Ecological Validity, External Validity, and Mundane Realism in Hearing Science.

Abstract

Tests of hearing function are typically conducted in conditions very different from those in which people need to hear and communicate. Even when test conditions are more similar, they cannot represent the diversity of situations that may be encountered by individuals in daily life. As a consequence, it is necessary to consider external validity: the extent to which findings are likely to generalize to conditions beyond those in which data are collected. External validity has long been a concern in many fields and has led to the development of theories and methods aimed at improving generalizability of laboratory findings. Within hearing science, along with related fields, efforts to address generalizability have come to focus heavily on realism: the extent to which laboratory conditions are similar to conditions found in everyday settings of interest. In fact, it seems that realism is now tacitly equated with generalizability. The term that has recently been applied to this approach by many researchers is ecological validity . Recent usage of the term ecological validity within hearing science, as well as other fields, is problematic for three related reasons: (i) it encourages the conflation of the separate concepts of realism and validity; (ii) it diverts attention from the need for methods of quantifying generalization directly; and (iii) it masks a useful longstanding definition of ecological validity within the field of ecological psychology. The definition of ecological validity first used within ecological psychology-the correlation between cues received at the peripheral nervous system and the identity of distant objects or events in the environment-is entirely different from its current usage in hearing science and many related fields. However, as part of an experimental approach known as representative design , the original concept of ecological validity can play a valuable role in facilitating generalizability. This paper will argue that separate existing terms should be used when referring to realism and generalizability, and that the definition of ecological validity provided by the Lens Model may be a valuable conceptual tool within hearing science.