Lawrence (Larry) Lesser and Martin Santos (The University of Texas at El Paso)
Abstract
Background. Informed by lexical ambiguity research (e.g., JSE and SERJ papers by Jennifer Kaplan’s team) on how statistics students let everyday or even slang usage of individual words overshadow statistical usage, we noticed that students in a statistical literacy course (using Utts’ Seeing Through Statistics) often (mis)apply statistics terms to people by referring to an “average person”, a “random person”, an “outlier person”, etc. Such incoherent, inconsistent or inaccurate phrases have also been observed both in broader society as well as among even researchers with quantitative training. This IRB-exempted study was conceived and conducted during the pandemic at a mid-sized research university (a Hispanic-Serving Institution) in the southwestern United States during the pandemic, so the fall 2020 instruction and data collection all occurred online. More details are in the associated 2024 paper (co-authored by Martin Santos) in the Journal of Statistics and Data Science Education.
Methods. The data collection instrument administered within the LMS during the penultimate week of the 15-week course (after relevant concepts had been covered) was an anonymous survey with 7 multiple-choice and 4 short essay items to assess how students perceived (the meaning or validity of) phrases such as “average person” or “random person.” Of 109 enrolled students, 67% participated. Two researchers with different backgrounds independently coded the open-ended items using a priori codes (with some additional codes emerging from the data) before discussing yielded higher levels of intercoder agreement.
Findings. Almost all students found “outlier person” and “random person” to be valid phrases, but only a small majority found “average person” valid (and only 1/5 for a contextualized usage of “average pedestrian”). Degrees of student approval among statements were aligned with respective researcher orderings of “correctness” for the term anonymous but not for the term random. For the latter, the clear researcher-rated best use of the word was seen as valid by only 48% of students. An overarching theme in the data was the distinction between process/method and person/product/outcome.
Implications For Teaching and For Research. The process versus person distinction can be further explored, acknowledging student pitfalls related to variable identification, agency, undue personalization, and essentializing.
Because instances of correct usage may still conceal misconceptions, we recommend exploring relationships to performance on assessment (e.g., ARTIST) items and conducting interviews. We also recommend examining possible demographic differences, as well as additional phrases (e.g., normal person, independent person, biased person). We recommend instructional sequences that start with the statistics word and its properties and misconceptions and then use vignettes and conceptual exercises to confront the further pitfalls when it is (mis)used in a phrase about a person.