Influence of Choice Architecture on the Preference for a Pro-Environmental Hotel

Factors That Influence Sustainable Travel and Moral Hypocrisy in General Research and in Travel Research

Numerous studies have reported travelers’ pro-environmental behavior; that is, travelers demonstrate high intentions to stay in a green hotel (Han, Hsu, and Sheu 2010; Han and Yoon 2015) and pay more for green products and services (Kang et al. 2012; Masau and Prideaux 2003). This condition suggests the growing concern of tourists over environmental issues and their consequences. A TripAdvisor-based survey reports that 62% of travelers consider environmental issues during their stay in hotels (Pathak 2015). The majority of hotel guests (i.e., 87%) appreciate the significance of eco-friendly hotels (Slye 2009). The increasing awareness of travelers and their participation in green management practices have prompted the lodging industry to actively offer green hotel attributes (e.g., use of recycled materials, waste reduction, water and energy conservation) and to engage in sustainable hotel management, which are believed to enhance guest loyalty (Han, Hsu, and Sheu 2010; Wu, Ai, and Cheng 2016) and increase profits (Singh, Cranage, and Lee 2014).

The pro-environmental behavior of travelers can be construed from value–belief–norm theory and theory of planned behavior. As adopted from previous studies (Anable, Lane, and Kelay 2006; Bamberg, Ajzen, and Schmidt 2003; Han, Hsu, and Sheu 2010; Shaw, Shiu, and Clarke 2000), the value–belief–norm theory of environmentalism (Stern 2000) and the theory of planned behavior elucidate that values, beliefs, and attitudes toward positive consequences of pro-environmental behaviors enhance individuals’ intention to behave sustainably and motivate actual eco-friendly behavior (Ajzen 1985). People hold values and beliefs about environmental issues; thus, after being exposed to green hotel attributes, they are responsible for protecting the environment, which leads to environmentally responsible behavior (Hopper and Nielsen 1991; Stern 2000).

However, theory of planned behavior is criticized for the weak relationship between behavioral intentions and actual behavior given that behavioral intentions do not necessarily translate to environmentally responsible behavior (Bergin-Seers and Mair 2009; McKercher and Tse 2012). This phenomenon is frequently observed in the context of sustainable tourism, wherein positive green attitude does not properly predict eco-friendly travel choices (Juvan and Dolnicar 2014) due to the attitude–behavior gap in environmentally responsible travel (Becken 2004; Bergin-Seers and Mair 2009).

One plausible explanation for the attitude–behavior gap in sustainable tourism activities is moral hypocrisy. Moral hypocrisy occurs when people would like to be seen as moral “while, if possible, avoiding the cost of actually being moral” (Batson and Thompson 2001, p. 54). This motivation is designed to hold a pretension of maintaining publicly approved attitudes, norms, and principles that people do not actually engage in. Moral hypocrisy is generally observed when individuals are engaged in self-deception (Batson et al. 1999). Individuals tend to be involved in self-deception, which allows them to avoid the perceived gap between their behavior and moral standards (Trivers 1985).

Baumeister and Newman (1994) used an intuitive lawyer to describe a person in self-deception. An intuitive lawyer selectively adopts evidence and infers from it to reach a predetermined conclusion. Self-deception, which is integral to moral hypocrisy, is closely associated with the reasoning of an intuitive lawyer when coming to a preferred conclusion (Baumeister and Newman 1994). Attribution theory, together with self-deception, can provide further explanation for the attitude–behavior gap in unsustainable vacation choices. Attribution theory states that travelers value and select environmentally sustainable hotels because they attribute the cause of environmental degradation and other direct/indirect harmful effects of tourism to themselves (Juvan and Dolnicar 2014). However, when travelers engage in self-deception, they justify their unsustainable tourism practices by attributing the source of climate change to others rather than to themselves, thereby exhibiting the attitude–behavior gap with moral hypocrisy.

In the next section, we suggest choice architecture as one of the key methods for reducing moral hypocrisy in travel and tourism situations.

Choice Architecture

The concept of “choice architecture” was first suggested by Thaler and Sunstein (2008). They explained choice architecture as different ways of presenting the choice problem to decision makers, thereby ultimately influencing choice outcome. Various forms of choice architecture have been suggested in different areas, such as the number of options (Chernev 2003; Iyengar and Lepper 2000; Scheibehenne, Greifeneder, and Todd 2010), the existence or form of default options (Brown and Krishna 2004; Johnson and Goldstein 2003), manner of option presentation (J. Kim, Hwang, Park, et al. 2019; Mogilner, Shiv, and Iyengar 2012; Simonson 1990), and decision mode.

