1 Reply Latest reply on Dec 30, 2021 9:54 AM by Lauren Algee

    Cut-off Newspaper Articles

    Diane Estes Tracker

      I have seen various methods of dealing with pages ripped from newspapers, and am hoping to get clarity from the LOC staff as to how best to handle these.

      To date these are the methods I have seen used: 

       

      1. Completely ignore the partial/ripped bit at edges, and transcribe only the relevant "core" article.

       

      2.  Transcribe partial words exactly as seen with no extrapolation -- (Made-up) example:

      "oosevelt met with Taft

      liament on Thursday

      ompers

      nt

      an"

       

      3. Transcribe words exactly as seen, but fill in whole words that can be determined from context. -- (Made-up) example:

      "Roosevelt met with Taft

      Parliament on Thursday

      Gompers

      nt

      an"

       

      4.  Insert [?} at missing edge to indicate unknown content.  This seems accurate, but also makes it appear as if there is unknown content waiting for another reviewer to diagnose and therefore incomplete - and may never move to "accepted" status. Example:

       

      https://crowd.loc.gov/campaigns/rough-rider-bull-moose-theodore-roosevelt/twilight/mss382990074/mss382990074-545/

       

      I personally do #3.  What is standard, and what should we accept?

       

      Thanks!

        • Re: Cut-off Newspaper Articles
          Lauren Algee Tracker

          This is a great question without a black and white answer. Method #2 (transcribing only the characters actually present on the page but not using ? for unknowns) probably conforms most closely to our directions. We do want to capture any information that is present and generally caution against guessing.

           

          But in practice, what I personally would do as a transcriber or reviewer really depends on *how* cut off the words are and therefore how much is actually missing and the transcriber/reviewer's confidence about the "missing" parts of letters/words. Depending on the page, I might accept something transcribed like 3 if most of the text is visible and I have a high level of confidence.

          1 person found this helpful