ESSA Tier 4 Evidence-Based Rationale

Introduction🔗

Passed in 2015, the Every Student Succeeds Act (ESSA) is a US law that governs K-12 public education policy. ESSA aims to provide all students with a fair, equitable, and high-quality education. One of the key components of ESSA is the classification of educational programs into tiers based on their effectiveness and evidence of impact.

Tier 4 programs are those that have not yet demonstrated evidence in well-designed and well-implemented studies but demonstrate a rationale for their potential effectiveness and are in the process of gathering evidence. This tier allows new and innovative programs to be recognized and be eligible for some federal funding while they continue to build their evidence base.

Pictures Are for Babies is a literacy software tutor designed to help student master every aspect of literacy, from alphabetic knowledge to reading and writing at the graduate and professional level. Given the extent of its scope, innovative technology, and uncompromising adherence to research-based practices, it is important that educators and policymakers understand how Pictures Are for Babies is based on a rigorous logic model that meets all the requirements for ESSA Tier 4 compliance and how it plans to gather evidence of its effectiveness.

Logic Model🔗

Problem Statement🔗

Despite overwhelming scientific consensus on the mechanisms that drive reading and writing acquisition, actual instruction often fails to align with its recommendations. Even in programs that align with research, much of its implementation is left to educators, who cannot reach fidelity even with extensive training. The state of literacy instruction is even more dire for struggling students, who require extensive individualized instruction. In the absence of what Elizabeth Tipton and Nicole Patton-Terry call an "engineering of reading", the "science of reading" is no more than a marketing badge that fails to translate to real outcomes even when the programs align with its recommendations.

National and international assessments of literacy confirm that literacy instruction is failing to meet the needs of students. The OECD Survey of Adult Skills (PIAAC) shows that 28% of adults have low literacy proficiency. The Program for International Student Assessment (PISA) shows that 26% of students in all member countries score below level 2 in reading, with the United States having 19% of students in this category. The National Assessment of Educational Progress (NAEP) shows 32% of 12th graders in the United States scoring below basic in reading.

These figures are often the focus of the discussion, but the results paint a more dire picture when focusing on the highest-performing students. Across all three assessments and all countries, the number of students demonstrating the highest levels of literacy is under 10%, with some assessments grouping the statistics for this group with the next highest level because the number is so small to be meaningfully reported as its own category. These assessments test only reading performance, so the number of students demonstrating the highest levels at both reading and writing is likely even lower.

Pictures Are for Babies is not designed for mere remediation of low-performing students or to meet some government standard. From its very inception and through every aspect of its design and implementation, it aims to engineer the conditions that make the highest levels of literacy the default for all students that complete it. In light of the continued failures of the education systems of every country in the world, this goal might seem impossible. However, as Louisa C. Moats states in her seminal article "Teaching Reading Is Rocket Science", "reading failure can be prevented in all but a small percentage of children with serious learning disorders. It is possible to teach most students how to read if we start early and follow the significant body of research showing which practices are most effective".

The rocket science of literacy instruction is settled, but without the corresponding aerospace engineering, the spaceship will not reach its faraway destination. What Tipton and Patton-Terry describe as the "engineering of reading" is a system designed from the ground up to solve user problems rather than test hypotheses, to anticipate variation across learners and contexts rather than control for it, to make errors the designer's responsibility rather than the teacher's, and to be refined through iteration rather than validated through a single study.

Unbeknownst to Tipton and Patton-Terry, the author, working independently and without institutional support, had already completed much of what they call for at the time of publication. That this system was built by a single person is not a coincidence. An effort this uncompromising could only have emerged outside the constraints of committees, investors, and institutional consensus. The forces that prevent established organizations from integrating research at this depth are the same forces that prevent them from naming a literacy program Pictures Are For Babies, let alone commit to the pedagogical and engineering decisions such an ambitious goal requires.

Logic Model Summary🔗

Pictures Are For Babies is built on a foundation of best-in-class pedagogy, drawing from a multidisciplinary body of research in cognitive science, expert performance, optimal learning, the science of reading, and the science of writing instruction. These disparate fields not only align, but reinforce each other, forming a coherent theoretical and practical framework for effective literacy instruction.

Unfortunately, while the research has been replicated and validated for decades, classroom instruction and most commercial programs are, at best, decades behind, and, at worst, actively hostile to it. Some programs correctly and thoroughly execute parts of the research, but none can claim to integrate all of it into a coherent whole. Pictures Are For Babies is the first program to do so. Before detailing how every aspect of the program is grounded in research, this section presents a high-level summary of the logic model that underpins it, both to provide a short summary for policymakers and educators, and to provide full context for the detailed logic model that follows.

The backbone of the logic model is the program's focus on deliberate practice. Deliberate practice is a well-established framework for explaining how individuals acquire expertise in any domain. This framework has been validated in a wide range of domains, from music to chess to sports to medicine. It states that expert performance across any domain is achieved through effortful, structured training with immediate feedback aimed at improving specific aspects of performance. Deliberate practice is not meant to be fun, inherently enjoyable, or easy. Deliberate practice succeeds because it relentlessly targets weaknesses at the edge of current ability and develops the learner's ability to self-monitor, self-correct, and their intrinsic motivation to improve.

Deliberate practice is implemented in Pictures Are For Babies through Trane, an engine that creates the conditions for deliberate practice in any domain. It surfaces optimized and personalized exercises based on the learner's past performance, a curriculum designed to target individual skills, and the dependence relationships between these skills. Trane uses many well-established techniques from cognitive science, such as spaced repetition, interleaving, and mastery learning that will be discussed in detail in the next section. Trane was designed from scratch by the creator of Pictures Are For Babies and validated in other fields like music before being applied to literacy instruction. Deliberate practice is not a buzzword or a retrofit, but core to the design and implementation of Pictures Are For Babies.

Existing literacy research and recommendations are well aware of the importance of deliberate practice, but the existing solutions have so far been unable to implement it at scale beyond simple adaptive algorithms and self-paced learning models. Without a proper and strict implementation of deliberate practice, we can expect that the number of students that reach the highest levels of literacy will remain very low. No other model of expert performance has been shown to be as effective as deliberate practice, so expecting students to reach those levels is akin to expecting novice pianists to become concert pianists by publishing graded sheet music without years of extensive structured practice and feedback.

