Half Marathon Trainer Editorial Methodology

Half Marathon Trainer publishes educational guidance for recreational runners, editorial analysis, free fixed resources, and information about a paid personalized training product. Our method is to answer a specific runner question, choose evidence appropriate to the claim, explain what the evidence cannot establish, and translate it into a decision a reader can use. We do not treat a citation as permission to sound certain.

Running research rarely produces one universal prescription. Studies differ in population, training status, outcome, duration, and how intensity is measured. An association between higher weekly volume and faster half-marathon times does not prove that increasing any runner's mileage will make that runner faster. A laboratory threshold result does not automatically define an outdoor training pace in summer heat. We preserve those distinctions because they change the decision.

This page explains how sources are selected, how causal and observational claims are separated, where AI assistance may be used, how product claims are kept apart from editorial claims, and how readers can request a correction. Readers can inspect the standard in practice across the training guide library. It is a publication standard, not a claim that every article is final or error-free.

The primary reader is an adult recreational runner preparing for a road half marathon in the United States. We use miles first and add kilometers when they improve understanding. Examples span beginners, run-walk participants, and experienced recreational runners, but they cannot represent every age, disability, environment, medical history, or performance level.

Scope controls the claim. Guidance for a healthy runner building from 20 miles per week cannot be applied unchanged to a person returning from surgery, an elite athlete, a pregnant runner, or a runner managing a condition or medication. When individual clinical context is necessary, the article should say so and stop short of an individualized answer.

The editorial team byline does not imply credentials we have not published and verified. Sources provide evidence; they do not confer the source author's credentials on the site. We prefer transparent limits over invented authority.

A meta-analysis of elite cyclists may be methodologically strong and still be an indirect source for a first-time half-marathon runner. A smaller prospective study of recreational half-marathoners may answer the population question more directly, even though it cannot prove causation. We consider design and relevance together.

The source should match the claim's scale. A public-health guideline can support the broad value of regular physical activity. It cannot justify an exact eight-week race progression. A sports-nutrition position statement can provide general endurance-fueling ranges; it cannot tell a reader with diabetes or gastrointestinal disease what to consume.

When evidence is indirect, we name the transfer. Phrases such as “in trained endurance athletes,” “observationally associated,” or “a practical starting point” are not timid writing. They prevent a narrow study from becoming a universal rule.

Disagreement is not solved by counting citations. We compare the studied population, protocol, outcome, and measurement method. Two intensity-distribution reviews can reach different practical conclusions because one counts sessions while another counts time, or because one studies elite athletes and another includes recreational runners.

When no approach clearly wins, the guide should present the decision criteria. A runner may choose pace for track repetitions and effort for a hot, hilly long run without declaring one metric universally superior. A plan may use three or five running days depending on total volume, schedule, and history.

We avoid false balance when evidence is not evenly divided. A single weak contrary source does not deserve equal weight with a coherent body of higher-quality evidence. The reason for prioritizing one conclusion should be visible in the wording or source note.

AI tools may help organize an outline, surface adjacent reader questions, detect repeated phrasing, propose structured examples, and flag missing source links. They may also assist with code that renders content. Assistance does not replace editorial responsibility for factual accuracy, source fit, safety boundaries, product facts, or final language.

AI output is not accepted as a source. A citation suggested by a model must be opened and checked against the claim. We do not invent studies, credentials, quotations, personal experience, or runner testimonials. Specific numbers need a resolvable source or must be removed or framed as an example.

The final article should not preserve generic model habits: repeated three-item lists, motivational filler, fake certainty, or the same FAQ on unrelated pages. A human-readable editorial standard—not detection avoidance—is the quality target. If AI assistance creates an error, responsibility remains with the publisher.

Published and updated dates are stored with each document. A date changes when a page receives a meaningful review—not automatically at build time. Priority updates include changed product facts, superseded guidance, broken sources, confirmed race information, and evidence that changes a practical recommendation.

Older evidence is not discarded because of age alone. A foundational position statement may remain relevant, while a recent small study may not outweigh it. We check whether the source is current for the claim and whether a later review has changed interpretation.

Corrections should identify the affected claim, replace or remove it, update connected pages when necessary, and change the document's updated date. Quietly changing a material product price or safety claim without reviewing related pages leaves the system inconsistent.

A limitation buried in the source list does not help the runner make a decision. When a study population is narrow, a measurement is self-reported, or an outcome is associated rather than tested, that constraint belongs beside the practical conclusion it limits. The page may respond by widening a range, presenting two defensible options, adding a checkpoint, or declining to give a universal number.

We distinguish uncertainty about the average effect from uncertainty about the individual. A meta-analysis can estimate an average across studies and still leave wide variation in who responds, how a recreational runner should apply the finding, or whether the result transfers to a different training history. “Supported by a review” is not equivalent to “predictable for you.”

Editorial usefulness comes from making the decision boundary explicit. The reader should know what evidence supports the recommendation, what assumption must be true for it to fit, what observation would justify changing course, and which question the article cannot answer. A precise limitation is more actionable than a final paragraph that says results may vary.

Use each guide as a structured starting point. The mileage, pace, heat, and fueling references on this methodology page illustrate how editorial claims are handled; they are not additional training prescriptions. Follow the specialist guide for the practical decision, follow the linked source when a claim matters, and use qualified professional advice for individual medical or nutrition questions. An article that says “this depends” should also explain what it depends on.

If you find a factual error, broken source, outdated product detail, overstated claim, or unclear safety boundary, report the page, exact passage, and supporting source. A useful correction is specific enough to reproduce. The editorial team should be able to explain whether the passage was corrected, clarified, or retained and why.