Abstract
This study compared the reading growth of students with and without learning disabilities, and students with and without reading deficits in response to tier 2 reading interventions within a response-to-intervention framework. Participants were 499 second- and third-grade students in six urban schools. Students who scored at or below the 10th percentile on the fall reading screening assessment were identified as having a severe reading deficit and received a tier 2 reading intervention that was targeted to their needs. Results showed a significant effect between groups on reading growth. Students with severe reading deficits receiving targeted tier 2 intervention grew at a rate that equaled the rate of growth of students without reading deficits and was significantly higher than students who were receiving special education services for reading. Implications for practice, suggestions for future research, and study limitations are discussed.
A large number of students in K-12 schools demonstrate significant reading difficulties. The percentage of the nation’s fourth-grade students who scored in the proficient range in reading has increased slightly over the last decade, but students scoring below the basic level of proficiency remained at 37% (U.S. Department of Education [USDE], Institute of Education Sciences, National Center for Education Statistics, National Assessment of Educational Progress [NAEP], 2017). A basic proficiency level indicates partial mastery of prerequisite skills that are fundamental to proficient reading (USDE, Institute of Education Sciences, National Center for Education Statistics, NAEP, 2017). Thus, one third to approximately one half of fourth-grade students in the United States do not demonstrate mastery of even prerequisite skills needed to read proficiently. There have been multiple approaches to reduce the number of students with significant reading difficulties, two of which include special education services for students with reading disabilities, and response to intervention (RtI). We will discuss research for both of these intervention systems below.
Special Education
The number of students with a disability who receive special education services in U.S. schools has almost doubled since the provision of special education services was codified in federal law in 1975, and the number of students identified with a specific learning disability (SLD) has almost tripled (National Center for Educational Statistics, 2017). The federal mandate for special education services requires schools to provide specialized instruction so that students with disabilities can access the general curriculum, which should lead to enhanced career opportunities (USDE Office of Special Education Programs, 2006). Special education services are the most intensive intervention that schools can provide. However, approximately two thirds (68%) of fourth-grade students with disabilities scored below the basic level of proficiency in reading (USDE, Institute of Education Sciences, National Center for Education Statistics, NAEP, 2017) despite receiving special education services. Research also suggested that special education services do not necessarily result in improved outcomes for students with SLD (D. Fuchs et al., 2010; Kavale & Forness, 2000).
A propensity-score analysis of the Early Childhood Longitudinal Study-Kindergarten (ECLS-K) database found negligible or significantly negative effects of special education services on the reading skills of students with disabilities (Morgan et al., 2010). Furthermore, an examination of reading growth among students receiving special education services found that they exhibited less growth across the school year than students receiving general education instruction only (Christ et al., 2010; Deno et al., 2001; L. S. Fuchs et al., 1993). However, these studies only examined the average reading growth rate of all students in these two groups and did not designate students based on whether or not students had reading difficulties specifically.
There are multiple hypotheses for why special education services have not led to more positive student outcomes including that the services simply may not be particularly effective (Kavale & Forness, 2000). Another hypothesis is that there could be a lack of evidence-based approaches implemented in special education because special education teachers frequently reported using less effective interventions (Burns & Ysseldyke, 2009). Previous intervention research found improved student outcomes for students with SLD in response to such interventions (L. S. Fuchs et al., 2013; Lovett et al., 2013), but perhaps those intervention practices are not used frequently enough in schools. Alternatively, perhaps students receiving special education services are students with severe deficits that are difficult to remediate (D. Fuchs et al., 2010; D. Fuchs & Fuchs, 2014; Kavale et al., 1994). Finally, it could also be argued that there is no consistency in how SLD is identified in policy or practice (Maki et al., 2015; Miciak et al., 2014; Stuebing et al., 2012), and students identified with SLD are not the ones who would best benefit from special education services.