First, the number of options can determine the outcome and process, such that decision makers tend to feel frustrated in finalizing their choice when numerous options are presented. Iyengar and Lepper (2000) provided the empirical evidence that shoppers’ actual purchase was higher when they were exposed to a limited number of jams (e.g., 6 jams with different flavors) than when they were exposed to large number of jams (e.g., 24 jams with different flavors). They theorized low purchase from large (vs. few) options as a result of choice overload because of the huge number of options available. Chernev (2003) extended this work by suggesting ideal point availability as the key moderating variable for the negative effect of large options.

Second, default option also significantly influences the final choice, such that organ donation rate across different nations can be significantly influenced by the default option. Johnson and Goldstein (2003) provided the empirical evidence that the actual donation rate was extremely high for countries where donating organs was the default option (i.e., opt-out for donation, such as Austria, Belgium, France, or Hungary), whereas the rate was extremely low for countries where the default option was not to donate (i.e., opt-in for donation, such as Germany, Denmark, or England). Finally, under situations with multiple available options, the manner of presentation (i.e., simultaneous presentation vs. sequential presentation) influences choice outcome differently, such that people demonstrate higher variety-seeking behavior (Simonson 1990) and a higher level of choice outcome satisfaction (Mogilner, Shiv, and Iyengar 2012) in simultaneous (vs. sequential) presentation.

In the travel and tourism setting, some researchers have investigated these factors of choice architecture in choice overload (J. Y. Park and Jang 2013; Sthapit 2017; Thai and Yuksel 2017), role of default option (Jin, He, and Song 2012; Lawson and Thyne 2001), and manner of presentation (J. Kim, Kim, and Kim 2018). For example, Thai and Yuksel (2017) suggested that self-confidence reduces the negative effect of choice overload. In particular, in situations wherein a large number of travel options are available, travelers with high (vs. low) self-confidence presented a positive evaluation of the final destination by reducing the uncertainty of their decision.

In the default setting, Jin, He, and Song (2012) investigated travelers’ preference for two package tour options: low-quality/low-price option versus high-quality/high-price option. They found that travelers’ preference was significantly influenced by the default option; that is, they preferred upgrading to downgrading. Kim et al. (2018) showed that the specific manner in which options are presented significantly affected final choice outcomes. In particular, when travelers booked different services (e.g., hotel and airline ticket) together in the form of a travel package, their final choice considerably differed when the package was selected from a combined format (i.e., simultaneous choice) rather than from a separated format (i.e., sequential choice). However, the explicit application of the concept of choice architecture to travel and tourism research remains limited (Hall 2013; Leon and Araña 2014). In this study, we focus on the decision mode (i.e., choice vs. rejection), which is discussed in the next section.

Two Types of Decisions: Choice versus Rejection

When individuals make decisions, they typically face two contrasting decision tasks: choice versus rejection. The literature on choice versus rejection suggests that decision-making processes and decision outcomes are dependent on the decision task in which an individual is engaged in, which arises from the task’s characteristics (Chernev 2009; Huber et al. 1982; Laran and Wilcox 2011; Shafir 1993). With regard to differences in the decision-making process between choice and rejection tasks, research has shown that people tend to consider the choice task more important than the rejection task given that they require stronger justifications for making their choice among alternatives (Wedell 1997).

With respect to differences in decision outcomes between choosing and rejecting tasks, research has reported that people search for different types of information when they choose versus reject available options (Ganzach 1995; Nagpal, Lei, and Khare 2015; Shafir 1993; Wedell 1997). Shafir (1993) proposed that in the context of a choice task, individuals place more emphasis on the positive features of the options because the goal is to choose the option that is most appealing to justify their choice decision. In a rejection task, the goal is to eliminate options that do not appeal to individuals. Consequently, individuals place more weight on the negative features of the options to justify their rejection decision.

Similarly, Laran and Wilcox (2011) suggested that individuals are likely to focus on preference-consistent alternatives in a choice task, whereas they tend to focus on preference-inconsistent features in a rejection task. For example, when the goal of a person is to buy a mobile phone for business, he or she focuses on business-use mobile plans in the choice task and on student-use mobile plans in the rejection task (Laran and Wilcox 2011). Mourali and Nagpal (2013) argued that people with high power are more likely to focus on the positive side of options, which results in option selection. By contrast, those with low power tend to focus on the negative side of options, thereby resulting in option rejection. In summary, individuals primarily weigh positive attributes in a choice task and negative attributes in a rejection task.

