Curiosity and interest for negative information

Although the topic of morbid curiosity has not received much attention, several authors have proposed insightful definitions and explanations of general curiosity. For example, Litman [10] defines curiosity as ‘a desire to know, to see or to experience, that motivates exploratory behavior directed towards the acquisition of new information’ (p.793). According to Loewenstein [11] people experience curiosity when there is an “information gap” between what they want to know and what they currently know. Curiosity is reflected in seeking out information that can reduce or resolve this discrepancy (see also [12]). Although these views leave the valence of the information that is sought unspecified, Silvia [13] has emphasized that both positive and negative stimuli can give rise to interest, a concept closely related to curiosity [14].

Although most studies investigating curiosity have focused on curiosity for positive stimuli or knowledge-oriented stimuli, such as trivia (e.g., [12, 15–17] see for an overview [11]), there are some studies that specifically targeted whether negative stimuli give rise to interest or curiosity. For example, Zuckerman and Litle [1] demonstrated a positive relationship between the personality trait of sensation seeking (i.e., the need to subject yourself to arousing and novel experiences) and self-reported curiosity for morbid events (e.g., enjoyment in watching violence or death in films, sport or news). A more recent example is the work of Turner and Silvia [18] who presented participants with images of calming art and disturbing art and asked participants for ratings of interest and pleasantness. Their findings demonstrated that there was no relationship between interest and pleasantness; people found disturbing art both unpleasant and interesting. Rimé and colleagues [3] focused on the phenomenon of “morbid fascination” and collected reports of attention and interest for pictures of the terror attacks of 9/11. According to their results, about half of the participants were characterized by their interest for the negative scenes. Some of these participants experienced an ambivalent mix of interest and specific emotional reactions (e.g., disgust or anxiety) whereas other participants experienced “cold” fascination for the negative scenes that was solely defined by interest and high attention.

Even though some work has been done on subjectively reported curiosity or interest for negative information, relatively little is known about the behavioral aspects of this phenomenon. Recently, Hsee and Ruan [19] performed a study that explicitly targeted people’s decisions to expose themselves to aversive stimuli. They found that participants showed more curiosity for aversive stimuli with an uncertain outcome, as compared to aversive stimuli or neutral stimuli with a certain outcome. This effect was found for decisions to expose oneself to electric shocks, unpleasant sounds and images of insects. Although these studies provide important new insights regarding the role of uncertainty in curiosity for negativity, this work does not allow for conclusions about choice for “morbid” stimuli that portray violence, death or harm. The present paper reports several studies that explicitly target people’s choice for stimuli with such content.

A focus on behavior

In daily life people commonly seek out negative information. The studies reported in the present paper aim to mimic this behavioral expression by implementing a paradigm that provides participants with a choice to approach or avoid stimuli that portray death, violence or harm. This operationalization connects to theoretical views that define curiosity in terms of explorative or information-seeking behavior [10, 14, 20].

The focus on behavior is important, because so far, most studies that explicitly targeted interest or curiosity for negative information examined subjective ratings (e.g., [3, 18]; see for exceptions [19, 21]) or studied the relationship between self-reported behavior and personality measures [1, 22, 23]. However, self-report and behavior do not always align [24], and it is therefore important to utilize a behavioral paradigm when testing the prevalence of curiosity for negative information. Other relevant work has focused on subjective reports of enjoyment of negative or violent information (e.g., [25–27]). Nevertheless, it has been suggested in the literature that morbid curiosity is a phenomenon characterized by “wanting” (i.e., incentive salience or approach motivation) as opposed to “liking” [10], and therefore a behavioral paradigm seems better suited to study this phenomenon.

Furthermore, it is important to distinguish curiosity from mere attention [11]. Although previous work has provided some insight into morbid curiosity by demonstrating increased viewing time for negative images over neutral images (e.g., [28–30]), viewing time does not necessarily indicate interest or curiosity, but may also reflect a state of general attentional vigilance or a state of anxiety. For example, individuals with generalized anxiety disorder or phobic individuals find it difficult to disengage their attention from stimuli relevant to their disorder, such as angry faces or pictures of spiders (see for an overview [31, 32]). Thus, enhanced attention towards negative stimuli may reflect curiosity or interest, but it can also indicate other affective processes.

