What 8 EPPP Score Reports Showed Us, and What We Changed

We read the score reports.

When a member tells us they did not pass the EPPP, we ask for a photo of their ASPPB score report. The report shows a bar for each of the eight domains. It does not show a percent correct. It shows how close you came to the passing line in each area.

Eight members sent reports we could read. This post is what those bars showed, what they cannot show, and what we changed in the product because of them.

Before the numbers, the limits. Eight is a small group. These are self-reported results, not a pass rate. Most of these members joined thePsychology.ai after their exam, so their bars mostly reflect how they prepared before they found us. The bars tell us where people who did not pass need help. They do not prove our content caused anything, good or bad.

The count as of September 22

As of September 22, 40 members have reported an EPPP result to us: 25 passed and 15 did not. That count has moved since our last post on practice scores, which used 36.

Eight of the 15 sent a score report photo we could read. Two of those eight were hard to read, so we trust them less. One more member sent photos in a format our reader could not open.

What the bars showed

There was no single weakest domain. Every report looked different. That alone matters: one study plan for everyone would not fit these reports.

But two domains kept showing up near the bottom.

Domain	Share of the exam	Average bar (of 100)	In a member's bottom 3
Research Methods and Statistics	7%	48	6 of 8
Assessment and Diagnosis	16%	51	4 of 8
Treatment, Intervention, Prevention and Supervision	15%	56	4 of 8
Biological Bases of Behavior	10%	57	3 of 8

Research had the lowest bars. It was in the bottom three for six of the eight members.

Assessment and Diagnosis had the second lowest bars. And because it is 16% of the exam, it cost the most points overall. A weak spot in a heavy domain hurts more than the same weak spot in a light one.

One more pattern surprised us. Treatment was in the bottom three for half the reports. But when members who did not pass ticked the areas they felt weakest in, only 3 of 11 picked Treatment. The domain that felt fine was not always fine on the report.

What this cannot tell you

These bars are ASPPB's picture of each domain, not a score we can compare line by line. Members who pass get no bars at all, so we cannot compare them with passers. And eight reports cannot tell you why someone did not pass. They can only tell you where the gap showed up.

So we did not read this as "our members are bad at Research." We read it as: if the gap keeps landing in the same two places, those two places need to be the clearest parts of the product.

What we changed so the next person in that spot knows what to do

Our own member ratings pointed the same way. Since late July, Research questions had the lowest average star rating of any domain: about 2 out of 5, with 4 of 6 ratings at 1 or 2 stars. That is a tiny sample. The complaints were about clarity: stems that packed in too much, and a question where the options did not match what the stem asked for. So we started there.

Research, Assessment, and Test Construction got a full audit. We checked all 1,362 questions in those areas for wrong keys, giveaway answers, options of very different lengths, and questions that left out the age group or test edition they meant. We fixed the facts or wording on more than 90 existing questions.

We added 144 new applied questions. That is 44 in Research, 38 in Assessment, and 62 in Test Construction. They are the kind the real exam leans on: which statistic fits this design, which threat to validity is this, which test answers this referral question, what does this reliability number mean for this client. Most are medium or hard. The correct answers are spread evenly across A, B, C, and D, so the position of an answer tells you nothing.

We checked every changed fact against a source. Each fact we changed in a lesson or explanation was checked against published research on PubMed (55 papers in all), or against the test publisher, the test manual, or the 2014 testing Standards. If we could not verify a claim, we did not write it. Our IQ lessons and questions now teach the WAIS-5, the current Wechsler adult scale, checked against its manual.

We edited 17 lessons and re-recorded their audio. Every Research, Assessment, and Test Construction question now quotes the exact lesson sentence it is built on, word for word. When you miss one, the lesson opens to the section that holds that sentence. That works for all but 3 of those 1,506 questions.

Your score report now reads right, even from a sideways photo. We found that photos taken sideways made our reader see every bar about 20 points higher than it was. That could send a retaker to the wrong domain. The reader now turns the photo upright and reads it again. If it still is not sure, it builds your plan from the areas you tick instead of guessing.

Your retake plan ranks by points, not by feel. When you upload your report, your plan puts each domain in order by how far below the line it is, times how much of the exam it covers. That is why Assessment often ranks first even when Research has the lower bar. And it is why a domain you feel fine about can still show up in your top three.

If your next attempt is coming up

If you did not pass, send us your score report on the retake page. Your study plan will be built from your real bars, not from a guess. If you are still deciding how to use your practice exams, our post on what practice scores predicted walks through what the numbers can and cannot tell you.

And if you have taken the exam, pass or not, please tell us. Every report makes the next plan a little sharper for the person behind you.