What is educational therapy?

Educational therapy takes place within specialized, one-on-one sessions tailored to the specific academic needs of the individual student. Various methods are used to optimize the student's strengths so that they may be used to fortify weaknesses. Because many students need concepts broken into manageable chunks, the educational therapist approaches sessions with methodical organization and structure. Educational therapy also teaches compensatory strategies that allow the student to experience success in the classroom and beyond, and can improve underlying processing skills so that learning new material is easier.

How is educational therapy different from tutoring?

Educational therapy addresses underlying learning skills such as visual and auditory processing, attention and focus, and memory. Academic skills such as reading and writing are remediated by teaching the strategies necessary to read, write, calculate and problem solve. An educational therapist has completed graduate and often post-graduate training in working with all types of learners. Tutoring, by contrast, focuses on re-teaching academic subject matter and helping with homework, and a tutor can be anyone proficient in the subject at hand.

Who needs educational therapy?

When a child continues to struggle despite parent-teacher interventions or private academic tutoring, they should be evaluated for underlying learning deficits by a licensed neuropsychologist or clinical psychologist who specializes in learning issues. Children do not outgrow learning problems, and it is important to begin intervention as early as possible to mitigate and potentially prevent self-esteem problems and unnecessary daily stress.

How long are educational therapy sessions and how often should they happen?

Sessions are one-on-one and run 50 minutes, with the educational therapist and student working intensively together. A minimum of two sessions per week is recommended, though each student's plan is discussed on a case-by-case basis during the intake session.

What learning differences do you work with?

Our network of independent professionals brings extensive experience with ADHD, autism spectrum disorders, dyslexia, dyscalculia, dysgraphia, language-based learning differences, processing weaknesses and anxiety, helping students identify and develop their strengths.

Which areas of Los Angeles do you serve, and can sessions be virtual?

We serve Calabasas, Santa Monica, Beverly Hills, Sherman Oaks, Encino, the South Bay, Los Angeles and surrounding areas. Sessions are available in-office, in-home, in-school, in the community, in college settings, or virtually.

How are students matched with a specialist?

We assess each student's needs, learning preferences and personality, then refer them to the right specialist - a qualified educational therapist, learning specialist or executive functioning coach - rather than assigning whoever is available.

Progress Monitoring Decision Rules: Where They Come From

Andrea Chernin

4 days ago

21 min read

Updated: 3 days ago

You have eight weeks of data on a student, an aimline on the graph, and a decision to make. The last four points sit under the line. Every handout you have ever been given says the same thing: change something.

Progress monitoring decision rules are among the most frequently applied numbers in intervention work, and among the least documented. Three points below the line. Four points below the line. A trend line flatter than the goal line. The median of the last three. These are repeated in training materials, built into vendor dashboards and written into district protocols, and the question of where any of them came from is almost never asked.

This post asks it. Not whether progress monitoring is worth doing — it is, and the measurement side of that case is strong. The question here is narrower and more awkward: who proposed these specific rules, on what evidence, and what has happened on the three occasions somebody has actually tested them.

The short answer is that the rules were proposed as practical conventions by researchers who said so at the time, that the federal materials which carry them today cite nothing at all, and that where the rules have been simulated against known growth, the common four-point rule produces a wrong answer more often than a right one.

The short version

The four-point rule is attributed to a 1990 practitioner-journal article, not to a validation study. The trace runs through Hintze and colleagues, who name it directly.
The National Center on Intensive Intervention's current decision-rules handout gives three methods and not one citation. It rates one method "highly accurate" and another "potentially not as accurate" with no reference attached to either judgement.
The two federal handouts you are most likely to be handed disagree about the data bar. NCII says three weeks and six data points for a four-point analysis. The IRIS Center says a minimum of six weeks.
A 2013 systematic review read 102 documents and found the practices rest on expert opinion, with what the authors called very limited psychometric or empirical support.
Simulated against known growth, a three-point rule was correct less than half the time in most conditions, and never better than chance when the student's true growth sat anywhere near the goal.
A four-point rule applied to typical probe error produced false-positive rates between about 0.49 and 0.61 — that is, a student who really was growing as expected was declared non-responsive on most occasions.
The federal practice guide that recommends monthly progress monitoring rates its own evidence Low, and says in print that the panel recommended the practice despite the lack of evidence supporting it.
The association that credentials educational therapists requires you to measure progress at regular intervals and names no interval, no rule and no threshold.

