Sample content: placeholder text for layout review. Names, numbers and claims are not real.
This page describes the rules every test on this site follows. Each article adds its own specific settings in its Methodology section.
Samples
- Bought at retail with our own money unless the article states otherwise.
- Recorded with brand, batch or lot number where available, and purchase date.
- Stored sealed with desiccant and dried before printing.
Printing
- One reference printer and slicer version per test series, recorded in the article.
- Settings held constant across materials except the parameters a test is meant to compare.
- Specimen geometry from the relevant ISO or ASTM standard.
Testing
- Conditioning at 23 °C and 50% relative humidity for at least 48 hours.
- Calibrated instruments; calibration dates kept in the dataset notes.
- At least five specimens per condition; failed or invalid specimens are kept in the raw data and flagged, never silently removed.
Statistics and numbers
- We report mean, standard deviation and 95% confidence intervals.
- Every number in an article is calculated by script from the published dataset. Nothing is typed from memory.
- Charts always come with a table or the CSV containing the same numbers.
Use of AI
We use AI tools to help clean data, run the analysis scripts and draft text and translations. The author checks the data, the calculations and every conclusion before publication. Articles say so in their Disclosures section.
Comparisons and prices
Each price shows its source and the date it was checked. Price articles are reviewed monthly; when numbers change, the article’s “Last updated” date and changelog are updated.
Open data
Datasets are published as CSV under CC BY 4.0 at stable URLs and archived on Zenodo with a DOI where possible. You may reuse them with attribution; each article has a ready “How to cite” entry.