The Fingerprint in the Text: How Structural Analysis Can Identify Any Newspaper

Every newspaper has a fingerprint — and it's not what you might think. Researchers at the National Superior University of Lima in Peru have discovered that the way a newspaper writes, not just what it writes about, can identify the source with near-perfect accuracy. Luciano Stucchi and colleagues found that structural patterns in how words are arranged — like how long words tend to follow each other, or how often certain word types appear — create a unique signature for each publication. The team analyzed articles from 14 Peruvian online newspapers using two different methods of converting text into numerical patterns. One method measured word length, tracking how often short words followed long words and vice versa. The other looked at how often different categories of words appeared. The researchers then ran these patterns through complexity analysis, measuring factors like how unpredictable the word sequences were and how organized the overall structure appeared. When they tested whether a computer could correctly guess which newspaper an article came from based purely on these structural patterns, the model achieved 99 percent accuracy. That means the writing style alone — the rhythm, flow, and word-choice habits — gave away the source almost every time. The researchers also found that the 14 newspapers naturally fell into three distinct groups based on their structural fingerprints, suggesting there are recognizable styles of Spanish-language journalism that share common organizational habits. What makes this finding especially interesting is what it reveals about writing itself. The patterns worked regardless of the article's topic, meaning the signature comes from how journalists at each outlet structure their sentences, not what they're reporting on. This extends similar research from English-language news into Spanish, an area that had been largely unexplored. The team believes these structural fingerprints reflect genuine habits of how each newsroom approaches writing, rather than being a quirk of their particular analysis methods. Both encoding approaches — word length and lexical frequency — produced consistent results, strengthening the case that they're measuring something real about how each newspaper operates.