RtI
Many states are implementing RtI frameworks and other multitiered systems of support (MTSS) to better support struggling learners and to both prevent and identify SLD (Berkeley et al., 2009). The research regarding effects of RtI on student outcomes has been inconsistent. A statewide evaluation of RtI outcomes found decreases in the number of students identified with a SLD and in those with serious reading difficulties (i.e., scored below the 10th percentile on district-administered reading screeners; Torgesen, 2009). Meta-analytic research also found large effects for both systemic (e.g., reduced number of students in special education) and student outcomes (Burns et al., 2005). However, a large-scale federally funded evaluation of RtI found negligible or even negative effects of RtI for students (Balu et al., 2015). The results of the Balu et al. (2015) study may have been due to inconsistent implementation of the interventions and contagion of the control group because control group students also participated in tiered interventions (Shinn & Brown, 2016). Thus, educators should exercise caution in determining that tiered interventions lack effectiveness, especially given evidence supporting their effectiveness (Gersten et al., 2009; Vaughn et al., 2010). Given the conflicting evidence on the impact of tiered intervention, additional research examining student growth in response to tiered intervention is needed.
Small-group reading interventions are the RtI component for which there is the strongest evidence (Gersten et al., 2009). Tiered interventions for reading were more effective if they directly targeted specific student reading needs in phonemic awareness (Ehri et al., 2007), phonics (Vaughn et al., 2006), fluency (Burns et al., 2016), and comprehension (Scammacca et al., 2007) than if they were more comprehensive in nature. Meta-analytic research supported the delivery of targeted tier 2 interventions (Hall & Burns, 2018; Piasta & Wagner, 2010), and students who participated in targeted tier 2 interventions demonstrated greater growth (1.33 and 1.23 weekly WRCM [words read correctly per minute] growth for second and third grades, respectively) over the course of the school year than did students participating only in tier 1 reading instruction (1.25 and 1.03 weekly WRCM growth for second and third grades, respectively; Burns et al., 2016). However, much of the research regarding RtI effects focused on students with moderate reading difficulties and did not examine students with disabilities or severe reading deficits. Thus, reading growth comparisons among students with severe reading deficits, students receiving special education services, and typically achieving students is relatively unknown.
Purpose
Reading growth is a commonly used metric in evaluating the effectiveness of education, special education, and RtI. Examining reading growth rates among students receiving reading intervention and students receiving special education services is necessary to provide insight into the effectiveness of tiered interventions and special education services. Moreover, previous research examining reading growth rates (e.g., Christ et al., 2010; Deno et al., 2001; L. S. Fuchs et al., 1993) did not examine the reading growth of students participating in a tiered reading intervention in comparison to students receiving special education services or students in general education literacy instruction only. Therefore, while these studies provide some insight into differences in reading growth between students receiving special education services and students only in general education, they highlight the need for additional research examining the comparative reading growth of students participating in tiered reading intervention within an RtI framework.
This study examined the reading growth of students with and without disabilities, and students with and without reading deficits in response to tier 2 reading interventions within an RtI framework in an effort to better understand response and non-response to intervention. The following research questions guided the study: (a) What is the effect of targeted reading interventions on the growth of students with severe reading deficits compared to typically achieving peers? and (b) What is the effect of targeted reading interventions on the growth of students with severe reading deficits compared to students receiving special education for reading?
Method
The study took place within the Path to Reading Excellence in School Sites (PRESS; PRESS Research Group, 2014) project, which was a partnership among six urban schools, a research university, a statewide service organization, and a national corporation. The current data were collected during the second year of the 3-year project. The PRESS partnership sought to assist every child to reach grade-level reading proficiency by the end of third grade by creating a tiered system of support that included tiered interventions, universal screening, and monitoring student progress, with emphasis on quality core instruction and embedded professional development. The PRESS model was the RtI implementation model for the participating schools, but primarily focused on implementing tier 2 interventions when the current data were collected.
Participants
The six participating schools served an average of 486.33 (SD = 125.64, range = 323–628) students, 16% of whom were White and 77% of whom were eligible for the Federal Free or Reduced Price Lunch program. On average, 51.67% (SD = 22.62) of third graders in the six schools scored in the proficient range on the state-mandated accountability test for reading before beginning the project. The majority of the schools (n = 4) implemented a balanced literacy curriculum for core reading instruction using the Fountas and Pinnell (1996) reading program and guided reading groups. The other two schools used Reading Mastery (SRA/McGraw Hill, 2008) as their reading curriculum.
