What does the research say about IEPs for autistic students?

The school plan every eligible child gets, and ways to make it better. Here’s what 8 sources can and can’t tell you.

Research synthesisUpdated 2026-10-038 sources includedProtocol v4.5 (Amdt. v4.6)
COMPASS teacher coaching Promising but contested — Low certainty, tested only by its makers
Having an IEP, and other ways to improve it No verdict possible, or not enough research yet

A care note before you read. This report is about research, not about your legal rights. In the US, an IEP is a legal right under a law called IDEA for children who qualify. Questions about your child’s rights belong with your school team or a parent advocacy center.

Podcast·Two-voice deep dive

Listen to the discussion

0:00
Chapters & transcript

This is an AI-generated discussion of this article, transcribed automatically. The written summary below is the reviewed source — if the two differ, go by the summary.

0:00

What is this discussion about?

Host A Imagine spending, like, months negotiating a legally binding contract to protect your child's future, only to find out that over 70% of it is written in a way that makes it completely unenforceable.

Host B Yeah, that's a genuine nightmare scenario for a lot of families.

Host A Right. And if you're listening to this and you are a parent or an educator who has ever sat in one of those, you know, tiny plastic school chairs for an IEP meeting, that statistic might validate every single frustration you've ever had.

Host B Oh, absolutely. The tiny chairs really set the mood for the whole process.

Host A They really do. So today we are doing a deep dive into the Individualized Education Program, or IEP, specifically for autistic students. But we aren't actually looking at the therapies inside the plan.

Host B Right, which is a big distinction.

Host A Exactly. We are looking at interventions designed to fix the actual process of making the plan itself, like how the sausage is made.

Host B Yes, the administrative mechanics of it.

Host A We're specifically diving into a teacher coaching program called COMPASS and an instruction method called the self-directed IEP, where students actually, well, they lead their own meetings.

Host B Which is such a fascinating concept.

Host A It really is. And we've got a great stack of sources today. We're pulling from an audit document and a findings report that draws on randomized controlled trials, secondary analyses, and some descriptive document studies.

Host B And before we get into the mechanisms of these interventions, I do need to state plainly that this is a summary of research findings.

Host A Always important to clarify.

Host B Right. It is not clinical or legal advice, and absolutely nothing discussed here today is a recommendation for any individual child.

Host A Good to have that out of the way.

1:37

What is an IEP?

Host A So let's establish what we are actually looking at here. For anyone outside the U.S. or maybe just new to this system, the IEP isn't like a specific reading program or a speech therapy module, right?

Host B No, it's not. It's really important to separate the container from what goes inside it.

Host A Okay, the container.

Host B Yeah. So the IEP is a statutory planning process. It's mandated under U.S. law, the Individuals with Disabilities Education Act, or IDEA. Gotcha. It is a written, legally binding plan created by a team, and that team includes the family, teachers, school specialists, everyone.

Host A Right, the whole squad in the tiny chairs.

Host B Exactly. And together, they detail a student's customized yearly goals and the very specific supports the school will provide.

Host A So it's universal in the U.S., but what about internationally?

Host B Other countries definitely have their own versions of this legal container. For instance, there's the Education, Health, and Care Plan, or EHCP, in England. But universally, you can think of the IEP as the blueprint for a child's education.

Host A The blueprint. I really like that analogy. The question we're exploring today isn't whether having a blueprint works. I mean, obviously, you need a plan before you build a house.

Host B That's a given.

Host A The question is whether bringing in an architectural consultant for the teacher or, you know, putting the kid in charge of drawing the blueprint actually leads to a better house.

3:01

What research is there on IEPs, and who ran it?

Host A So how do researchers even go about studying that?

Host B Well, the evidence base we are drawing from is quite specific. And honestly, it's a bit narrow. Really? How so? We didn't find any broad, systematic reviews on this specific topic. Instead, the evidence relies heavily on three randomized controlled trials, or RCTs, for that COMPASS coaching program.

Host A Okay, three RCTs. What age groups?

Host B Those span from young children all the way up to transition-age teenagers. There is also one secondary analysis of teacher outcomes, one RCT for the self-directed IEPs, and some descriptive document analyses that basically evaluate the physical quality of the written IEP documents themselves.

