High-Impact Tutoring to Accelerate Learning for Intermediate Students: A Pilot Study

Abstract

High-impact tutoring is a promising way to address reading and mathematics achievement declines following years of pandemic-interrupted schooling. This pilot study seeks to determine the impact of small-group tutoring, provided by preservice teachers, on students in grades 2–5. Using beginning-, middle-, and end-of year benchmark assessments, descriptive statistics were calculated for tutored and non-tutored groups and compared against the norming sample. The results indicate that students receiving mathematics tutoring (fourth and fifth grades) and third-grade students receiving reading tutoring demonstrated growth at rates that exceeded their non-tutored peers. Second-grade students who received reading tutoring improved at a greater rate than the norming sample but less than their non-tutored peers. The results of this pilot study indicate the potential impact of tutoring by preservice teachers on reading and mathematics outcomes for intermediate students.

1. Introduction

After the COVID-19 pandemic, schools were drastically different for both students and teachers. Students experienced increased challenges with schoolwork (Johnson et al., 2024). Special education services decreased for high-needs students (Sonnenschein et al., 2022). Standardized test scores displayed sharp declines in both reading and mathematics, with even greater declines in achievement for low-performing students and students with disabilities (Irwin et al., 2022). Many stakeholders are worried about learning loss and increasing student achievement (Willen, 2022); the lasting effects of school shutdowns linger (Anderson, 2021) with concerns about long-term impacts on students’ academic success (Dusseault et al., 2021). Accordingly, policymakers are interested in finding effective ways to positively impact student learning (Fahle et al., 2023), particularly through high-impact tutoring (The White House, 2024).

1.1. Tutoring in the United States

Broadly, tutoring refers to a variety of services in which students receive one-on-one or small-group learning support (Willen, 2022). Private tutoring and tutoring services exist outside of the traditional school day. Tutoring within the school day has included many initiatives in the United States over the last 30 years. Prior large-scale tutoring initiatives in the United States have produced mixed results on student achievement. For example, the Corporation for National and Community Service was founded in 1993. One component of this government agency, AmeriCorps, offers reading and mathematics tutoring as part of its programming. AmeriCorps tutors receive significant training, offer dedicated tutoring each week using high-quality instructional materials, and receive supervision and support during tutoring. Consequently, several studies have shown AmeriCorps tutoring has positive impacts on student achievement in both reading and mathematics (Markovitz et al., 2022; Parker et al., 2019).

America Reads is another large-scale initiative focused on tutoring. Founded in the late 1990s, the program leverages work study resources for college students to tutor K-12 students in reading and mathematics. Worthy et al. (2003) noted that America Reads has not consistently produced positive student achievement outcomes. The studies that did demonstrate positive student achievement outcomes from America Reads included significant tutor training, tutors with high motivation, and ongoing support and supervision (Borges & McCarthy, 2015; Worthy et al., 2003). More recently, nonprofits such as Saga Education have entered the tutoring space. Saga Education partners with school districts and provides trained tutors in mathematics. Evaluations of Saga Education tutoring have shown significant gains, particularly in secondary education settings (Cook et al., 2015).

1.2. High-Impact Tutoring

The varied designs and qualities of the tutoring models in the United States have led to many studies on effective and non-effective methods. High-impact tutoring leads to significant learning gains for students receiving it. It is more successful when tutoring is conducted in small groups (Nickow et al., 2020) and provided by education professionals (e.g., preservice teachers, paraprofessionals, and teachers) (Nickow et al., 2024). Using materials aligned with classroom instruction also leads to better student achievement (Robinson et al., 2024). Dosage and frequency are also critical factors, with a recommendation of 30 min at least three times per week (Robinson et al., 2024). Based on this research, the National Student Support Accelerator (NSSA) offers high-impact tutoring program design considerations to ensure that tutors are trained to implement the best practices, build relationships with students, and that the programs provide ongoing support to ensure quality tutoring (Robinson et al., 2024).

Some states require tutoring to accelerate student learning. For example, Ohio state law requires the Ohio Department of Education and Workforce to provide school districts with a list of approved vendors of high-quality tutoring programs that meet the criteria aligned with these research-based standards (Ohio Revised Code, 2022). Furthermore, the Ohio Revised Code indicates that opportunities for high-dosage tutoring must be provided for students who have not achieved proficiency in third-grade reading (Ohio Revised Code, 2023). As policy initiatives emphasize high-impact tutoring, it is important to align efforts with established effective tutoring practices (Robinson et al., 2024) and identify more opportunities to meet students’ learning needs through high-impact tutoring.

