The Feelings of Extreme Risk: Exploring Emotional Quality and Variability in Skydiving and BASE Jumping

The quality and intensity of emotional experiences are important motivators for extreme sport, but surprising little is known scientifically about the feelings of leisure activities that involve high risks. In the current study the emotions of 13 BASE-jumpers and 18 skydivers were measured during and immediately after a jump, and again 24 hour later, during and immediately after watching a film of themselves jumping. Verbal reports, visual reports and heart rate measures were collected. Results showed that feelings like engagement and interest varied more between the jump, the film and a neutral baseline condition, as compared with feelings of pleasure and happiness. During the jump, however, participants experienced huge variation in pleasure, but relatively stable levels of interest. The findings further suggest that extreme sport episodes produce both very intense and very heterogeneous feeling states with a clear distinction between otherwise relatively similar feelings. The results are discussed with reference to so-called eudaimonic and hedonic feeling states, to memory bias, and to the issue of reliability in self-reported emotions during high levels of arousal.






Publication: Journal of Sport Behavior
Author: Hetland, Audun
Date published: June 1, 2012

Is it possible to identify the quality and intensity of the best feelings imaginable? According to a group of extreme sport athletes interviewed on television news (CBSNews, 2009), the answer is yes: BASE jumping is the hallmark of peak experiences, so intense and fascinating that it legitimates the danger of throwing oneself off cliffs or out of airplanes. Inspired by the idea that the intense feelings of extreme sport are strong enough to prepare individuals to take the high risks often involved in such activities, the current study set out to investigate the claim that risk taking provides some of the best feelings that human possibly can experience. Hence, the two questions propelling the current research were these. What does it feel like to throw your self off cliffs or out of airplanes, and can such experiences be captured by scientific methods?

Intense feelings have repeatedly been suggested as the chief motivator for skydivers and other extreme sport athletes (Willig, 2008). However, even if extreme sport is a rapidly growing phenomenon in the western world (Campbell & Johnson, 2005; Celsi, Rose, & Leigh, 1993; Puchan, 2004; Slanger & Rudestam, 1997; Soreide, Ellingsen, & Knutson,2007; Willig, 2008), surprisingly few researchers have investigated the trend scientifically. For instance, a search in the database PsycINFO in July 20 1 0 gives only 8 hits for "extreme sport" and no hits for combining "extreme sport" with terms such as "emotion" or "subjective experiences". It seems reasonable then, to start filling some of the knowledge gaps in the areas of the emotional life of extreme sport events.

Extreme Sport

The term "extreme sport" is not easily defined nor is it easily delimited, but it may be characterized as recreational physical activity that carries a risk of serious physical injury or even death (Willig, 2008). The term "extreme sport" has become a popular label for a range of relatively new activities like climbing, bungee jumping, free ride skiing and snowboarding, surfing, hang gliding and paragliding, kayaking, rafting, small plane aerobatics, full contact marital art, skydiving and BASE-jumping (Slanger & Rudestam, 1997; Soreide, et al., 2007; Willig, 2008). The notion of competition is a frequent qualifier for sport, however, not all of the activities mentioned above are mainly focused on competition. A more correct definition could be "Extremely dangerous physical recreation". Despite the somewhat imprecise definition we still use the term "extreme sport" in this article because this is a well-known label for these activities. These new activities are becoming increasingly popular. In the USA, participation in traditional sports grew by 1 .8% between 1 978 and 2002, while what is called alternative sports rose by more than 244 per cent during the same period (Puchan, 2004). Campel and Johnson (2005) reported that more than five per cent of the adult population was taking part in at least one adventure activity on regular basis, and further that 12% would like to participate. In addition Gelsi and colleges report that the demographics of the participants have broaden to include people of all ages and an increasing number of women (Gelsi, et al., 1993).

However, not all of these activities are truly risk associated and the risk within the same category of activity can vary extensively. As an example, Dahl (2008) estimates that about 30,000 people in Norway conduct some form of climbing on a regular basis. But that is not to say that Norway is a nation of extreme sport athletes. Only a handful of these are doing high altitude rock climbing, or free-solo rock climbing, ice-climbing or difficult technical climbing and big wall climbing. Whereas in- and outdoor sport climbing are low risk activities, high altitude mountain climbing is associated with a significant mortality rate. For instance, K2 - the world's second highest mountain - is also one of the most dangerous, with a summit/ fatality rate of nearly 21%. In other words, for every 100 climbers returning after successfully having reached the summit, 26 have died trying (Perˇn, 2009).

The above review suggests that it is difficult to make a distinction among risk -associated activities, segregating the extreme from the high or medium risk activities. The main reason is that the subjective experience of extreme sports depends on factors other than the objective calculation of risk. Thus, the question that arises is whether extreme sport can be defined as a set of activities, or if we rather should turn to the personal experience of the activity. The current study leans towards the former option, as we picked two activities to represent extreme sport activities.

The two activities we have chosen are skydiving and BASE-jumping. BASE is an acronym for Building, Antenna, Span and Earth, and thus represents jumping from fixed objects, and is similar to skydiving in several respects. But there are also distinct differences between BASE jumping and skydiving. Unlike regular skydiving, the BASE-jumpers normally carry no reserve parachute, and the jumps are often made from lower altitude and consequently the chute is deployed closer to ground. As for the danger involved, the risk of skydiving has been estimated to be around 5 deaths and 140-170 injuries per 100.000 parachute jumps, and annually leading to 30 - 40 deaths in United States (Soreide, et al., 2007). By contrast, BASEjumping is considerably more extreme and dangerous, ranked among the most dangerous sports in the world (Griffith, Hart, Goodling, Kessler, & Withmire, 2006; Pedersen, 1997).

