top of page
Search

Reading Program Group Size: What the Manuals Permit

Writer: Andrea Chernin
Andrea Chernin
Oct 17
25 min read

Updated: 4 days ago

Reading Program Group Size: What the Manuals Permit


The short version

If you are deciding whether to run a reading intervention with one student or three, the published record splits in a way nobody warns you about. Program publishers answer the reading program group size question precisely and in writing — Wilson names a number, Barton names a maximum, Corrective Reading names two different numbers in the same federal report.


The professional bodies that credential you answer it not at all. And the number AET actually counts, 1,500 direct service hours, is recorded on a form that never asks who else was in the room.


Three findings are worth having before your next intake call.


The manuals specify group size, and no two specify it in the same unit. Across seventeen programs documented by the National Center on Intensive Intervention, the group size field is filled in as a single number, a word, a range, a ratio, a per-adult ceiling that changes by grade band, a whole class, and a ceiling-or-one-to-one. There is no common scale.


Adding a student does not save you time, and two publishers say so in print. Wilson's published optimal schedule for small-group delivery asks for more weekly minutes than its optimal schedule for one-to-one. Barton puts a figure on it and then says plainly that you will not save any time.


The evidence does not support the intuition that smaller is automatically better. One federal guide says reduce the group to intensify the dose. Another counts group size 116 times and recommends nothing about it. A meta-analysis of extensive interventions after grade 3 found no significant differences in outcomes related to instructional group size — a between-studies finding with a boundary you need to read before you use it.


Nothing here is legal advice. Caseload caps, IEP service minutes and anything a district or state requires are law and contract, not program specification, and they are out of scope below.


Every program publishes a group size, and no two publish it in the same unit

Start with the thing that is actually written down. NCII publishes Academic Intervention Taxonomy Briefs for seventeen programs, and every one of them carries a Group size row in its program-information table. Here is what those rows actually say, verbatim.


The reading programs


  • Leveled Literacy Intervention — "3 students," at 30 minutes, five days a week, for 14–18 weeks (brief).

  • Lexia Core5 — "Individual" (brief).

  • Lexia PowerUp — "Individual; small group" (brief).

  • Read Naturally Live — "Up to six students per adult in Grades 1–5; up to eight students per adult in Grades 6–adult" (brief).

  • Seeing Stars — "Up to 5:1 in homogeneous groups based on diagnostic data. Whole class, developmentally K–2" (brief).

  • Visualizing and Verbalizing — "5:1 in homogeneous groups based on diagnostic data" and, as a second bullet, "Whole class, developmentally" (brief).

  • Sound Partners — "1:1 (research also supports its use with dyads)" (brief).

  • PACT — "Whole class of 15 or fewer students" (brief).


The mathematics programs


  • Number Rockets, Pirate Math and Hot Math — "1–4."

  • Fraction Face-off! — "2–4."

  • Fusion (Whole Number Foundations Level 1) and ROOTS (Level K) — "2–5 students."

  • Strategic Instruction in Multiplication, both the partial-products and the standard-algorithm versions — "1–4 students."

  • Spring Math — "No more than six students per group or individual one on one. Classwide intervention is available when a majority of the students are below targets" (brief).


Read that list again as a set of units rather than as a set of numbers. One program gives you a single integer. Two give you a word. Six give you a closed range. Two give you a ratio. One gives you a per-adult ceiling that changes at grade six.


One gives you a class with a cap. One gives you a ceiling, an alternative, and a third mode that depends on how many students in the class are below target. One tells you its format is one-to-one and then tells you the research also covers pairs.


The field has a shared vocabulary for group size and no shared scale. "Small group" in one manual is three students; in another it is six; in a third it is anything under fifteen.

That matters the moment you try to compare two programs, which is the same problem that runs through programs compared on four factual axes: the axes exist, but the publishers do not fill them in the same units.


