CAI Psychs

Have you heard someone say that a PSW method is no better than a coin flip? The studies that led to that statement left out some important factors…

The "coin flip" line comes from simulation studies that compared true scores and observed scores. Whenever the two landed on opposite sides of a cut point, they called it a diagnostic error. Stack up enough of those errors, and the conclusion writes itself: PSW doesn't work.

We looked closer. Here's what we found.

First, most of those "errors" were minimal. The simulation studies didn't control for measurement error. When you do account for measurement error, the diagnostic accuracy improves significantly.

Second, the studies used population base rates. But you're not pulling random kids out of the general population to assess. You're seeing kids who were referred. Use referral base rates and the numbers shift again.

Diagnostic accuracy should be measured at the individual case level. Not "how accurate is PSW for a random person from the general population," but "how confident can this practitioner be that the decisions they made on this student are correct?"

That work is now built into X-BASS 3.0. 🙌

When you run a PSW analysis, you can see probability statistics for the decisions you made to ensure the most accurate possible results.

Want to try X-BASS 3.0?

👉 Comment DISCOUNT below and we'll send you a coupon code for 25% off through Wiley.

❌ “You don't need cognitive tests to identify a specific learning disability."

Specific learning disabilities are related to and caused by weaknesses in cognitive and linguistic processing… There is enough research to say that with a high degree of certainty.

So when Dr. Flanagan hears "you don't need cognitive tests to identify SLD," her response is straightforward. You are cutting out an important piece of the puzzle!

For example, when using achievement data only, phonological processing gets called "academic" by some and "cognitive" by others. The label changes. The construct stays exactly the same.

Arguing about what to call the test is not the same as arguing about whether the test is useful. It is.

What do you think about cognitive testing and SLD identification?👇

Teachers get three years of structured induction. School psychologists get a desk and a caseload.

In teaching, there's a research base showing that three years of induction support keeps teachers in the profession. Without it, they leave.

School psychology has no equivalent. You finish your program, you start your job, and you're often the only one in the building doing the work.

The result is predictable. You're running case to case, person to person, with no time to think and nowhere to debrief. It stops feeling like a job and starts feeling like an island.

Dr Alonso’s advice for year one and year two: build a professional network before you need it. Find real mentorship, not just a name on a form. Attend the webinars, the workshops, the conferences. Stay in conversation with people doing the same work you're doing.

The isolation is structural, not personal. The fix has to be too.

What’s your best advice for new school psychs? 👇

How many courses on specific learning disabilities did you take in grad school?

For most school psychologists, the answer is zero. Not because the topic isn't important. SLD evaluations are probably the single most common thing you do. But somehow, most graduate programs still don't require a dedicated course on it.

Dr. Flanagan says it plainly: most entry-level school psychologists aren't yet fully equipped to do effective SLD evaluations. Continuing education is crucial in bridging the gap between graduation and evaluation expertise.

Educational psychology. Cognitive psychology. Cognitive neuroscience. There's a massive body of research on SLD spread across disciplines. But it can be hard for programs to keep up with applying this research to the curriculum.

That's why CAI Psychs exists. We built this for the school psychologist who knows there is more to learn and pursues deeper knowledge to better support their students.

👉 Comment SLDFREE to access a free webinar about specific learning disabilities!

AI can’t testify in court for you.

Yes, AI tools can help you phrase things, tighten language, and work through certain sections of a report more efficiently. That part is currently fair game when used judiciously and ethically.

What AI cannot do is replace the examiner (you!).

It cannot replace the clinical judgment that goes into selecting the measures, interpreting the data, and writing about a specific child. The professional responsibility to proofread, verify, and stand behind every word in that report belongs to the person whose name is on it.

If you get called in to defend an evaluation, whether it is a due process hearing, an IEP dispute, or anything else, AI is not the one on the stand. You are.

How do you use AI at work as a school psychologist?👇