Big Data logic rejects theory-driven hypothesis testing in favor of raw pattern extraction

Evidence

Some, such as Wired editor Chris Anderson, turn this opacity into a positive. Anderson argued controversially in 2008 that Big Data (the principle that says N must be all) undercuts the basic idea that historically underlay statistics as a science: that of formulating and carefully testing hypotheses to establish precisely stated correlations and, where possible, causal relations between variables. … Out with every theory of human behaviour, from linguistics to sociology. Forget taxonomy, ontology and psychology. The point is that they do it, and we can track and measure it with unprecedented fidelity. With enough data, the numbers speak for themselves. — Nick Couldry & Ulises A. Mejias, The Costs of Connection, 394

Categorization

• DEFINE: Crisp statement of the ‘N must be all’ Big Data principle and what it discards from classical statistics.
• XREF: Connects to the reader’s attention/surveillance school reading—this is the epistemological rival that data colonialism critiques. — connects to: Big Data epistemology replaces population sampling with continuous extraction from whole populations — Nick Couldry & Ulises A. Mejias, The Costs of Connection: How Data Is…, 398; Big data correlations alone cannot deliver causal knowledge to prediction systems — Shoshana Zuboff, The Age of Surveillance Capitalism, 930; Any measurable data can be rendered as sound to enable rapid pattern exploration (+6 more)
• SEED: Angles for an essay: the death of the hypothesis as a form of epistemic deskilling, and what ‘the numbers speak for themselves’ erases.

Broader Connections

• Big Data epistemology replaces population sampling with continuous extraction from whole populations #expands (0.85) — Nick Couldry & Ulises A. Mejias, The Costs of Connection, 398
• Big data correlations alone cannot deliver causal knowledge to prediction systems #qualifies (0.82) — Shoshana Zuboff, The Age of Surveillance Capitalism, 930
• Any measurable data can be rendered as sound to enable rapid pattern exploration #expands (0.75)
• Coding categories derived from prior hypotheses risk circular validation #qualifies (0.75) — John T. Jost and Jim Sidanius, Political Psychology, 521
• Seamless data flows breed opaque inequality and uncontrolled surveillance #complicates (0.72) — Nick Couldry & Ulises A. Mejias, The Costs of Connection, 607