This paradigm works much better when the instruction and feedback are individualized to the learner. Trane solves the problems of providing individualized instruction at scale. Pictures Are For Babies goes further by implementing a model of reading and writing instruction that is both scientifically grounded and delivered through simple protocols any fluent reader can follow. Through this combination, teachers, parents, and volunteers can provide individualized instruction to students without needing to be experts and without needing to know the prior knowledge of the student. Schools can thus implement both small group and one-on-one instruction without needing to hire additional staff or provide extensive training.

By treating deliberate practice as an engineering problem, Pictures Are For Babies is able to implement and enforce the science of reading and writing instruction in a way that is impossible for human instructors and printed materials to do. For example, by keeping track of the performance of every word lesson, the software can identify when a student should receive a higher tier intervention and alert the tutor. While a teacher might be able to identify that a student is struggling, Pictures Are For Babies can perform this analysis with precise performance data and at the lesson level on a word curriculum that spans over 17,500 words. In another example, the concept of Scarborough's reading rope explains how the many different skills that contribute to literacy acquisition are interdependent and interact. Many products pay lip service to this concept but fail to implement it in practice. In Pictures Are For Babies, the concept is implemented without any human intervention because Trane makes sure that the exercises shown to the student come from a variety of courses and lessons testing different skills. Many more of these examples will be discussed in the next section.

Beyond allowing for implementing and enforcing the science of reading and writing instruction, Trane also allows Pictures Are For Babies to implement a curriculum that is far more extensive than any other program and to thoroughly adhere to the science without fear that it will require large amounts of effort on the part of the tutor. The curriculum of Pictures Are For Babies covers every aspect of literacy, namely:

• A symbols track that teaches alphabetic knowledge and knowledge of basic punctuation and symbols.
• A words track that teaches word-level reading and writing. By having Trane take care of spaced review and interleaving, the words track can cover over 17,500 words with a simple intervention protocol that does not require the tutor to have any knowledge of linguistic concepts, detailed scripts, or lesson plans. The only requirements for a tutor are to be a fluent reader, the phonetic skills all fluent readers possess, and the willingness to learn and follow this simple protocol.
• A sentences track that teaches sentence-level reading and spelling and implicitly introduces syntax and grammar. Each sentence course follows a corresponding word course and is designed to reinforce the words learned in that course.
• A reading comprehension track that teaches reading comprehension through a combination of reading a large variety of texts of many types, genres, and topics and developing the background knowledge necessary to understand them. The texts in this track are carefully selected and annotated with example questions for the tutor and footnotes that provide additional context and background knowledge.
• A writing track that teaches writing by having students study and analyze worked examples of explicit writing instruction exercises at the sentence, paragraph, and multi-paragraph level. Each writing course follows a corresponding reading comprehension course and its material is derived from the texts in that course to further reinforce it. Writing production is outside the scope of the deliberate practice framework for reasons explained in the next section, but tutors can use the writing exercises to provide writing production practice and feedback to students.

In terms of adherence to the science of reading and writing instruction, Pictures Are For Babies follows all the standard recommendations for literacy instruction. These choices are thoroughly explained in the next section, but a few examples include:

• Systematic and explicit instruction of alphabetic knowledge, phonemic awareness, and phonics.
• Teaching spelling along with reading to reinforce each other.
• Reading comprehension instruction based on the simple view of reading by teaching decoding and language comprehension skills.
• Explicit writing instruction that teaches students how to write at the sentence, paragraph, and multi-paragraph level.

In order to create a system that engineers the conditions for expert literacy performance in most students, it is necessary to consider the needs of the lowest-performing students. The scientific base has shown that reading difficulties mostly stem from deficits in phonological processing, and that with early and adequate intervention, over 90% of students can achieve grade-level reading proficiency. By focusing its multi-tiered word course intervention protocol on developing the required phonological processing skills, Pictures Are For Babies helps the students with deficits in this area. For students with deficits in other areas, professionals can use Pictures Are For Babies to provide individualized instruction and deliberate practice while they focus on delivering the specific interventions that these students require. By requiring mastery of the lower-level skills before moving on to higher-level skills, Pictures Are For Babies ensures that any gap in skills is addressed before it becomes a larger academic problem. In this way, Pictures Are For Babies can serve as a universal literacy intervention that handles the needs of all students, from those with learning and phonological deficits to the highest-performing group.

Detailed Logic Model🔗

This section details the research and evidence that underpins the design and implementation of Pictures Are For Babies. It is organized into multiple sections for each of the fields that inform the program's design, and each section shows a principle from the research followed by an explanation of how Pictures Are For Babies implements it. For simplicity, Pictures Are For Babies will be referenced as PAFB in the rest of this section. Trane is the name of the practice engine that powers PAFB.

Cognitive Science🔗

• Learning and immediate performance are not the same. Immediate performance in a task is not a guarantee of long-term retention or transfer of that skill.

  • PAFB asks students to repeat the same task multiple times at optimal intervals. Mastered exercises are shown occasionally to make sure foundational skills remain solid.

• Working memory is the type of memory used for immediate performance. This type of memory is very limited in capacity and duration. Any type of instruction that requires working memory incurs a cognitive load that must be effectively managed.

  • PAFB follows this principle in many ways. For example, it scaffolds exercises to ensure only a small amount of new information is introduced at a time, uses a simple UI to minimize distractions, and does not teach complicated phonic rules by name.

• Just like working memory, attention is limited and must be guided.

  • PAFB uses a simple UI to minimize distractions. The sessions include explicit instructions to take breaks to replenish attention. Instructions guide students to focus on the correct aspects of literacy and avoid activities not backed by research.

• In contrast, long-term memory has a vast capacity and retains information for long periods of time. Complicated tasks are performed by moving building blocks to long-term memory. Very complex skills like literacy involve repeated instances of this process.

  • PAFB starts at the very beginning with learning to read and write single letters and very basic words. Once they in long-term memory, sentences and more complex words are introduced. Then real texts and writing tasks in increasing level of complexity.

