Exploration of a Blended Learning Approach to Reading Instruction for Low SES Students in Early Elementary Grades

Abstract

This study investigated the potential benefits of a blended learning approach on the reading skills of low socioeconomic status students in Grades 1 and 2. Treatment students received English language arts instruction that was both teacher-led and technology-based. Comparisons were made with control students who received the same English language arts instruction without the blended learning component. Results showed significantly greater pretest/posttest gains on a standardized reading assessment for the treatment students compared to the control students. The greatest discrepancy occurred in reading comprehension. A sub-analysis of low-performing English language learner students in the treatment group revealed the largest reading gains. At posttest, these students performed at the level of non-English language learner students in the control group. Results indicated a blended learning approach can be effective in enhancing the reading skills of low socioeconomic students.

KEYWORDS:

• reading
• blended learning
• low SES
• technology
• English language learners

The National Institutes of Health have long recognized high illiteracy rates among elementary school children. According to The National Center for Education Statistics' (2011) Nation's Report Card (2011), data indicate that 34% of fourth graders read below a basic level on tests of reading skills and more strikingly, around 50% of Black and Latino students read below a basic level. Students who were eligible for free/reduced lunch and/or were English language learners (ELLs) were also more likely to be low performers on reading assessments. Reading comprehension has long been identified as an essential component of reading success (National Reading Panel, 2000) and becomes increasingly more important as children move into the upper elementary grades and beyond. In addition, early identification of reading difficulty is critical and most children who have not received quality reading intervention by Grade 3 never catch up with their typically achieving peers (Torgesen, Rashotte, & Alexander, 2001). Given the severity of these literacy problems, there is a need for more research on interventions to improve reading skills, including reading comprehension, particularly in the early elementary grades (Toste et al., 2014).

Historically, computer-assisted instruction (CAI) has been found to be a valuable supplementary aid to support reading acquisition, particularly for struggling students (see MacArthur, Ferretti, Okolo, & Cavalier, 2001). CAI allows students to work at their own pace so that they can receive sufficient, independent practice that may not be possible within a traditional classroom setting (Johnson, Perry, & Shamir, 2010). Individualized instruction, together with use of pictorial displays and consistent, positive feedback provided by CAI, is often highly effective and motivating (Saine, Lerkkanen, Ahonen, Tolvanen, & Lyytinen, 2011).

A meta-analysis by Cheung and Slavin (2013) investigated the use of CAI for struggling readers. Overall, a small, positive effect of CAI on acquisition of reading skills was found (effect size =.14). The largest effect sizes were obtained in small-group studies, while comprehensive models produced the smallest effect sizes. Based on this analysis, additional studies on CAI and reading are needed in order to validate the effectiveness, generalizability, and areas of impact of these interventions (see Chambers et al., 2008).

Most previous research on CAI in elementary grades has focused on early reading skills. For instance, a number of studies have examined the effects of CAI on phonological awareness. Positive effects have been reported for students in general (Cassady & Smith, 2004; de Graaff, Bosman, Hasselman, & Verhoeven, 2009; Macaruso & Rodman, 2011b; Savage et al., 2013; Segers & Verhoeven, 2005; Wild, 2009), as well as for students identified as at risk or low performers (Hecht & Close, 2002; Lonigan et al., 2003; Macaruso & Walker, 2008; Mitchell & Fox, 2001; Mioduser, Tur-Kaspa, & Leitner, 2000). In addition to phonological awareness, some studies have found evidence that CAI benefits acquisition of word attack (i.e., letter-sound) knowledge (Macaruso, Hook, & McCabe, 2006; Segers & Verhoeven, 2005), word identification skills (Hecht & Close, 2002; Macaruso & Rodman, 2011b; Saine et al., 2011; Shelley-Tremblay & Eyer, 2009; van Daal & Reitsma, 2000), and rapid naming/word fluency (Johnson, Perry, & Shamir, 2010; Saine et al., 2011).

