Skip to content
No Vape

An E-value in vaping research is a confounding question, not a causal score

Read the estimate, confidence-limit target and unmeasured-confounding scope behind an E-value without turning it into a personal risk prediction.

No Vape · Updated

Books arranged on a shelf
Illustrative image · Pramod Tiwari / Unsplash

Name the association and the target being challenged

First locate the exposure comparison, outcome, period and reported effect measure. Then find the authors’ E-value description. A value attached to the main estimate and one attached to its confidence limit answer related but different questions. Keep their labels instead of presenting the larger number as the only sensitivity result.

VanderWeele and Ding’s original methodological paper defines the E-value through associations an unmeasured confounder would need with both exposure and outcome, conditional on measured covariates. It also discusses a confidence-limit calculation. This is a question about hypothetical confounding strength, not a measurement of a confounder discovered in the dataset.

VanderWeele and Ding, Annals of Internal Medicine (2017): Sensitivity Analysis in Observational Research: Introducing the E-Value

Retain both links in the explanation

An omitted characteristic merely being associated with the outcome does not complete the question. Its relationship with the exposure also matters. A reading note should therefore retain both links and the already measured covariates. Do not replace that joint requirement with the claim that any unmeasured variable of the reported size would necessarily erase the result.

For an invented example, a report compares two support arrangements and gives one E-value for its estimate and another for a confidence-limit target. Your worksheet has two rows, each naming its target. It leaves the hypothetical common cause unnamed unless the study discusses one. It does not invent its prevalence or claim that such a characteristic exists.

Recognize a real vaping-paper use without importing its conclusions

An original longitudinal study of US adults who smoked, published in the American Journal of Epidemiology, explicitly reported E-values for potential unmeasured confounding and separately identified confidence-limit quantities. That supplies a real example of the reporting task. We use no numerical outcome estimates, cessation findings or personal-care implications from that paper.

This differs from simply repeating an analysis under another inclusion rule. Here your output identifies the hypothetical confounding strength and which target it challenges. You are reading the authors’ calculation, not recomputing it. Other effect scales, conversions and particular study assumptions require appropriate methodological interpretation.

Cho et al., American Journal of Epidemiology 195(5):1300–1309 (2026; online2025): Association of naturalistic e-cigarette use and smoking cessation in US adults

Leave other sources of error visible

The original method does not assess measurement error, selection bias or selective reporting. A large value cannot certify that all these problems are absent. Nor does a small value prove that an association is entirely explained by an omitted factor. The calculation describes a sensitivity question; it does not observe the answer.

Finish with a bounded sentence naming the target and unmeasured-confounding scope. Leave the causal verdict open to the complete design and evidence. This reading aid cannot establish treatment suitability or turn a study statistic into your individual risk. A factual quitting record remains a different kind of information from a hypothetical bias calculation.

VanderWeele and Ding, Annals of Internal Medicine (2017): Sensitivity Analysis in Observational Research: Introducing the E-Value

Sources and scope

Sources inform the guidance. The worksheets and examples are No Vape editorial suggestions, not tested treatments.

  1. VanderWeele and Ding, Annals of Internal Medicine (2017) — Sensitivity Analysis in Observational Research: Introducing the E-Value
  2. Cho et al., American Journal of Epidemiology 195(5):1300–1309 (2026; online2025) — Association of naturalistic e-cigarette use and smoking cessation in US adults
  3. 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