A few studies have concerned themselves with the meaning and motivation behind extreme sport (Gelsi, et al., 1993; Larkin & Griffiths, 2004; Willig, 2008). They found that instead of falling pray to impulsivity, extreme sport participants carefully construct the context in their search for thrills. The athletes do stretch themselves toward the edge of their skills, yet they insist on doing so within existing capabilities. By this strategy the athletes can test their own limits without being overwhelmed, which enables optimal functioning and may lead to the unique experience of risk taking.

According to Willig (2008), motivation for extreme sport seems to be twofold. The first is typically rational and purposeful, linked to the desire for challenges and to acquire increasing levels of mastery and skills. The other motivating force seems to operate on a more emotional level, motivating the participants to repeat an activity that produces otherwise inaccessible feelings of joy and elation.

Feelings and emotions

Emotions comprise several distinct components, such as physiological processes, bodily expressions, action readiness and feeling quality (Oatley & Jenkins, 1996). The current paper is mainly concerned with the subjective feeling part of emotions, although we pay some attention to the physiological processes as well.

Theories of emotions disagree about the causes and categorization of the emotional feeling states. For example, according to the core affect approach (e.g., Russell, 2003), the quality of a feeling state depends on the two basic dimensions of pleasure (valence) and arousal (activation). Feeling states can further be lumped together into the two broad categories of positive affect and negative affect. By contrast, theories of basic emotions (e.g., Ekman, 1984; Oatley, 1992; Panksepp, 1998) defend the notion that a limited number of innate, hardwired affect programmes is the fundamental properties of emotional feelings. A set of discrete, or basic, emotions such as anger, fear, sadness, and happiness give rise to the physiological and phenomenological components of our emotions.

A dominant premise of basic emotions theories is the idea that emotions are functional. They assist in the management of human action, by guiding our attention and prompting action in relation to events that have implications for our concerns. Because of the concern for functionality, basic emotions theorists are uncomfortable with the broad categorizations, such as positive and negative emotions, and they prefer a more fine-grained taxonomy. For instance, pleasure and interest are believed to be quite distinct feeling states, with different physiology, phenomenology and functionality. This distinction is nicely summarized in a recent article by Barbano and Cador (2007), entitled: "Opioids for hedonic experience and dopamine to get ready for it". Barbano and Cador, following earlier work by Berridge (e.g., 2003), claim that dopamine appears to be much more involved in the approach, anticipation, and "wanting" aspects of behavior, whereas the pleasantness of hedonic evaluations seems to be regulated by other brain systems, such as endogenous opioids.

In social psychology, and in the literature on human well-being, the "liking" and "wanting" parts of positive emotions are sometimes referred to as hedonic emotions and eudaimonic emotions, respectively (Ryan & Deci, 2001 ; Vitterso, Saholt, Hetland, A., & Raysamb, 20 10). Although these terms are less precise than liking and wanting, hedonic and eudaimonic emotions carry fewer connotations to the physiological aspects of emotions, and are more concerned with the psychological aspects of emotions. For example, eudaimonic emotions, such as interest, engagement and vitality, are emotional states related to processes of fulfilling one's potentials, experiencing meaning in life and developing a sense of being oneself (for a review, see Vitters0, in press). In investigating emotional experiences among BASE jumpers and skydivers, we believe that the distinction between hedonic and eudaimonic emotions makes sense, and will make use of these concepts in our analyses (see also Beedie, Terry, & Lane, 2000; Delle Faveaaa, Bassi, & Massimini, 2003)

Emotions at High Levels of Arousal

In everyday terms, to be aroused means to be wide awake, excited, vigorous and alert. To be unaroused means to be relaxed, sleepy or tired (Revelle & Loftus, 1992; Thayer, 1989). Although there is evidence that arousal consists of a range of separate arousal systems, McGaugh and colleges (1990) have shown how separate arousal systems serve the same function.

Revelle and Loftshus ( 1 992) argue that variations in arousal may be serving the function of varying the resources available for information processing. They point out that high arousal at encoding facilitates both detection and encoding for long-term retrieval, but may also lead to an inability to retrieve information for a short period of time, up to 30 minutes, after the original experience (Revelle & Loftus, 1 992). For example, Kleinsmith and Kaplan ( 1 963) found that arousing words were better remembered after one week than they had been two minutes after learning. This is in line with the research on stress hormones, in which high arousal has been shown to increase the ability to retrieve information in the long run (Civettini &Redlawsk,2009).

This deficit of immediate retrieval may not be entirely due to high arousal at encoding. Because, if a participant has a high level of arousal during an emotional episode, a high level of arousal will likely spill over to any assessment collected immediately after the episode as well. For that reason, the change in immediate memory performance may be due to the influence of arousal upon retrieval as much as upon learning (Millar, Styles, & Wastell, 1980).

Given the effect arousal has on the immediate retrieval, self-reported emotions may be severely biased if administered immediately after an extreme sport experience (for further arguments see Corson & Verrier, 2007). However, memory is grinded by time, making a delayed report of an experience prone to an increasing range of biases.

Arousal may be measured in a variety of ways, from cortical activity like measures of EEG, to autonomie measures like Skin Conductance (SC) and Heart Rate (HR) (Thayer, 1989).

Sources of Information In Self-Report

Robinson and Clore (2002) argue that people prefer to use the most specific source of information when reporting their emotions. They suggest that people access at least four types of knowledge, ranging from most to least specific: experiential knowledge, episodic memory, situation-specific belief and identity-related belief. Experiential knowledge is a direct access of current emotions. This information can neither be stored nor retrieved. However, through the episodic memory, people can attempt to retrieve specific moments or contextual details from the past. Although past emotional experiences cannot be re-experienced, they can often be reconstructed, aided by this information.