Sokolova and Krishna (2016) suggested that rejection and choice tasks differ in terms of information processing level. Their study provided empirical evidence that participants in a rejection task tend to use a more deliberative information processing strategy that requires cognitive resource or effort. By contrast, participants in a choice task rely on a less deliberative information processing strategy that is unaffected by the limitations of cognitive resource.

Unlike in other disciplines, such as psychology (Shafir 1993) and marketing (Laran and Wilcox 2011; Nagpal, Lei, and Khare 2015; C. W. Park, Jun, and Macinnis 2000; Sokolova and Krishna 2016), only a few studies have addressed the issue of decision tasks (i.e., choice vs. rejection) and their impact on travelers’ decision making in the tourism literature. In fact, the terms choice and rejection have, by default, been frequently regarded as homogenous in the tourism literature, with minimal attention given to the differences between choice and rejection tasks (Karl, Reintinger, and Schmude 2015; Woodside and King 2001). Karl, Reintinger, and Schmude (2015) assumed that the reasons for the rejection of destinations in a choice task are the same as those for destination selection in their analysis of tourists’ destination choice processes. Similarly, Woodside and King (2001) considered choice (buying) decisions similar to rejection decisions in travelers’ decision-making processes. Furthermore, the issue of booking cancellation (cf. a rejection task) has frequently been examined in the tourism literature from the perspective of operators (Murphy and Chen 2016; Smith, Parsa, Bujisic, and van der Rest 2015) rather that of customers.

A few exceptional studies in the tourism literature have regarded option choice differently from option rejection (Oppewal, Huybers, and Crouch 2015; Perdue and Meng 2006). Perdue and Meng (2006) examined the impacts of choice and rejection tasks on the importance of destination attributes in a set of decision considerations. They found that when travelers were asked to choose their winter vacation destination, quality attributes (e.g., snow and lodging quality) were more important than others, whereas price attributes (e.g., lodging prices and price of access) were more important than others when they were asked to reject destination options. Oppewal, Huybers, and Crouch (2015) noted that prescreening information during the initial stage of decision making (cf. a rejection task) is not as important to a traveler as the information considered in the final stage of decision making (cf. a choice task). In summary, the topic of different decision tasks has received minimal attention from tourism researchers compared with that in other disciplines.

Main Prediction: The Impact of the Decision Mode on the Preference for a Sustainable Option

In this study, we assume that selection by either choosing or rejecting in the final stage of travel booking is closely related to the behavioral (rather than attitudinal) aspect, by following other studies (Ajzen and Fishbein 1969; Lam and Hsu 2006; Lanken et al. 1994). On the basis of this assumption, we predict that the decision mode will influence the preference for a sustainable hotel option based on two main reasons.

First, as suggested in the previous literature (Sokolova and Krishna 2016), travelers in the rejection decision mode are likely to engage in deliberative information processing, whereas those in the choice decision mode tend to rely on less deliberative information processing. Additional research has shown that a sustainable option will be preferred, particularly by highly involved or deliberative individuals (De Boer, Hoogland, and Boersema 2007; Hedlin and Sunstein 2016; Vermeir and Verbeke 2006); thus, we expect that the preference for a sustainable hotel option will be higher in the rejection than in the choice decision mode.

Second, assume that a traveler is facing two options: one option is superior in terms of sustainable attributes (e.g., environment-friendly services), whereas the other one is superior in terms of nonsustainable attributes (e.g., better service quality or price by disregarding the environmental factor). If the traveler chooses the nonsustainable option over the sustainable one, then he or she may feel guilty for sacrificing social value for economic or personal value. However, the salience of not choosing a sustainable hotel option can differ for the choice and rejection modes. In the rejection mode, people explicitly reject the socially right option, and thus, they may feel a relatively strong emotion of guilt. By contrast, people in the choice mode implicitly reject the sustainable option, and thus, their feeling of guilt will not be as strong. The literature of default option will support this argument by showing that the participating rate for organ donation will be significantly higher when the donation decision is opt-out rather than opt-in (Johnson and Goldstein 2003; Theotokis and Manganari 2015).