1.3. Evidence-Based Practices

Evidence-based practices (EBPs) are instructional strategies established to be effective for most students through rigorous research methods (Fuchs et al., 2021). The tutoring model described in the study begins with a framework of EBPs drawn from What Works Clearinghouse’s Practice Guides. Preservice teachers serving as tutors received university coursework in these foundational practices that are applicable across curricula.

For reading, EBPs include phonemic awareness, word recognition, reading connected texts, academic language in grades K-3 (Foorman et al., 2016) and explicit fluency, vocabulary, and comprehension instruction in grade 4 and beyond (Vaughn et al., 2022). Tutors developed plans for each group related to students’ academic needs within established systematic instruction routines that have the strongest evidence of effectiveness. Primary lessons focused on developing phonemic awareness through recognizing the segments of sounds in speech, learning how sounds link to letters, decoding words, analyzing word parts, and writing words (Foorman et al., 2016). Intermediate lessons focused on advanced word study, with an explicit routine for decoding multisyllabic words and instruction in morphemic analysis for generative vocabulary growth (Vaughn et al., 2022).

Furthermore, tutors modeled explicit comprehension strategies with guided, collaborative, and independent practice opportunities. Multiple strategy instruction is important for comprehension (National Reading Panel, 2000); tutors provided scaffolded support for comprehension-building practices related to background knowledge, vocabulary, questioning, summarizing, and comprehension monitoring (Vaughn et al., 2022). All students had the opportunity to apply their learning in connected text that incorporated grade level content at an appropriately complex level. When appropriate, students reread texts to develop fluency (Vaughn et al., 2022) with corrective feedback on oral reading to increase accuracy (National Reading Panel, 2000).

For mathematics, evidence-based practices include systematic instruction, representations, mathematical language, number lines, word problems, and timed activities (Fuchs et al., 2021). Using systematic instruction to teach math to struggling learners delivers positive math outcomes (Bryant et al., 2016). Therefore, math tutoring plans used systematic instruction that consisted of (1) identifying learner intentions, (2) sharing learning intentions to inform students’ academic progress, (3) providing explicit models through visuals and concrete representations during guided practice, (4) providing independent practice while checking for understanding, (5) providing closure with feedback, and (6) evaluating student performance to make instructional decisions (Allsopp et al., 2018).

This process aligns with the best practices identified by the National Council of Teachers of Mathematics to increase students’ mathematical proficiency (National Council of Teachers of Mathematics, 2014). For each component, the tutor provided evidence-based instruction emphasizing problem-solving tasks, using visuals, focusing on being strategic in reasoning, justifying answers, and engaging in discourse around the mathematical tasks (Allsopp et al., 2018). This emphasis on problem-solving, reasoning, discourse, and visuals engages students with the Standards for Mathematical Practice (Koestler et al., 2013).

1.4. High-Quality Instructional Materials

Research, vetted by the What Works Clearinghouse, has established that curriculum choices have measurable impacts on students’ academic success in reading (Borman et al., 2008) and mathematics (Hirschhorn, 1993). High-quality instructional materials (HQIM) are commonly defined as those that align to standards, have clear learning outcomes, reflect evidence-based practices, are content rich and culturally and linguistically relevant, and provide all necessary teacher and student materials (Accelerated Learning Work Group, 2023). HQIM can reduce workload for new teachers because they are already sequenced in ways to help students meet standards (Kwok et al., 2024). The nonprofit organization EdReports provides evaluations of curriculum materials to enable districts to select high-quality materials to meet their students’ needs (EdReports, n.d.). However, it is not enough to simply have the materials; schools must carefully consider implementing them, including supporting educator professional development (Short & Hirsh, 2020).

