What this section is
Wensa Research publishes original work on hiring, resumes and the job market. It is separate from our blog on purpose. A blog post answers a question someone searched for; the pages here are meant to be cited, argued with, and checked — which means they carry things a blog post doesn’t, and they are held to rules a blog post isn’t.
We build resume and cover letter tools, so we have a commercial interest in this subject. That is exactly why the standards below are written down and public rather than assumed.
The rules we hold ourselves to
The data-collection date is published separately from the publication date
Every study here shows when its underlying data was actually gathered, in its own field, next to the date the page went up. These are routinely conflated across this industry — a page stamped with the current year over data nobody has revisited in four years is the single most common defect our first study documented. A section that asks others to make that distinction has to display it.
Methodology is published, including the parts that weaken the finding
Sample size, how it was gathered, when, and what the limits are. If a result rests on a proxy rather than a direct measurement, we say which proxy and why. If we could not establish something, it is recorded as unestablished rather than rounded into a conclusion.
Our own errors are published alongside everyone else’s
Research that audits other people’s work and hides its own mistakes is worth nothing. Our first study lists twelve errors we made while producing it, including one where our own tooling manufactured a false consensus — the exact failure the study is about. Those stay in the piece permanently. They are not an appendix of contrition; they are evidence about how the failure works, and the most useful evidence we have is the kind we generated ourselves.
We audit our own site by the same protocol, and fix it first
Before the first study published, every numeric claim on wensaresume.com went through the identical process. It produced 38 corrections across 19 URLs, several of them to claims we had stated confidently and repeatedly. Where a claim of ours had no sound source, we removed the number rather than softening the wording.
One boundary is worth stating precisely: this covers claims about hiring, resumes and the job market. Our product performance claims — what our own users experience — are advertising, and are marked as such rather than audited as research.
Data ships with the study
Where a study produces a dataset, it is downloadable, and every figure in the write-up can be re-walked from it. A finding you cannot check is an assertion with a chart next to it.
Corrections are logged, never quietly edited
When something changes, the study’s changelog records what changed and when, and the version number moves. We do not silently revise a number and leave the page looking as though it always said that. If you find an error in our work, tell us and we will publish the correction.
Everything is reviewed at least once a year
Each study displays the date it was last genuinely re-checked — not the date the file was touched. We review annually at minimum, and sooner when a source we depend on is superseded.
That date is public so it can be used against us. If a study here says it is reviewed annually and its review date is two years old, you can see that. We would rather carry that risk than publish an unfalsifiable promise.
How to read our verdicts
When we trace a claim to its source, it gets one of seven labels. They are not synonyms for true and false — most bad numbers are neither.
| Verdict | What it means |
|---|---|
| Verified | Traces to a real primary source; the number, the population and the unit all match |
| Stale | Correctly quoted from a real source, but the data is old and is being presented as current |
| Distorted | A real source exists, but the number, the population or the unit has drifted from what it says |
| Circular | Every “source” in the chain is another aggregator; no primary source exists at the end of it |
| Orphaned | Attributed to a named study or organisation that cannot be found, or whose source no longer resolves |
| Unsourced | No attribution anywhere in the chain |
| Inconclusive | We could not settle it. Recorded as unsettled, with the steps we took, rather than forced into a bucket |
A claim is only downgraded on positive evidence. Failing to find a source quickly makes something orphaned or inconclusive — never fabricated.
And a claim is judged as it was published on the page carrying it. A correctly sourced, correctly stated figure stays verified even when other people misuse it downstream.
Differing populations and differing questions legitimately produce differing numbers. Establishing which of those you are looking at, before calling anything a contradiction, changes the answer more often than you would expect — three times in our first study, a publisher we assumed had been careless turned out to be the careful one.
Who Wensa Research is
Studies here are published under a collective byline, and every one names the people who did the work. The team is small and it is not anonymous:
Artur Lopato is a product designer and the founder of Wensa, with nearly a decade of experience designing digital products for startups. He leads product and design at Wensa, focused on making resume writing simple, clear, and effective for job seekers everywhere.
LinkedIn
Teanca Holt is a recruiting expert and digital content writer from Utah with over five years of experience helping people craft resumes that get noticed. At Wensa, she shares practical tips on writing, career growth, and landing the job you want.
LinkedIn