We predict that travelers’ preference for a sustainable (over a less sustainable) option will be higher in the rejection than in the choice decision mode based on this logic. Our argument can be additionally supported by other factors. First, the current literature suggests that people tend to choose a socially acceptable option over a less acceptable option to justify their decision (Okada 2005). We expect that travelers in the rejection (vs. choice) decision mode will have lower justification in their decision if they selected the less sustainable over the sustainable option. Second, morality is generally related to the form of “do not act” in religion. For example, eight of the “Ten Commandments” are expressed in negative form (i.e., do not do something rather than do something). Therefore, people in the rejection mode feel a strong moral emotion of guilt.

In summary, we expect different preferences for a sustainable hotel option for the choice versus rejection task based on the two aforementioned reasons. The formal predictions regarding the main effect of decision mode on the preference for a sustainable travel option and its underlying mechanism are as follows:

Method: Subjects, Design, and Procedure

A total of 130 adults from the United States participated in the study through an online panel (Amazon MTurk) for a nominal compensation (M_age = 35.11, SD = 9.99, 38.2% female). The participants were randomly assigned to one of the two (decision modes: choice vs. rejection) under between-subjects experimental conditions.

First, the participants were asked to read the description that they had planned a weeklong vacation. They were also requested to imagine that they had two hotel options, but they could no longer retain their reservation for both hotels. The hotels differed in terms of several attributes, including quality based on customer rating, location, linen service, bathroom toiletries, and price, as shown in Figure 1. The sustainable attributes were manipulated with two attributes (i.e., linen service and bathroom toiletries). In particular, hotel A was better in terms of sustainable attributes, such as “linen service: sheets changed upon request only” and “bathroom toiletries: refillable shampoo and soap dispensers,” whereas hotel B was less sustainable in terms of both attributes, such as “linen service: sheets changed every day” and “bathroom toiletries: disposable shampoo and soap bottles.” The quality of the hotels was relatively high (i.e., 4.5 customer rating out of 5), and the price per day was also high (i.e., US$250). After reading the description of the hotels, the participants were asked to proceed to the next page where they finalized their choice based on two experimental conditions, which provide the same information of the hotels. Participants under the choice condition were asked to choose one vacation spot (e.g., “Which reservation would you choose? Please review carefully and select one.”). By contrast, participants under the rejection condition were requested to indicate their rejection option (e.g., “Which reservation would you cancel? Please review carefully and reject one.”). We consider the participants’ choices as selecting the sustainable hotel when they selected hotel A and selecting the less sustainable hotel when they selected hotel B.

OPEN IN VIEWER

Figure 1.

Stimuli for experiments 1A and 1B.

Subsequently, the participants were asked to rate their relative perceived sustainability of the two hotels (i.e., “Which hotel do you perceive to demonstrate a more sustainable practice?”) using an 11-point scale (0 = hotel A, 10 = hotel B). Finally, the participants were asked to provide their background variables, including ethnical background, income level, education level, gender, and age. The profile of the participants across all studies is provided in Table 1.

OPEN IN VIEWER

Table 1.

Profile of Participants.

	Experiment 1A(n = 130)	Experiment 1B(n = 122)	Experiment 2(n = 103)	Experiment 3(n = 155)	Experiment 4(n = 193)
Gender
Male	60.8	49.2	50.5	67.7	49.7
Female	39.2	50.8	49.5	32.3	50.3
Age, years
18–29	33.1	36.9	35.9	51.6	32.1
30–39	36.9	39.3	30.1	29.7	34.2
40–49	22.3	10.7	23.3	11.0	20.2
50–59	4.6	8.2	6.8	3.2	6.2
60–	3.1	4.9	3.9	4.5	7.3
Race
White	73.1	73.8	78.6	48.4	75.2
African American	8.5	4.1	3.9	5.8	9.3
Hispanic	4.6	6.6	2.9	2.6	5.7
Asian	10.8	10.7	10.7	38.7	6.7
Others	3.1	4.9	3.9	4.4	3.1
Educational level
Did not complete high school	0.0	1.6	0.0	1.9	0.0
High school graduate or college level	36.9	33.6	36.9	30.3	43.0
College graduate (4 years)	52.3	54.1	49.5	51.6	42.5
Postgraduate degree	10.8	10.7	13.6	16.1	14.5
Family income
≤$30,000	20.0	20.5	24.3	21.9	17.6
$30,001–$60,000	30.8	37.7	28.2	25.8	39.9
$60,001–$90,000	20.0	24.6	18.4	20.6	22.3
$90,001–$120,000	16.2	11.5	17.5	14.2	11.4
≥$120,001	13.1	5.7	11.7	17.4	8.8

Results

First, the manipulation of sustainable hotels was successful given that the perceived sustainability was less than the neutral value of 5 in the 11-point scale, M = 4.32, SD = 3.36, t(129) = −2.30, p <.05. As expected in hypothesis 1, the decision mode significantly influenced the final selection between the two hotels, χ²(1) = 6.98, p <.01. In particular, the sustainable hotel (i.e., hotel A) was selected by 35.9% (23/64) of the participants in the choice decision mode and 59.1% (39/66) of participants in the rejection decision mode, as shown in Figure 2.

