that represented the performance of the students 6% who attained a value converted to 0.
The authors comment about age and grade equivalent scores and the concerns raised about their use in the chapter on interpreting scores (Wiederholt & Bryant, 2001, p. 37). They state their position as follows, “We prefer you use GORT-4 standard scores, quotients, and percentiles rather than age and grade equivalents when reporting results to parents and other professionals” (p. 37). They cite several references including Aiken (2000), and Anastasi and Urbina (1997). In addition to the use of norm-referenced scores, the authors present a procedure for conducting a discrepancy analysis.
Reliability
Internal consistency of items: Cronbach’s coefficient alphas were computed for the entire norming sample by age groups. Large value alphas were reported for the subtests and composite (ORQ) ranging from .91 to .97.
Test-retest: Forty-nine students in elementary (n=30), middle school (n=10), and high school (n=9) participated in retesting. The students, 22 males and 27 females, ranged in age from 6 years to 18 years. The majority, 78% were European Americans with participation from 16% African American, and 2% Other. The interval between tests was “approximately two weeks” (Wiederholt & Bryant, 2001, p. 66). Correlation coefficients were corrected for restricted sample range where needed. The results indicated a range from .85 to .95 for the four subtests and ORQ.
Inter-rater: Thirty randomly selected completed test records were independently scored by two Pro-Ed staff. The results indicated inter-rater reliability from .94 to .99.
Other: none
Validity: The extensively detailed chapter on validity is not surprising given the long history and research related to the test.
Content: The authors note that using stories with corresponding comprehension questions was the format originally chosen by Gray. The story format continues to be a practical and logical way to assess reading abilities as they are used extensively in the students’ classroom experiences. Other reading tests also use the story format. Therefore, face validity is established. The composition of the stories reflected general interest, controlled syntactic structure, and age and grade appropriate vocabulary. The authors describe the specific procedures and materials they consulted in this process. They note that the stories for all but the youngest children remain unchanged from earlier editions. The only new story is the one provided for the entry level where emerging oral reading ability is assessed.
Comment: This is a credible process, a number of sources were consulted for vocabulary in particular.
In terms of developing the comprehension questions, the authors describe both the types of questions used and the research to support their choice of comprehension questions (i.e., literal, inferential, critical, and affective). Also, vocabulary was controlled so that “the vocabulary would not be more difficult than that used in the stories” (Wiederholt & Bryant, 2001, p. 75).
Reading fluency has received substantial research attention according the authors. In particular, they note that “poor fluent word recognition is the most frequently cited factor in reading disorders” (Wiederholt & Bryant, 2001, p. 75). The GORT-4 combines rate and accuracy to reflect the evidence indicating that good readers are proficient at both of these aspects.
The sequence of the stories is arranged in ascending order of difficulty. The controversy over readability formulas is stated with references to the literature. The authors, who wished to provide some indication of readability, chose the Flesch-Kincaid Readability Formula to do so. Table 6.1 (Wiederholt & Bryant, 2001) provides the results for the stories. However, the authors also used the data from the normative sample to determine the final order of stories.
In terms of deviations from print, the term preferred by the authors, a synopsis of the research history on reading errors and classifications is given. On the basis of the research, the authors decided to use their system of five types of miscues described elsewhere in this review. Their system, they comment, is similar to others currently in use (references are given).
Relationship to tests of intelligence or aptitude: The GORT-4 was correlated as follows: WISC-III (individual scales and full scale) ranged from .40 and .67, Scholastic Aptitude Scale ranged from .34 to .83, Short Form Test of Academic Aptitude ranged from .26-.50, WISC-R (individual scales and full scale) ranged from .43 to .80, and Median correlations for all studies ranged from .45 to .62. Relationship to RAN: Three studies report the relationship of RAN to GORT-4. These studies found median correlation coefficients to be .49 for Fluency and .48 for Comprehension. The authors state that these results compare favourably with the .41 reported in the RAN literature. Group differentiation Using the mean standard scores for the normative sample and subgroups, the predicted outcomes were demonstrated in regard to gender, race/ethnicity (European Americans, African Americans, and Hispanic Americans), and disability (learning disability, serious emotional disturbance, and attention-deficit/hyperactivity disorder). (N.B. The sample for serious emotional disturbance was small n=15). The learning disability group performed Below Average whereas the gifted children were Above Average as expected. All results are presented in Table 6.11. Additionally, other studies reported significant differences between the groups of students studied (Wiederholt & Bryant, 2001, pp. 94-95). Each study is briefly described in terms of the sample, sample size, and the comparisons made. Change over time: Intervention studies reported the ability of GORT-4 to detect change due to intervention. These studies are summarized both in the text and in Table 6.13. The authors state that, “all three studies reported in this section support this use of the GORT-4 and clearly provide convincing evidence of the GORT-4’s construct identification validity” (Wiederholt & Bryant, 2001, p. 96). Item validity: The discriminating powers previously reported in Table 6.2 provide evidence of construct-identification validity. The authors note that these results are replications of those found in earlier studies using the GORT-R.
Differential Item Functioning: Swaminathan and Rogers’s (1990) logistic regression procedure was used to detect DIF. The authors explain their rationale and procedures undertaken. The results of 560 comparisons found that nine were significant at the .001 level. Effect size was investigated and none of the nine comparisons indicated moderate or large effect sizes. On the basis of this evidence the authors conclude that “the test is unbiased in regard to gender, race, and ethnicity” (Wiederholt & Bryant, 2001, p. 82).
Other: none Summary/Conclusions/Observations: This is a comprehensive and well-researched test, attesting to its continued relevance and
use.
Clinical/Diagnostic Usefulness: On the basis of this review, I would use this test for its intended purposes trusting that it would yield important and clinically useful information. I was particularly interested in the miscue analysis as I think that clinicians would use this information as a beginning point for exploring the individual student’s approach to reading.
References
Aiken, L. R. (2000). Psychological testing and assessment (8th^ ed.). Needham Heights, MA: MIT Press.
Anastasi, A., & Urbina, S. (1997). Psychological testing (7th^ ed.). Upper SaddleRiver, NK: Prentice-Hall.
Catts, H. (1993). The relationship between speech-language impairments and reading disabilities. Journal of Speech and Hearing Research, 36 , 948-958.
Cornwall, A. (1992). The relationship of phonological awareness, rapid naming, and verbal memory to severe reading and spelling disability. Journal of Learning Disabilities, 25 (8), 532-538.
Crumpton, N.L., & Miller Whitehead, M. (2003). Test review of the Gray Oral Reading Test, Fourth Edition. In B. S. Plake, J.C. Impara, and R.A. Spies (Eds.), The fifteenth mental measurement yearbook (pp. 417-421). Lincoln, NE: Buros Institute of Mental Measurements.
Kolen, M. J., & Brennan, R. L. (1995). Test equating methods and practices. New York: Springer.
Linn, R. L., & Gronlund, N. E. (1995). Measurement and evaluation in teaching (7th^ ed.). Upper Saddle River, NJ: Merrill.
MacEachron, A. E. (1982). Basic statistics in the human sciences. Austin, TX: Pro-Ed.