Relatively few studies have investigated the benefits of CAI for reading comprehension, particularly in the early elementary grades. Studies addressing reading comprehension have typically utilized a blended learning approach that includes intensive intervention programs that incorporate teacher-led instruction along with CAI. A blended learning approach is defined by a student learning, in part through a self-controlled digital content delivery method, combined with traditional, teacher-led instruction in a classroom setting (Staker & Horn, 2012). In one study Chambers et al. (2008) used Success for All with first graders who were non-native speakers of English and from low socioeconomic status (SES) families. Success for All is a year-long program with over 100 minutes per day of instruction, including whole class lessons, computer-assisted tutoring groups, multimedia activities, and individual daily tutoring. Students in the intervention program were found to make significant gains in word attack, word identification, passage-level fluency, and reading comprehension. Similar outcomes were reported by Torgesen, Wagner, Rashotte, Herron, and Lindamood (2010) using two computer-based intervention programs—Read Write and Type and Lindamood Phoneme Sequencing—with at-risk first graders. The year-long programs included four 50-minute sessions per week of small group instruction with teachers recruited and trained to implement the interventions. Significant effects of the interventions were found in phonological awareness, word attack, rapid naming, word identification, and reading comprehension (see also Cassady & Smith, 2005; but see Paterson, Henry, O’Quin, Ceprano, & Blue, 2003, for an exception).

In the current study, we examined the efficacy of a blended learning approach for children in early elementary grades as compared to control students. Like Chambers et al. (2008), this study considered students from low-SES families, including ELL students. In conducting this study, we assessed the benefits of blended learning in the context of a typical daily classroom schedule, which closely reflected authentic use patterns found in school-based implementations.

THE BLENDED LEARNING APPROACH

The blended learning approach in the current study integrated Lexia Reading Core5 (Core5) into the school's basic curriculum for English language arts (ELA) instruction. The curriculum, LEAD21 (Wright Group, n.d.), is a comprehensive program for kindergarten through Grade 5 that is aligned with Common Core State Standards (National Governors Association for Best Practices & Council of Chief State School Officers, 2010). The LEAD21 program addresses literacy-based skills (phonemic awareness, phonics, fluency, vocabulary, language acquisition, comprehension, and writing) and other essential 21st-century skills (e.g., communication and collaboration). Curriculum units rotate through various themes in conjunction with the larger areas of humanities, science, and social studies. Readings assigned in each grade link multiple topics with a mixture of nonfiction and fiction texts. LEAD21 provides Theme Readers and Differentiated Readers in paper and ebook format, which can be used with interactive whiteboards.

Recently released in school systems, Core5 integrates online activities, ongoing assessment to guide instruction, and targeted resources for teacher-led instruction as well as independent offline work. Core5 was developed as an enhanced version of an earlier technology-based program called Lexia Reading. The earlier program focused strongly on building phonological awareness and word attack skills for words in isolation, as well as in context. Macaruso and Walker (2008) found that Lexia Reading helped develop phonological awareness in kindergartners, with the greatest benefits seen in students identified as low performers at pretest. Follow-up studies showed significant gains in word reading for a larger sample of students identified as low performers (Macaruso & Rodman, 2011b), and in both phonological awareness and word reading for ELL students (Macaruso & Rodman, 2011a).

Core5 provides systematic and personalized instruction through six strands of reading skills: phonological awareness, phonics, structural analysis, automaticity/fluency, vocabulary, and comprehension. The content aligns with recommendations from the National Reading Panel (2000) and Common Core State Standards. Core5 contains activities organized into 18 levels: preschool (Level 1), kindergarten (Levels 2–5), first grade (Levels 6–9), second grade (Levels 10–12), third grade (Levels 13–14), fourth grade (Levels 15–16), and fifth grade (Levels 17–18). Each level consists of five activities (four in Level 1) with multiple units designed to address various combinations of the six strands listed previously. Table 1 provides a summary of Core5's contents by strand.

