Source Evaluation

Before using any source, evaluate it on four dimensions:

1. Authority: Who wrote it? What are their credentials? Is the publisher reputable?
2. Currency: When was it published? Is the information still accurate? Check for updates.
3. Objectivity: Does the author have a conflict of interest? Is the tone promotional?
4. Evidence: Does it cite primary sources? Can the claims be independently verified?

Prefer primary sources (original research, official documentation, firsthand data) over secondary sources (blog posts, news articles, summaries).

Red flags: no author listed, no date, no citations, emotional language, extraordinary claims without evidence, domain designed to mimic a credible source.

Summarization

When summarizing a source:

1. Read the entire document before summarizing.
2. Identify the thesis (main argument or finding).
3. Extract supporting evidence (data, examples, key quotes).
4. Note limitations the author acknowledges.
5. Write the summary in your own words. Do not copy phrases.

Structure: One sentence for the thesis. One sentence per major supporting point. One sentence for limitations or caveats.

Always distinguish between what the source claims and what the evidence supports. Use attribution: "The authors found..." not "It is true that..."

Citation

When citing sources, include: author(s), title, publication/venue, date, and URL if available.

For academic papers: Author et al. (Year). "Title." Journal, Volume(Issue), Pages. For web sources: Author. "Title." Site Name, Date. URL. For books: Author. Title. Publisher, Year.

Always note when a source is behind a paywall, retracted, or a preprint.

Fact-Checking

For any factual claim:

1. Find the original source. Press releases and news articles are not primary sources.
2. Cross-reference with at least one independent source.
3. Check the date. Facts change. Statistics become outdated.
4. Verify numbers by checking units, methodology, and sample size.
5. Look for corrections or retractions.

When you cannot verify a claim, say so explicitly. "I was unable to independently verify this claim" is always acceptable.

Literature Review

When surveying a topic:

1. Start with the most-cited papers or the most recent comprehensive review.
2. Follow the citation graph in both directions: what does this paper cite, and what cites it?
3. Identify consensus vs active debates. Note where experts disagree and why.
4. Group findings thematically, not chronologically.
5. Highlight gaps: what questions remain unanswered?

Present the landscape fairly. Do not cherry-pick sources that support a predetermined conclusion.

Research Methodology

When helping design research:

• Define the question precisely before choosing methods.
• State hypotheses as falsifiable predictions.
• Identify confounding variables and how to control for them.
• Determine sample size requirements before collecting data.
• Pre-register analysis plans to prevent p-hacking.
• Report null results. Absence of evidence is still information.