OPEN IN VIEWER

Figure 2.

Results of experiments 1A, 1B, 2, and 3.

In theorizing, we assume that the participants in the rejection decision mode will rely on relatively elaborative information processing. To empirically test this assumption, we analyzed the decision time (i.e., total duration of decision time) across two experimental conditions. First, we excluded one participant whose decision time would be the outliers, with data over three standard deviations from the mean (new n = 129). The results indicated that the covariate, F(1, 126) = 21.76, p <.001, and the main effect of the decision mode, F(1, 126) = 4.02, p <.05, were significant. In particular, the decision time for the rejection task (M = 22.51 s, SD = 15.66) took longer than that for the choice task (M = 19.05 s, SD = 12.64). On the basis of these data, we could infer that different levels of deliberative processing from the two decision modes influenced the selection results.

Method: Subjects, Design, and Procedure

A total of 122 adults from the United States participated in the study through an online panel (Amazon MTurk) for a nominal compensation (M_age = 34.97, SD = 11.26, 50.8% female). The participants were randomly assigned to one of the two (decision modes: choice vs. rejection) under between-subjects experimental conditions.

The general procedure for this study was relatively similar to that for experiment 1A with a few modifications. The hotels were the same in terms of the key attributes. However, the detailed information of the hotels was different. The quality of the hotels was relatively high (i.e., 3.0 customer rating out of 5.0), and the price per day was also high (i.e., US$120). Other information was the same as that in experiment 1A, as shown in Figure 1.

Results

As expected in hypothesis 1 and similar to experiment 1A, the decision mode significantly influenced the final selection of the two hotels, χ²(1) = 7.26, p <.01. That is, the sustainable hotel (i.e., hotel A) was selected by 33.3% (21/63) of the participants in the choice decision mode and 57.6% (34/59) of the participants in the rejection decision mode, as shown in Figure 2.

For the decision time, we excluded one participant whose decision time would be the outliers, with data over 3 standard deviations from the mean (new n = 121). The results of analysis of covariance (ANCOVA) indicated that the covariate (i.e., time duration of reading the instructions) was significant, F(1, 118) = 23.55, p <.001. Moreover, the main effect of the decision mode was also significant—F(1, 118) = 42.30, p <.001—given that the decision time for the rejection task (M = 17.78 s, SD = 9.88) took longer than that for the choice task (M = 7.65 s, SD = 7.82). The results of this study showed that the different levels of deliberative processing from the two different decision modes influenced the selection results.

Method: Subjects, Design, and Procedure

A total of 103 adults from the United States participated in the study through an online panel (Amazon MTurk) for a nominal compensation (M_age = 35.76, SD = 10.95, 49.5% female). The participants were randomly assigned to one of the two (decision modes: choice vs. rejection) under between-subjects experimental conditions.

The general stimuli and procedure for this study were relatively similar to those of study 1B but include a few mediators. First, a sustainable hotel was positioned on the left side of the information figure in all the previous studies. Now that the specific location of an item can influence choice (Bar-Hillel 2015; J. Kim, Hwang, Park, et al. 2018; Romero and Biswas 2016), we used different positions for the sustainable hotel (e.g., the sustainable hotel is located at the right side of the less sustainable hotel) as shown in Figure 3. Second, with regard to the general procedure, the participants were presented with two hotels and asked to either choose or reject one. After the selection, they were requested to answer questions related to their feeling of guilt, that is, “I feel guilty if I do not participate in the linen reuse program/I feel accountable about not helping to protect the environment,” modified from Dahl, Honea, and Manchanda (2005) using a 7-point scale (1 = strongly disagree, 7 = strongly agree). Since two items were highly correlated (r = .85, p <.001), we used the mean for further analysis. We also measured their perceived sustainability using the 7-point scale for a manipulation check.

