This interdisciplinary, mixed-methods research uses non-experimental data to examine how social factors of an instructor would influence written feedback provided to FYC students by examining written feedback in the form of end comments, marginal comments, and rubrics from seven FYC instructors at a midsize midwestern public university. Using inductive coding and discourse analysis, the salient trends in the written feedback were quantified and tested for statistical significance using three grouping variables: instructor gender (i.e., male or female), instructor position (i.e., GTA or non-GTA), and the amount of time since structured instruction on providing feedback in graduate school (i.e., more than 10 years ago or less than 10 years ago). To contextualize patterns, discourse-based interviews are conducted with each instructor. Thirteen salient usage trends are discussed. These include five statistically significant gendered trends: feedback from female instructors containing a larger percentage of neutral comments and positive upgraders and feedback from male instructors containing a larger percentage of speech acts classified as directives, a larger percentage of Negative Comments and Negative Upgraders; and three statistically significant differences in usage based on instructor position: GTAs directives are issued more frequently in the form of questions and questions employing other mitigation techniques, and non-GTAs issuing directives more frequently with a hearer perspective. This research illustrates the (un)reliability of assessment practice (Clark, 2011) through feedback variations, examines the communication between people in different positions of power in higher education (Brinko, 1993), assists in filling the large gap in non-experimental data examination in composition classrooms, and helps to bring variation in instructor feedback to the level of consciousness for writing educators (Rubin, 1993). Implications for FYC programs and beyond are discussed.