One publisher defines "whole class" in a parenthesis

EPS Learning's own S.P.I.R.E. Quick Start Lessons contains the most quietly useful sentence in this whole area:


"S.P.I.R.E. is intended to be implemented with small groups of students, but to save time, teachers may wish to teach the Quick Start Lessons to a whole class (more than 5 students) at once, before placing students into smaller groups for the full implementation of the S.P.I.R.E. program."

A whole class, to this publisher, is more than five students. That is the only place in the document where either term is given a number, and the number arrives inside a parenthesis in a sentence about saving time on the preliminary lessons. The publisher never states a ceiling for its own small groups at all.


If you have ever sat in a meeting where two people used "small group" to mean four and fourteen and neither noticed, this is why.


The ceiling is often a statement about you, not about the students

Here is the detail that changes how you read every one of those rows. Several of the published ceilings are conditioned on the practitioner's skill rather than on the students' needs.


Wilson's own implementation guidance for the Wilson Reading System says instruction "may occur one-on-one or in small, homogeneous groups (ideally up to four students)," and then adds that "a group size of up to six learners is appropriate in some settings (e.g., with a highly skilled WRS teacher)."


The same sentence appears on Wilson's own multi-tier comparison chart, which gives WRS's group size as "1:1 or small homogeneous groups (ideally up to 4 students). Group size up to 6 appropriate in some settings (i.e., with a highly skilled WRS teacher)" (chart).


Read that as a practitioner rather than as an administrator. The ceiling moves from four to six not because a particular group of students can tolerate it, but because of who is standing in front of them.


Read Naturally's row works the same way in a different direction — "up to six students per adult" in grades 1–5 and "up to eight students per adult" in grades 6 and up — where the variable is the students' age band and the unit is explicitly per adult, not per group.


Barton is blunter still. Its published FAQ states "The maximum group size allowed with the Barton System is three-on-one," and then rules out an entire setting: "The Barton System is not designed to be classroom curriculum." The same page says that if you are a regular education teacher, "do not purchase the Barton System."


A program can therefore permit a group and forbid a classroom, which are different statements, and the difference is exactly the one a practitioner moving from a school to a private practice has to understand.


Adding a student does not save you time, and two publishers say so in print

This is the finding with the most direct consequence for anyone running a private practice, and it is the opposite of why most practices form groups.


Wilson's own numbers go up, not down


Wilson publishes separate intensity schedules for one-to-one and small-group delivery of the same program. For one-to-one, the published options are 45 minutes four to five times a week, 60 minutes three to five times a week, or 75–90 minutes two to five times a week, with an optimal intensity of "90 minutes/3 times a week or 75 minutes/4–5 times a week."


For small-group delivery, the options are 45–60 minutes four to five times a week or 75–90 minutes two to five times a week, with an optimal intensity of "90 minutes, 5 times per week."


Put the optimal figures side by side. One-to-one at its published optimum is 270 minutes a week at one end of the range, or 300 to 375 minutes at the other. Small group at its published optimum is 450 minutes a week. The curriculum is identical; the publisher's recommended weekly contact time is substantially higher when students are added.


Barton puts a number on the conversion and then says the quiet part


Barton's FAQ goes further than any other publisher found this run. It states that "the intensity in 3-on-1 tutoring is not the same as 1-on-1 tutoring," and then supplies the conversion:


"To get the same intensity, you must tutor a group of 3 students for 6 hours each week."

And then, in case the implication has not landed:


"Please be aware you will not save any time tutoring in a group."

The same page adds that "the more students in a group, the longer it will take to complete the Barton System," and raises a constraint that no research document mentions: "you cannot present new material if any one of the students is absent."


That last sentence is worth sitting with. A group of three does not merely divide your attention; it couples three families' attendance into one instructional sequence. One absence stalls the lesson for everybody.


If you have built a practice schedule on the assumption that a group of three is three times the revenue for the same hour, both of these publishers are telling you in writing that the hour is not the same hour.


