We tested the reliability of sensory evaluations of tactile sensation on bare skin and investigated the reliability among evaluation attributes by trained and untrained assessors. Two trained professional panelists and two untrained researchers evaluated skin in terms of several attributes: smooth–rough, elastic–not elastic, soft–hard (surface), soft–hard (base), moisturized–dry. Twenty-two women aged 25–57 years were evaluated, and the sensory evaluation was repeated twice. Correlation coefficients and intraclass correlation coefficients (ICCs) were used to examine intra- and inter-assessor reliability. The sensory evaluation and physical quantities acquired by commercial and non-commercial instruments were moderately correlated. Smooth–rough and elastic–not elastic showed high or moderate inter-assessor reliabilities with mean correlation coefficients between panelists of 0.81 and 0.58, respectively. Further, the ICC (2,1) values were 0.64 and 0.51, respectively, and the ICC (2,2) values were 0.77 and 0.67, respectively. Conversely, the reliabilities of soft–hard (surface), soft–hard (base), and moisturized–dry were low; the mean correlation coefficients between the panelists were 0.36, 0.23, and 0.22; the ICC (2,1) values were 0.27, 0.23, and 0.17; and the ICC (2,2) values were 0.42, 0.29, and 0.26, respectively. Reliability differed between attributes. We found no meaningful differences between the trained and untrained panelists regarding intra- or inter-assessor reliability.
Read full abstract