Reactions to memorable experiences of sad music were studied by means of a survey administered to a convenience (N = 1577), representative (N = 445), and quota sample (N = 414). The survey explored the reasons, mechanisms, and emotions of such experiences. Memorable experiences linked with sad music typically occurred in relation to extremely familiar music, caused intense and pleasurable experiences, which were accompanied by physiological reactions and positive mood changes in about a third of the participants. A consistent structure of reasons and emotions for these experiences was identified through exploratory and confirmatory factor analyses across the samples. Three types of sadness experiences were established, one that was genuinely negative (Grief-Stricken Sorrow) and two that were positive (Comforting Sorrow and Sweet Sorrow). Each type of emotion exhibited certain individual differences and had distinct profiles in terms of the underlying reasons, mechanisms, and elicited reactions. The prevalence of these broad types of emotional experiences suggested that positive experiences are the most frequent, but negative experiences were not uncommon in any of the samples. The findings have implications for measuring emotions induced by music and fiction in general, and call attention to the non-pleasurable aspects of these experiences.
One possibility is that studies involving music and sadness may not differentiate the emotions experienced in sufficient detail, thus conflating inherently incompatible emotional experiences that have distinct mechanisms, outcomes and reasons. Failure to address the key moderating variables such as the focus of the emotions (expressed or experienced), the emotion-induction mechanisms , (particularly the role of the episodic memories, or selection and familiarity with the music ), may have led the field to consider only a limited range of emotional responses to sad music: responses from individuals who actually enjoy and derive pleasure from it.
These questions were considered to be best explored through a structured survey, using both existing instruments and a new series of questions derived from previous findings on music and sadness. We differentiated between generic and specific questions, the latter referring to a specific memorable example that the participants are asked to specify, since this is known to provide more accurate and definitive information than generic questions . Collecting data from different samples enabled us to carry out exploratory and confirmatory factor analyses of the main conceptual structures (reasons, mechanisms, and emotions), and offered a better estimation of the prevalence and relevance of these themes for sadness in music.
The first section dealt with broad attitudes towards sad music, implemented through an existing instrument, Attitudes towards Sad Music  that assesses six different types of attitudes linked with engaging in sad music. The rationale for employing this instrument was to establish links between the attitudes towards sad music and the memorable sad music experiences.
The second section explored the importance of 24 reasons to listen to sad music, as well as the emotional functions of listening to sad music, derived from previous studies [6, 19, 22]. These 24 reasons, shown in Table 1, were provided as checklists from which the participant could choose all that applied.
Three samples were used in the present study. The first was a convenience sample from a single country (Finland), hereafter S1. The second was a representative sample of the UK population to qualify and expand the findings of the convenience sample (S2). The third sample (S3) utilised quota sampling of the Finnish population to provide a selection of listeners similar to the representative UK sample in terms of gender and amount of listening to sad music in the representative of UK sample. The purpose of this was to facilitate the comparison between the two countries. A numerical summary of the samples S1-S3 is given in S1 Table.
Representative sample (S2) To explore the reliability and generalisibility of the findings of the convenience sample, the same survey was administered to a representative sample from a different country. A stratified sample of UK citizens was taken, in which region, age and gender formed the individually controlled strata. The sample was obtained from SurveyMonkey Audience (www.surveymonkey.com/mp/audience), and consisted of 445 participants. 53.3% of the sample were women, and age range followed the UK age distribution (18-24 9.19%, 25-34 17.26%, 35-44 22.42%, 45-54 25.56%, 55-64 22.87% 65-74 2.24% 75+ 0.45%). The frequency of listening to sad music by the representative sample is shown in Supporting Information (S1 Table).
Quota sample (S3) The third sample utilised quota sampling to create a sampling in which there were equal numbers of each gender and those listening to sad music (5 categories) as in the representative UK sample. The purpose was to seek a balanced representation of men and women who would not be especially keen on music and sadness in order to compare the findings with other samples that will vary in this respect. This data collection was, of course, carried after the representative (UK) sample was made.
