Michal Kozak: On the conception of human subject in subjective survey data
This paper is meant to be a critical reflection of certain epistemological problems related to analysis and interpretation of so called subjective survey data, i.e. all the quantitative data referring to respondents’ subjective mental worlds, states and representations (e.g. values, attitudes, opinions or satisfactions). I argue that while specificity and difficulty of subjective data is routinely recognized in most of the survey research, its true philosophical sources and implications remain still relatively unexplored. The main message of the paper is that fundamental problems of subjective data are the implicit ontological assumptions made about their authors, the respondents themselves. To analyze linguistic self-reports about respondents’ subjectivity as facts, a perfect cognitive ability of respondents to contemplate on and communicate such content has to be assumed at first place. While this approach renders the factual analysis of subjective data methodologically possible, it also leads to empirical results which show that such data are highly imprecise, at least when subjected to the same technical criteria as data that are of factual nature (e.g. validity, reliability etc.). In other words, subjective data are said to contain an enormous share of measurement error and it is therefore suggested to either meticulously refine their measurement before the data collection process, or to avoid usage of such data whenever it is possible.
Read more in the attached paper.