Using self-report and skin conductance measures to evaluate theme park experiences

Abstract

This study examines the impact of the key affective moments of a theme park experience on visitors’ post-trip evaluations measured immediately after their visits. One hundred and twenty-three participants visited a theme park while their real-time skin conductance and self-report data were collected. Results indicate that visitors’ pleasure levels (i.e. average, beginning, peak, and end) consistently correlate with satisfaction, which in turn, positively influences behavioural intention. In particular, visitors’ satisfaction levels are better aligned with the affective intensity at the end moment and the average emotion intensity of a theme park experience. Arousal, however, was not found to be a significant indicator of post-trip evaluation. The study extends literature on key moments and retrospective evaluation by illustrating how visitors rely on affective moments of a theme park experience to construct overall evaluations. The study concludes with practical implications and scope for future research.

Keywords

Key moments psychophysiological measure skin conductance measure theme park visitor emotion

Introduction

Emotion is an important aspect of the tourist experience, and pursuing emotional arousal is a major motivation for the purchase of tourism products (Labarge et al., 2004). Tourism, as a typical type of hedonic consumption (Ma et al., 2013), involves fun (Otto and Ritchie, 1996), happiness (Nawijn, 2010), and pleasure (Floyd, 1997), and tourism-related products, such as theme parks, accommodation, and restaurants, are, in many cases, exemplars of hedonic services (del Bosque and Martín, 2008). The emotional outcome elicited from an experience can be seen as the enjoyment that it brings and the consequent feeling of pleasure that it generates (Ma et al., 2013); Mitas et al. (2012) even posited travel as a source of positive emotions in one’s life.

Numerous studies have examined the pivotal role of emotion in tourists’ retrospective evaluations. For example, tourists’ emotions are linked to a series of post-trip evaluation measures such as satisfaction levels (Bigné et al., 2005; Ma et al., 2013), intention to recommend (Hosany et al., 2017), revisit intention (Nawijn and Fricke, 2013), positive meaning in life (Nawijn et al., 2017), and willingness to pay more (Bigné et al., 2005). Despite the importance of emotion in shaping tourists’ retrospective evaluations, extant research has examined emotion as a static construct (Li et al., 2015). In particular, existing studies consistently used an average emotion score as a proxy measure to indicate consumers’ emotional states and few studies have investigated which components or characteristics of tourists’ emotional experiences are better determinants of their overall post-trip judgements. This is problematic because both a flat emotion pattern and emotion curves with positive or negative slopes could produce an identical mean score (Hughes, 1992), even though tourists may not assess them equally.

The literature outside of the tourism field has suggested that key moments of an experience, such as those at the end (Mukherjee and Lau-Gesk, 2016), peak (Baumgartner et al., 1997), or beginning (Biswas et al., 2014) are better predictors of customer preferences or choice than the average emotion felt during such an experience. However, little effort has been expelled in the tourism experience field to understand how tourists integrate key affective moments in constructing their post-trip retrospective evaluations.

In this research, we examine the influence of key affective moments on overall post-trip assessments, based on a combination of self-reporting and psychophysiological measures. The context of this study is a theme park consisting of 13 different sub-themed caves. Following Mukherjee and Lau-Gesk’s (2016) work on examining the key moments of a game experience, the beginning and end moments in this study were defined as a visitor’s emotional experience in the first cave and last cave, respectively. The peak moment was defined as a visitor’s emotional experience in a particular cave (it varied across different individuals) where his or her maximum affective score throughout the entire theme park experience was identified. Additionally, the mean emotion score throughout all of the 13 caves was used as the indicator of average emotion intensity.

Two important contributions of the current study are worth highlighting. First, whereas previous studies investigated emotion as a static construct, this study aims at extending literature on key moments and retrospective evaluation by illustrating how visitors rely on affective moments of a theme park experience to construct their overall evaluations. Specifically, this study investigates which distinct aspects of visitors’ emotional experiences (e.g. the average, beginning, end, or peak moments) are the key determinants of their satisfaction and behavioural intention in a theme park experience. Second, this study is one of the first to use a combination of self-report and physiological measures to track emotion in a theme park, which contributes to a comprehensive and accurate understanding of visitor emotions. In particular, the wearable physiological monitor used in this study serves as the alternative pathway to validate the reliability of the self-report measure.

Literature review

Approaches to the study of emotion

Two prominent approaches to the describing of emotions have frequently been used in tourism literature: the dimensional and the basic approach. The dimensional approach distinguishes emotions based on affective dimensions of pleasure and arousal (Johnson and Stewart, 2005, Li et al., 2015). The pleasure dimension indicates the pleasantness of an experience and arousal refers to the activation level elicited by an experience or event. Conversely, the basic approach identifies a limited number of key human emotions and considers more complex emotions as the combination of these; for example, the Consumption Emotion Set (SET) proposed 16 emotions relevant to the consumption context, such as anger, worry, joy, excitement, and surprise (Richins, 1997). However, there is not enough evidence supporting the notion that discrete emotions are categorically different from each other (Barrett and Wager, 2006). The dimensional approach is more prevalent than its counterpart in tourism research (Prayag et al., 2015) as it abandons the long lists of emotive items and conceptualised emotions along two dimensions. In addition, the dimensional approach has greater explanatory value than the basic emotion approach in the domains of facial behaviours, vocal characteristics and physiological responses (Mauss and Robinson, 2009). Thus, the present study adopted the dimensional approach for describing visitors’ emotions in a theme park because it captures the greatest variance of emotional responses (Mauss and Robinson, 2009).

A few recent studies have examined the antecedents of tourism consumer’s emotions using cognitive appraisal theory (Jiang et al., 2020; Le et al., 2020; Ma et al., 2016). CAT claims that emotions arise as a consequence of evaluative judgement of events and situations as opposed to the events or situations themselves (Bagozzi et al., 1999, Scherer, 1993). Based on CAT, a series of cognitive appraisal dimensions have been identified as the antecedents of emotions such as goal congruence, goal importance, certainty, novelty and agency (Ma et al., 2013).

Key moments of an experience

The peak-end rule theorises that individuals base their overall retrospective evaluations of an experience on the peak moment (i.e. the maximum affective intense moment) or the final moment (Kahneman et al., 1993), rather than the total amount of emotions felt during an experience. It was later extended to include the beginning moment of an experience (Bhargave and Montgomery, 2013) – its rationale lying in the belief that these key moments are most accessible in memory at any given recall time due to individuals’ memory instability (Kemp et al., 2008). Thus, the average affective intensity of an experience is not as important as the key moments in shaping an experience: when an individual is making a retrospective assessment of an experience, both the total duration of the experience (Kahneman, 2000) and the number of peaks during the experience (Schreiber and Kahneman, 2000) are insignificant.

Numerous studies have demonstrated the robustness of the peak-end rule in certain situations, including negative stimuli such as aversive sounds (Schreiber and Kahneman, 2000), pain (Kahneman et al., 1993), pleasant stimuli (Fredrickson and Kahneman, 1993), and mixed-valence events (Thomas et al., 2018). However, several other studies have reported deviations from this rule; for example, Miron-Shatz (2009) established that the average rating of the emotion experienced, rather than the peak or end moments, was a better predictor of retrospective evaluations of multi-episode events. Likewise, the role of certain emotive patterns, such as the ‘peak and end’ pattern, has been observed to be limited in the context of advertising (Hui et al., 2013; Polsfuss and Hess, 1991). These conflicts warrant a closer investigation of the generality and boundary conditions of the peak-end rule in different contexts, especially in a fun-maximisation and hedonic type of experience (e.g. theme park experience).