Delivery format is a fidelity question, in other words, and not only a scheduling one — the same conclusion the record forces about session frequency and duration, where published dosage turns out to be a specification rather than a minimum effective dose.


NCII's own arithmetic runs backwards from the intuition

Now the part that genuinely surprised me.


Every NCII taxonomy brief computes, for its program, how many opportunities each student gets to respond and receive corrective feedback. The method is stated plainly:


"Dosage is the number of opportunities a student has to respond or practice and receive corrective feedback. Dosage may be impacted by the size of the instructional group, the number of minutes each session lasts, the number of student-teacher interactions built into lessons, and the number of sessions provided per week."

So NCII has done the arithmetic the intuition rests on. If group size drives opportunities to respond, the smallest groups should produce the highest counts. They do not.


The numbers, as published


  • Sound Partners, at a group size of one, in 20–30 minutes: 73 opportunities.

  • Pirate Math, assuming four students, in 30 minutes: 106.88.

  • Fraction Face-off!, assuming four, in 30 minutes: 97.14.

  • Hot Math, assuming four, in 20–30 minutes: 93.75.

  • Fusion, assuming five students, in 30 minutes: 92.13.

  • Number Rockets, assuming four, in 40 minutes: 88.23.

  • Spring Math, assuming no more than six, in a 15-minute session: 76.

  • ROOTS, assuming two students, in 20 minutes: 58.

  • Leveled Literacy Intervention, assuming three students, in 30 minutes: 53.

  • Strategic Instruction in Multiplication, assuming three students, in 30 minutes: 44.

  • Visualizing and Verbalizing and Seeing Stars, assuming five students, in sessions of one to four hours a day: 30.

  • Lexia PowerUp, assuming one student online and two offline: 29.

  • Lexia Core5, same assumption: 26.

  • PACT, assuming fifteen students in a 45-minute class period: nine, per class period.


Lay that against group size and the relationship falls apart. A five-student group running Fusion gets each student an estimated 92 opportunities in half an hour. A one-student session of Lexia Core5 gets 26. A five-student group in Visualizing and Verbalizing gets 30 — in a session that may run four hours. A three-student LLI group gets 53, fewer than the five-student Fusion group.


Group size does not predict opportunities to respond. Lesson structure does. Within these seventeen programs, the variable that moves the per-student response count is how many scripted teacher-student exchanges the lesson contains, not how many students are sharing them.

What the decimal places tell you


Those figures are reported to two decimal places — 106.88, 97.14, 93.75, 92.13, 88.23 — and the decimals are the most informative part of them. A count of real responses in a real session cannot have a fractional part. These are arithmetic properties of the written lesson: a count of scripted exchanges, divided across an assumed group. They describe what the manual contains, not what happened in your room on Tuesday.


That is a limitation and a gift at once. It means you cannot cite these numbers as observed dosage. It also means you can compute the same quantity yourself for any program you own, because the input is the lesson script, which is sitting on your shelf.


Group size is not a dimension of intervention intensity — it is an input to one

Here is the structural point that explains why the professional conversation about group size is so muddled.


Every NCII brief opens with the same set of taxonomy questions, and group size appears in exactly one of them:


"Will the group size, duration, structure, and frequency provide sufficient opportunities for students to respond and receive corrective feedback (dosage)?"

Group size has no independent standing in the Taxonomy of Intervention Intensity. It is one of four inputs to the dosage dimension, alongside duration, structure and frequency. There is no "group size" dimension to rate, no threshold to clear, and no rubric entry for it.


So when a practitioner asks "is a group of three intensive enough?", the framework the field uses to answer intensity questions cannot answer that question as asked. It can only answer "do the group size, the minutes, the lesson structure and the weekly frequency together produce enough opportunities to respond?" — which is a different question with four knobs, and which is why the arithmetic above comes out the way it does.