Table 1 Summary of Core5's Contents

Strand	Content
Phonological awareness	Rhyming, blending and segmenting syllables and sounds, manipulating sounds
Phonics	Upper- and lower-case letters, alphabetizing, letter/sound correspondence, six syllable types, syllable division
Structural analysis	Prefixes, roots, and suffixes; spelling rules; Greek combining forms
Automaticity/fluency	Foundational concepts, high-frequency words, paragraph level text
Vocabulary	Categorizing, spatial concepts, advanced adjectives, multiple meanings, shades of meaning, synonyms, antonyms, similes, metaphors, idioms, analogies, affix and root meanings
Comprehension	Listening comprehension, picturing, signal words, strategies for narrative and informational text (main idea, details, prediction, inferences, conclusion, cause/effect, compare/contrast, summarizing, paraphrasing, perspective, fact/opinion)

The student experience begins with an embedded auto-placement test, which is used to place students at an initial level consistent with their reading ability. For example, a second-grade student may place in any level, including Level 6 (below grade level), Level 10 (in grade level) or Level 13 (above grade level). After completing the auto-placement, students begin working on activities within the level where they are placed. Students progress to subsequent units and levels only when they demonstrate mastery of the content, which is 90%–100% accuracy. However, if students struggle in a unit, a system of automatic branching moves them to a scaffolded practice task with fewer stimuli and more structure. If students continue to struggle, they receive direct, targeted instruction that explicitly addresses the type of error the student made when attempting to master the reading skill. After successful completion of all activities in a level, students are promoted to the next level. Motivating visual graphics include progress bars that fill up as students successfully complete units within an activity, allowing students to monitor their own progress through the program. Core5 also maximizes time-on-task by providing immediate feedback after each response without distraction or delay.

Teachers have access to online reports that identify students who are struggling with a particular skill. Those students are flagged for individual or small group instruction. Core5 offers skill-specific materials called Lexia Lessons that provide step-by-step instruction following the gradual release of responsibility model for a teacher or paraprofessional to address students' skills. Upon completion of all units in an activity, students have access to Skill Builders—paper and pencil tasks developed to build automaticity and extend the skills presented in the activity. Students can transition between Lexia Lessons, Skill Builders, and the online program as directed by the teacher.

A key component of Core5 is the availability of online reports that include actionable data and are updated in real time. One component of the online report is monthly Performance Predictors that show students' percent chance of reaching end-of-year, grade-level benchmark (i.e., completing all activities at the students' grade level). For each Performance Predictor, online reports also present a Prescription of Intensity (i.e., recommended weekly minutes of online work) designed to improve the students' chance of reaching benchmark. For students who need additional support to reach benchmark, suggestions and materials for targeted instruction are provided. In the current study, we examined the efficacy of a blended learning approach, including Core5, for children in early elementary grades as compared to control students within the context of a typical daily classroom schedule.

METHOD

Participants

The current study was conducted in an urban elementary school in western Massachusetts. Students in two first-grade classes and two second-grade classes were invited to participate in this study. For each grade, one class was randomly assigned to the treatment group and the other class to the control group. Teachers in the treatment and control classes had comparable qualifications. Three teachers had master's degrees in education (one treatment, two control) and one teacher had a Bachelor of Science degree. The two treatment teachers had 6 and 12 years of teaching experience, and the two control teachers had 9 and 15 years of teaching experience.