Finally, most work reviewed above involved research methods in which participants were confronted with negative stimuli. Nevertheless, in the context of morbid curiosity it is important to emphasize that people are active agents. People choose, voluntarily, to view images on a website, watch documentaries, or read articles with morbid content. Therefore, the present studies do not measure how people respond to a morbid stimulus once it is there, but rather, as a more direct indicator of morbid curiosity, whether people voluntarily expose themselves to stimuli that portray violence, death or harm.

The present studies

The present paper has two aims. The first aim is to examine, in a carefully controlled experimental setting, whether people deliberately choose to view images that portray death, violence or harm when they have a non-negative alternative option that would allow them to avoid the negative images. The main hypothesis tested with regard to this aim follows from theoretical accounts that emphasize that curiosity is characterized by exploratory behavior [10, 14, 20]. Although the existence of morbid curiosity could be argued when some exploration of images portraying death, violence or harm occurs, a strong hypothesis regarding morbid curiosity would state that people prefer to explore images with such content over non-negative alternatives. This hypothesis was tested by asking participants to choose between viewing a negative image and a neutral alternative (Study 1, 2 and 3) and between viewing a negative image and a positive alternative (Study 2 and 3). To evoke an information gap [11] participants were either cued to make their choice by two briefly presented images in small format (Study 1 and 2) or by two verbal descriptions of the negative and alternative image (Study 3). When the choice was made, participants were presented with the chosen image in full screen format.

In order to advance the understanding of morbid curiosity, it is important to engage in hypothesis-generating research. Therefore, the second aim of the present paper is to explore whether different types of negative information affect curiosity in a similar (or dissimilar) way. For this purpose, the choice paradigm included three categories that reflected different portrayals of death, violence or harm in terms of contextual information, social interaction and the presence of human actors. The negative social category involved images of death, violence or harm in a social context, such as interacting victims and/or perpetrators of violence (e.g., a man beaten up in a subway), or people responding to or observing dead or wounded people (e.g., a group of people standing around a dead body). These images always displayed multiple people and emphasized interpersonal interaction. The negative physical category involved stimuli that displayed close-up, graphic portrayals of death or physical harm, such as mutilations (e.g., a severed hand) or single dead bodies (e.g., a body with a slit throat in a morgue). These images always displayed a single person or body part and did not involve interpersonal interaction. The negative nature category (only included in Study 1 and 2) involved images that displayed close-up, decontextualized images of attacking animals (e.g., attacking shark) or passive insects (e.g., a group of cockroaches). No humans were present in the negative nature images.

Although there was no theoretical basis for distinguishing between these three categories, there were two practical observations that lead to examining choice behavior for social, physical and nature images separately. The first observation was that these three categories connect to particular types of negative content that people can encounter in daily life; namely negative content with a focus on a social narrative (e.g., crime shows/documentaries, thrillers); negative content with a focus on graphic harm or gore (e.g., horror movies, online videos of harm); and negative content that involves “dangerous” animals (e.g., shark videos/ documentaries). The second observation was that stimuli from these three categories are used in affective science to target different discrete emotions [7, 33]. For example, images from the negative physical category have been used in research that specifically targeted disgust [6, 30, 34], whereas images from the negative nature category have been used in research that specifically targeted fear or threat [35, 36]. Images that display humans in negative social situations are associated in the literature with a variety of discrete negative emotions, such as sadness, disgust and fear [7] and have been used in studies that specifically target empathy [37, 38]. Considering these two observations, it was decided not to collapse across negative stimuli, but instead to measure choice behavior for social, physical and nature images separately.

Since the second aim of this paper is exploratory in nature, no specific hypotheses regarding curiosity for the three categories were formulated at the start of the project. These hypotheses were based on the findings of a pilot study and subsequently pre-registered and tested in confirmatory fashion in Study 1. This procedure is consistent with current guidelines in the literature for exploratory or hypothesis-generating research [39, 40].

Method

Procedure.

The paradigm was introduced as a study on exploring visual information. No mention of (morbid) curiosity was made. Participants were encouraged to participate seriously and to ignore social-desirable considerations in making their choice.