• Subconscious processes do not require active effort and are performed much faster than tasks that require it. Moving information to long-term memory enables the skills to become automatic and free working memory and attention to work on more advanced skills.

  • Trane supports this process by providing spaced review and gating of progress until mastery of easier skills is proven. Together with the scaffolded lessons, it supports the process of turning new skills into automatic habits.

Expert Performance🔗

• Deliberate practice is the essence of expert performance. It is effortful, structured training with immediate feedback aimed at improving specific aspects of performance. It is not meant to be fun, inherently enjoyable, or easy.

  • PAFB uses Trane to primarily surface exercises at the edge of the student's current abilities. The lessons break literacy into small skills and students receive immediate feedback on their performance.

• Experts continually seek to improve their performance instead of falling into automatic patterns.

  • Trane is always gently pushing students past their current edge of ability. Over time, students build on their current skills until they master more and more difficult ones.

• Experts organize skills into large chunks to process complicated information faster.

  • Trane allows skills to be split into smaller skills and to explicitly encode the relationships among them. As students progress through this connected graph, they are given tasks that require they build these mental structures to demonstrate mastery.

• Expert performers engage in self-regulated learning, setting specific goals, monitoring progress, and adjusting strategies when necessary.

  • Trane tracks learner performance and adapts exercise difficulty accordingly. Tutors can guide students toward deliberate strategy use, especially when progress stalls.

• Experts possess strong metacognitive skills, allowing them to reflect on their performance and make effective adjustments.

  • PAFB trains students to identify confusion and rely on tutors to clarify misunderstandings, building foundational metacognitive habits early.

• Motivation plays a critical role in sustaining the effort required for long-term improvement.

  • Trane and the PAFB tutor model aim to foster intrinsic motivation by tying progress directly to increased competence and ensuring a clear sense of purpose in all tasks.
  • The manual allows the tutor to make use of reading breaks in which they read interesting texts to the student when they detect frustration or boredom. These reading breaks allow students to relax while reinforcing their love of literacy and developing their background knowledge and vocabulary.

• Experts develop flexible, adaptable skills that allow them to transfer knowledge across situations while still relying on deep domain-specific knowledge.

  • Trane’s graph structure allows students to master discrete skills and apply them in increasingly varied contexts, from decoding novel words to understanding complex texts and expressing themselves in writing.

• Early advantages and perceived talent are not predictive of long-term outcomes. Students who appear "gifted" early on often plateau without the right kind of input, while those who build strong foundations through sustained, structured, and deliberate practice eventually surpass them.

  • PAFB produces this kind of quality input for all students, regardless of initial ability. It does not cater to "gifted" early learners or the fastest students. Instead, it supports all students, providing the same rigor, structure, and opportunities for growth. The only difference is individual pace, adjusted based on mastery, not superficial measurements of early ability. Excellence is developed over years, not discovered or accelerated in a race for early achievement.

Optimal Learning🔗

• Mastery learning ensures that students do not progress until they have demonstrated mastery of the current material. This prevents gaps from accumulating and supports long-term growth.

  • Trane automatically withholds access to more difficult material until the student has demonstrated consistent success with prerequisite skills. Mastery is based on performance, not on completion.

• Spaced repetition improves retention by revisiting material in optimal intervals, just before it is forgotten.

  • Trane uses time-aware review scheduling to resurface lessons at optimal intervals based on each student’s history, ensuring long-term retention without overloading them with review.

• Retrieval practice is the act of recalling information from memory, and it has been shown to be more effective than reviewing material or taking notes. The act of trying to recall information strengthens memory and highlights gaps in knowledge.

  • All PAFB exercises are a form of retrieval practice. Students are first asked with recalling how individual words are read and spelled. Higher level tasks require recall of individual words, background knowledge, word meaning, and text structures.

• Interleaving different types of tasks strengthens retention and transfer by forcing students to repeatedly retrieve and distinguish between skills.

  • PAFB implements interleaving through randomized review selection and frequent alternation between decoding, spelling, sentence writing, and comprehension tasks.

• Mixing tasks of varying difficulty improves engagement and learning efficiency. Too much difficulty leads to failure and too much ease leads to stagnation.

  • Trane surfaces a blend of easier and harder tasks based on the student’s performance scores, maintaining an optimal difficulty range for learning.

• Implicit repetition strengthens learning by increasing practice on related tasks when a student struggles, and decreasing practice when a student does well.

  • Trane implements a reward system. When a student performs well, simpler and related skills are given a positive reward so that they are practiced less often. When a student performs badly, related, more complicated skills are given a negative reward so that they are practiced more often.

• Flow is a state of deep concentration and enjoyment that occurs when a task's difficulty is well-matched to the learner’s current skill level. It is characterized by sustained focus, loss of self-awareness, and high intrinsic motivation.

  • PAFB and Trane use all the strategies above to best match the exercises to the student's current abilities. Flow is the result of implementing optimal learning strategies, not of turning learning into gamified experiences or hacking motivation.

• Accelerating the timeline for mastering a complex skill is possible, but only under specific conditions. Research shows that true acceleration does not come from shortcuts, gamification, or early performance boosts. Rather, it comes from compressing high-quality, representative experience into tighter timelines using structured feedback, gradual complexity, and expert-like modeling.

  • Similarly to how PAFB achieves flow, acceleration is not a goal in itself, but the result of implementing optimal learning strategies, combining them with the best reading pedagogy, and a deep and wide curriculum that progressively exposes learners to all aspects of literacy.

The Science of Reading: Word and Sentence Reading and Writing🔗

• Phonological awareness is the ability to recognize and manipulate the sounds of spoken language. Phonemic awareness is a more specific skill involving the ability to isolate and work with individual phonemes, the smallest units of sound.

  • PAFB builds these skills from the beginning. Students learn to hear and manipulate sounds in words before mapping them to letters.

• Most reading difficulties are caused by weaknesses in phonological processing, not vision or motivation. Students who cannot hear, segment, or manipulate sounds will struggle to decode and map words into memory.

  • PAFB targets these core deficits directly. Every word and sentence lesson includes tasks that strengthen the phonological system and link it to written language. No guessing strategies or visual cueing systems are used.