The rule is not a finding. It is a convention, and the people who proposed it said so. What has happened since is that the convention stopped being described as one.

What the three rules are, in the federal handout

The cleanest current statement of the rules is a two-page handout from the National Center on Intensive Intervention, produced at the American Institutes for Research under an award from the Office of Special Education Programs. The handout carries a 2024 copyright and describes three options.

Four-point analysis. "If at least three weeks of instruction have passed and six data points have been collected, examine the four most recent data points." All four above the goal line, increase the goal. All four below, "consider adapting the intervention." A mix of above and below, keep collecting.

Trend line analysis. "If at least four weeks of instruction have passed and eight data points have been collected, determine the trend of current performance and compare it with the goal line." Steeper than the goal line, raise the goal. Flatter, "consider adapting the intervention and assess fidelity."

Median of the last three data points. Same data bar as the trend line — four weeks, eight points — then compare the median of the three most recent scores with the goal line.

Two things in that handout are worth noticing before anything else.

The first is the data bar. A four-point analysis is not four data points. It is six collected and the last four examined, which is a materially higher bar than most people apply.

The second is what the handout says about accuracy. Four-point analysis is listed as "Potentially not as accurate as Trend Line Analysis." Trend line analysis is listed, flatly, as "Highly accurate." Median of the last three "may be more accurate with more weeks of data collection."

Those are comparative empirical claims. The handout attaches no reference to any of them. There is no reference list. The only citation on the document is the federal award number.

A two-page federal handout is allowed to be a two-page federal handout. The problem is that for most practitioners it is the whole published basis of a decision they make every six weeks.

Which rule you get depends on which document you opened

The vendor systems disagree with each other about this, and that disagreement is covered in detail in our comparison of the four progress monitoring systems practitioners actually encounter, which quotes each manual's own rule. Three points for Acadience. Four for DIBELS 8, except for oral reading fluency, where three will do. A different aimline origin again for easyCBM.

That post owns the question of what each system specifies. This one is about something the manuals do not address: where the number came from in the first place.

But the federal materials disagree too, and that is less widely known. The IRIS Center's data-based individualization module describes the four-point method on its analysis page as "an easy method for examining the relationship between the four most recent data points and the goal line" — the same method, and a looser threshold, since it asks whether "most" of the four points fall below the line rather than all four.

IRIS also sets a different data bar. It tells the practitioner to collect data "for a minimum of six weeks (or six data points if data are collected more than once per week)." NCII says three weeks and six points. The same method, from two centers funded by the same department, with the waiting period doubled.

Neither page gives an attribution for the rule itself.

Where the four-point rule actually comes from

The paper trail runs through a 2018 simulation study in the Journal of Psychoeducational Assessment by John Hintze, Craig Wells, Amanda Marcotte and Benjamin Solomon, which is one of the very few papers to test the four-point rule directly. In setting up the problem, the authors name its source: "the frequently referenced 4-point rule suggested by L. S. Fuchs et al. (1990)."

The article sits behind a paywall at the journal, and a freely readable copy is hosted by a Reading League state chapter if you want to check the tables yourself.

The reference list resolves that to Fuchs, L. S., Fuchs, D., and Hamlett, C. L. (1990), "Curriculum-based measurement: A standardized long-term goal approach to monitoring student progress," in Academic Therapy.

Two observations about that citation, and they are the point of this section rather than a footnote to it.