The choice paradigm consisted of 60 trials and was presented with the stimulus presentation software E-prime. For an overview of the choice paradigm see Fig 1. Each trial started with a fixation cross presented for 500 ms, followed by a combination of two images (128 x 96 pixels or approximately 3.5 cm x 2.5 cm) presented in the middle of the screen. To make viewing time equal for each participant, the combination was presented for 2000 ms. Previous research has shown that a short exposure time of 250 ms is sufficient to categorize the valence of IAPS images [46]. Furthermore, participants are able to categorize whether an image contains a human or an animal when images are presented in very small format (similar to the size used in the present study) for 100 ms [47]. Based on these findings, it is assumed that presenting two small images on a computer screen for 2000 ms is sufficient for participants to process the valence of the images (but not long enough to take in all the details of the image). After 2000 ms the two small images were removed and the word “choice” was presented on the screen. Participants could choose with the ‘A’ (left) or the ‘L’ (right) key on the keyboard whether they wanted to see the left or the right alternative. When the participant made the choice, the chosen image appeared in full-screen format (1024 x 768 pixels) on the computer screen for 4000 ms. Trials were separated by a 2500 ms inter-trial-interval.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Overview of the choice paradigm.

https://doi.org/10.1371/journal.pone.0178399.g001

The choice paradigm consisted of the following combination conditions: negative social—neutral (10); negative physical—neutral (10); negative nature—neutral (10); negative social—negative physical (10); negative social—negative nature (10); negative physical—negative nature (10). The presentation of the different combination trials was fully randomized. The position of the images was counterbalanced (i.e., negative images were presented equally often on the left and on the right side). Moreover, the combination between specific images was varied in two different presentation lists, which alternated between participants. The statistical analyses regarding the choice between two negative images are reported in S1 Study; the mean choice proportions for these conditions are presented in Table 1.

After the choice task, participants were presented with all negative and neutral stimuli in full-screen format, and in random order and asked to rate these on interest, negativity, intensity and complexity. Each rating was made on a 0 (not interesting/negative/intense/complex) to 100 (very interesting/negative/intense/complex) continuous rating scale using a slider. Interest was explained to participants as reflecting the extent to which an image grabbed attention and was experienced as interesting. Negativity reflected to what extent an image was found negative. Intensity was explained as reflecting the extent to which an image evoked an activated feeling inside the body. Complexity was explained as reflecting the extent to which an image contained many details and/or was hard to understand.

Results

Correlations between subjective ratings and choice.

A correlation analysis examined the relationship between subjective ratings of interest, negativity, intensity and complexity and the proportion chosen negative images for the three negative–neutral choice conditions. All correlations are presented in Table 2. It is important to note that two correlations did not reach the pre-registered significance threshold. Choice for negative social images correlated with interest for negative social images, ρ = .39, p = .005; choice for negative physical images correlated with interest for negative physical images, ρ = .32, p = .026 (significant only at uncorrected threshold); choice for negative nature images correlated with interest for negative nature images, ρ = .30, p = .035 (significant only at uncorrected threshold). There were no significant (negative) correlations between the choice scores and interest ratings for neutral images, nor were there any correlations between the choice scores and subjective ratings of negativity, intensity or complexity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Correlations between choice proportions and subjective ratings.

https://doi.org/10.1371/journal.pone.0178399.t002

Discussion

Study 1 demonstrated that participants did not consistently avoid images portraying death, violence or harm, but instead chose to explore some of them. This exploration behavior is consistent with theoretical models of curiosity [10, 14, 20], and therefore interpreted as a reflection of “morbid curiosity”. In line with the “strong” hypothesis of morbid curiosity, Study 1 demonstrated that participants preferred to view social images that portrayed death, violence or harm over neutral images. Furthermore, and as hypothesized, morbid curiosity was more strongly expressed for negative social images than for the other categories of negative content. Physical harm images were chosen equally often as the neutral alternatives; negative nature images were chosen less often than the neutral alternatives.