• Orthographic mapping is the central process that allows readers to store words in memory for instant recognition. It links the pronunciation, spelling, and meaning of words, and depends on phonemic awareness and letter–sound knowledge. Without successful orthographic mapping, students will not become fluent readers, no matter how much they are exposed to print.

  • PAFB's lessons on word and sentences are built on this concept. Tutors are instructed to mark words and sentences as mastered only when reading and writing is effortless and automatic.
  • Trane ensures that progress is gated behind accurate mapping of the words, not on mere decoding or guessing.

• Orthographic mapping should extend beyond simple words to include irregular spellings, multisyllabic words, and morphologically complex forms. Fluent reading requires rapid retrieval of tens of thousands of words, including those not easily decoded by rule.

  • PAFB is designed to support deep orthographic mapping. The system introduces increasingly complex word forms only after foundational skills are in place and goes all the way to ensure students master words needed to proficiently read undergraduate and graduate material.
  • The word curriculum of PAFB covers over 17,500 unique words. Through repeated exposure to real texts in more advanced courses, their vocabulary continues to grow beyond the words explicitly taught.

• The self-teaching hypothesis posits that once a student develops efficient decoding skills, successive encounters with new words allow them to apply their existing knowledge to map them in long-term memory with fewer trials. Through this process, fluent readers develop a sight vocabulary that contains tens of thousands of words.

  • PAFB supports this process by providing ample opportunities for students to encounter words in varied contexts, promoting deeper understanding and retention, and enabling generalization of spelling patterns. As students progress through the word curriculum, they steadily use their previous knowledge to expand their sight vocabulary.
  • Unlike the raw application of the hypothesis, PAFB does not leave this to chance. The curriculum and the practice engine ensure students efficiently develop a large sight vocabulary.

• With proper instruction that targets phonemic skills and supports orthographic mapping, the vast majority of children can learn to read proficiently. Research estimates that 90 to 95 percent of children can reach grade-level reading ability with effective interventions.

  • PAFB is built on the methods shown to produce these outcomes. By addressing phonemic awareness, decoding, spelling, and comprehension in the correct order, it ensures that reading failure is not treated as inevitable, but as preventable.

• A hallmark of the most successful interventions is teaching phonemic awareness to proficiency in students that exhibit persistent reading difficulties, including tasks like phoneme deletion, substitution, and reversal. These skills are essential for supporting orthographic mapping and are often missing in students with persistent reading difficulties.

  • PAFB gives explicit instructions on how and when to address deficits in phonemic awareness with these advanced tasks.

• Research supports a tiered model of instruction, where all students receive core instruction and only those who struggle are given additional support.

  • PAFB follows this model closely. Every student receives explicit, structured lessons, but when a student struggles with a task, additional instruction and targeted phonemic work is introduced. This ensures that time is spent efficiently and support is provided only where needed.
  • PAFB uses the student's performance data to determine when to suggest these higher tier interventions and shows a graphical warning.

• Phonics is essential for teaching how letters represent sounds, but it is not enough to produce fluent readers. Letter-sound knowledge must lead to permanent connections between sounds, spellings, and meanings.

  • PAFB uses phonics to organize instruction and support orthographic mapping, but does not teach rules explicitly. Concepts like vowel types or syllable labels are avoided to reduce cognitive load. Every word, regular or irregular, is taught by mapping its phonemes to its spelling, and no word is ever treated as an unbreakable whole to be memorized.

• Phonemic awareness instruction works better when it is integrated with phonics instruction, not as standalone units that precede phonics instruction by months or years. On the other hand, some like Kilpatrick argue that introducing letters too early can interfere with phonemic awareness development.

• PAFB solves the tension between these two positions by teaching phonemic awareness first followed by teaching the mapping of letters to sounds within the same lesson. That way, both skills are developed in parallel while allowing the tutors to notice deficits in phonemic awareness.

• The higher tiers of the intervention model are designed to address these deficits, and tutors are instructed to only use sounds and tokens in deletion, substitution, and reversal tasks. Without the use of letters in these tasks, students are forced to rely entirely on their phonemic awareness skills.

• Reading and spelling are two sides of the same process, each reinforcing the other. Spelling is more difficult because it demands exact recall of letter sequences.

  • PAFB treats spelling as a core component of literacy. Each word and sentence reading lesson is followed by a lesson where students must correctly write the words or sentences from the tutor's dictation.

• Word-level spelling develops along a predictable path. Early attempts are phonetic but imprecise. Over time, students learn to spell words more accurately by internalizing common letter patterns and morphological endings.

  • PAFB follows this trajectory deliberately. Students begin with simple one-syllable words and progress toward irregular spellings and affixed forms once the foundations are secure.

• Sentence-level reading and writing introduce new challenges. Students must coordinate multiple word forms, apply grammar, and maintain fluency.

  • PAFB begins sentence work only after students can reliably decode and spell the words involved. Sentences are used to reinforce word reading in context and to teach syntax and grammar through reading and taking dictation.

The Science of Reading: Comprehension🔗

• The Simple View of Reading defines reading comprehension as the product of two components: decoding and linguistic comprehension. If either is weak, comprehension will suffer.

  • PAFB ensures decoding is fully mastered before comprehension is emphasized. Real texts are only introduced after the student can reliably read and spell all the words involved.
  • PAFB tackles linguistic comprehension by annotating texts with footnotes and briefings, with tutors available to cover any gap.

• Most comprehension problems in struggling readers are caused by poor word reading, not poor reasoning.

  • PAFB follows this principle by focusing early instruction on decoding, spelling, and syntax. Comprehension tasks are introduced only after students can fluently access the text.

• Vocabulary and background knowledge are critical predictors of comprehension. Readers must know the meaning of the words and the context behind the ideas they are reading.

  • PAFB includes a wide range of text types to develop both vocabulary and domain knowledge. Students are exposed to literature, nonfiction, poetry, essays and more, curated and leveled to support gradual knowledge growth.

• Skilled readers construct a mental model of the text by integrating ideas across sentences. This process depends on decoding, vocabulary, working memory, and attention.