Host A Okay, so randomized control trials are usually the gold standard. But whenever we look at an evidence base on this show, we always have to ask, you know, who is actually doing the research?

Host B Yeah, and that brings up a crucial caveat that we really must keep in mind throughout this whole discussion. All three of the COMPASS RCTs were run by the program's developers. Oh, okay. Furthermore, the consultations themselves were delivered by members of the research team.

Host A Wait, so there's no independent verification?

Host B In the sources we have, no. There are no independent replication studies.

Host A I mean, that's a bit like a chef rating their own restaurant, right? The food might genuinely be fantastic and their five-star rating might be totally honest, but you still want an independent food critic to review it just to be sure.

Host B That is a very fair way to look at it. It doesn't mean the intervention is flawed. It simply tells researchers that the science is in a very early localized stage.

4:38

How well are IEP goals written?

Host A Okay, keeping that early stage context in mind, let's look at the actual quality of these blueprints. Because earlier I mentioned that wild stat about 70% of the contract being unenforceable. Where does that actually come from?

Host B It comes from a 2022 descriptive study by Findley, Ruble, and McGrew, which is detailed in our sources. They analyzed 20 transition age IEPs for autistic students.

Host A Transition age meaning high schoolers, right?

Host B Right. These are teenagers, typically 14 to 18, preparing to leave the school system and enter adulthood. So the stakes are incredibly high here.

Host A Absolutely.

Host B So what did they find in these documents?

Host A The findings were stark.

Host B Only 26.4% of the goals in those documents were written in a way that could actually be measured in behavioral terms.

Host A Wow. Barely a quarter.

Host B Exactly. And even fewer, about 9%, clearly stated the actual method of measurement.

Host A Let's just pause on that for a second because the implications are massive. If only 26% of the goals can be measured, how does the team even know if the child is making any progress at all?

Host B That's the exact problem.

Host A Like a goal like be a better listener isn't exactly something a teacher can just chart on a graph.

Host B No, it's not. When goals lack measurable metrics, progress tracking becomes highly subjective. And that is dangerous for a legal document meant to guarantee specific support.

Host A Yeah, that sounds like a liability.

Host B But the researchers found an even more striking gap in the data.

Host A Okay, lay it on me.

Host B Almost 90% of those IEPs documented that the student had a clear need for social skills support. However, only 22% of the IEPs that documented that need actually incorporated an IEP objective to address it.

Host A Wait, I'm trying to wrap my head around that. 90% of the time, the school and the parents sit around a table and agree, yes, this student needs help with social skills.

Host B Correct.

Host A But then almost 80% of the time, they just, they don't write a goal for it.

Host B That's what the data showed.

Host A I mean, it sounds like they are acknowledging the kitchen is on fire, but not actually writing down a plan to put it out. Why does that happen?

Host B The sources suggest this points to a massive disconnect between identifying a child's needs and legally committing the resources to address them.

Host A Is it a time thing? A budget thing?

Host B Often the IEP process can just devolve into a compliance checklist. Administrators are under immense pressure to complete the paperwork legally and efficiently. Writing a highly specific custom social skills goal requires specialized knowledge, time, and data tracking that an overwhelmed special education teacher simply might not have.

Host A Which perfectly sets up the need for an intervention, right? If the teachers are overwhelmed and the documents have these glaring gaps, we need to train the people writing them.

Host B Exactly.

7:29

What is COMPASS teacher coaching?

Host A That is where COMPASS comes in. What is this program, mechanically speaking? What does it actually do?

Host B COMPASS stands for Collaborative Model for Promoting Competence and Success. It is essentially a consultation and coaching program for teachers. And it's delivered either face-to-face or via the web.

Host A But what is the coach actually saying to the teacher? Are they, you know, telling them how to rearrange their classroom?

Host B No, they're targeting the mechanics of the IEP process itself.

Host A Okay, so the paperwork side of it.

Host B Right. The consultant works with the teacher to assess the student's present levels, identify the most critical needs, like those ignored social skills we just talked about, and then rigorously design measurable goals.

Host A Got it. So instead of a vague goal...