For example, after landing a BASE jumper may remember exit position, the relative speed of the track, the deployment altitude and following canopy ride. Further the BASE jumper may also remember thinking: "this is a good track" or "this exit is unstable and dangerous". This labelling of the feeling during the jump, but not the feeling itself, will be available for retrospection. But emotion related memory is no different from other types of memory, and therefore the ability to recall contextual details will decline quickly with the passage of time.

Situation-specific beliefs are people's belief that certain emotions are likely to be experienced in a particular type of situation. For instance, most of us believe that vacations are associated with happiness and death of a loved one lead to sadness. Finally, Identity-related beliefs are the beliefs people hold about their emotional experiences in general, like their emotional traits, but also the normative social beliefs.

All of these four sources of information give potentially different information about the individuals' emotional experiences. This may cause people to report differently when completing a self-report, depending on the kind of information being accessed. The online reports will most likely access the state emotions, while retrospective reports will, if no episodic cues are available, most likely access the trait emotions.

Aims of t he study

The major aim of this paper was to investigate emotional experiences among BASEjumpers and skydivers. We wanted to identify what it feels like to jump offa cliff or an airplane, and to study how reliable measures of emotions are under such circumstances. To this end we designed the study to measure emotions at different points of time, and with different methods. In addition we also recorded the participants heart rate (HR) both during the jump and 24 hours later, after watching them self on film. We also wanted to investigate the so-called "Repeated assessment bias", which may occur if questions about the same event is asked several times.

Research questions

Grounded in the above aims, the paper set out to investigate the following research questions:

1. Repeated assessment bias: Are there any differences between individuals who report both their emotions immediately after the jump, and after the film, compared with individuals who only report their emotions after the film?

2. Retrospective bias: Are there any differences in emotions reported immediately after the jump compared to after the film?

3. Different expressions of emotions: What are the similarities and differences between extreme sport experiences as measured verbally, visually and physiologically?

4. Emotional distinctiveness: Do the overarching categories of positive emotions and negative emotions make sense for the high aroused extreme sport experiences?

Method

Participants

A total of 3 1 extreme sport athletes (3 women and 28 men), were investigated during the summer of 2008. The sample comprised 13 BASE jumpers and 1 8 skydivers, with an age range from 18 to 39 years (M=28.50). The BASE jumpers were recruited by the first author at the camping ground in Lysefjorden, below the Kjerag massif, outside Stavanger, in Rogaland county. The Kjerag massif attracts jumpers from all over the world, and the Stavanger BASE association runs a scheduled taxi service bringing the jumpers by car into the mountain, and by boat from the landing area at the fjord and back to the camping ground.

The skydivers were recruited at the drop zone in Voss, in Hordaland county, near the city of Bergen. The drop zone at Voss is the second biggest in Norway, and jumping is conducted seven days a week from 1 May to 1 September. Only extreme sport athletes already jumping with a helmet mounted camera were asked to participate in the study. All participants used their own helmet mounted camera gear. Except from that, there were no other conditions for the selection of participants.

Materials and Procedure

The data for our study came from five different sources: three questionnaires and two heart rate measures. The questionnaires assessed (1) background variables (Questionnaire A); (2) subjective experiences immediately after the jump (Questionnaire B); and (3) subjective experiences after 24 hours (Questionnaire C). The two heart rate measures were recorded (4) during the jump, and (5) after 24 hours while watching the film.

One day before the jump

One day before the jump the participants completed Questionnaire A. The participants were first briefed on the study one by one, gave their informed consent to participate, and were informed that they could withdraw at any time. All information was given anonymously.

Immediately before and during the jump

The procedure on the day of the jump, from start until exit, was slightly different for the BASE-jumpers and sky divers and will be described separately.

BASE-jumping: Before they participants geared up, they were given the heart rate monitor (HRM), and assisted by the first author if needed in mounting it. The first author then withdrew after checking that the HRM was working properly. When the participants were ready to jump, the first author set the HRM on record, and also made sure that the helmet mounted camera was recording. The participants were filmed using two cameras. The first was mounted on the participants' helmet, capturing their point of view, and the other camera was operated by the first author, who filmed the exit from a third person's point of view. The third person's point of view was shot as a total picture 5-10 meters to the side of the exit, with the exit point (cliff edge) defining the middle of the frame. From the exit point to the very edge of the frame there was approximately ten meters. The authors' camera was left steady, not following the jumpers as they exited.

Skydiving: The first author met up with the skydivers no less than 30 minutes prior to departure. They were given the HRM and assisted, if needed, in mounting it. After they had geared up, the author followed them into the plane and up in the air. No less than five minutes prior to exit the author started the HRM. Before exit, we also made sure that the helmetmounted camera was recording. The exit was also here shot as a total picture and the camera was left steady, not following the jumpers as they exited the plane.

Immediately after the jump

The rest of the procedure was the same for all participants. Five to ten minutes after the participants had landed, and as they were gearing of, their HRM and helmet mounted camera were turned of. Half of the participants where then given a second questionnaire (Questionnaire B), which they filled in within the first 30 minutes after landing. The films from both cameras were transferred in full DV quality to the authors Macbook Pro using Firewire and iMovie.

The film was then edited with two clips. The first clip (10 sec.) showed the participants' exit. As the participants left the frame of the first camera, the view shifted to the second clip, recorded by the helmet-mounted camera, which showed the rest of the jump until 10 seconds after landing. The film lasted between two and five minutes, depending on the flight time under canopy.

One day after the jump, the participants met up with us again. First the HRM was mounted. The participants then completed the initial part of Questionnaire C, asking about emotional feelings before the film, in order to get a measure of emotional baseline. Next, the participants were shown the film from their own jump. Lastly, they completed the rest of Questionnaire C. HRM was measured during the film, not while the participants were answering the questions.

