Consistency of perceiving odors: Inter- and Intra- Individual Differences in Odor Similarity Ratings

This study conducted odor pair similarity ratings twice and used the replicability of the ratings, as indicated by the correlation coefficients between the two sets of ratings, to examine the consistency of odor perception. The results showed that the correlation coefficient between the two ratings of the participants was generally low, and many participants did not rate the odor pairs’ similarly between the two sessions. Therefore, we concluded that participants did not consistently perceive the similarity of an odor quality when making repeated ratings, indicating low consistency in odor perception. In particular, participants showed more errors of rating, such as rating different odor pairs as "similar" in the second session. The lower correlation coefficient between the two ratings suggests that reduced carefulness in the rating due to familiarity may have caused the low rating consistency. Despite the low individual correlation coefficient between the two ratings, we observed a high mean correlation between the two ratings for all the experimental participants. Furthermore, the influence of outliers and errors of rating disappeared when we calculated the ratings of approximately seven participants in Experiments 1 and 3. In addition, despite significant individual differences in ratings, the mean similarity ratings for odor pairs with no overlapping components between them were low, whereas the similarity ratings for overlapping pairs were high. These findings indicate the possibility of perceiving the general characteristics of odor pairs to be evaluated, but not the subtle details.