Some newspaper distortions are not caused by missing text, but by too much text. Many historic newspapers were printed on thin, inexpensive paper that allowed ink from the reverse side of the page to show through. To a human reader, the effect is usually a minor distraction. To OCR software, however, bleed-through can be disastrous. Characters from the back page may blend with the foreground text, creating phantom words, scrambled names, and confusing article boundaries. The result is an OCR transcription that bears only a partial resemblance to what was actually printed.
Distortion Type
Bleed-Through from Reverse Pages
Why it Happens
Thin paper lets opposite-side text seep through.
Why it Matters
Bleed-through doubles text or merges two articles, scrambling OCR.
Search Strategies
- Switch to image mode; look for mirrored or doubled text.
- Cross-check with the continuation page for the same notice.
- Confirm with page layout view to isolate the correct column.
Example
A researcher locates what appears to be a death notice for Sarah Whitaker, but the OCR transcription contains random words and fragmented phrases that seem unrelated to the article. When viewing the newspaper image, the cause becomes obvious: text from the reverse side of the page is visible through the paper and overlaps portions of the notice. OCR mistakenly interpreted characters from both sides of the sheet as a single block of text. In another issue, bleed-through causes two unrelated articles to appear merged together in the OCR version, obscuring key names and dates. By studying the page image and isolating the correct column, the researcher reconstructs the original article and identifies the family members mentioned in the notice.
Key Takeaway
When a newspaper page contains more than one layer of visible text, OCR may read both at once. Sometimes the real story is hidden behind the shadow of another page.
The Newspaper Distortions Field Guides
Discover all 25 Newspaper Distortions, companion Field Notes, videos, and audio lessons in the Field Guides Collection