I am working on the following projects, which will become the chapters of my dissertation. They are all works in progress. Comments and pointers to new literature are welcome.
Sample Representation in the Social Sciences
This project examines what it means for a sample to be “representative”, discusses the challenges associated with sample gathering in the social sciences, and proposes practical mitigations for those challenges. I begin by clarifying the distinction between a random sampling strategy and a representative sample such that an ideally random sampling strategy results in an ideally representative sample, in the limit, through the Law of Large Numbers. However, I argue that ideally random sampling is impossible to achieve with human subjects, and that any deviation from this ideal prevents the guaranteed link to sample representation. I then introduce a graded conception of sample representation, where a sample may be more or less representative within a specific sampling context, for a specific research purpose. I argue that this conception presents a much more realistic picture of sampling practice and how it may be improved. I end with a few practical proposals for such improvements.
Realism, Reliability, and Item Response Theory
This project argues the current way of measuring psychological traits or aptitudes is inadequate for providing the kind of evidence we need to draw the kind of conclusions we would like to draw from these psychological constructs. I start with a theoretical overview of the nature of traits. By looking at discussions around personality, intelligence, and well-being, I identify four important features associated with calling something a “trait”: Stability, Biological Basis, Universality, and Causal Efficacy. Second, I discuss some of the existing operationalizations used for measuring these features and criteria by which a particular measuring instrument is deemed valid or reliable. I assess the degree to which these measurement strategies can successfully produce evidence for the existence of the four features identified above. Third, I introduce a pragmatic perspective, according to which whether a trait has a certain property is important just in case we have plans to use this property of this trait for some practical goal. This picture is not meant to replace the epistemic perspective of discovering the nature of trait for the sake of psychological understanding. Instead, it is meant to provide more manageable goals in the face of measurement difficulties discussed above. I end with some reflections on what a realism means in this context as well as the kind of evidence that might decide such realism.
[scheduled for next year]
(last updated: 04-10-2019)