Louisiana State University
LSU Scholarly Repository
LSU Doctoral Dissertations Graduate School
10-27-2023
THE EFFECTS OF AMIRA LEARNING ON LITERACY
DEVELOPMENT IN EARLY CHILDHOOD EDUCATION
Caroline C. Tolentino
Louisiana State University at Baton Rouge
Follow this and additional works at: https://repository.lsu.edu/gradschool_dissertations
Recommended Citation
Tolentino, Caroline C., "THE EFFECTS OF AMIRA LEARNING ON LITERACY DEVELOPMENT IN EARLY
CHILDHOOD EDUCATION" (2023). LSU Doctoral Dissertations. 6280.
https://repository.lsu.edu/gradschool_dissertations/6280
This Dissertation is brought to you for free and open access by the Graduate School at LSU Scholarly Repository. It
has been accepted for inclusion in LSU Doctoral Dissertations by an authorized graduate school editor of LSU
Scholarly Repository. For more information, please contactgradetd@lsu.edu.
THE EFFECTS OF AMIRA LEARNING ON LITERACY
DEVELOPMENT IN EARLY CHILDHOOD EDUCATION
A Dissertation
Submitted to the Graduate Faculty of the
Louisiana State University and
Agriculture and Mechanical College
in partial fulfillment of the
requirements for the degree of
Doctor of Philosophy
in
The School of Education
by
Caroline C. Tolentino
B.A., University of the Philippines, 2004
M.Ed., Louisiana State University, 2017
Ed.S., Louisiana State University, 2020
December 2023
© 2023/copyright
Caroline C. Tolentino
All rights reserved
ii
ACKNOWLEDGEMENTS
I would like to express my deepest gratitude to the many individuals who have supported
and guided me throughout the journey of completing this dissertation. First and foremost, I am
profoundly thankful to my dissertation advisor, Dr. Cynthia DiCarlo, for her unwavering support,
invaluable expertise, and endless patience. Her guidance and mentorship have been instrumental
in shaping the direction of this research. I extend my heartfelt appreciation to the members of my
dissertation committee, Dr. Eugene Kennedy and Dr. Jennifer Qian, for their constructive
feedback, insightful suggestions, and dedication to ensuring the quality of this work. To Dr.
Andrew Webb and Dr. Shawndaya Thrasher, for their interest in my work and willingness to
serve on my committee.
I am grateful to my friends who provided me with a network of support and camaraderie
throughout this journey. Their encouragement and shared experiences have lightened the load.
I am indebted to my family for their boundless encouragement and understanding during
the demanding phases of this academic pursuit. Their unwavering belief in my abilities has been
my constant motivation. I am truly fortunate to have had such a strong support system, and for
that, I am deeply thankful.
To God Almighty, I am forever grateful.
iii
TABLE OF CONTENTS
ACKNOWLDEGEMENTS ....................................................................................................... iii
ABSTRACT ................................................................................................................................ v
CHAPTER ONE. INTRODUCTION ......................................................................................... 1
CHAPTER TWO. LITERATURE REVIEW ............................................................................. 11
CHAPTER THREE. METHODOLOGY ................................................................................... 40
CHAPTER FOUR. RESULTS ................................................................................................... 56
CHAPTER FIVE: DISCUSSION AND IMPLICATIONS ....................................................... 67
APPENDIX A. CENTRAL TENDENCY TABLES AND FIGURES ……………..……........ 83
APPENDIX B. INSTITUTIONAL REVIEW BOARD ............................................................. 87
APPENDIX C. INFORMED CONSENT FORMS ….................................................................88
REFERENCES ........................................................................................................................... 92
VITA .......................................................................................................................................... 101
iv
ABSTRACT
The aim of the present study was to investigate the effects of an Artificial Intelligence
software, Amira Learning, on literacy development in early childhood education. Specifically,
the research investigated the impact of Amira Learning software usage, the feedback it provided,
and teachers' perspectives on its alignment with the Science of Reading. The study employed
quantitative analyses of oral reading fluency outcomes and compared the pretest and posttest
scores of the students after the 6-week usage of the software.
v
CHAPTER ONE. INTRODUCTION
The Covid-19 pandemic posed a challenge in education that no nation in the world was
prepared to tackle. It took a toll on human resources, risking physical and mental wellbeing of
students and teachers, disrupted the normal teaching and learning cycle, and had everyone
thinking of ways to safely continue educating students. During this unprecedented event, leaders
at the school, district, and state levels all turned to educational technology to try to create
normalcy for teachers and students.
Although there were conflicting thoughts on the acceptance of educational technology
(Straker, et al., 2018), the threats caused by the pandemic to the educational system opened the
doors to what it has to offer. Technology was used in varied ways in education, even in early
childhood education (ECE). It was used to deliver remote learning and to instruct in real time, to
electronically store resources, assessments, and data, and for interventions or supplemental
activities through applications and software. More recently, artificial intelligence (AI) for
educational use also garnered attention from students, teachers, school leaders, and stakeholders
because of its promising capabilities. The power of AI in education (AIEd) lies in its ability to
allow teachers and students to use machines that mimic human intelligence in computer
programs (McCarthy, 2007).
In ECE where high-quality learning experiences and interactions are a golden standard,
programs that can replicate human intelligence such as intelligent tutoring systems (ITS) sound
encouraging especially because Covid-19 depleted human capital in education with schools not
allowing volunteers and members of the community to provide additional services for students.
Covid-19 has been predicted to cause learning loss (Sabates, et al.., 2021) and believed to widen
1
already existing learning gaps especially for the most vulnerable groups (Hoofman & Secord,
2021; Van Lancker & Parolin, 2020). The present study aimed to investigate the effects of an AI
program, Amira Learning software, on literacy development in early childhood education. Amira
Learning software claims to help young learners develop their literacy skills through constant
practice and interaction with the animated tutor, Amira.
Significance of Study
The significance of the present study was aimed at understanding the potential impact of
using Amira Learning software on young learners' literacy development. As AIEd continued to
play an increasingly significant role in the learning process (Broda & Frank, 2015; Touretzky, et
al., 2019; Yang, 2022) and as educators drew more interest in using ITS in educational setting, a
number of researchers offered meta-analysis and review of ITS and its effects on learning
outcomes (Ma, et al. 2014; Mousavinasab, et al., 2021; Nickow, et al., 2020,; Su & Yang, 2022).
The present study aimed to explore some of its important implications.
The findings of the present study may shed light on the effectiveness of an ITS in
improving overall literacy of young learners as evidenced by improvement in oral reading
fluency (ORF). This knowledge can be valuable for educators and educational institutions
considering the integration of such technologies into their instructional practices. With Amira
Learning software’s promise of literacy growth, the present study can also contribute to
evidence-based literacy instruction strategies and may provide insights into how ITS can support
and enhance students' reading abilities. If the study revealed a positive relationship between time
spent using Amira Learning software and oral reading fluency, it could emphasize the
importance of personalized learning experiences tailored to individual students' needs and
2
learning preferences, through means other than the traditional, human tutor. Based on what the
findings will be, educational institutions may consider allocating resources and time to
incorporate Amira Learning software or similar technologies as part of their literacy curriculum.
Purpose of the Study
The purpose of this present study was to investigate the potential relationship between the
amount of time students spend using Amira Learning software and their literacy development as
measured by oral reading fluency (ORF) achievement levels. It also aimed to investigate how
effective Amira Learning software’s feedback was in improving students’ oral reading fluency.
The present study intended to collect and analyze survey data from teachers to understand their
views on how well Amira Learning software incorporates evidence-based reading instruction
methods outlined in the Science of Reading (SOR) framework. The present study hoped to have
achieved the following objectives:
Examine the Use of Amira Learning Software. The present study gathered data on the
time spent by students using Amira Learning software while practicing reading grade level texts.
This data will help determine the extent of exposure to this ITS. The amount of feedback in
minutes that Amira Learning software provided will also be collected and compared to students’
oral reading fluency levels.
Assess Oral Reading Fluency Levels. The present study aimed to measure students' oral
reading fluency levels before and after using Amira Learning software using the assessment and
data management system, DIBELS Next (www.acadiencelearning.org, n.d.). By comparing the
fluency levels to Amira Learning software usage and feedback given, the study sought to identify
any potential relationship between the variables.
3
Analyze the Relationships. The present study statistically analyzed the data collected to
determine whether there was a significant relationship between the time spent using Amira
Learning software and learners' oral reading fluency achievement levels as well as amount of
Amira Learning software feedback and oral reading fluency achievement levels.
Interpret Teacher’s View on Amira Learning software and SOR Alignment. The
present study investigated and assessed the level of agreement among teachers regarding Amira
Learning software's alignment to SOR and its effectiveness in fulfilling the SOR objectives
through an online survey.
Contribute to Educational Research. The present study investigated the impact of
Amira Learning software, an AI-enabled ITS that claims to be aligned with SOR, to literacy
development of students in first grade level in a suburban public school in Louisiana. Being
aligned with the state’s literacy vision and with the forward trend in AIEd, it was hoped that this
study will set the foundation for further, wider scale research in the future. The impact of Amira
Learning software to literacy development and teachers' perspectives on its alignment to SOR
can have implications for the successful integration of AIEd, in the form of ITS, helping
educators and researchers make informed decisions about the use of this technology in literacy
instruction and supporting evidence-based practices in educational settings.
Research Questions
In order to investigate the effects of Amira Learning software, an Artificial Intelligence
(AI) program, on the literacy development in early childhood education, the following questions
supported this study:
4
RQ1: What is the relationship between the time spent using Amira Learning software and
learners' oral reading fluency achievement levels?
RQ2: How effective is the feedback provided by Amira Learning software in improving
students’ oral reading fluency?
RQ3: What is the level of agreement among teachers regarding Amira Learning
software’s alignment to the Science of Reading and its effectiveness in fulfilling the Science of
Reading objectives?
Theoretical Framework
To fully explore the relationship of Amira Learning software and early learners' reading
skills, it was essential to consider two overarching concepts: the mastery learning (Bloom, 1968)
and skilled reader (Gough & Tunmer, 1986; Scarborough, 2001; National Reading Panel, 2000).
The mastery learning theory aims to address individual learning differences, promote
deeper understanding, and ensure that students are well-prepared for more advanced topics. It
has been influential in shaping the field of education, particularly in the design of personalized
and competency-based learning models (Guskey, 2005). Mastery learning is regarded as a means
to promote educational fairness by enabling every student to attain a high level of expertise in the
subject matter.
The Skilled Reader theory, often associated with cognitive psychology and reading
research, seeks to understand and describe the cognitive processes and skills that proficient
readers use when they read and comprehend text. It is supported by three distinct models , the
Simple View of Reading (SVR) (Gough & Tunmer (1986), Scarborough's Reading Rope (2001),
5
and the Five Pillars of Reading (National Reading Panel, 2005), which collectively provide a
comprehensive insight into the various components involved in the reading process. Although
each of these models has its own unique characteristics, they all share a fundamental recognition
that reading is a complex undertaking characterized by two essential elements: decoding and
comprehension. Additionally, these frameworks all acknowledge that reading skills develop in a
hierarchical fashion, with foundational skills forming the building blocks for more advanced
ones.
Limitations & Delimitations
Although the real-world, quasi-experimental design has several advantages, specifically
when conducting experimental control is challenging or unethical (Grimshaw, et al., 2000), it has
some limitations that may affect the results of the present study. One of the limitations included
not involving random assignment of participants to groups, which can lead to potential biases
and can make it difficult to establish causal relationships between the independent variable and
the outcomes. The non-random assignment of participants to groups may also result in
systematic differences between the groups, affecting the internal validity of the study. Another
limitation involved the participants selected based on extreme scores at the beginning of the
study who may naturally move toward the average on subsequent measurements, leading to
regression to the mean and potentially misinterpretation of the intervention's impact (Harris, et
al., 2006).
Feasibility and practicality are two of the delimitations of quasi-experimental studies. It
is often more practical and feasible than true experimental designs, especially in situations where
random assignment to groups is not possible or ethical. This makes them more accessible for
researchers working in real-world settings with existing groups or natural occurrences. This
6
design also allows researchers to examine cause-and-effect relationships without compromising
ethical principles (Grimshaw, et al., 2000). Because the topic of AIEd in the form of ITS in early
childhood education has not been fully explored yet, quasi-experimental studies can serve as a
basis for generating hypotheses that can later be tested using more controlled experimental
designs. Quasi-experimental design is best suited for the present study for these reasons.
Overall, quasi-experimental designs offer researchers a valuable approach to study causal
relationships in situations where true experimental control is not feasible (Grimshaw, et al.,
2000; Harris, et al., 2006). By understanding the strengths and limitations of these designs,
researchers can use them effectively to address important research questions and contribute to
evidence-based practices in various fields.
Definition of Key Terms
The following terms were used as they applied to this study. They are defined hereafter.
Artificial Intelligence (AI). McCarthy (2007), defined AI as “the science and engineering
of making intelligent machines, especially intelligent computer programs'' (p. 2). This definition
seems to encompass a broad description of what AI truly is. It can be challenging to find a
specific, direct definition for the term since AI has been around for more than a decade in many
forms and applications (Lindner & Romeike, 2019). The common description of AI lies in its
ability to understand and mimic human actions and thinking through reasoning, adapting,
recognizing patterns, making predictions, recommendations, and decisions to solve complex
problems (ISTE, 2019, 0:28; Lindner & Romeike, 2019; Unicef, 2020).
Artificial Intelligence in Education (AIEd). AIEd refers to the application of artificial
intelligence (AI) technologies and techniques in the field of education. It aims to enhance and
7
improve various aspects of the educational process by leveraging AI's capabilities to analyze
data, make predictions, and adapt to individual learners' needs (Artificial Intelligence and the
Future of Teaching and Learning Insights and Recommendations, 2023)
Intelligent Tutoring Systems (ITS). ITS is an advanced educational technology that uses
artificial intelligence (AI) and computer-based instruction to provide personalized and adaptive
learning experiences to students. ITS is designed to simulate the role of a human tutor, delivering
individualized instruction and support to learners based on their specific needs and progress
(Luckin, et.al., 2016).
Science of Reading (SOR). The Science of Reading is an evidence-based approach to
reading instruction that draws on extensive research from various disciplines, including cognitive
psychology, linguistics, neuroscience, and education. It seeks to understand how reading
develops in the brain and how best to teach reading skills effectively (Moats, 1999; Moats 2020;
Science of Reading: Defining Guide, 2022).
Early literacy skills. Early literacy skills are the fundamental abilities that young children
develop during their early childhood years, which are essential for future reading and writing
success. These skills serve as the foundation for becoming proficient readers and writers as
children progress through their education. The Science of Reading includes the following
measures of early literacy skills: phonemic awareness, phonics, fluency, vocabulary,
comprehension.
Dynamic Indicator of Basic Early Literacy Skills (DIBELS) Next edition. DIBELS Next
are a set of research-based procedures and measures for assessing the acquisition of early literacy
skills. They are designed to be short fluency measures used to regularly monitor the development
8
of early literacy skills and early reading skills (www. Acadiencelearning.org, n.d.). In first grade,
DIBELS Next assess the measures of phoneme segmentation, nonsense word fluency, and oral
reading fluency. It uses four levels of risk to categorize students' performance on the assessment.
These risk levels help educators identify students who may need additional support or
intervention in developing their early literacy skills. The risk levels are typically classified as
follows: low risk (green and blue; students who are performing at or above grade level), some
risk (yellow; students who are performing slightly below grade level), and at risk (red; students
who are performing significantly below grade level).
Oral Reading Fluency (ORF). ORF as measured by DIBELS Next refers to a specific
assessment used to evaluate a student's reading proficiency. In the DIBELS ORF assessment, a
student is asked to read a grade-level passage aloud for one minute. The evaluator records the
number of words the student reads correctly within that time frame. It focuses on accuracy and
rate of reading.
Word Correct Per Minute (WCPM). WCPM is a metric used to measure ORF in
educational assessments, particularly in the context of DIBELS Next. WCPM represents the
number of words a student reads correctly in one minute during an ORF assessment. The
assessment typically involves having the student read a grade-level passage aloud, and the
evaluator records the number of words read accurately within the one-minute time frame.
Summary
This chapter presented the global challenges that Covid-19 posed in education. Amidst
the challenges, educational technology emerged as a crucial tool to continue teaching and
learning, although there were differing opinions on its acceptance, specifically in ECE.
9
Educational technology has advanced with AIEd which has opened a whole new way of
enhancing learning experiences for young students with human intelligence capabilities. An AI
program, Amira Learning software, promises to close learning gaps and support students’
literacy development using explicit instruction and tutoring that is aligned with SOR. As
Louisiana focused on improving the literacy rates of students from kindergarten to third grade
through SOR, the impact of Amira Learning software on literacy development was worth
investigating.
Chapter Two, a review of literature, the existing literature on the predicted effects of
Covid-19 on education, the role of educational technology in ECE in relation to the pandemic,
AI and ITS, Amira Learning software and SOR, as well as the theoretical framework for the
present study are explored.
In Chapter Three, details about the present study and how the topic will be investigated
are presented. The intended research paradigm, setting and context, ethical considerations, data
sources, data analysis, and research design are included.
