PC1 on a vaping research plot is a derived component, not a health score
Identify PCA scores, loadings and explained variance without turning a vaping research component into a personal health score.
No Vape · Updated

Find the inputs behind the component names
A label such as PC1 does not name a chemical, symptom or kind of person. Start with the measurements that entered the analysis and the preparation described by the authors. NIST’s original statistical handbook explains PCA as dimension reduction using linear combinations of variables to capture variation. The result depends on the actual analyzed data.
A real vaping-related metabolomics paper used PCA for an exploratory display of measured signals, with its caption separately identifying participant samples and pooled quality-control injections. That example establishes why the point legend matters. We import no chemical finding, biological interpretation, exposure estimate or health conclusion from that paper.
NIST/SEMATECH e-Handbook of Statistical Methods: 6.5.5. Principal Components · Hsiao and colleagues / Chemical Research in Toxicology: Untargeted Metabolomics to Characterize the Urinary Chemical Landscape of E-cigarette Users
Keep scores and loadings in different boxes
Jolliffe and Cadima’s methodological account distinguishes component scores from loadings. Scores locate analyzed observations on a component; loadings describe the original variables’ contributions to that component. A plot of sample scores and a display of variable loadings therefore answer different questions, even when both carry PC1 and PC2 labels.
Make two boxes in your reading note: observations represented and original variables contributing. Copy the figure’s definitions into the appropriate box. Do not call every mark a participant or every arrow a measured change. If the publication combines scores and loadings in one display, retain the documented meaning of each rather than reading them all as the same kind of quantity.
Read explained variance as a data-summary statement
The component order concerns captured variation, not an automatic ranking of health importance. If a component has an explained-variance percentage, keep that label intact. It is not the percentage of people who quit, the proportion diagnosed or an estimate of how much harm a product causes. A two-component display also does not show every dimension of the original measurements.
In a fictional example, a caption reports how much variation two components retain, but a repost relabels that percentage as the share of participants with a problem. Your correction restores the quantity being summarized and leaves the clinical question unestablished. No real percentage, risk threshold or research result is supplied by the example.
NIST/SEMATECH e-Handbook of Statistical Methods: 6.5.5. Principal Components · Jolliffe and Cadima / Philosophical Transactions of the Royal Society A: Principal component analysis: a review and recent developments
Do not assign yourself a position from a screenshot
Visible grouping can prompt a research question without creating permanent categories of people. Likewise, nearby points in the displayed components do not establish identical original measurements or identical health. Interpreting a particular analysis requires its inputs, processing, retained components and full methods; the plot alone is insufficient.
Finish with a component map that names the inputs, point unit, scores or loadings, and displayed variance. This is a reading aid, not a new PCA, instrument validation or biological assessment. Your own quitting record can describe an observed event without borrowing a research component or locating you among study points.
Sources and scope
Sources inform the guidance. The worksheets and examples are No Vape editorial suggestions, not tested treatments.
- NIST/SEMATECH e-Handbook of Statistical Methods — 6.5.5. Principal Components
- Jolliffe and Cadima / Philosophical Transactions of the Royal Society A — Principal component analysis: a review and recent developments
- Hsiao and colleagues / Chemical Research in Toxicology — Untargeted Metabolomics to Characterize the Urinary Chemical Landscape of E-cigarette Users
- No Vape — No Vape — app features and store links
Written with AI assistance and checked against the cited sources. No clinical review is claimed. This guide covers practical planning, not individual medical treatment. How we research and write