Abstract: Does sensory measurement deserve the label of measurement? I argue that it does. First, it complies with an epistemological view of measurement held in current philosophy of science. Sensory measurement faces the same epistemological challenges as ‘nomic measurement’ (e.g., temperature): the problem of coordination and the problem of circularity. These problems are resolved by similar procedures of ‘epistemic iteration’, i.e., a continually progressing coordination between theoretical concepts and their empirical basis. Second, I address the apparently insufficient reliability of sensory measurement. I argue for separating the problem of standardization from the problem of coordination. The problem of standardization characterizes the ambiguity of many sensory performance tests, while the problem of coordination relates to the correlation of theoretical concepts with an observational grounding in the measurement of sensory perception.
To exemplify my claims, I draw on olfactory performance tests, especially studies linking olfactory decline to neurodegenerative disorders. Changes in smell perception are a first symptom and a potential diagnostic tool for the pre-clinical detection of major neurodegenerative disorders. Differences in the course of hyposmias (reduced ability to smell) may also aid in differentiating disorders with similar clinical symptoms such as Parkinson’s, Lewy bodies and Alzheimer’s. A basic requirement for comparing abilities and differences of perception between healthy and ill test subjects is the design of standardized identification sets of test odors, and a reliable way to assess human sensory performances in identifying and discriminating odor qualities. Nonetheless, the measurement of odor perceptions in humans is distinctly difficult. Little agreement exists in how to best measure odor quality and how to deal with the apparent impossibility of eliminating bias in human sensory performances.
Distinguishing between the problem of standardization and the problem of coordination will prove a starting point for further analysis of measurement techniques in psychophysical studies. The practitioners’ concerns are primarily methodological; however, their methodological concerns comprise three different epistemological issues that can be subsumed under the following three questions:
1. First, how can we link changes in qualitative sensory experiences to indicate and distinguish neurological disorders (coordination)?
2. Second, given the inherent variability of sensory experiences, how can we compare different tests and results (standardization)?
3. And, third, how reliable are our measurements, i.e., how objective are our strategies addressing the first two questions (reliability)?
To address the epistemological issues underlying these questions, I take a closer look at the particular design and development of different olfactory performance tests: the University of Pennsylvania Smell Identification Kit (UPSIT), largely used in North America, and the Sniffin’ Sticks, predominantly used in central Europe. In fact, an extensive study of olfactory loss found that the choices of olfactory test kits produce different research outcomes. But what do such differences in measurement actually mean? Looking at their methodological design, studies of sensory performances make clear that when we analyze human test subjects as measurement instruments and the laboratory setting as measurement apparatus, the separation of standardization and coordination issues helps pinpointing sources of defect measurement.