There were 47 students in treatment classes and 41 in control classes. Of these students, two treatment and three control students were eliminated from analyses because they were absent at posttest. The final sample consisted of 45 treatment students (23 male, 22 female) and 38 control students (14 male, 24 female). Dates of birth were made available for 30 treatment and 27 control students. There was no difference (p >.05) between treatment and control students in mean age: 86.8 months (SD = 8.6) and 85.7 months (SD = 9.3) for treatment and controls, respectively. Most students in each group were Hispanic: 89% treatment, 82% control. The remaining students were Black (9% treatment, 13% control) or White (2% treatment, 5% control). Nearly all students (95%) qualified for free or reduced lunch. Fourteen treatment and seven control students were classified as Limited English Proficiency. For the purpose of this study, these students were considered ELL. Due to the additional oral language challenges associated with ELL students, separate analyses were conducted for this group. Two treatment students received special education services. Results were nearly identical with the two special education students in or out of the analyses, so they remained in the analyses.

Materials and Procedures

As described previously, both treatment and control classrooms used LEAD21 as the ELA curriculum. This curriculum had been used at the school for three years prior to the study. In LEAD21, concepts, skills, and strategies are introduced through whole class instruction, and then students rotate through small group instruction and individual activities to reinforce lessons. As described above, Core5 includes an online program along with integrated teacher-led Lexia Lessons and paper and pencil independent, student-led activities (Skill Builders) as appropriate. Teachers took part in a half-day orientation session prior to implementation, where they were trained in best practices for integrating the online learning and offline instructional materials of Core5 into their classroom instruction.

Core5 served as part of the blended learning component of the ELA instruction for treatment classes. Each treatment classroom had six computers and students used the online program as a center activity following a rotating schedule. Teachers were instructed to have students use the online program in accordance with recommended minutes (20 to 80 per week, depending on risk) based on the Prescriptions of Intensity. Treatment students began using Core5 in October 2012 and continued into June 2013. During the time when treatment students were using Core5, control students were engaged in regular ELA instruction, resulting in all students receiving 150 minutes per day of ELA instruction.

Over the course of the study, the students in the treatment group showed strong use of Core5’s online program: Their average login time was 85 minutes per week. Looking across all student sessions, the minimum and maximum times were 28 and 203 minutes per week, respectively. These strong use patterns resulted in over 90% of students meeting usage recommendations for at least three months and 62% met usage recommendations for five or more months.

As part of ongoing monitoring, two members of the research team visited the treatment classes in February 2013 to observe students using the online program. During the session, 12 students were observed in each classroom. Independent observations were made at two time points in a session. For the first-grade class, there was 88% (21/24) agreement across time points. The observations agreed upon by both raters were as follows: 15 cases in which students were engaged with the program, four cases in which students were seated but not engaged, and two cases in which students were not seated. Raters disagreed on three cases, where students were seated but it was unclear if they were engaged with the program. For the second-grade class there was 100% (24/24) agreement that all students were engaged with the program.

In addition to observations and data analysis, interviews were conducted to assess use of Core5 as part of classroom instruction. Both teachers of the treatment classes reported that the online program was a key center activity and they used the Lexia Lessons when students were struggling with specific skills in the online program. The first-grade treatment teacher said she printed out lessons (with directions) for parents to use at home. The Skill Builders were used as morning activities in the first-grade class and as homework for second graders. The second-grade teacher reported that when all students had mastered a skill in the online activities, she used that information to direct her whole class instruction to other skill areas.

Measures

The Group Reading Assessment and Diagnostic Evaluation (GRADE), Level 1 (first grade) or Level 2 (second grade) (Williams, 2001) was administered to treatment and control students as a pretest in September 2012 and as a posttest in May 2013. Separate forms of the GRADE were used at pretest and posttest. Both levels of this assessment contain two domains:

Vocabulary and Comprehension. The Vocabulary domain consists of two subtests—Word Reading and Word Meaning. On Word Reading, the student hears a word spoken by the tester (e.g., “here”) and is asked to select one of four words (e.g., have, had, here, he) that matches the spoken word. This subtest contains 20 items in Level 1 and 28 items in Level 2. On Word Meaning, the student sees a word (e.g., drum) and selects one of four pictures that represents the meaning of the word (e.g., gum, drawer, door, drum). This subtest contains 27 items in both Level 1 and Level 2.