Mean interest ratings for each negative category correlated positively with how often people chose images from that negative category during the choice paradigm (albeit tested against an uncorrected threshold). This finding is relevant because, following the assumption that interest and curiosity are closely related concepts [14], it supports the interpretation that choice for negative images reflects morbid curiosity. Nevertheless, it is important to note that because the interest ratings were made after the choice paradigm was performed, it is possible that choice behavior influenced the interest ratings (i.e., participants may have judged the images as more interesting because they chose them). In future research, an experimental design that would contrast an active choice condition with a random computer-generated choice condition may shed more light on this possibility. Interestingly, there was no correlation in the present dataset between choice for negative images and the judged intensity of those images. This finding is inconsistent with an explanation of morbid curiosity merely based on a motivation to seek out arousing experiences [1].

One possible interpretation of the findings from Study 1 is that the proportion chosen negative stimuli does not reflect a choice for a negative stimulus, but a choice against a neutral stimulus. Although there was no correlation between rated complexity and choice for negative images, the relatively higher visual complexity of the negative images as compared to the neutral images may have influenced choice. Moreover, previous research has shown that human content is overrepresented in high/arousal negative IAPS images as compared to low arousal/neutral IAPS images [50]. Thus, participants may have preferred social negative stimuli over neutral stimuli, because of the presence of interacting humans, and not because of the negative content per se. Therefore, Study 2 assessed choice behavior for negative images when the alternative choice was well matched in terms of content and visual configuration. Furthermore, Study 2 assessed choice behavior for negative images when the alternative choice was either relatively neutral or positive (e.g., pairing a negative social image with a positive social image).

Study 2 tested the hypothesis that participants would prefer to view negative social images over neutral social images. Furthermore, the study also tested the hypothesis that, overall, the proportion chosen negative social images would be significantly higher than the proportion chosen negative physical and negative nature images. Regarding the choice between negative physical images and neutral physical images, it was hypothesized that participants would not prefer negative physical images. The study did not test specific hypotheses regarding negative nature stimuli and for the choice between negative stimuli and positive alternatives. Finally, based on the findings of Study 1 regarding the interest ratings, the study tested the hypothesis that mean interest ratings for the images in each negative category would correlate positively with the proportion of chosen negative images of that category.

Method

Results

Discussion

Study 2 examined whether participants chose to view images portraying death, violence or harm when those negative images were combined with neutral or positive alternatives that were similar in content. In line with the hypotheses, curiosity was most strongly expressed for negative social images; this category was preferred over neutral social images with similar low-level visual properties, and chosen significantly more often than the two other negative categories. Study 2 clearly demonstrated that choice for negative social images changed depending on the valence of the alternative (i.e., it dropped from 58% to 43% depending on whether it was paired with a neutral or positive alternative respectively). In contrast to the findings of Study 1, images displaying physical harm were chosen less often than the alternative, irrespective of whether the alternative was neutral or positive. Images displaying natural threat were also chosen less often than the neutral or positive alternative. Finally, as in Study 1, there was a positive correlation between interest ratings for the images in each negative category and the proportion of chosen negative images of that category.

Method

Results

Discussion

Study 3 examined choice for images portraying death, violence or harm with verbal descriptions instead of visual information as choice cues. As hypothesized, people chose to view negative social images more often than negative physical images. Furthermore, and in line with Study 1 and 2, negative social images were preferred over neutral social alternatives, but not over positive social alternatives. In contrast to Study 2, negative social images were chosen equally often as positive social alternatives. And finally, negative physical images were chosen less often than the neutral and positive physical alternatives. Thus, participants made largely similar choices as in Study 1 and 2, when they based their choice on verbal descriptions of negative images.

What drives morbid curiosity?

So far morbid curiosity has mainly been explained in terms of a sensation seeking motivation [1]. Applying such an interpretation to the present findings would imply that people choose to view negative images because they seek out the sensations or experiences that they expect will be evoked by the selected stimulus [53]. An alternative, though not mutually exclusive, interpretation is that people choose negative images because of an epistemic motivation. In other words, people may be curious about negative stimuli because these stimuli allow people to acquire knowledge about the world [20]. This interpretation is consistent with models that characterize curiosity as a “drive state for information” ([54]; p. 450), as a state characterized by a motivation to explore information [10], or as a state characterized by a motivation to close an “information gap” [11].