10
CHAPTER TWO. LITERATURE REVIEW
In December 2019, the entire world was shocked by the news of a new virus, Covid-19,
that was highly contagious and caused infectious illness and sudden death. It claimed close to
seven million lives worldwide and infected more than half a billion people from all walks of life
(World Health Organization, 2023). In the first quarter of 2020, Covid-19 caused the world to
completely shut down, disrupting every industry and challenging every aspect of the society,
including the educational system. Since then, technology has come a long way and its adoption
in different industries has speeded as part of the responses to the pandemic (LaBerge, et.al.,
2020). This has also resulted in rapid advancements in educational technology. Currently, the
most popular yet most controversial educational technology is artificial intelligence in education
(AIEd). AIEd has raised both praises and disagreements especially in early childhood education
(ECE). Nonetheless, developers continued to attempt to create appropriate AIEd tools in ECE
(Broda & Frank, 2015; Touretzky, et al., 2019; Yang, 2022).
AIEd has made its way to many states. In Louisiana, the Department of Education
(LDOE) promoted the use of Amira Learning software, an AIEd Intelligent Tutoring System
(ITS) program to help English Learners (ELs) to master literacy skills necessary for school
success. Amira Learning software claims to provide effective tutoring to young learners and to
close learning gaps through instruction and feedback that are aligned with the Science of
Reading (SOR). With about 900 school districts in 15 countries using Amira Learning software
and a program that alleges to be backed by research (www.amiralearning.com, n.d.),
independently published studies on the effects of Amira Learning software did not seem to be
available on academic search engines yet. The present study aimed to investigate the effects of
an AI program, Amira Learning, on literacy development in early childhood education.
11
To better understand and explain the importance of this study, this review of literature
explored and analyzed current research associated with the effects of Covid-19 to the educational
system, the different views on educational technology and AIEd, studies involving ITS and
literacy development, and a brief overview of SOR. A wide range of scholarly sources including
academic journals, books, and reputable online databases were used to provide a comprehensive
understanding of this topic and to identify gaps or areas for future research. For the purpose of
fairness, Amira Learning software inhouse, company-funded research on their program will not
be presented in this chapter, but will be used to analyze data for the present study.
Covid-19 and Early Childhood Education
In December 2019, an unfamiliar strain of the airborne virus Severe Acute Respiratory
Syndrome (SARS) was reported about 7,000 miles away from the United States in Wuhan,
China. In a few months’ time, the virus, now known as Covid-19, has caused a global pandemic
and has shut the world down, significantly affecting all industries and sectors such as healthcare,
manufacturing and retail, travel and tourism, labor and employment, and education.
Covid-19 undeniably threatened the physical and mental well-being of teachers and
students, but research suggested that there were other pressing issues that the pandemic posed on
the education system. Although the effects of Covid-19 can be observed in all levels in
education, the extent to which the pandemic impacted each level of education varied. In ECE,
where developmentally appropriate high-quality learning experiences are linked to children’s
holistic development and future success (Britto, 2015; Bredekamp, 2020, UNICEF, 2020),
school closures in 2020 proved to be problematic.
12
Early Childhood Education, Remote Learning, and Educational Technology
Remote Learning
School closures disrupted social interactions, isolated young students, and limited their
learning experiences which presented threats to their overall welfare and academic achievement
(Golberstein & Miller, 2020). At the midst of the pandemic, the majority of the countries in the
world opted for remote learning to continue educating students safely from their home. However,
only 60% of these countries implemented support for ECE, despite recognizing the crucial
development of young children during these early stages (UNICEF, 2020). At that time, young
children were believed to be at a lower risk of developing and transmitting the virus, hence, they
were not at the center of public discussions and planning (Dias, et al., 2020).
For countries that implemented remote learning in ECE, the challenge was on adult
engagement because young students were not able to manage remote learning independently,
therefore, their online learning experiences mainly depended on the adults at home. During this
time, the inequities in ECE were also observed in the areas of teacher training, technology
availability, and quality of resources and materials for use while students were at home (Dias, et
al., 2020; Ford, et al., 2021). All of these affected the standard of teaching and learning during
the pandemic.
The absence of high-quality enriching activities and experiences in ECE during the peak
of the pandemic was beyond concerning. At that time, young children were not believed to be
super transmitters of Covid-19, which naturally led state leaders to deal with the more vulnerable
groups putting ECE on the backburner of Covid-19 discussions(Dias, et al., 2020). Interestingly,
a published study by the Journal of Pediatrics “refuted the idea that children were at minimal
13
risk, providing new evidence that children can have higher levels of virus and transmit COVID-
19 more than adults in intensive care units” (Dias, et.al., 2020, p. 39). This played a role in
decisions about continuing remote learning or reintroducing face-to-face instruction.
Learning Loss
It was presumed that when students are given extended breaks, whether it be weather-
related disruptions or summer vacation, they experience the usual learning loss where they are
predicted to regress and lose some knowledge and skills that were previously acquired during the
school year (Dorn, et al., 2021; Poletti, 2020). With about five months of school closures and
remote learning, education experts predicted that Covid-19 will cause learning loss (Sabates, et
al., 2021) and will widen already existing learning gaps especially for the most vulnerable groups
(Hoofman & Secord, 2021; Van Lancker & Parolin, 2020).
In an analysis of Covid-19’s effect on K-12 students, Dorn, et al.. (2021), found that the
impact on their learning was substantial with an average setback of five months in mathematics
and four months in reading by the end of the 2020-2021 school year. The authors also confirmed
that the pandemic exacerbated the already existing learning gaps due to factors such as economic
status, disabilities, students’ primary language, and other physical, mental, or emotional issues.
The widening of the existing academic gaps was also observed by Dias, et al. (2020). The
authors interviewed 26 early childhood teachers from different countries across Latin and North
America about their experiences in teaching during the pandemic. They emphasized that remote
learning, to provide young children with meaningful learning experiences, had to be “more
dynamic and less teacher-centered” (p. 41). In a study that Hoffman and Secord (2021)
conducted, they suggested that Covid-19 “will continue to affect the delivery of knowledge and
14
skills at all levels of education” (p. 1076) and although some students may adapt to new modes
of learning, they still need the guidance and support to continue with the new normal.
Educational Technology
Despite facing the adverse effects of the pandemic, early childhood educators and
advocates strived to find ways to provide the youngest students the developmentally appropriate,
high-quality learning experiences that they deserve while maintaining their safety from the virus
(Ford, et al., 2021). As a solution to this dilemma, world leaders opted for remote learning using
educational technology during the pandemic.
For the past few decades, there has been an ongoing dispute about the benefits and
consequences of integrating technology in the classroom, and whether or not to accept
technological transformation (Straker, et al., 2018). Although this topic remained open to
discussions (Jeong & Kim, 2017), current developments and circumstances have shifted the
focus of educational technology. As early as April 2020, UNESCO (2020) identified the use of
information and communications technology (ICT) to continue student learning while the
schools were closed and expressed the urgency to “ensure that the youngest learners are not
neglected and receive the stimulation they need to set the foundation for learning in their future”
(p. 1).
Although promising, not everyone believed that technology was the solution in closing
learning gaps as a result of the pandemic. In a literature review conducted by Jalongo (2021)
about Covid-19 research and resources in ECE, the author noted that while relying on technology
allowed for continuous learning during the pandemic, it may intensify the already existing gap
especially for disadvantaged students- those who belong to high-poverty, low income families,
15
who belong to subgroups such as ELs or those with disabilities (Bailey, et.al., 2021; Dorn, et.al.,
2021; UNICEF 2020). Therefore, it was critical to consider the educational technology that will
support the implementation of high-quality learning experiences for all students (UNICEF, 2020)
and will help them become productive citizens in the future (NAEYC, 2022; UNICEF, 2017).
After the pandemic, educational technology has been continuously used in many different
ways. In ECE, teachers and students utilized it as online platform tools for synchronous and
asynchronous learning, as learning management systems or hubs for resources, materials, and
apps, as tools to communicate with parents and families, as assessment portals, as and even as
personalized intervention programs. Due to the special developmental characteristics and needs
of young children, selecting the most appropriate educational technology in ECE can be tricky.
The developments in neuroscience have offered a better understanding of how the young brain
works and how the quality of early learning experiences can either leave a positive, lifelong
impact on children’s growth, development, and success (NAEYC, 2022; UNICEF, 2017) or
cause lifetime adverse effects (NAEYC, 2022; Knudsen et al., 2006). ECE is also linked to a
grander idea of global development. It is associated with sustainable development and future
economic growth (Britto, 2015; Samuelsson & Kaga, 2008) and is believed to contribute around
$163 billion to the US gross domestic product (GDP) (LeMoine, 2020). With investing in high-
quality ECE for children considered being high stakes, there is a need to ensure that the
technology that will be used by young children will support them in reaching their full potential.
A prospective educational technology that supported these objectives is artificial intelligence in
education (AIEd).
16
Artificial Intelligence in Early Childhood Education
Recently, the use of computers and digital devices in the early childhood classrooms has
been taken over by the use of AI. Artificial Intelligence (AI) has enabled machines to act and
think like humans (ISTE, 2020, 0:28; Lindner & Romeike, 2019; Unicef, 2020), something that
was thought to be impossible until the last few decades. AI is seen as a major contributor to
economic growth with the expected total contribution of $15.7 trillion by 2030 (PwC, 2017). In
2025, it is predicted that AI will displace 85 million jobs, but at the same time, add about 12
million AI-related jobs globally (Russo, 2020). Most businesses will have to reskill or upskill
their current employees (Russo, 2020) to ensure that they are equipped to do the work of the
future, today. With the shift in skills in workforce and training, the top universities from around
the world have increased their AI investments over the past few years, offering a 102.9%
undergraduate and 41.7% graduate increase in the number of courses that teach necessary skills
to build or use a practical AI model (Zhang, et al., 2021).
With the universities acting quickly to support this fourth industrial revolution, most k-12
schools were called to support this movement, hence, many k-12 schools have started using AI
enabled programs to facilitate teaching and learning. Among the top skills needed to thrive in
2025 included analytical thinking, creativity, and flexibility (Russo, 2020), all of which are skills
that AI enabled programs are capable of supporting and honing (ISTE, 2019).
AI has already made its way into every doorstep of homes, job sites, and schools around
the world. It can be visibly seen in gadgets such as smartphones, smart speakers, and
programmable robots or it could be working invisibly in the background in search engines and
social media search recommendations, in email filters for spam or inbox-worthy messages, or in
17
the amount of support tutoring software provides students based on their responses. Whether we
see it or not, or whether we like it or not, AI is here and will also be in the future.
AI Capabilities
Most k-12 students will enter the workforce when AI is deeply rooted in jobs and has
been well-established (Passow, 2019), this means that all students, including those who are in
ECE, have to explore and manipulate AI enabled programs to acquire soft skills associated with
working with them in the future. It is also necessary that in their young age they become fully
aware and begin to understand how these programs work and how they can be used in their
chosen careers in the future. AIEd brings a lot of possibilities and offers capabilities that
promise to provide differentiation and support for all students. In early childhood settings, the AI
capabilities below are the most commonly used.
Data Collection and Management using Educational Data Mining (EDM)
Part of the most challenging tasks of a teacher, aside from the actual act of teaching,
includes collecting, recording, analyzing, and reporting data. EDMs are programs and
applications that are used to collect, analyze, and track students’ achievement and behavior and
convert them to useful and meaningful information (Luckin, et.al., 2016; Zorić, 2020). This
information is readily available for teachers, administrators, and parents to use. EDM is useful
for early childhood teachers because of its ability to collect uniform data for all students and to
precisely and efficiently analyze them. Some EDM programs like ESGI and DIBELS Next have
the ability to produce detailed reports based on student data that can be used for intervention and
early identification. Reports can also be used to communicate student progress with parents and
18
other professionals. EDM can also collect and analyze students’ attendance and behavior and
predict possible issues that may occur based on the trends and patterns.
Production of Smart Content
Many early childhood students and teachers have yet to discover individual students’
learning preferences and needs. In order to provide differentiation and promote a love for
learning, teachers need to provide students with choices on how to access and learn new
information. AI has enabled teachers to customize lessons by producing smart contents through
digital platforms, information visualization, and learning content update (Plitnichenko,
2020). Several digital platforms offer lessons in various formats, complexities, and languages,
customizable visuals to represent the lessons, and automatically updates the information to keep
abreast with the latest, most relevant information. Learning A-Z, NewsELA, and Zearn Math are
some examples of digital platforms that offer smart contents.
Accessibility and Support
Perhaps one of the best capabilities of AIEd is its accessibility. The Covid-19 pandemic
caused several sectors, including education, to lean on technology tools, specifically the use of
human-like helpers like AI (Zorić, 2020). The need to remotely work and study forced many
school districts to offer synchronous and asynchronous virtual lessons for all students in k-12.
The Louisiana Department of Education (LDOE) opted not to require synchronous lessons for
pre-k and younger students for many reasons, including the lack of parental availability to
support students in online learning communities during school hours (J. Board, personal
communication, February 2020). AI has made it possible for parents to conveniently and easily
access lessons, materials, and resources in their children’s grade level that students can do at
19
their own time and pace. Chatbots and online assistants are also able to support parents and
students in answering questions they may have at any time.
Automated Assessments
Teachers and students can take advantage of AI’s ability to provide automated
assessments that will immediately grade and provide feedback to both teachers and students.
Google Forms, Mentimeter, and Kahoot are some programs that can be used in the early
childhood classroom that will do this task while adding additional tools for students such as links
to videos or other resources. Some of these programs can even be fun and celebrate student
success by enabling teachers to add special animations or sounds when students get correct
answers.
Personalization and Feedback through Intelligent Tutor System (ITS)
In ECE, it is important that one-on-one attention is given to students because of their
unique developmental characteristics and learning differences. Unfortunately, in some states like
Louisiana, kindergarten to 2nd grade classrooms in public schools have an average teacher-to-
student ratio of 1:25. This makes it almost impossible for the teacher to provide this one-on-one
attention consistently and on a daily basis. Luckily, AIEd has made it possible for ITS to provide
all students the attention and support they need, whenever they need it. ITS uses “AI techniques
to simulate one-to-one human tutoring, delivering learning activities best matched to a learner’s
cognitive needs and providing targeted and timely feedback, all without an individual teacher
having to be present” (Luckin, et.al., 2016, p. 25).
20
Amira Learning Software
Benefits and Challenges in Tutoring
The present study focused on an ITS program, Amira Learning software, and its impact
on young children’s literacy development. Before exploring the impact of intelligent tutors, it is
imperative to understand the significant role of a human tutor to the academic growth of
students. Merriam- Webster defines a tutor as a person in charge with the instruction and
guidance of another. In K-12, learning from a knowledgeable and skilled tutor is an invaluable
tool. It can particularly be beneficial for children who are experiencing academic challenges so
they can build resilience and perseverance. In some cases, tutoring can be what separates
academic achievement from academic failure. Perhaps the most astounding characteristic of
tutoring is the ability of tutors to provide immediate feedback and adjust instruction accordingly
to meet the needs of the individual student. Just like in any capacity, expertise plays a big role in
the success of tutoring. A highly-skilled tutor can influence learners in many positive ways, but
on the other hand, “human tutors who are not experts and yet tutor in academic areas can do
more harm than good” (Graesser, et al., 2012). It is expected that when tutors have the expertise
in the subject they teach, students will yield positive results.
In a meta-analysis with 96 past sample studies, Nickow, et al.. (2020) confirmed the
outstanding claims that tutoring produced significant positive outcomes for students. In addition
to this, they emphasized that tutoring was more effective when facilitated by teachers and
paraprofessionals as compared to paraprofessionals and parents, which again highlights the point
of a subject-area expert tutor. The authors also expressed that tutoring has a greater effect in
early grades and that customization, the ability to gear instruction based on the students’ level,
21
may have been the main reasons why tutoring is effective. They also added that the one-on-one
nature of tutoring allows students to be more engaged in the lesson while at the same time
enabling the teacher to provide more feedback. This leads to more opportunities for students to
practice the skills they are learning. Nickow, et al. (2020) also believe that the interactions
between the tutor and the students allow them to develop a healthy, positive relationship which
results in academic success.
Despite the positive effects of tutoring in K-12, there are a few challenges in this form of
instruction. The first issue can be linked to the cost of the program. While traditional tutoring has
been proven to work, the one-on-one sessions can be very costly compared to small group
instructions. Another hurdle is time management. In a classroom setting with roughly between
20-30 students with varying levels and needs, it can be difficult to find the time to meet with
individual students for tutoring with only one teacher present without sacrificing regular
instruction time. The last but not the least of the concerns will be customization (Bloom,
1984). Planning for actual tutoring instruction for each individual student can also be taxing to
teachers, especially for lower grade teachers who teach all contents.
Intelligent Machine
The conditions explained above make AI a perfect candidate for a human-like tutor that
will be more cost effective and manageable than employing real teachers. As a result of Covid-
19 pandemic, educators and stakeholders have been exploring ways to help close learning gaps
without over exhausting the human and financial resources the school systems have. AI was
initially explored to serve the role of a tutor in the 1970’s when J. Carbonell introduced
SCHOLAR, an intelligent tutor (Guo, et al.., 2021; Mousavinasab, et al.., 2021). With the known
22
success rate of one-on-one tutoring, many academic institutions have been working hard to
develop new and polish existing ITS.