Host B Yeah, instead of something vague like improve peer interactions, the coach helps the teacher rewrite it to something very specific, like the student will initiate a verbal request with a peer twice during a 15-minute recess in four out of five days.

Host A Oh, wow. That is super specific.

Host B It is. And then they coach the teacher on specific evidence-based teaching plans to actually hit that metric.

Host A That makes total sense. You upgrade the architect, you get a better blueprint. But does a better blueprint actually mean the student builds a better house? Like, how did the studies even measure success?

Host B Well, this is where we have to look at how researchers evaluate evidence for highly individualized programs.

Host A Right, because autism is a spectrum.

Host B Exactly. You can't just give every autistic child in the study a standardized math test to see if COMPASS worked. Standardized tests often fail to capture meaningful, individualized progress. So the COMPASS trials use what's called an idiographic measure.

Host A Ideographic, meaning unique to the individual?

Host B Yes, exactly.

Host A But if kid A's goal is to raise their hand before speaking and kid B's goal is to independently navigate the loud cafeteria, how on earth do researchers compare those outcomes in a single study? It sounds like trying to mathematically compare apples and oranges.

Host B It does. But researchers use a specific tool called PET-GAS, which stands for goal attainment scaling. Think of it kind of like a currency exchange rate.

Host A Okay, I'm with you.

Host B You can't directly compare 100 Japanese yen to 100 euros. But if you convert them both into a standardized baseline, you can compare their relative purchasing power.

Host A Oh, that's a great analogy.

Host B Thanks. PET-GAS scales every individual goal on a standardized numerical grid, from baseline to expected outcome to better than expected outcome. It standardizes the degree of progress rather than the specific task itself.

Host A That is brilliant. It levels the playing field so you can actually crunch the data.

10:07

Does COMPASS help children reach their IEP goals?

Host A So let's look at that data. When they tested COMPASS, what did they find?

Host B Let's look at the 2013 randomized control trial by Ruble and colleagues. They evaluated 49 teacher-child dyads.

Host A So 49 pairings of a teacher and a young autistic student.

Host B Right. And they randomized them into three groups. One group of teachers got face-to-face COMPASS coaching. One got web-based coaching. And a placebo control group just received standard online autism training.

Host A And the results? Did the kids actually hit those PET-GAS goals?

Host B They did. Children whose teachers received the COMPASS coaching reached significantly more of their individualized goals compared to the placebo group.

Host A Who evaluated that, though?

Host B The coach? No, the progress was rated by a blind observer, meaning the person evaluating the child didn't know which group the teacher was in. And the effect sizes were quite large.

Host A Okay, let's talk about effect size for a second because we see this in the sources a lot.

Host B They use something called Cohen's d. Yes, Cohen's d is a statistical metric used to indicate the standardized difference between two means.

Host A So how do we interpret that?

Host B Generally speaking, in this type of research, a d of 0.2 is considered a small effect, 0.5 is medium, and anything over 0.8 is considered a large effect.

Host A To put that in perspective for you listening, in education research, an effect size of 0.4 is often considered a solid standard year of learning.

Host B Right.

Host A The sources note that the face-to-face COMPASS coaching had a D of 1.41, and the web coaching had a D of 1.12.

Host B Which is massive.

Host A Yeah. An effect size of 1.41 isn't just a slight bump. It is an absolute earthquake of an improvement. It implies the intervention entirely changed the trajectory of the student's year.

Host B It is a truly remarkable finding. And it appears to hold across different age groups, too. A later study abstract from 2018 looking at those transition age students we discussed earlier also showed massive gains, reporting an effect size of 2.1.

Host A If my child is suddenly hitting their goals because my teacher got some web coaching, I'm absolutely thrilled. But I have to imagine the teacher is pretty thrilled, too, right?

Host B Oh, definitely.

Host A Because special education is notorious for high teacher turnover and burnout. Did the sources look at how this impacted the educators?

Host B They did. And the mechanism there is fascinating. A secondary analysis by Love and colleagues in 2020 looked at 44 teachers from these trials.

Host A What did they find?

Host B They found that teachers who received COMPASS coaching reported significantly higher self-efficacy.

Host A Self-efficacy. So, like, confidence?