Assessments

Verbally Reported Emotions

State emotions were measured immediately after the jump, at baseline, and immediately after the film. Our instrument was the Basic Emotions State Test (BEST - Vitters0, Oelman & Wang, 2009). The scale consists of nine items capturing five categories of emotional feelings; pleasure, engagement, fear, anger and sadness. The BEST items were presented after the introduction: "Try to recall how you felt during the episode (alternatively "right now" for the baseline condition), and draw a circle around the number that best describes your emotions", and for each item participants then reported on a Likert-like response scale running from 1 (Not at all) to 7 (Very true).

The items were initially collapsed into three subscales: Pleasure (contentment, enjoyment and happiness) (Jump a = .82; Baseline α = .93; Film a = .92, F(13, 30) = 2.25, p < .05)1 , Engagement (engagement, interest and enthusiasm) (Jump α = .53 ; Baseline α = .90; Film α = .85, F(13, 30) = 3.07, p < .0 1 )) and Negative Emotions (fear, anger and sadness) (Jump α = .33 ; Baseline α = .61; Film α = .40, F( 13,3O) = 1.17, p > . 1 0)). Due to the lower reliability observed for positive emotions with regard to the Jump condition compared with the Film condition, and to the general low reliability for the negative emotions with respect to all conditions, the emotion-items will be treated as separate emotions, and not as sumscore variables.

Visually Reported Emotions

The visually reported scores of pleasure and interest were obtained by a "Feelometer". This instrument enables the participants to provide a moment-to-moment report from the episode. The yaxis shows the intensity of the emotions, and the x-axis is the timeline of the episode. The result is a schematic emotion report that gives the researcher the opportunity to do comparisons at different stages during the episode. Figure 1 shows the five measures we extracted from the Feelometer (1) Height of starting point , (2) Number of peaks, (3) Height of end point, (4) Height of highest point, and (5) Height of lowest point.

Figure 1 provides a compressed example of the Feelometer used in the questionnaire, with the size of the x-axis being 11.5 cm, and the size of the y-axis being 6.5 cm.

Heart rate measures

Heart rate (HR) was recorded with a Polar AXN 500 heart rate monitor (HRM). This records both altitude and HR every 5 seconds. The HR measures were imported to a PC using the program Polar Pro Trainer 5. The HR scales were adjusted so that all HR measures started 30 sec prior to exit, and stopped 3 minutes after landing. The recorded altitude defined exit and landing points. After processing the HR was exported to SPSS 16.0 for Mac via Microsoft Exel 2008 for Mac Version 12.1.5.

The film was recorded using a DV camera with a 0.5 wide angel lens. The film was then imported to iMovie HD 6.03 for editing. The film was also shown in full screen using iMovie on a Macbook pro 2,5 GHz 15", using Mac earphones.

Analyses

Data were analyzed using SPSS 16.0 for Mac. With a sample size of 30, there is an imminent danger of conducting Type I error. In an attempt to remedy this situation we will consider results with p-values in the range between .05 and .10 as "parasignificant".

Results

Repeated Assessment Bias

There were no significant differences between the group that had completed both the questionnaire immediately after the jump (Jump condition) and the one distributed the following day, after the film (Film condition) and the group that only completed the report after the film. For these two groups, non-significant t-values were found for the verbally reported emotions (5 variables), for the visually reported emotions (12 variables), and heart rate (3 variables). Thus, there is no indication that reporting one's emotions immediately after the jump affects the report given after the film.

Emotions during jump, film and baseline

Verbal reports. For three of the jump variables, distribution deviated severely from normality (defined as skewness < |2.00|, see West et al 1995). Visual inspection of box-plots revealed that for pleasure and engagement, non-normality was caused by an extreme score (outlier). These variables became normally distributed when the outlier was excluded. For anger, the non-normality was due to a cluster of low scores and several higher scores. Neither exclusion of two extreme scores nor a log transformation reduced skewness to below |2.00|, and we kept the anger variable untransformed (and highly skewed).

For the film variables, the distribution for anger and sadness deviated severely from normality. Again, neither exclusion of two extreme scores nor a log transformation reduced the skewness to below |2.00[, and we kept both variables untransformed (and highly skewed). However, as a remedy for the non-normality we re-ran all the analysis in Tables 1 and 2 with a non-parametric statistics ( Wilcoxon), finding only trivial differences in significance levels when comparing the nonparametric statistics with the traditional t-tests.

Table 1 shows that the participants tended to report more pleasure during the jump than during the film, with Cohen 's d= 0.47, t(1 3) = 1.79, p =.096). The results for fear showed a similar pattern, with a tendency to report reduced fear for the film (d= 0.60, t[14] = 1-75, p =.103). Further, none of the participants reported any sadness during the jump, whereas they did report some sadness during the film, and the difference appeared to be para-significant (d= -0.92, t[14] = 1 .74,p = . 1 04). Jump and film fear were uncorrelated (r = .t 1,p = .704), whereas anger correlated strongly and significantly between jump and film (r = .97, p <.001). Jump pleasure correlated significantly with film pleasure (r = .55, p = .040).

The differences between the jump emotion and the baseline emotions are characterized by changes in the "eudaimonic" emotions. Engagement, interest and enthusiasm were much higher for the jump compared with baseline O's < .01, < .01 and < .05, respectively). On the other hand, the stability between the positive emotions from the jump to the baseline was low and non-significant. The only exception was interest, for which the correlation between jump and baseline was para-significant (r = .48, p = .073). As for the negative emotions, the tendencies between jump and baseline were pretty similar to those found for the jump-film differences. The mean differences were stronger though, with the p-values for fear and sadness both lower than .05.