Existing studies have demonstrated the significant role of emotion in evaluating visitors’ quality of their experience in a theme park (Bigné et al., 2005; Ma et al., 2013). However, extant research focused solely on tourists’ overall emotions and did not examine which specific aspects of tourists’ emotional experiences (e.g. the average, beginning, end, or peak moments) are the key determinants of their satisfaction and behavioural intent in a theme park experience. Therefore, this study proposed the following research question: Which specific aspects of visitors’ emotional responses are the key determinants of their retrospective evaluations of their theme park experience?

Emotion measurement in tourism

The self-report measure is still the dominant method for measuring tourist emotions in the literature (Bastiaansen, 2019; Li et al., 2015), although, in recent years, scholars have begun to see the potential in using physiological tools to track tourism consumers’ real-time emotions (Kim and Fesenmaier, 2015; Shoval et al., 2018). Consistent with the dimensional approach, skin conductance measurement has been used to record tourism consumers’ arousal level. For example, Kim and Fesenmaier (2015) presented the feasibility of examining the unbiased emotions of tourists visiting a destination by measuring their skin conductance (SC) levels over time. By combining GPS with SC, Shoval et al. (2018) extended the work of Kim and Fesenmaier (2015) by exploring tourists’ real-time emotions in time and space. To better capture tourists’ emotions, Stadler, Jepson, & Wood (2018) combined SC measures with post-experience narrative discussion groups to illuminate the process of collective memory creation.

The present study used a combination of self-report questionnaire and skin conductance measures to track visitors’ emotions while visiting a theme park. SC, measured in Micro-Siemens, is an important index of individual arousal levels, is capable of detecting implicit emotional responses that are beyond one’s conscious awareness (Braithwaite et al., 2013), and is ‘the only autonomic psychophysiological variable that is not contaminated by the parasympathetic activity of the body such as the functioning of major organs’ (Jepson, 2018). SC data reports two types of components: skin conductance level (SCL) and skin conductance responses (SCR). SCL refers to an individual’s longer-term responses to stimuli and reflects one’s general changes in autonomic arousal, which usually last for at least 30 seconds (Algie, 2005). SCR, a portion of long-term SCL data, is fast-changing and is elicited by a specific stimulus, which lasts for just a few seconds. Previous studies have demonstrated the capability of using SCL data to track tourists’ longer-term emotional arousal during their experience, ranging from a few hours to 1 day (Jepson, 2018; Kim and Fesenmaier; 2015; Shoval et al., 2018). Thus, this study utilised SCL data to track visitors’ real-time arousal while visiting the Shell Dream World theme park.

Emotion, satisfaction, and behavioural intentions

Consistent with previous research (Hosany and Prayag, 2013; Ma et al., 2013; Su and Hsu, 2013), this study measured post-trip evaluations using satisfaction, intention to recommend, and revisit intention. Satisfaction is the evaluation or comparison between what individuals have and what they think they deserve or expect. If the discrepancy is small, they are satisfied, while a sense of dissatisfaction may be evoked if the discrepancy is large (Campbell, 1980). A robust relationship between emotions and satisfaction has been demonstrated in different contexts, such as sporting events (Martin et al., 2013), higher education (Zhang and Zhu, 2008), and the performing arts (Hume and Mort, 2010).

Positive emotions have previously been positively linked with satisfaction, and negative emotions have been negatively related to satisfaction (Su and Hsu, 2013). In particular, previous studies have revealed the positive relationship between tourists’ emotions (pleasure and arousal) and satisfaction (Yuksel and Yuksel, 2007). Thus, we proposed the following hypotheses:

H1/2: Pleasure/arousal evoked by a theme park experience has a positive impact on satisfaction.

In environmental psychology, Russell and Mehrabian (1974) proposed the approach/avoidance model, which posited that individuals’ reactions to an environment or stimuli are influenced by their emotional states. As pleasure or arousal increases, an individual’s approach behaviour increases. In tourism studies, intention to recommend and revisit intention are thus considered two behavioural consequences of emotions (Qu et al., 2011). Tourists’ positive emotions were found to influence their revisit intention and intention to recommend in the context of a Chinese natural heritage site (Su and Hsu, 2013), festivals (Lee et al., 2008), and dark tourism sites (Nawijn and Fricke, 2015). These empirical perspectives informed the following hypotheses:

H3/4: Pleasure/arousal evoked by a theme park experience has a positive impact on intention to recommend.

H5/6: Pleasure/arousal evoked by a theme park experience has a positive impact on revisit intention.

The relationship between satisfaction and behavioural intention is well documented in tourism literature (Ma et al., 2016; Prayag et al., 2013, 2015). Satisfaction is an important service factor in forming one’s behavioural intentions (Cronin and Taylor, 1994): satisfied tourists are more likely to revisit and recommend the destination or attraction to others. Given that the cost of attracting a new consumer is greater than retaining an existing customer (Spreng et al., 1995), it is imperative to keep tourists satisfied to prompt positive behavioural intentions. Thus, we propose:

H7a: Visitors’ satisfaction has a positive impact on intention to recommend.

H7b: Visitors’ satisfaction has a positive impact on revisit intention.

In examining the relationship between emotions and behavioural intention, previous studies have utilised satisfaction as a key construct. Generally, tourists’ positive emotions are positively related to their satisfaction levels, which in turn, influence their follow-up behavioural intentions. For example, Su and Hsu (2013) found that satisfaction partially mediated the effects of positive emotions on revisit intention and intention to recommend when tourists visited a natural heritage site. However, satisfaction was found to fully mediate the influence of emotions on behavioural intentions in the context of a restaurant (Han et al., 2009). This mixed evidence might be explained by the complexity of tourist experiences (Prayag et al., 2013). However, little research has been devoted to understanding the mediating effects of satisfaction in the relationship between emotions and behavioural intentions in the context of theme parks. Thus, this study proposed:

H8a: Visitors’ satisfaction levels mediate the relationship between pleasure on intention to recommend.

H8b: Visitors’ satisfaction levels mediate the relationship between pleasure on revisit intention.

H8c: Visitors’ satisfaction levels mediate the relationship between arousal on intention to recommend.

H8d: Visitors’ satisfaction levels mediate the relationship between arousal on revisit intention.

The model proposed by this study is shown in Figure 1.

Figure 1.

Conceptual model and hypotheses.

Method

Research context

A theme park was chosen as the context for the investigation of visitors’ key affective moments and our corresponding hypotheses. Theme parks are experience- and fun-oriented tourism products, which have been recognised as typical examples of hedonic consumption (Bigné et al., 2005). Further, the emotive outcome of this consumption is considered imperative to the evaluation of the overall quality of the experience (Ma et al., 2013); hence, a theme park provides an ideal environment for examining visitors’ emotions. Considering the potential effect of temperature on visitors’ skin conductance levels (Braithwaite et al., 2013), this study selected an indoor theme park – Shell Dream World – as the data collection venue.