The Comprehension domain consists of two subtests—Sentence Comprehension and Passage Comprehension. Sentence Comprehension requires the student to read a sentence silently (e.g., Paul had five ______________ in his pocket.) and select one of four words (e.g., shiny, minutes, money, pennies) that best completes the meaning of the sentence. This subtest contains 19 items in both Level 1 and Level 2. On Passage Comprehension, the student reads a passage silently and then answers three questions (Level 1) or four questions (Level 2) about the passage. Questions are presented in a four-choice format and the passage remains in view while the student answers the questions. The questions address factual information, identifying the gist of the passage, and making inferences from the passage. This subtest contains 24 items in Level 1 and 28 items in Level 2.

The GRADE was administered to each class in two sessions in accordance with test guidelines. Each session lasted between 30–45 minutes. The Word Reading and Sentence Comprehension subtests were given in the first session, and the Word Meaning and Passage Comprehension subtests were given in the second session.

Raw scores on the Word Reading and Word Meaning subtests were added to obtain a Vocabulary Composite score, and raw scores on the Sentence and Passage Comprehension subtests were added to obtain a Comprehension Composite score. Adding these two composite scores results in a Total Test score. The GRADE provides standard scores for the Total Test score and the composite scores. Separate analyses were conducted for Total Test, Vocabulary, and Comprehension composite standard scores.

RESULTS

The first section examines the extent to which treatment students advanced in Core5 over the school year. The second section compares treatment and control groups in terms of pretest and posttest scores on the standardized assessment, and the third section provides sub-analyses for ELL students.

Advances in Core5 for Treatment Students

In general, treatment students were able to systematically complete activities and advance in Core5 over the school year (see Table 2). Of the 18 students who started in Core5 two grade levels below their grade in school (e.g., a second grader placed in a kindergarten level), 83% advanced to the next grade level of material and 17% advanced more than one grade level of material over the school year, ending the year in their grade level of material. Most of the students (81%) who started one grade level below or in grade level were able to advance to the next grade level or beyond. Overall, 20% of students finished their grade level material and reached end-of-year benchmark.

Table 2 Advances in Core5 Over the School Year

Placement level	Number of students	No change (%)	Advanced to next grade level of material (%)	Advanced more than one grade level of material (%)
Two grades below	18 (40%)	0	83	17
One grade below	18 (40%)	17	72	11
In grade level	9 (20%)	22	78	0

Comparison of Treatment and Control Students

Table 3 presents mean pretest and posttest standard scores on the GRADE for the treatment and control groups. Independent t tests showed that there were no significant differences between groups at pretest on Total Test scores (t(81) = 0.80, p =.43) and for each domain: Vocabulary (t(81) = 0.19, p =.85) and Comprehension (t(81) = 0.73, p =.47). These results indicate that the two groups showed comparable literacy skills prior to initiation of the treatment program.

Table 3 GRADE Standard Scores for All Students

	Treatment (N = 45)				Control (N = 38)
	Pretest		Posttest		Pretest		Posttest
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	d
Total test	91	17.1	106.6	14	93.9	14.8	102.9	14.1	0.53
Vocabulary	93.7	20.8	105.6	11.4	94.5	16	105.1	12.1	0.09
Comprehension	91.7	13.9	105.9	14.5	93.8	11.9	100.8	15	0.52

One-sample t tests revealed that both groups showed significant gains in Total Test scores: treatment (t(44) = 7.50, p <.01); control (t(37) = 5.10, p <.01). However, the extent of gain was larger for the treatment group (15.6) than the control group (9.0). An independent sample t test showed that the difference in gain scores favoring the treatment group was significant (t(81) = 2.38, p =.02). In addition, an analysis of covariance comparing Total Test scores at posttest using Total Test pretest scores as covariates confirmed the significant group effect (F(1, 80) = 5.23, p =.03). An effect size of.53 was obtained when comparing mean gain scores for the two groups. This effect size is moderate.