ITS are advanced AIEd tools that resulted from the collaborative efforts of researchers in
the fields of Education, Psychology, and AI (Atun, 2020). ITS are human-like computer
programs or tutors that have the ability and knowledge of the subject they teach (Alkhatlan &
Kalita, 2018). They are designed to facilitate personalized adaptive learning experiences and
unlike conventional educational technology tools, ITS possess a distinct characteristic of being
able to modify their own behavior based on the information they receive from the student and
teacher users, programmed to learn the cognitive patterns of students, and provide individualized
instruction to help them learn specific skills in different subject areas (Atun, 2020; Ma, et al.,
2014). ITS “know what they teach, whom they teach, and how to teach” (Alkhatlan & Kalita,
2018, p.2). What sets ITS apart from other AIEd programs is that these tutors have the ability to
process the students’ learning and provide instant, automated feedback to support learning
(Alkhatlan & Kalita, 2018; Roscoe, et al., 2014) similar to human tutors.
Intelligent Reading Tutor
As a reading tutor, ITS has proven its positive impact on elementary-aged students.
Wijekumar (2012, 2013, 2018) has been a part of a number of studies on ITS and its impact on
reading comprehension. In 2012, Wijekumar, et al. conducted a randomized controlled trial to
investigate if 4th grade ITS classrooms outperform control classrooms in expository reading
comprehension as measured by standardized and researcher-designed tests. There were a total of
60 rural and 71 suburban classrooms that participated in the study. The experimental conditions
were randomly assigned to the schools. After the teachers have been notified of the assignments,
23
the ITS classroom teachers were asked to use the program for 30-45 minutes each week as a
substitute for a portion of the regular language arts instruction. Total instruction time for both
ITS and control classrooms were planned to be the same. The teachers who were assigned to ITS
classrooms were given professional development training prior to the beginning of the next
school year. A pretest was administered to all participants and a posttest at the end of the school
year after the experiment.
Pretests revealed no significant differences in students’ comprehension level, which make
them comparable groups before the experiment. Results of the study suggested that students who
were in the ITS group who used the program for about 30-45 minutes per week over a span of 6
months displayed significantly improved performance, as evidenced by the research-designed
posttest, compared to their counterparts in the control group. On the other hand, a small effect
size with the ITS group was observed on the standardized multiple choice comprehension test
given during the posttest.
In 2013, Wijekumar, et al.. conducted a multi-site cluster randomized trial to investigate
whether there was a connection between fidelity in using ITS and improvement in reading
comprehension of 4th and 5th grade students. The researchers used both standardized and
researcher designed measures in their study. The experimental and control groups used the same
curriculum, with the exception of one class period each week because of the random ITS and
control classroom assignments. For the ITS classroom, one language arts class was substituted
with the web-based program for 30-45 minutes each week.
A total of 131 4th grade teachers and 128 5th grade teachers were randomly assigned to
ITS and control groups. The ITS classrooms that were selected in the study after random
24
assignment were the ones who implemented the ITS software with fidelity. Students were
administered a pretest at the beginning of the academic year and a posttest at the end of the
academic year. Researchers noted that there were no significant differences between the ITS and
control groups on the pretest, which indicated that both classrooms were comparable before the
implementation of the experiment.
The analysis revealed that the use of ITS with a high level of fidelity, including sufficient
time used, teacher monitoring, and teacher engagement in 4th and 5th grade classrooms resulted
in moderate to large effects on the researcher-developed measures. On the standardized test, the
effect sizes were small, but still notable considering the nature of the multiple choice format of
the reading comprehension test. The effect sizes were particularly impressive for 4th grade
students.
In a similar study from 2012, Wijekumar, et al. (2018) conducted a large scale
randomized controlled trial to examine the impacts of a web-based instruction structure on the
reading comprehension of struggling fourth and fifth grade readers. The researchers used
standardized and researcher designed tests. A total of 45 schools, consisting of 22 rural and 23
urban schools, volunteered to participate in the study. There were a total of 725 fourth grade
students and 717 fifth grade students from both school settings. The students were administered a
pretest at the beginning of the school year and a posttest at the end of the school year. Students
who scored at the lowest 25% percentile on reading comprehension pretest were chosen to
participate in the study.
The regular instruction was based on the structure strategy, an instructional approach that
was text structure-based for improving reading comprehension in the content area. The ITS
25
modeled how to identify signaling words, classify text structure, write a main idea, construct a
recall of text, generate inferences, and monitor comprehension using an AI agent, I.T. Both the
ITS and the control classrooms had an equivalent amount of language arts instructional time,
averaging 450 minutes per week or around 90 minutes per day. In the ITS classrooms, the
students received 30-45 minutes of language arts instruction each week using the software,
replacing traditional teacher instruction.
The findings in the study of Wijekumar, et al. (2018) indicated that the implementation of
the web-based ITS for teaching structure strategy has demonstrated a positive impact and holds
potential for further improving instruction for students who struggle with reading
comprehension.
The studies presented above investigated the effectiveness of ITS in reading
comprehension of upper elementary grades. Although the results were significant and
instrumental in the present study, the positive outcomes of ITS in these levels cannot be used to
generalize and suggest positive impacts on foundational literacy skills, which the present study
investigated. Searches on academic databases for studies on the impact of ITS on foundational
literacy skills did not yield a wide range of results, but Amira Learning software’s webpage
(https://www.amiralearning.com/research.html) listed titles of articles and field trials published
and funded by the company, in conjunction with Amira Learning software, or were about Project
LISTEN, which where Amira Learning software reportedly originated from. Some of these
studies were used as references in the present study but were purposely not presented in the
literature review for fairness.
26
Amira the Intelligent Tutor
Amira Learning software claims to be the world's first intelligent reading assistant that
offers personalized tutoring to help students improve their reading fluency. Amira, an animated
young lady, interacts with the students by initiating conversations, asking questions, and
providing feedback while students are reading. Amira has the ability to filter discussions and is
programmed to respond and talk only about the story students are reading
(www.amiralearning.com, n.d.).
Amira Learning software's technology has its roots in research spanning two decades
under the sponsorships of Carnegie Mellon University's (CMU) Project LISTEN (Literacy
Innovation that Speech Technology ENables). Originally known as "RoboTutor '' during its
tenure within Project LISTEN, this software underwent rigorous development and refinement
after being licensed from CMU. The objective was to make sure that students, educators, and
families could access and utilize this extensively researched and evidence-based platform on a
large scale (Meta-Analysis of Research on Amira Intelligent Tutoring’s Impact, 2022).
Project LISTEN is a research initiative developed by CMU. The primary focus of Project
LISTEN is to improve literacy skills, especially reading, using speech technology. An overview
of the technology typically used in Project LISTEN included (https://www.cs.cmu.edu/~listen/):
Automatic Speech Recognition (ASR). ASR technology is used to convert spoken
language into written text. In the context of Project LISTEN, ASR is employed to transcribe and
analyze students' spoken responses to reading exercises.
27
Intelligent Tutoring Systems (ITS). Project LISTEN often utilizes intelligent tutoring
systems to provide personalized instruction to students. These systems adapt to individual
learners and provide feedback and guidance based on their performance.
Speech Processing Algorithms. Various speech processing algorithms are used to
analyze students' pronunciation, fluency, and comprehension while reading. These algorithms
can help identify areas where students may need improvement. Project LISTEN converts speech
input (Sphinx) to usable data.
Machine Learning and Data Analytics. Project LISTEN employs machine learning and
data analytics to assess and track student progress, identify learning patterns, and adapt
instruction accordingly.
Educational Data Mining. The technology in Project LISTEN includes interactive
educational software and applications designed to engage students in reading activities and
provide a dynamic learning experience. It uses data to learn about students and instruction to
predict students’ behavior, assess their needs, and evaluate the intelligent tutor’s teaching.
Amira Learning Platform
Amira claims to offer support to both students and teachers. When students log in to the
portal, Amira, the animated tutor, provides them with a variety of books to choose from. She
then activates her AI technology to listen to students read. Amira begins conversations with the
students and takes notes of reading patterns and their responses to help identify reading skills
gap. Amira uses AI-powered, perfectly-timed micro-interventions aligned with SOR to support
foundational reading skills (phonics, phonemic awareness, decoding, and vocabulary) and later
28
on, comprehension, to provide tutoring to the students. As students progress with their tutoring
sessions, Amira celebrates the completion and presents their reading data (average speed and
average correct) to show their reading progress.
To support teachers, after students’ reading sessions, updated, detailed reports which
track student progress and identify struggles while students are reading are readily available for
analysis. Aside from the reports, Amira Learning software also offers the teachers access to
instructional links, resources, and activities that they can use when students are not practicing
with the software.
The Science of Reading
Louisiana Literacy Initiatives
During its 2021 Regular Session, the Louisiana Senate approved Act no. 108. This new
law mandated early literacy training for all subject area K-3 teachers, as well as principals and
assistant principals who serve students in these grade levels. The training focused on SOR, an
extensive and interdisciplinary collection of scientifically-grounded research on literacy
instruction (Moats, 1999; Moats 2020; Science of Reading: Defining Guide, 2022). It has been
an area of study for several decades now, but its implications on education have been questioned
and a subject of debate (Petscher, et al.., 2021; Seidenberg, 2013; Shanahan, 2020). It was not
until the late 1990’s when SOR began to gain popularity and recognition in the field of education
and literacy instruction. During this time, there was a growing emphasis on evidence-based
practices and scientific research in understanding the processes involved in reading development.
Additionally, there was an emphasis on the importance of effective reading instruction to match
the knowledge of reading (Shanahan, 2020; Moats 2020).
29
Back in 1999, Moats stressed that even with the scientific knowledge of how the brain
works when one reads and despite Scientists’ claim that 95% of children can learn how to read,
“statistics reveal an alarming prevalence of struggling and poor readers that is not limited to
anyone segment of society” (p. 7).
With the promising research but insignificant outcomes, SOR must not only provide
information to educators and stakeholders, instead, it should support in identifying the most
effective methods to expand literacy opportunities, enhance literacy achievement levels, promote
equitable access, and improve efficiency and effectiveness of reading instruction (Shanahan,
2020). This should include focusing on teacher preparation programs and professional
development for teachers to provide them with a more rigorous and closely aligned guidance to
instructing through SOR (Moats, 2020).
It was not until after more than a decade when the first state in the US embarked on SOR.
Mississippi was the first state to support literacy initiatives aligned with SOR in 2013. The state
mandated that school systems provide teachers with training in evidence-based reading
instruction and intervention while obtaining support from school-based reading interventions.
Students with reading difficulties received individualized reading plans and retention policy and
support for third grade students who were retained were enforced. The US Department of
Education, National Assessment of Education (NAEP) report showed that Mississippi went from
the second lowest ranked state in reading scores for low-income 4th grade students in 2013 to the
21st state in the rankings (Lurye, 2023). This significant improvement has gradually influenced
other states to adopt SOR, but in 2021, year after the Covid-19 school year, there was a
significant increase in the number of states in a year (11 states) that got on board with SOR. This
included the state of Louisiana. At present, there are a total of 31 states together with the District
30
of Columbia, that have mandated laws or introduced new policies regarding evidence-based
reading instruction (Schwartz, 2022).
As established earlier in the present study, the impact of Covid-19 on education extends
beyond the physical and mental health risks faced by teachers and students. It has also resulted in
learning loss and gaps that have adversely affected students’ academic progress, specifically
growth in math and literacy. Mastery in both math and literacy can be considered as essential
components for students to succeed in various aspects of life. However, literacy skills, including
reading, writing, and comprehension, are crucial for effective communication, critical thinking,
and accessing information across various subjects. Literacy is fundamental for academic success,
as it forms the basis for learning and understanding other subjects beyond just language arts
(Moats, 1999; Moats 2020). Therefore, anything that negatively affects literacy development is
never good news for any state in the nation, but its effects are most daunting on the states like
Louisiana which have been at the bottom of the national literacy ranking list for decades
(Chavez, 2021). Thus, it was a turning point for many educators when Louisiana decided to be
on board with training K-3rd grade teachers and administrators with SOR.
Theoretical Framework: Amira Learning
The present study aimed to investigate the effects of an AI program, Amira Learning, on
literacy development in early childhood education. To fully explore this topic, it was essential to
consider two overarching concepts: mastery learning (Bloom, 1975) and skilled reader (Gough &
Tunmer, 1986; Scarborough, 2001; National Reading Panel, 2000).
31
Theory of Mastery of Learning
The theory of mastery learning in ITS stemmed from the work of educational
psychologist Bloom, who developed the concept of mastery learning in the 1960s. Mastery
learning emphasized that all learners can achieve mastery of a subject if provided with
appropriate instructional strategies, feedback, and time for learning (Guskey, 2005). Bloom's
(1968) theory of mastery learning presumed that students should attain a specified level of
mastery on prerequisite knowledge or skills before moving on to more advanced material. This
approach contrasted with traditional instructional models that relied on fixed time periods for
learning. Mastery learning promoted individualized instruction and allowed learners to progress
at their own pace, ensuring a solid foundation before advancing to higher-level concepts.
In the context of ITS, mastery learning theory influenced the design and implementation
of instructional sequences, assessment strategies, and feedback mechanisms. ITS based on
mastery learning provided learners with tailored instruction and practice opportunities,
continually assessing their understanding and providing personalized feedback to address
knowledge gaps or misconceptions. These systems monitor learner progress, identify areas of
weakness, and offer remediation or additional learning resources to support mastery.
Mastery learning was built upon several fundamental principles that align with the
pedagogical principles of ITS. These principles included the following (Abdelsalam, 2014):
● Time is a crucial factor in the learning process. Teachers should allocate sufficient
time for each learner, recognizing their varying levels of achievement.
32
● The quality of teaching significantly influences learning outcomes. If teaching
quality is low, learners may require more time to grasp the material effectively.
● Motivation and the learners' ability to comprehend are essential foundations for
successful learning. Tasks are accomplished by considering these basics.
● Learners progress at different rates based on their individual capabilities and
characteristics.
● Immediate feedback assists learners in identifying and rectifying their mistakes
promptly.
● Alignment of the learning objectives with the breakdown of topics and tasks to
help in learning the materials.
● Continuous assessment is crucial for attaining the objectives of mastery learning.
Diagnostic, formative, and summative assessments are employed to ensure
ongoing evaluation.
The present study delved into the alignment between Amira Learning software’s
instructional design and three of the principles of the mastery of learning. These principles
served as a guide when assessing the effectiveness of Amira Learning software in facilitating
mastery-oriented learning experiences and improving students’ literacy outcomes. The three
principles that were used were 1) Learning time and achievement: Bloom's theory recognized the
variability in learners' pace of learning and the importance of allowing sufficient time for
mastery. The present study investigated the relationship between the time spent using Amira
Learning software and learners' achievement levels. 2) Immediate feedback: Bloom's theory
highlighted the importance of continuously assessing students’ learning to monitor their progress
and provide timely feedback. The present study explored the effectiveness of the feedback
33
provided by Amira Learning software in guiding learners towards mastery and facilitating their
learning progress. 3) Alignment of learning objectives with the topics and tasks: Bloom’s theory
emphasized the importance of clearly defined learning objectives. The present study explored
how teachers rate Amira Learning software’s feedback and lessons as they align with SOR and
its objectives.
The Skilled Reader
The Simple View of Reading (SVR), Scarborough's Reading Rope, and the Five Pillars
of Reading are three frameworks that provide a comprehensive understanding of the various
components involved in the process of reading. While they have distinct characteristics, they all
recognize that reading is a complex process that involves two fundamental practices, decoding
and comprehension. They also recognize that reading skills develop hierarchically, with
foundational skills serving as building blocks for more advanced ones.
Simple View of Reading
SOR instruction supports the delivery of explicit, systematic literacy lessons to students
followed by multiple opportunities to practice reading and obtain feedback from the teacher
(Shanahan, 2020; Moats, 2020, Petscher, et al.., 2021; Science of Reading: Defining Guide,
2022). While SOR is not attached to a single ideology or philosophy or program of instruction
(Science of Reading Defining Guide, 2022), there are two theoretical models that have been used
in professional developments about SOR to explain the complexity of reading, the Simple View
of Reading (SVR) and Scarborough’s Reading Rope.
34
Gough & Tunmer (1986) developed and proposed SVR. SVR is a theoretical model that
explains reading comprehension by breaking it down into two components: decoding and
language comprehension. SVR suggested that reading comprehension is a product of the
interaction between these two essential factors (Decoding x Language Comprehension= Reading
Comprehension). In this model, both decoding and language comprehension are considered
critical for reading success. A reader must possess adequate decoding skills to accurately read
words, and at the same time, they need strong language comprehension abilities to understand
and make meaning from the text.
Decoding refers to the ability to convert written or printed words into spoken words. It
involves recognizing and understanding the individual sounds (phonemes) that make up words
and blending them to form meaningful units, such as syllables or words. Decoding skills are
primarily associated with phonics instruction, which teaches the relationship between letters and
their sound correspondence. Decoding involves phonological awareness, decoding, sight word
reading, and fluency.
