A Study Analyzed Nearly 450,000 IEP Goals. Here's the Pattern That Keeps Showing Up.

Most providers have a private suspicion about IEP goal quality: that it varies more than it should, that some goals in some files would not hold up if anyone looked closely, and that this is not really anyone's individual fault. A recent large-scale academic study gives that suspicion something it rarely gets: actual numbers.

Researchers at the Wheelock Policy Center analyzed 448,533 IEP goals belonging to 184,960 students, in what its authors describe as the first comprehensive statewide study of IEP goal content across all disability categories. It is not a study about compliance checklists in the abstract. It is a systematic look at what is actually written in real goals.

A separate body of research, including goal-quality studies by Boavida and colleagues, Ruble and colleagues, Sanches-Ferreira and colleagues, and others going back to the early 2000s, has repeatedly converged on the same handful of failure modes across different states, disability categories, and decades of practice:

• Not disability-specific: generic enough to apply to almost anyone with a similar label.
• Missing components: no clearly observable behavior, no stated condition, no measurable criterion.
• Not functional: targets skills that do not reflect the student's actual daily or academic life.
• Does not generalize: rarely addresses transfer across home, classroom, and community settings.
• Lacks measurability: the single most cited problem across nearly every study on this topic.

That last one deserves particular attention, because it is not a cosmetic issue. Under IDEA 2004, a goal that cannot be objectively measured is more than weak writing. It cannot demonstrate progress, cannot justify continued or adjusted services, and is unlikely to survive scrutiny if a parent's advocate or a due process hearing officer reads it closely.

It would be easy to read this as a story about individual carelessness. The research says otherwise. A 2023 practitioner-focused analysis in a major speech-language pathology journal reported that goal writing is difficult even for trained, experienced providers, difficult enough that many turn to internet or software-based goal banks to keep up with caseload demands.

What appears to drive the pattern:

• Caseload pressure pushes providers toward the fastest available option, not necessarily the most individualized one.
• Goal banks scale the writing process by averaging across students, which is the opposite of what measurability requires.
• Some providers unintentionally write goals that mirror specific items from norm-referenced assessments, which can invalidate future testing and may not reflect what is relevant to the student.
• The tools built to save time can be the same tools producing the goals the research keeps flagging.

None of this is a story about providers not caring enough or not knowing better. It is a story about a workflow problem.

Takeaway: this is a systemic pattern with a systemic cause: caseload pressure meeting tools that trade individualization for speed. Fixing it means changing the workflow, not asking providers to try harder.

If these problems are structural rather than individual, the fix has to be structural too: not another goal bank with more templates, but a process that keeps the goal tied to the student's actual data at every step.

That is the design problem IEP Pilot was built around. Every goal it generates is built from the student's own uploaded assessment or entered baseline data, not selected from a bank of pre-written options. Here is how that maps to what the research found:

• Goals are not specific to the disability: goals are built from the student's own uploaded assessment or baseline data.
• Goals lack measurability: goals are structured to include an observable behavior, condition, and criterion every time.
• Skills are not functional: goals, present levels, and services are generated together rather than in isolation.
• Goals do not generalize: goals are individualized per student rather than pulled from a shared bank.

None of this replaces the provider's judgment. Several of the same studies note that generic AI or templated tools can reproduce these failure patterns at even greater scale if they are not grounded in real assessment data. The distinction that matters is not automated versus manual. It is whether the goal is built from the student in front of you or from an average of every student who came before.

Takeaway: the fix is not less automation or more automation. It is automation tied to real data.

At nearly half a million goals, this is not a story about a few overworked providers cutting corners. It is a documented, widespread pattern, which means the fix has to work at the level of process, not effort. Providers already know what a measurable, functional, individualized goal looks like. The research suggests the gap is not knowledge. It is whether the workflow makes that standard achievable at caseload scale, every time, not just when there is room to slow down.

Sources: Wheelock Policy Center, "Understanding Individualized Education Program (IEP) Goals at Scale" (2024 working paper); Boavida et al. (2010); Boavida, Aguiar & McWilliam (2014); Ruble et al. (2010); Sanches-Ferreira et al. (2013); Catone & Brady (2005); Pretti-Frontczak & Bricker (2000); Kurth & Mastergeorge (2010); Rakap (2015); ASHA Perspectives of the Special Interest Groups, "Writing Measurable and Academically Relevant IEP Goals."