One-sample t tests were also used to examine standard score gains in each domain separately. The treatment group showed significant gains in Vocabulary (t(44) = 4.51, p <.01) and Comprehension (t(44) = 6.59, p <.01). The control group also showed significant gains in both domains—Vocabulary (t(37) = 6.42, p <.01) and Comprehension (t(37) = 3.24, p <.01). The two groups showed similar gains in Vocabulary (treatment: 11.9; control: 10.6), whereas gains were greater for the treatment group (14.2) than for the control group (7.0) in Comprehension. As might be expected, independent sample t tests showed no significant group difference in Vocabulary (t(81) = 0.42, p =.67) but a significant group difference in Comprehension (t(81) = 2.33, p =.02). Analyses of covariance comparing posttest scores using pretest scores as covariates confirmed the significant group effect in Comprehension: (F(1, 80) = 4.89, p =.03). Effect sizes comparing mean gain scores for the two groups were.09 and.52 in Vocabulary and Comprehension, respectively. While the effect size in Vocabulary is low, the effect size in Comprehension is moderate.

Sub-analyses for ELL Students. This section examines the performance of ELL students. Table 4 compares GRADE scores for ELL students in the treatment group (N = 14) with ELL students in the control group (N = 7). At pretest, the Total Test mean scores for both groups were low, falling more than one standard deviation below the normed mean (100). There were no significant group differences at pretest on Total Test scores (t(19) =.12, p =.86) or scores in each domain: Vocabulary (t(19) =.17, p =.87) and Comprehension (t(19) = 0.15, p =.88).

Table 4 Standard Scores on the GRADE for ELL Students

	Treatment (N = 14)				Control (N = 7)
	Pretest		Posttest		Pretest		Posttest
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	D
Total (N = 21)	82.7	17	102.9	11.8	84.1	16.9	96.3	22.8	0.85
Vocabulary	83.8	15.8	100.4	8.3	82.6	16.1	97.4	13.9	0.21
Comprehension	88.1	12.4	102.4	14.5	89.0	14.2	93.6	23.0	0.75

Significant gains in Total Test scores were found for both ELL groups: treatment (t(13) = 8.30, p <.01); control (t(6) = 3.33, p =.02). The extent of gain was more dramatic for the treatment group (20.2) than the control group (12.2). However, due to low power associated with small sample sizes, the difference in gain scores failed to reach significance (t(19) = 1.88, p =.08). An effect size of.85 was obtained when comparing mean gain scores for the two groups. This effect size is high.

Both ELL groups showed significant gains in Vocabulary. The treatment group showed a gain of 16.6 (t(13) = 5.17, p <.01) and the control group's gain was 14.8 (t(6) = 7.45, p <.01). The difference in gain scores was not significant (t(19) = 0.37, p =.72). The effect size of.21 in Vocabulary is low. In the area of Comprehension, the treatment group showed a significant gain of 14.3 (t(13) = 4.89, p <.01) whereas the control group's gain of 4.6 was not significant (t(6) = 0.81, p =.45). The difference in gain scores failed to reach significance due to limited power (t(19) = 1.71, p =.10); however, the effect size of.75 in Comprehension falls in the moderate-to-high range.

Overall, ELL students in the treatment group demonstrated substantial gains on the GRADE. Total Test scores improved from more than one standard deviation below the normed mean (82.7) to above the normed mean (102.9). Figure 1 compares gains made by ELL students in the treatment group with non-ELL students in the control group. At pretest, ELL treatment students scored significantly below non-ELL control students (t(43) = 2.82, p <.01). By posttest, ELL treatment students closed the performance gap and showed nearly identical scores to non-ELL control students (t(43) = 0.39, p =.70).

Figure 1 GRADE total test standard score changes for ELL treatment and non-ELL control students.

