The Google Books corpus has strong limits for cultural inference

Evidence

Pechenick, Eitan Adam, Christopher M. Danforth, and Peter Sheridan Dodds. “Characterizing the Google Books Corpus: Strong Limits to Inferences of Socio-cultural and Linguistic Evolution.” PLoS ONE 10, no. 10 (2015): e0137041. — Andrew deWaard, Derivative Media, 1369

Categorization

• SHARE: For someone doing quantitative cultural analysis: this is the canonical caveat paper on using Google Ngram data, warding against naive frequency-based causal claims.