i-Ready Finally Has Independent Research. What Does the Study Actually Say?

For years, parents and educators have raised questions about i-Ready: How effective is it? How much instructional time should students spend using it? Is it intended to supplement teacher instruction or replace portions of it? And, perhaps most importantly, what does independent research actually show?

A newly published study finally gives us considerably more information.

In June 2026, Jennifer Darling-Aduana and K. Jurée Capers of Georgia State University published “Evaluating a Personalized Online Learning Tool for Academic Recovery: Lessons from i-Ready” in Online Learning, Volume 30, Issue 2. The study examined i-Ready implementation across a large southeastern metropolitan school district and included nearly 28,000 student-year observations of students who used the program.

Then, on July 29, 2026, education researcher Jared Cooney Horvath examined the study in his Substack, The Digital Delusion, in an article provocatively titled: “i-Ready Finally Has Real Research…And It Changes Nothing.”

His argument deserves attention. But parents should also go one step further: read the actual study.

Because the research itself tells a much more nuanced story about i-Ready—one that includes modest overall effects, limitations on what researchers can conclude, the importance of teacher involvement, questions about instructional time, and explicit warnings against allowing online programs to operate as replacements for classroom instruction.

What The Digital Delusion Says

Horvath begins by acknowledging something important: he considers the new Georgia State study the strongest evaluation of i-Ready to date.

He describes it as a careful, sophisticated study involving tens of thousands of students and published in a reputable peer-reviewed journal. His disagreement is not primarily with the quality of the research. It is with what the results actually mean.

According to his analysis, the study's overall results amount to approximately 0.04 standard deviations of additional achievement in both reading and mathematics. He translates that into approximately a 1.5-percentile-point movement for an average student—from roughly the 50th percentile to 51.5th percentile.

That is statistically detectable. But statistical significance and educational magnitude are not the same question. And that distinction is central to his criticism.

“Weeks of Learning” Sounds Considerably Bigger

The study translates some of its findings into approximately two additional weeks of instructional growth in mathematics and nearly three weeks in reading.

Horvath challenges that presentation. His concern is that the study does not adequately explain how the effect-size estimates were converted into “weeks of learning.” He argues that this language may make a relatively small standardized effect appear more substantial to a general audience.

That does not mean the researchers fabricated the result. It means parents and policymakers should understand the difference between the measured statistical effect and a subsequent attempt to express that effect in more intuitive terms.

https://nepc.colorado.edu/blog/i-ready-finally

Now Let's Look at the Actual Study

The researchers examined i-Ready as a COVID-19 academic recovery strategy.

The district had made i-Ready available to participating schools, but teachers retained discretion over its use. Students could use it during the school day and outside of school. The researchers analyzed the staggered rollout of the program along with district administrative data and i-Ready usage records.

Approximately 67,602 student-year cases were eligible to use i-Ready, while 27,717—about 41%—actually used it.

The researchers themselves note that participation was not uniform. For example, usage was lower in middle school, and some elementary teachers with access apparently chose not to use the platform.

That matters because this was not a randomized controlled trial.

The researchers used sophisticated statistical methods—including regression, school fixed effects and inverse probability-weighted regression adjustment—to try to account for differences between students.

But sophisticated statistics cannot turn an observational study into a randomized experiment.

And the researchers are transparent about that.

They explicitly describe their analyses as descriptive rather than causal, explaining that they could not completely account for which students received the intervention or other academic-recovery interventions occurring simultaneously.

The study found an association between i-Ready use and improved achievement. It did not establish that i-Ready caused all of the observed difference.

The Teacher Matters

This may be the most important finding for parents and educators.

The study did not find that putting students in front of an adaptive computer program and allowing the algorithm to take over was the optimal implementation model.

Quite the opposite.

Students whose teachers assigned at least one lesson in addition to the lessons assigned automatically by the program experienced greater positive associations with achievement.

The researchers ultimately conclude that “teacher input is critical” to the effectiveness of online learning platforms.

The study's abstract itself says these platforms appear useful when implemented in a manner that integrates teacher expertise and uses an appropriate dosage.

And in their discussion, the researchers become even more explicit. They caution districts against using online platforms as an “auto-pilot” replacement for classroom instruction.

Instead, the researchers recommend that these platforms function as complementary tools to in-person instruction, with teachers monitoring usage and strategically assigning lessons.

i-Ready as a supplemental instructional tool is one proposition.

i-Ready replacing meaningful teacher-led instruction is another.

This study should not be used to conflate the two.

Interestingly, Curriculum Associates itself now publicly describes i-Ready in essentially the same way: as a targeted, time-limited complement to teacher-led instruction, not a substitute for it.

More i-Ready Is Not Necessarily Better

Another finding deserves attention.