It is a practitioner article, not a validation study. The title describes an approach to monitoring student progress. It is a methods proposal in a practitioner journal, which is a perfectly respectable thing to be. It is not an experiment in which four points was compared against three or five and found superior.

The citation itself is hard to pin down. Hintze and colleagues give the pages as 615 to 632. Douglas Fuchs's own curriculum vitae lists the article at pages 263 to 276.

That is the same page range the CV gives for a different 1990 Academic Therapy article by an overlapping set of authors, which points to a transcription slip somewhere in the record rather than to two articles. We could not resolve which pagination is correct from any freely readable source.

That is a small thing. It is also a fair measure of how thin the documentation is behind a number that gets applied to individual students several times a term.

The wider lineage is clearer and better documented. The whole apparatus — graph, goal line, data-driven adjustment — comes out of the work Stanley Deno and colleagues did at the University of Minnesota in the late 1970s and 1980s, published as data-based program modification in Deno and Mirkin (1977).

Fry and colleagues set out that history in an open-access 2024 review. What that lineage establishes is the practice of monitoring and adjusting. It does not establish a threshold.

What the evidence actually supports, and what it does not

There is a strong-sounding number that gets cited whenever the decision rules are questioned, and it is worth handling carefully because it is real and it does not say what it is used to say.

A 1986 meta-analysis by Lynn Fuchs and Douglas Fuchs in Exceptional Children estimated a mean effect size of 0.70 for systematic formative evaluation.

Van Norman and Christ, summarising it in their 2016 study of data point rules, note the four components the original authors tied to that effect: ongoing data collection, graphic display, explicit decision rules for when to change a program, and behaviourally based interventions.

The original authors wrote that "when teachers were required to employ data utilization rules, effect sizes were higher than when data were evaluated by teacher judgment."

Read that carefully. It is evidence that having an explicit rule beat having none. It is not evidence for any particular rule.

Van Norman and Christ make exactly that distinction: the results "do not confer support for CBM specifically; rather, they confer support for progress monitoring generally."

That is the gap this whole post sits in. The case for structured progress monitoring is good. The case for four points rather than three, or five, or a trend line, is a separate case, and it has been examined far less often.

You can believe the meta-analysis and still have no basis for the threshold. Those are two different claims and the field routinely spends the first to pay for the second.

What happened when somebody went looking, in 2013

In February 2013, the Journal of School Psychology published a systematic review by Scott Ardoin, Theodore Christ, Laura Morena, Damien Cormier and David Klingbeil, titled "A systematic review and summarization of the recommendations and research surrounding Curriculum-Based Measurement of oral reading fluency (CBM-R) decision rules." The ERIC record is EJ1001681 and the article is doi 10.1016/j.jsp.2012.09.004.

The authors identified 171 documents and evaluated the 102 that met their inclusion criteria. Their finding, in the abstract's own terms, is that most decision-making practices rest on expert opinion, with very limited psychometric or empirical support for such practices.

The review also did something useful and structural: it sorted the rules into two families. Data point rules, which compare individual observations against the goal line. Trend line rules, which compare the slope of the student's data against the slope of the goal line. Everything in the NCII handout is one or the other.

The review is indexed at the University of Minnesota and summarised at Semantic Scholar for anyone without journal access.

That review is twelve years old. It is still the most comprehensive statement of the problem, and the handouts in circulation today do not cite it.

When the three-point rule was simulated, it was worse than chance

Three years later, two of that review's authors tested one of the rules.

Ethan Van Norman and Theodore Christ published "Curriculum-Based Measurement of Reading: Accuracy of Recommendations From Three-Point Decision Rules" in School Psychology Review in 2016. The full text is open at ERIC under record EJ1141160.

The design is a probability model rather than a classroom trial. They specified a goal line of 1.50 words read correctly per minute per week, six levels of the student's true rate of improvement from 0 to 3 words per minute per week, and two levels of dataset quality — a residual of 5 and of 10 words per minute.