  • PAFB supports this process through carefully leveled texts. As students progress, the texts become longer and more conceptually demanding, giving them practice with sustained reading and inference-making.

• Good readers monitor their comprehension as they read and notice when something does not make sense.

  • PAFB trains students to say “I don’t know” when confused. Tutors are instructed to stop, clarify, and re-read as needed, helping students develop this metacognitive skill.

• Reciprocal teaching is a well-researched method for improving comprehension, especially in struggling readers. It involves the tutor modeling what skilled readers do when they encounter confusion by pretending not to understand and asking the student to explain or clarify.

  • PAFB uses this approach by having tutors ask questions aloud during reading, even when they know the answer. This strategy is used only when the tutor suspects the student may be confused, and is faded over time as the student becomes more proficient.

The Science of Writing🔗

• Reading and writing reinforce each other. Reading builds the vocabulary, syntax, and background knowledge that writing draws on, while writing strengthens understanding of how texts are structured, and exercises word and sentence reading and spelling skills.

  • PAFB treats reading and writing as interdependent skills. Each level of text reading courses is followed by a writing course that introduces students to increasingly complex composition tasks.

• Writing is not a natural skill. It must be taught explicitly, in a structured progression, beginning with sentence construction and extending toward paragraphs and full texts.

  • PAFB and Trane are built on small and repeatable exercises that are easily verifiable by a tutor. Writing production tasks do not fit this mold. They cannot be repeated many times without boring students, take much longer than reading and spelling tasks, and have open-ended answers that would require more training and effort from the tutor.
  • This problem is solved by turning writing production tasks into reading tasks. Instead of asking students to solve writing tasks, PAFB's writing track shows students how expert writers tackle the problems by having them read the problem and the solution.
  • The strategy makes the exercises repeatable, shorter, and removes their open-ended nature. Writing production tasks must be undertaken outside the program. PAFB supports production by exposing students to the same type of tasks and letting tutors use the writing questions as a starting point for their own production tasks.

• Writing instruction is most effective when directly tied to reading content. Disconnected prompts do not build transferable skills. The content of the curriculum drives the rigor of the writing activity.

  • Whenever possible, PAFB links each writing task to the texts from the previous writing comprehension course. The model answers reuse the vocabulary and ideas from those texts, helping students internalize structure and content together.
  • As the texts become more complex, so do the writing tasks associated with them.

• Students must learn to write complete sentences before they can write meaningful paragraphs or essays.

  • PAFB begins with structured sentence tasks that train students to express relationships clearly and precisely. Students are shown how to use because–but–so, appositives, and sentence combining to add depth and control to their writing.

• Writing tasks at the paragraph and text level must be addressed after sentence-level writing. The two most important phases of the writing process are planning and revising.

  • PAFB teaches planning by showing examples of structured outlines. These outlines make the internal structure of writing visible and teach students how to organize ideas before writing begins.
  • Revising exercises show what well-revised writing looks like. Tutors are encouraged to ask students how the response could be improved, giving them an opportunity to reflect on clarity, logic, and structure without having to generate a full rewrite.

• Grammar and mechanics should be taught in the context of actual writing, not through isolated drills.

  • PAFB exposes students to correct usage by having them read and study well-constructed sentences. Each task models how grammar supports clarity and meaning in real communication.

Expected Outcomes🔗

Another advantage of using Trane in Pictures Are For Babies is that its mastery learning model allows us to make stronger assertions about the expected outcomes of students that complete the program than programs that lack its features. After all, a student can only be shown exercises from the very last courses in the curriculum if they have shown mastery of all previous courses. Before proceeding, it is important to mention that while the symbol, word, and sentence courses are complete, the reading comprehension and writing courses are still being developed and only the courses spanning early grades have been released. However, the missing courses (spanning from those early grades to graduate and professional material) are not expected to change the design and software of the program in any way. At the current pace of development, the full curriculum is expected to be completed and released in one to two years from the time of this writing (mid 2026). Therefore, all the assertions in this section are made about a complete version of the program. With this out of the way, we can say that students that complete the program must possess:

• Knowledge of every letter and common symbol in the English language, including their names, sounds, and common usages.
• Fluent and accurate reading of most words in the English language, including irregular spellings, foreign loan words, proper nouns, and morphologically complex forms. A sight vocabulary of this size is present in the best adult readers and is necessary for all other literacy skills.
• Fluent and accurate spelling of that same sight vocabulary.
• Fluent and accurate reading and spelling of those words in connected text.
• Ability to comprehend a wide variety of texts, including literature, nonfiction, poetry, essays, plays, scientific articles, journalistic writing, and more. This level of comprehension is not enabled by cookie cutter comprehension strategies, but by excellent word-level reading and the background knowledge required to understand their content and conventions, in line with the Simple View of Reading.
• Ability to comprehend such texts at the undergraduate, graduate, and professional level. These texts require that the reader integrate information across multiple dense and often conflicting sources, evaluate the strength of evidence and the credibility of arguments, infer meaning from indirect or embedded information, recognize the author's purpose, tone, and stylistic choices, and apply prior knowledge to interpret novel or abstract ideas. These skills are what national and international assessments define as needed to be included in the top reading proficiency levels.
• A strong mental model of how to write at the sentence, paragraph, and multi-paragraph level. This model contains the knowledge of how to use conjunctions and transitions to chain sentences together, how to perform revision of single sentences, and how to plan and revise single paragraph and multi-paragraph texts, and more.
• Those students that complete writing production tasks outside the software with the aid of their tutor and the writing exercises will have ensured the mental model has been transferred to actual writing production.

It is still an open question to what extent students that do not complete writing production tasks outside the software are able to transfer the mental model developed through the writing courses to their writing production. Research on the reading-writing relationship shows that reading comprehension and writing quality are moderately correlated (r ≈ .44), with skilled readers consistently outperforming weaker readers on writing tasks. However, the reading measured in these studies is normal independent reading, not reading of extensively worked examples showing the mechanics of the writing process delivered through a deliberate practice engine. Therefore, it is possible that the correlation between reading and writing in graduates of Pictures Are For Babies is much higher than the correlation found in the general population.