OPEN IN VIEWER

Figure 3.

Stimuli for experiments 2 and 3.

Results

First, the manipulation of sustainable hotels was successful given that the perceived sustainability was more than the neutral value of 4 in the 7-point scale, M = 5.72, SD = 1.75, t(102) = 9.99, p <.001. Second, similar to experiment 1B, the decision mode significantly influenced the final selection of the two hotels, χ²(1) = 6.80, p <.01. In particular, the sustainable hotel (i.e., hotel B) was selected by 32.7% (18/55) of the participants in the choice decision mode and 58.3% (28/48) of the participants in the rejection decision mode. These results supported hypothesis 1, as shown in Figure 2.

Third, the decision mode significantly influenced expected guilt. Expected guilt was higher among participants in the rejection task (M = 3.88, SD = 1.91) than those in the choice task (M = 3.15, SD = 1.70), F(1, 101) = 4.23, p <.05. To test the mediating role of guilt on the choice decision mode, we conducted a mediation analysis based on Hayes’s SPSS macro-modules (with 5,000 bootstrapping for model 4; Hayes, Montoya, and Rockwood 2017). The mediator was expected guilt. The results showed that the indirect effect of expected guilt (95% bootstrap confidence interval [CI]: 0.0267, 1.253) did not contain zero and was significant. However, the direct effect of the decision mode on the preference was insignificant (95% CI: −0.060, 1.828). Therefore, the result supported hypothesis 2.

Method: Subjects, Design, and Procedure

A total of 155 adults from the United States participated in the study through an online panel (Amazon MTurk) for a nominal compensation (M_age = 32.53, SD = 10.70, 32.3% female). The participants were randomly assigned to one of the two (decision modes: choice vs. rejection) under between-subjects experimental conditions.

The general procedure and stimuli for this study were relatively similar to those of experiment 1B, with a few modifications. The hotels had the same attributes except for price level. The price of the sustainable hotel was US$130, whereas that of the nonsustainable hotel was US$120, as shown in Figure 3. After the selection, the participants were asked to rate their mood using a 7-point scale (1 = very sad/bad, 7 = very happy/good, r = 0.76, p <.001).

Results

First, mood was not influenced by our experimental factors (M_{_chioce} = 5.56, SD = 1.24, vs. M_{_rejection} = 5.50, SD = 1.18), F (1, 153) = 0.11, p = 0.74. Therefore, we will not discuss mood further. As expected in hypothesis 1 and similar to experiment 1A, the decision mode significantly influenced the final selection between the two hotels, χ²(1) = 9.61, p <.01. The sustainable hotel (i.e., hotel A) was selected by 16.7% (13/78) of participants in the choice decision mode and 39.0% (30/77) of the participants in the rejection decision mode, as shown in Figure 2. The general preference for the less sustainable hotel (i.e., hotel B) over the sustainable hotel (i.e., hotel A) was expected given the relatively cheaper price.

For decision time, we excluded three participants whose decision time would be outliers, with data over 3 standard deviations from the mean (new n = 152). The results of ANCOVA indicated that the covariate was insignificant, F(1, 149) = 0.05, p >.10. Moreover, the main effect of the decision mode was significant, F(1, 149) = 4.48, p <.05, given that the decision time for the rejection task (M = 21.00 s, SD = 11.44) took longer than that for the choice task (M = 17.20 s, SD = 10.43). The results of this experiment showed that the different levels of deliberative processing for the two different decision modes influenced the selection results.

Method: Subjects, Design, and Procedure

A total of 193 adults from the United States participated in the study through an online panel (Amazon MTurk) for a nominal compensation (M_age = 36.81, SD = 11.88, 50.3% female). The participants were randomly assigned to one of the two (decision modes: choice vs. rejection) × two (response modes: dichotomous selection vs. relative rating) under between-subjects experimental conditions.

The general procedure and stimuli for this study were relatively similar to those of experiment 2, with few modifications. The hotels had the same attributes except for price level. In this experiment, we used hotels with a relatively high price/quality. The price of the sustainable hotel was US$260, whereas that of the less sustainable hotel was US$250, as shown in Figure 2.

The participants under the dichotomous selection condition were given the same instructions as those in experiment 2, whereas participants under the relative selection condition were asked to express their choice or rejection using a 7-point scale (1 = definitely hotel A, 7 = definitely hotel B). Afterward, all the participants were requested to answer their relative perceived sustainability for the two hotels (i.e., “Which hotel do you perceive demonstrate better sustainability practice?”) using a 7-point scale (1 = hotel A, 7 = hotel B).