Shell Dream World opened in 2014 in Gulang Island, Xiamen, Fujian province, China and is the first cave-style adventure park themed around the ocean and shells in China. Combining advanced multimedia technology, 3D phantom imaging, sound, light, games, and electricity, the park creates a unique audio-visual experience for visitors. Shell Dream World consists of 13 different sub-themed caves, such as the ‘Shell Family Museum’ (No. 1), ‘Mermaid Dance’ (No. 10), and ‘Shell wood’ (No. 6) and each cave offers visitors a unique experience. For example, ‘Shell Family Museum’ provides visitors with a memorable learning experience, enabling them to view more than 6800 types of shells collected all over the world including a giant clam that weighs more than 250 kilograms. Using 3D phantom imaging technology, ‘Mermaid Dance’ offers visitors an elaborate interactive experience – a visual ‘mermaid’ appears to be ‘alive’ and visitors feel as if the mermaid is swimming towards and dancing around them. Combining advanced multimedia technologies of sound and light, ‘Shell wood’ presents a music show that blended ‘Hollywood special effects’ and ‘Broadway song and dance’ with a ‘shell’ theme. Additionally, Shell Dream World engaged visitors with a series of interactive activities. For example, visitors are enticed to ‘fight’ with a visual 3D shark and octopus in cave 9, ‘Adventurer’s Submarine’, and may experience a spinning sensation and feel a sense of weightlessness in cave 11, ‘Dizzy Vortex’.

As illustrated in Figure 2, visitors begin their adventure in the ‘Shell Family Museum’ and exit the theme park after experiencing the final sub-themed cave, ‘20,000 Miles under the Sea’. Shell Dream World shares cafes, restaurants and rest zones with other attractions on Gulang Island, thus, the geographical scope of visitors’ experiences in Shell Dream World exclusively refers to the conjoined 13 caves.

Figure 2.

Shell Dream World’s adventure route map (Source: Obtained from Shell Dream World with permission).

Participants

A pre-test of 10 Chinese visitors was conducted to determine if the meaning of any question was unclear and to ensure that the research assistants were familiarised with the data collection procedure. In the formal data collection stage, 140 visitors volunteered to participate in this study in exchange for 20 RMB. Of the original 140 participants, eight respondents were recognised as ‘non-responders (hypo-responsive)’ (Braithwaite et al., 2013) in SC and nine respondents forgot to press the event marker in some caves and thus their data were discarded. Consequently, data from 123 respondents were used. Given that the physiological data collection is time-consuming and complex, leading to what can be an onerous task for participants, a small sample is common in studies involving psychophysiological measures (Ravaja and Kallinen, 2005; Ravaja and Somervuori, 2013) and is not intended to be representative. Additionally, this sample size did not present any validity issues concerning the necessary analysis techniques (i.e. PLS path modelling accommodates small samples).

Convenience sampling was adopted to collect data. However, considering the effect of age on skin conductance levels (Venables and Mitchell, 1996), only those above 18 years old were recruited as participants in this study. Adults accompanied by a child or children were also excluded due to their high drop-out/reject rate when tested during the pre-test stage.

Emotion measures

Participants’ self-reported pleasure and arousal levels were captured via a non-verbal measure: Self-Assessment Manikin (SAM). Constructed by Lang (1980), SAM is a visual representation of the pleasure-arousal model, in which participant pleasure or arousal is represented by five cartoon-like figures arranged along a nine-point scale. Participants select the point on the scale below the manikin that best represents their emotions. SAM has been frequently used to capture consumer emotions (e.g., Brodien Hapairai et al., 2018), and has been demonstrated as more valid than verbal self-report scales (e.g. a battery of emotion items) in reducing cognitive processing and introspection (Morris et al., 2002).

This study also used skin conductance (SC) to provide an alternative pathway to validate the self-reported arousal level captured by SAM. Consistent with the guidelines of SC measure, real-time SC data was collected using a clinical device, the Empatica E4 wristband. This device was designed primarily for clinical purposes, has shown high validity and accuracy (Empatica, 2019), and has been demonstrated to be a reliable and valid tool for recording visitors’ real-time skin conductance levels in the field (Kim and Fesenmaier, 2015; Shoval et al., 2018). Skin conductance data from the Empatica E4 wristband can be synced to an online cloud platform, where research assistants can access the data in its raw form.

Questionnaire

The questionnaire consisted of three sections covering visitors’ pleasure and arousal (i.e. SAM), retrospective evaluation (i.e. satisfaction, intention to recommend and revisit intention), and demographic information. To ensure content validity, measurement items were adapted from previously verified measures where possible (Appendix 1). All constructs were measured on a 5-point Likert scale ranging from 1 (‘strongly disagree’) to 5 (‘strongly agree’).

Data collection procedure

Data were gathered continuously from April to August 2019. Visitors were approached at the entrance of Shell Dream World, and those who agreed to participate were given an introduction to the research project before signing the consent form. Each participant was fitted with an Empatica E4 wristband. For the purpose of SC score calculation within individual caves, each participant was instructed to press the event marker button on the Empatica E4 wristband whenever she or he entered a new cave. When exiting the last cave (‘20,000 Miles under the Sea’), participants were immediately requested to complete a self-report questionnaire regarding their self-report emotions, retrospective evaluation measures and demographic information. In particular, participants were required to indicate their pleasure/arousal levels at the ‘Shell Family Museum’ (beginning moment), ‘20,000 Miles under the Sea’ (end moment), and their overall pleasure/arousal level (overall emotion intensity) in Shell Dream World by filling out SAM scales separately. They were also asked to recall the caves where they felt the strongest emotions (i.e. pleasure and arousal respectively) and to rate their emotions in that particular cave on the SAM. The entire process (touring plus pre, during and post survey) lasted approximately 1 to 1.5 hours per individual.

Establish ecological validity of SC

Precautions were taken to minimise the noise or artefacts that may threaten the ecological validity of SC data. To ensure good electrode contact and establish the baseline level for each participant, a period of 5 min was allowed between attaching the Empatica E4 wristband to participants’ wrists and starting official data collection (Braithwaite et al., 2013). Further, SC data could be affected by the artefacts (caused by movement) only when the wristband was worn loosely (Empatica, 2019). Therefore, each participant was carefully checked by a trained research assistant to ensure that the wristband was not worn snugly before the participant entered the first cave. Finally, each participant’s continuous SC data (especially the moment when the participant pressed the event marker button) was closely examined by a research assistant to ensure that ‘tagging’ would not cause data disruption. We did not observe any visible data alteration or disruption at the moment of ‘tagging’ among the 123 participants included in the data analysis.

Data analysis

As per previous research (Shoval et al., 2018), the original skin conductance level score for each individual was converted to a z-score to enable comparison between individuals. The mean SC scores from the first cave and the last cave were used as indicators of the beginning and end moment arousal scores, respectively. The peak moment arousal score was represented by the mean SC score in the cave where the maximum SC score was identified. The mean of the SC scores for all of the 13 caves was used as the indicator of average emotion intensity.

To evaluate the structural model, partial least squares (PLS) path modelling (Hair et al., 2012) was used with SmartPLS 3.0 software. The PLS path modelling method has the advantage of being able to handle small sample sizes and process data that is not normally distributed. PLS was thus considered more suitable for this study than the commonly used, covariance-based SEM technique. To compare the differences between two correlation coefficients that involve a common variable, Steiger’s one-tailed z-test was used with an online calculator (Hoerger, 2013; Steiger, 1980).

Results

Sample characteristics

Of the 123 participants, 72 were female and 51 were male. The largest age group was the 18 to 24-year-old group (68%), followed by those aged 25–34 years (27%), 45–55 years (2%), and 35–44 years (3%). Participants varied widely in terms of educational level; the majority held a bachelor’s degree (57%), followed by high school education (40%), master’s degree (1%), doctoral degree (1%), and other (1%). Most participants (122 out of 123) had not previously visited Shell Dream World. Thirty-seven percent of the respondents visited Shell Dream World alone, whereas 63% travelled with companions. As the interest of the current research is in construct associations and not descriptive insights, we refrained from weighting the sample elements.