Estimated Effect Size🔗

While Pictures Are For Babies claims to be an innovative and effective literacy intervention, none of its innovation lies in inventing or developing new methods of literacy instruction. Instead, it is a carefully designed and engineered synthesis of the best practices across the fields of cognitive science, expert performance, optimal learning, and the science of reading and writing. Therefore, while it is impossible to provide a direct estimate of the effect size of Pictures Are For Babies without conducting a large-scale randomized controlled trial, we can estimate the expected effect size based on the known effect sizes of the individual components of the program. These effect sizes are shown in the list below and sources are provided in the Annotated Bibliography section to avoid cluttering the text.

• The most famous and one of the most controversial of these studies was published by Benjamin Bloom on the effects of mastery learning and one-to-one tutoring. Bloom found that students that received both interventions scored two standard deviations higher than students that received traditional instruction. Bloom's study proposes the famous problem that bears his name: finding a scalable solution that approximates the effects of one-to-one tutoring and mastery learning.

  • Attempts to replicate Bloom's results have never produced the same effect size, with recent meta-analyses of tutoring programs showing pooled effect sizes of 0.42 across all studies and 0.16 to 0.22 for large-scale programs using standardized tests. However, the enormous labor costs of properly implementing one-to-one tutoring and mastery learning make it difficult to know whether the lower effect sizes reflect the ceiling of these interventions or the inevitable decay of implementation fidelity at scale. Pictures Are For Babies combines tutoring with automated mastery learning and deliberate practice in a single system and is uniquely positioned to capture the full effect of these interventions at scale.

• Mastery learning alone was found to have an effect size of 1.0 by Bloom. A meta-analysis found the overall effect size to be 0.59. Given that the mastery learning component of Pictures Are For Babies is implemented in a more rigorous and automated manner than in those studies, we can expect the effect size of the mastery learning component of Pictures Are For Babies to be at least as large.

• Retrieval practice has an overall effect size of 0.50, with feedback increasing the effect size to 0.73. Every lesson in Pictures Are For Babies is a form of retrieval practice, and the intervention protocol ensures that students receive immediate feedback every time they make a mistake. Therefore, we can expect the effect size of the retrieval practice component of Pictures Are For Babies to be close to that of the studies that included feedback.

• Interleaving shows an overall effect size of 0.42, with larger effects for more complex material. Trane interleaves exercises of varying difficulty and type to ensure students receive a healthy mix of practice, review, and challenge. Given the complexity of the literacy skills being taught, we can expect the effect size of the interleaving component of Pictures Are For Babies to be at least as large as the overall effect size found in the meta-analysis.

• A meta-analysis of spaced repetition practice found an overall effect size of 0.54. Another one found an overall effect size of 0.74 when combined with retrieval practice. Trane implements spaced repetition based on a stateless power-law forgetting model inspired by the state-of-the-art in the field. Therefore, we can expect the effect size of the spaced repetition component of Pictures Are For Babies to be closer to the second meta-analysis that combined it with retrieval practice.

• David Kilpatrick has categorized reading interventions into three tiers of increasing effectiveness. The most effective interventions are those that offer explicit instruction in phonics, develop phonemic awareness to proficiency, and present opportunities for reading in connected text. He identifies the second component as the secret ingredient that separates these interventions from less effective ones. Kilpatrick reports that these interventions result in a gain of 12-25 standard score points in nationally normed reading tests, with the majority falling in the 14-17 range, which is roughly an effect size of 1.0. Pictures Are For Babies implements all three components and is therefore suited to capture the full effect of these interventions. These programs also integrate spelling instruction, as does Pictures Are For Babies, so the effect of integration reading and spelling is included in this estimate.

• A study evaluating Core Knowledge Language Arts (CKLA), a program that builds reading comprehension through a knowledge-rich curriculum, found an effect size of 0.47 on reading across grades 3-6. Pictures Are For Babies implements a similar approach to building background knowledge but goes further by capping the curriculum at the graduate and professional level. Therefore, we can expect at least a similar effect size for the knowledge-building component of Pictures Are For Babies.

• A meta-analysis of explicit writing instruction found that teaching explicit writing strategies produced an effect size of 1.02. A study on using the sentence strategies outlined in The Writing Revolution, the specific writing strategies methodology used in Pictures Are For Babies, found an effect size of 0.84 in 9th graders. This second study has not been peer-reviewed, but it shows effect sizes within the range of the meta-analysis. Given that the writing courses in Pictures Are For Babies are based on teaching the same strategies, we can expect a similar effect size for the writing component of the program, provided that the tutor ensures that students complete the writing production tasks that are assigned outside the flow of the deliberate practice engine.

Below is a table summarizing the expected effect sizes of the individual components of Pictures Are For Babies:

Component	Effect Size	Source
One-to-one tutoring + mastery learning	0.42 (at scale) to 2.0 (Bloom, narrow)	Bloom (1984); Kraft et al. (2026)
Retrieval practice with feedback	0.73	Rowland (2014)
Interleaving	0.42	Brunmair & Richter (2019)
Spaced repetition with retrieval practice	0.74	Latimier et al. (2021)
Advanced PA + phonics + connected text	~1.0	Kilpatrick (2015)
Knowledge-building comprehension	0.47	Grissmer et al. (2023)
Explicit writing strategy instruction	1.02	Graham et al. (2012)

These components are not independent, and an estimate of the actual effect size of an intervention that combines them cannot be derived by a mere sum. Otherwise, the effect size would be over 4.0 even with the most conservative estimates for mastery learning and one-on-one tutoring, which is clearly a ridiculous number. However, integrating these components into a single, coherent whole creates the conditions for synergistic effects. Some of these synergistic effects are already known, such as the integration of phonics and phonemic awareness, reading and spelling, and comprehension and writing. Given the breadth of research synthesized by Pictures Are For Babies, the integrated system may exhibit synergistic effects beyond those documented for individual combinations of instructional components.

