Methodology

The thesis methodology describes the design of the studies the thesis reports. This page describes the programme-level method: how a question becomes a hypothesis, a hypothesis becomes an experiment, and a result becomes a claim.

The loop

Rules the loop is held to

1. A hypothesis is written before the experiment that tests it runs. The EXP- record names the HYP- it tests, and the hypothesis page is dated earlier. Reordering these is how a post-hoc observation becomes a false prediction.
2. A failed run is recorded, not deleted. Its status becomes failed or invalidated, with the reason. The set of things tried is part of the result.
3. A number that appears in prose exists in a registered artifact. If it cannot be pointed at, it cannot be written.
4. A claim names its evidence. Claims with no evidence are conjectures and are labelled as such.

Evidence standards

Uncertainty

Every reported quantity carries an uncertainty. A single-run number with no spread is reported as a single run, never as a measurement.