Data collection duration was varied across thirteen levels, from 3 to 15 weeks. That gives 158 unique conditions, and for each one they derived the probability that a three-point rule returns the correct recommendation.

Their own summary: "recommendations are often inaccurate" and "the probability of a correct recommendation is below chance in most situations."

The published table is more useful than the summary, because it shows where the failure sits.

At six weeks of weekly data, with a true growth rate of 0.84 or 2.16 words per minute per week, the probability of a correct decision was 0.35 when residual was 5 — and 0.22 when residual was 10.
When the student's true growth sat close to the goal line, at 1.25 or 1.75 words per minute per week, the probability of a correct decision never exceeded chance at any duration, at either level of data quality, out to fifteen weeks.
The probability of a correct decision passed fifty percent only when the gap between the student's true growth and the goal exceeded 0.25 words per minute per week.

That last line is the one to keep. The rule works tolerably when the intervention is either obviously working or obviously not. In the region where you actually need help — the student whose growth is near but not quite at the goal — it is a coin toss at best.

The authors' own practice recommendation is blunt: practitioners "should abstain from making decisions until collecting 12–14 weeks of data, especially if only one CBM-R is collected per week," and short durations are "rarely advisable, particularly not" fewer than twelve to fourteen weeks.

They also report which way the error runs. When the student's true growth was below the goal, "the inaccurate decision to maintain instruction was more likely than the correct decision to change instruction for the first 5 to 15 weeks." In their words, "the incorrect decision is often to maintain the presumably ineffective intervention."

One disclosure the paper makes itself, and it belongs in any honest summary: the second author declares equity and royalty interests in a company commercialising a formative assessment system, reviewed and managed under his university's conflict-of-interest policy.

The study's conclusion runs against data point rules and toward trend lines, which is also the direction a vendor dashboard would favour. The disclosure is in the paper. We mention it because leaving it out would be worse.

When the four-point rule was simulated, it misfired on the majority of occasions

The 2016 study tested three points. Hintze, Wells, Marcotte and Solomon tested four, five, six and seven, in Journal of Psychoeducational Assessment in 2018.

They simulated 20,000 progress monitoring datasets modelling growth of two words per minute per week across fifteen consecutive weeks, then asked how often each rule falsely declared a student non-responsive when that student was in fact growing at the expected rate. They varied the standard error of estimate from 2 to 18 words per minute, which spans the range reported in the published reliability literature for these probes.

For the four-point rule, the false-positive rates they report are:

Standard error of 2: 0.36
Standard error of 4: 0.49
Standard error of 6: 0.55
Standard error of 8: 0.57
Standard error of 10: 0.59
Standard error of 12 to 18: 0.60 to 0.61

The three-point rule, at the same error levels, runs from 0.74 to 0.93.

The authors locate typical probe quality in the range of 6 to 10, citing the reliability estimates in the literature. At that level, applying a four-point rule means that on more than half of the occasions it fires, it fires on a student who was growing as expected.

Their conclusion: "commonly used 3- and 4-point decision-making rules lead to unrealistically high levels of false positives which result in prematurely altering an otherwise effective intervention." Requiring five or six consecutive points "reduces errors in decision making by a minimum of 50%."

There is a cost to that, and the authors state it. The true-positive rate falls as the rule gets stricter. At a standard error of 10, the four-point rule catches a genuine non-responder about 93 percent of the time; a six-point rule catches about 66 percent; a seven-point rule, about 52 percent. You are trading false alarms for missed cases, and there is no setting of the dial that gives you both.

Notice that the two studies point in opposite directions about which error you are making.

Van Norman and Christ modelled a student whose growth was genuinely short of the goal, and found the likely mistake was to leave an ineffective intervention running. Hintze and colleagues modelled a student who was growing as expected, and found the likely mistake was to tear up an intervention that was working.

Both are true, of their own conditions. The honest reading is that a short run of points carries very little information either way.

The federal practice guide rates the practice Low and says so

The underlying recommendation to monitor at all has a published evidence rating, and it is worth reading rather than assuming.

