Inside Kesler's student data · Study 01

Is Hammering More MCQs Enough to Pass the CPA Exam?

The myth: if you just hammer enough multiple-choice questions (MCQs), you'll be ready for the CPA exam.

It's an appealing plan. You have a question bank, a counter, and the ability to sacrifice another perfectly good evening. More questions feels like a clear path forward.

But what happens when you compare passing and non-passing study records? In a selected sample of Kesler students who reported exam results, the passing group initially averaged far more MCQ answers. But the question-count comparison changed sharply when we excluded records with fewer than 200 MCQs answered. The average counts became almost identical. Accuracy and structured-practice use still differed.

Here's what that changes: your question total can't be the whole study plan. The better question is what you do with the mistakes those questions reveal.

We investigated 50 selected section results from 34 students who reported outcomes: 37 passes and 13 non-passes. These are not all Kesler students or all students who passed with Kesler. The comparisons describe this sample; the study routine below is our recommendation, not a proven cause of its results. Methods and source details.

A Thousand Answers Can Still Leave You with the Same Problem.

In the sample with no MCQ minimum, passing records averaged 942.9 answers versus 542.2 for non-passing records. “Do more questions” seems like the obvious advice.

Then we excluded records with fewer than 200 answers. The passing average became 1,114.9; the non-passing average was 1,099.5. Almost the same.

Average Recorded MCQ Answers per Section Result

No MCQ Minimum

37 passing / 13 non-passing section results

Reported pass942.9
Reported non-pass542.2

200+ Recorded MCQs

31 passing / 6 non-passing section results

Reported pass1,114.9
Reported non-pass1,099.5
Both panels use the same zero-based scale. The 200+ view is a subset of the first panel, not a second study. Source: Kesler reporting snapshot, September 29, 2026.

The middle values were still different: medians of 991 and 496.5, respectively. Larger histories pulled the non-passing average upward. This doesn't make practice volume irrelevant. It means an impressive total alone doesn't tell you whether you've addressed your weak areas. Two hundred is an analysis filter, not a recommended workload.

Accuracy provides another clue. In that subset, passing records averaged 71.2% MCQ accuracy, compared with 62.9% for non-passing records. Both groups were answering questions; the passing group was getting a greater share correct on average.

Your decision: if the same concept keeps showing up in your misses, review it before adding another batch of attempts. A rising question counter is not evidence that this particular gap has closed. The repair needs its own check.

The Strongest Practice-pattern Difference Was in Tasks, Loops, and Funnels.

Among the 31 passing section records with 200+ MCQs, study tasks appeared in 27, study loops in 21, and study funnels in 23. Each appeared in only one of the six non-passing records. It was the same non-passing record for all three tools.

This is the observation we'd pay attention to as a candidate: structured practice was much more common in the passing histories we reviewed. It gives you a specific part of your routine to inspect beyond “How many questions did I do?”

Which Types of Practice Appear in the 200+ MCQ Records?

Recorded use within each outcome group. Blue: reported pass (31 results). Striped gray: reported non-pass (6 results). Every bar runs from 0% to 100%.

Study Tasks

Reported pass27/31 · 87.1%
Reported non-pass1/6 · 16.7%

Study Loops

Reported pass21/31 · 67.7%
Reported non-pass1/6 · 16.7%

Study Funnels / Retakes

Reported pass23/31 · 74.2%
Reported non-pass1/6 · 16.7%

Custom Quizzes

Reported pass21/31 · 67.7%
Reported non-pass5/6 · 83.3%

Smart Practice

Reported pass19/31 · 61.3%
Reported non-pass6/6 · 100.0%

Games

Reported pass14/31 · 45.2%
Reported non-pass4/6 · 66.7%

Mock MCQ Testlets

Reported pass13/31 · 41.9%
Reported non-pass4/6 · 66.7%

Other Classified Quizzes

Reported pass3/31 · 9.7%
Reported non-pass1/6 · 16.7%
A section can appear in several rows. These are usage proportions, not pass rates among feature users. The same single non-passing section supplies all three task, loop, and funnel observations. No recorded use is not proof of no use.

The complete chart also shows that custom quizzes, Smart Practice, games, and mock MCQ testlets appeared in a larger proportion of non-passing records. These are usage comparisons within outcome groups, not evidence that one feature causes passing or failing. Starting knowledge, section mix, other courses, and timing weren't controlled.

So Should You Stop Doing Custom Quizzes or Games?