This is the same structural move as the one that makes dosage itself two different definitions in NCII's own documents, and it has the same practical upshot: you cannot optimise one input in isolation and claim you have intensified the intervention.


The two federal guides answer the question differently, and one of them does not answer it at all

The 2009 guide: three to four students, and reduce the group to intensify


The WWC practice guide Assisting Students Struggling with Reading: Response to Intervention and Multi-Tier Intervention in the Primary Grades is where the familiar number comes from. Its Recommendation 3 states that for tier 2, "the small groups should be homogenous and range from three to four students," with sessions three to five times a week.


The guide's own companion presentation on systematic teaching in tiers 2 and 3 is more explicit about the mechanism, and it states the intuition directly:


"This dosage can be intensified by reducing the size of student groups and increasing the frequency and duration of instruction."

And on group size specifically: "Reducing the number of students in learning groups provides them more opportunities to practice new skills and respond to what they are learning. Tier 2 provides instruction to small groups of three to four students, while Tier 3 offers even more intensity through daily one-on-one tutoring."


That is the clearest statement of the three-to-four convention in any federal document, and it is worth knowing two things about it: it is a 2009 document about primary grades, and the mechanism it asserts is the one NCII's own arithmetic, published later, does not bear out across programs.


The 2022 guide: group size counted 116 times, recommended zero times


The current federal guide for older students, Providing Reading Interventions for Students in Grades 4–9 (WWC 2022007), handles group size very differently, and the pattern only shows up if you count.


The phrase "group size" appears 116 times in the guide. Of those:


  • 112 are the row label "Group size:" inside Appendix C, the appendix that records the characteristics of every study in the evidence base.

  • three are inside the title of a single cited study.

  • one is a background sentence: "Group size typically varies, with smaller groups in grades 4 and 5 than in elective reading courses in middle and high schools."

  • None is in the text of a recommendation.


The guide makes four recommendations — multisyllabic word reading, fluency building, comprehension building, and stretch texts — and not one of them specifies, suggests or bounds a group size.


What the 112 entries actually record


The distribution is the part nobody writes about. Across those 112 study entries, the most frequent value recorded is not a small group at all:


  • Whole-class instruction, in one phrasing or another: 33 entries.

  • Individual: 14.

  • 10–15 students: 12.

  • The remainder spread across 2, 2–3, 2–4, 3, 3–4, 3–6, 3–8, 4–5, 4–6, 4–7, 7–16, 9–21, 15, and the bare word "Small."


Whole-class delivery is recorded more than twice as often as individual delivery in the evidence base of the guide practitioners cite for adolescents. And the word "one-on-one" appears exactly once in the entire guide — not in a recommendation, not in an appendix, but in the glossary, inside the definition of the word intervention:


"Interventions are used to provide focused, often more intense, instruction to students who are falling behind in core instruction, usually provided one-on-one or in small groups."

The phrase "one-to-one" appears zero times.


So the federal guide for grades 4–9 records group size as a property of 112 studies, mentions one-to-one delivery once in a glossary, and recommends nothing. The age-band post here notes that the guide specifies neither dosage nor group size; the counts above are what that silence looks like when you measure it.


The one study designed to answer the question is in the evidence base three times

This is the sharpest thing in the 2022 guide and it is three lines in a reference list.


Among the studies the guide reviewed is Vaughn, Wanzek, Wexler, Barth, Cirino, Fletcher, Romain, Denton, Roberts and Francis (2010), "The relative effects of group size on reading progress of older students with reading difficulties", Reading and Writing 23(8), 931–956. The title is the question this article is about.


The guide enters that study three separate times, with three different contrast labels:


1. the large-group arm against business as usual, 2. the small-group arm against business as usual, and 3. the small-group arm against the large-group arm — the head-to-head comparison.


And then, in Appendix C, where the evidence for each recommendation is tabulated, only one of those three entries appears: the large-group arm. It is recorded as a randomised controlled trial of 420 grade 7–8 students with reading difficulties across six urban schools in the southwestern United States, 45–50-minute sessions five days a week for one year, group size 10–15 students, rated "meets WWC standards with reservations," against business-as-usual English language arts.