Curriculum is considered a key component in whether students have equitable access to high-impact tutoring (White et al., 2022). The NSSA Tutoring Program Design Principles (Robinson et al., 2024) indicate the importance of aligning the HQIM used in classroom instruction with the tutoring materials. Research suggests that using HQIM aligned with grade-level standards is more effective than using simplified or remedial material (Zimmer et al., 2010). Preservice educators often do not have the opportunity to interact with curriculum materials before their student teaching experience and so can benefit from learning about HQIM in their preparation program. Training modules created by Deans for Impact (2023), used by tutors in this study, provide an introduction to HQIM, use the framework of lesson internalization to develop familiarity with a curriculum, develop awareness of how to use HQIM to engage students in effortful thinking, and establish how an understanding of cognitive load is essential for supporting all learners. By completing these modules, the tutors were more prepared to engage with HQIM provided in the schools.

2. Materials and Methods

2.1. Research Aims

Partnerships between schools and universities can leverage resources to positively impact all stakeholders and student learning (Jackson et al., 2018). In 2022, our university received a grant from the Ohio Department of Education to develop, implement, and evaluate a high-impact tutoring program that leverages preservice teachers as tutors. Through this grant, preservice elementary teachers served as tutors in reading and mathematics in partner schools. The preservice teachers learn about building relationships, the importance of high-quality instructional materials, and how to facilitate small group work using evidence-based practices. Course instructors review tutoring plans for preparation, and tutors receive two observations with feedback during their tutoring placements. These procedures, coupled with ongoing professional learning through coursework, provided support for implementation fidelity. The purpose of this paper is to describe the early effectiveness of the tutoring model on student achievement in reading and mathematics. Specifically, the research questions we will answer are as follows. (1) What outcomes did high-impact tutoring have on elementary students’ reading achievement? (2) What outcomes did high-impact tutoring have on elementary students’ mathematics achievement?

2.2. Participants

The research participants were 711 students (129 tutored; 582 non-tutored) in grades 2–5 in public elementary or intermediate schools from one school district in the Midwest region of the United States. The students were selected for tutoring by their classroom teachers, with a suggestion that tutoring would focus on students who needed additional support. Secondary data analysis focused specifically on students’ benchmark test scores to understand how tutoring groups progressed over time.

Table 1. Student Demographics.

Grade Level	Characteristic	Tutored	Non-Tutored	Total
2nd	Gender
	Female	21	84	105
	Male	19	88	107
	Race
	Asian	0	4	4
	Black	6	14	20
	Hispanic	4	18	22
	White	25	123	148
	Two or more	13	5	18
	Individualized Education Plan	4	29	33
3rd	Gender
	Female	15	51	66
	Male	21	66	87
	Race
	Asian	0	2	2
	Black	5	14	19
	Hispanic	2	15	17
	White	28	76	104
	Two or more	10	1	11
	Individualized Education Plan	6	30	36
4th	Gender
	Female	19	69	88
	Male	14	79	93
	Race
	Asian	0	2	2
	Black	4	28	32
	Hispanic	3	11	14
	White	23	95	118
	Two or more	3	12	15
	Individualized Education Plan	5	28	33
5th	Gender
	Female	10	76	86
	Male	10	69	79
	Race
	Asian	0	0	0
	Black	2	18	20
	Hispanic	1	9	10
	White	14	104	118
	Two or more	3	14	17
	Individualized Education Plan	5	30	35
TOTAL	Gender
	Female	65	280	345
	Male	64	302	366
	Race
	Asian	0	8	8
	Black	17	74	91
	Hispanic	10	53	63
	White	90	398	488
	Two or more	29	32	61
	Individualized Education Plan	20	117	137

shows participant demographics.

2.3. Setting

High-impact tutoring occurred in two primary schools (grades K–3) and one intermediate school (grades 4–5) from one school district in the Midwest region of the United States. The tutors were preservice teachers supported by coursework, faculty, and graduate assistants with training, feedback on tutoring plans, and observations with actionable feedback. The tutors provided thirty-minute tutoring sessions, three days a week, to groups of no more than four students per tutor. Reading instruction occurred in grades 2–3 and mathematics instruction in grades 4–5, as these were the grades and subjects requested by the district partner. The tutoring materials aligned with the core instruction. For example, in mathematics, the core instruction used the Bridges^® mathematics curriculum and tutoring used Bridges Intervention^®.

