Beyond Grade 2
Annual Achievement Tests as Screens. There are surprisingly few studies of screening measures beyond grade 2. Where it is possible, results form district- or state-wide annual achievement tests can be used to identify at risk students. This practice should result in reasonably good predictions, given that spring-spring and fall-spring achievement correlations are typically strong. For example, Jenkins & Jewell (1992) reported a correlation of .89 between Gates-McGinitie Reading Test (fall) and Metropolitan Achievement Test-Reading (spring).
CBM-Maze. Espin, Deno, Maryuma, and Cohen (1989) developed the Basic Academic Skills Samples (BASS) which includes a maze reading test designed to screen students at risk for failure. For the BASS, students read 3 passages (Spache readability = 2.3) with every 7th word deleted. For the missing words, students attempt to select the correct word from among 3 choices. Distracters are clearly incongruous with the context of the story. Each passage is scored for the number of words correctly restored in one min. The student's score is the median of the score of the three passages. For a sample of grade 3-5 students, Espin et al. reported a concurrent validity coefficient of .81 for the BASS-Maze, using CBM-ORF as the criterion test. Jenkins and Jewell (1992) examined the concurrent validity of the BASS Maze and its degree of overlap with the Gates-McGinitie Reading Test and the Metropolitan Achievement Test--Reading. Concurrent validity ranged from .63 to .78, depending on grade level and the achievement test used as the criterion (Table 6). On average there was approximately 60% overlap between the bottom 15% of students on BASS and the bottom scoring 15% on the achievement tests. Unfortunately, Jenkins and Jewell did not report sensitivity and specificity levels. Research is needed on the classification accuracy of CBM-Maze screens.
| Measure/Study | Sample | Type of Evidence | Result | |||
| CBM-ORF Stage & Jacobson, 2001 |
173 Sept Grade 4 | Classification | At Risk 32% |
Sensitivity 66% |
Specificity 76% |
Criterion Not meeting standard on state test |
| BASS-MAZE Jenkins and Jewell, 1991 |
322 Fall Grades 2-6 | Concurrent Validity | Gates-McGinitie Total Reading .65 Grade 2 .63 Grade 3 .75 Grade 4 .76 Grade 5 .72 Grade 6 |
|||
| 374 Spring Grades 1-6 | Concurrent Validity | Metropolitan Achievement Test Total Reading .78 Grade 1 .76 Grade 2 .66 Grade 3 .72 Grade 4 .72 Grade 5 .67 Grade 6 |
||||
| 322 Fall Grades 2-6 | Concurrent Percent Overlap |
Bottom 15% of students on BASS-MAZE and Gates-McGinitie Total Reading 57% Grade 2 75% Grade 3 54% Grade 4 50% Grade 5 65% Grade 6 |
||||
| 374 Spring Grades 1-6 | Concurrent Percent Overlap | Bottom 15% of students on BASS-MAZE and Metropolitan Achievement Test Total Reading 38% Grade 1 57% Grade 2 62% Grade 3 54% Grade 4 62% Grade 5 60% Grade 6 |
||||
CBM-ORF. Although studies too numerous to report have documented strong criterion validity for CBM-ORF, there is little information on the classification accuracy that can be achieved using ORF cut-points. In one of the few screening classification studies, Stage and Jacobson (2001) used ORF to screen beginning fourth-graders for unsatisfactory performance on the state-mandated standards based reading test. Using an ORF fall cut-score of 100 correct words, they reported low sensitivity and specificity (66% and 76%, respectively). In an analysis using an ORF of 50 words correct (the marker for the lowest scoring 10% of the sample), Stage and Jacobson reported sensitivity of 31% and specificity of 96%. More research is needed to determine if CBM-ORF can produce reasonable specificity levels once sensitivity requirements are satisfied.
Previous Page | Next Page
(End of First/Beginning of Second) | (Conclusion)
