Bringing ethnomethodology to thinking-aloud research protocols
A compact reading list for applying ethnomethodology, the study of how people make sense of activity, to the practice of thinking-aloud research protocols.
21 entries with this tag.
A compact reading list for applying ethnomethodology, the study of how people make sense of activity, to the practice of thinking-aloud research protocols.
Claude walks through the real statutes blocking mass housing in a national park, then designs a scaled-down 15,000-unit plan with transit and a marquee tower.
A precise legal breakdown of what statute would need to change to bar Chinese state broadcasters from the US, and why courts keep it narrower than expected.
A sprawling design session builds a microfinance-based electrification program, then folds in lessons from military bases, Ukraine, Gaza, and refugee camps.
A request for fifteen grouped survey questions about an online intellectual community turns into a full research design, with predicted top answers for each.
A close legal read of whether an ex-EU Parliament aide's move to Anthropic, the AI Act's chief lobbying target, triggers conflict-of-interest disclosure rules.
Tests a blogger's harsh verdict on a well-known tech journalist against his own interview transcripts, and finds the evidence backs the criticism.
A researcher tests a fact-checking prompt on a Dutch political claim, and Claude finds the real number was higher but the credit given was misplaced.
Claude checks Freakonomics' famous claims against primary sources, finding the '7 million missing children' fraud story and others were oversold.
Claude walks through a list of nine blunt axioms about human behavior, showing how the post uses each one on the reader while explaining what it means.
Claude reads a Chinese state-TV screenshot of a cross-region compute map and unpacks it as propaganda for a domestic-chip AI network spanning thousands of km.
A rendered markdown proposal arguing for caps on corporate ownership and for redirecting profit from institutions built on public infrastructure and labor.
The user corrects the assistant's sloppy reading of a prediction market twice, forcing precise, cited answers about the 118th Congress's real 2024 actions.
A form-based calculator that scores Singapore PR chances from inputs like age, education, occupation, salary, years in Singapore, marital status, and property.
A club officer works out how to fit bylaws-mandated quorum, treasurer, and membership business into half an hour using parliamentary shortcuts and time limits.
Asked why governments are more efficient, Claude pushes back on the premise, then explains where public provision actually beats markets, and where it doesn't.
A breakdown of which countries shield vaccine makers from lawsuits versus the more common no-fault programs that compensate injury without assigning blame.
Terms of service hand you the output, but copyright law wants a human author, so the protection you end up with depends on how much you reshaped the draft.
A proposed state bill taxing a single federal payout at full value gets tested against two doctrines, and the older one turns out to be the cleaner kill.
The user and Claude test whether an economy where AI outperforms humans at most jobs can function, probing UBI, ownership, and post-scarcity models.
The user claims a mockingbird secretly co-wrote Jay-Z lyrics; the assistant responds only with pointed questions that model how to evaluate outlandish claims.
We use cookies for anonymous analytics.