Its effect sizes as the guide reports them: passage reading fluency (silent) −0.02, word and pseudoword reading 0.07, reading comprehension 0.03.


The small-group arm and the head-to-head contrast are listed among the reviewed studies and contribute to no recommendation's evidence table.


I could not retrieve the study's own report of the head-to-head result this run. The ERIC record resolved to the wrong document and the journal is behind access controls, so this article does not state what that comparison found.


If the answer is load-bearing for a decision you are about to make, the citation above is the one to chase through a library. The result you want is in the paper rather than in the guide.


What the research syntheses found, and what you may not conclude from them

One meta-analysis does speak to this directly and can be reported.


Wanzek, Vaughn, Scammacca, Metz, Murray, Roberts and Danielson, "Extensive Reading Interventions for Students With Reading Difficulties After Grade 3", in Review of Educational Research, synthesised nineteen studies of extensive interventions for students in grades 4 through 12, ten of which met criteria for meta-analysis and reported 22 distinct treatment/comparison differences.


Mean effect sizes ranged from 0.10 to 0.16 for comprehension, word reading, word reading fluency, reading fluency and spelling outcomes. On group size, the finding is a null:


"No significant differences in student outcomes were noted among studies related to instructional group size, relative number of hours of intervention, or grade level of intervention."

Read the boundary in the same breath as the finding, because it is doing a great deal of work. This is a between-studies moderator analysis across ten studies, not a randomised within-study comparison of group sizes holding everything else constant. Studies differ in program, duration, grade, interventionist and sample as well as in group size.


A null moderator result in that design means the synthesis could not detect an association between group size and outcome across this body of work.


It does not establish that group size makes no difference to a given student in a given program, and you may not tell a family that the research shows group work is just as good. Nor can it be used in the other direction, as evidence in favour of shrinking a group.


What it does license is narrower and still useful: the published research record does not currently supply a group-size threshold you can point to, and anyone who tells you it does is overreading it.


There is a second relevant synthesis — Hall and Burns, "Meta-analysis of targeted small-group reading interventions," in the Journal of School Psychology — and I could not read it this run.


PubMed served a browser challenge and the publisher's page refuses automated access. Its findings are therefore not reported here at all, rather than guessed at. That is the honest state of the record from where I am sitting, and it is the kind of gap that a university library login closes in two minutes.


The broader problem is the one running through whether educational therapy is evidence-based and through how five systems rate reading programs: the evidence attaches to components and conditions, and delivery format is one of the conditions least often isolated.


The WWC contradicts itself about one program inside one report

A small but instructive case, and it belongs to the same family as the two-document contradictions that keep turning up in this field.


The WWC's intervention report on Corrective Reading describes the program on its first page like this: it "can be implemented in small groups of four to five students or in a whole-class format," and is "intended to be taught in 45-minute lessons four to five times a week."


The report's own Teaching section says this:


"The program's 45-minute lessons are designed for groups of up to 20 students up to five times a week."

Four to five students, or up to twenty, in one federal intervention report about one program. Neither statement is flagged as a correction of the other.


If you are comparing programs on group size and taking your figures from clearinghouse summaries rather than from publishers, this is the kind of thing you will inherit without noticing — the same hazard that makes a publisher's own documentation the better primary source whenever one exists.


The word "group" is not in AET's governing documents in a service sense

Now turn from the programs to the profession, because this is where the record goes quiet.


I word-counted every document the Association of Educational Therapists publishes that governs what the work is and how it is credentialed. The results:


The Code of Ethics


The AET Code of Ethics, Revision #9, September 2025 runs to roughly 3,000 words. The word "group" appears in it exactly once, in the preamble, in the phrase "protection of the individuals and groups with whom the professional works" — a statement about who is protected, not about how instruction is delivered. The document never uses "small group," never uses "one-to-one," never uses "caseload," and never uses "class."