DISCUSSION

This study examined the efficacy of a blended learning approach, combining teacher-led and technology-based ELA instruction to teach reading to low SES and ELL students in Grades 1 and 2. Comparisons were made with a control group receiving the same ELA instruction without the blended learning component. All students in treatment and control groups received the same total amount of ELA instructional time. Blended learning was implemented in the context of a typical classroom schedule to mirror use patterns of school-based blended learning programs. Treatment and control groups did not differ on pretest scores on a standardized test of reading skills and both groups made significant gains in reading. However, results indicated that treatment students demonstrated gains that were significantly higher than gains achieved by the control students, particularly in the area of reading comprehension. An ELL sub-sample displayed even greater gains than the group as a whole, with a high effect size, but the results failed to reach significance due to a small sample size. These findings demonstrate the potential benefit of integrating digital instruction into an ELA curriculum utilizing a blended learning approach for low SES and ELL students and strengthen previous reports showing benefits of computer-aided instruction for ELL students (Chambers et al., 2008; Macaruso & Rodman, 2011a).

It is encouraging that both groups made substantial gains in standard scores regardless of whether they participated in the treatment. This progress can be attributed to the fact that the LEAD21 program used in the school was comprehensive, addressing the five components of reading, in conjunction with theme-based readers that likely contributed to gains in vocabulary and comprehension. To find significant differences in gains when the traditional curriculum is strong further reinforces the effectiveness of a carefully structured and implemented blended learning model over the traditional curriculum. Notably, the treatment students demonstrated better performance in posttest reading comprehension after using Core5 as a part of the ELA curriculum. This finding supports previous studies by both Chambers et al. (2008) and Torgesen et al. (2010) that effectively used a blended learning approach to bolster the reading comprehension abilities in early elementary grades. Successful early intervention in reading comprehension is particularly valuable for supporting reading growth in later grades.

As discussed previously, Cheung and Slavin's (2013) meta-analysis has shown that CAI studies have produced modest outcomes with generally low effect sizes. One possible reason for this is that some studies inadequately integrated the technology component into the classroom curriculum. For instance, Paterson et al. (2003) failed to find benefits of the Waterford Early Reading Program in a large scale kindergarten study. Unlike Paterson et al., however, other studies have reported significant benefits of the Waterford Early Reading Program in early elementary grades (Cassady & Smith, 2004, 2005; Hecht & Close, 2002). According to Cassady and Smith (2005), the null findings of Paterson et al. (2003) were related to low levels of effort or interest on the part of the teachers in actively integrating the materials into their instructional literacy program. In a more recent study of students in grades K–5, Ness, Couperus, and Willey (2013) did not include the teacher-led aspects of the program Lexia Reading as part of the implementation model. Intervention students received no additional support from the teachers, either through the use of the performance data to guide instruction or through any targeted offline instruction. The study found no significant benefits of the intervention program; however, as noted earlier, studies of Lexia Reading have reported positive outcomes when the program was implemented with fidelity (Macaruso & Rodman, 2011b; Macaruso & Walker, 2008).

Although there have been positive outcomes with a blended learning approach, the use of this approach in the typical classroom raises some challenges that should be considered; including the time it takes to plan for students of all the ability levels, the intensity of instructional time for each student, and the amount of professional development for the teacher. Previous studies showing benefits of a blended learning approach for reading comprehension involved substantially more intensity than commonly found in the regular classroom setting. In the Torgesen et al. (2010) study, students spent 200 minutes a week in small group specialized instruction, primarily held outside the classroom reading block. All teachers had previously worked with the researchers and received an additional 18 hours of preservice training plus biweekly three-hour staff meetings for the duration of the school year. In the Chambers et al. (2008) study, tutors had intensive support, including pre-training, follow-up support, in-classroom visitations from both school facilitators and Success for All trainers, and a comprehensive training manual. Tutors also utilized real-time tutor support in the form of video vignettes of intervention strategies. The time commitment for at-risk students was also intense, with daily 20-minute, one-on-one sessions, until they reached their classroom reading level.