Differences between baseline and film emotions are shown in Table 2. The most striking result is the huge difference between the eudaimonic emotions, which are much lower for baseline than for film (all p 's < .001). The rank order stability (i.e., autocorrelations) was also higher, particularly when compared with the correlations reported in Table 1 between jump and baseline.

To better illustrate the distinction between hedonic and eudaimonic feelings, we computed two sumscore variables for each of the three conditions. The hedonic sumscore (comprising satisfaction, pleasure and happiness) did not differ between the jump and the film (M^sub j^ = 6.23, SD = 0.79; M = 5.87, SD = 0.86, t[14] = p A6,p= .168). The eudaimonic sumscore (comprising engagement, interest and enthusiasm) did vary between jump and film (M^sub j^ = 6.3 1, SD^sub f^ = 0.49; M = 5.82, SD^sub f^ = 1 .0 1 , t[ 1 4] = 2. 1 7, p = .048). The eudaimonic baseline sumscore was lower than for the other emotions reported (see Figure 2).

Visual reports. Compared with the visually reported emotions for the film, participants showed less pleasure at the start of the jump and more pleasure at the end of the jump (Table 3). A paired-samples t-test confirmed that the differences were significant and para-significant, respectively, with d= -0.42, /(14) = - 2.28, (p = .039) forthe starting point, and d= 0.30, t(14) = 1 .9 1 , p = .077) for the ending point. The peak pleasure was higher for the jump than for the film (d= 0.72, t[14] = 2.08, p = .057) whereas pleasure also dipped to a lower point during the jump than during the film (d= -0.47, t[ 1 4] = - 1 .95, p = 0,71 ). Comparing the difference between the highest and lowest point on the pleasure curve between jump and film gave a significant result, with d= 0.72, t( 14) = - 2.60, p = .029, (this result is not shown in the table). We found only two significant correlations between the pleasure profile during the jump as compared with the pleasure profile during the film: Height of the end points and height of the lowest points correlated r = .67, p = .007; and r = .57, p = .027, respectively.

We did not find any significant differences between the interest variables obtained after the jump and after the film. As for the rank order stability, the end points, the highest points and the lowest points were significantly correlated, with r = .6 1 (p < .0 1 7) and r = .79 (p < .00 1 ), and r = .48 (p < .069), respectively.

Prototype scoring of visual reports. All the visualized emotional plots were examined, and eight prototypes were created (cf. Ross, 1989). Figure 3 shows the prototypes that we arrived at. Table 3 reveals that many cells are empty or have fewer than 5 responses. In the interest of making a cross-tabulated comparison with chi-square significance testing, we decided to reduce the number of low-response cells by recoding the 8 prototypes into three main categories. By increasing the number cells with an adequate number of responses, this procedure made it possible to test selected patterns of responses (so-called partitioning tests, see Rosnow & Rosenthal, 2008).

The first category covers profiles with a high ending (i.e., collapsing all responses for prototypes 3 -5 into one category). The next category was "low end" (collapsing responses from prototypes 1, 2 and 8), and the third category was labelled "flat" (collapsing responses from prototypes 6 and 7). Table 4b shows the re-categorized responses and reports the result of the significant tests.

The prototypes revealed a significant difference between pleasure and interest experienced at the end of the jump. Whereas half of the participants reported high pleasure at the end of the jump, only one reported high-end interest (p =.008). We see the same pattern after the film (p = .053). Finally, the participants reported an overall lower variation for both pleasure and interest after the film, compared to the jump.

Comparing different expressions of emotions

Verbally vs. Visually Reported Emotions. Verbally reported pleasure showed no significant correlations with visually reported pleasure during the jump. Yet, during the film both the height of the highest point and the height of the lowest point correlated with overall pleasure reported verbally for the jump (r = A3, p = .016; and r = .37, p = .039, respectively.)

We observed more stability between verbal and visual reports for interest. As for the jump, verbally reported interest correlated positively with the height of the staring point and negatively with the number of peaks reported (r = .68, p = .004; and r = -.54,p= .030, respectively). During the film, the overall interest reported verbally correlated significantly with the highest point on the interest curve (r = .5 1 , p = .003). Please cf. Table 5 for further details.

Heart Rate. Table 6 shows descriptive statistics and the correlations between minimum, average and maximum heart rate recorded during the jump (HRJ), and when the participants later watched the film of themselves jumping (HRF). We observed one outlier for maximum HR, which caused the HR max variable to be severely skewed (Skew = -3.44). After removing that particular participant, HR max was normally distributed (Skew - -0.28). The HR variables during the jump correlated highly with each other (ally's < .001), and the HR variables during the film correlated highly with each other (all p 's < .00 1). However, we did not find any significant correlations between jump HR and film HR (ally's > .24 1). All HR means were significantly different from each other (ally's < .001).

Heart rate and verbally reported emotions. The correlations between heart rate and emotions are shown in Table 7. Most of the correlations failed to reach significance, and for those who did, we have not been able to identify a meaningful explanation for the observed associations. For example, most of the positive emotions correlated positively with heart rate, but for the HR^sub max^ during the film, baseline emotions tended to be negatively correlated. Thus, film happiness correlated positively with HR^sub max^ r = .48, p = .03 1) whereas baseline happiness correlated negatively with HR^sub max^ (r = -.16,p= .490). The difference between the two correlations (after being z-transformed) was significant (z = 2.00, p = .023). Another puzzling finding relates to heart rate and visually reported interest. There was a tendency for a positive correlation between these measures during the jump, and a negative correlation during the film. For instance, the lowest point on the interest curve correlated positively with all the heart rate measures during the jump, while they correlated negatively with heart rate during the film. A similar tendency was found for the height of the staring point, and we do not have any obvious explanation for this result.