No. The useful change is to make them serve a study purpose. A custom quiz can check a specific weak topic or mix it back into broader practice. Games give you another practice format. Neither needs to carry your entire study plan.

Games also show why “used it” isn't enough information: the passing group logged 1,100 game answers; the non-passing group logged 34. The latter had a higher proportion of records with some game use, but much less recorded game activity.

Our interpretation is to build a routine that includes learning, follow-up, and broader application, then choose the tools for those jobs. Clicking every feature once won't do it. This is CPA prep, not a software scavenger hunt.

See the complete answer mix and accuracy tables
200+ subset: recorded answers and share of each group's answers
Practice type Reported pass Reported non-pass
Study Tasks 10,287
(29.8%)
654
(9.9%)
Study Loops 3,770
(10.9%)
172
(2.6%)
Study Funnels / Retakes 5,037
(14.6%)
169
(2.6%)
Custom Quizzes 8,109
(23.5%)
1,122
(17.0%)
Smart Practice 3,493
(10.1%)
3,854
(58.4%)
Games 1,100
(3.2%)
34
(0.5%)
Mock MCQ Testlets 2,726
(7.9%)
584
(8.9%)
Other Classified Quizzes 40
(0.1%)
8
(0.1%)
Total 34,562 6,597

Shares pool answers, so larger histories count more. These are not study-time allocations or a recommended mix. Rounding can prevent totals from reaching exactly 100%.

200+ subset: mean section accuracy; n = scored section results
Practice type Reported pass Reported non-pass
Study Tasks 72.9%
n = 27
61.6%
n = 1
Study Loops 71.1%
n = 21
59.3%
n = 1
Study Funnels / Retakes 68.5%
n = 23
55.0%
n = 1
Custom Quizzes 70.3%
n = 21
52.6%
n = 5
Smart Practice 65.9%
n = 19
64.2%
n = 6
Games 53.6%
n = 14
24.6%
n = 4
Mock MCQ Testlets 73.7%
n = 13
60.2%
n = 4
Other Classified Quizzes 66.7%
n = 3
87.5%
n = 1

These are different groups and question histories, not before-and-after improvements. Small rows can be volatile: “other” non-passing accuracy represents eight answers in one section. Quiz difficulty and topics aren't held constant.

Use Your Mock Exam to Write the Repair List.

The latest completed full-mock averages were 77.6% for 13 passing section records and 75.7% for two non-passing records in the 200+ subset. A reassuring mock percentage appeared alongside a reported non-pass. With only two non-passing cases, this is an observation, not a prediction model.

After your next mock, write down your three most useful follow-up tasks. For example: revisit a misunderstood rule, practice extracting information from an exhibit, and rehearse a section you didn't finish. Then schedule the work. Posting the score in a “was I ready?” thread is optional; fixing the identified gaps is the part you can act on.

A full mock includes MCQ testlets and sims. The usage chart's “mock MCQ testlets” can include partial practice and does not mean a full mock was completed. See Kesler's practice-exam format.

Here's How to Change Your Next Week of Studying.

Keep your existing study schedule. Replace one unplanned question grind with this sequence. The timing below is a suggested way to organize the work, not a schedule tested by this dataset.

  1. First session: find and repair one recurring mistake. Open a recent quiz. Include questions you guessed correctly, not just the red ones. Choose a concept that keeps causing trouble. Work through its explanation, then close it and write the rule in your own words. In Kesler, use the question's Learn N GO MCQ Explanation. Your output is a specific error and its correction—not “review FAR.”
  2. Next session: check it without the explanation open. Use follow-up practice, such as a study loop or a targeted custom quiz. Use a different question on the concept where available. Explain why your answer fits and why the alternatives don't. If you still need the explanation, keep the topic on your repair list. Recognizing “C” from yesterday is not the check you're after.
  3. Later in the week: put the topic back among other topics. Use mixed practice to see whether you can identify when that rule applies without the topic label giving it away. If you can do a focused set but lose the thread in mixed questions, make choosing the right rule your next task.
  4. At week's end: decide what stays on the list. For each topic, record whether you can explain the rule, apply it on another question, and recognize it in mixed practice. Keep unresolved items scheduled. Continue covering the rest of your section; repairing one weakness doesn't replace broad preparation or simulations.

Let the Mistake Choose the Next Action.