Measurement model

Data analysis results indicated that all measures met the commonly suggested requirements for measurement model assessment (Table 1). First, all the multiple-item constructs’ AVE values were well above 0.5, and, therefore, the convergent validity was acceptable (Henseler et al., 2009). Additionally, all indicators met reliability requirements because they were higher than 0.7 (p < 0.01). Table 1 indicates good internal consistency (Henseler et al., 2009) via a high Cronbach’s alpha (α) and composite reliability (ρ) values (all greater than 0.7). Table 2 demonstrates that all measures achieved satisfactory discriminant validity because all values were less than 0.9 in the Herotrait-Monotrait ratio (HTMT) of correlations.

Table 1.

Mean and factor loadings for individual items.

Construct	Mean	Item	Loadings	Cronbach’s Alpha	Composite Reliability	Average Variance Extracted (AVE)
Satisfaction	3.5203	SAT1	0.937	0.921	0.949	0.861
	3.3739	SAT2	0.921
	3.5203	SAT3	0.926
Intention to recommend	3.3252	ITR1	0.921	0.916	0.947	0.857
	3.4634	ITR2	0.902
	3.2113	ITR3	0.953
Revisit intention	2.6585	RI1	0.949	0.945	0.965	0.901
	2.6504	RI2	0.952
	2.6016	RI3	0.947
Pleasure	6.2967	PL1	n/a	n/a	n/a	n/a
Arousal	5.9349	AL1	n/a	n/a	n/a	n/a

Note: All loadings are significant at p < 0.001.

Table 2.

Mean and Heterotrait-Monotrait (HTMT).

	Arousal	Intention to recommend	Pleasure	Revisit intention
Arousal
Intention to recommend	0.461
Pleasure	0.673	0.57
Revisit intention	0.299	0.704	0.405
Satisfaction	0.425	0.837	0.553	0.564

Note: shaded boxes are the standard reporting format for HTMT procedure.

Structural model evaluation (hypotheses testing)

In this section, the results will be reported separately for the structural model, with average emotion and beginning/peak/end moment scores as independent variables. The hypotheses were tested by applying a nonparametric bootstrapping analysis (5,000 sub-samples and 123 cases), which is the standard method to test the significance of PLS path-modelling results (Henseler et al., 2009). The results of hypothesis testing with different emotion indexes are displayed in Table 3.

Table 3.

Hypotheses testing.

Hypothesis	Emotion index	Path coefficient	T values	Supported
H1: Pleasure → Satisfaction	Average	0.508**	6.433	Yes
	Beginning	0.339**	3.925	Yes
	Peak	0.367**	5.085	Yes
	End	0.509**	7.868	Yes
H2: Arousal → Satisfaction	Average	0.093^ns/0.021^ns	0.858/0.377	No
	Beginning	0.281^ns/0.181^ns	1.309/1.928	No
	Peak	−0.081^ns/0.034^ns	0.085/0.455	No
	End	−0.277^ns/0.049^ns	1.783/0.678	No
H3: Pleasure → IR	Average	0.175**	2.597	Yes
	Beginning	0.165**	2.572	Yes
	Peak	0.067^ns	1.174	No
	End	0.187**	2.686	Yes
H4: Arousal → IR	Average	0.075^ns/0.011^ns	0.949/0.231	No
	Beginning	0.065^ns/0.012^ns	0.504/0.182	No
	Peak	−0.006^ns/0.015^ns	0.073/0.291	No
	End	0.123^ns/−0.082^ns	0.822 /1.415	No
H5: Pleasure → RI	Average	0.167^ns	1.748	No
	Beginning	0.236**	3.226	Yes
	Peak	0.012^ns	0.133	No
	End	0.184**	2.058	Yes
H6: Arousal → RI	Average	0.007^ns/−0.059^ns	0.061/1.369	No
	Beginning	0.102^ns/−0.085^ns	0.573/1.208	No
	Peak	−0.202^ns/−0.071^ns	1.711/1.068	No
	End	−0.09^ns/0.044^ns	0.466/0.586	No
H7a: Satisfaction → IR	Average	0.657**	11.122	Yes
	Beginning	0.688**	12.226	Yes
	Peak	0.722**	13.891	Yes
	End	0.654**	10.442	Yes
H7b: Satisfaction → RI	Average	0.445**	4.983	Yes
	Beginning	0.461**	5.892	Yes
	Peak	0.522**	6.446	Yes
	End	0.431**	4.729	Yes
H8a: Pleasure → Satisfaction → IR	Average	0.334**	4.926	Yes
	Beginning	0.223**	3.581	Yes
	Peak	0.265**	4.646	Yes
	End	0.333**	5.769	Yes
H8b: Pleasure → Satisfaction → RI	Average	0.226**	3.526	Yes
	Beginning	0.156**	2.979	Yes
	Peak	0.192**	3.767	Yes
	End	0.219**	3.734	Yes
H8c: Arousal → Satisfaction → IR	Average	0.062^ns/0.014^ns	0.854/0.375	No
	Beginning	0.197^ns/0.124^ns	1.274/1.947	No
	Peak	−0.059^ns/0.025^ns	0.691/0.456	No
	End	−0.213^ns/0.032^ns	1.775/0.667	No
H8d: Arousal → Satisfaction → RI	Average	0.041^ns/0.009^ns	0.798/0.374	No
	Beginning	0.122^ns/0.083^ns	1.251/1.942	No
	Peak	−0.042^ns/0.018^ns	0.701/0.457	No
	End	−0.16^ns/0.021^ns	1.783/0.664	No

Notes: IR = Intention to recommend; RI = Revisit intention; ns = non-significant.

For hypotheses involving arousal, the first value is based on self-report measure while the second is based on skin conductance.

**p < .01, *p < .05 (one-tailed test; number of observations = 123).

Average emotion score

Figure 3 displays the outcome of the structural model test using the average emotion score. The PLS path model estimation gave an R² value of 0.623 for intention to recommend (IR), 0.302 for revisit intention (RI), and 0.289 for satisfaction. In support of H₁ and H₃, we found that pleasure had a significant positive impact on satisfaction and IR, respectively. In addition, it was revealed that satisfaction had a direct effect on both IR and RI, which supported H_7a and H_7b. However, the direct effect of pleasure on revisit intention (H₅) was not supported. Further, the direct effect of self-report arousal on satisfaction, IR, and RI was not found. To further validate the results, we replaced the self-report arousal with physiological arousal. Similarly, the direct effect of self-report arousal on satisfaction, IR, and RI was not found, which rejected H₂, H₄, and H₆.

Figure 3.

Structural model with average emotion intensity (Notes: **p < .01, *p < .05; for simplicity, nonsignificant paths have been removed).

To test H₈, we conducted a mediation analysis using a nonparametric bootstrapping procedure. In support of H_8a and H_8b, we found that satisfaction partially mediates the relationship between pleasure and IR, and fully mediates the relationship between pleasure and RI. H_8c and H_8d, predicting that satisfaction will mediate the effects of arousal (both self-report and physiological arousal) on IR and RI, were not supported. The statistical analysis above suggested that visitors’ overall emotion intensity throughout the 13 caves partially predicted their immediate post-visit evaluation. In particular, the pleasant feelings (i.e. pleasure) evoked by visitors’ experiences in Shell Dream World had a significant impact on intention to recommend and revisit intention via satisfaction. Arousal (i.e. excitement), however, was not a significant predictor of visitors’ post-visit evaluations.