2.4. Data Sources

The deidentified data was shared with the research team by the district partners. The data included demographic information (e.g., ethnicity, IEP status, and gender), student Star Assessments™ reading and mathematics scaled scores, and percentile rank test scores from beginning-, middle-, and end-of-the-year assessments. Star Assessments™ “are backed by extensive research, and for years have been favorably reviewed as reliable, valid, and efficient by independent groups, including the National Center on Intensive Intervention and the National Center on Response to Intervention” (Renaissance Learning, 2020). This computer adaptive test was selected by the state as an approved assessment to evaluate the state learning standards and support data-drive instructional decisions (Renaissance, 2020).

Star Assessments™ provides an introduction, content and item development, validity, norming information, and score definitions in their 2024 Mathematics Technical Manual (Renaissance Learning, 2024a) and 2024 Reading Technical Manual (Renaissance Learning, 2024b) for all tested grades (K-12). Descriptive statistics for scaled scores by grade are provided in the manual for each norming sample in mathematics and reading. This study focuses on scaled scores, as they serve as appropriate benchmarks for each student and provide multiple data points throughout the school year to measure growth.

2.5. Data Analysis

Following the suggestion of Hopkin et al. (2015), due to sample size, we are limiting inferential analyses and focusing more on descriptive statistics, rich description of trends, and data visualizations. Moreover, a statistical comparison between tutored and non-tutored groups (e.g., ANCOVA) is not appropriate due to the large differences in sample sizes and great variation within the non-tutored group (Tabachnick & Fidell, 2019). Therefore, tutored and non-tutored student scores were compared using growth on benchmarks to understand the impact of the tutoring program on student learning.

3. Results

3.1. Research Question 1: Reading Achievement

For grade 2, over 1.1 million students across the United States were assessed in computing Star Reading assessment growth norms (Renaissance Learning, 2024b). For the norming sample, the beginning-of-the-year (fall) scaled scores for students resulted in an overall average reading score of 864 (SD = 93), with a median score of 872. The end-of-year (spring) scaled scores in the norming sample showed an overall average reading score of 927 (SD = 87), with a median score of 940. Accordingly, the norming sample demonstrated an expected average growth of approximately 60 points from the beginning- to the end-of-the year assessments (Renaissance Learning, 2024b, p. 102). See Figure 1.

In our sample of second-grade students, the non-tutored students had a wide range of scores ranging from the 1st percentile to the 98th percentile (scaled score range: 615–1039; median: 883). The non-tutored students began the school year slightly above the norming sample, yet remained within one standard deviation (M = 872.67). The non-tutored students grew an average of 85.66 points from the first benchmark test to the last benchmark test (M = 958.33), and their growth was higher (within one standard deviation) than the normed sample. See Figure 1.

The tutored students were approximately equivalent to the non-tutored students, with scores ranging from the 1st to the 97th percentiles (scaled score range: 615–1024) and a median of 882. This group of students began the school year slightly below but within one standard deviation of the norming sample. The tutored students showed an average growth of 82.29 points from the first benchmark test (M = 853.61) to the last benchmark test (M = 935.9), ending slightly above, but still within, one standard deviation of the normed sample. Interestingly, the percentile ranks among the tutored students maintained the same range, with end-of-the-year scores falling between the 1st and 97th percentiles, while the non-tutored group expanded slightly to between the 1st to 99th percentiles. However, the slope of the growth lines indicates that the non-tutored students grew at a faster rate than tutored students, gaining 1.7 more points per benchmark test. Figure 1 below shows the growth trends of the two groups.

For grade 3, growth norms for the Star Reading assessment were computed using data from over 1.2 million students across the United States (Renaissance Learning, 2024b). For the norming sample, the beginning-of-the-year (fall) scaled scores for students resulted in an overall average reading scaled score of 929 (SD = 87), with a median score of 941. The end-of-the-year (spring) scaled score samples in the norming sample resulted in an overall average reading score of 970 (SD = 84), with a median score of 983. As such, the overall expected average growth from the beginning- to the end-of-the-year assessments in the norming sample was approximately 50 points (Renaissance Learning, 2024b, p. 102).

In our sample of third-grade students, non-tutored students had a wide range of scores ranging from the 1st percentile to the 99th percentile (scaled score range: 615–1090; median: 968). Non-tutored students began the school year higher, yet within one standard deviation of the norming sample (M = 943.57). The non-tutored students grew an average of 52.48 points from the first benchmark test to the last benchmark test (M = 996.05), and their growth was higher (within one standard deviation) than that of the normed sample. See Figure 2.