A study decision guide—not a readiness score
What you notice What to do next
You can't explain why the answer is right. Return to the explanation or lesson. Write the missing rule before another set on that concept.
You know the rule but overlooked a fact. Name the missed condition. On the next attempt, identify the facts that change the answer before choosing it.
You succeed only when you know which topic is being tested. Mix the topic with others. Explain why this rule applies instead of a competing one.
MCQs feel manageable, but sims fall apart. Practice a task-based simulation (TBS). Identify where the breakdown starts: instructions, exhibits, setup, calculation, or time.

A hypothetical example: you finish 30 MCQs and miss several questions about the same concept. Instead of immediately launching 30 more, review one explanation and write the distinction you missed. At your next session, try another question on that distinction. Later, mix it with other topics. Now you have something useful to compare: whether you can apply the rule after the original answer has left the screen.

For a fuller routine, use our guide to planning from practice results. If you're rebuilding after score release, start with the retake plan.

A Better Recommendation: Use the Misses Before Adding More Reps.

The passing group did not simply have a much bigger average question count once we applied the 200+ filter. It had higher average accuracy and more recorded use of tasks, loops, and funnels. That's the distinction this sample adds to the usual “just hammer MCQs” advice.

Replace “hammer enough MCQs and I'll be ready” with this: answer questions, use the misses to choose what to relearn, revisit it later, and check it in mixed practice.

If you're putting in the reps but repeating the mistakes, change what happens between sets before simply raising your daily question target. Your next study session already has a starting point: the concept that keeps coming back wrong.

Kesler gives you tools for that routine: Learn N GO MCQ Explanations for the question you're working through, structured practice for follow-up, and Study Funneling to narrow final review toward weak, high-priority areas. If you're already a member, start with a recent miss and put the follow-up into your plan. If your current course supports the same routine, you can do it there.

Try the Routine on a Topic You're Stuck On.

The Kesler CPA Review Free Trial provides a limited preview of current course features, with paid content behind an in-app paywall. Use it to evaluate how the explanations and study tools fit your work.

Explore the Free Trial

Sample, Sources, and Limits

Kesler product analytics and selected student-reported section outcomes, exported September 29, 2026 at 12:45 UTC. We analyzed saved snapshots, not a live dashboard. This is descriptive first-party research, not a customer pass-rate or causal-effect study.

Read the measurement and verification details
  • Selection: 50 current section results from 34 selected reporting students with paid Kesler use; 37 passes and 13 non-passes. Internal/staff accounts excluded. The nested 200+ MCQ sample has 37 results from 26 students: 31 passes and six non-passes. These samples do not enumerate all Kesler users or successful students. Outcomes are self-reported, not independently authenticated here. Recruitment and submission-incentive details weren't established in the supplied snapshots.
  • Outcome and activity: current reported section status, excluding earlier failures on sections now marked passed. Two people contribute different sections to both broader-sample groups. Activity is cumulative through the reported exam day and includes repeat answers and potentially earlier study cycles. Counts are answers, not unique questions or study hours. Recorded use does not establish a review routine; absent logging doesn't prove nonuse. Resumed games may retain another source classification.
  • Volume and accuracy: broader-sample mean MCQs 942.9/542.2 and medians 747/184, pass/non-pass. The 200+ figures are 1,114.9/1,099.5 and medians 991/496.5. Mean section accuracy is 70.5%/64.5% in the broader sample (usable n=36/11) and 71.2%/62.9% at 200+ (n=31/6). Each scored section gets equal weight in accuracy averages. Quiz-type tables use normalized analytics. No percentage is an exam-readiness cutoff.
  • Comparability: 200+ pass/non-pass section mix: AUD 8/3, FAR 8/2, ISC 1/0, REG 12/1, TCP 2/0; no BAR records. Other-course use, prior preparation, question difficulty, and timing aren't controlled. Exam months in that subset span April 2025–September 2026; the retained detail files don't establish the full broader-sample date range.
  • Full mocks: each section's latest eligible completed mock, with both MCQ and all three TBS testlets. One passing record uses approximate completion timing. Partial mock MCQ practice is a separate measure.
  • Verification: source hashes and aggregate/quiz totals matched. The complete 200+ cohort and full-mock aggregates were independently recomputed from retained details. Broader-sample medians come from the source aggregate because detailed records don't cover every one of its 50 results. No non-Kesler control group or causal estimate is available.

Editorial record: Kesler CPA Review sells the course discussed here. Analysis and drafting used AI assistance and scripted arithmetic checks; source analysis completed September 29, 2026. No independent or peer review is claimed. Individual identities and profiles aren't published. For questions or corrections, contact the Kesler team and reference Study 01.

Back to Articles