In the 1930s, Dr. Anna Julia Cooper, renowned educator and foremother of Black feminist thought, responded to a Survey of Negro College Graduates. Cooper's response to one question about her "racial philosophy" exceeded the limits of the form and was later published as an essay, removed from its original context. In Fall 2020, Elisa Beshero-Bondar's undergraduate class in Text Encoding worked on the challenge of modeling in TEI the entire survey, its typescript and the handwritten input, with guidance from researchers who digitized Cooper's collection of papers at Howard University (see https://dh.howard.edu/ajcooper). We hoped to represent how a distinctive manuscript was composed on a printed, circulated document, and found this interesting and complicated to model in TEI. We came up with a solution suitable for the semester project, which involved students' learning much code for the first time: TEI, XSLT to transform their TEI to HTML, and CSS highlighting to guide the reading experience. See https://alicer98.github.io/DIGIT-110-AJC-Survey/document.html for our coauthor's digital edition, designed purposefully to look like an interactive PDF form. For the TEI conference, we presented this edition's TEI data model and discussed with the TEI community some alternatives for modeling documents like this, in which the encounter with a primary source gains interest from describing and even documenting tensions between survey questions and respondent answers. We are curious about the possibility of representing a survey as a dialogue between print and manuscript hands. And we are curious about the notion of documenting by how much a survey respondent exceeds the spaces provided on a form to reply to a question. What is the best TEI modeling for a historically significant survey form, with a historically significant response?