If we consider the possibility that curiosity is about the informational value of a stimulus, then what do images that portray death, violence or harm have to offer in informational terms? According to Baumeister and colleagues [55] negative events have greater psychological impact (e.g., better memory, more cognitive elaboration) than neutral or positive events, because it is evolutionary adaptive to be sensitive to negative information. Another interpretation is that learning about emotional events (both positive and negative) may expand the conceptual system that according to some emotion models is crucial in the experience and regulation of emotional states [56]. Thus, people may explore stimuli that portray death, violence or harm because it gives them handholds that are useful in dealing with future negative situations.

Another relevant suggestion made by Unkelbach and colleagues [57] is that positive information is very much alike, whereas negative information is often uniquely negative. Translating this to the current findings, this means that people may choose stimuli that portray death, violence or harm because they deviate from the norm, and thus are associated with a relatively strong gain in information. This is in line with suggestions from other authors that people are curious about stimuli that present a relatively rare source of information [20] or a relatively uncertain or contradictory source of information [18, 19].

In addition, it is also relevant to address what negative social stimuli, specifically, have to offer in informational terms. Interestingly, Kashdan and Silvia [14] suggest that curiosity for people and social situations may offer psychological and social benefits. It is important to note, however, that this suggestion is made within the context of discussing curiosity as a positive state, whereas the present paper targets curiosity for negative social information. That said, Kashdan and Silvia’s argument that social situations are often ambiguous and challenging can be applied to negative social situations, and may explain why this type of information evokes exploratory behavior. Moreover, social information may have a relatively high learning potential. Developmental studies, for example, show that young infants have a preference for social stimuli (see for an overview [58]). Applying these interpretations to morbid curiosity, it may be the case that the exploration of negative social information helps an individual to acquire and encode important knowledge about the social environment.

Furthermore, an image of a negative event placed within a social context may evoke a more complex narrative [59] as compared to an image that portrays a decontextualized attacking animal or a decontextualized portrayal of physical harm. Recent research within literature science has shown that negative narratives can be attractive [60]. Furthermore, the narrative suggested by a negative social scene may evoke questions, such as What happened? What are the relationships between these people? What will happen after this scene? Participants may choose the negative social images more often as compared to the images portraying graphic physical harm or attacking animals, because they want to find answers or clues to these questions (and close the information gap; [11]).

Finally, the urge to inform oneself about the details of a negative social situation may not only be driven by an information-seeking motivation, but potentially also by a motivation to experience empathy or sympathy for the people portrayed in the images (see for a similar point [50, 61]). Indeed, some studies on empathy have used stimuli with similar content as the negative social images used in the present studies (e.g., [37, 38, 62]). Moreover, people may experience a sense of moral responsibility to inform oneself about the reality of other people’s suffering [63]. The question how different motivations to empathize might influence people’s choice to expose themselves to the negative experiences of others is an important avenue for future research (see also [64]).

Conclusion

The present paper advances new insights into the phenomenon of morbid curiosity and implies that there is a dynamic relationship between negative stimuli and behavior. Images that portray death, violence or harm are not always avoided, but can evoke curiosity driven approach behavior. These findings expand theoretical models of curiosity that have focused mainly on the reward value of positive stimuli (e.g., [65]; see for a similar point [10, 18, 19]). Furthermore, these findings connect to current models of emotion that argue for individual and situational diversity in how affective stimuli are experienced [56, 66].

In practical terms, the present findings also have methodological implications. First of all, the present paper presents a paradigm that provides a behavioral index of (morbid) curiosity that can be used in future studies to test situational and contextual moderators of curiosity and interest-related constructs. Furthermore, with some adaptations the paradigm is suited to be combined with physiological and neurological measures to understand how curiosity for negative information (or other types of information) is represented in the brain and body. Another important methodological implication concerns the fact that all negative images that were used in the present paradigm were selected from standardized image sets, such as the IAPS database [44]. These same images are routinely used in affective science to study a variety of emotional reactions, such as fear, disgust, arousal or negative affect. When using this kind of material it may be important for affective scientists to take into consideration that some of the participants in the sample may actually want to view images that portray death, violence or harm.

In short, the present series of studies provide new insights about the occurrence of morbid curiosity, a largely neglected topic of investigation. Although many questions remain unanswered, and more research is needed, the present paper aims to put this phenomenon back on the research agenda of affective science.