How we would report a pass rate
Pass-Rate Methodology
We do not publish a pass rate yet. This page explains exactly what we would have to see before we did, and what the number would and would not mean.
Current status
No published pass rate
Freehold has not collected enough reported outcomes to state a pass rate. Until the sample reaches
385 reported outcomes for a track, the figure is not computed at all — not
internally, not privately, and not through our API, which returns null for the rate below
that threshold.
The rules, set in advance
Committing to these before there is a number is the point — it removes the option of choosing a flattering definition later.
- Minimum sample. 385 reported outcomes for a track before any figure is published. That is the sample that gives roughly a ±5 point margin at 95% confidence.
- Everyone who reports is counted. Passes and failures both increment. There is no mechanism to drop a result, because outcomes are stored as two counters with nothing attached to identify or revisit an individual report.
- No filtering by engagement. We do not restrict the figure to people who studied "enough". Selecting the diligent subset is the most common way a prep pass rate is inflated.
- The sample size is published with the number. A rate without an n is not a claim, it is decoration.
- The self-selection caveat is published with the number. See below.
What the number will not mean
People who pass an exam are more likely to report back than people who do not. That makes any self-reported pass rate an upper bound, not a measurement of how well the app works. It also cannot show causation: someone who passes may have used three other resources. A pass rate reported this way is evidence that people who use Freehold pass — not evidence that Freehold is why they passed. We will print that alongside the figure, in the same size type.
What is stored
Two integers per track per month: a passed count and a failed count. No account, no device identifier, no IP address, no timestamp finer than the month, and no link to your practice history. Two people reporting the same outcome in the same month are indistinguishable in storage, so there is nothing to re-identify later.
Questions about this number
What is Freehold's pass rate?
We do not have one yet, and we will not print a number until at least 385 people have reported an outcome. Below that sample a single extra pass or fail moves the figure by more than a percentage point, which is not a number anyone should make a decision on.
Why do other apps publish a pass rate and you don't?
Published rates in this category generally arrive without a sample size, a definition of who was counted, or a way to check them. We would rather show the method first and the number second.
How is an outcome counted?
If you tell us how your exam went, the site increments one of two counters for your track in the current month: passed, or failed. No identifier of any kind is stored alongside it.
Doesn't self-reporting bias the result?
Yes. People who pass are more likely to report than people who do not, so any figure produced this way is an upper bound, not a neutral measurement. That caveat will be printed next to the number itself, not buried here.