The Fact Sheet


The Educational Therapy Fact Sheet runs to about 400 words and contains the word "group" zero times. Its description of the role says the educational therapist will "provide individualized intensive intervention for clients who present with a wide range of learning disabilities and learning differences."


The supervision packet and the ET/P application


The AET Supervision Packet, revision 4, 2026 uses the word "group" once, and it is not about clients: it appears in a list of supervision discussion topics, as "joining or forming study groups," which is about the supervisee's own professional development.


More to the point, both the supervision packet and the 2026 ET/P application enumerate what counts as a qualifying direct service hour, and the first two items on both lists specify a delivery format:


  • "Individual sessions of remediation or teaching with students who are the focus of supervision. (Note: this does NOT include time spent on review of records.)"

  • "Individual sessions of assessment. (Note: this does NOT include time spent on report preparation.)"


The remaining items are classroom observations on behalf of a client, parent conferences and IEP attendance, consultations with allied professionals, qualifying fieldwork or practicum hours, and direct service completed as part of supervision. No item on either list names a group session. The only word describing a delivery format anywhere in the enumeration is "individual."


Across five AET documents — a code of ethics, a fact sheet, a supervision packet, an application and an hours form — the only delivery format named for direct service is the individual session. The word "group" never appears in that sense at all.

I want to be careful about what that does and does not establish, because the temptation to over-read it is strong. It does not say that group hours are disqualified. AET has published no clause saying so.


What the published record shows is an enumeration that names only individual sessions and a set of governing documents that never address group delivery — and the gap between "not named" and "not permitted" is exactly the kind of gap you should not fill with a confident sentence of your own.


If your hours are mostly group hours and your ET/P application is in front of you, ask AET directly and get the answer in writing. This is the same shape as the problem in what "supervised hours" actually means across six bodies, where six organisations count in six different units.


The 1,500-hour form never asks who else was in the room

Then there is the form itself, and it is the single cleanest illustration of the whole problem.


To qualify as an ET/P, AET requires documentation of at least 1,500 direct service hours completed within the past ten years, verified by letters. The verification instrument is AET's own Direct Service Hours Verification form. It runs to roughly a hundred words in total, and here is everything it asks for:


  • who is submitting it, and the date

  • the number of hours accumulated

  • the applicant's job title

  • the name of the school, clinic or private practice

  • the year range in which the hours were accumulated

  • one free-text line: "During these hours, I performed the following Educational Therapy related activities:"

  • whether the verifier is a work site administrator, a professional supervisor, a colleague or "other"

  • the verifier's name, title, email, address and phone


There is no field for delivery format. No field for group size. No field for the number of students served. The one quantitative requirement in the profession's own credential ladder is counted in hours of your time, and the form that counts them is indifferent to how many students were in the room while the clock ran.


That is not an oversight you should read as permission, and it is not an oversight you should read as a trap either. It is simply the state of the published record: the programs specify group size to the student, and the profession specifies it not at all.


If you want the wider version of that pattern, it is the same one that runs through what the codes require of session notes — five codes requiring accurate records and none defining what a record is — and through the consent machinery five codes require and barely define.


Four factual axes for deciding a group size

This is a decision sequence, not a decision. Nothing below tells you what group size to run, because the published record does not support anyone telling you that, and because the choice turns on facts about your students that no document knows.


Axis one: what the manual permits, in its own units


Open the implementation guidance for the program you actually own and find the group size statement. Write down the number and the unit: is it per group, per adult, a ratio, a ceiling, or a class? Note whether the ceiling is conditioned on anything — the practitioner's skill level, the students' grade band, the homogeneity of the group.


If the publisher names no ceiling, that is a finding to record, not a licence — and check the edition you own is the current one, because instrument and program documentation goes stale faster than practitioners expect.