Host B Essentially, yes. It's a person's internal confidence in their ability to effectively execute a task. In this case, teaching autistic students.

Host A And confidence changes behavior.

Host B Completely. The researchers found that this increase in teacher self-efficacy directly correlated with lower teacher stress and, crucially, higher student goal attainment.

Host A It's a positive feedback loop.

Host B Exactly.

Host A The coach helps the teacher write a better plan. The student succeeds. The teacher feels competent instead of overwhelmed. Their stress drops, and they become a better teacher for that student.

Host B It really highlights how interventions that support the teacher often indirectly but powerfully support the student.

Host A I love that.

13:28

Can students lead their own IEP meetings?

Host A But so far, we've only talked about the adults in the room. We have the parents, the teachers, the coaches. There is someone missing from this architectural planning meeting, and it's the person who actually has to live in the house.

Host B The student.

Host A Right, the student, which brings us to the second intervention in our sources, the self-directed IEP.

Host B Yes, and this intervention takes a completely different philosophical approach. How so? Instead of adults sitting around a table talking about the student, the student is given specialized instruction on how to actively participate in and eventually lead their own IEP meetings.

Host A This completely flips the power dynamic of the room. I mean, we're talking about transition age kids preparing for adulthood.

Host B Right. They're about to leave the nest.

Host A Exactly. If they leave the school system and go to college or a workplace, there won't be a team of five adults advocating for their accommodations. They have to do it themselves.

Host B That's the core idea behind it.

Host A So how does this program actually work?

Host B The evidence here centers on a 2006 randomized trial by Martin and colleagues involving 130 high school students. The intervention is basically a curriculum that teaches students the specific mechanics of the meeting.

Host A Mechanics like what?

Host B They are instructed on how to introduce the team members, state the purpose of the meeting, review their past goals, present their own data on their current performance, and actively advocate for the supports they believe they need.

Host A It shifts the student from being the subject of a stressful evaluation to, like, the CEO of their own boardroom.

Host B That's a great way to picture it.

Host A What happens when a teenager actually does that? Does the dynamic of the meeting change?

Host B The data shows a profound behavioral shift. The trial demonstrated that the self-directed IEP instruction had a strong effect on increasing the percentage of time students spent talking, starting the meeting, and leading the discussion.

Host A So it actually works.

Host B Yes, the mechanics of the program successfully transferred agency back to the student.

Host A That is fantastic.

15:28

Why is the evidence rated low?

Host A It sounds like a total win, but, you know, this is the deep dive. We have to look at the blind spots. We've seen some huge effect sizes and some great concepts, but how strong is this evidence really? What are the limitations that the sources themselves point out?

Host B To evaluate this evidence through a rigorous scientific lens, we have to acknowledge several structural limitations.

Host A Okay, let's hear them.

Host B First, at a fundamental level, researchers cannot test having an IEP versus not having an IEP. Because the IEP is a universal legal entitlement for qualifying students in the United States, there is no ethical or legal way to have a completely unexposed control group.

Host A Right. You can't just deny a kid their special education services for a year just to see what happens. That would be entirely unethical.

Host B Exactly.

Host A So we are always just measuring different flavors of the IEP process against each other.

Host B Precisely. Beyond that baseline limitation, we must look at sample sizes. The COMPASS trials rely on very small cohorts. For example, just 49 teacher-child pairings in that 2013 study.

Host A That is pretty small.

Host B And as we established with your chef analogy earlier, the fact that these trials are entirely developer-run introduces a recognized limitation.

Host A Right. The coaches were the researchers.

Host B Yes. The consultants providing the coaching were members of the research team itself. Until we see independent replication by researchers who have no stake in the program's success, scientists will treat these massive effect sizes with a degree of cautious optimism.

Host A The sources also mentioned an issue with the demographics in that 2013 COMPASS study, didn't they?

Host B They did. Despite randomizing the participants, the baseline IQ levels of the children were dissimilar between the intervention group and the control group.

Host A Why does that matter?

Host B Well, when groups aren't perfectly matched at the start, it introduces a confounding variable. It just makes it harder to isolate the intervention as the sole cause of the progress.