Discussion

This study explored the feeling qualities of extreme sport, and ways of measuring these experiences. By separating eudaimonic emotions from hedonic emotions, and by including a wider range of emotions measures than used in prior work, we extended past research on the dynamics of emotional experiences of skydiving and BASE jumping. For example, the most prominent difference between a parachute/ BASE jump, and less extreme activities had to do with eudaimonic rather than hedonic emotions. Some indicators suggest that it is eudaimonic feelings like engagement, interest and enthusiasm that make extreme sport so special, and not the level of hedonic feelings defined as satisfaction, pleasure and happiness (Vitterso, et al., 2010). Even if the level of hedonic emotions were high during the jump, these feelings were reported to be equally high during the film and also quite high in the baseline report. For the eudamonic feeling states, on the other hand, a significant difference was found between the jump itself and the two other conditions. A possible interpretation of these results is that the difference between the jump and the other two conditions is more explicit for eudaimonic feelings than for hedonic feelings. It should be noted though, that no difference in intensity between the eudaimonic feelings and the hedonic feelings was found during the jump, so our interpretation is not a claim about the profiles of subjective experiences during the jump. Rather we suggest that compared with hedonÝa, eudaimonic states seem to be more sensitive in terms of differentiating between extreme events and normal conditions.

Actually, high levels of hedonic emotions are typically reported from almost all walks of life from a broad range of nations (Biswas-Diener, Vitterso, & Diener, 2009). Moreover, participants in our study reported a complete absence of sadness during the jump. We are not aware of other studies reporting such a complete lack of sadness. We also found that the three negative emotions of anger, fear and sadness operated rather independently during the jump. Quite understandable, fear reached a peak during the jump, and was significantly lower during the film and for the baseline condition. Both anger and sadness were higher for baseline than for the film (significant) and for the jump (a tendency).

Another interesting result from our study relates to the visual expression of the jump and the film as they unfolded through different phases. Pleasure during the jump was experienced more unstable, with a lower staring point, higher peaks and deeper dips as compared with the visual profile of pleasure during the film. We did not observe these differenced for visualized interest, although both interest and pleasure were given a flatter profile during film than during the jump.

Finally, measures of heart rate showed huge differences between the jump and the film. Heart rates were, of course, much higher during the jump than during the film, but we also discovered a less obvious difference between the two conditions. Whereas the positive verbal emotions basically correlated positively with maximum heart rate during film, several of the positive baseline emotion correlated negative with maximum heart rate during the film. Such a pattern was not observed for the jump condition. Another oddity was between visualized interest and heart rate. For the jump these correlations were strong and positive, for the film they were strong and negative.

Pleasure varied more across our different measures than the other emotions. For example, visually reported pleasure had a larger difference between the highest and lowest point on the drawn curve, it had significantly lower starting and end points than interest. Compared with interest, pleasure varied more between the jump condition and the film condition as well. Adding to the variability over different measures, verbalized pleasure was more stable across the jump, the baseline and the film than engagement and interest.

Memory bias

Kahneman and his colleagues (2004) argue that people are able to give a fairly accurate description of a particular emotional episode, given that they are primed into the episode. Priming takes place when participants revive their memories of the episode, for instance by systematically reconstruct the activity of the episode, or, as in the current paper, by watching a film of the activity. Kahneman, Krueger, Schkade, Schwarz, and Stone (2004) refer to this measurement technique as "day reconstruction", and if aimed towards a restricted set of particular episodes, as "episode reconstruction" (Schwarz, Kahneman, & Xu, 2009).

In the current study we observed some important discrepancies between momentary measures of emotions and those measured after a 24 hours delay. For example, when asked to visualize their remembered pleasures, participants in the current study reported less emotional variations, compared with the online reports. Moreover, our results showed that the intensity of eudaimonic emotions was reduced from the real-time measure to the reconstruction measure. Interestingly, no differences between real-time and reconstruction were found for the hedonic emotions. It is also worth mentioning that none of our participants reported any sadness at the real-time measure, while some of them did so for the reconstructed measure. The reliabilities for the instruments were also significantly smaller for the real-time measures compared with reconstructed measures.

The discrepancies between real-time feelings and reconstructed feelings may throw some light on the essence of the extreme sport experience. Our results suggest that the participants do experience fewer, stronger and more clearly defined emotions during the jump, as compared with both the reconstructed emotions and with the baseline emotions. Despite the almost trivial result showing that emotions during risk taking episodes are experienced as very intense, to our knowledge no studies have pointed out the importance of clearly defined emotions. An important task for future research will be to further illuminate the phenomenological consequences of experiencing clearly defined feeling states of high intensity.

Arousal

This study suggests that there is a negative correlation between pleasantness and arousal. We know from previous studies that high levels of arousal is associated with feelings of unpleasantness (Thayer, 1 989). Such a negative association was corroborated in the current study, showing low levels of pleasure immediately before the jump. At the same time a careful inspection of the heart rate data showed that the participants, without exceptions, reached maximum heart rate at the point of exit or immediately after they had stepped of the cliff or air plain. But during the jump the participants experience both high levels and pleasure peaks. After landing they still report high levels of pleasure, although a new inspection of the heart rate showed that the participants, without exceptions, reached their minimum recorded heart rate some minutes after landing.

The low reliabilities of the verbally reported jump emotions, may be explained with reference to the high level of arousal. High arousal has been reported to cause an immediate memory deficit, lasting up to 30 minutes after the experience (Revelle & Loftus, 1992). In this study all participants were assessed within this time frame. Given this difficulty in retrieving information immediately after a high arousal experience, one should believe that assessments are better administered some time after the episode. The fact that high arousal actually improve the encoding into the long term memory is even another argument to post pone the assessment.