Recommendation 4 of the 2008 Institute of Education Sciences practice guide Assisting Students Struggling with Reading: Response to Intervention and Multi-Tier Intervention for Reading in the Primary Grades — Gersten and colleagues, NCEE 2009-4045, available from IES — tells schools to monitor tier 2 students at least once a month, at least eight times a school year.

The guide's own evidence appendix rates that recommendation's level of evidence as Low, and explains why in terms that are hard to improve on:

Of the 11 randomized controlled trials and quasi-experimental design studies that evaluated effects of tier 2 interventions and that met WWC standards or that met WWC standards with reservations, only 3 reported using mastery checks or progress monitoring in instructional decision-making.

None of the studies demonstrate that progress monitoring is essential in tier 2 instruction. However, in the opinion of the panel, awareness of tier 2 student progress is essential for understanding whether tier 2 is helping the students and whether modifications are needed.

The appendix goes further: in the three studies that did include a progress monitoring component, "neither of the studies evaluated progress monitoring as an independent variable. Thus, no inferences can be drawn about its effectiveness based on the research reviewed." And then, in the panel's own voice: "Despite the lack of evidence supporting use of progress monitoring, the panel decided to recommend this practice."

The guide also disposes of a frequency assumption that most practitioners carry. "Many tier 2 intervention programs available recommend weekly assessment of student progress, but there is little research evidence that weekly assessment is more beneficial than monthly." A free one-page summary of the recommendation is hosted by the Nebraska Department of Education for anyone who wants the text without the full guide.

The single study the guide cites for a positive effect of progress monitoring on teachers' instructional decisions is Fuchs, Deno and Mirkin (1984) — the same research group whose later article is the source of the four-point rule.

This is a guide that recommends the practice, states the evidence is thin, and labels its recommendation accordingly. That is the federal record behaving well. Our post on how evidence ratings for reading programs are actually assigned covers the rating machinery in more detail, and our piece on whether educational therapy itself is evidence-based applies the same reading to our own field.

How often the practitioner literature discusses this at all

There is one more published finding worth having, because it explains why none of the above is common knowledge.

In 2024, Erica Fry, Jessica Toste, Beth Feuer and Christine Espin published an open-access systematic review in Journal of Learning Disabilities asking how much of the curriculum-based measurement content written for practitioners is actually about decision-making. The full text is free at PubMed Central; the article is doi 10.1177/00222194231215031.

They searched twelve journals. Twenty-nine articles across five of those journals met the criteria — 106,784 words in total. Of that content, approximately 18 percent was devoted to decision-making. A companion review of professional development materials by Espin and colleagues had found 12 to 14 percent.

Within those 29 articles:

35 percent (10 articles) gave specific guidance on when to make an instructional change.
38 percent (11 articles) gave non-specific guidance on the same question.
17 percent (5 articles) referenced when to make a change without saying how.

So roughly a third of the practitioner literature on this topic states a rule, and the review's own illustration of a typical statement is a three-point rule given without derivation.

The authors' conclusion is a training conclusion: there should be more attention to decision-making in curriculum-based measurement materials. Read alongside the 2013 review, it suggests something slightly different as well. The rules are under-documented and under-discussed, and those two facts keep each other in place.

What the profession that credentials you says about it

Nothing. And that is worth saying precisely rather than rhetorically.

The Association of Educational Therapists Code of Ethics, Revision 9 of 2025, addresses the obligation directly in Section Two, I.N. Members are to "establish baseline data about the skills and needs of new clients and maintain accurate data for the purpose of decision making and consultation," and under I.N.1, to "establish ways to measure and determine client progress towards those goals at regular intervals."

That is an unambiguous requirement to measure progress, at intervals, and to make decisions from the data. It names no interval. It names no rule. It sets no threshold at which a decision is owed.

Section Two, I.B is also live here, and it is the clause that bears on the choice of rule: members are to "select and use appropriate assessment instruments, recognizing their limitations with respect to reliability, validity, and bias."

