CORESTA Meeting, Smoke Science/Product Technology, 2023, Cancun, CROM 00

Science starts with measurement: essential measurement science for self-report in tobacco and nicotine product research

(1) Juul Labs Inc., Washington DC, U.S.A.; (2) Yale School of Medicine, New Haven, CT, U.S.A.; (3) Philip Morris Products S.A., Neuchatel, Switzerland; (4) Pinney Associates, Inc., Bethesda, MD, U.S.A.; (5) Altria Client Services LLC, Richmond, VA, U.S.A.

Consumer reported outcome measures (CROM) are a critical component of tobacco regulatory science. Examples of CROM include measures of tobacco product dependence and risk perceptions. Like other fields of science, behavioral researchers require instruments (i.e., CROM) that are reliable and valid for measuring a given construct (e.g., risk perceptions), to ensure accuracy and replicability of measurement, and to support the validity of study conclusions. Researchers must also consider the context in which a CROM is applied, and whether a CROM valid for use in one context is valid in another. Such concerns are the focus of psychometrics, the science of behavioral measurement.

The CORESTA CROM Task Force (TF) is charged with establishing best practices and guidelines pertaining to the use of CROM in tobacco research. This Symposium, organized by the CROM TF, began by introducing the TF’s guidelines regarding the identification, development, validation, and implementation of CROM (McCaffrey). An overview of psychometrics was provided and the importance of considering CROM psychometric functioning when used in a different context from which it was developed will be discussed (McCaffrey). To illustrate this point, two speakers (Morean and Afolalu) presented the application of CROMs to different products (e-cigarettes and IQOS) and the languages from which they were developed. Next, Black demonstrated how invariance testing can be used to evaluate CROM functioning when used with different populations and discuss the consequences when such testing reveals non-invariance. Shiffman presented an approach for evaluating the appropriateness of numerical rating scales for individuals with limited numeracy to determine whether scores from such scales are comparable across those with limited vs. adequate numeracy. The discussant (Sarkar) elaborated on the analogy between the methods used to establish inferences from clinical outcome measures compared to those used for self-reported measures.

CROM Symposium

Programme (PDF)

Papers presented