Participants for the study were 499 students in second and third grade attending one of the six urban elementary schools participating in the PRESS project. Of the 499 students, 252 (50.5%) were in second grade and 247 (49.5%) were in third grade. The total sample consisted of 53.6% females, and 81% of the students were eligible for the Federal Free or Reduced Price Lunch program. A total of 65.2% of the students were African American, 22.3% were White, 11.2% were Native American, and 1.3% were Hispanic.
Measures
Three measures were used within the study. Measures of Academic Progress for Reading (MAP-R; Northwest Evaluation Association, 2003) was the primary screener, the dependent variable was curriculum-based measures of reading (CBM-R), and the PRESS Decoding Inventory (PRESS Research Group, 2014) was used to help determine intervention targets.
MAP-R—screener
The MAP-R is a computerized adaptive achievement test that was used to screen reading skills of the participants. Each student obtains a Rasch Unit Scale (RIT) score, which is equal interval and independent of grade. The RIT score was converted to a percentile rank. Students who scored at or below the 10th percentile on the national norm at the fall screening were identified as having a severe reading deficit and were selected to receive a reading intervention (n = 92), unless the student already received special education services (n = 22). Students who scored at or above the 50th percentile on the fall assessment were identified as Tier 1 Readers (n = 385). Data from MAP-R correlate well with other measures of reading comprehension and internal and alternate-form reliability coefficients all meet or exceed .89 (Brown & Coughlin, 2007). Students took the computer adaptive reading test in the computer lab in group format in the fall.
CBM-R
Every student was administered grade-level probes from the AIMSweb (2012) assessment system using standard CBM-R administration procedures (Shinn, 1989). Each student read three passages and the number of WRCM was documented for each passage. Then, the median WRCM score from the three probes was recorded as the student’s final score. All CBM-R screeners were administered by trained research assistants as part of the participating schools’ screening activities 3 times per year (fall, winter, and spring).
CBM-R data were also used to monitor student progress for the students receiving a tier 2 intervention, but the data were not analyzed to address the research questions because they were only collected for students receiving a tier 2 intervention. One grade-level probe was administered each week by the research assistant who worked with the student and the number of WRCM was recorded. The assessment was conducted one-on-one. Response to the interventions was judged by the rate of growth as compared to the norm group, and the student was dismissed from intervention if three consecutive progress monitoring data points scored above the next seasonal benchmark criterion as published by AIMSweb (2012). Interventions were implemented until the student scored above the next seasonal benchmark criterion reported by the CBM-R publisher (AIMSweb, 2012), which happened in 10 weeks or less for six students (6.5%), after 13 weeks for seven students (7.6%), 17 weeks for 53 students (57.6%), and 26 students (28.3%) continued the intervention for the full 18 weeks without reaching the next benchmark criterion.
CBM-R data have demonstrated adequate psychometric evidence for students in second and third grade. A review of research found median estimates of alternate-form and test–retest reliability of .89 and .95, respectively (Yeo, 2011). CBM-R has also exhibited adequate concurrent (r = .67–.82, Baker et al., 2008) and predictive validity (r = .62–.92, Harn et al., 2008).
The dependent variable was the growth score for each student that was computed by calculating the slope of the seasonal benchmark scores across the fall, winter, and spring assessments. Slope was computed with ordinary least squares (OLS) because OLS is consistent with previous slope research (e.g., Burns et al., 2016; Deno et al., 2001). The OLS growth score equaled the average weekly increase in WRCM which functioned as an indicator of rate of improvement. The resulting reading growth score served as the dependent variable for the research questions.