On the other hand, Robinson and Clore (2002) have in several studies showed that the contextual details, which aid the reconstruction of emotional experiences, decrease with the passage of time, and therefore impoverish the emotional recall. In the current study we have used film to aid the reconstruction of emotions, but still there is significant differences in the report given immediately after the jump, compared with the following day.

Even when primed by a film of the episode, the emotions are affected by retrospective biases like the peak/end rule. Further, we also see less variation in the filmed condition. On the other hand, the immediate memory deficit caused by high arousal may impair the report immediately after. In sum, time of assessment seems to be a trade off between reliability and retrospective biases.

Limitations and future research

The present study had a relatively low number of participants (N = 31). Such lack of statistical power increases the risk of conducting Type II errors. For example, we detected very few differences between BASE jumpers and skydivers, even if one might expect some to exist, given the distinctiveness of the two activities. In just 7 of our 82 variables (9%) did we observe significant differences between the two groups, which is only slightly above change level of the conventional p < .05 significance level. However, in every of the 7 instances did the BASE jumpers reveal higher levels of emotional intensity than did skydivers. Hence, future research may benefit from taking a closer look into the issue of how different kinds of risk taking activities may diverge emotionally.

We partly used questionnaire data in this study, which unavoidably have some inherent limitations as discussed and illustrated throughout the paper. As for the heart rate measures, we did not record any heart rate baseline, which would have given us a possibility to see changes on a within participant level during the film. The heart rate, especially the one recorded during the jump, is affected of a range of different factors, which we have had no possibility to control. For example heart rate may be seen as an indicator of the physical shape of the participants. The heart rate for a person in good physical shape would drop quite fast when he or she stop being physically active. Further, for a person with a lower level of fitness the hike to the exit point will take more time, leaving the more fit participants resting for a longer time before the jump.

Future research would benefit from using larger samples, by including more physiological data, and by analyzing a broader range of different activities. Future research should also re-examine some of the results reported in this paper. For example, the nature of negative emotions in extreme sport situations deserves additional attention. Moreover, the difference between visual moment-to-moment reports and the verbally reported episode experiences needs to be further analyzed with an emphasis on the high arousal nature of extreme sport.

Conclusions

The current article contributes to the literature on the emotions of risk taking in several ways. First, it has revealed that different measures of emotions often give diverging results, even when comparing diverse forms of self-reports. Second, the difference between real-time feelings of risk taking and the reconstructed feelings are more pronounced for eudaimonic feelings as compared with hedonic feelings. Third, it turned out that emotions fluctuated dramatically during a risk taking episode, to the extent that an adjective scale or a Likert-like was unable to capture the dynamics of the emotion. The feeling of pleasure seemed particularly exposed to such swings, although it remains to be determined whether this is due to the particular activity under investigation, or if more ordinary episodes have a similar ebb and flow dynamics. Forth, the paper speaks more generally to the issue of real-time experiences as opposed to reconstructed experiences. The current data cast some doubt on the assumption that emotions are properly reported by episode reconstruction techniques, as they hint at substantial differences between real-time emotions and the recalled emotions. But great care must be taken when interpreting this result, given the particularity of the activity we have been studying, and the modest size of our sample.

A final and promising message from our study relates to the quality and intensity of extreme sport experiences. Compared with emotions reported from other activities and episodes, as these are reported in the literature, risk taking seems to produces fewer, clearer and more intense feelings. We speculate that these qualities might contribute to the essence of extreme sport experiences, and suggest that future studies look more closely into what a restricted range of intensely and clearly felt emotions mean for risk taking experiences in particular, and for peak experiences in human lives in general.

Endnote

1 The difference between alpha Jump and alpha Film. According to van de Vijver and Leung (1997) the differences between two alpha values can be tested for significance, using the equation ( 1 -a 1 )/( 1 -a2). For large samples the statistics follows an F distribution with N 1 - 1 and N2 - 1 degrees of freedom (Nl and N2 are the sample sizes).

References

Barbano, M., & Cador, M. (2007). Opioids for hedonic experience and dopamine to get ready for it. Psychopharmacology, 191(3), 497-506.

Beedie, C. J., Terry, P. C., & Lane, A. M. (2000). The profile of mood states and athletic performance: Twometa-analyses. Journal of Applied Sport Psychology, 12(1), 49-68.

Berridge, K. C. (2003). Pleasures of the brain. Brain and Cognition, 52(1), 106-128.

Biswas-Diener, R., Vitterso, J., & Diener, E. (2009). Most people are pretty happy, but there is cultural variation: The Inughuit, the Amish, and the Maasai. Culture and well-being: The collected works of Ed Diener (pp. 245-260). New York, NY: Springer Science + Business Media; US.

Campbell, D., & Johnson, E. (2005). If it can't kill you, it just isn't sporting Retrieved 10.th April, 2009, from http://www.guardian.co.uk/uk/2005/mar/27/sport.deniscampbell

CBSNews (2009, October 22, 2009). The birdmenn. 60 - Minutes, from hup://www.cbsnews.com/video/watch/?id=5377317n&tag;=contentBody:housing

Celsi, R. L., Rose, R. L., & Leigh, T. W. ( 1993). An exploration of high-risk leisure consumption through skydiving. Journal of Consumer Research, 20(1), 1-23.

Civettini, A. J., & Redlawsk, D. P. (2009). Voters, emotions, and memory. Political Psychology, 30(1), 125-151.

Corson, Y, & Verrier, N. (2007). Emotions and false memories: Valence or arousal? Peer Reviewed Journal, 2007-06128-004.

