Intensive Summer Intervention Drives Linear Growth of Reading Skill in Struggling Readers

Abstract

A major achievement of reading research has been the development of effective intervention programs for struggling readers. Most intervention studies employ a pre-post design, to examine efficacy, but this precludes the study of growth curves over the course of the intervention program. Determining the time-course of improvement is essential for cost-effective, evidence-based decisions on the optimal intervention dosage. The goal of this study was to analyze reading growth curves during an intensive summer intervention program. A cohort of 31 children (6-12 years) with reading difficulties (N = 21 with dyslexia diagnosis) were enrolled in 160 h of intervention occurring over 8 weeks of summer vacation. We collected behavioral measures over 4 sessions assessing decoding, oral reading fluency, and comprehension. Mixed-effects modeling of longitudinal measurements revealed a linear dose-response relationship between hours of intervention and improvement in reading ability; there was significant linear growth on every measure of reading skill and none of the measures showed non-linear growth trajectories. Decoding skills showed substantial growth [Cohen's d = 0.85 (WJ Basic Reading Skills)], with fluency and comprehension growing more gradually [d = 0.41 (WJ Reading Fluency)]. These results highlight the opportunity to improve reading skills over an intensive, short-term summer intervention program, and the linear dose-response relationship between duration and gains enables educators to set reading level goals and design a treatment plan to achieve them.

Keywords: dyslexia; growth curves; literacy; response to intervention; summer intervention.

FIGURE 1

Schematic diagram of experimental design. Squares indicate experimental laboratory visits at the timepoints (T).

FIGURE 2

Significant growth across reading measures. (A) Mean growth of composite reading skills. Growth curves are plotted using the intercept and slope estimates from a linear mixed-effects model with session as a categorical variable. The dashed lines represent measurements during the baseline period. Results show growth across reading measures during the intervention period, and no change (or a decline) in scores during the baseline period. Error bars represent ± 1 SEM across participants. (B) Longitudinal growth of basic reading skills. Basic reading skills, measured by the Woodcock Johnson IV Basic Reading composite standard score, plotted for each individual child as a function of hours in the intervention. Participants completed up to 160 h of intervention. The bold line represents the linear fit based on a linear mixed-effects model (p = 3.53 × 10^–13). (C) Growth rates across reading measures. Bar heights depict growth in skills per hour of intervention estimated based on a linear mixed-effects model. Error bars depict standard errors from the linear mixed-effects model. (D) Hypothetical growth trajectories. When reading skills are measured at multiple time-points over the course of an intervention, we might observe different patterns of growth that would be detected by adding quadratic terms to the model. (E) Comparison of a non-linear model of reading growth trajectories. Coefficients for the quadratic effects with error bars representing ± 1 SEM across participants. These effects were not significant for any of the reading measures, confirming that growth is predominantly linear. (F) A hypothetical dose-response curve. Our findings of linear growth in reading skills, without any significant deviations from linearity, indicate that 160 h of Seeing Stars lives in the shaded gray area of the dose response curve. Code and data to reproduce each figure is available in the online repository (e.g., https://github.com/yeatmanlab/growthcurves_public/blob/master/figure2.m for code to reproduce the figure). The tests abbreviated include the Woodcock-Johnson IV Tests of Achievement (WJ) Letter-Word Identification (WJ LWID), Word Attack (WJ WA), Basic Reading Skills composite (WJ BRS), Oral Reading (WJ OR), Sentence Reading Fluency (WJ SRF), Reading Fluency composite (WJ RF), Math Facts Fluency (WJ MFF), Calculation (WJ CALC), Math Calculation Skills composite (WJ MCS), Test of Word Reading Efficiency (TOWRE) Sight Word Efficiency (TOWRE SWE), Phonemic Decoding Efficiency (TOWRE PDE), and composite index (TWRE INDEX).