The tutored students had less variation than the non-tutored students, from the 3rd to 95th percentiles (scaled score range: 804–1056; median: 949). Tutored students began the school year slightly below but within one standard deviation of the norming sample (M = 949.85). The group of tutored students grew an average of 54.51 points from the first benchmark test to the last benchmark test (M = 1004.36), ending above, but within one standard deviation of the normed sample. The slope of the growth lines shows that the non-tutored students grew at a faster rate than the tutored students, growing 1.02 points more per benchmark test. The range of percentile rank achieved between the tutored students narrowed in a positive direction, with the end-of-the-year scores falling between the 9th and 94th percentiles; interestingly, the non-tutored group range decreased to scores in the 1st to 97th percentile range. Figure 2 shows the two groups’ growth.

To answer our first research question, “What outcomes did high-impact tutoring have on elementary students’ reading achievement?,” the outcomes of high-impact tutoring on reading achievement were mixed. Second-grade students in the tutored group of our sample demonstrated growth at a rate greater than the national average used to create the scaled scores but not greater than the non-tutored group (Figure 1). Third-grade students in the tutored group demonstrated growth at a rate that exceeded the national average and the non-tutored group in our sample (Figure 2).

3.2. Research Question 2: Mathematics Achievement

For fourth-grade students, over 730,000 students across the United States were assessed in computing growth norms (Renaissance Learning, 2024a). For the norming sample, the beginning-of-the-year (fall) scaled scores for students resulted in an overall average of 967 (SD = 66), with a median score of 974. The end-of-the-year (spring) scaled samples in the norming sample resulted in an overall average of 1017 (SD = 73), with a median score of 1026. The overall expected average growth from the beginning- to the end-of-the-year in the norming sample was approximately 50 points (p. 86). See Figure 3.

In our sample of fourth-grade students, non-tutored students had a wide range of scores ranging from the 1st percentile to the 99th percentile (scaled score range: 767–1196; median: 990). The non-tutored students began the school year scoring slightly higher than, yet within one standard deviation of, the norming sample (M = 975.47). The non-tutored students grew an average of 37.262 points from the first benchmark test to the last benchmark test (M = 1012.732). This group of students also scored lower than the norming group, but within one standard deviation (see Figure 3).

The tutored students had smaller variation than the non-tutored students, from the 2nd to 85th percentiles (scaled score range: 827–1025; median: 940). The tutored students began the year scoring much lower than the norming sample (M = 935.88), but still within one standard deviation of the norming sample. The tutored students grew an average of 41.03 points from the first benchmark test to the last benchmark test, aligning more closely with normed expected growth. Meanwhile, the tutored group’s end-of-the-year scores were lower than the non-tutored group (M = 976.91), this average was still within one standard deviation of the normed sample. The percentile rank achieved between the tutored students expanded in a positive direction, with the end-of-the-year scores falling between the 2nd and 94th percentiles; the non-tutored group stayed in the 1st to 99th percentile ranges. Additionally, the slope of the growth lines shows that tutored students grew at a faster rate than non-tutored students, growing 1.884 points more per benchmark test. Figure 3 below shows the two groups’ growth.

For fifth-grade students, over 730,000 students across the United States were assessed in computing growth norms (Renaissance Learning, 2024a). For the norming sample, the beginning-of-the-year (fall) scaled scores for students resulted in an overall average of 1006 (SD = 70), with a median score of 1013. The end-of-the-year (spring) scaled samples in the norming sample resulted in an overall average of 1046 (SD = 76), with a median score of 1055. As such, the overall expected average growth from the beginning- to end-of-the-year assessments in the norming sample was approximately 50 points (Renaissance Learning, 2024a). See Figure 4.

In our sample of fifth-grade students, non-tutored students had a wide range of scores, encompassing the full range of student possible performance from the 1st percentile to the 99th percentile (scaled score range: 776–1196; median: 1019). Non-tutored students began the school year with slightly higher scaled scores, yet within one standard deviation of the norming sample (M = 1011.48). The non-tutored students grew an average of 36.34 points from the first benchmark test to the last benchmark test (M = 1047.82). This group of students was within one point of the norming group. The tutored students had smaller variation than the non-tutored students, from the 1st to 63rd percentiles (scaled score range: 806–1031; median: 953). This group of students began the year much lower than the norming sample (M = 944.28) but still within one standard deviation of the norming sample.