Host A Makes sense. What about that PET-GAS measurement tool? The currency exchange analogy is great, but is there a flaw in measuring idiographic individualized goals?

Host B There is a known structural risk when evaluating idiographic data, especially when the intervention itself involves setting the goals.

Host A Okay, I think I see where this is going.

Host B If the COMPASS coaching arm is actively helping the teachers write the goals, researchers have to ask a difficult question. Are the children genuinely making more developmental progress? Or did the intervention simply train the teachers to choose more realistic, easily attainable goals?

Host A Oh, wow. Yeah. Like if I hire a fitness coach and they tell me, hey, your goal shouldn't be to run a marathon next month. It should just be to run a 5K.

Host B Right.

Host A I am way more likely to hit that 5K goal. I succeeded and my goal attainment score is super high, but only because the coach moved the goalposts closer to me.

Host B That is an excellent way to visualize the risk. Now, PET-GAS is designed mathematically to equate goal difficulty to try and prevent that very thing. But researchers acknowledge it remains a limitation when analyzing this kind of data.

Host A And what about the self-directed IEP study? Any limitations there?

Host B The primary limitation noted in the sources for the Martin 2006 study is the participant pool. That trial of 130 students was conducted on a mixed disability population.

Host A Oh, so it wasn't just autistic students.

Host B No, it was not exclusively focused on autistic students.

Host A So we can't definitively say that an autistic student with, say, profound communication challenges would experience the exact same empowerment benefit from that specific curriculum as a student with a different learning disability.

Host B Correct. The findings cannot be cleanly transferred to the entire autistic population without more targeted research.

19:11

Does a better-written IEP lead to better results?

Host A Okay, so those are the limitations of the data we actually have. But what about the gaps? What are the big questions that these sources completely leave unanswered?

Host B There are several glaring gaps in the science, and the most profound one is this. Does a better written, higher quality IEP document actually lead to better real-world results for the child?

Host A Wait, really? That's wild to think about. We know that currently only 26% of goals are measurable, but you're saying we don't actually have longitudinal data proving that the kids with the measurable goals end up doing better in life than the kids with the poorly written, unmeasurable ones?

Host B Exactly. It seems highly intuitive that a better plan leads to a better outcome. But none of the sources we reviewed have studied the empirical link between the physical quality of the document and the long-term developmental outcome of the child.

Host A What else is missing? I'm noticing we haven't talked much about the parents, even though they are legally mandated members of the team.

Host B That is another major gap. Does any of this improve parent participation or parent outcomes?

Host A Because the IEP process is notoriously taxing on families.

Host B Absolutely. So do interventions like COMPASS reduce parent stress or improve their engagement with the school? It is totally unstudied in the data we have.

Host A It's hard to imagine fixing the whole process without measuring how it impacts half of the team. Are there any other gaps?

Host B We lack long-term adult outcome data.

Host A Like post-graduation?

Host B Yes. If a child experiences a phenomenal self-directed IEP process in high school, does that translate to better employment rates or independent living at age 25? We don't have that data.

Host A And what about other countries?

Host B That's the last big one. Cross-jurisdictional applicability is unstudied. If you live in England and use an EHCP, which operates under a completely different legal framework, does teacher coaching yield the same results? We simply don't know from these sources.

21:05

What are the key takeaways?

Host A Okay, we have covered a lot of ground today. Let's pull all these threads together. If you're listening to this and trying to synthesize all this information for your next meeting, what are the core takeaways?

Host B We can distill this down into five key findings. And again, I want to reiterate that these are research findings, not specific instructions for your life or your child.

Host A Right.

Host B Number one, many standard IEP goals are currently written in ways that cannot be measured. And a significant portion of standard plans fail to document goals for identified needs, particularly social skills.

Host A Number two, the COMPASS Teacher Coaching Program shows large positive gains in helping children reach their individualized goals, and the mechanism behind that involves actively increasing teacher confidence and lowering their stress.

Host B Number three, student-led IEP instruction successfully changes the dynamics of the meeting, actively increasing student participation and leadership, though the current data draws from a mixed disability population.

Host A Number four, our limitation. The evidence for COMPASS is very promising, but it currently relies entirely on small tests run by the program's creators. The scientific community needs independent replication to confirm these large effect sizes.