Beginning moment score

Figure 4 displays the outcome of the structural model test using the beginning moment score. The PLS path model estimation provided an R² value of 0.624 for IR, 0.334 for RI, and 0.189 for satisfaction. H₁, H₃, and H₅ were strongly supported when emotion was indexed by the beginning moment score. Specifically,, pleasure had a significant positive impact on satisfaction, IR, and RI. With regard to H_7a and H_7b, we found that satisfaction had a direct effect on both IR and RI, which supported H_7a and H_7b. The direct effect of self-report arousal on satisfaction, IR and RI was not supported; similarly, the direct effect of arousal on satisfaction, IR and RI was not supported when self-report arousal was replaced by physiological arousal. Thus, H₂, H₄, and H₆ were rejected.

Figure 4.

Structural model with beginning moment score (Notes: **p <.01, *p < .05; for simplicity, nonsignificant paths have been removed).

In terms of indirect effects, we found that satisfaction partially mediated the relationship between pleasure and IR, and RI, which supported H_8a and H_8b. H_8c and H_8d, predicting that satisfaction would mediate the effects of arousal (both self-report and physiological arousal) on IR and RI, were not supported. The statistical analysis above indicated that visitors’ emotions in the first cave (i.e. ‘Shell Family Museum’) partially predicted their immediate post-visit evaluation. In particular, the pleasant feelings (i.e. pleasure) evoked by visitors’ experiences in ‘Shell Family Museum’ (the first cave in Shell Dream World) had a direct impact on satisfaction, intention to recommend and revisit intention. Arousal, however, was not a significant predictor of visitors’ post-visit evaluations.

Peak moment score

Figure 5 displays the outcome of the structural model test using the peak moment score. The PLS path model estimation gave an R² value of 0.605 for IR, 0.283 for RI, and 0.169 for satisfaction. We found that pleasure had a significant positive impact on satisfaction, which supported H₁. However, H₃ and H_5, predicting the direct impact of pleasure on IR and RI, were not supported. Satisfaction had a direct positive effect on both IR and RI, which supported H_7a and H_7b. With regard to H₂, H₄, and H₆, the direct effect of self-report arousal on satisfaction, IR and RI were not supported. Similarly, the direct effect of self-report arousal on satisfaction, IR and RI were not supported when we replaced the self-report arousal with physiological arousal.

Figure 5.

Structural model with peak moment score (Notes: **p <.01, *p < .05; for simplicity, nonsignificant paths have been removed).

In terms of indirect effects, we found that satisfaction fully mediated the relationship between pleasure and IR, and RI, which supported H_8a and H_8b. H_8c and H_8d, predicting that satisfaction would mediate the effects of arousal (both self-report and physiological arousal) on IR and RI, were not supported. The statistical analysis above showed that visitors’ emotions in the ‘peak’ cave (where his or her maximum emotion score throughout the 13 caves was identified) partially predicted their immediate post-visit evaluation. In particular, the pleasant feelings (i.e. pleasure) in the ‘peak’ cave exerted an impact on intention to recommend and revisit intention via satisfaction. Arousal, however, was not a significant predictor of visitors’ post-visit evaluations.

End moment score

Figure 6 illustrates the outcome of the structural model test using end moment score. The PLS path model estimation delivered an R² value of 0.635 for IR, 0.302 for RI, and 0.291 for satisfaction. Pleasure had a significant positive impact on satisfaction, IR, and RI, respectively, which supported H₁, H₃ and H₅. Satisfaction had a direct positive effect on both IR and RI, which supported H_7a and H_7b. H₂, H₄, and H_6, positing the direct effect of self-report arousal on satisfaction, IR and RI, respectively, were not supported. The direct effect of physiological arousal on satisfaction, IR and RI were also not supported when we replaced self-report arousal with physiological arousal.

Figure 6.

Structural model with end moment score (Notes: **p <.01, *p < .05; for simplicity, nonsignificant paths have been removed).

In terms of indirect effects, we found that satisfaction partially mediated the relationship between pleasure and IR, and RI, which supported H_8a and H_8b. H_8c and H_8d, predicting that satisfaction will mediate the effects of arousal (both self-report and physiological arousal) on IR and RI, were not supported. The statistical analysis above demonstrated that visitors’ emotions in the last cave (‘20,000 Miles Under the Sea’) partially predicted their immediate post-visit evaluation. In particular, the pleasant feelings (i.e. pleasure) in the last cave had a direct impact on satisfaction, intention to recommend and revisit intention. Arousal, however, was not a significant predictor of visitors’ post-visit evaluations.

Comparison of key moments

The results presented above demonstrate that emotions indexed by average/beginning/peak/end moments are consistently positively related with satisfaction. Thus, we further compared the path coefficients between average/beginning/peak/end moments and satisfaction to investigate whether statistically significant differences existed based on Steiger’s one-tailed z-test. However, we focused solely on the pleasure level as arousal did not predict any of the outcome variables. For simplicity, only statistically significant differences in path coefficients were reported.

A comparison between average emotion/beginning moment scores and satisfaction yielded a z-value of 2.51, p = 0.006; similarly, we compared the correlation coefficients between average emotion/peak moment score and satisfaction and obtained a z-value of 2.125 with p = 0.016. In the comparison between end moment score/beginning and satisfaction, a z-score of 2.078, p = 0.019 was obtained. Further, we compared the correlation coefficients between end moment/peak moment score and satisfaction and yielded a z-value of 1.874 with p = 0.03. These results reveal that average emotion intensity and end moments are more effective than beginning and peak moments in predicting satisfaction in a theme park experience.

Discussion and conclusion

Major findings

This study provided empirical evidence of a link between pleasure and satisfaction in a theme park context, irrespective of the specific ‘pleasure’ index being used: either average emotion intensity, beginning, end, or peak emotional experience exerted a significant impact on visitors’ satisfaction. This result supports previous findings, which have posited that the greater the pleasant feelings evoked by an experience, the more likely a consumer will feel satisfied towards it (Bigné et al., 2005; Floyd, 1997), and extends the current literature by demonstrating that key moments of pleasure also hold considerable explanatory power in relation to visitors’ satisfaction levels, capturing the dynamic aspects of the pleasure-satisfaction relationship in the context of theme parks.

Within the affective dimension of pleasure, this study found that the end moment was more influential than the beginning and peak moments in determining visitors’ satisfaction in a theme park. In this study, the end moment refers to the experience of visitors in the last cave, ‘20,000 Miles Under the Sea’. The end moment is considered as the most defining feature of a theme park experience because it may already partly reflect overall evaluation (Hui et al., 2013). Most human experiences are goal-directed (Spiegel, 1998), and individuals can evaluate whether their experiences in Shell Dream World are consistent with their goals (i.e. the abstract benefits sought by the visitors) by the end of their visit. According to CAT, if a theme park visitor perceives that the experience has moved him or her closer to a desired goal, positive emotions and evaluations are more likely to be evoked, and vice versa (Ma et al., 2016). Moreover, endings convey relative certainty and offer individuals one route to comprehend what their global impression really is (Fredrickson, 2000), while for participants who report their evaluation immediately after a theme park visit, ‘recency effects’ could also drive the influence of end moments (Biswas et al., 2014). Specifically, the time interval between the ‘last cave’ (i.e. ‘20,000 Miles Under the Sea’) and the moment when the overall evaluation was made was the shortest, which may explain why the most recent segments had better recall and greater impact.

It was also found that average pleasure intensity was a better predictor of satisfaction than beginning and peak moments, which dovetails with extant findings that have highlighted the important influence of average emotion intensity in overall judgements (Mukherjee and Lau-Gesk, 2016; Polsfuss and Hess, 1991). In this study, average pleasure intensity is a sum of a visitor’s moment-to-moment affective response and reflects an individual’s overall ‘impression’ of his or her entire visit to Shell Dream World. Similarly, satisfaction also captures a visitor’s general evaluation of the visit to the theme park. The similar nature of these measures may explain the stronger correlation between average pleasure intensity and satisfaction.