PRESS decoding inventory
The school’s instructional coach administered the PRESS Decoding Inventory to students demonstrating phonics needs within 1 week of completing the fall reading screener to determine the appropriate phonics skills with which to intervene. The assessment consisted of the following seven areas, letter-sound correspondence, short-vowel sounds in CVC words, short-vowels and diagraphs, consonant blends with short vowels, long-vowel spellings r- and l-controlled words, and variant vowels and diphthongs. The letter-sound correspondence area involved showing the students 24 letters (x and q were excluded) and asking them to provide the corresponding sound. The students had to correctly answer 21 of the 24 items to demonstrate proficiency and move to the next area. Each of the remaining six areas had 10 items, five that were examples of real words and five that were pseudowords. The students were each asked to read the 10 example words from every area, and student proficiency was demonstrated for the area if 9 or 10 of the items were read correctly. The assessment started with the first area (letter-sound correspondence) and continued until the student did not pass an area (read eight or fewer words correctly). The intervention targeted the most basic area in which the student did not demonstrate proficiency. The decoding inventory was administered one-on-one and was not timed.
Interventions
Students who scored at or below the 10th percentile on the fall MAP-R assessment and who were not receiving special education services were selected to receive a standard protocol reading intervention. The interventions were matched to individual student reading needs based on a diagnostic assessment of the student’s reading skills to determine if a phonics, fluency, or comprehension intervention was most appropriate for each student. The initial diagnostic assessment included an examination of the students’ CBM-R WRCM and accuracy (i.e., WRCM divided by total words read). If the student demonstrated adequate accuracy (i.e., 93%; Treptow et al., 2007), but a below benchmark CBM-R score, then it was determined the student was in need of a fluency intervention. If the student’s CBM-R accuracy was less than 93%, then the PRESS Decoding Inventory was administered within 1 week of completing the fall reading screener to determine the appropriate phonics skills with which to intervene. All students who received a tier 2 intervention were in need of a phonics or fluency intervention.
Phonics interventions
Students with a phonics deficit received one of six phonics interventions. The first two interventions explicitly taught letter sound relationships where students matched picture cards with letters. The third and fourth interventions used Elkonin (1971) boxes for word building activities and were used to match sounds with magnetic letters and to create new words. Elkonin boxes included three or four consecutive boxes on a white board with each box used to represent the space for one letter in a word. In the fifth intervention, students used white boards to write words with a specific phonics skill (e.g., vowel teams, diphthongs), and then they read a passage and identified words with the targeted phonics combination. The last phonics intervention concentrated on analysis of words where students sorted words into categories based on a specific phonics combination.
Fluency interventions
Two fluency interventions were implemented with students demonstrating mastery with phonemic awareness and phonics (i.e., reading with at least 93% reading accuracy; Treptow et al., 2007) but needing additional support with fluency in speed, accuracy, and/or expression in reading (National Reading Panel, 2000). Using the supported cloze procedure (Rasinski, 2003), students worked in pairs with a peer or interventionist to read an instructional level narrative passage 3 times for 1 min each time, alternating reading every other word. An interventionist also followed along and provided support or error correction as needed. Repeated reading (Samuels, 1979) was implemented as the second fluency intervention to assist with rate and expression. Students read an instructional level narrative passage 4 times, for 1 min each time, and the literacy assistant provided error correction at the end of each reading. Comprehension questions were asked after the second, third, and fourth readings about the most important information in the passage and what might happen next in the reading.
Procedures
Students were categorized by their fall MAP-R score and special education status. First, students who scored at or above the 50th percentile on the fall MAP-R assessment were designated as typically achieving and labeled Tier 1 (n = 385). Second, students who read at or below the 10th percentile on the fall assessment were in the Severe Reading Deficit group (n = 92; Torgesen, 2009). Students with fall MAP-R scores that fell between the 11th and 49th percentiles were excluded from the sample. The third group was the Special Education group and consisted of students who were receiving special education services for reading who scored at or below the 10th percentile on the fall assessment (n = 22). Table 1 provides frequencies for each reading category in the sample.
Descriptive Statistics for CBM-R for the Groups.
Note. CBM-R = curriculum-based measures of reading.