Results

First, the manipulation of sustainable hotels was successful given that the perceived sustainability was less than the neutral value of 4 in the 7-point scale, M = 2.52, SD = 1.76, t(192) = −1.48, p <.001. Second, to compare the two experimental conditions of selection and rating, we calculated the inferred choice based on the relative rating (we excluded three participants who responded with 4 in the 7-point scale in further analysis).

We conducted a bilogistic analysis with the dependent variable (i.e., final selection outcome; 1 = choice of sustainable hotel, 2 = choice of nonsustainable hotel), independent variable (i.e., decision mode), and moderator (i.e., response mode). The results of the bilogistic analysis indicated the significant interaction effect (b = 1.81, Wald = 8.11, p >.01) as shown in Figure 4. In particular, when the participants were asked to express their preference from the dichotomous selection, we replicated the previous studies given that the sustainable hotel (i.e., hotel A) was selected by 20.8% (10/48) of the participants in the choice decision mode and 49.0% (24/49), χ²(1) = 8.44, p <.01, of the participants in the rejection decision mode. By contrast, the significant difference disappeared when participants were asked to express their preference from the relative rating. In particular, the sustainable hotel (i.e., hotel A) was similarly selected given that 39.1% (18/46) of the participants in the choice decision mode selected the sustainable hotel, whereas 27.7% (13/47), χ²(1) = 1.38, p = 0.24, of the participants in the rejection decision mode selected the same sustainable hotel.

OPEN IN VIEWER

Figure 4.

Results of experiment 4.

Summary of the Experiments

To test the research hypotheses, we have deliberately designed a series of experimental studies. In experiment 1A, we found that the main effect of decision modes (choice vs. rejection) on the preference for a pro-environmental hotel was significant and decision time in the rejection task took longer than that in the choice task. Experiment 1B was designed to address a potential limitation of the findings of experiment 1A given that the quality and price levels of hotels may influence travelers’ behavioral intention. Therefore, we replicated experiment 1A under a relatively low-quality and low-price hotel selection situation. We found that the main effect of decision modes (choice vs. rejection) on the preference for a sustainable hotel was significant and decision time in the rejection task took longer than that in the choice task. This result suggests that people engaged in a rejection task take more time to decide between a sustainable option and a less sustainable option and, consequently, are more likely to choose a sustainable option over a less sustainable one.

In experiment 2, the research focus was on the underlying mechanism of the effect of decision modes on the preference for a pro-environmental hotel over a less pro-environmental one by adding a mediator, that is, the guilty feeling for choosing a less sustainable hotel instead of a sustainable hotel. As expected, the guilty feeling mediated the link between decision modes and the preference for a sustainable hotel. That is, people are more likely choose a sustainable hotel option in the rejection task because they feel guiltier than those in the choice task when they choose a less sustainable hotel. Experiment 3 was designed to enhance the realism of our research manipulations. A less sustainable hotel mostly tends to have better economical values (i.e., ratio between benefits and costs) than a sustainable hotel; thus, we manipulated the price of hotel options, that is, a high price was tagged on a sustainable hotel. As hypothesized, the price difference between the two hotel options did not confound the impact of the decision modes on the preference for a pro-environmental hotel. Thus, people are more likely to choose a sustainable hotel even at a higher price in the rejection task than in the choice task.

Finally, we tested a moderating role of response modes (i.e., dichotomous selection vs. rating) on the effect of the decision modes on the preference for a sustainable hotel. We found that the impact of the decision modes on the preference for a pro-environmental hotel was significant in the case of selecting one of the two options, whereas the effect of the decision modes on the preference was insignificant in the case of rating the likelihood of choosing either option. This result implies that considering preference revealing formats is important in influencing people’s preference for a sustainable hotel over a less sustainable one in any contextual situation, such as the decision modes in this study.

Theoretical and Practical Implications

This study has several theoretical contributions. First, the results of this study can extend our understanding of pro-environmental behavior in travel and tourism. In contrast to current factors that influence this behavior, our theoretical predictions and empirical findings showed the importance of the choice architecture of decision modes (i.e., choice vs. rejection) in determining the selection of a sustainable hotel over a less sustainable one.

Second, this study could contribute to our understanding of the decision mode. That is, we could extend our understanding of the different roles of choosing and rejecting in option selection in pro-environmental travel behavior. In particular, multiple experiments showed that the decision mode could be related to the guilty feeling arising from selection options, which was not covered in the existing literature.