Host B And number five, the open question. Science hasn't yet proven if a perfectly written, legally compliant IEP document automatically translates to better real-world, long-term outcomes for the child.

Host A This has been incredibly eye-opening. As a final reminder to you listening, decisions about any individual child belong in a conversation with that child's pediatrician or care team, and this deep dive cannot make that call. Research on this topic continues, and Spectrum Connect will update its coverage as new evidence appears.

Host B Spectrum Connect translates autism research into plain language for parents and checks findings for accuracy against misinformation. If you are a credentialed expert, MD, PhD, PsyD, or DO, and would like to help validate our findings, email research@spectrumconnect.us.

Host A So we started by talking about blueprints. We've seen how messy the drafting process can be and how coaching the architect or empowering the resident can radically change the design. But I want to leave you with this final thought to mull over.

Host B Let's hear it.

Host A If the IEP is ultimately a legal contract, you know, a rigid statutory document meant to enforce compliance, how much of a child's actual real-world success depends on the exact wording printed on the page versus the confidence, the coaching, and the human connection of the people sitting around the table?

Key Takeaways

There is no research verdict on “having an IEP.” In the US every child who qualifies gets one, so there is no group without one to compare. What has been tested is ways to make the IEP better.

One way of improving the IEP has promising results. A teacher-coaching program called COMPASS helped children reach more of their IEP goals in three small fair tests. In the one test where we could read the details (49 teacher and child pairs), the person rating progress did not know which group each child was in, and the difference was large.

All three COMPASS tests were run by the people who made it. The goals that were measured were each child’s own IEP goals, not a standard test. We found no independent test, so we rate our certainty as low.

Most other questions have not been studied enough. We found no study checking whether a better-written IEP leads to better results for a child, and no test of any program for parents. One study found that only about 1 in 4 goals in 20 high school IEPs could be measured, but that describes the paper plan, not how the child did.

If you are preparing for an IEP meeting, it is reasonable to ask how each goal will be measured, how progress will be reported to you, and whether your child can take part in the meeting.

What this means for you

An IEP is a written plan that a US public school makes for a child who qualifies for special education. A team writes it with the family. It lists the child’s goals for the year and the help the school will give, and it is reviewed at least once a year. Other countries have their own plans; in England it is called an Education, Health and Care Plan (EHCP). Because an IEP is a legal process and not a treatment, we did not ask “does an IEP work?” We asked four smaller questions: how good IEPs are for autistic students, whether parents and students take part in the meetings, whether programs that improve the IEP process help children reach their goals, and whether better IEPs lead to better results. Where an autistic child is taught (mainstream class or special school) is a separate question, covered in our report on inclusive education. Teaching methods that an IEP might name, such as peer-mediated instruction or TEACCH, have their own reports. We did not re-judge any of them here, and none of their studies are counted on this page.

The strongest result is for COMPASS, a coaching program for teachers. In a fair test with 49 teachers and their young autistic students, children whose teachers got COMPASS reached more of their IEP goals than children in the comparison group, which got online autism training instead. The size of the difference (an effect size, where bigger means a bigger gap between the groups) was 1.41 for face-to-face coaching and 1.12 for web-based coaching, both counted as large. The person who rated progress did not know which group each child was in. A test with teenagers preparing to leave school reported a difference of 2.1, and an earlier test with young children reported about 1.5. We could only read the 2018 test’s summary, and we know the earlier one only through the makers’ own later paper, so we do not have their sample sizes.

Here is why we are not more confident. All three tests were run by the team that created COMPASS, with that team’s own coaches. In the first two the authors say the coaching came from consultants on the research team, the 2013 test reached about three fourths of the sample it needed, and the groups differed in children’s IQ even after random assignment. What was measured was how far each child moved toward their own IEP goals, which is not a standard test. We read author-posted copies and summaries, not independent full texts. A later test with school-based consultants exists, but we could not read it. In one COMPASS study of 44 teachers, teachers who got the coaching rated themselves more confident about teaching autistic students, but the teachers filled that in themselves and knew which group they were in.