A four-point rule applied to a probe with a large standard error is a reliability limitation, exactly in the sense that clause means. The AET fact sheet sets out the scope of the role more generally.

So the structure is the same one this site has found in group size, in session dosage and in session documentation: the programs specify a number, in incompatible units, and the body that credentials you specifies the duty and leaves the number to you.

When the manual gives you a number and your ethics code gives you a duty, the number is not the answer to the duty. Writing down which rule you used, and why, is.

When several organisations share a vocabulary

This is the recurring finding across the assessment cluster on this site, and it is worth collecting in one place, because each post found it independently.

Progress monitoring tools, compared found four systems with four different decision rules and two different ways of drawing the aimline.
Standard error of measurement and confidence intervals found that the width of the band on a score report is a setting the publisher chose, not a property of the child.
Oral reading fluency norms found the most-used norm table in the field publishes no error term at all.
Outdated test norms found publishers disagree about when a sample stops describing the population.
Grade equivalent scores found a score type the publishers themselves warn against using.
Reading program evidence ratings found clearinghouses rating the same program differently because they ask different questions.
What "the science of reading" actually names found an umbrella phrase no organisation adjudicates.
Math fact fluency found the field's central claim about automaticity in the federal record exactly once, labelled as panel opinion.

Decision rules belong on that list. The pattern is not that anybody is being careless. It is that a convention proposed for one purpose, in one measurement system, gets lifted out and applied as though it were a general fact.

A sequence, not a decision

What follows is not advice about which rule to use. It is the sequence that keeps your reasoning legible to you, to a parent and to a school.

Write down which rule you are using, and where you got it. Name the document. The rule in your vendor dashboard, the rule in the NCII handout and the rule in the IRIS module are not the same rule.
Write down the data bar you are holding yourself to — how many points collected before you will look, not just how many you will look at. The NCII handout's six-then-four is stricter than most practice.
Know the error on your own measure before you trust a short run of points. Where the publisher reports a standard error for the score, read it. Where it does not, say so in your notes. Our post on score bands and measurement error covers where to find these figures and which publishers print them.
Treat a four-point run as a prompt to look, not a verdict. On the published simulations, four points below the line at typical probe quality is slightly more likely to be noise than signal.
Prefer a longer run, or a trend line, when the decision is consequential. Both of the simulation studies point that way, and the 2016 study names twelve to fourteen weeks for weekly data.
Separate the fidelity question from the data question. NCII's trend line rule says to "assess fidelity" alongside adapting the intervention, and a flat line can mean the intervention was not delivered as written.
Record the decision and the reason, not just the change. This matters most when the next practitioner, or a school team, inherits the graph. Our notes on session notes and progress reports cover the documentation side.

What is out of scope here

Several things, deliberately.

Nothing here is legal advice. This post does not address what any statute, regulation or education plan requires of anyone. Progress monitoring data are used in eligibility and placement decisions in school settings, those uses are governed by federal and state law, and the questions they raise belong with an attorney and with the student's district — not with a blog post and not with a decision rule.

This is not guidance on how to interpret a particular student's graph. The published simulations tell you about the long-run behaviour of a rule across many datasets. They do not tell you what is happening with the student in front of you. That judgement is yours, and it should be informed by the measure's error, the fidelity of delivery and everything else you know about the case.

This is not a claim that any named program or system works or does not work. The studies discussed here are about decision rules, not about interventions.

It is also not a complete survey of the literature. Three empirical papers and two reviews are the core of the published record on this specific question, which is itself part of the finding, but there is other work — on trend-line estimation methods, on visual analysis, on monitoring frequency — that a reader going deeper should expect to find.

The rules on this page operate on a general outcome measure and answer a question about a whole program. The threshold inside a program's own lesson sequence is a different number in a different measurement paradigm, and where the mastery criteria in the manuals come from sets out what each publisher specifies, in whose unit, and what happened when somebody tested a threshold instead of inheriting it.