The researchers found a nonlinear relationship between usage and outcomes.

Their analysis suggests online instructional tools may be most effective within an approximate range of 40–100 hours per school year, with diminishing returns beyond certain thresholds.

And the researchers explicitly caution that expanding use beyond those thresholds may not be a wise investment of instructional time or student effort.

That changes the question schools should be asking.

It should not simply be: “Did the student complete enough i-Ready minutes?”

It should also be: “What instruction did those minutes replace?”

Because instructional time has an opportunity cost.

Forty minutes spent on a computer program is forty minutes that cannot simultaneously be spent receiving something else.

That “something else” might be explicit teacher instruction, small-group intervention, specially designed instruction, writing instruction, hands-on mathematics, peer interaction, or individualized feedback.

A positive effect from a computer program does not automatically establish that it is the best use of that instructional time.

And Then There Is the Cost

This is another area where The Digital Delusion raises an important question—but we need to distinguish Horvath's analysis from the study itself.

The researchers estimated approximately $831,510 in subscription costs over two years, using an assumed cost of $30 for each of the 27,717 students who actually used i-Ready.

Horvath questions that calculation.

Why?

Because 67,602 student-year cases had access to i-Ready, even though only approximately 41% actually used it.

If a district purchases licenses based upon access rather than actual student usage, Horvath argues, calculating costs only for students who logged into the program could substantially underestimate what the district actually spent.

He further notes that the study does not provide a source for the $30 figure or clearly establish whether that price represents one subject or both reading and mathematics.

From those uncertainties, he constructs alternative cost scenarios that could be substantially higher than the study's estimate.

Those alternative figures are Horvath's analysis—not findings of the Darling-Aduana and Capers study.

But the underlying question is legitimate:

What did the district actually purchase, what did it actually cost, and what educational benefit did students receive in return?

That is precisely the kind of question school boards and communities should be able to answer when evaluating educational technology.

What About Students With Disabilities?

This is where parents of students receiving special education services should exercise particular caution.

The study examined disability status among its subgroup variables and reported some intriguing findings, including stronger associations for students identified with intellectual disabilities.

But those subgroup findings should not be transformed into a broad conclusion that i-Ready is an effective special education intervention for every student with a disability.

The study's own literature review acknowledges that prior research involving students receiving special education services and technology-based learning has produced positive, null and negative associations across different blended-learning environments.

And, critically, this was an academic-outcomes study.

It was not an IDEA compliance study.

It did not determine whether i-Ready constitutes specially designed instruction under IDEA.

It did not determine whether i-Ready can replace instruction identified in an IEP.

It did not determine whether an i-Ready Diagnostic alone is sufficient to measure progress toward an individualized annual goal.

And it did not determine whether assigning i-Ready lessons fulfills a district's obligation to provide the services written into an individual student's IEP.

Those are different questions.

For a student with an IEP, the question cannot simply be whether a commercially available program is supported by research.

The question must remain: What does this individual child require to make appropriate progress in light of the child's circumstances?

The Research Is Valuable—But Read What It Actually Says

There is a temptation in education to divide programs into two categories: It works. Or it doesn't work.

This study does not support such a simplistic conclusion about i-Ready.

• The researchers found positive associations with reading and mathematics achievement.

• The overall effects were relatively small.

• The study was observational rather than randomized, and the researchers expressly caution against causal interpretation.

• Teacher involvement improved outcomes.

• Usage had diminishing returns.

• The researchers caution against allowing online platforms to become an “auto-pilot” replacement for classroom instruction.

• And the cost-benefit question remains very much open.

That is considerably more nuanced than either “i-Ready doesn't work” or “research proves i-Ready works.”

Perhaps the most important finding is actually much simpler:

Technology appears to work better when teachers remain actively involved in the teaching.

For parents—especially parents of children with IEPs—that distinction matters.

Personalized software is not the same thing as individualized education.

An algorithm can select a lesson.

It cannot replace the IDEA process of identifying a child's unique needs, developing measurable annual goals, determining appropriate specially designed instruction and related services, implementing those services, collecting meaningful progress-monitoring data, and revising the program when the child is not making appropriate progress.

i-Ready may have a place in that educational program.

But this new research does not establish that the computer should become the educational program.

https://www.researchgate.net/scientific-contributions/K-Juree-Capers-2351709711

Sources

Jennifer Darling-Aduana & K. Jurée Capers, “Evaluating a Personalized Online Learning Tool for Academic Recovery: Lessons from i-Ready,” Online Learning, Vol. 30(2), 459–490 (June 2026).

Jared Cooney Horvath, “i-Ready Finally Has Real Research…And It Changes Nothing,” The Digital Delusion (July 29, 2026).