In contrast with the prevailing peak-end rule, these results cast doubt on the power of the peak moment in driving retrospective evaluations (Thomas et al., 2018). One possible explanation lies in a lack of distinctiveness of the peak moment. Montgomery and Unnava (2009) posited that the peak moment failed to lead to better recall if that moment was not clearly distinguished from surrounding stimuli. In this study, theme park visitors may have perceived their visit to a particular ‘cave’ in Shell Dream World as the most pleasant, but that moment may not necessarily have carried a distinct emotional meaning to fulfil their ‘goal’. As noted by Fredrickson (2000), the meaningfulness of a moment depends on consumer goal assessment, and peak moments outweigh a consumer’s overall feelings only when those moments carry a wealth of self-relevant information. When assessing the entire Shell Dream World experience, visitors may have recalled their goals and evaluated whether they had been obtained. Therefore, the role of the most pleasant part of a theme park visit in impacting post-trip evaluation may be weakened.

The notion that the first moment of an experience is inherently over-weighted in driving retrospective judgement was also questioned by this study. Previous studies have attributed the importance of an experience’s beginning moment to the ‘primacy effect’; however, this effect only holds true when consumers experience a series of similar, rather than dissimilar, stimuli due to habituation (Biswas et al., 2014). When consumers experience a sequence of products with a dissimilar nature, on the other hand, they tend to prefer the last product. Our study’s participants visited a series of ‘caves’ with different themes. As described above, each cave featured a distinct theme and the dissimilar nature of these 13 theme park caves may explain the attenuation of the ‘primacy effect’: the salience and accessibility of the first moment (i.e. the ‘Shell Family Cave’) might become overshadowed by other key moments, such as the end moment (Mukherjee and Lau-Gesk, 2016). According to the CAT, theme park visitors are not capable of assessing whether their visits to Shell Dream World is goal-congruent at the very beginning stage, which may weaken the capability of the beginning moment in predicting their overall post-trip evaluations (Ma et al., 2016).

Consistent with the previous studies positing that feeling good and delighted have a more vital role in generating satisfaction than excitement and stimulation (Moon et al., 2016), this study found that self-report arousal was not a significant predictor of retrospective evaluation in the theme park. Arousal refers to bodily activation and ranges from sleepy to highly energised (Li et al., 2016); however, arousal, as an independent dimension of emotion, does not specify the emotional valence and feeling confused or disappointed can also be accompanied by high arousal level. Therefore, the complex and variable nature of arousal may explain its ineffectiveness in influencing visitors’ post-trip evaluation of their Shell Dream World experience. On the other hand, the ‘broaden-and-build’ theory considers an individual’s behavioural intention as ‘the consequences of emotions that support the continuing pursuit of a successful goal’ (Ma et al., 2016); however, arousal is positively and more strongly correlated with goal importance than with goal congruence (Johnson and Stewart, 2005) and the elicitation of arousal does not signal the fulfilment of a visitor’s goal. An examination of the psychological mechanism through which arousal is evoked based on CAT may further explain the insignificant relationship between arousal and post-visit evaluation as found in this study.

To enhance the reliability and validity of the results, this study also used skin conductance to measure visitors’ continuous physiological arousal levels. While Ravaja (2004) described skin conductance as an excellent operational definition of arousal, a significant influence of arousal on post-trip evaluations was not identified. This finding is partially consistent with one prior study that failed to reveal a significant link between physiological arousal and visit intention in tourism advertising (Li et al., 2018) and fast moving consumer goods (e.g. toothpaste) (Ravaja and Somervuori, 2013). The present study extended previous research by demonstrating that physiological arousal did not have a significant impact on satisfaction or behavioural intention in the context of an on-site theme park experience.

Conclusion

Extant tourism research has only minimally addressed the dynamic process of tourists’ emotions. However, this study represents an attempt to examine visitors’ real-time emotions during their on-site experience and, more importantly, to investigate which aspects of their emotions are better determinants of post-trip, retrospective evaluations. The results indicated that visitors’ pleasure levels can predict their post-trip satisfaction, which, in turn, influences their behavioural intentions. The superiority of average emotion intensity and end moment over peak and first moments in determining satisfaction was also empirically demonstrated herein. However, arousal did not predict either satisfaction or resulting behavioural intention in the context of the theme park experience.

Practical implications

The research reported here not only provides clear and actionable directions for theme park management but is also applicable to museums, heritage sites, and festivals. Identifying key moments of an experience can help marketers optimise resources at critical points to induce higher levels of satisfaction. Because we found that the end moment of a theme park visit can enhance positive post-trip evaluation, managers should devote more resources to enhancing the latter part of a visit. A good example is IKEA, which elicits heightened pleasurable emotions from customers at the end of their shopping trip via the provision of low-cost ice cream. IKEA’s store design has also been demonstrated to be useful in counteracting the feelings of fatigue at the end of customers’ lengthy shopping sessions (Knudsen, 2017). This principle can also be applied to tourist attractions with fixed (e.g. indoor adventure theme parks, museums or heritage sites) or unfixed (e.g. Disneyland, Universal Studio or Chimelong Ocean Kingdom) visiting sequences. Our findings imply that integrating ‘highlights’ of a tourism attraction at the end or final moment of a visit can enhance tourists’ positive post-trip evaluations. Thus, tourist attraction managers (in a tourist attraction with a fixed visiting sequence) can shift the most fun activities or create a ‘positive surprise’ (e.g. offer visitors free or low-cost gifts) towards the end of an experience. For tourism attractions without a fixed touring route, staff can assist visitors in optimising their visiting sequence (e.g. provide verbal advice on an optimal touring route or add touring suggestions on the brochures) and saving the most fun event or activity for last.

While the end moment of an experience is vital, our findings further indicate that average emotion intensity is still a key factor in determining tourists’ overall evaluation of an event or experience. Average emotions reflect visitors’ overall affective impression of an experience and can be used as a proxy to predict their retrospective evaluation (Miron-Shatz, 2009). Thus, managers in different types of theme parks may also need to improve the quality of the entire experience, instead of only specific moments. In addition to evoking a memorable end experience among visitors, theme park managers should keep visitors focused for the duration of their visit through use of engaging elements such as virtual and augmented reality technology, storytelling, theme-related activities, and greater interaction between the theme park and its visitors.

Consistent with the dimensional approach, visitors’ emotions in this study were categorised by two independent dimensions, i.e. pleasure and arousal (Li et al., 2016). The results of this study highlight the importance of using appropriate strategies to evoke visitors’ positive emotions (i.e. pleasure). Thus, marketing campaigns should emphasise the pleasurable emotional experience that can be evoked in different types of tourist attractions (such as joy and delight). In addition to designing and developing more amusing activities, tourism practitioners could enhance visitors’ on-site pleasure levels by improving the ‘servicescape’ (i.e. physical environment) of a theme park. It’s been documented that environmental cues, such as decor, facility aesthetics, and layout accessibility, can significantly evoke customers’ positive emotions (Moon et al., 2016). Nevertheless, this study does not necessarily indicate that arousal is not important in an experience design. We argue that visitors’ positive retrospective evaluation cannot be achieved by merely ‘activating’ or ‘stimulating’ visitors. Rather, a pleasurable emotional experience is more likely to result in positive post-trip judgments.

