Lab · Method
About the Lab
One question per file, answered with numbers someone actually measured. Where nobody has measured it, the file says so.
What a Lab file is
A Lab file starts from a question people ask about cold brew, like how long to steep or whether it's really less acidic. It answers in one line, shows the numbers behind that line, explains how they were measured, and lists what the evidence doesn't cover. Then it turns the finding into a setting you can use, usually as a link that opens the calculator already set.
Files marked Literature review are built from published research. Files marked Original experiment are built from our own measurements, with the raw data published alongside.
How sources are chosen
- Peer-reviewed papers first, preferably open access, so you can read what we read.
- Industry or technical sources only when they publish their method and their numbers.
- We read the full paper, not the abstract, and check the tables against the text. When they disagree, the file says so.
- A finding is credited to the paper that measured it. If a paper repeats a claim from earlier work, we cite the earlier work.
How the numbers are handled
- Every figure in a file's Key numbers panel links to its source.
- Charts are re-plotted from values printed in a paper's text or tables. We never trace numbers off a published figure, and we never copy one.
- If a value only exists as a dot on someone else's graph, we describe the result in words instead.
- Each chart has its data table directly under it, and the caption names the table it came from.
- We paraphrase. Quotes are rare and short.
Rules for our own tests
Most papers test one setup: one ratio, one temperature, a handful of coffees. The Limits section of each file lists what they left out, and those gaps are the starting list for our own tests. An original Lab file has to meet these rules before it publishes:
- At least three replicates per condition, run on a stated date.
- Coffee and water weighed on a 0.1 g scale; temperatures logged, not assumed.
- Strength measured as total dissolved solids with a refractometer, and every instrument named on the page.
- The raw data published as a downloadable file next to the write-up.
- A link back to the literature file it follows up, and a link forward from that file.
We don't run sensory panels, so taste claims stay attributed to the studies that did.