Language comprehension encompasses the broader understanding of spoken or written
language, including vocabulary knowledge, syntax (grammar), semantics (word meanings), and
background knowledge. It involves making sense of the words, sentences, and ideas encountered
while reading. Language comprehension incorporates background knowledge, syntax,
vocabulary, and text structure
Scarborough’s Reading Rope
Scarborough (2001) offered a more thorough representation of the skilled reader through
the Reading Rope Model or Scarborough's Reading Rope. The present study referred to it using
35
the latter. In Scarborough's Reading Rope, the two essential components of reading
comprehension have been broken down into individual strands that are interwoven to represent
the strong bond that needs to be present in order for students to become skilled readers. This
model presented a view of reading comprehension which includes various sub-skills within word
recognition (phonological awareness, phonics, fluency) and language comprehension
(vocabulary, background knowledge, verbal reasoning, literacy knowledge).
Scarborough’s Reading Rope emphasized the reciprocal relationship between word
recognition and language comprehension, with both strands intertwined and mutually supportive.
It has helped inform instructional practices for phonological awareness, phonics, fluency,
vocabulary development, background knowledge, and comprehension strategies.
Figure 1. The Scarborough's Reading Rope
The Five Pillars of Reading
Another theoretical framework that has been linked to SOR is the concept of the Five
Pillars of Reading. This was originally based on research by the National Reading Panel ([NRP]
National Institute of Child Health and Human Development [NICHD], 2000), a group of highly-
qualified scholars bringing together their expertise in the field of science-based research reading
and instruction which started in 1997. In NRP NICHD’s final report published in 2000, they
36
identified the five essential components of reading instruction, phonemic awareness, phonics,
fluency (also oral reading fluency), vocabulary, and comprehension (Shanahan, 2005).
Shanahan (2005) summarized and explained each component of effective reading
instruction below.
Phonemic awareness. Phonemic awareness is the ability to identify, manipulate, and
work with individual phonemes, which are the smallest units of sound in spoken language. It is a
critical precursor to reading and is considered one of the foundational skills for literacy
development. Phonemic awareness goes beyond simply recognizing individual letters or sounds.
It involves understanding that words are made up of individual phonemes and being able to
manipulate and blend them to form words. Phonemic awareness tasks do not involve printed
text; they are focused solely on oral language.
Phonics. Phonics is a method of teaching and learning reading that focuses on the
relationship between letters (graphemes) and sounds (phonemes). It involves teaching children to
understand and use the systematic connections between written letters and the sounds they
represent. In phonics instruction, children learn to associate individual or groups of letters with
specific sounds. They learn the phonetic rules and patterns that govern the English language,
such as letter-sound correspondences, syllable division, and spelling patterns. Phonics instruction
typically starts with simple letter-sound relationships and progresses to more complex ones.
Oral Reading Fluency. Oral reading fluency refers to the ability to read aloud with
accuracy, speed, and expression. It involves the smooth and fluent delivery of text, where the
reader demonstrates proficiency in decoding words, maintaining an appropriate pace, and using
expressive intonation and phrasing. Oral reading fluency is typically assessed by having students
37
read a passage aloud while being timed. The assessment measures factors such as reading words
correct per minute (WCPM), accuracy, and prosody (expression, phrasing, and intonation). The
goal is to determine how well a reader can read aloud with both accuracy and fluency.
Vocabulary. In this context, vocabulary pertains to the meanings of words, and
vocabulary instruction involves teaching those meanings. However, it is worth noting that
reading instruction often concentrates heavily on various aspects of words, such as recognizing
them, sight words, strategies for approaching unfamiliar words, word structure, and organizing
words based on shared characteristics. As a result, the term "vocabulary" is sometimes used
interchangeably to refer to both word recognition and word meaning.
Reading Comprehension. Reading comprehension involves the process of
comprehending and interpreting the information presented in a text. Rather than passively
remembering information, comprehension focuses on actively constructing meaning. It requires
active and dynamic thinking, including interpreting information based on one's own knowledge
and beliefs, using the text's organizational structure to analyze information (or applying one's
own structure to the ideas), making inferences beyond explicit statements by the author, and
engaging in various cognitive processes. Effective comprehension depends on the engaged
interaction between the reader and the text, where the reader actively processes and interacts
with the information presented.
Using these three literacy concepts as a theoretical framework, the present study analyzed
the impact of Amira Learning software on literacy development of first grade students. Oral
Reading Fluency (ORF) is an important component of reading as it allows students to practice
fluency and develop vocabulary and comprehension.
38
Summary
The impact of Covid-19 affected everyone around the world, disrupting all aspects of life.
In the field of education, it has widened pre-existing learning gaps and has created new
challenges (Sabates, et al.., 2021; Hoofman & Secord, 2021; Van Lancker & Parolin, 2020),
including providing individualized instruction to meet students’ new needs. Fortunately, AIEd
has the ability to think and function like a human (McCarthy, 2007) as an intelligent tutor and
provide personalized instruction geared to the individual needs of students (Atun, 2020; Ma, et
al.., 2014). In states like Louisiana where reading proficiency has traditionally been low (Sentell,
2022), reliance on SOR, a research-based scientific knowledge in teaching students how to read,
has been necessary. This present study investigated the impact of Amira Learning software, one
of the pioneering intelligent tutors aligned with SOR, to young students’ literacy development.
The second chapter of this research presented an extensive review of relevant literature
about intelligent tutors and their influence on literacy development. In Chapter Three, various
components of the study are introduced such as the research paradigm, setting and context,
ethical considerations, data sources, data analysis, and research design.
39
CHAPTER THREE. METHODOLOGY
The present study aimed to investigate the effects of an AI program, Amira Learning, on
literacy development in early childhood education. Amira Learning software is a web-supported
Artificial Intelligence in Education (AIEd) Intelligent Tutoring System (ITS) designed to
improve oral reading fluency of early learners. With the rapid development of AI technologies
and a continuously growing interest in their application in educational contexts, there has been
significant growth in the scientific literature in relation to the application of AIEd (Broda &
Frank, 2015; Ma, et al. 2014; Mousavinasab, et al., 2021; Nickow, et al., 2020,; Su, et. al 2022;
Touretzky, et al., 2019; Yang, 2022). The threats on human capital caused by Covid-19 has also
increased research interest in AI as intelligent tutors. To make informed decisions about the
research methods to be used in the present study, previous studies were examined. Quantitative
research involves the collection and analysis of numerical data to answer research questions.
This method provided a systematic and objective way to gather and analyze data, as it allowed
for the use of statistical analysis to draw conclusions (Eyisi, 2016; Queiros, et al.., 2017).
A quantitative methodology was used to answer the research questions, with the focus on
whether an AI application was effective in improving oral reading fluency in early learners. As
planned, the present study focused on following a statistical approach to ensure fairness, validity,
and reliability of the data analysis. The research paradigm outlined below provided the
theoretical foundation for the study and shaped its direction and goals. The research design
constituted a comprehensive plan that systematically addressed the research question. This
design served as an overarching framework that guided the selection of appropriate data
collection and analysis methods (Spatz, 2019). The research design also outlined the general
40
approach that was employed in answering the research questions which provided a roadmap for a
quantitative investigation. By adhering to a well-articulated research design, it was hoped that
the present study ensured the rigor and coherence of the data collection and analysis processes,
which led to meaningful conclusions. The data collection section included methods that will be
used to retrieve the necessary information to conduct the research (Spatz, 2019). The statistical
section explicitly explained the tests that were used to determine the statistical approach that
most effectively answered the research questions. The statistical tests that were chosen for the
present study were based on the type of data being analyzed and the research questions being
asked. The statistical tests helped the study determine the statistical methods that were deemed to
provide the most accurate and meaningful results, allowing it to draw valid conclusions about the
differences in the data. The careful selection of the appropriate statistical tests helped to produce
accurate and reliable results that were useful in answering the research questions and ensured
that the findings were as accurate and reliable as possible (Spatz, 2019).
Research Paradigm
The present study used quantitative research methods to examine the effectiveness of the
Amira Learning software among first grade students within a suburban school district in
Louisiana. As an intelligent tutor, Amira has been designed to provide students with feedback to
improve their oral reading fluency (ORF). ORF has been identified as one of the literacy
measures that predict literacy success (Shanahan, 2020; Moats, 2020, Petscher, et al.., 2021;
Science of Reading: Defining Guide, 2022). The Dynamic Indicators of Basic Early Literacy
Skills Next edition (DIBELS Next) has measured ORF beginning in first grade. In the present
study, DIBELS Next was used to assess and compare the students’ progress before and after the
use of the Amira Learning software tutoring software for six weeks.
41
Because the present study aimed to focus on establishing a specific relationship between
two measurable variables (time spent using Amira Learning software and ORF achievement
levels), a positivist approach was valuable. Positivism emphasized the use of quantitative
methods to gather data and establish causal relationships. The hope was that the findings of the
study will be useful and applied in other similar situations to impact research in literacy
education. The positivist paradigm enabled researchers to observe occurrences within the
particular phenomenon they have studied and draw conclusions about what can be expected in
similar cases (Kivunja & Kuyini, 2017).
While there was available research on the effectiveness of AIEd learning applications
(Chen et al.,2020; Dong et al., 2022), there was limited information about the Amira Learning
software (Loble & Hawcroft, 2022), with the exception of a few research conducted by Amira
Learning software in conjunction with other educational organizations. The present study wished
to contribute to the body of research that explored the effects of the ITS in improving literacy
development in early childhood education (ECE).
Research Questions
The research questions that were investigated in the present study were:
RQ1: What is the relationship between the time spent using Amira Learning software and
learners' oral reading fluency achievement levels?
RQ2: How effective is the feedback provided by Amira Learning software in improving
students’ oral reading fluency?
42
RQ3: What is the level of agreement among teachers regarding Amira Learning
software’s alignment to the Science of Reading and its effectiveness in fulfilling the Science of
Reading objectives?
Setting and Context
Setting
The present study investigated the literacy development of first grade students who used
Amira Learning software for six weeks. The data that was used for the present study was
obtained from one of the suburban elementary schools in Louisiana serving students from Pre-K
3 to 5th grade. Students at the school had access to high-speed internet as well as 1:1 technology
device (Chromebook) from kindergarten to 5th grade. The school was the first in the district to
use Amira Learning software and pilot-tested the program for all students in K-5th grade during
the 2022-2023 school year. During the same school year, the state of Louisiana piloted the
program to support English Learners (ELs) and currently had sixteen districts using Amira
Learning software for their EL population (www.amiralearning.com).
Population
During the 2022-2023 school year, the school site had a total population of 598 students,
with 52% considered economically disadvantaged who received free or reduced lunch. The
school was comprised of 54.8% white students, 32.6% black, 11% Hispanic, and 1.6% Asian/
Pacific Islander. For two consecutive years, the school has had the highest EL population in the
district. The teacher-to-student ratio in K-5th grade classes ranged from 1:18 to 1:25. The ratios
of the five first grade classes that were used in the present study were 1:17, 1:19, 1:21, 1:21, and
1:22.
43
Participants
The data that was used for the present study was obtained from five classes in the first-
grade level. Although data for first grade students was used, the students were not directly
involved in the study. The present study used data collected on the DIBELS Next and Amira
Learning software platforms. The researcher of the present study adhered to ethical guidelines
for using secondary data and identified the original data sources (see Ethical Considerations,
below).
There were a total of 101 students in first grade at the school site, but data was analyzed
only from 79 students. Data from 21 students were excluded if they fell under one or more of the
following categories: did not take the DIBELS Next ORF pretest or posttest, was not present, at
school, or in class for a total of 5 or more instructional days (equal to 1 week of Amira Learning
software usage) during the six-week pilot testing period for Amira Learning software, or had an
average Amira Learning software usage of 0 minutes in 6 weeks.
To answer research question 3, purposive and convenience sampling were used to
identify the teacher participants. The researcher of the present study chose participants who,
because of their background and expertise, were believed to have valuable input on the topic.
Teachers at the school site were included if they were certified in early childhood or elementary
education (grades pre-kindergarten to 3rd grade), had experience in the early childhood classroom
(grades pre-kindergarten to 3rd grade), had experience using Amira Learning software, and had
the training in the Science of Reading (SOR) and Amira Learning software. Under these
conditions, 17 teachers at the school site were included to answer the two-question online survey.
44
Ethical Considerations
The researcher of the present study applied for an approval to conduct the study from the
Institutional Review Board (IRB) at Louisiana State University (LSU) before conducting the
research. An IRB exemption application was granted for the present study after IRB determined
that the study posed little to no risk to human subjects.
Even with using secondary data from two online platforms, DIBELS Next and Amira
Learning software, the researcher of the present study ensured that safety and security measures
were observed throughout the process. A signed consent form was sent to the district
superintendent of schools, data administrator, and school administrator. The consent forms
described the process of the research, including its purpose, risks, benefits, and their right to
withdraw at any time without consequences. The same was done for the teacher participants. The
anonymity and confidentiality of data and participants were also discussed. After the consent was
signed by the district and school administrators, the data was provided to the researcher of the
present study.
No identifying information was used in the study and data security was ensured to protect
it from unauthorized access, loss, or disclosure. The data was saved in a password protected
electronic device and online information was obtained and stored in a secured online data
management system. It was known to the teacher participants that the researcher of the present
study was the person who reached out to Amira Learning to pilot-test the software. The
researcher of the present study completed the SOR and Amira Learning software training prior to
teachers’ completion. The participants were informed that the researcher of the present study did
not receive any benefits from the company at any time during the pilot testing and while
45
conducting the present study. In using the secondary data, appropriate credit to the original data
collectors or sources in any research outputs or publications were cited as needed.
Amira Learning Software
Amira Learning software claims to offer support to both students and teachers. When
students log in to the portal, Amira, the animated tutor, provides them with a variety of books to
choose from. She then activates her AI technology to listen to students read. Amira begins
conversations with the students and takes notes of reading patterns and their responses to help
identify reading skills gaps. Amira uses AI-powered, perfectly-timed micro-interventions aligned
with SOR to support foundational reading skills (phonics, phonemic awareness, decoding, and
vocabulary) and later on, comprehension, to provide tutoring to the students. As students
progress with their tutoring sessions, Amira celebrates the completion and presents their reading
data (average speed and average correct) to show their reading progress.
To support teachers, after students’ reading sessions, updated, detailed reports which
track student progress and identify struggles while students are reading were readily available for
analysis. Aside from the reports, Amira Learning software also offers the teachers access to
instructional links, resources, and activities that they can use when students are not practicing
with the intelligent tutor.
Tools
The present study investigated the use of Amira Learning software
(www.amiralearning.com) on student reading skills as measured by DIBELS Next ORF measure
(www. acadiencelearning.com).
46
Dynamic Indicator of Basic Early Literacy Skills (DIBELS) Next edition
DIBELS Next assessments are a set of research-based procedures and measures used to
evaluate students’ acquisition of early literacy skills. They are designed to be short fluency
measures used to regularly monitor the development of early literacy skills and early reading
skills (www. acadiencelearning.org, n.d.). In first grade, DIBELS Next assess the measures of
phoneme segmentation, nonsense word fluency, and oral reading fluency. It uses four levels of
risk to categorize students' performance on the assessment. These risk levels help educators
identify students who may need additional support or intervention in developing their early
literacy skills. The risk levels are typically classified as follows: low risk (green and blue;
students who are performing at or above grade level), some risk (yellow; students who are
performing slightly below grade level), and at risk (red; students who are performing
significantly below grade level).
Oral Reading Fluency (ORF). ORF as measured by DIBELS Next refers to a specific
assessment used to evaluate a student's reading proficiency. In the DIBELS ORF assessment, a
student is asked to read a grade-level passage aloud for one minute. The evaluator records the
number of words the student reads correctly within that time frame. It focuses on accuracy and
rate of reading.
Procedures
The present study used a quasi-experimental design and did not alter any grouping,
testing, data collection, and Amira Learning software usage protocols. The school site strictly
followed the DIBELS Next (www. acadiencelearning.org, n.d.) testing protocols and had
expectations in place for these processes. Student testing and collection of data for DIBELS Next
47
ORF pretest and posttest followed standardized procedures that included scripts for teachers to
use to give the timed test for all literacy measures. This ensured that each student was given the
same directions, equal amount of time to complete the tasks, exact materials, similar conducive
environment, and consistent score reporting. The test was administered by DIBELS Next trained,
certified teachers. Teachers were not allowed to test their own students.
At the school site, the researcher of the present study was responsible for disseminating
information to teachers and administrators about Amira Learning software. To ensure that
teachers and students get equal opportunities in using the software and that they receive a
uniform message about the usage, limitations, and benefits of the program, a 30-minute training
was conducted by the researcher of the present study to communicate the program expectations
set by the school administrator. The school expectations for usage included allowing students to
use Amira Learning software for at least 30 minutes per week for six weeks. The teachers were
expected to allot 15-20 minutes each day during their small group English Language Arts (ELA)
block for students to use Amira Learning software.
After discussing the usage expectations for the school site, the teachers were given an
overview of Amira Learning software using presentation slides shared by the representative from
the company. The slides covered topics such as how Amira Learning software works, what the
student experiences look like, what reports can be generated from the platform, and how to
troubleshoot common issues about the use of the program.