Limitation and recommendations for future research

This study was limited in its capacity to use a large sample size due to the time-consuming and complex nature of physiological measures. Adults accompanied by a child or children were excluded in this study due to their high drop-out and reject rate. As the study was attempting to theorise, and not generalise, this does not present a significant issue for its conclusions. Nevertheless, future research could enlarge the sample size and compare the differences between solo travellers and travellers with companions, which may provide more conclusive evidence on the influence of affective responses on visitors’ overall retrospective evaluations. Portable eye tracking could also be utilised, in addition to skin conductance measurement, to identify the external stimuli that evoked visitors’ emotional responses. This study was further limited to just one form of tourism experiences, the theme park visit. Future research could replicate this study in other contexts, such as festivals or national parks, to investigate the role of key moments in driving one’s global retrospective judgement. Finally, the focus of this study was theme park visitors’ emotion outcomes in the 13 caves of Shell Dream World. However, guest experiences could also be evoked by other experiential consumption activities such as food offerings, merchandise, rides, shows, bathroom facilities, parking, and more (Tasci and Milman, 2019). Thus, future research could extend the geographical scope of the tourist attraction being investigated and examine the relationship between various environmental cues and visitors’ emotions.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the [National Natural Science Foundation of China #1] under Grant [number 71804154] and [Fundamental Research Funds for the Central Universities #2] under Grant [number 20720181030].

ORCID iD

Shanshi Li

Appendix 1: Measures of post-trip retrospective evaluations

Measure	Item	Reference
Satisfaction	‘Overall, I am satisfied with my experience at Shell Dream World’; ‘My decision to visit Shell Dream World was a wise one’; ‘As a whole, I have really enjoyed myself at Shell Dream World’	Ali et al. (2016)
Intention to recommend	‘I will recommend Shell Dream World to other people’; ‘I will say positive things about Shell Dream World to other people’; ‘I will encourage friends and relatives to visit Shell Dream World’	Prayag et al. (2015)
Revisit intention	‘I intend to revisit Shell Dream World again’; ‘It is very likely that I will revisit Shell Dream World in the future’; ‘The likelihood of my return to Shell Dream World for another travel is high’	Su and Hsu (2013)

References

Algie

(2005) A comparison of viewers’ cognitive and psychophysiological responses to threat-appeal TV advertisements. In: Ghallachir

(ed) Academy of Marketing: Marketing: Building Business, Shaping Society Conference. Dublin: Academy of Marketing, pp.1–12.

Ali

Kim

Ryu

(2016) The effect of physical environment on passenger delight and satisfaction: Moderating effect of national identity. Tourism Management 57: 213–224.

Bagozzi

Gopinath

Nyer

(1999) The role of emotions in marketing. Journal of the Academy of Marketing Science 27: 184–206.

Barrett

Wager

(2006) The structure of emotion: evidence from neuroimaging studies. Current Directions in Psychological Science 15: 79–83.

Bastiaansen

(2019) Emotions as core building blocks of an experience. International Journal of Contemporary Hospitality Management 31: 651–668.

Baumgartner

Sujan

Padgett

(1997) Patterns of affective reactions to advertisements: the integration of moment-to-moment responses into overall judgments. Journal of Marketing Research (JMR) 34: 219–232.

Bhargave

Montgomery

(2013) The social context of temporal sequences: why first impressions shape shared experiences. Journal of Consumer Research 40: 501–517.

Bigné

Andreu

Gnoth

(2005) The theme park experience: an analysis of pleasure, arousal and satisfaction. Tourism Management 26: 833–844.

Biswas

Labrecque

Lehmann

, et al. (2014) Making choices while smelling, tasting, and listening: the role of sensory (dis)similarity when sequentially sampling products. Journal of Marketing 78: 112–126.

10.

Braithwaite

Watson

Jones

, et al. (2013) A guide for analysing electrodermal activity (EDA) & skin conductance responses (SCRs) for psychological experiments. Psychophysiology 49: 1017–1034.

11.

Brodien Hapairai

Walters

(2018) The effectiveness of ad-induced emotion in reducing tourist risk perceptions towards politically unstable destinations. Tourism Recreation Research 43(4): 483–496.

12.

Campbell

(1980) The Sense of Well-Being in America. New York, NY: McGraw-Hill.

13.

Cronin

JJJ

Taylor

(1994) SERVPERF versus SERVQUAL: reconciling performance-based and perceptions-minus-expectations measurement of service quality. Journal of Marketing 58: 125–131.

14.

Del Bosque

Martín

(2008) Tourist satisfaction a cognitive-affective model. Annals of Tourism Research 35: 551–573.

15.

Empatica (2019) E4 Wristband [Online]. Available at: https://www.empatica.com/research/e4/ (accesssed 20 July 2019).

16.

Floyd

(1997) Pleasure, arousal, and dominance: exploring affective determinants of recreation satisfaction. Leisure Sciences 19: 83–96.

17.

Fredrickson

(2000) Extracting meaning from past affective experiences: the importance of peaks, ends, and specific emotions. Cognition & Emotion 14: 577–606.

18.

Fredrickson

Kahneman

(1993) Duration neglect in retrospective evaluations of affective episodes. Journal of Personality & Social Psychology 65: 45–55.

19.

Hair

Marko

Christian

, et al. (2012) An Assessment of the Use of Partial Least Squares Structural Equation Modeling in Marketing Research. Journal of the Academy of Marketing Science 40(3): 414–433.

20.

Han

Back

Barrett

(2009) Influencing factors on restaurant customers’ revisit intention: the roles of emotions and switching barriers. International Journal of Hospitality Management 28: 563–572.

21.

Henseler

Ringle

Sinkovics

(2009) The use of partial least squares path modeling in international marketing. Advances in International Marketing 20: 277–319.

22.

Hoerger

(2013) ZH: An updated version of Steiger’s Z and web-based calculator for testing the statistical significance of the difference between dependent correlations. Retrieved from http://www.psychmike.com/dependent_correlations.php

23.

Hosany

Prayag

(2013) Patterns of tourists’ emotional responses, satisfaction, and intention to recommend. Journal of Business Research 66: 730–737.

24.

Hosany

Prayag

Veen

RVD

, et al. (2017) Mediating effects of place attachment and satisfaction on the relationship between tourists emotions and intention to recommend. Journal of Travel Research 56(8): 1079–1093.

25.

Hughes

(1992) Realtime Response Measures Redefine Advertising Wearout. Journal of advertising research 32(3): 61–77.

26.

Hui

Meyvis

Assael

(2013) Analyzing moment-to-moment data using a Bayesian functional linear model: application to TV show pilot testing. Social Science Electronic Publishing 33: 222–240.

27.

Hume

Mort

(2010) The consequence of appraisal emotion, service quality, perceived value and customer satisfaction on repurchase intent in the performing arts. Journal of Services Marketing 24: 170–182.

28.

Jiang

Huang

, et al. (2020) Worry and anger from flight delay: antecedents and consequences. International Journal of Tourism Research 22: 289–302.

29.

Johnson

Stewart

(2005) A Reappraisal of the Role of Emotion in Consumer Behavior: Traditional and Contemporary Approaches. Bingley: Emerald Group Publishing Limited.

30.

Kahneman

(2000) Evaluation by moments: past and future. In: Kahneman

DATA

(ed) Choices, Values, and Frames. New York, NY: Russel Sage Foundation.

31.

Kahneman

Fredrickson

Schreiber

, et al. (1993) When more pain is preferred to less: adding a better end. Psychological Science 4: 401–405.

32.

Kemp

Burt

CDB

Furneaux

(2008) A test of the peak – end rule with extended autobiographical events. Memory & Cognition 36: 132–138.