The tutored students grew an average of 45.87 points from the first benchmark test to the last benchmark test, aligning more closely with the normed expected growth. Meanwhile, the tutored group’s end-of-the-year scores were lower than the non-tutored group and the norming sample, (M = 990.15), this average was still within one standard deviation of the normed sample (see Figure 4). Demonstrating further success, the percentile rank achieved between the tutored students expanded in a positive direction, with the end-of-the-year scores falling between the 2nd and 91st percentiles; the non-tutored group stayed in the 1st to 99th percentile ranges. Additionally, the slope of the growth lines shows that the tutored students grew at a faster rate than the non-tutored students, growing 4.766 points more per benchmark test. Figure 4 shows the two groups’ growth.

To answer our second research question, the outcomes of high-impact tutoring on mathematics achievement were positive. The fourth-grade students in the tutored group began and ended the year with scores lower than the national average and the non-tutored group; however, their growth rate was similar to the norm group and exceeded that of the non-tutored group (see Figure 3). Fifth-grade students in the tutored group demonstrated growth at a rate similar to the national average that exceeded the growth rate of the non-tutored group in our sample (see Figure 4).

4. Discussion

The results demonstrate that this pilot tutoring model positively impacted student achievement for those receiving tutoring. In three out of four groups (third-grade reading, fourth- and fifth-grade mathematics), students in the tutoring groups grew consistently at rates that exceeded their non-tutored peers. In the case of second-grade reading, the students who received tutoring services did not grow more than non-tutored students but still grew more than the national average assessed to create the scaled scores. These trends suggest that aligning the model design with prior research recommendations for high-impact tutoring regarding tutor selection and support (Nickow et al., 2024), group size (Nickow et al., 2020), recommended dosage, and alignment with core classroom instruction (Robinson et al., 2024) has led to similarly successful outcomes.

As the majority of tutoring research has evaluated early literacy (K-2) efforts (Nickow et al., 2024), this study adds to the literature on tutoring for students in the intermediate grades for both reading and mathematics. Partner educators used school data to determine a need for reading tutoring in grades 2–3 and mathematics in grades 4–5. Although these decisions were made to work in the context of district partners, the focus aligns with meta-analytical research that shows stronger effect sizes for reading tutoring for younger students (Vadasy et al., 2007).

4.1. Tutors’ Use of Evidence-Based Practices

Curriculum and pedagogical approach may have an impact on the efficacy of tutoring programs (Nickow et al., 2024). In our context, second- and third-grade teachers set different instructional focuses for tutors. Second-grade tutors were primarily assigned groups who needed additional practice with basic phonics skills. The lessons were based on the scripted curriculum provided to the whole class earlier in the day or week. Third-grade tutors had a wider range of skills to address in fluency, comprehension, and vocabulary. These lessons were also aligned with classroom high-quality instructional materials but reflected more varied teaching practices. Across grade levels, the tutors adjusted for individual and group needs, but these larger focus areas persisted. Previous research has indicated that tutors may experience tensions when navigating a need to implement heavily scripted intervention curricula with fidelity and also collaborate with classroom educators to ensure instructional alignment (Hallgren et al., 2017). These tensions may explain some of the differences in tutoring sessions between the second- and third-grade groups and, subsequently, their outcomes.

Regardless of grade level, the tutors implemented evidence-based practices including explicit instruction in word recognition (Foorman et al., 2016), fluency, vocabulary, and comprehension strategies (Vaughn et al., 2022). Reading tutoring programs that use explicit instruction in grades 2 and 3 have had significant effects when focused on basic phonics (Vadasy et al., 2007) and word study through more advanced structural analysis (Vadasy et al., 2006). There is some evidence that comprehension interventions and mixed-strategy interventions are more effective with students beyond first grade (Suggate, 2010). However, later studies complicate whether foundational skills or mixed-component interventions are more effective for second- and third-grade students (Denton et al., 2022).