Despite the impossibility of providing a precise estimate of the effect size of Pictures Are For Babies, listing the expected effect sizes of its components provides a strong indication that absent an unfaithful implementation of the research, the program should be able to produce an effect size at least as large as one of its major components. Given several independent components have an effect size close or above 1.0, we can reasonably estimate that Pictures Are For Babies will produce an effect size of at least 1.0. Whether the synergistic effects of combining these components will produce an effect size close or above the 2.0 found by Bloom in his original study is an open question that can only be answered through a properly designed and executed randomized controlled trial.

A complementary way to estimate the expected effect size uses the expected outcomes described in the previous section and compares them with the highest levels of literacy measured in the National Assessment of Educational Progress (NAEP). In this assessment, students are classified into four levels of proficiency: Below Basic, Basic, Proficient, and Advanced, with only 5% of 12th graders reaching the Advanced level in 2024. Given that Pictures Are For Babies guarantees that completion and mastery are equivalent, and that the scope of the curriculum goes beyond the 12th grade, we can expect that a significant proportion of a randomly selected group of students given access to the program and working through it in a one-to-one tutoring setting will score at the Advanced level at the end of their K-12 education.

We can estimate the effect size implied by various proportions of Pictures Are For Babies graduates reaching that threshold. Assuming that reading scores are approximately normally distributed and that the treatment and general-population distributions have equal variance, the cutoff for NAEP Advanced corresponds approximately to the 95th percentile of the overall distribution, which sits at roughly 1.645 standard deviations above the general mean. The resulting effect size d for a given proportion p of PAFB graduates at or above Advanced can be calculated using the formula:

d = 1.645 + Φ⁻¹(p)

where Φ⁻¹ is the inverse cumulative distribution function of the standard normal distribution.

The table below shows the implied effect size for various proportions of Pictures Are For Babies graduates reaching the NAEP Advanced level, compared to the 5% observed in the general population. These calculations assume that the intervention primarily shifts the mean of the achievement distribution while leaving its variance approximately unchanged.

% of PAFB graduates at NAEP Advanced	Implied effect size (d)
20	0.80
30	1.12
40	1.39
50	1.65
60	1.90
64	2.00
70	2.17
80	2.49
90	2.93
95	3.29
99	3.97

Even conservative outcomes imply effects close to the largest components in the logic model. If half of Pictures Are For Babies graduates reach NAEP Advanced, the implied effect size is 1.65, which exceeds every individual component in the summary table except Bloom's original finding. If approximately two-thirds of graduates reach Advanced, the implied effect size matches Bloom's 2.0 effect size. Larger values produce effect sizes that are difficult to imagine in a real-world setting, but that cannot be dismissed as impossible given that the test only measures reading and that the reading courses cover material well beyond the 12th grade. Whether Pictures Are For Babies can achieve these outcomes and remove variability through its use of a deliberate practice engine and one-to-one tutoring is the central empirical question to be addressed by a future randomized controlled trial.

Planned Efforts for Further Evaluation🔗

While Pictures Are For Babies has been designed to be a research-backed and effective literacy intervention and its claims are deeply rooted in well-researched and widely accepted principles, the depth of integration, the scope of the curriculum, and the ambition of its goals make further evaluation a critical next step. To that end, the author is engaged in an ongoing effort to find a research partner to conduct a multi-year longitudinal randomized controlled trial to evaluate the program's effectiveness in a real-world setting that has or can apply for the necessary grants. Full access to the program will be provided at no cost to such a research partner and to all schools and students involved in the study. Due to the large efforts needed to study the full efficacy of the program, Pictures Are For Babies does not plan to conduct or coordinate any shorter-term studies that would qualify for ESSA tiers 3 or 2.

The deep integration of Kilpatrick's orthographic mapping model in Pictures Are For Babies opens the possibility to also study the effect additional phonemic tasks have on the overall effect size of the program. A study with two intervention groups, one using only the first tier of the word protocol and the other using all three tiers, could provide valuable insight into Kilpatrick's categorization of reading interventions and settle the debate on the extent to which phonemic proficiency drives fluent reading acquisition.

While such a research partner is recruited, the author is collecting the full data from students whose parents agree to share. Since Pictures Are For Babies runs locally and does not store any personal data, such sharing does not require any data-sharing agreements or compliance infrastructure. The author is also looking to conduct school and individual free pilots that can provide feedback and iron out any issues in the program before a large-scale study is conducted. However, based on the extent and fidelity to which the research has been followed, the author is confident that any issues solved by user feedback are minor and will mostly affect the manual and other training materials, not the core of the program's design or software.

Annotated Bibliography🔗

Books🔗

The pedagogy outlined in Pictures Are For Babies is largely derived from the sources below. These are some of the most widely respected and research-backed works in their fields. Their authors are recognized leaders in reading science, writing instruction, and expert learning. These texts were not used for surface-level validation. They shaped the design and sequencing of every component in the program from the very beginning. Books were preferred over individual journal articles because they provide a more comprehensive and integrated view of the research that is difficult for an outside researcher to construct from individual articles.

• Ericsson, K. A., Hoffman, R. R., Kozbelt, A., & Williams, A. M. (Eds.). (2018). The Cambridge Handbook of Expertise and Expert Performance (2nd ed.). Cambridge University Press.

  • The definitive reference on deliberate practice and expert performance. Informed Trane's design as a deliberate practice engine and the program's focus on mastery through effortful, structured training with immediate feedback.

• Kilpatrick, D. A. (2015). Essentials of Assessing, Preventing, and Overcoming Reading Difficulties. Wiley.

  • Synthesizes research on orthographic mapping, advanced phonemic awareness, and intervention effectiveness. Informed the tiered word-level intervention protocol and the categorization of reading interventions by standard score gains used in the Estimated Effect Size section.

• Snowling, M. J., Hulme, C., & Nation, K. (Eds.). (2022). The Science of Reading: A Handbook (2nd ed.). Wiley-Blackwell.

  • Comprehensive reference on the science of reading. Informed the program's adherence to research on phonological awareness, decoding, orthographic mapping, comprehension, and reading difficulties.

• Hochman, J. C., & Wexler, N. (2024). The Writing Revolution 2.0: A Guide to Advancing Thinking Through Writing in All Subjects and Grades. Jossey-Bass.

  • The basis for the program's writing track. Informed the worked-examples approach, sentence-level to paragraph-level progression, and integration of writing with reading content.