A follow-up online training sponsored by Amira Learning University, the company’s
professional development platform (www.amiralearning.com), was completed by every teacher
who was asked to use the program. The follow-up teacher training took about 30 minutes to
complete. Topics covered included usage logistics, sample videos of students interacting with
48
Amira, information on the program’s alignment with the Science of Reading (SOR), and
additional training courses available for teachers and leaders. Teachers started using Amira
Learning software the following week after the teacher training was completed.
Data Sources
Aligned with quasi-experimental uncontrolled before and after research methods, the
effects of Amira Learning software on the literacy development of first grade students were
explored without a control group for comparison (Grimshaw, et al., 2000). The following data
sources were used in the present study.
First data source. DIBELS Next Oral Reading Fluency (ORF) pretest and posttest were
administered to all first-grade students by certified teachers who were members of the school
level DIBELS Next team. The data was collected before and after the six-week Amira Learning
software usage. Student data were entered to the DIBELS Next portal where pre-generated score
reports were downloadable. The reports included all measures of DIBELS Next assessment that
were administered to students. The researcher of the present study was provided with the
DIBELS Next report by the school district data administrator.
Second data source. The second data source that was used was the total amount of time
students used Amira Learning software in minutes. The total amount of time spent in the
software was calculated by adding the amount of time students spent reading books and the
amount of time students interacted with Amira the avatar during tutoring and feedback sessions .
49
This data was collected by and stored on the Amira Learning software platform. The researcher
of the present study was provided a copy of the report by the school district data administrator.
Third data source. After the six-week period of using Amira Learning software, the
teacher participants were asked to answer these questions: “How would you rate the
effectiveness of Amira Learning software in accomplishing the goals of the Science of
Reading?” and “How aligned is Amira Learning software with the Science of Reading?”.
Teachers were provided with an online link to the survey to share their insights.
Data Analysis
In this section, the statistical methods employed to investigate the research questions and
hypotheses in this study were outlined. The present study focused on the effectiveness of Amira
Learning software usage and feedback in improving students' oral reading fluency and its
alignment with the SOR framework. Descriptive statistics were initiated to gain a comprehensive
understanding of the central tendencies of the study variables, which provided an initial insight
into the data's characteristics. Subsequently, an assessment of the normality of the data
distribution was conducted through histogram visualizations and the Kolmogorov-Smirnov test.
These assessments indicated that the study variables did not adhere to a normal distribution,
leading to the selection of the Wilcoxon Signed Rank test, a non-parametric statistical test that is
better suited to analyze the change in oral reading fluency scores from pretest to posttest.
For research question one, the goal was not only to assess the relationship between
DIBELS Next ORF pre and post test scores (dependent variable) and Amira Learning software
usage in minutes (independent variable) but to also predict and model the oral reading fluency
achievement levels based on the time spent using Amira Learning software. To explore the
50
relationship between Amira Learning software usage and learners' oral reading fluency, a simple
regression analysis was conducted. This enabled the ascertainment of the extent to which Amira
Learning software usage was associated with improvements in oral reading fluency. The
regression analysis provided insights into how the time spent on the platform affected the oral
reading fluency scores, considering other potential factors that might be influencing fluency.
Regression analysis allowed the researcher of the present study to estimate the strength and
direction of the relationship, identify any significant predictors, and generate a regression
equation that was used for predictions (Hinkle, et al., 2003).
To address the second research question regarding the effectiveness of Amira Learning
software's feedback in enhancing oral reading fluency, new variables representing the difference
between pretest and posttest oral reading fluency scores were derived. This allowed the
quantification of the degree of improvement in oral reading fluency over the study duration. The
paired t-test was initially chosen as a statistical method in the present study to compare the
means of two paired measurements taken from the same individual. Two statistical analyses,
linear and multiple linear regressions, were conducted to ensure validity of the results of the
study.
For research question 3, teachers' insights on the alignment and effectiveness of Amira
Learning software with the SOR framework were examined through chi-square goodness-of-fit
analyses. These analyses provided valuable insights into teachers' perceptions regarding Amira
Learning software. Pie charts were used to show how teachers responded to the two questions.
All statistical analyses were performed using the Statistical Package for the Social
Sciences (SPSS) version 29.0.0.0. This program has been believed to facilitate a thorough
examination of the research questions and enabled a robust and systematic analysis of the
51
collected data. The SPSS software has been a widely used statistical analysis tool in the social
sciences and has offered a comprehensive suite of techniques for analyzing and visualizing data.
The utilization of this software in the present study ensured a rigorous and systematic
examination of the data, thereby contributing to the robustness and credibility of the findings. A
predetermined significance level of 0.05 was used in the present study. The chosen analytical
approaches were carefully selected to ensure the validity of the findings and to effectively
address the specific research questions.
Research Design
The present study used a quasi-experimental quantitative research design to explore the
research questions. Quantitative designs are fixed and deductive, with variables and hypotheses
designed prior to data collection. For research questions one and two, quasi-experimental design
was used; specifically, an uncontrolled quasi-experimental before and after design to measure the
improvement in oral fluency due to time spent within Amira Learning software, and Amira
Learning software feedback. The present study did not alter any protocols at the school site to
obtain data for the study. Instead, secondary data from the school was used and analyzed. After
receiving an exemption for the study from the Institutional Review Board (IRB), the researcher
of the present study obtained permission from school and district level administrators to use their
previously collected DIBELS Next and Amira Learning software data. The data used in the
present study were analyzed using the Statistical Package for the Social Sciences (SPSS) version
29.0.0.0.
For research question 1, the impact of Amira Learning software usage on students’ ORF
scores, the student data that was used were collected immediately before, and immediately after
the use of the Amira Learning software using DIBELS Next ORF measure. Tests of assumptions
52
were performed to ensure the validity and reliability of the regression model and its results.
Regression analysis made several assumptions about the data and the relationships between
variables. By testing these assumptions, the researcher was able to assess whether the model was
appropriate for the data and whether the results can be trusted (Hinkle, et al., 2003). Any
observed differences in performance were assumed to be due to the Amira Learning software
usage and tutoring. It was hypothesized that, assuming there was a moderate to high positive
relationship between ORF scores and Amira Learning software usage, high usage of Amira
Learning software will result in high posttest ORF scores.
For research question 2, the effects of Amira Learning software's feedback on students'
reading skills, data used was also collected from DIBELS Next ORF and Amira Learning
software platforms. The degree of correlation between ORF score and Amira Learning software
feedback in minutes was originally planned to be analyzed using paired t-tests on SPSS 29.0.0.0.
Tests of assumptions were performed using the program. Similar to regression analysis,
conducting assumption checks are also important when performing a paired t-test. A paired t-test
is a statistical test used to compare the means of two related groups or conditions, often before
and after an intervention or treatment (Hinkle, et al., 2003), and in this case, after receiving
feedback from Amira Learning software. Because the assumptions for normality for this data
were not met, an alternative non-parametric test, the Wilcoxon signed-rank test was considered
and the data was analyzed using linear regression.
For research question 3, Amira Learning software’s alignment with SOR and its
objectives, classroom teachers were asked to participate in a brief online survey. Teachers were
sent a Microsoft Forms online link to answer a two-question survey about their insights on the
level of alignment of Amira Learning software and the Science of Reading (SOR). The survey
53
asked the teachers, “How would you rate the effectiveness of Amira Learning software in
accomplishing the goals of the Science of Reading?” and “How aligned is Amira Learning
software with the Science of Reading?”. For each question, the teachers were provided the
categorical measures to represent their responses. The first question had teachers choose one of
the following responses: not effective, effective, highly effective. The second question had the
teachers choose from the following responses: no alignment, moderate alignment, high
alignment. Descriptive statistics were used to analyze the teachers’ responses on this survey.
Summary
The present study sought to investigate the effects of an AI program, Amira Learning
software, on literacy development in early childhood education. Amira Learning software is an
Artificial Intelligence in Education (AIEd) Intelligent Tutoring System (ITS) designed to
improve oral reading fluency for early learners. The study used a quantitative research
methodology to systematically collect and analyze numerical data to answer the research
questions. The focus was on whether the AI application was effective in enhancing oral reading
fluency in early learners. To ensure rigor and coherence, the study followed a statistical approach
and adhered to a well-articulated research design, which served as a comprehensive plan to
address the research questions. The data collection methods retrieved the necessary information
for the study, and the chosen statistical tests were based on the type of data and research
questions to provide accurate and meaningful results. The present study's use of quantitative
methods and statistical analysis aimed to draw valid conclusions and produce accurate and
reliable findings to contribute to the growing interest in AIEd and its potential applications in
education, especially in response to the challenges posed by the Covid-19 pandemic on human
capital.
54
This chapter provided an overview of the present study, covering essential aspects like
research paradigm, definition of key terms, setting and context, ethical considerations, data
sources, data analysis, and research design. In Chapter Four, the quantitative results of the
present study are discussed.
55
CHAPTER FOUR. RESULTS
Effects of Amira Learning on Pre and Post Test Oral Reading Fluency Scores
The present study aimed to investigate the effects of an AI program, Amira Learning, on
literacy development in early childhood education. The impact of Amira Learning was explored
through analysis of the difference in the students’ pretest and posttest oral reading fluency scores
following the 6-week usage of the software.
A Wilcoxon Signed Ranks Test was conducted to evaluate the efficacy of the Amira
Learning software in enhancing the oral reading fluency (ORF) of first-grade students. This
analysis compared pretest and posttest ORF scores, providing valuable insights into the impact of
the usage. The results revealed a substantial and statistically significant difference in ORF scores
between the pretest (Mdn = 19.0, IQR = 0-126) and posttest (Mdn = 47.0, IQR = 4-149) conditions,
Z= -7.67, p < .001. There was a notable increase in the median ORF score from 19.0 before the
usage to 47.0 after its completion. The accompanying rank results provided additional insights into
the comparison of ORF scores: notably, there were no negative ranks, indicating that in all
instances, the ORF posttest scores surpassed their corresponding pretest scores. A total of 78
positive ranks were observed and one tie was recorded between ORF pretest and ORF posttest
scores, which was a singular instance of identical scores. The improvement in scores ranged from
0 (from pretest and posttest tied) to 61 words correct per minute (WCPM). Table 1 presented Oral
Reading Fluency (ORF) scores for first-grade students in a suburban Louisiana school district (N
= 79) before (Pre) and after (Post) a six-week usage of Amira Learning software.
56
Table 1. Comparison of Oral Reading Fluency (ORF) Scores Before and After Amira Learning
Usage
Pretest Posttest
Variable M (SD) Mdn (IQR) M (SD) Mdn (IQR)
Oral Reading Fluency
29.57 (30.02) 19.0 (0-126) 49.05 (31.63) 47.0 (4-149)
Score
The box plot, Figure 2, illustrated the distribution of ORF scores for first-grade students
before (pretest) and after (posttest) usage of Amira Learning Software, which was conducted over
a 6-week period. The central line within each box represented the median ORF score, while the
box itself spanned the interquartile range (IQR). Whiskers extend to the minimum and maximum
values within 1.5 times the IQR. This revealed a substantial improvement in ORF scores among
first-grade students following the six-week usage of the Amira Learning software. The median
ORF score post-usage notably surpassed the pre-usage median.
Figure 2. Effects of Amira Learning Usage on First-Grade Students' Oral Reading Fluency
(ORF) Scores
57
The pie chart, Figure 3, showed the percent per category increase in the ORF scores of
first grade students. The trajectory for average growth was to see an improvement in ORF scores
by about 2 WCPM per week (Hasbrouck & Tindal, 2017). A total of 34% (27/79) of the students
gained between 0-11 word correct per minute (WCPM) in ORF after the 6-week usage of Amira
Learning. Out of the 79 students, 33% (26/79) were reported to have increased their scores by
12-23 WCPM, 22% (17/79) increased their scores by 24-35 WCPM, and 11% (9/79) gained 36
or more WCPM in their ORF posttest.
Figure 3. Growth in Oral Reading Fluency (WCPM scores)
11% 0--11
34%
12 -- 23
22%
24--35
36 & above
33%
Relationship Between Time Spent on Amira Learning and Oral Reading Fluency
The first research question addressed in the present study was “What is the relationship
between the time spent using Amira Learning software and learners' oral reading fluency
achievement levels?” This research question aimed to examine the relationship between the
amount of time students spent using the Amira Learning software over a six-week period (Amira
Learning average usage in minutes) and the improvement in their ORF scores as measured by
DIBELS Next (www.acadeincelearning.org, n.d.). Specifically, it sought to understand whether
there was a connection between the extent of software usage and the degree of improvement in
reading fluency. Additionally, the research question aimed to assess the predictive value of Amira
58
Learning average usage in determining the enhancement in oral reading fluency scores. In essence,
it explored the role of software usage in influencing learners' reading fluency improvements.
A linear regression analysis was performed to assess the relationship between Amira
Learning software usage, specifically Amira Learning average usage in 6 weeks (in minutes),
and learners' ORF improvement scores. The study included a sample of 79 participants, and ORF
improvement scores denoted the difference between pretest and posttest assessments. The
dependent variable in this analysis was ORF improvement.
Table 2 displayed the results of a linear regression analysis examining the relationship
between Amira Learning average usage in 6 Weeks (in minutes) and learners' ORF improvement
scores. Figure 4, the scatter plot, illustrated the relationship between Amira Learning average
usage in 6 weeks (in minutes) and learners' ORF improvement scores. Each point on the plot
represented an individual participant within the sample of 79. The x-axis denoted the time spent
using Amira Learning, while the y-axis represented the improvement in ORF scores. The scatter
plot demonstrated a scattered distribution of points, suggesting that while there was a modest
positive association between Amira Learning average usage and ORF improvement, there was
variation in improvement scores among participants.
The results of analysis revealed a significant model effect, F (1, 77) = 4.79, p = .032. This
indicated that the model, which included Amira Learning average usage as a predictor, provided
a statistically significant explanation for variance in the ORF improvement scores. R2 value of
0.059, suggesting that approximately 5.9% of the variance in the ORF improvement scores can
be attributed to the independent variable, Amira Learning average usage. Subsequently,
unstandardized coefficients demonstrated that, after controlling for the constant, each additional
minute allocated to using Amira Learning over a six-week period was associated with an
59
increase of 0.20 units in ORF improvement scores (B = 0.20, SE = .09, t = 2.19, p = .032).
Additionally, a standardized coefficient (β = 0.24) indicated a positive and moderately strong
relationship between Amira Learning average usage and ORF improvement scores.
Table 2. Linear Regression Analysis: Relationship Between Amira Learning Usage and ORF
improvement
Unstandardized Standardized
Coefficients Coefficients
Predictor B SE β t R2 F
0.06 4.79*
Constant 14.90 2.48 6.01***
Amira Average .20 .09 .24 2.19*
Usage in 6 Weeks
(in Min)
Figure 4. Scatter Plot - Relationship Between Amira Average Usage and ORF improvement
60
Amira Learning’s Feedback and Improving Oral Reading Fluency
A second research question was investigated in the present study. This research question
asked “How effective is the feedback provided by Amira Learning in improving students’ oral
reading fluency?”. Amira Learning tutoring was equated to the feedback that Amira Learning
provided the students, hence, the association between Amira Learning tutoring and the
improvement in ORF scores among learners was explored. This research question sought to
understand whether increased engagement with Amira Learning tutoring, as indicated by
additional tutoring time, was linked to significant improvements in ORF scores. Additionally, the
question studied the predictive value of Amira Learning tutoring time in determining the
enhancement in learners' oral reading fluency.
Table 3 displayed the results of a linear regression analysis examining the relationship
between Amira Learning tutoring time in 6 weeks (in minutes) and learners' ORF improvement
scores. The sample size consisted of 79 participants. ORF improvement scores represent the
difference between pretest and posttest scores. Dependent variable was the ORF improvement. b
represented unstandardized regression weights. SE indicated standard error of b. β indicated the
standardized regression weights. R2 indicated variances predicted by the independent variables.
*Indicated p < .05, **indicated p < .01** indicated p < .001
Figure 5, a scatter plot, showcased the connection between Amira Learning tutoring time
in 6 weeks (in minutes) and learners' ORF improvement scores. Each data point corresponds to a
participant from the sample of 79 individuals. The x-axis represented the time invested in Amira
Tutoring, while the y-axis displayed the enhancement in oral reading fluency scores. The scatter
plot revealed a dispersed arrangement of data points, indicating variability in improvement
scores among participants. However, the linear regression analysis uncovered a significant
61
relationship (p < .05), demonstrating that greater time spent with Amira Tutoring corresponded
to substantial improvements in oral reading fluency.
The second results of a linear regression analysis examining the association between
Amira Learning tutoring time in 6 weeks (in minutes) and learners' ORF improvement scores
were presented in Table 3. The model's overall statistical significance was evaluated using an F-
test, which yielded a significant result, F (1, 77) = 4.46, p = .038. This indicated that the model,
including Amira Learning tutoring time as a predictor, held explanatory power. The R2 value
(0.05) illustrated that approximately 5.5% of the variance in ORF improvement scores can be
attributed to the independent variable, Amira Learning tutoring time. The regression coefficients
showed that, after controlling for the constant, for every additional minute spent using Amira
Tutoring over a six-week period, there was an associated increase of 0.30 units in ORF
improvement scores (B = 0.30, SE = 0.14, t = 2.11, p = .038). The standardized coefficient (β =
0.23) indicated a positive and moderate relationship between Amira Learning tutoring time and
ORF improvement.