The blended learning approach used in the current study was less time intensive for both student and teacher and required less specialized teacher training than the above studies. Students spent an average of 80 minutes a week with the computer-based components of the instructional intervention depending on their personal usage prescription (usage targets ranged from 20–80 minutes per week, depending on overall risk level). Foorman and Torgesen (2001) suggested that intervention for at-risk students must be intensive. Although increased intensity usually implies more time, it could be argued that the specific targeting of individualized skill needs serves a similar purpose and allows students to make significant progress without requiring them to spend more time. As noted previously, treatment and control students all received the same total amount of ELA instructional time. In addition, students were identified in the online program as needing additional support and flagged for specific lessons that targeted their personal skill gaps. Because the targeted lessons in this program provide highly structured language, they can be delivered without additional training by a teacher or an aide. This less time intensive approach to teacher training and student daily lessons is more aligned with the realities that teachers face in a typical classroom setting.

It is important to consider unique elements of this blended learning model which may have contributed to student growth over and above the traditional ELA curriculum. Effective instruction involves clear goals for the student, content that is explicit, systematic, and scaffolded with immediate modeling and corrective feedback, familiar instructional routines, and ongoing monitoring of student performance (Foorman & Torgesen, 2001; Foorman et al., 2003; Scammacca et al., 2007; Vaughn, Denton, & Fletcher, 2010). This description of effective instruction can seem overwhelming to teachers of underperforming students who have not yet mastered concepts presented at lower levels. Teachers may not be as familiar with the scope and sequence of material typically taught in earlier grades, so it is difficult to provide material that is systematic and targeted at each student's specific level, give immediate feedback, monitor student performance, and intervene when necessary. Technology offers a platform to deliver an individualized learning path, as well as provide the teacher with performance data to help target teacher-led instruction and monitor progress. Additionally, scripted lessons can allow teachers to target instructional content that is less familiar to them when necessary.

The individualized, systematic, and structured approach of the Core5 program was particularly effective in accelerating the reading growth of ELL students, a majority of whom started well below grade level at the beginning of the year. As discussed above, the ELL treatment students had a growth of 20 standard score points as compared to a gain of 12 standard score points for ELL control students. By the end of the year, ELL treatment students had closed the gap and were performing at the levels of non-ELL students in the control group. ELL students are often exposed primarily to grade-level material in the standard classroom but need to begin reviewing skills at a lower level. The initial placement tool in the Core5 program is designed to start students on skills at the appropriate level and then systematically move them into grade level skills. Another benefit of online learning is that students are not able to remain passive in their learning. Instead, they are required to respond and interact with material. This forced active response may work to counteract a classroom-based phenomenon for ELL students where they are less likely to engage in classroom discussion than their native English speaking peers (Mohr & Mohr, 2007; Pappamihiel, 2002).

Limitations and Future Directions

A main limitation of the current study is that overall sample size is relatively small. However, integrity in the study design was maintained as classrooms in the same grade were randomly assigned to treatment and control. While significant findings were obtained for the sample as a whole, the subsamples of ELL students were too small to reach significance. Future studies will aim to address larger samples of interest in underperforming populations, particularly ELL students, as considerable academic skill gaps remain between ELL and non-ELL students (Grantmakers in Education, 2011).

CONCLUSION

A blended learning program integrated into a strong ELA curriculum was shown to be more effective in bolstering the reading skills of low SES and ELL students than the same ELA curriculum without the advantages of digital technology. Previous studies that failed to find significant benefits tended to perceive digital technology as operating in a vacuum instead of playing an important role in driving the ELA curriculum. Results of the current study indicate that digital technology can leverage teachers' time, allowing teachers to identify and address areas of need through a time efficient but individualized data-driven approach that can be implemented within the schedule and time constraints of a typical classroom.