The Tier 1 and Special Education groups served as business as usual control groups. Students in the Tier 1 group received regular reading instruction from the Fountas and Pinnell reading program (n = 278, 72.2%) or Reading Mastery (n = 107, 27.8%). The Special Education group included students who were identified with a SLD in reading by school personnel using district guidelines and had an individualized education program (IEP) goal that addressed reading. Students in the Special Education group received reading intervention from their special education teacher based on their IEP goals, but did not participate in the tier 2 reading interventions implemented with the Severe Reading Deficit group. Moreover, all of the students in the Special Education group participated in district-mandated benchmark assessments.
Students scoring at or below the 10th percentile on the fall MAP-R received a phonics or fluency intervention (based on the results of the diagnostic reading assessment as described above) 4 times per week for 20 min in groups of two to four students for up to 18 weeks. Students were grouped by skill level and reading deficit (i.e., phonics or fluency) area. Graduate education students served as reading interventionists to implement and support interventions. All of the interventions began in the fall within 4 weeks of completing the fall screener and within 1 week of each other.
Treatment Fidelity
All interventionists received a 3-hr training on how to implement interventions and administer assessments. The interventionists had to demonstrate at least 95% fidelity with all of the interventions before they began implementing them with students and were observed at least 4 times throughout the year with an intervention checklist to assess continued fidelity. All fidelity observations were unannounced. The observations yielded 90% or higher intervention fidelity for all of the interventions.
Research Design and Analyses
Because the participants in this study made up pre-existing groups, the study used a quasi-experimental research design to compare the growth of the different groups of students. Data were analyzed with an analysis of covariance (ANCOVA) using OLS slope of reading growth as the dependent variable and the group (Severe Reading Deficit, Tier 1, or Special Education) as the independent variable. Reading curriculum (Fountas and Pinnell or Reading Mastery), school, and grade (second or third) were used as the covariates. Planned follow-up comparisons were conducted for the three groups, using an adjusted alpha level of .017 to find significance, and Hedges’s g as an estimate of effect.
Results
Table 1 lists the means and standard deviations of OLS slope by comparison groups. The OLS slope for the Severe Reading Deficit group was first compared to the slope of growth from progress monitoring data to validate the estimate used as the dependent variable. The mean slope of growth from the three benchmark assessments was 1.09 (SD = 0.72), and the mean for the slope from the progress monitoring data was 1.43 (SD = 1.32). The difference between the two scores was a small (g = −0.32) and nonsignificant, t (91) = 1.63, p = .12 effect.
A total of eight students moved from the participating schools after the study began. Four of the students were from the Severe Reading Deficit group, four were from the Tier 1 group, and none of the students in the Special Education group moved, which resulted in attrition rates of 4.3%, 1.0%, and 0.0%, respectively. Thus, the attrition rates for the three groups were low and comparable. The fall CBM-R scores were compared between students who moved and those who stayed with a Mann–Whitney U, which resulted in a nonsignificant effect, U = 712.00, p = .09. Therefore, attrition did not seem to be a major consideration before conducting the analyses to address the research questions.
As shown in Table 2, the effect by comparison group was significant F (2, 493) = 4.59, p < .05 after factoring out the effect of curriculum, school, and grade. The planned follow-up comparisons led to a small (g = 0.22) and nonsignificant effect for the mean scores of the Severe Reading Deficit and Tier 1 groups, t(475) = 1.83, p = .07. However, the difference in mean score between the Severe Reading Deficit and Special Education groups was moderate to large (g = 0.74) and significant t (112) = 3.10, p < .017, and the difference between the Tier 1 and Special Education groups was also moderate (g = 0.68) and significant, t (405) = 3.10, p < .017.
Test of Between Subject Effects for the Analysis of Covariance of Reading Growth.
p < .017.
The data are also displayed in Figure 1 to graphically show the growth for each group. As can be seen in the figure, the Severe Reading Deficit group had the lowest mean score in the fall, but grew at a faster rate than the Special Education group, and accelerated from winter to spring. The mean score for the Severe Reading Deficit group was higher than that for the Special Education group during the spring assessment. Both the Severe Reading Deficit and Special Education groups started lower than the Tier 1 group and remained lower in the spring. The difference between the Tier 1 and Severe Reading Deficit decreased by 5.6% over the three assessments, but the difference between Tier 1 and Special Education increased by 19.4%.