Table 3. Linear Regression Analysis: Relationship Between Amira Learning Tutoring Time and
ORF Improvement
Unstandardized Standardized
Coefficients Coefficients
Predictor B SE β t R2 F
0.05 4.46*
Constant 14.98 2.52 5.94***
Amira Learning .30 .14 .23 2.11*
tutoring time in 6
Weeks (in Min)
62
Figure 5. Scatter Plot - Relationship Between Amira Learning tutoring time and ORF
improvement
To further investigate the relationship between Amira Learning tutoring time in 6 weeks
(in minutes) and learners' ORF posttest (Post-ORF) scores, while controlling for ORF pretest
(Pre-ORF) scores, a multiple linear regression was conducted . A two-step multiple linear
regression analysis was conducted to explore the relationship between Amira Tutoring Time in 6
weeks (in minutes) and learners' post-ORF scores, while considering pre-ORF scores. In the first
model, it was observed that both pre-ORF scores (B = 0.95, SE = .05, β = 0.90, t = 20.46, p <
.001) and Amira Tutoring Time (B = 0.34, SE = .15, β = 0.10, t = 2.33, p = .022) significantly
predicted post-ORF scores. This initial model accounted for a substantial proportion of the
variance in post-ORF scores (R² = 0.86). Subsequently, a second model introduced an interaction
term between pre-ORF scores and Amira Tutoring Time to examine whether Amira Learning
feedback impact on performance differed depending on the learners' initial fluency levels. In this
second model, pre-ORF scores (B = 0.95, SE = .05, β = 0.90, t = 19.19, p < .001) and Amira
Tutoring Time (B = 0.34, SE = .15, β = 0.10, t = 2.31, p = .023) remained significant predictors
63
of post-ORF scores, while the interaction term was not significant (t = -0.005, p = .996). Table 4
showed the results of the multiple linear regression.
Table 4. Multiple Linear Regression Analysis: Relationship Between Amira Tutoring Time and
Post Oral Reading Fluency Score While controlling Pre- Oral Reading Fluency Score
Unstandardized Standardized
Coefficients Coefficients
Predictor B SE β t R2 F
Model 1
0.86 237.04***
Constant 49.05 1.34 36.61***
Pre-ORF .95 .05 .90 20.46***
Amira Tutoring .34 .14 .10 2.33**
Model 2
0.86 155.95***
Constant 49.05 35.53***
Pre-ORF .95 .90 19.19***
Amira Tutoring .34 .10 2.31**
Pre-ORF X Amira -.0002 .01 .000 -.005
Tutoring
Teachers’ Perceptions on Amira Learning and Science of Reading
A third research question focused on the topic of Amira Learning and its alignment with
the Science of Reading. The question that was explored in the present study was: “What is the
level of agreement among teachers regarding Amira Learning’s alignment to the Science of
Reading (SOR) and its effectiveness in fulfilling the Science of Reading objectives?”. The
question sought to understand the collective viewpoint of teachers regarding Amira Learning's
alignment with SOR and its effectiveness in fulfilling SOR objectives. It explored teachers'
64
perspectives on the extent to which Amira Learning has aligned with the Science of Reading and
its perceived role in achieving these educational goals.
In the final research question, two separate chi-square goodness-of-fit analyses were
conducted to assess the level of agreement among teachers regarding Amira Learning's
alignment with the Science of Reading and its effectiveness in achieving Science of Reading
goals.
Alignment Rating
Regarding the alignment of Amira Learning with the Science of Reading, participants
were asked to rate the alignment as no alignment, moderate alignment, or high alignment.
Results indicated that 3 teachers (17.6%) perceived moderate alignment, while a significant
majority of 14 teachers (82.4%) perceived high alignment. None of the teachers who participated
perceived low alignment with SOR. A chi-square test was performed, revealing a statistically
significant result χ2(1) = 7.12, p = .008.
Effectiveness Rating
For the question concerning the effectiveness of Amira Learning in achieving Science of
Reading goals, teachers were asked to rate the AIEd as either not effective, effective, or highly
effective. The observed frequencies indicated that 8 teachers (47.1%) rated it as effective, while
9 teachers (52.9%) rated it as highly effective. A chi-square test yielded a non-significant result
χ2(1) = .059, p = .808, indicating no significant deviation from the expected distribution. These
results suggest that teachers generally agree on the effectiveness of Amira Learning in achieving
SOR goals. A chi-square test yielded a non-significant result χ2(1) = .059, p = .808, indicating no
significant deviation from the expected distribution. These results suggest that teachers generally
agree on the effectiveness of Amira Learning in achieving SOR goals.A chi-square test yielded a
65
non-significant result χ2(1) = .059, p = .808, indicating no significant deviation from the
expected distribution. These results suggest that teachers generally agree on the effectiveness of
Amira Learning in achieving SOR goals.
Table 5. Teachers' Ratings of Amira Learning's Alignment and Effectiveness in Achieving Science
of Reading Goals
Variable N %
Alignment Rating of Amira Learning with
the Science of Reading
Moderate Alignment 3 17.6
High Alignment 14 82.4
Effectiveness Rating of Amira Learning in
Achieving Science of Reading Goals
Effective 8 47.1
Highly Effective 9 52.9
Figure 6. Amira Learning’s Alignment and Effectiveness Rating
66
CHAPTER FIVE. DISCUSSION AND IMPLICATIONS
The purpose of the present study was to investigate the effects of Amira Learning, an
Artificial Intelligence (AI) program, on the literacy development in early childhood education.
Specifically, it sought to determine 1) the relationship between the time spent using Amira
Learning and learners' oral reading fluency (ORF) achievement levels; 2) the effectiveness of the
feedback provided by Amira Learning software in improving students’ oral reading fluency; and
3) the level of agreement among teachers regarding Amira Learning software’s alignment to the
Science of Reading and its effectiveness in fulfilling the Science of Reading objectives.
Amira Learning and Oral Reading Fluency
The present study delved into investigating the effects of an Artificial Intelligence in
Education (AIEd) program on oral reading fluency (ORF) of first grade students. ORF has been
established in the literature as a critical measure of overall reading success and an indicator of
reading progress (Hintze et.al., 2002; Jenkins, et.al., 2003; Schatschneider et.al., 2002; Speece
& Ritchey, 2005). In the early stages of reading, children learn to decode words and understand
the relationship between sounds and letters. Fluent reading at this stage is crucial because it
allows students to focus less on decoding individual words and more on understanding the
meaning of the text. Fluent readers are better equipped to comprehend what they read, which
has also been a fundamental goal of reading instruction.
The results of the Wilcoxon Signed Ranks Test (Wilcoxon, 1945) to evaluate the
efficacy of Amira Learning software in enhancing ORF of first-grade students suggested a
statistically significant difference in ORF scores between the pretest and posttest median scores.
The substantial increase in postest median ORF scores of the students after the 6-week usage of
Amira Learning software denotes the positive effect of using AIEd to improve ORF, one of the
67
foundational literacy measures. What was also notable about this increase in median scores was
the change in the DIBELS Next benchmark bands from Below Benchmark at pretest to At
Benchmark at posttest.
DIBELS Next (www.acadiencereading.org, n.d.) establishes benchmark scores for each
grade level to serve as reference points for assessing students' reading skills. The benchmark
categories have been used to help educators assess and monitor students' reading proficiency
to determine when intervention is needed. These categories are based on grade-level
expectations and provide a framework for understanding students' progress in early literacy
skills. Each category indicates how students perform on the assessment based on grade-level
expectations and each score represents the predicted literacy outcomes for the students. The
DIBELS Next categories Above Benchmark, At Benchmark, Below Benchmark, and Well
Below Benchmark.
Students in Above Benchmark category perform at a level that exceeds grade-level
expectations. They demonstrate strong proficiency in the assessed skill, and their performance
suggests a high level of competence. Students in Above Benchmark may not require additional
intervention or support in the assessed skill. Students in At Benchmark category perform at a
level that aligns with grade-level expectations. They meet the expected standard for their grade
in the assessed skill. Being At Benchmark is an indicator that a student has achieved a reasonable
level of proficiency and is on track for reading success at their grade level.
Students in Below Benchmark category perform below the grade-level expectations for
the assessed skill. They may be struggling with the skill to some extent, and their performance
suggests that they may benefit from additional support or targeted instruction. Being Below
Benchmark indicates a need for closer monitoring and possibly intervention.
68
Students in Well Below Benchmark category are significantly below the grade-level
expectations for the assessed skill. They have typically been at the greatest risk for reading
difficulties or literacy challenges. Students in Well Below Benchmark require intensive and
specialized intervention and support to improve their skills and catch up to their peers
(www.acadiencereading.org, n.d.).
Prior to the use of Amira Learning software, the average score of the students was Below
Benchmark which predicted that typical students in the grade level will likely need additional
support, such as intervention or tutoring, to become successful in reading. Students who were
unable to receive the support needed will likely be at more risk of reading deficits in the future.
Providing this level of support to individual students can be labor-intensive for teachers and their
class size and daily schedule may prevent them from providing the level of support and feedback
that each student needs. After the 6-week usage of and interaction with Amira Learning software,
the average ORF score of the students was a level higher which was At Benchmark category,
which predicted successful achievement of reading outcomes provided that the students receive
high-quality core English Language Arts (ELA) instruction. Using the Amira Learning software
as an intelligent reading tutor allowed all students the support and feedback that they needed to
progress in their reading skills without providing additional teaching staff or using extensive
class periods to attend to individual student needs.
Another noteworthy effect of Amira Learning software on the students’ ORF scores was
the amount of growth on individual scores that were recorded from pretest to posttest. Amira
Learning software was able to positively affect the students’ accuracy and rate of reading as
measured by the number of words correct per minute (WCPM). Consistent with the principles
of the Skilled Reader, these gains in ORF posttest meant that students will be more equipped
for language comprehension. This result suggested positive effects of Amira Learning software
69
on ORF even when used in a relatively short period.
To address the first research question “What is the relationship between the time spent
using Amira Learning software and learners' oral reading fluency achievement levels?” A linear
regression analysis was conducted to assess this connection, which focused specifically on
Amira Learning software average usage over a six-week period. The analysis demonstrated a
significant model effect and implied that the inclusion of Amira Learning software usage as a
predictor in the model offered a statistically significant explanation for the variance observed in
the ORF improvement scores. In other words, there was a meaningful relationship between
how much time students spend using Amira Learning software and their ORF score
improvements. This underscored the potential benefits of incorporating Amira Learning
software into reading instruction to support students in developing their reading skills
effectively.
These findings were consistent with the results of a study by Consortium for Policy
Research in Education (2021) when Savannah-Chatham County Public School (SCCPSS)
partnered with Amira Learning software to improve literacy outcomes. The study explored the
effects of Amira Learning software on four literacy outcomes, including ORF. The study span
over two different periods, fall to winter and winter to spring. The findings of this study shed
light on the connection between the frequency of Amira Learning software usage and the
growth in WCPM over the fall-to-winter period. Specifically, for each additional week of
usage, there was a statistically significant increase of 0.030 standard deviations in WCPM
scores (p < .001).
The results of the present study found a correlation between the amount of time students spent
using Amira Learning software and the notable improvement in their ORF scores. This implied
Amira Learning software, as an ITS, has acknowledged that learning should be an ongoing
70
process. Amira Learning software's approach has encouraged continuous improvement by
adapting content and difficulty levels based on individual progress, thus supporting the idea that
learning to read has never been a one-time event but a continuous journey that will impact
overall reading success with consistent use. These can be related with the theory of mastery of
learning which has acknowledged the differences in the pace at which learners acquire
knowledge and has highlighted the significance of allowing students ample time for achieving
mastery (Bloom, 1968; Guskey, 2005).
Amira Learning Feedback and Oral Reading Fluency
To fully investigate Amira Learning software’s impact on students’ ORF scores, two
statistical analyses were conducted. These ensured the credibility of the results that the
improvement in ORF posttest scores can be attributed to Amira Learning feedback. Results of
the present study demonstrated that Amira Learning software tutoring time or feedback, was
responsible for the increase in DIBELS Next ORF scores of the first grade students. This result
aligned with the theory of mastery learning (Bloom, 1968) which has underscored the
significance of ongoing evaluation of students' learning to track their advancement and deliver
prompt feedback. The present study investigated the efficacy of the feedback offered by Amira
Learning software in directing learners toward mastery and supporting their learning outcomes
in early reading. When students do not achieve mastery at first, Bloom (1968) suggested that
feedback and opportunities for remediation should be offered to students. In the light of the
present study, Amira Learning software’s feedback mechanisms and adaptive nature allowed
students to receive feedback and additional practice when needed, supporting the concept of
revisiting and reinforcing learning until mastery has been achieved. Although not all students
achieved At Benchmark status on the DIBELS ORF posttest, a notable 66% of the students who
used Amira Learning software and engaged in the feedback process increased their ORF scores
71
by 12 points or more, denoting a positive literacy outcome in just six weeks.
The findings also supported Blooms’ (1968) theory which emphasized progressing from
lower-order thinking skills (e.g., in literacy decoding and blending) to higher-order thinking
skills (e.g., fluent reading and comprehension). Amira Learning software's approach has been
designed to scaffold learning, starting with foundational skills and gradually moving students
toward more complex reading tasks and higher-order thinking, thus aligning with Bloom's
taxonomy. Amira Learning has claimed to offer feedback and support, while also allowing the
students the opportunity to engage in productive struggle and to practice their reading skills
(www.amiralearning.com, n.d.).
In the context of the Skilled Reader theory, the findings of the present study indicated
that Amira Learning software, as an Intelligent Tutoring System (ITS), effectively offers
instruction and feedback to support the development of individual students into skilled readers.
The study's results highlighted the efficacy of Amira Learning software software’s feedback in
improving students' ORF. Amira Learning software has followed the research around skilled
readers and has conformed to SOR principles as evidenced by the results of the present study,
which has also been consistent with the research by Consortium for Policy Research in
Education (2021). Amira Learning software has taken pride in encouraging students to use their
abilities to read and not rely on the cueing system to guess and has worked on individual
students to strengthen their ORF skills (www.amiralearning.com, n.d.).
The results of the present study demonstrated that Amira Learning software’s feedback
was consistent with the principles of Simple View of Reading (SVR), Scarborough’s (2001)
Reading Rope, and the Five Pillars of Literacy as evidenced by the growth in WCPM across
the board, regardless of students’ pretest score. In other words, Amira delivered the mix of
72
what was necessary for reading success, decoding and language comprehension which resulted
in increased ORF posttest scores.
Teachers’ Insights on Amira Learning and the Science of Reading
The third research question of the present study explored teachers’ insights regarding
Amira Learning software’s alignment with the Science of Reading (SOR) and its effectiveness in
fulfilling SOR objectives. The teachers who participated in the present study recently completed
SOR training that consisted of 50 hours of asynchronous, modular work and 6 1-hour
synchronous discussion sessions with other participants and the facilitator of the training. At the
time of participation, the teachers had completed Amira Learning software implementation
training and had the opportunity to use Amira Learning software for a period of six months.
The teachers’ responses to the online survey showed teachers’ belief in Amira
Learning’s ability to teach young students according to SOR. In both of the questions they had
to answer, there was no recorded response that negatively associated Amira Learning software
with SOR. This supported the idea of teachers’ insights that Amira Learning software was
designed and implemented in a way that was consistent with the scientific understanding of
how reading skills are acquired and developed. SOR has an evidence-based approach to
literacy instruction that has drawn from decades of research in cognitive psychology,
linguistics, neuroscience, and education (Moats, 1999; Moats 2020; Science of Reading:
Defining Guide, 2022). Based on these responses, teachers felt that the software has followed
the evidence-based practices and principles that have been shown to be effective in teaching
reading skills to young students.
The results of the teacher survey also suggested that the teachers had high confidence in
allowing Amira Learning software to instruct and provide feedback to their students despite the
73
controversies surrounding using AIEd in early childhood education. Although the teachers
were not asked to explain their responses, the inclusion criteria ensured that all teachers who
participated had sufficient knowledge of the objectives of SOR and Amira Learning
software, following the intensive training they completed for SOR and Amira Learning
software implementation.
As teachers expressed their confidence regarding the alignment of Amira Learning
software with SOR and its effectiveness in fulfilling its objectives, the present study has
also acknowledged its connection to the mastery learning theory (Bloom, 1978). Consistent
with Bloom's (1978) theory, the teachers' responses highlighted how Amira Learning
software provided students with instruction using well-defined learning goals, sequential
progression, individualized pacing, and using assessments to drive instruction.
In essence, both SOR and mastery learning principles have emphasized the importance
of clearly defined and specific learning objectives. SOR has placed great importance on
foundational skills such as phonological awareness, phonics, vocabulary development, and
fluency, recognizing them as essential components for achieving overall reading proficiency.
This alignment with the mastery learning approach, which has established explicit and
measurable goals for students, was evident. While Amira Learning software has not offered
whole group, explicit instruction, as an intelligent tutor, it has provided individualized
opportunities for students to practice their reading skills while receiving timely, focused
feedback based on their reading needs.