33.

Kim

Fesenmaier

(2015) Measuring emotions in real time implications for tourism experience design. Journal of Travel Research 54: 1–11.

34.

Knudsen

(2017) Ikea and the peak-end rule: how to secure that last impressions are lasting impressions. Nyhedsbrevet om Forbrugeradfærd 27: 3–9.

35.

Lang

(1980) Behavioral Treatment and Bio-behavioral Assessment: Computer Applications. In: Sidowski

Johnson

Williams

(eds) Technology in Mental Health Care Delivery Systems. Norwood, NJ: Ablex, pp. 119–137.

36.

Labarge

Dacin

Minor

, et al. (2004) Competitive paper session: hedonic consumption. Advances in Consumer Research 31: 316–328.

37.

Pratt

Wang

, et al. (2020) How to win the consumer’s heart? Exploring appraisal determinants of consumer pre-consumption emotions. International Journal of Hospitality Management 88: 102542.

38.

Lee

, et al. (2008) Festivalscapes and patrons’ emotions, satisfaction, and loyalty. Journal of Business Research 61: 56–64.

39.

Scott

Walters

(2015) Current and potential methods for measuring emotion in tourism experiences: a review. Current Issues in Tourism 18: 805–827.

40.

Walters

Packer

, et al. (2016) Using skin conductance and facial electromyography to measure emotional responses to tourism advertising. Current Issues in Tourism 21(15): 1761–1783.

41.

Walters

Packer

, et al. (2018) A comparative analysis of self-report and psychophysiological measures of emotion in the context of tourism advertising. Journal of Travel Research 57: 1078–1092.

42.

Gao

Scott

, et al. (2013) Customer delight from theme park experiences: the antecedents of delight based on cognitive appraisal theory. Annals of Tourism Research 42: 359–381.

43.

Scott

Gao

, et al. (2016) Delighted or satisfied? Positive emotional responses derived from theme park experiences. Journal of Travel & Tourism Marketing 34: 1–19.

44.

Martin

O’neill

Hubbard

, et al. (2013) The role of emotion in explaining consumer satisfaction and future behavioural intention. Journal of Services Marketing 22: 224–236.

45.

Mauss

Robinson

(2009) Measures of emotion: a review. Cognition & Emotion 23: 209–237.

46.

Miron-Shatz

(2009) Evaluating multiepisode events: boundary conditions for the peak-end rule. Emotion 9: 206–213.

47.

Mitas

Yarnal

Adams

, et al. (2012) Taking a ‘peak’ at leisure travelers’ positive emotions. Leisure Sciences 34: 115–135.

48.

Montgomery

Unnava

(2009) Temporal sequence effects: a memory framework. Journal of Consumer Research 36: 83–92.

49.

Moon

Yoon

Han

(2016) Role of airport physical environments in the satisfaction generation process: mediating the impact of traveller emotion. Asia Pacific Journal of Tourism Research 21: 193–211.

50.

Morris

Woo

Geason

, et al. (2002) The power of affect: predicting intention. Journal of Advertising Research 42: 7–18.

51.

Mukherjee

Lau-Gesk

(2016) Retrospective evaluations of playful experiences. Journal of Consumer Marketing 33: 387–395.

52.

Nawijn

(2010) The holiday happiness curve: a preliminary investigation into mood during a holiday abroad. International Journal of Tourism Research 12: 281–290.

53.

Nawijn

Brüggemann

Mitas

(2017) The effect of Sachsenhausen visitors’ personality and emotions on meaning and word of mouth. Tourism Analysis 22: 349–359.

54.

Nawijn

Fricke

(2013) Visitor emotions and behavioral intentions: the case of concentration camp memorial Neuengamme. International Journal of Tourism Research 17: 221–228.

55.

Nawijn

Fricke

(2015) Visitor emotions and behavioral intentions: the case of concentration camp memorial Neuengamme. International Journal of Tourism Research 17: 221–228.

56.

Otto

Ritchie

JRB

(1996) The service experience in tourism. Tourism Management 17: 165–174.

57.

Polsfuss

Hess

(1991) ‘Liking’ through moment-to-moment evaluation; identifying key selling segments in advertising. Advances in Consumer Research 18: 540–544.

58.

Prayag

Hosany

Muskat

, et al. (2015) Understanding the relationships between tourists’ emotional experiences, perceived overall image, satisfaction, and intention to recommend. Journal of Travel Research 56(1): 41–54.

59.

Prayag

Hosany

Odeh

, et al. (2013) The role of tourists’ emotional experiences and satisfaction in understanding behavioral intentions. Journal of Destination Marketing & Management 2: 118–127.

60.

Kim

(2011) A model of destination branding: integrating the concepts of the branding and destination image. Interdisciplinary Journal of Research in Business 32: 465–476.

61.

Ravaja

(2004) Contributions of psychophysiology to media research: review and recommendations. Media Psychology 6: 193–235.

62.

Ravaja

Kallinen

(2005) Effects of the rate of computer-mediated speech on emotion-related subjective and physiological responses. Behaviour & Information Technology 24(5): 365–373.

63.

Ravaja

Somervuori

(2013) Purchase behavior and psychophysiological responses to different price levels. Psychology & Marketing 30: 479–489.

64.

Richins

(1997) Measuring emotions in the consumption experience. Journal of Consumer Research 24: 127–146.

65.

Russell

Mehrabian

(1974) An Approach to Environmental Psychology. Cambridge: MIT Press.

66.

Scherer

(1993) Studying the emotion-antecedent appraisal process: an expert system approach. Cognition & Emotion 7: 325–355.

67.

Schreiber

Kahneman

(2000) Determinants of the remembered utility of aversive sounds. Journal of Experimental Psychology General 129: 27–42.

68.

Shoval

Schvimer

Tamir

(2018) Real-time measurement of tourists’ objective and subjective emotions in time and space. Journal of Travel Research 57: 3–16.

69.

Speigel

(1998) Getting there is half the fun: Relating happiness to health. Psychological Inquiry 9: 66–68.

70.

Spreng

Harrell

Mackoy

(1995) Service recovery: impact on satisfaction and intentions. Journal of Services Marketing 9: 15–23.

71.

Stadler

Jepson

Wood

(2018) Electrodermal activity measurement within a qualitative methodology: Exploring emotion in leisure experiences. International Journal of Contemporary Hospitality Management 30(11): 3363–3385.

72.

Steiger

(1980) Tests for comparing elements of a correlation matrix. Psychological Bulletin 87: 245–251. doi:10.1037/0033-2909.87.2.245

73.

Hsu

(2013) Service fairness, consumption emotions, satisfaction, and behavioral intentions: the experience of Chinese heritage tourists. Journal of Travel & Tourism Marketing 30: 786–805.

74.

Tasci

ADA

Milman

(2019) Exploring experiential consumption dimensions in the theme park context. Current Issues in Tourism 22: 853–876.

75.

Thomas

Olsen

Murray

(2018) Evaluations of a sequence of affective events presented simultaneously: an investigation of the peak-end rule. European Journal of Marketing 52(3–4): 866–881.

76.

Venables

Mitchell

(1996) The effects of age, sex and time of testing on skin conductance activity. Biological Psychology 43: 87–101.

77.

Yuksel

(2007) Shopping risk perceptions: effects on tourists’ emotions, satisfaction and expressed loyalty intentions. Tourism Management 28: 703–713.

78.

Zhang

Zhu

(2008) Exploring emotion in teaching: emotional labor, burnout, and satisfaction in Chinese higher education. Communication Education 57: 105–122.