The IEP-quality research describes plans; it does not test them. In one study of 20 IEPs for autistic students in their last year of high school, only about 1 in 4 goals (26.4%) was written in a way that could be measured. About nine in ten of the IEPs said the student needed help with social skills, but only 22% of those had a social skills goal. The authors did not look at whether this affected how students did, and we found no study that did. A plan that is well written on paper is not the same as a plan that helps a child.

For students taking part in their own meetings, one fair test with 130 high school students found that a program called the Self-Directed IEP made students talk more and lead more of their meetings. The students were not chosen because they were autistic, so we do not know how well it applies to autistic students. For parents, we did not find any test of whether a program changes how parents take part or how parents feel. We also found no research on England’s EHCP or on plans in other countries in this short search.

There is no research verdict on having an IEP, but one way of improving it, COMPASS coaching, helped children reach more of their own IEP goals in three small tests, all run by its makers.

Who was studied. COMPASS was tested with young autistic children (ages 3 to 8 in the teacher study) and with teenagers preparing to leave school. The IEP-quality study looked at 20 IEPs from students in their last year of high school. The Self-Directed IEP test had 130 secondary students who were not chosen for autism. Everything here concerns the US system; we did not capture where each study took place, how long each program ran, or the sample sizes for two of the three COMPASS tests. We read summaries and author-posted copies, not independent full texts. Whether these results apply to a particular child or school is not established.

Where the studies landed

Promising but contested

Eight sources are shown below. Only the COMPASS tests compare groups on how well children reach their IEP goals. None of the other sources tests whether something helps an autistic child reach goals.

3 COMPASS tests: children reached more of their own IEP goals

3 Helpful signals, but a different question or too limited

2 Describe IEPs, but don’t test whether anything helps

Tap the cover to open it. Six more records were found but only their titles were captured, so they are not counted. Two sources here (Love 2020 and Findley 2022) appear to reuse the children of the COMPASS tests, so they are not independent evidence.
One study

Tap any tile to read it. Six more records were found but only their titles were captured, so they are not counted. Two sources here (Love 2020 and Findley 2022) appear to reuse the children of the COMPASS tests, so they are not independent evidence.

COMPASS tests: children reached more of their own IEP goals3
Three fair tests of COMPASS, all by its developers. Ruble 2013 (49 teacher and child pairs) found large gains in goals met against online autism training, rated by someone unaware of group. Ruble 2018 (teenagers preparing to leave school) reported a larger difference of 2.1, from its summary only. A 2010 test with young children reported about 1.5, and we know it only through the developers’ later citation. We rate the 2013 test provisionally as having some concerns about bias, and could not rate the other two.
Helpful signals, but a different question or too limited3
Love 2020 found that COMPASS teachers rated themselves more confident, but this was self-reported by teachers who knew their group. Martin 2006, a fair test of the Self-Directed IEP, found students talked and led more in their meetings, but the students were not chosen for autism and it measures participation, not how the child did. Allen 2001 is a small study of students with moderate intellectual disability, and we read only its record, so it does not reach this question.
Describe IEPs, but don’t test whether anything helps2
Findley 2022 and Ruble 2010 (JADD) look at how IEPs are written. Findley found measurable goals in about 1 in 4, and said it did not examine links to student outcomes. We could not read the 2010 paper’s abstract, so we do not know whether it links quality to results. Document quality is not the same as benefit to the child.

Spectrum Connect reviews published research on interventions parents are exploring for their autistic children — so you can see where the evidence actually stands. No agenda, no selling, no cherry-picking. Just the studies, our method, and what it means for you.

See the research behind this For anyone who wants to check our work · 8 sources
01

Where we looked

This was a web-based scoping search, not a full multi-database search: 2 web searches plus 10 page-read attempts. Of those 10, 8 returned content (4 publisher or author-posted pages and 4 Crossref lookups) and 2 failed (an ERIC full-text PDF was refused, and a Springer page was rate-limited and not retried). One of the Crossref lookups, for England’s EHCP, returned only unrelated records, and another returned details but no abstract. No database was searched directly, and PubMed and PMC pages were not used because of a CAPTCHA, so there are no per-database counts. The search of record for a full run covers ERIC, APA PsycINFO, PubMed/MEDLINE, Education Source, Cochrane CENTRAL, the Campbell Library and Epistemonikos; trial and study registers (ClinicalTrials.gov, the IES/NCEE funded-project database and the Registry of Efficacy and Effectiveness Studies); and UK grey literature for EHCP (DfE, Ofsted, NFER). Here is the PubMed search of record.

