# Content rules for lecture notes

Rules from Ezra, starting 2026-10-16.
Each has a canonical prose rule in [`psw`](https://github.com/Morrison-Lab/psw/), possibly still in review.
This fragment only points agents to it.

## State the estimand

- **Do:** name what is being estimated or predicted.
  A prediction's estimand is the new observation's outcome, not its mean.
- **Don't:** leave the target implicit or swap the mean for the outcome.

Canonical rule: [psw#219](https://github.com/Morrison-Lab/psw/pull/219).
Tracked in [ai-config#4556](https://github.com/Morrison-Lab/ai-config/issues/4556).

## Link prerequisite definitions

- **Do:** link a prerequisite definition from its home site: math in mds, probability in pds, statistics in sds.
- **Don't:** restate a definition the prerequisite course already gives.

Canonical rule: [psw#214](https://github.com/Morrison-Lab/psw/pull/214).
Tracked in [ai-config#4563](https://github.com/Morrison-Lab/ai-config/issues/4563).

## Talk about the ideas, not the notes

- **Do:** write about the concepts themselves.
- **Don't:** refer to "the notes" or "this document" as the subject of a sentence.

Canonical rule: [psw#216](https://github.com/Morrison-Lab/psw/pull/216).
Tracked in [ai-config#4564](https://github.com/Morrison-Lab/ai-config/issues/4564).

## Qualify "mean" and "average" as population or sample

- **Do:** write "population mean" or "sample mean" (likewise "average"),
  and pMSE vs sMSE, pRMSE vs sRMSE, pMAE vs sMAE.
- **Don't:** write a bare "mean", "average", MSE, RMSE or MAE,
  unless it is completely clear from context.

Canonical rule: [psw#235](https://github.com/Morrison-Lab/psw/pull/235).
Tracked in [ai-config#4584](https://github.com/Morrison-Lab/ai-config/issues/4584).

## Include graphical methods in exploratory data analysis

- **Do:** pair the numerical summaries of an exploratory data analysis exercise or lab with plots:
  a histogram or boxplot for a distribution, a scatterplot or time plot for a trend or relationship.
- **Don't:** limit an exploratory data analysis to counts, means and correlations.

Ezra, 2026-11-04, on the logistic regression lab in lds.
No check for it exists yet, so this is a prose rule for now.
Canonical rule: [psw#478](https://github.com/Morrison-Lab/psw/pull/478).
Tracked in [ai-config#4938](https://github.com/Morrison-Lab/ai-config/issues/4938).
