Every prep book advertises a number. Three hundred questions. Five hundred. A thousand. The number is doing a lot of work in that sentence and almost none of it is honest, because a question bank is not a pile — it is a covering of something, and nobody tells you what it is covering.
Here is the arithmetic nobody publishes.
The unit is a statement, not a topic
ASWB does not write its exams from chapters. It writes them from applied knowledge statements — the individual lines in the published content outline, each one a specific thing you are expected to be able to do. Every question on your exam traces back to exactly one of them.
There are more of these than people expect, and how many depends on which exam you sit:
| Exam | Applied knowledge statements |
|---|---|
| Bachelors (LBSW) | 110 |
| Masters (LMSW) | 109 |
| Advanced Generalist (LISW) | 115 |
| Clinical (LCSW) | 116 |
That is the denominator. It is the only denominator that matters, and it is the one the marketing never divides by.
What "500 questions" actually buys you
Divide.
| Bank size | Questions per statement (Masters) |
|---|---|
| 300 | 2.8 |
| 500 | 4.6 |
| 1,000 | 9.2 |
Three questions on a statement. That is what a 300-question book gives you on mandatory reporting, or on assessing suicide risk, or on the difference between a dual relationship and a boundary crossing.
Ask yourself what three questions can tell you. They can tell you that you got three questions right. They cannot tell you that you understand the statement, because the exam will not ask you the same three. It will ask a fourth, built from the same statement and a different vignette, and you will meet it for the first time on exam day.
Coverage is not the same as depth, and neither is the same as confidence. A bank that averages three questions per statement gives you no signal at all: get two of three right and you have a 67% on a sample far too small to mean anything.
The number you actually need is set by how you learn
There is no magic figure, and anyone quoting one is guessing. But there are two floors worth naming, and they come from how learning works rather than from marketing.
The first floor is discovery. You need enough questions per statement to find out what you do not know. One question tells you nothing; two tells you nothing. Somewhere around five or six per statement, patterns start appearing — you notice you are reliably wrong about capacity, or about which layer of obligation controls.
The second floor is repair. Discovery is the cheap half. Once you have found a weakness you need more questions on that statement — fresh ones, not the ones you have already seen — to check whether the repair took. Every question you burn on discovery is one you cannot reuse for repair.
This is why the total matters more than it looks. A bank of 500 spread over 109 statements has no depth left after the first pass. You will have seen every question, and re-answering a question you remember measures memory, not understanding.
Then there is the part the number cannot tell you
A question is only as useful as its explanation. The ASWB exam is a best-answer exam: most of its questions offer more than one defensible option, and the whole difficulty is choosing between two answers that are both arguably right.
So the explanation has to say why the wrong ones are wrong. Not "B is correct because the Code says X" — that tells you nothing you could not read yourself — but why A is attractive, what it would be right for, and what in the vignette rules it out.
If you answer a question correctly for the wrong reason, the explanation is the only place you will ever find out. A bank without full rationales is a scoring machine, not a teacher, and its question count is beside the point.
How to check a book before you buy it
Four questions, and every one is answerable from the product page:
- How many questions, and against how many statements? Divide. Under five per statement, treat the bank as a diagnostic rather than a course of study.
- Is every wrong option explained, or only the right one? Look at a sample. If the sample does not show you a full rationale, assume there is not one.
- Which blueprint is it written to? The outline changed on 3 August 2026. A guide written to the old one is not wrong so much as differently weighted — see what actually changed.
- Can you see a real extract before paying? Not a table of contents. An actual lesson with its actual questions.
Where the guides on this site sit
We publish these figures because the argument above is only fair if we apply it to ourselves.
| Guide | Statements | Questions | Per statement |
|---|---|---|---|
| Bachelors | 110 | 1,808 | 16.4 |
| Masters | 109 | 1,779 | 16.3 |
| Advanced Generalist | 115 | 1,824 | 15.9 |
| Clinical | 116 | 1,828 | 15.8 |
Every question carries a full rationale covering every option. Every lesson corresponds to one applied knowledge statement, so nothing is padding and nothing on the outline is missing.
For comparison, and to be fair to the best-known alternative: Dawn Apgar's guides, published by Springer and updated for the 2026 blueprint, advertise 500+ questions with more than 150 of those online only. They are the books most social workers have heard of, they carry a reputation ours does not, and if you want a name people trust they are the obvious choice. On questions per statement, the arithmetic above is the arithmetic above.
The honest caveat
More questions is not automatically better. A thousand bad questions are worse than three hundred good ones, and volume can hide a lack of care as easily as it can show depth.
Which is why the only test that settles it is reading some. Every guide here has a free sample — a complete lesson and every question that follows it, with the full explanations, exactly as they appear in the book. Read one and judge the writing, not the number.
Then divide.