Updated daily · last updated
Our data & methodology
What 18,351 students got right and wrong on their first try — across 102 certification and licensing exams — and exactly how we count it.
Question-level data since Aug 29, 2026 · full-exam and exam-date signals since Aug 1, 2026 · data through
Inputs
What we collect
Career Employer is free and needs no account. As people study, the page records what they do with each question and card — never who they are. These are the signals behind every number on this site:
Practice-question answers
Whether each answer was right on the first try, the time spent on the question, whether the student switched answers, and which wrong option they chose.
Flashcard grades
“Know it” or “still learning” for each card, per session.
Study-guide checkpoint answers
The quick check questions at the end of study-guide modules.
Exam dates people set
When someone sets their exam date, we learn how far ahead people start studying — the lead-time figures on each practice page.
Approximate state and local hour
Coarse location (state level) and the local hour of study. No city, no precise location.
Never: names, emails or accounts
The data has no names, no email addresses and no city. A “student” is a browser, not a person.
Cleaning
What we exclude
Repeat attempts
Only each student’s first attempt at each question counts. A question answered for the fifth time says more about memory than difficulty.
Answers after revealing the explanation
If someone peeked at the answer first, that answer is dropped.
Bots
Automated traffic flagged by our bot checks never enters the numbers.
Staff and test accounts
Our own team’s study sessions and QA runs are excluded.
Retired question versions
When we rewrite a question set, answers to the old version stop counting toward the current figures (see hardening, below).
Rigor
Sample-size rules
Every block publishes on its own minimum. Below the minimum, the block simply doesn’t appear — we never show a number built on a handful of people. Blocks fill in as data grows, so some exams show a few sections before they show the whole exam.
| We publish… | …only once it has |
|---|---|
| An exam section | 300+ first-try answers and 50+ students |
| An exam's headline accuracy | 150+ students |
| Median first full-length practice exam | 100+ students' first full exams |
| Full-length practice exam finish rate | 150+ students who started one (starts from the last 2 days excluded — they may still be mid-exam) |
| Exam-date lead time | 50+ students who set an exam date |
Versions
How hardening works
We find the easy questions
When our data shows a question set is easier than the real exam, we rewrite those questions to be harder.
The new version starts from zero
Answers to the old wording are frozen. Current figures only count answers to the question set you see today.
Sections never go dark
Until the rewritten section reaches its own minimum, it keeps the previous version's numbers — clearly labelled “data from the previous question set”.
Read this
Caveats
- Our questions, not the official exam.These are Career Employer practice questions written to each exam’s official outline. They are not the official exam, and our accuracy is not an official pass rate.
- A self-selected sample. Everyone in the data chose to study with us. They are not a random sample of test-takers.
- A browser is not a person. Someone who studies on a phone and a laptop counts as two students; a shared computer counts as one.
Updated daily
Every figure is recomputed once a day from the full history, and each number on a practice page carries its own “data through” date. The current data runs through Oct 1, 2026.
Free to reuse — CC BY 4.0
Journalists, educators and researchers may reuse this data under the Creative Commons Attribution 4.0 license. Please cite it as:
Career Employer practice-test data, careeremployer.com/data, accessed Oct 1, 2026.
Every exam
Exams with published data
102exams have at least one published block. “Full” means the headline accuracy and every weighted section are published; “partial” means some blocks are still below their minimum. Open any practice test for its sections, charts and its own CSV.
Download all data (CSV)
One row per published number — exam, section, metric, value, sample size, students, source and data-through date.
Questions about the data? Contact us · Read our editorial process.
