Resources

Reference pages to keep open while you read the weekly notes

These four pages are the course’s reference shelf. They do not introduce new material; they collect, in one place, the vocabulary and the decision logic that recur across the weekly notes, so you can look something up quickly and keep the analysis blueprint in view. That blueprint is the spine of the whole course — for every method, you walk the same six steps: (1) Question (are you comparing, explaining, or predicting?), (2) Structure (the unit of analysis; response vs explanatory/grouping/covariate; the outcome type; the design), (3) Method (the analysis that matches that structure, and why this one), (4) Assumptions & diagnostics (what it assumes and how you check), (5) Estimate & uncertainty (what the model estimates — a mean difference, an effect size, a slope, an odds ratio — reported with a confidence interval, not a bare p-value), and (6) Conclusion (statistical vs practical significance; association vs causation; what the analysis cannot support). Keep these pages open in a second tab while you read the notes and work the labs.

Two disciplines run through every one of these pages, the same two that run through every weekly note: report the estimate with its uncertainty, not just a verdict, and keep statistical significance, practical significance, and causation distinct (observational data buy you association, not causation). The resources exist to make those two habits automatic.

The four resources

• Methods glossary — the vocabulary used across the course, organized by the blueprint and by method family: the structure vocabulary (unit of analysis, response vs explanatory/grouping/ covariate, outcome type, design); estimation and uncertainty (point estimate, standard error, confidence interval, p-value, effect size, practical significance); and the method-specific terms — paired and two-sample \(t\), one-way and two-way ANOVA and the interaction, planned contrasts and family-wise error, simple and multiple (partial) regression, ANCOVA and adjusted means, the contingency table and its risk difference, relative risk, and odds ratio, and logistic regression on the log-odds scale. Start here when a term is unfamiliar.

• Method chooser (decision guide) — a walk from a question and a data structure to a defensible method. It is deliberately not a flowchart that picks “the” test; it lays out how the structure — paired vs independent, one factor vs two, a quantitative vs a categorical or binary outcome, a covariate to adjust for — points you to the matching analysis, what each one assumes, and what each one estimates, so you can choose for a purpose and say why. This is the page that keeps the course from becoming a disconnected catalog of named tests: every test here is one instance of the blueprint.

• Assumptions & diagnostics guide — step 4 of the blueprint, in one place: what each method assumes (independence, the right design, approximate normality of residuals, equal variances, parallel slopes in ANCOVA, expected counts \(\ge 5\) for chi-square, linearity on the log-odds scale for logistic regression) and how you actually check it — QQ plots, residual plots, Levene’s test, leverage and influence, VIF for multicollinearity. It carries the course rule for a surprising point: investigate, do not auto-delete.

• Reporting & interpretation guide — step 5 and step 6 of the blueprint: how to report an estimate with its uncertainty (an effect size and a confidence interval, never a lone p-value), and how to write a conclusion that keeps statistical significance, practical significance, and causation separate. This is the page for the recurring distinctions — a small p-value is not a large effect, a confidence interval that excludes zero is not a causal claim, and an observational association (students who chose the support center, students who self-selected into a program) buys association, not causation.

Read the four together and they reconstruct the blueprint end to end: the glossary names the pieces, the chooser maps a question and a structure onto a method, the assumptions guide checks that the method is fair to the data, and the reporting guide turns the output into an honest, bounded conclusion. None of them is a software manual — R and Quarto carry out the fit and produce the output, but what stays central here is the method’s logic: what is compared, what is assumed, what is estimated, and what can be concluded.

A reminder about the numbers

Every numeric value mentioned on these pages — and across the weekly notes — comes from the course’s synthetic Cypress Ridge College Student-Success datasets (the five related structures P, G, F, X, and R, all generated with set.seed(35203)) and is provisional pending review. This is a draft course site: R is not executed here, the example data are not real records, and every drafted statistic — the means and SDs, the paired and two-sample \(t\)’s and their confidence intervals, the ANOVA \(F\)-ratios and \(\eta^2\), the Tukey and contrast estimates, the regression slopes and \(R^2\), the ANCOVA adjusted means, the chi-square, the risk difference and relative risk, and the logistic odds ratios — is provisional and not independently checked. Use the resources for the vocabulary and the reasoning, not as a table of certified results.

Public vs. graded

These resource pages are public and ungraded — study reference only. For graded specifics — deadlines, submissions, and policies — Blackboard (the LMS) is authoritative.