4.2. Tutors’ Use of High-Quality Instructional Materials

In mathematics, tutors also focused on evidence-based practices (Fuchs et al., 2021). Mathematics tutors had a unique experience in that they utilized consistent curriculum materials that aligned between general classroom instruction and the intervention setting. Mathematics is also the subject area in which tutors received extensive training in using HQIM and in the curriculum specifically. The curriculum focused on using consistent representations. With all school partners using this curriculum, the tutors entered their placement with knowledge of the curriculum and how to enact it with the use of evidence-based practices.

The common mathematics curriculum also allowed tutors to experience HQIM. Understanding how well-designed curriculum can be enacted is an ongoing struggle in teacher education; however, quality clinical experiences supported by coursework positively impacts teacher candidates’ skills and knowledge (Polly & Colonnese, 2022). In this experience, the tutors learned about HQIM for mathematics, evidence-based practices, and enacted the curriculum in clinical experiences with support from coursework. Due to the curricular alignment, the tutors were also able to connect intervention with core instruction. This combination of consistently aligned HQIM (Robinson et al., 2024) and tutors with professional learning in the curriculum (Short & Hirsh, 2020) may relate to tutored students’ stronger growth in math outcomes than in reading.

4.3. Limitations

Despite the limitations posed by the unequal sample sizes between the tutored and non-tutored groups, the preliminary results of this pilot study indicate that our high-impact tutoring model was effective in supporting students’ growth in reading (second and third grades) and mathematics (fourth and fifth grades). As the sampling of tutored students was not random, we suspect that selection bias impacted the true improvement of scores. For instance, the range of scores for the non-tutored students had a much wider range of scores at any point: fifth-grade mathematics scores for the non-tutored group at the beginning of the year ranged from the 1st to 99th percentiles, while the tutored group ranged from the 1st to 66th percentiles. Meanwhile, at the end of the year, the non-tutored group continued to range from the 1st to 99th percentiles and the tutored group resulted in scores expanding positively from the 2nd to 91st percentiles, increasing the range of student scores. Comparing the students in our sample to the normed data of the Star mathematics and reading assessments provides additional comparison to the success of the program. Examination of model implementation provides some insight into why some areas (e.g., second-grade reading) did not reflect the desired accelerated learning and provides implications for program design and future research.

4.4. Implications

A unique feature of this tutoring model is the collaboration between preservice teachers and the cooperating teachers in whose classroom they were working. To support the development of this partnership, several implementation factors were left to the schools to determine. We shared research-based recommendations for the best practices and allowed school personnel to make decisions that best suited their contexts. While this flexibility proved valuable to school partners and the ongoing partnership, it did create some challenges for researching outcomes. These contextual variations are important to consider in the further development and research of this program. The results of this study bring two implementation factors to light.

First, it is important to consider how students were identified as eligible to receive tutoring. The teachers were allowed to determine their own criteria for identifying the students. Teachers or grade-level teams used different processes, which may have been responsible for the group differences. We can infer that teachers used different criteria for identifying students for tutoring. In second, fourth, and fifth grades, the students in the tutoring group began with lower baselines than their non-tutored peers. This suggests that most teachers identified students for tutoring as those they deemed in need of remediation. In the third grade, the students in the tutoring group began with a higher baseline than their non-tutored peers. This indicates that the third-grade teachers chose to work with higher-need students themselves and assigned tutors to work with students in need of enrichment. The lack of clarity in these patterns suggests that greater attention on student selection is needed to evaluate effects.

A second factor to explore is actual tutoring dosage. Although the tutors planned to provide a minimum of 30 min of small-group instruction three times a week, some practicalities interrupted these plans. As in standard operation of elementary schools in the United States, instruction was disrupted by inclement weather, special events, and state-mandated testing. Combined with student and teacher absences, more information is needed to make stronger claims about tutoring effectiveness. This pilot study included some measures of dosage, but data collection was not sufficiently robust to analyze these metrics.

5. Conclusions

Overall, the results of this pilot study demonstrate the potential of this model for positively impacting student learning in reading and mathematics. Students who received tutoring regularly demonstrated more academic growth than those who did not. Future research will include additional data collection with a larger sample size to allow for more advanced analysis. Furthermore, this pilot study affirms that, like tutoring provided by educational professionals, tutoring provided by preservice teachers can impact student learning. Universities with educator preparation programs may consider taking advantage of their access to a cadre of preservice teachers who, when provided with initial and ongoing support, can serve as effective tutors.