210 likes | 528 Views
Closing the Literacy Achievement Gap with Early Intervention. Emily M. Rodgers Chuang Wang Francisco X. Gomez-Bellenge The Ohio State University AERA, San Diego, April 2004. The Problem.
E N D
Closing the Literacy Achievement Gap with Early Intervention Emily M. Rodgers Chuang Wang Francisco X. Gomez-Bellenge The Ohio State University AERA, San Diego, April 2004
The Problem Mounting statistical evidence continues to document that a literacy achievement gap exists between students along race/ethnicity and economic lines (Lee, 2002; West, Denton & Reaney, 2001; Zill & West, 2000)
Theoretical Frame • Research shows that certain demographic groups under perform academically relative to other groups (Lee, 2002) • Closing this gap will take more than a quick fix (Bainbridge & Lasley, 2002) • One-to-one teaching is widely used for children having reading difficulties (Wasik & Slavin, 1993) • Can one-to-one teaching make a difference to the achievement gap?
Research Question 1 Does a literacy achievement gap exist along race/ethnicity and economic lines within a random sample of first grade students?
Research Question 2 Do students who have had an opportunity for a full treatment of Reading Recovery close the literacy achievement gap along race/ethnicity and economic lines with a random sample of first grade students?
Research Question 3 Do students who have been successfully discontinued from Reading Recovery (a subset of the treatment group) close the literacy achievement gap along race/ethnicity and economic lines with a random sample of first grade students?
200 school districts and 536 schools in a Midwestern state Children served (n= 6,447) Treatment Group (n=4,764) Discontinued Group (n = 3,499) Comparison Group (n = 1,038) 2 students selected at random from each RR school Some were served by RR All Data collected in 2002-2003 Participants
Measure: Observation Survey (Clay, 2002) Concepts About Print (CAP) Range: 0 - 37 Cronbach’s alpha = .78 (Clay, 1993) Hearing and Recording Sounds in Words (HRSW) Range: 0 - 37 Cronbach’s alpha = .96 (Pinnell et al, 1994) Text Reading Level (TRL) Range: 0 - 30 Cronbach’s alpha = .83 Item separation reliability = .98
Experimental Design Independent t-tests were conducted to determine if fall and spring gaps for the three reading measures existed between relevant groups on each measure. The alpha level was pre-set at .05 and Bonferroni adjustments were used to adjust the p-values since these are pre-planned comparisons. A standardized effect size estimate (Cohen’s d) for each significant difference was calculated with the pooled standard deviation (Rosnow & Rosenthal, 1996).
Fall No significant difference on HRSW or TRL Significant difference for CAP (d =.48) Spring Significant differences on all 3 measures: CAP (d = .37) HRSW (d = .26) TRL (d = .32) Random Sample by Race
Fall Significant differences on all 3 measures: CAP (d = .70) HRSW (d = .70) TRL (d = .57) Spring Significant differences on all 3 measures: CAP (d = .46) HRSW (d = .40) TRL (d = .69) Random Sample by Economic Status
Fall Significant differences on all 3 measures: CAP (d = 1.05) HRSW (d = .96) TRL (d = .79) Spring No significant difference on CAP or HRSW Significant difference on TRL, d = .63) Discontinued Group versus Random Sample by Race
Fall Significant differences on all 3 measures: CAP (d = 1.12) HRSW (d = 1.51) TRL (d = .97) Spring No significant difference on CAP or HRSW Significant difference on TRL (d = .63) Discontinued Group versus Random Sample by Economic Status
Discussion • A gap exists in fall and is either maintained or opens further in the general population when groups are disaggregated by Race or EconomicStatus • Students served by Reading Recovery either close the gap or trend towards closing the gap
Closing the Gap • Reading Recovery spoils the predictions of failure for low achieving first grade students. • An effective intervention can have a statistically measurable impact, even on a large scale. • Group comparisons were made across two dimensions simultaneously; low achieving readers versus average readers and at-risk groups versus not at-risk groups.
Limitations & Future Studies • Only independent t-tests were used for comparison purposes • This is a study of RR only. No comparisons were made to other reading interventions • Need to investigate factors such as race/ethnicity, Economic Status, RR and school district that are related to students’ reading achievement with regression analysis or hierarchical linear modeling.