Growth rates for the three groups.
Discussion
The research questions inquired about the effect of interventions on the reading growth of second- and third-grade students with severe reading deficits as compared to typically achieving peers and students identified with a reading disability. Students who participated in the reading intervention grew at a rate that equaled the Tier 1 group (typically achieving peers) but grew at a significantly higher rate than same-age students in the Special Education group.
The data were consistent with previous research that found positive effects for small-group interventions (Burns et al., 2016; Gersten et al., 2009; Vaughn & Wanzek, 2014). Moreover, the interventions were targeted to student needs, which focused on phonics or fluency and previous research also found that it was effective to target those areas of reading deficit among students who struggle with the code-based aspects of reading (Piasta & Wagner, 2010; Scammacca et al., 2007; Vaughn et al., 2006). Burns et al. (2016) study directly compared the effects of targeting an intervention to a more comprehensive intervention that addressed phonics, fluency, and comprehension and found the former to be more effective. This study is consistent with the Burns et al. study regarding the effectiveness of targeted reading interventions but did not directly compare targeted versus comprehensive interventions. Direct comparisons may be useful to help guide interventions and reading theory, which suggests a direction for future research.
A total of 92 students received targeted interventions that were delivered by trained research staff, and 23 special education students received instruction by their special education teachers. The research design did not allow for the examination of the effect of targeted interventions beyond the potential positive effect of receiving intervention from research staff. Scammacca et al. (2015) found 38 studies in which teachers implemented reading interventions with students above the third grade and found a mean effect size of 0.35, but the 29 studies that used research staff to deliver interventions led to a larger mean effect size of 0.68. However, meta-analytic research found equal or larger effect sizes (ES) for interventions implemented by school personnel on standardized (k = 42, ES = 0.50) and nonstandardized reading measures (k = 20, ES = 0.70) than for those implemented with research staff (standardized k = 18, ES = 0.52, nonstandardized k = 13, ES = 0.55) among students in kindergarten through third grade (Wanzek et al., 2016). Future researchers could replicate the current design, but take steps to factor out the effect of research staff versus teachers as implementers.
Implications for SLD
We only directly examined one of the multiple hypotheses for why special education services have not led to more positive student outcomes. We did not examine the quality of the interventions used for special education to determine if they were effective or evidence-based (Burns & Ysseldyke, 2009; Kavale & Forness, 2000), nor did we consider the method with how SLD was identified (Maki et al., 2015) or how well it detected students who would actually benefit from special education. However, we compared students who scored below the 10th percentile in reading to those receiving special education, and some have hypothesized that students receiving special education services had severe deficits that were difficult to remediate (D. Fuchs et al., 2010; D. Fuchs & Fuchs, 2014; Kavale et al., 1994). The Special Education group had higher mean fall CBM-R measure than the Severe Reading Deficit group, but the latter group grew at a significantly faster rate despite the lower starting scores.
This study’s findings were also consistent with previous research suggesting that students receiving special education services made less reading growth across the school year than did students in general education only (Christ et al., 2010; Deno et al., 2001; L. S. Fuchs et al., 1993). However, this study extended this line of research by also examining the reading growth of students participating in a tiered reading intervention in comparison to students in general and special education. Although students in the Severe Reading Deficit group made typical growth across the school year (Deno et al., 2001), they did not demonstrate adequate growth to catch up to their typically achieving peers (L. S. Fuchs et al., 1993). Thus, future research could examine how systematically intensifying intervention (e.g., Lemons et al., 2014) could help students with severe reading deficits meet grade-level benchmarks.
The greater reading growth of the Severe Reading Deficits group compared to the Special Education group suggested that systematically implementing evidence-based targeted reading interventions with students that have severe reading difficulties can result in improved reading outcomes for students. Although the special education reading instruction was not directly examined in this study, these results may support the hypothesis that less effective instructional practices were implemented in special education (Burns & Ysseldyke, 2009), because that group of students made significantly less growth than the students in the Severe Reading Deficits group. However, this hypothesis could be examined by directly comparing the instructional practices of special education and tier 2 interventions and the reading growth of the students receiving those services. Future research could also examine the effectiveness of special education teachers implementing such tiered reading intervention with students with disabilities.
