2.4. Data Analysis

The mean “recurrence score” of each of the 79 sound items of the survey, comprised between 1 (never) and 5 (daily), was obtained by averaging across the 620 participants.

Sixteen participants gave either no or unclear responses regarding the sounds they considered as stressful/fearful for their dogs (e.g., “my dog is stressed by/scared of unpredictable sounds”, “…loud sounds”, etc.). Among the 595 remaining participants, 62 responded that their dogs were negatively sensitive to absolutely no sound. Four of them indicated that their dogs had been desensitized to sounds. The 542 remaining participants (i.e., 87% of the total cohort) reported at least one sound, either listed in the survey or not, to which their dogs were negatively sensitive. We assessed the percentage of reports of negative emotional sensitivity for each (listed and freely reported) sound across these 542 participants.

For the set of 79 sound items for which recurrence scores were obtained, the relationship between recurrence score and emotional sensitivity report was evaluated using the Spearman correlation coefficient.

It is frequently assumed that dogs are more likely to exhibit negative emotional sensitivity to sounds that have very high and, conversely, very low pitch. Therefore, using Spearman correlation coefficients, we assessed the relationship between emotional sensitivity reports and two spectral parameters: F0 averaged across the whole signal duration (total of 60 sounds) and the center frequency of the 10 dB bandwidth (total of 84 sounds). R software 3.6.2 was used for statistical analysis. The two-tailed p values reported in the text and figures were adjusted using Holm’s correction for multiple comparisons [47].

The relevance of 15 environmental sounds that have been used in past academic or clinical investigations of dog/cat hearing status was examined according to the three following variables of the present study: emotional sensitivity report and the width and center frequency of the 10 dB bandwidth.