SOR has recognized the significance of a structured progression aligned with the
developmental stages of reading acquisition. This progression was based on research into how
children learn to read and the cognitive processes involved. While the exact progression may
74
vary based on student individual needs, the general progression has been from less complicated
reading skills to the more sophisticated ones. This suggested that phonemic awareness comes
first before fluency, which will come first before comprehension. Similarly, mastery learning has
often organized objectives in a sequential manner, ensuring that students have achieved mastery
of prerequisite skills before moving on to more advanced ones. This alignment has reinforced
the concept that students established a solid groundwork before engaging in more complex
reading tasks. Amira Learning software has pledged to provide the right amount of scaffolds
through feedback that has been aligned with SOR to meet students’ reading needs.
Both SOR and mastery learning have promoted individualized pacing. SOR has
recognized that students may progress at different rates and require different levels of support.
In the same manner, mastery learning has allowed students to proceed at their own pace,
receiving additional instruction and support as needed until they achieve mastery. This
individualized approach has ensured that all students have had the opportunity to master the
content. Amira Learning, as it has advertised, has offered personalized instruction while
adapting content and difficulty levels based on each student's performance. This individualized
approach has been aligned with the customization of instruction in both SOR and mastery
learning.
SOR and mastery learning have both emphasized the alignment of assessments with
learning objectives and have used them to guide future instruction. In SOR, assessments
have been used to monitor student progress and identify areas of strength and weakness in
reading skills. SOR has placed a strong emphasis on providing feedback and support to
students to help them develop reading skills. In mastery learning, assessments have been
closely tied to specific objectives, ensuring that students meet predefined criteria for
75
mastery. Mastery learning has also highlighted the importance of feedback and has offered
opportunities for remediation when students do not achieve mastery on assessments. This
feedback loop has been essential for both approaches to support student learning. Both SOR
and mastery learning have recognized that learning is an ongoing and continuous process,
therefore, assessments should be used to inform future instruction. Amira Learning software
actively listened as students read aloud, capturing the session and producing an objective
running record that was free from subjective inaccuracies and testing biases. Its AI
capability made it possible for Amira Learning software to adjust the rigor of stories and
feedback based on previous reading sessions.
Ultimately, the goal of reading instruction has been to develop skilled readers who are
fluent, are able to comprehend what they read, and apply their knowledge of reading into
different context or content, demonstrating mastery of their learning. The principles of skilled
reader and mastery learning both have supported the idea that all learners can become skilled
readers when provided with appropriate instruction, feedback, assessment, and time for
learning. They have recognized that reading proficiency is a developmental process and that
strong foundational skills are crucial for success in reading and overall academic achievement.
Limitations
The present study has acknowledged the possible biases in data analyses due to the
limited access to data. Other variables and external factors not considered in this study could
have influenced the outcomes of the study. Some of these factors include student
demographics such as age, socioeconomic status, or educational background, variations in
teacher competence, differences in the teaching pedagogies, and additional services received
by students such as small group instruction, interventions, or human tutoring outside of
76
school. Given that the present study did not use a control group nor controlled variables when
running data analyses, the results yielded may have been affected. The duration of the study
was relatively short and could have been extended to capture longer-term effects or changes
that could occur over time.
Responses from teacher participants may have been partial because of the teachers’
relationship with the researcher, who was their curriculum specialist. Although the participants
were assured that the survey was anonymous and confidential, their responses might have been
influenced by their worry of being judged because of their knowledge of Amira Learning
software or SOR or being identified based on their responses because of the small number of
participants in the study.
Clinical Implications
The positive effects of Amira Learning software on literacy development in early
childhood education are multifaceted and can have several significant clinical implications for
educators, administrators, policymakers, stakeholders, and AI program creators. As Amira
Learning has been considered a determinant in improving ORF scores, the present study has
highlighted the importance of early intervention and feedback in literacy development
(Guskey, 2005). The present study supports previous research that recognizes that identifying
and addressing reading difficulties in early childhood can prevent future, long-term academic
challenges (Shanahan, 2020; Moats, 2020, Petscher, et al.., 2021; Science of Reading: Defining
Guide, 2022).
The results of the present study implied that incorporating an AIEd intelligent tutor into
early childhood education can lead to improved reading and literacy outcomes for young
learners. Amira Learning software's adaptive nature could allow early childhood educators to
77
better tailor their instruction to individual student needs, promoting personalized learning
experiences (Wijekumar, et al., 2018) without the burden of planning and facilitating one-on-
one sessions with every student every day. Using Amira Learning software can also suggest that
early intervention using technology-based literacy programs can be effective. Educators can
consider incorporating software such as this into early childhood interventions to prevent
having students who will be at risk of literacy difficulties in the future. Amira Learning software
may underscore the importance of data-driven assessment and progress monitoring. Educators
can use the reports generated from the platform to track children's literacy development over
time, enabling more informed decisions about intervention strategies and classroom instruction.
Amira Learning software's data collection and analysis capabilities can enable educators a low
labor-intensive, yet accurate measure to make data-driven decisions about students' progress.
This implies a need for data literacy among educators.
Administrators can benefit from the results of the present study by using the information
presented in making informed decisions about technology purchases and staffing.
Administrators who wish to use Amira Learning software should consider other factors related
to the use of the software such as access to high-speed, reliable internet, dependable electronic
devices, and other additional supplies such as headsets with a microphone. The aftermath of
Covid-19 depleted the workforce and made it hard for administrators to find teachers and
support staff. Amira Learning software can be utilized as a low-cost, non-labor intensive
personalized reading tutor to address the staffing issue without sacrificing the quality of support
offered to the students.
This study also has the potential to influence policymakers in their decisions to allocate
resources for schools and early childhood education centers to acquire and implement
educational technology like Amira Learning software. Although the present study only
78
measured results for first grade students, the results show promise and could be scaled across
the state for all students. The present study's findings may also shed light on potential
disparities in access to educational technology. In light of the positive results found in this
study, policymakers should consider issues of equity to ensure that all students, regardless of
their background, can have access to AIEd and that technology infrastructures, such as the
internet and computers, are available to our most vulnerable populations.
The positive effects of Amira Learning software may have encouraged parents to engage
with their children's education by using the software or similar technology at home. Using Amira
Learning software or other literacy-based AIEd may provide parents the supplemental tool to
support their child's literacy development at home. It can serve as an additional resource to
reinforce what children are learning in school. The interactive nature of Amira Learning
software and the immediate feedback given to students may help in motivating young children
to engage in educational activities at home. At the time of writing the present study, home
access to Amira Learning software was available for students via the district’s online learning
platform. The platform can only be accessed through the district-provided Chromebook, which
was not sent home to students in the early grades. The results of the present study may drive
parents to become advocates for equitable access to technology-based learning tools, ensuring
that all children have the opportunity to benefit from effective resources regardless of their
background or location.
AI creators may have also been influenced by the alignment of Amira Learning software
with SOR and be inspired to create similar technology to support early readers. The positive
impact of AI-powered educational tools like Amira Learning software on early childhood
literacy development suggests that AI creators should focus on continuous improvement, ethical
considerations, and collaboration with educational stakeholders to maximize the benefits of AI
79
in education while addressing challenges and ensuring accessibility and inclusivity.
Recommendations for Future Research
The findings from the present study revealed critical information on using an AIEd in the
form of an intelligent tutor to improve literacy outcomes for young children. Several future
research can be done based on the results of the present study to further expand understanding of
the potential benefits and challenges associated with integrating AIEd during this crucial stage
of learning.
Longitudinal studies to examine the long-term effects of Amira Learning software on
literacy development can help determine whether early gains in literacy skills are sustained over
time. An investigation on whether the effectiveness of Amira Learning software varies across
different demographic groups, such as students from diverse socioeconomic backgrounds or
with varying levels of prior literacy skills would be valuable. Future researchers can conduct
comparative studies to evaluate the effectiveness of Amira Learning software in comparison to
other literacy interventions and instructional approaches commonly used in early childhood
education. Exploring the impact of teacher training and ongoing support in maximizing the
benefits of Amira Learning software can also be a topic of interest in the future. Additionally,
research can focus on how well-prepared educators are in integrating AIEd effectively into their
teaching practices. A study on motivation and engagement to investigate the impact of Amira
Learning software on students' motivation, engagement, and attitudes towards reading and
literacy can also be explored.
Conducting qualitative studies to gain deeper insights into the experiences of teachers
and students using Amira Learning software should be considered. Qualitative data can provide
80
a richer understanding of the learning process and can enable researchers to investigate the
challenges and barriers that educators may face when integrating Amira Learning software into
their curriculum, including technical issues, time constraints, and resource limitations.
By addressing these research areas, future studies can provide valuable insights into the role
of an intelligent AIEd tutor like Amira Learning software in early childhood education and
inform best practices for leveraging technology to support literacy development in young
learners.
Conclusion
Overall, the results of data analyses and the evaluation of the connections to the
theoretical frameworks of the present study have indicated that Amira Learning software, as
an AI intelligent tutor, provided students with a strong foundation in ORF, feedback, and
ongoing assessment information to teachers. Additionally, teachers confirmed this program as
being in alignment with the Science of Reading literature. The Amira Learning software
platform’s individualized instruction has prioritized student learning and achievement as the
central goals of its tutoring sessions while continuously growing individual students based on
their unique needs. Amira Learning has given the students the 1-on-1 attention they need,
something that would have not been possible if not for the power of AI.
Summary
The present study sought to answer the following questions to explore the overarching
goal of the study which was to investigate the effects of Amira Learning software on literacy
development in early childhood education.
81
RQ1: What is the relationship between the time spent using Amira Learning and
learners' oral reading fluency achievement levels?
RQ2: How effective is the feedback provided by Amira Learning in improving
students’ oral reading fluency?
RQ3: What is the level of agreement among teachers regarding Amira Learning’s
alignment to the Science of Reading and its effectiveness in fulfilling the Science of
Reading objectives?
This chapter discussed the results of the data analyses, answered the research questions,
and offered explanations on the findings as they relate to the theoretical frameworks of mastery
learning (Bloom, 1968) and the skilled reader. It also discussed the limitations of the study, the
implications for practice, and recommendations for future research to inform practice in the
areas of early childhood education, educational technology, and literacy development.
82
APPENDIX A. CENTRAL TENDENCY TABLES AND FIGURES
Figure A1. Frequencies of the Oral Reading Fluency (ORF) Scores at Pretests
Figure A2. Frequencies of the Oral Reading Fluency (ORF) Scores at Posttests
Figure A3. Frequencies of Amira Average Usage Minutes in 6 Weeks
83
Figure A4. Frequencies of Amira Tutoring Time Average in 6 Weeks
Regression Model 1 – IV: Amira Average Usage in 6 Weeks (in Min
Figure A5. Histogram of Regression Standardized Residual of ORF Improvement
Figure A6. Normal P-P Plot of Regression Standardized Residual of ORF Improvement
84
Figure A7. Testing the Homoscedasticity Assumption for Regression of ORF Improvement
Regression Model 2 – IV: Amira Tutoring Time in 6 Weeks (in Min)
Figure A8. Histogram of Regression Standardized Residual of ORF Improvement
Figure A8. Normal P-P Plot of Regression Standardized Residual of ORF Improvement
85
Figure A9. Testing the Homoscedasticity Assumption for Regression of ORF Improvement
86
APPENDIX B. INSTITUTIONAL REVIEW BOARD
87
APPENDIX C. INFORMED CONSENT FORMS
School/ District Administrator Consent Form
1. Study Title: The Effects of Amira Learning on Literacy Development in Early Childhood
Education
2. Purpose of the Study: The main purpose of this study is to assess the effectiveness of an
Artificial Intelligence program, Amira Learning, on literacy development of first grade
students. Secondary data will be used to delve into this goal. As part of exploring this main
goal, the study also aims to understand teachers’ insights on Amira Learning’s alignment
with the Science of Reading. Teachers will be asked to take a two-question survey to rate
their agreement on the alignment of Amira Learning with the Science of Reading. The survey
should take no more than 5 minutes to take and will be administered using a secured,
protected platform (MS Forms). The teachers will not be asked any identifying information
during this survey.
3. Risks: There is no anticipated risk associated with participation in the study.
4. Benefits: There will be no direct benefit to the subjects for participating in this study.
However, it is hoped that the information obtained from this study may drive district’s future
decisions on obtaining programs that support teachers and in exploring additional online
classroom resources and interventions.
5. Investigators: The following investigators are available for questions about this study, please
contact Dr. Cynthia DiCarlo (cdicar2@lsu.edu) or Ms. Caroline Tolentino (ctolen2@lsu.edu)
6. Number of Subjects: 17
7. Inclusion Criteria: Teachers who hold a valid Louisiana teaching license, who are over the
age of 18, and who have been trained in the Science of Reading. To participate in this study,
the requirements of both the inclusion and exclusion criteria must be met.
13. Exclusion Criteria: Teachers without a valid Louisiana teaching license, are under the age of
18, and have not been trained on the Science of Reading.
14. Right to Refuse: Subjects may choose not to participate or to withdraw from the study at any
time without penalty or loss of any benefit to which they might otherwise be entitled.
information will be included for publication.
15. Privacy: Results of the study may be published, but no names or identifying information will
be included in the publication. Subject identity will remain confidential unless disclosure is
required by law.
88
16. Financial Information: There is no cost for participation in the study, nor is there any
compensation to the subjects for participation.
17. Signatures:
______ I consent for the researcher to use secondary data requested and for teachers to
participate in the study if they wish to. The study has been discussed with me and all my
questions have been answered. I may direct additional questions regarding study specifics to
the investigator. If I have questions about subjects' rights or other concerns, I can contact
Alex Cohen, Chairman, Institutional Review Board, (225) 578-8692, irb@lsu.edu, or
www.lsu.edu/research. I will allow the use of students’ data in the study described above
and acknowledge the investigator's obligation to provide me with a signed copy of this
consent form.
______ I do not consent to participation in this study.
Superintendent Name (printed or typed): _________________________________________
Superintendent Signature: ________________________________ Date:________________
89
Teacher Consent Form
1. Study Title: The Effects of Amira Learning on Literacy Development in Early Childhood
Education
2. Purpose of the Study: The main purpose of this study is to assess the effectiveness of an
Artificial Intelligence program, Amira Learning, on literacy development of first grade
students. Part of exploring this main goal is to understand teachers’ insights on Amira
Learning’s alignment with the Science of Reading. Your participation is needed to
accomplish this secondary goal.
3. Risks: There is no anticipated risk associated with participation in the study.
4. Benefits: There will be no direct benefit to you for your participation in this study.
However, it is hoped that the information obtained from this study may drive district’s
future decisions on obtaining programs that support teachers and in exploring additional
online classroom resources and interventions.
5. Investigators: The following investigators are available for questions about this study, Dr.
Cynthia DiCarlo, (cdicar2@lsu.edu) or Ms. Caroline Tolentino (ctolen2@lsu.edu)
6. Number of subjects: 17
7. Inclusion Criteria: Teachers who hold a valid Louisiana teaching license, who are over
the age of 18, and who have been trained in the Science of Reading. To participate in this
study, you must meet the requirements of both the inclusion and exclusion criteria.
8. Exclusion Criteria: Teachers without a valid Louisiana teaching license, are under the age
of 18, and have not been trained on the Science of Reading.
90
9. Right to Refuse: Subjects may choose not to participate or to withdraw from the study at
any time without penalty or loss of any benefit to which they might otherwise be entitled.
10. Privacy: Results of the study may be published, but no names or identifying information
will be included in the publication. Subject identity will remain confidential unless
disclosure is required by law.
11. Signatures:
The study has been discussed with me and all my questions have been answered. I may
direct additional questions regarding study specifics to the investigators. For injury or
illness, call your physician, or the Student Health Center if you are an LSU student. If I
have questions about subjects' rights or other concerns, I can contact Alex Cohen,
Institutional Review Board, (225) 578-8692, irb@lsu.edu, or www.lsu.edu/research. I
agree to participate in the study described above and acknowledge the investigator's
obligation to provide me with a signed copy of this consent form.
Subject Signature: ____________________________ Date: ________________
The study subject has indicated to me that he/she is unable to read. I certify that I have
read this consent form to the subject and explained that by completing the signature line
above, the subject has agreed to participate.
Signature of Reader: ____________________________ Date: _______________
91
REFERENCES
Abdelsalam, U. M. (2014). A proposal model of developing intelligent tutoring systems based on
mastery learning. In The Third International Conference On E-Learning In
Education (pp. 106-118).
Alkhatlan, A., & Kalita, J. (2018). Intelligent tutoring systems: A comprehensive historical
survey with recent developments. arXiv preprint arXiv:1812.09628.
Amira. (n.d.). www.amiralearning.com. https://www.amiralearning.com/
Amira Learning. (n.d.). White paper: Amira is Science of reading [White paper].
https://amiralearning.com/documents/White-Paper-Amira-Is-The-Science-Of-
Reading.pdf
Atun H. (2020). Intelligent tutoring systems (its) to improve reading comprehension: a
systematic review, Journal of Teacher Education and Lifelong Learning, 2(2), 77-89.
Bailey, D. H., Duncan, G. J., Murnane, R. J., & Au Yeung, N. (2021). Achievement gaps in the
wake of COVID-19. Educational Researcher, 50(5), 266-275.
