Beginning First-Grade Screens.
Table 4 shows results of screening conducted in first-grade to predict ending first-grade performance. Findings from the different studies are hard to compare because of design differences--e.g., the definition of unsatisfactory reading outcomes (i.e. reading disabled vs. reading skills in the lowest quartile), the method for determining screening cut-points (post hoc vs. arbitrary cuts), and the proportion of the population considered at risk (i.e., ranging from 17% to 48%). As before, higher sensitivity was reported for screens that combined phonological and letter knowledge measures and that used post hoc procedures to identify cut-scores on the screening instrument. Based on the classification studies, the best candidates for screening measures appear to be a combination of LNF, phoneme blending, and sound repetition (O'Connor & Jenkins, 1999) and a combination of phoneme blending, LN/S, and word reading (Foorman et al., 1998).
| Measure/Study | Sample | Type of Evidence | Result | |||
| Letter-Sound Fluency (LSF) Speece and Case (2001) |
142 Fall Grade 1 | Classification | At Risk 25% on LNF |
Sensitivity 55.9 |
Specificity 83.7 |
Criterion Measure Dually (Level and Slope) Discrepant (-1 Standard Deviation) on CBM-ORF |
| Combination of: (1) Letter Naming Fluency (2) Phoneme Segmentation (3) Sound Repetition O'Connor and Jenkins (1999) |
215 Oct. Grade 1 | At Risk 17% |
Sensitivity 100% |
Specificity 87% |
Criterion Measure Same |
|
| TRPI Combination of: (1) Word Reading (2) Blending Phonemes |
599 Fall Grade 1 | Classification | At Risk 48% |
Sensitivity 93% |
Specificity 63% |
Criterion WJ-Broad Reading - Grade 1 |
| DIBELS-NWF Fuchs et. al. (2003) |
151 Fall and Spring Grade 1 at-risk (defined by Letter Naming Fluency) | Concurrent Validity | ||||
| Fall Grade 1 | .58 WRMT-R Word ID .50 WRMT-R Word Attack |
|||||
| Spring Grade 1 | .64 WRMT-R Word ID .51 WRMT-R Word Attack .80 CRAB Fluency |
|||||
| Predictive Validity | ||||||
| Fall-Spring Grade 1 | .57 WRMT-R Word ID .46 WRMT-R Word Attack .64 CRAB Fluency |
|||||
| DIBELS-NWF (Good et al., 2001) |
342 Winter Grade 1 | Concurrent Validity | .36-.59 WJ-R Reading Readiness | |||
| Predictive Validity | ||||||
| Spring Grade 1 | .82 CBM-ORF | |||||
| Spring Grade 2 | .60 CBM-ORF | |||||
| Spring Grade 2 (?) | .66 WJ-Reading Cluster | |||||
| WIF (Fuchs et. al., 2003) |
151 Fall and Spring Grade 1 at-risk | Concurrent Validity | ||||
| Fall Grade 1 | .77 WRMT-R Word ID .59 WRMT-R Word Attack |
|||||
| Spring Grade 1 | .82 WRMT-R .52 WRMT-R .93 CRAB Fluency |
|||||
| Predictive Validity | ||||||
| Spring Grade 1 | .63 WRMT-R Word ID .45 WRMT-R Word Attack .80 CRAB Fluency |
|||||
The addition of word identification to the TPRI screen is notable (Foorman, et al., 1998), because it is the first time that word reading emerges as a sufficiently sensitive measure for discriminating risk groups. Fuchs et al. (2003) criterion validity study of at risk first-graders (Table 4) also bolsters the case for using word identification to screen. Their Word Identification Fluency measure (CBM-WIF) produced strong concurrent and predictive validity coefficients with WRMT-R Word Identification and Word Attack. Remarkably, fall CBM-WIF was a stronger predictor of spring WRMT-R Word ID than was fall WRMT-R Word ID (.63 vs. .49). Validity coefficients of the CBM-WIF surpassed those of DIBELS NWF on several measures for the same group of first-graders, suggesting that word identification fluency may be a better screening measure than either WRMT-R or DIBELS NWF for marking risk status of first graders. A comparative classification study could settle this question.
Previous Page | Next Page
(End of Kindergarten) | (End of First/Beginning of Second)