Axis two: what the manual asks of your schedule at that group size


Check whether the published intensity changes with delivery format, as Wilson's does. If it does, the honest comparison is not "one hour with three students versus one hour with one student." It is the publisher's recommended weekly minutes for a group against the publisher's recommended weekly minutes for an individual. Do that arithmetic before you quote a schedule to a family.


Axis three: what the lesson structure gives each student


Count the scripted teacher-student exchanges in a representative lesson and divide by the group size you are contemplating. That is exactly NCII's method, it is reproducible on your own materials, and as the figures above show it frequently disagrees with the intuition that fewer students means more practice.


A tightly scripted group lesson can out-deliver a loosely structured individual one on this measure, which is also why a program's published evidence rating tells you nothing about the format you are running it in.


Axis four: what grouping does to homogeneity and to the sequence


Several publishers condition their group size on homogeneity — Wilson says "small, homogeneous groups," Lindamood-Bell's briefs say "homogeneous groups based on diagnostic data." A group is only a group if the students belong at the same point in the sequence, which constrains who you can put together far more tightly than the ceiling does. And remember Barton's attendance point: in a mastery-sequenced program, one student's absence can hold the sequence for everyone.


Where this leaves you is a sequence to run at intake and at every data review, which is the same discipline that progress monitoring outside a district demands — and the grouping decision should be revisited when the graph changes, not set once at intake and inherited for a year.


What changes in the room when a second student arrives

A short section on the things the documents do not quantify but practitioners deal with daily, offered as considerations rather than as findings.


  • The sequence couples. In a mastery-based program, the group advances together. Barton states this plainly; most publishers do not, and it is the constraint most likely to surprise someone new to group delivery.

  • Error correction becomes public. A correction delivered to one student is heard by the others. Whether that helps or costs you is not something the published record settles.

  • Diagnostic information thins out. In an individual session every response is attributable. In a group of four, a chorus response tells you much less — which matters directly for anything you intend to write in a progress report, and for how much weight a single data point can carry once you remember the error band around any score.

  • The scheduling problem multiplies. Three families' calendars must agree, and the group is only as reliable as its least reliable attender.

  • The consent and confidentiality picture changes. Students in a group know who else is in it. None of the five codes compared in the consent post addresses group delivery specifically, which means the service agreement is yours to draft carefully.


Where this sits next to what is already published here

One sentence each, so the thread is visible.



Three more that bear on the practice side: how to become an educational therapist covers the hours ladder the form above belongs to, starting a private practice covers the business questions a grouping decision touches, and writing a treatment plan is where a delivery-format decision should be recorded.



Group size is published by every program in a different unit. So is the threshold that decides when a student leaves a lesson: the mastery criteria in reading intervention manuals arrive as percentages, error counts, rates per minute and proportions of occasions, and no authority anywhere converts between them.

Where the published record runs out

The honest inventory, because this is a thinner record than it looks.


The source universe here is small, and it is small because of the subject. The primary sources for this article are seventeen NCII taxonomy briefs, two publishers' implementation pages, two federal practice guides, one WWC intervention report, one published meta-analysis, and five AET documents.


That is fewer distinct domains than most posts here carry. Padding it would have meant citing other people's blog posts about these manuals instead of the manuals, and the manuals are the whole point.


The one study designed to answer the question could not be read. Vaughn et al. (2010) is the only study in either federal guide's reference list whose title is about group size, and its head-to-head result is not reported above because the record resolved to the wrong document and the journal is gated.


One relevant meta-analysis could not be read either. Hall and Burns is named above and its findings are deliberately absent.


Take Flight could not be verified. No publicly readable Scottish Rite specification of Take Flight's permitted group size was retrievable this run, so it is not in the table. It is one of the most widely used dyslexia therapy curricula in the country and its grouping requirements belong in any serious comparison; if you have the curriculum guide, you have the answer and I do not.