Dahl, M. (2008). Fra gentleman-sporttil folkeidrett Retrieved 9th May, 2008, from http://www.sportsbransien.no/sport.asp?act=read&recno=679

Delle Faveaaa, A., Bassi, M., & Massimini, F. (2003). Quality of Experience and Risk Perception in High-Altitude Rock Climbing. Journal of Applied Sport Psychology, 75(1), 82-98.

Ekman, P. (1984). Expression and the nature of emotion. In K. Sheerer & P. Ekman (Eds.), Approaches to emotion. Hillsdale, NJ: Lawrence Erlbaum.

Griffith, J. D., Hart, C. L., Goodling, M., Kessler, J., & Withmire, A. (2006). Responses to Sport Inventory for Pain among BASE jumpers. Journal of Sport Behavior, 29(3), 242-254.

Kahneman, D., Krueger, A. B., Schkade, D. A., Schwarz, N., & Stone, A. A. (2004). A Survey Method for Characterizing Daily Life Experience: The Day Reconstruction Method. Science, 306(5702), 1776-1780.

Kleinsmith, L. J., & Kaplan, S. (1963). Paired-associate learning as a function of arousal and interpolated interval. Journal of Experimental Psychology, 65(2), 190-193.

Larkin, M., & Griffiths, M. D. (2004). Dangerous sports and recreational drug use: Rationalising and contextualizing risk. Journal of Community and Applied Social Psychology, 14, 215-232.

McGaugh, J. L. (1990). Significance and remembrance: The role of neuromodulatory systems. Psychological Science, 1(1), 15-25.

Millar, K., Styles, B. C., & Wastell, D. G. (1980). Time of day and retrieval from long-term memory. British Journal of Psychology, 71(3), 407-414.

Oatley, K. (1992). Best laid schemes: The psychology of emotions. New York, NY ; Paris, France: Cambridge University Press; Editions de la Maison des Sciences de l'Homme; US; France.

Oatley, K., & Jenkins, J. M. (1996). Understanding emotions. Maiden: Blackwell Publishing.

Panksepp, J. (1998). Affective neuroscience. The foundations of human and animal emotions. Oxford: Oxford University Press.

Pedersen, D. M. (1997). Perception of high risk sports. Perceptual and motorskills 85, 756-758.

Peron, R. G. (2009). Karakoram 2008 Season's end Chronicle. K2 climb.net Retrieved 10th April, 2009

Puchan, H. (2004). Living extreme: Adventure sports, media and commercialisation. Journal of Communication Management, 9(2), 171 - 178.

Revelle, W., & Loftus, D. A. (1992). The implications of arousal effects for the study of affect and memory. Hillsdale, NJ, England: Lawrence Erlbaum Associates, Inc.

Robinson, M. D., & Clore, G. L. (2002). Belief and feeling: Evidence for an accessibility model of emotional self-report. Psychological Bulletin, 128(6), 934-960.

Rosnow, R. L., & Rosenthal, R. (2008). Assessing the effect size of outcome research. Evidence-based outcome research: A practical guide to conducting randomized controlled trials for psychosocial interventions (pp. 379-401). New York, NY: Oxford University Press; US.

Ross, M. (1989). Relation of implicit theories to the construction of personal histories. Psychological Review, 96(2), 341-357.

Russell, J. A. (2003). Core affect and the psychological construction of emotion. Psychological Review, 110(1), 145-172.

Ryan, R. M., & Deci, E. L. (2001). On happiness and human potentials: A review of research on hedonic and eudaimonic well-being. Annual Review of Psychology, 52, 141-166.

Schwarz, N., Kahneman, D., & Xu, J. (2009). Global and episodic reports of hedonic experience. In R. F. Belli, F. P. Stafford & D. F. Alwin (Eds.), Calendar and time diary. Methods in life course research (pp. 157-174). London: Sage.

Slanger, E., & Rudestam, K. E. (1997). Motivation and disinhibition in high risk sports: Sensation seeking and self-efficacy. Journal of Research in Personality, 31(3), 355-374.

Soreide, K., Ellingsen, C. L., & Knutson, V. (2007). How Dangerous is BASE Jumping? An Analysis of Adverse Event in 20,850 Jumps From the Kjerag Massif, Norway The Journal of Trauma, 62(5), 1 1 13-1117.

Thayer, R. E. (1989). The biopsychology of mood and arousal. New York, NY: Oxford University Press.

van de Vijver, F., & Leung, K. (1997). Methods and data analyisisfor cross-cultural data. London: Sage. London: Sage.

Vitterso, J., S0holt, Y., Hetland, ?., ?., T. I., & R0ysamb, E. (2010). Was Hercules happy? Some answers from a functional model of human well-being. Social Indicators Research, 95(1), 1-18.

Vitters0, J. (in press). Functional well-being: Happiness as feelings, evaluations and functioning. In I. Boniwell & S. David (Eds.), The Oxford hanbook of happiness. Oxford UK: Oxford University Press.

West, S. G., Finch, J. F., & Curran, P. J. (1995). Structural equation models with nonnormal variables: Problems and remedies. Thousand Oaks, CA: Sage Publications, Inc.

Willig, C. (2008). A phenomenological investigation of the experience of taking part in 'extreme sports'. Journal of Health Psychology, 13(5), 690-702.

Author affiliation:

Audun Hetland and Joar Vitterse

University of Tromse, Norway

Author affiliation:

Address correspondence to: Audun Hetland, Department of Psychology, University of Troms0, 9037 Troms0, Norway. Tel: +47 93041612(Voice), + 47 77 64 52 91 (Fax) E-mail: audun@hetland.as

The use of this website is subject to the following Terms of Use