Three members reported first-attempt EPPP passes in a single week in August. A 661, a 600, and a 555.

They prepared almost nothing alike. One took eight full-length practice exams in seven days. One had been working with us for months and is the most active question-feedback reporter we have. One signed up the day before the exam and never took a full sim at all.

That third one is why this post exists. When three results land at once, some of them confirm what you already published and some of them make you narrow it. Both happened here.

How we collect this, so you can judge it yourself: we ask every member who sits for the real exam to write back, pass or fail. Twenty-three people have. Fifteen passed, eight did not. These are collected reports, not a pass rate, and only people who chose to tell us are in them. Details below are kept vague on purpose. We never publish anything that could identify someone from their data.

Case one: 661, eight sims inside a seven-day trial

This one signed up for the free seven-day trial and used every day of it. Eight full-length practice exams in that week. Best score 93%. One more full simulation on the morning of test day, 92%. Real exam on the last day of the trial, 661, first attempt.

Before us there were months of other work: PrepJet, free practice tests, flashcards. Their own framing was that we were the last tool they used to hone their knowledge before the exam. That is the right description of what a final week does, and we wrote a whole post about the limits of it: the last week before the EPPP is not for learning new material.

<!-- verbatim quote slot: case one (661), pending testimonial permission -->

Case two: 600, months of work and an exam that felt harder

This is the long-haul version. 2,218 logged study minutes, about 37 hours on the platform, spread over months. This member also reports more feedback on individual questions than anyone else in our data, which is its own kind of studying: you do not argue with a question unless you have thought hard about it.

The exam was in late August. Score 600, first attempt.

What they told us afterward matters more than the number. The real exam felt much harder than most of the questions on our platform. Not the same, not easier. Harder. I will come back to that, because it is the most useful thing anyone told us this month.

We do not have their full-length sim history, so I am not going to imply one. Unknown means unknown.

<!-- verbatim quote slot: case two (600), difficulty comment, pending testimonial permission -->

Case three: 555, thirty-six hours, zero sims

This member found us on TikTok and signed up one day before the exam. In roughly 36 hours they logged 287 study minutes and 18 quizzes. Full-length practice exams taken: zero. The only timed measurements were two short diagnostics, 68% and 74%.

Behind that were months of preparation somewhere else: ChatGPT, ASPPB materials, PrepJet. Then a first-attempt pass at 555.

<!-- verbatim quote slot: case three (555), pending testimonial permission -->

What this does to the 83% finding

We published a finding in am I ready for the EPPP: every passer we could measure had taken a full-length, timed practice exam before test day and peaked at 83% or better, while none of the earlier fails had taken one before test day. Three new results, three different verdicts on it.

Case one fits it exactly. Eight full sims, a peak of 93%, a 92% on the morning of the exam, then a 661. That is the pattern in its cleanest form, just compressed into a week instead of spread across months.

Case two neither fits nor breaks it. We cannot see their sim history, so they are not evidence in either direction. Counting an unknown as support would be the easiest way to make a finding look stronger than it is.

Case three narrows it. They passed on a first attempt with no full-length sim anywhere in our data. That is a genuine problem for the sentence as written, and I am not going to explain it away.

Here is the narrower version, which is the one I will use from now on. Among passers whose preparation we can actually see, every one put a full-length, timed score on the board before test day. Case three's preparation happened almost entirely off our platform, across months, in tools we have no view into. They may have run full-length practice exams elsewhere. They may not have. We do not know, and a finding that cannot tell you which should say so out loud rather than quietly count the case as a win.

The finding is still the strongest signal in our data. It is just a claim about measured prep, not a law about human beings.

What the margin looked like

One more thing worth saying carefully. The two who reached us in their final week both passed, and the sim number tracked how far above passing they landed rather than whether they landed there. Eight sims with a 93% peak, 661. No sims, 555.

That is n=2. It is a hint, not a finding, and I would not want anyone rearranging their study plan around it. But it does line up with what we saw in the what EPPP passers actually do differently data, where the practice number was never pure noise.

When someone says the real exam was harder

Case two's comment is the one I keep thinking about, and it points the opposite way from a report we published earlier this summer, when a passer told us our practice questions ran more detailed than the real thing and over-prepared them on content.

Both of those are true reports from real people. That is what twenty-three data points look like when you do not sand them down. Question difficulty is not one number; it moves by domain, and it moves with whichever version of the exam you happened to sit.

Here is what we do about it, without dressing it up. We made our questions harder in July, and that work is still going. The direction we want is the one the earlier passer described: practice that runs slightly harder than the real exam, so test day feels like a step down instead of an ambush. A report that says the real thing felt harder than our questions means we are not there yet in at least one place, and it is worth more to us than a compliment. If you have taken the real exam recently, tell us where it outran the practice. That is the feedback that actually changes the bank.

What we take from a week like this

Three pass reports in one week is a nice thing to see and a bad thing to over-read. Three is not a trend, twenty-three is not a study, and the people who report back are not a random sample of the people who test.

What the week actually gave us is one clean confirmation, one honest unknown, and one case that made a published sentence narrower and more accurate. That last one is the most valuable of the three.

If you sat the EPPP recently, write in either way. Pass or fail, the reports are the only reason any of these posts have numbers in them, and the fail reports do more work than the passes. thePsychology.ai is free for a week, which is long enough to take one full-length simulated exam and find out where you actually stand.