The glossary

Every term The Companion has named, kept in one place. Terms are linked where they appear in an issue, so nothing gets defined twice in the letter. The list grows every week.

  • The paper’s own term for what it says happens when a tool does the thinking: the output arrives fine, and the capability behind it goes unexercised. The study measured brain activity and recall. The debt is the authors’ interpretation of what those measurements mean. Entered in Issue 001.

  • How strongly different regions of the brain fire together. An EEG measures it at the scalp. It is a measure of activity, and reading it as a measure of effort is an interpretation the researcher has to earn. Entered in Issue 001.

  • Two things moving together, versus one of them making the other move. A study that only watches can show the first. Only a design that intervenes can earn the second. This is the oldest question in the toolkit and it does most of the work. Entered in Issue 001.

  • A paper posted publicly before peer review. A real result and an unchecked one. Reason to read it carefully, and not reason to dismiss it. Entered in Issue 001.

  • A study that sorts people into groups by chance, so the groups differ only in the thing being tested. It is the design that earns the right to talk about cause. Entered in Issue 001.

  • Whether a study has enough participants to detect the effect it is looking for. An underpowered study can still be right. It cannot show that it was right rather than lucky. Entered in Issue 001.