Bloom, B. S. (1968). Learning for mastery. Instruction and Curriculum. Regional Education
Laboratory for the Carolinas and Virginia, Topical Papers and Reprints, Number
1. Evaluation comment, 1(2), n2.
Bloom, B. S. (1975). Evaluation, Instruction and Policy Making. IIEP Seminar Paper: 9.
Bloom, B. S. (1984). The 2 sigma problem: The search for methods of group instruction as
effective as one-to-one tutoring. Educational researcher, 13(6), 4-16.
Bredekamp, S. (2020). Developmentally appropriate practice in early childhood programs
serving children from birth through age 8 (4th edition). National Association for the
Education of Young Children. 1313 L Street NW Suite 500, Washington, DC 22205-
4101.
Britto, P. (2015, December 15). Why early childhood development is the foundation for
sustainable development. Www.unicef.cn. https://www.unicef.cn/en/stories/why-early-
childhood-development-foundation-sustainable-development
92
Broda, M., & Frank, A. (2015, October). Learning beyond the screen: assessing the impact of
reflective artificial intelligence technology on the development of emergent literacy
skills. In E-Learn: World Conference on E-Learning in Corporate, Government,
Healthcare, and Higher Education (pp. 753-758). Association for the Advancement of
Computing in Education (AACE).
Chavez, R. (2021, November 16). Louisiana public schools grapple with learning lost to
pandemic surges and storms. PBS. https://www.pbs.org/newshour/nation/90-percent-of-
kindergarteners-in-louisiana-arent-ready-for-school-thanks-to-pandemic-disruptions
Chen, X., Xie, H., & Hwang, G. J. (2020). A multi-perspective study on artificial intelligence in
education: Grants, conferences, journals, software tools, institutions, and
researchers. Computers and Education: Artificial Intelligence, 1, 100005.
Consortium for Policy Research in Education. (2021).
https://www.amiralearning.com/documents/Matches-Human-Tutoring-After-30-
Sessions.pdf
Department for Business Innovation & Skills. (2016). Success as a knowledge economy:
Teaching excellent, social mobility and student choice [White paper]. Crown.
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/523396/bis
-16-265-success-as-a-knowledge-economy.pdf
Dias, M. J., Almodóvar, M., Atiles, J. T., Vargas, A. C., & Zúñiga León, I. M. (2020). Rising to
the Challenge: Innovative early childhood teachers adapt to the COVID-19
era. Childhood Education, 96(6), 38-45.
Dong, Y., Yu, X., Alharbi, A., & Ahmad, S. (2022). AI-based production and application of
English multimode online reading using multi-criteria decision support system. Soft
Computing, 26(20), 10927-10937.
Dorn, E., Hancock, B., Sarakatsannis, J., & Viruleg, E. (2021). COVID-19 and education: The
lingering effects of unfinished learning. McKinsey & Company, 27.
Dynamic Measurement Group. (2011). Phase 2 DIBELS Next Data Interpretation Workshop.
Eyisi, D. (2016). The usefulness of qualitative and quantitative approaches and methods in
researching problem-solving ability in science education curriculum. Journal of education
and practice, 7(15), 91-100.
93
Ford, T. G., Kwon, K. A., & Tsotsoros, J. D. (2021). Early childhood distance learning in the US
during the COVID pandemic: Challenges and opportunities. Children and Youth Services
Review, 131, 106297.
Fuchs, L. S., Fuchs, D., Hosp, M. K., & Jenkins, J. R. (2001). Oral reading fluency as an
indicator of reading competence: A theoretical, empirical, and historical
analysis. Scientific studies of reading, 5(3), 239-256.
Golberstein, E., Wen, H., & Miller, B. F. (2020). Coronavirus disease 2019 (COVID-19) and
mental health for children and adolescents. JAMA pediatrics, 174(9), 819-820.
Gough, P. B., & Tunmer, W. E. (1986). Decoding, reading, and reading disability. Remedial and
special education, 7(1), 6-10.
Grimshaw, J., Campbell, M., Eccles, M., & Steen, N. (2000). Experimental and quasi-
experimental designs for evaluating guideline implementation strategies. Family
practice, 17(suppl_1), S11-S16.
Guo, L., Wang, D., Gu, F., Li, Y., Wang, Y., & Zhou, R. (2021). Evolution and trends in
intelligent tutoring systems research: a multidisciplinary and scientometric view. Asia
Pacific Education Review, 22(3), 441-461.
Guskey, T. R. (2005). Formative classroom assessment and Benjamin S. Bloom: Theory,
research, and implications. Online submission.
Harris, A. D., McGregor, J. C., Perencevich, E. N., Furuno, J. P., Zhu, J., Peterson, D. E., &
Finkelstein, J. (2006). The use and interpretation of quasi-experimental studies in medical
informatics. Journal of the American Medical Informatics Association, 13(1), 16-23.
Hasbrouck, J., & Tindal, G. (2017). An update to compiled ORF norms (No. 1702). Technical
report.
Hinkle, D. E., Wiersma, W., & Jurs, S. G. (2003). Applied statistics for the behavioral
sciences. (No Title).
Hintze, J. M., Callahan III, J. E., Mathews, W. J., Williams, S. A. S., & Tobin, K. G. (2002).
Oral reading fluency and prediction of reading comprehension in African American and
Caucasian elementary school children. School Psychology Review, 31, 540–553.
Hoofman, J., & Secord, E. (2021). The effect of COVID-19 on education. Pediatric
Clinics, 68(5), 1071- 1079.
94
International Society for Technology in Education. (2019). ISTE standards.
https://cdn.iste.org/www-root/PDF/ISTE%20Standards-One-Sheet_Combined_09-
2021_vF3.pdf
International Society for Technology in Education (2020, January 10). Innovative designer 4D:
Open-ended problems (ISTE standards for students). [Video file]. YouTube.
https://www.youtube.com/watch?v=FvCc3D6XNrg&list=PL6aVN_9hcQEGNw8OUN-
5Cc_dogE1Q4gqo&index=4
Jenkins, J. R., Fuchs, L. S., van den Broek, P., Espin, C., & Deno, S. L. (2003). Sources of
individual differences in reading comprehension and reading fluency. Journal of
Educational Psychology, 95, 719–729.
Jeong, H. I., & Kim, Y. (2017). The acceptance of computer technology by teachers in early
childhood education. Interactive Learning Environments, 25(4), 496-512.
Ji, X. R., Beerwinkle, A., Wijekumar, K., Lei, P., Malatesha Joshi, R., & Zhang, S. (2018).
Using latent transition analysis to identify effects of an intelligent tutoring system on
reading comprehension of seventh-grade students. Reading and Writing, 31, 2095-2113.
Knudsen, E. I., Heckman, J. J., Cameron, J. L., & Shonkoff, J. P. (2006). Economic,
neurobiological, and behavioral perspectives on building America’s future
workforce. Proceedings of the national Academy of Sciences, 103(27), 10155-10162.
Kivunja, C., & Kuyini, A. B. (2017). Understanding and applying research paradigms in
educational contexts. International Journal of higher education, 6(5), 26-41.
LaBerge, L., O’Toole, C., Schneider, J., & Smaje, K. (2020). How COVID-19 has pushed
companies over the technology tipping point—and transformed business
forever. McKinsey & Company, 5.
LeMoine, S. (2020). Power to the Profession Task Force: Foundational to Early Childhood
Education. ZERO TO THREE, 40(3), 20-23.
Lindner, A., & Romeike, R. (2019). Teachers' perspectives on artificial intelligence. In 12th
International conference on informatics in schools,“Situation, evaluation and
perspectives”, ISSEP.
95
Loble, L., & Hawcroft, A. (2022). Shaping AI to Tackle Australia’s Learning Divide
Lock, R. H., & Welsch, R. G. (2006). Increase oral reading fluency. Intervention in School and
Clinic, 41(3), 180-183.
Louisiana Department of Education. (2022). Act no. 108.
https://legis.la.gov/legis/ViewDocument.aspx?d=1232840
Luckin, R., Holmes, W., Griffiths, M., & Forcier, L. B. (2016). Intelligence unleashed: An
argument for AI in education.
Lurye, S. (2023, May 17). “Mississippi miracle”: Kids’ reading scores have soared in Deep
South states. AP News. https://apnews.com/article/reading-scores-phonics-mississippi-
alabama-louisiana-5bdd5d6ff719b23faa37db2fb95d5004
Ma, W., Adesope, O., Nesbit, J., & Liu, Q. (2014). Intelligent tutoring systems and learning
outcomes: A meta-analysis. Journal of educational psychology, 106(4), 901.
McCarthy, J. (2007). From here to human-level AI. Artificial Intelligence, 171(18), 1174-1182.
Meta- Analysis of Research on Amira: Intelligent tutoring's Impact Amira Learning. (2022,
July).Retrieved from https://www.amiralearning.com/research.html
Moats, L. C. (1999). Teaching reading is rocket science: What expert teachers of reading should
know and be able to do.
Moats, L. C. (2020). Teaching reading "is" rocket science: What expert teachers of reading
should know and be able to do. American Educator, 44(2), 4.
Mousavinasab, E., Zarifsanaiey, N., R. Niakan Kalhori, S., Rakhshan, M., Keikha, L., & Ghazi
Saeedi, M. (2021). Intelligent tutoring systems: a systematic review of characteristics,
applications, and evaluation methods. Interactive Learning Environments, 29(1), 142-
163.
National Reading Panel. (2000). Teaching children to read: An evidenced-based assessment of
the scientific research literature on reading and its implications for reading instruction.
https://www.nichd.nih.gov/sites/default/files/publications/pubs/nrp/Documents/report.pdf
96
Nickow, A., Oreopoulos, P., & Quan, V. (2020). The impressive effects of tutoring on prek-12
learning: A systematic review and meta-analysis of the experimental evidence.
Passow, A. (2019, January 3). How k-12 schools have adopted artificial intelligence: AI
integration offers new potential to improve student outcomes and security. EdTech Focus
on K-12. https://edtechmagazine.com/k12/article/2019/01/how-k-12-schools-have-
adopted-artificial-intelligence
Petscher, Y., Cabell, S. Q., Catts, H. W., Compton, D. L., Foorman, B. R., Hart, S. A., ... &
Wagner, R. K. (2020). How the science of reading informs 21st‐century
education. Reading research quarterly, 55, S267-S282.
Plitnichenko, L. (2020, May 30). 5 main roles of artificial intelligence in education. eLearning
Industry. https://elearningindustry.com/5-main-roles-artificial-intelligence-in-education
Poletti, M. (2020). Hey teachers! Do not leave them kids alone! Envisioning schools during and
after the coronavirus (COVID-19) pandemic. Trends in neuroscience and education, 20,
100140.
PwC. (2017). Sizing the prize: What’s the real value of AI for your business and how can you
capitalise? https://www.pwc.com/ gx/en/issues/analytics/assets/pwc-ai-analysis-sizing-
the-prize -report.pdf
Queirós, A., Faria, D., & Almeida, F. (2017). Strengths and limitations of qualitative and
quantitative research methods. European journal of education studies.
Roscoe, R. D., Allen, L. K., Weston, J. L., Crossley, S. A., & McNamara, D. S. (2014). The
Writing Pal intelligent tutoring system: Usability testing and development. Computers
and Composition, 34, 39-59.
Russo, A. (2020, October 20). Recession and automation changes our future of work, but there
are jobs coming, report says. World Economic Forum.
https://www.weforum.org/press/2020/10/recession-and-automation-changes-our-future-
of-work-but-there-are-jobs-coming-report-says-52c5162fce/
Sabates, R., Carter, E., & Stern, J. M. (2021). Using educational transitions to estimate learning
loss due to COVID-19 school closures: The case of Complementary Basic Education in
Ghana. International Journal of Educational Development, 82, 102377.
97
Samuelsson, I. P., & Kaga, Y. (Eds.). (2008). The contribution of early childhood education to a
sustainable society (pp. 1-136). Paris: Unesco.
Scarborough, H. S. (2001). Connecting early language and literacy to later reading (dis)abilities:
Evidence, theory, and practice. In S. Neuman & D. Dickinson (Eds.), Handbook for
research in early literacy. New York: Guilford Press.
Science of Reading: Defining Guide. (2022). The Reading League.
Schatschneider, C., Carlson, C. D., Francis, D. J., Foorman, B. R., & Fletcher, J. M. (2002).
Relationship of rapid automatized naming and phonological awareness in early reading
development. Implications for the double-deficit hypothesis. Journal of Learning
Disabilities, 35, 245–256.
Schwartz, S. (2022, July 20). Which states have passed “Science of Reading” laws? What’s in
them?. Education Week. https://www.edweek.org/teaching-learning/which-states-have-
passed-science-of-reading-laws-whats-in-them/2022/07
Seidenberg, M. S. (2013). The science of reading and its educational implications. Language
learning and development, 9(4), 331-360.
Sentell, W. (2022, January 18). New report highlights literacy problems in Louisiana: “It is
unfathomable.” The Advocate.
https://www.theadvocate.com/baton_rouge/news/education/new-report-highlights-
literacy-problems-in-louisiana-it-is-unfathomable/article_661a2d1c-77cb-11ec-a678-
23cdb6577031.html
Shanahan, T. (2020). What constitutes a science of reading instruction?. Reading Research
Quarterly, 55, S235-S247.
Spatz, C. (2019). Exploring statistics: Tales of distributions. Outcrop Publishers.
Speece, D. L., & Ritchey, K. D. (2005). A longitudinal study of the development of oral reading
fluency in young children at risk for reading failure. Journal of Learning Disabilities, 38,
387–399.
Straker, L., Zabatiero, J., Danby, S., Thorpe, K., & Edwards, S. (2018). Conflicting guidelines on
young children's screen time and use of digital technology create policy and practice
dilemmas. The Journal of pediatrics, 202, 300-303.
98
Su, J., & Yang, W. (2022). Artificial intelligence in early childhood education: A scoping
review. Computers and Education: Artificial Intelligence, 3, 100049.
Touretzky, D., Gardner-McCune, C., Martin, F., & Seehorn, D. (2019, July). Envisioning AI for
K-12: What should every child know about AI?. In Proceedings of the AAAI conference
on artificial intelligence (Vol. 33, No. 01, pp. 9795-9799).
UNESCO. (2020). National education responses to Covid-19 summary report of UNESCO’s
online survey. https://unesdoc.unesco.org/ark:/48223/pf0000373322.
Unicef. (2017). Programme guidance for early childhood development Unicef programme
division 2017. https://www.unicef.org/media/107616/file/UNICEF-Programme-
%20Guidance-for-Early-Childhood-Development-2017.pdf
Unicef. (2020). Policy guidance on AI for children.
https://www.unicef.org/globalinsight/media/1171/file/UNICEF-Global-Insight-policy-
guidance-AI-children-draft-1.0-2020.pdf.
Van Lancker, W., & Parolin, Z. (2020). COVID-19, school closures, and child poverty: a social
crisis in the making. The Lancet Public Health, 5(5), e243-e244.
Wijekumar, K. K., Meyer, B. J., & Lei, P. (2013). High-fidelity implementation of web-based
intelligent tutoring system improves fourth and fifth graders content area reading
comprehension. Computers & Education, 68, 366-379.
Wijekumar, K., Meyer, B. J., Lei, P. W., Lin, Y. C., Johnson, L. A., Spielvogel, J. A., ... & Cook,
M. (2014). Multisite randomized controlled trial examining intelligent tutoring of
structure strategy for fifth-grade readers. Journal of Research on Educational
Effectiveness, 7(4), 331-357.
Wilcoxon, F. (1945). Individual comparisons by ranking methods. Biom Bull 1 (6): 80–83.
World Health Organization. (2023, May 18). WHO COVID-19 dashboard. World Health
Organization. https://covid19.who.int/
Yang, W. (2022). Artificial Intelligence education for young children: Why, what, and how in
curriculum design and implementation. Computers and Education: Artificial
Intelligence, 3, 100061.
Zhang, D., Mishra, S., Brynjolfsson, E., Etchemendy, J., Ganguli, D., Grosz, B., ... & Perrault, R.
(2021). The ai index 2021 annual report. arXiv preprint arXiv:2103.06312.
99
Zorić, A.B. (2020). Benefits of educational data mining. Journal of International Business
Research and Marketing, 6(1), 12-16.
100
VITA
Caroline C. Tolentino, a native of the Philippines and naturalized citizen of the United
States, received her bachelor’s degree from the University of the Philippines, Diliman in 2004.
She taught in the Philippines from 2004 until right before she migrated to Baton Rouge
Louisiana in 2008. She became a kindergarten teacher and instructional specialist in Baton
Rouge where she served for more than 10 years before moving to Ascension Parish. Caroline’s
passion for educating young students and her love for learning made her decide to pursue her
graduate degree. In 2015, she was admitted to the master’s program at the Louisiana State
University, College of Human Sciences and Education. She obtained her master’s degree in 2017
and educational specialist degree in 2020.
Caroline has taught undergraduate courses, teaching preservice teachers in PK3 Early
Childhood Program since 2020. In 2021, she became an instructional coach in Ascension Parish
supporting teachers and students in the elementary grades. In her career as an educator, she has
presented at numerous conferences, published academic articles, received several teaching
awards, and wrote a number of grants. Caroline intends to obtain her Ph.D. in Curriculum and
Instruction in December 2023.
101