("Autism Spectrum Disorder"[MeSH] OR autis*[tiab] OR ASD[tiab]) AND ("individualized education program*"[tiab] OR IEP[tiab] OR "individualised education plan*"[tiab] OR "education, health and care plan*"[tiab] OR EHCP[tiab] OR COMPASS[tiab] OR "self-directed IEP"[tiab] OR "student-led IEP"[tiab] OR "goal attainment scaling"[tiab]) Run on PubMed →

Counts for every database, and the dates searched, are pending, and no protocol was registered in advance. This search is not reproducible from the record we kept.

02

What we did with what we found

32records found (19 web results + 13 database records)
−3the same papers found on different sites
29judged by title (not formally screened)
−12blogs, a sample-IEP page, a practice article, a dissertation, an unopened paper and 6 off-topic records
14cited as sources
−6title only, page not captured; listed in the record
8shown here (3 COMPASS fair tests + 1 teacher analysis + 1 student-led IEP trial + 1 single-case study + 2 studies of how IEPs are written)

This is a scoping count, not a formal screening flow. No second reviewer screened these works, and none of the independent trial full texts was read; three author-posted full texts were read through a summary. The counts above are recorded as stated: 29 works judged less 12 not selected does not equal 14 cited, and we have not adjusted them. The 2010 COMPASS test with young children is shown even though we know it only through the developers’ later citation, because the findings above rely on it.

03

What the strongest evidence says

Children reaching their own IEP goals (COMPASS)

Large gains in three small fair tests, all by the program’s makers. In the 2013 test the rater did not know the group. The goals measured are each child’s own.

How sure
Low
Teacher confidence (COMPASS)

Teachers who got COMPASS rated themselves more confident (44 teachers). Self-reported by teachers who knew their group, in a follow-up analysis.

How sure
Very low
Students leading their own meetings

A fair test with 130 high school students found they talked and led more, but they were not chosen for autism.

How sure
Very low
Whether a better-written IEP leads to better results

Not tested in what we found. One study counted what is written in 20 IEPs but did not look at how students did.

How sure
Not studied
Programs that change how parents take part or feel

Not studied in what we found.

How sure
Not studied
Whether having an IEP works

Cannot be tested: every child who qualifies gets one, so there is no group without one.

How sure
No verdict
Burdens or harms

None measured. Meeting time and teacher stress were not studied, and no data is not the same as no harm.

How sure
No data
Plans outside the US, such as England’s EHCP

Nothing found in this short search.

How sure
Not found
Ray Kawai · Protocol v4.5 BCAT · Open record
Spectrum Connect is not a medical provider, and nothing here is medical advice. This page shows where the research stands and how we got there. It is not a recommendation, and it is not a substitute for your child’s doctor or care team. What you do with it is yours to decide, together with them.

What would you like to do?

Want to talk to a parent who’s been there?

We match you with families whose kid is like yours. Free, takes two minutes.

Join Free
Heard of something we haven’t covered?Suggest an article

If another parent, a clinic or a video mentioned a therapy you can’t find here, tell us. We’ll read the research and review it next.

Suggest an article
Think we got something wrong?Report a factual error

We publish the whole record so it can be checked, and that only counts if we act on what you find. If a number looks wrong, a study is missing or has been retracted, or we’ve read a finding in a way the evidence doesn’t support, tell us.

You don’t need a research background to file one. “This doesn’t match what our doctor told us” is a useful report. Every one reaches a person: we reply within seven days, and within thirty we have either corrected the page or told you when we will. Substantive reports send the affected steps back through the protocol and need fresh sign-off before anything here changes.

Report a factual error
Are you a clinician or researcher?Review this article

If you have read this and the evidence holds up, you can sign off on it. We send you a personal review link where you confirm your credentials, and our team checks every review before anything is published. Once approved, your name and a link to your profile appear on this article.

Ask to review this