The 24 statements related to reasons for engaging with sad music yielded a different overall number of reasons across the samples (an average of 9.95 reasons in S1, 4.62 in S2, and 8.33 in S3). The rankings of the reasons also vary between the samples, but there is an overall agreement of the most important reasons. Within all samples, the top five reasons include To listen to music privately (#2, #1 and #1 for S1, S2 and S3 samples, respectively), closely followed by beauty of the music (#1, #2,#3), and to get comfort (#3, #4, and #5 reason across samples), and to reminisce (#5, #2,#5). It is also apparent that some reasons were infrequently chosen, including those relating to sharing emotions or choosing sad music because it produces feelings of belonging. A full list of the ranked frequency of the 24 statements related to reasons for engaging with sad music across the three samples is given in Supporting Information (S2 Table).
To explore the notions behind the factors underlying the reasons for engaging with sad music in more detail, the responses to the specific statements associated with theory-driven list of six categories of affects (pleasure, nostalgia, hurt, melancholia, comfort, and pain) are shown in S3 Table. For broadly positive emotions (pleasure, comfort, nostalgia), the most frequently nominated reasons are similar to the general reasons outlined in S2 Table, consisting of beauty, reminiscing, and relaxing. For the negative emotions (hurt, pain, and melancholia), which were generally less frequently nominated, the pattern of the most often mentioned reasons is perhaps more interesting, since it highlights the importance of personal losses and being reminded of loved ones that have passed away, emphasising that music often conveys a tragic or hopeless narrative.
To expand the reasons underlying choosing to listen to sad music, the relevance of various mechanisms responsible for music-induced emotions were explored. The list of mechanisms, though largely derived from Juslin and Västfjäll , also included additions such as beauty of the music , sharing emotions with others, and expectations of re-experiences the emotion. It is clear that the emotional expression of the music itself (#1, #2, and #1 ranked reason for S1, S2 and S3), the beauty of the music (#2, #3, and #2), and memories (#4, #1, and #3) are the most often implicated mechanisms. These have already been implicated in several studies [6, 10, 19]. Interestingly, most of the mechanisms linked with music itself such as surprising events, strong captivating rhythm or expectations about how it will unfold, seem to play a relatively minor role here (from 6 to 16 percent of the nominations across the samples), as can be seen in summary provided in S3 Table.
After the generic and unspecific questions about attitudes towards sad music and reasons involved in listening to sad music, the remainder of the questions addressed a memorable experience with sad music.
According to the responses, these experiences had typically occurred more than a year ago (41.9%), involving music chosen by the participants (71.8%) that was extremely familiar (48.7%) or at least very familiar (26.1%) to the respondents. In the situation remembered, the participants listened to sad music for a wide range of durations (from less than 5 minutes, 15.1%, to 5-10 minutes 21.0%, 10-30 minutes 23.6%, 30-60 minutes 19.0%, and more than 60 minutes 20.4%). The exact structure of the emotions experienced whilst listening to sad music in this memorable example are analysed in detail in the next section, but their emotions were typically highly intense (M = 5.56, SD = 1.20 on a scale of 1-7) and pleasurable (5.35, SD = 1.57), although these differed across samples, see Fig 1.
Other physical reactions such as dancing, being immobile, and singing were mentioned in the open comments. When participants were asked to compare how they felt mentally after listening to the music compared to how they felt before, most of the samples indicated that they felt better than before, although again, samples differed in proportion of answers given to each category (see Fig 1 for details). A related question about the changes in physical states painted a similar but more subdued picture, since a only those keen on sad music (S1) reported feeling physically better than before whereas the other samples commonly reported having similar physical state after listening as before the episode. It is also worth highlighting the proportion (11.8% and 11.7% for S2 and S3) of people reporting that their mental state actually worsens after these memorable experiences, suggesting either that these experiences are not entirely pleasant or that they may also lead to rumination . 59ce067264