Scientific Studies🔗

• 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), 4-9, 39. https://www.aft.org/ae/summer2020/moats

  • Seminal article arguing that reading failure is preventable in all but a small percentage of children with serious learning disorders and that classroom instruction is the critical factor.

• Tipton, E., & Patton-Terry, N. (2026). We need an "engineering of reading": Why the "science of reading" may not be enough. Mind, Brain, and Education. https://doi.org/10.1111/mbe.70053

  • Argues that the science of reading requires an accompanying "engineering of reading" to translate research into effective classroom practice.

• 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.

  • Found one-to-one tutoring plus mastery learning produced an effect size of 2.0 on narrow tests. Mastery learning alone produced ~1.0. Originated the "2 sigma problem."

• Kulik, J. A., Kulik, C.-L. C., & Bangert-Drowns, R. L. (1990). Effectiveness of mastery learning programs: A meta-analysis. Review of Educational Research, 60(2), 265-299.

  • Meta-analysis of 108 controlled studies. Overall effect size of 0.52. LFM (Bloom's method) = 0.59. Follow-up exams (8 weeks later) = 0.71.

• Kraft, M. A., Schueler, B. E., & Falken, G. (2026). What impacts should we expect from tutoring at scale? Exploring meta-analytic generalizability. Review of Educational Research. https://doi.org/10.3102/00346543261446660

  • Meta-analysis of 263 RCTs. Pooled effect size of 0.42 for all tutoring studies, dropping to 0.16-0.22 for large-scale programs (400+ students) using standardized tests.

• Rowland, C. A. (2014). The effect of testing versus restudy on retention: A meta-analytic review of the testing effect. Psychological Bulletin, 140(6), 1432-1463. https://doi.org/10.1037/a0037559

  • Meta-analysis of 61 studies on retrieval practice. Overall d = 0.50. Retrieval practice with feedback produced larger effects (d = 0.73) than without feedback (d = 0.39).

• Brunmair, M., & Richter, T. (2019). Similarity matters: A meta-analysis of interleaved learning and its moderators. Psychological Bulletin, 145(11), 1029-1052. https://doi.org/10.1037/bul0000209

  • Meta-analysis of 59 studies (238 effect sizes). Overall interleaving effect g = 0.42. Stronger effects for complex material, longer retention intervals, and categories with high between-category similarity and within-category dissimilarity.

• Grissmer, D., White, T., Buddin, R., Berends, M., Willingham, D., DeCoster, J., Duran, C., Hulleman, C., Murrah, W., & Evans, T. (2023). A kindergarten lottery evaluation of Core Knowledge charter schools: Should building general knowledge have a central role in educational and social science research and policy? (EdWorkingPaper No. 23-755). Annenberg Institute at Brown University. https://doi.org/10.26300/nsbq-hb21

  • Lottery-based RCT of Core Knowledge curriculum in 9 charter schools (N = 2,310). Long-term treatment-on-treated effect on reading/ELA = 0.47 across grades 3-6.

• Graham, S., McKeown, D., Kiuhara, S., & Harris, K. R. (2012). A meta-analysis of writing instruction for students in the elementary grades. Journal of Educational Psychology, 104(4), 879-896. https://doi.org/10.1037/a0029185

  • Meta-analysis of 115 studies on writing interventions. Strategy instruction ES = 1.02; SRSD = 1.17; text structure instruction = 0.59.

• Vroom, T.-A. (2021). The effects of sentence-level writing instruction based on the Hochman Method on the writing of ninth-grade students [Doctoral dissertation, St. John's University]. St. John's Scholar. https://scholar.stjohns.edu/theses_dissertations/213/

  • Quasi-experimental study of The Writing Revolution's sentence-level strategies. d = 0.84 for writing quality after 4 months of instruction with 9th graders.

• Latimier, A., Peyre, H., & Ramus, F. (2021). A meta-analytic review of the benefit of spacing out retrieval practice episodes on retention. Educational Psychology Review, 33(3), 959-987. https://doi.org/10.1007/s10648-020-09572-8

  • Meta-analysis of 29 studies on spaced retrieval practice. Overall g = 0.74 after correcting for publication bias. No significant difference between expanding and uniform spacing schedules.

• Mawson, R. D., & Kang, S. H. K. (2025). The distributed practice effect on classroom learning: A meta-analytic review of applied research. Behavioral Sciences, 15(6), 771. https://doi.org/10.3390/bs15060771

  • Meta-analysis of 22 classroom-based studies (31 effect sizes). Overall d = 0.54 for distributed vs. massed practice with curriculum-relevant materials.

• Kim, Y.-S. G., Wolters, A., & Lee, J. W. (2023). Reading and writing relations are not uniform: They differ by the linguistic grain size, developmental phase, and measurement. Review of Educational Research. https://doi.org/10.3102/00346543231178830

  • Meta-analysis of 395 studies (N = 120,669). Found reading comprehension and written composition are moderately correlated (r ≈ .44).

Literacy Assessments🔗

The data on the state of literacy instruction cited in the problem statement is drawn from the following three assessments. Together they cover school-age students (NAEP, PISA) and adults (PIAAC), providing a complete picture of literacy outcomes across grades and countries.

• National Center for Education Statistics. (2024). NAEP Report Card: Reading — National trends and student skills, grade 12. U.S. Department of Education. https://www.nationsreportcard.gov/reports/reading/2024/g12/national-trends/

  • 5% of 12th graders scored at the NAEP Advanced level in 2024. 32% scored below Basic.

• OECD. (2023). PISA 2022 Results (Volume I): The state of learning and equity in education. OECD Publishing. https://ilsa-gateway.org/studies/results/1612

  • 7% of 15-year-olds scored at the highest proficiency levels (5 or 6) across OECD countries. 26% scored below Level 2.

• OECD. (2024). Survey of Adults Skills 2023: United States — Country note. OECD Publishing. https://www.oecd.org/en/publications/survey-of-adults-skills-2023-country-notes_ab4f6b8c-en/united-states_427d6aac-en.html

  • 13% of US adults scored at the highest levels (4 or 5) in literacy. 28% scored at Level 1 or below.