Third, we integrate the literature on moral hypocrisy into choice architecture to further explain travelers’ decision between less sustainable and sustainable hotels. Furthermore, we suggested the key underlying mechanism for the feeling of guilt in selecting a less sustainable hotel, particularly under the rejection (vs. choice) task. In addition, the measurement of time spent in making a decision supported the underlying process given that the rejection (vs. choice) task generated relatively deliberative processing.

Fourth, in relation to moral hypocrisy, this study could extend our understanding of the factors that could reduce moral hypocrisy. In contrast to the previous literature that identifies factors that eliminate moral hypocrisy, such as self-awareness (Batson et al. 1999), being powerless (Lammers, Stapel, and Galinsky 2010), or religious motivation (Carpenter and Marshall 2009), we suggested that decision modes can reduce moral hypocrisy in a pro-environmental hotel choice. However, our findings are parallel with those of Polman and Ruttan (2012) given that the feeling of guilt can reduce the moral hypocrisy level.

Finally, our results are related to those of Kern and Chugh (2009), who argued that ethical decisions depend on the framing of information. In particular, they provided empirical evidence that people are more likely to behave unethically when the information of action is presented in the framing of losses (e.g., 75% chance of losing X) rather than of gains (e.g., 25% chance of gaining X). These results contradict the results of the current study given that that the current study found that travelers are more likely to behave ethically (e.g., selecting a pro-environmental option) when the decision is framed negatively (e.g., rejecting options) rather than positively (e.g., choosing options). We infer that the specific target of framing, that is, the attribute framing of Kern and Chugh (2009) versus the decision framing of our study, can generate this inconsistency. Future studies should test the difference between the two types of framing in selecting a pro-environmental hotel.

This study has several practical and managerial contributions. First, the most straightforward message of our research is concerned with the fact that decision modes (choice vs. rejection) influence the likelihood of travelers selecting a sustainable hotel in their booking process. This study suggests that the websites of online travel agencies (OTAs) can be designed with multiple booking stages, with a cancellation task in the last stage to boost the possibility of travelers selecting a sustainable hotel over a less sustainable one. For example, travelers can be allowed to select multiple hotel options, including a sustainable hotel, in the initial stage and then asked to delete their less preferred hotel options in the later stage using an OTA’s booking site. In this situation, travelers may initially choose multiple hotel options and then make a cancellation decision later. In the cancellation stage, travelers may select a sustainable hotel over a less sustainable one given the strong feeling of guilt inherent in a rejection task.

Second, the findings of this study can assist in understanding the effect of promotional message when tourism businesses develop tourism packages or products. The findings of this research can be used to effectively frame product packages and persuade and attract customers. For example, marketers may develop promotional information for a sustainable hotel, including situations wherein travelers are involved in a rejection decision. Third, the results of this study provide diverse managerial insights into hotel departments that require sustainability or environmentalism, such as restaurants, kitchen, swimming pool, or sauna. For example, the results of this study can be utilized to develop menus or to establish pricing strategies because they can help understand the varied psychological structures of customers in advance. As concepts of environmentalism, eco-friendliness, sustainability, and corporate social responsibility become more prevalent in the hotel industry, the results of this study will offer managerial values to other tourism business practices, such as resorts, airlines, conventions, and events.

Limitation and Future Research

The current study has several limitations, which suggest the direction of future research. First, this research mainly used the scenario-based method although this method is frequently used in the travel literature (J. Kim, Kim, and Kim 2018; J. H. Kim and Jang 2014). Future studies should investigate this issue by using other methods, such as analyzing actual data from travelers. Second, we mainly used choice as the key measurement. Future studies may use other measurements, such as monetary value (e.g., maximum reservation price) or actual behavior. Third, although we suggested one moderating variable (i.e., response mode: selection vs. rating) that influences the impact of the decision mode on pro-environmental hotel choice in experiment 4, future studies may investigate additional boundaries or moderating factors, such as travel purpose or duration of stay. Fourth, the decision-making structure in choosing a green hotel can be influenced by customers’ cultural background (Chou and Chen 2014), generation (Tang and Lam 2017), social norm, or general environmental behavior. In addition, attitude toward the environment or sustainability can influence green hotel selection (Han and Yoon 2015; Y. Kim and Han 2010). Therefore, future studies should assess the moderating effect of these factors in determining hotel product options.