Where the published record runs out

The honest boundaries of everything above.

The original proposal is not freely readable. The 1990 Academic Therapy article that Hintze and colleagues name as the source of the four-point rule is not available online in any version we could reach, and the two records we could check disagree about its pagination.

We have described what its title and its citing literature say about it. We have not read it. If the derivation of the number is in that article, we cannot tell you so.

We did not re-verify the 1986 meta-analysis directly. The effect size of 0.70 and the quoted sentence about data utilization rules are taken from Van Norman and Christ's summary of Fuchs and Fuchs (1986), which is a peer-reviewed source quoting a page number. The original article is behind a paywall.

Both simulation studies are simulations. Neither followed real students through real decisions. They model the statistical behaviour of a rule given assumptions about growth and error. Those assumptions are stated in both papers and are drawn from published reliability estimates, which is the right way to do it, but a probability model is not a trial.

The error ranges may not describe your measure. Hintze and colleagues locate typical probe quality at a standard error of 6 to 10 words per minute, citing reading fluency studies from the 2000s. If you are monitoring something other than oral reading fluency, or using a measure whose technical manual reports different figures, the published false-positive rates do not transfer directly.

We found no decision-rule guidance from any educational therapy or learning specialist credentialing body. Not from AET beyond the ethics clauses quoted above. That is a negative finding as of October 2026, from the publicly readable documents, and it is not the same as a statement that no such guidance exists anywhere.

Nobody has published a validation of the five-point or six-point rule either. Hintze and colleagues show those thresholds behave better in simulation. That is a modelling result, not a field trial, and the stricter rules carry the missed-case cost described above.

Frequently asked questions

What is the four-point rule in progress monitoring?

It is a decision rule that asks you to compare the four most recent progress monitoring scores with the goal line. All four below, consider changing the intervention. All four above, raise the goal. A mix, keep collecting. The NCII handout adds a data bar most practitioners skip: at least three weeks of instruction and six data points collected before you examine the last four.

Who came up with the four-point rule?

Hintze, Wells, Marcotte and Solomon, in their 2018 simulation study, attribute it to Fuchs, Fuchs and Hamlett (1990) in Academic Therapy. That is a practitioner-journal methods article rather than a validation study. The broader framework it sits in comes from Deno and Mirkin's data-based program modification work in 1977.

Is the four-point rule accurate?

On the only published simulation that tested it directly, not very. Hintze and colleagues report false-positive rates of about 0.55 to 0.59 at probe error levels typical of oral reading fluency measures. More than half the time the rule fired, it fired on a student who was in fact growing as expected. Requiring five or six consecutive points cuts decision errors by at least half.

How many weeks of progress monitoring data do I need before deciding?

More than most handouts imply. Van Norman and Christ recommend abstaining from decisions until twelve to fourteen weeks of data have been collected if you are measuring once a week. The NCII handout's own bars are three weeks and six points for a four-point analysis, four weeks and eight points for a trend line. The IRIS module says a minimum of six weeks.

Should I use a trend line instead of counting data points?

Both published simulations lean that way. Van Norman and Christ recommend trend line rules "until evidence for improved decision rules emerges," while noting trend lines are not perfect and are themselves affected by data quality. NCII lists trend line analysis as its most accurate option but attaches no citation to that judgement. Trend lines also need a longer run of data and a tool that will calculate the slope for you.

Does the Association of Educational Therapists specify a decision rule?

No. Its Code of Ethics, Revision 9 of 2025, requires members to establish baseline data, maintain accurate data for decision making, and establish ways to measure client progress at regular intervals. It does not name an interval, a rule or a threshold. The choice, and the responsibility for documenting it, sits with the practitioner.

Join the Illuminate educator network

We write for educational therapists, learning specialists and executive function coaches, and we try to say plainly where the published record stops. If a number is a convention, we would rather call it one.

If that is useful to you, join the Illuminate educator list and the next one will come to you.