Electronic Cigarette Device and Liquid Characteristics: Comparison of Self-Reports to User-Provided Pictures.

INTRODUCTION

Self-reports of electronic cigarette (ECIG) device and liquid characteristics are not always accurate or consistent with characteristics as measured by researchers. Two methods for measuring ECIG characteristics were compared: user self-reports and rater-coded pictures.

METHODS

Exclusive ECIG users ( = 321) reported on device (disposable, refillable, adjustable power, brand) and liquid (nicotine concentration, formulation, flavor) characteristics. To measure device type, they chose the term that best described their device ("cig-alike," "vape pen," "mod," "pod," "don't know") and the picture that best resembled their device (cig-alike, vape pen, box mod, USB-shaped pod, teardrop-shaped pod, none). Respondents uploaded device and liquid pictures, and independent raters coded these same features. Agreement between methods was examined with Cohen's kappa and intra-class correlations, including with "don't know" responses included and excluded from analyses.

RESULTS

Regardless of how "don't know" responses were treated, agreement was highest for disposable (95.3-97.7%), refillable (96.3%), adjustable power (83.6-88.7%), and brand (77.9-80.4%), and lower for nicotine concentration (72.7%), nicotine formulation (58.6-79.4%), and flavor (66.2%). For device type, agreement was moderate using both term-based (67.9-78.8%) and picture resemblance-based (71.7%) items. For terms, the greatest discrepancy was for devices classified as "vape pens" by self-reports; of these, 70.6% were classified as "pods" by raters. For picture resemblance, ∼13% of users reported that their device resembled none of the pictures; raters classified these devices as USB-shaped pods (50.0%) and mods (23.8%).

CONCLUSIONS

Self-reports may be sufficient for measuring some characteristics (brand, disposable, refillable, adjustable power), but not others (nicotine concentration and formulation, and some flavor).