The base‑rate problem refers to the difficulty of determining how common or rare a linguistic feature or style marker is within the relevant population of writers, which is essential for assessing its evidential weight in forensic authorship attribution. Without “base‑rate knowledge” – i.e., data on the frequency of a feature in a comparable corpus – an examiner’s intuition may lead to methodological subjectivity and bias (Coulthard et al.). Large, specially designed corpora can provide the population‑level distinctiveness needed to address this problem by grounding the rarity of markers in empirical data.
Definition drawn from An Introduction to Forensic Linguistics: Language in Evidence (Coulthard, Johnson & Wright, 2nd ed.). Extracted text: /Volumes/mu-not/projects/zodiac/books/an_introduction_to_forensic_linguistics_language_in_evidence/an_introduction_to_forensic_linguistics_language_in_evidence.txt.
Bears on: authorship & stylometry · the letters · sources