The programs with no published ceiling are a real gap. S.P.I.R.E. says "small groups" and defines whole class but never bounds its own small groups. Several NCII rows give a range without saying what happens above it.


AET's silence is a gap in the record, not a ruling. Everything above about AET is a statement of what its documents do and do not contain. Whether group hours count toward the 1,500 is a question only AET can answer, and the answer is not published.


If you are at the stage of assembling that application, the hours ladder and the continuing-education clocks that follow it are the two places this question will surface again.


Nothing here is legal advice. Caseload limits, class-size regulation, what an IEP's service minutes oblige a district to deliver, and what constitutes specially designed instruction under federal law are all questions of statute, regulation and contract rather than of program specification.


If a grouping decision in your setting turns on any of them, that is a conversation for your state education agency and for an attorney, not for a program manual and not for this article.


Common questions

How many students can be in a reading intervention group?


It depends entirely on the program, and the published answers range from one to a whole class. Wilson names four as ideal and six as acceptable "with a highly skilled WRS teacher." Barton caps it at three and says it is not classroom curriculum.


The 2009 federal RTI guide says three to four for tier 2 in the primary grades. NCII's seventeen documented programs give everything from "Individual" to "Whole class of 15 or fewer students." There is no cross-program standard, so the only defensible answer is the one in the manual for the program you are running.


Does a smaller group mean more intensive instruction?


Not automatically, and NCII's own published arithmetic is the reason to be careful. Group size is one of four inputs to the dosage dimension of the Taxonomy of Intervention Intensity, alongside duration, lesson structure and frequency.


Across NCII's briefs, a five-student Fusion group is credited with an estimated 92 opportunities to respond per student in 30 minutes while a five-student Visualizing and Verbalizing session is credited with 30 in a session that may run four hours. Lesson structure moves that number more than group size does.


Does the research say group instruction is as effective as one-to-one?


The research does not support that claim as stated. The closest finding is from Wanzek and colleagues' synthesis of extensive interventions after grade 3, which reported "no significant differences in student outcomes… related to instructional group size" across ten meta-analysed studies in grades 4 through 12.


That is a between-studies moderator result, not a controlled comparison of group sizes, so it means the synthesis could not detect an association in that body of work — not that group and individual delivery are equivalent for a given student in a given program. Do not tell a family the research shows group work is just as good.


Will running a group save me time?


Two publishers say in writing that it will not. Wilson's published optimal schedule is 90 minutes five times a week for small groups against 90 minutes three times a week, or 75 minutes four to five times, for one-to-one.


Barton states that "to get the same intensity, you must tutor a group of 3 students for 6 hours each week" and adds "please be aware you will not save any time tutoring in a group." Those are the publishers' own statements about their own curricula, and they are the figures to reason from.


Do group hours count toward AET's 1,500 direct service hours?


AET has not published an answer. The qualifying-hours lists in both the 2026 ET/P application and the 2026 supervision packet name "individual sessions of remediation or teaching" and "individual sessions of assessment," and no item on either list names a group session.


The Direct Service Hours Verification form never asks how many students were present. "Not named" is not the same as "not permitted," and the gap is not one to fill with an assumption — ask AET directly and keep the reply.


What should I record about group size in my notes?


Record the group size, the program, the published ceiling you are working within, and your reason for the grouping, alongside the usual session data. None of the five professional codes compared here specifies this, so it is a professional judgement rather than a requirement.


The practical argument for recording it is that attribution thins out in a group, so a future reader of your file — including you — needs to know how many students were sharing the responses behind a data point.


If this is the work you do

Illuminate is building a referral network of educational therapists, learning specialists and executive function coaches, and a resource library written for practitioners rather than for parents — posts like this one, sourced to the manuals and the governing documents rather than to other people's summaries of them.


If that is useful to you, join the educator list. You will get new practitioner resources as they are published, and an invitation to the referral network when it opens in your area.

Comments

bottom of page