Implications for RtI
The current data supported the effects of implementing small-group targeted reading interventions, even among students with very low reading skills. Therefore, practitioners should consider implementing small-group interventions that focus on foundational reading skills with students who have not yet mastered those skills, like students in this study. We also implemented several components of an effective small-group intervention including closely monitoring progress, keeping the groups relatively small, and implementing them in early elementary (Gersten et al., 2009; Vaughn & Wanzek, 2014). Future research is needed to identify which aspects of an intervention are most important, but until that research is conducted, practitioners should implement all aspects of an effective targeted intervention package. Moreover, the current data, in combination with previous research (Burns et al., 2016; Hall & Burns, 2018), support targeting the tier 2 intervention to student need.
Limitations
Although the study resulted in data that may be of interest to practitioners and researchers, they should be considered within their limitations. First, we only used CBM-R to monitor reading growth. Although CBM-R is closely linked to reading proficiency (L. S. Fuchs et al., 2001), and is a quick and reliable measure, future research should replicate the current design with different or additional measures of reading proficiency. In addition, we only considered growth to measure the dependent variable and did not compare proficiency levels. Students who were below the 10th percentile made more growth, but remained behind their peers in the Tier 1 group. We also did not have psychometric data on the PRESS Decoding Inventory.
We examined the three data points from the seasonal benchmark assessments because those data were collected for all three groups and did not examine the weekly progress monitoring data. Growth across 1 school year is not linear and tends to decrease from winter to spring as compared to fall to winter (Nese et al., 2013). The OLS slopes for the Tier 1 and Special Education groups followed that pattern, and the difference in slopes from weekly progress monitoring data and benchmark scores was not significant. Moreover, the slope of growth for the Severe Reading Deficit group accelerated from winter to spring. Future researchers could replicate this design while also conducting weekly progress monitoring for all students or by taking into account the curvilinear nature of the data across 1 year.
The study used a quasi-experimental design and did not randomly assign students to groups, which resulted in substantial differences in group size. Future researchers could seek out more students with disabilities to better equate group sizes. We also used school-based identification of disabilities and did not have access to the actual data with which the identification decisions were made. Future research should either confirm the presence of a disability or use the data with which the school-based identification decisions were made within the statistical model to determine identification. We also focused on second and third grades, and future research is needed across the grades or with different types of reading deficits (i.e., phonemic awareness, vocabulary, and comprehension).
The validity of our conclusions was also limited by the lack of information about the intervention and instruction. We did not examine the instructional practices for students in the Special Education group and did not examine how the students were identified with a disability. Thus, we did not consider hypotheses for the reported poor outcomes for students receiving special education other than to compare students with reading deficits but who did and did not receive special education. It is unknown to what extent effective instructional practices were implemented to remediate the special education students’ reading difficulties or the extent of their reading difficulties beyond a percentile rank score on a fall reading screener. Relatedly, we do not know about the instruction for students in the Tier 1 group other than the curriculum that was reportedly implemented and could not compare growth between the two types of reading interventions (phonics or fluency) that students did receive because of movement between the two groups.
Finally, we used the 10th percentile as the cut point for our groups to serve as an indication of severe reading deficit but students who performed between the 11th and 49th percentiles may also have severe reading needs that were not identified on the benchmark assessment. Moreover, in this study, students below the 10th percentile were compared to those at or above the 50th percentile. Future research could implement a regression discontinuity design around the 10th percentile and compare the effects of various cut points to the analysis.
Conclusion
The current data supported the importance of implementing small-group reading interventions, even with students who have the lowest reading skills. The students who received the targeted interventions grew at a rate that equaled the growth rate of students without reading difficulties and exceeded that of their peers receiving the most intensive intervention that a school has to offer (i.e., special education), which has implications for schools implementing tiered systems of support. Given the current findings and the frequency with which schools across the country are implementing MTSS, this study highlights a clear need for continued research in this area.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: The study was implemented with a grant from the Target Corporation.