**3.2. Data collection**

A non-probability sample was used, based on quotas, proportional to the tourist destinations visited in Chile according to the National Tourism Service [81], divided into beaches (36%), lakes (25%) and other attractions (39%). The surveys were conducted in the main capital cities of Chile. Data were collected in the third and last stage. The survey was conducted on a total sample of 750 people, using the last tourist destination where they stayed as reference. A psychometric analysis of the data was conducted in order to obtain scales with a good level of reliability, validity and dimensionality. At the same time, a separate analysis of the multidimensionality of the hedonic benefit scale was conducted (see **Table 2**). The results of the partial analysis and the global model were both satisfactory.


#### **Affective evaluation (Affev)**

**Satisfaction (Sat)**

**Functional benefit (Fube)**

**Hedonic benefit visual attractions (Hbviat)**

**Hedonic benefit escape from routine (Hbero)**

**Hedonic benefit intrinsic pleasure (Hbinple)**

**Cognitive perception environment (Copen)**

**Hedonic benefit recreation (Hbere)**

**Symbolic benefit (Sben)**

Satisfaction Sat 1 This is the best place I have visited

42 Mobilities, Tourism and Travel Behavior - Contexts and Boundaries

Hedonic benefit visual attractions Hbviat 1 I love the appearance of this place

Hedonic benefit escape from routine Hbero 1 This place is an escape from routine

Hedonic benefit recreation Hbere 1 This is a great place to have fun

Symbolic benefit Sben 1 This place reflects who I am

Cognitive perception environment Copen 1 This place is known to be very safe

Functional benefit Fube 1 Vacationing in this place was just what I needed

Sat 2 This place is what I expected Sat 3 This place fulfills my expectations Sat 4 This place is exactly what I imagined Sat 5 This place was my best choice

Hbviat 2 I love the aesthetics of this place Hbviat 3 This place is a pleasure for my senses

Hbere 3 This is a place to enjoy life

Hedonic benefit intrinsic pleasure Hbinple 1 This place is fascinating to visit, compared to other possible activities

Fube 2 In this place, I found the vacation I was looking for Fube 3 Vacationing in this place is always convenient

Hbero 2 This place makes me feel I'm in a different world

Hbere 2 This place is very exciting, which is contagious

Hbinple 2 I wish I could be in this place all the time

Sben 2 This place is in tune with the way I see myself Sben 3 I identify with the people who choose this place

Copen 3 This place is known for its good transport system Copen 4 This place is known to have adequate signage

Sben 4 This place is visited by people like me

Copen 2 This place is known to be clean

Fube 4 Compared to similar places, this is the best vacation spot


**Table 1.** Measurement scales.

Then, an exploratory factor analysis and a confirmatory factor analysis (SEM) were conducted, along with different reliability analyses applying Cronbach's alpha, construct reliability and variance extracted (AVE). In order to identify items unattached to their dimension, principal component factor analyses were conducted with varimax rotation [82]. Following this procedure, there was no need to eliminate indicators from the scales analyzed (see **Table 3**). In fact, they all featured a good level of unidimensionality, with factor loadings well over 0.4 [83]. Considering the different scales, through structural equations, a confirmatory factor analysis was developed in order to confirm if the indicators or variables were adequate for an appropriate adjustment of the model. The requirements considered were the three criteria proposed by [84]. The first is to eliminate the indicators with a weak condition of convergence with their corresponding latent variable. A Student's *t* higher than 2.28 (*p* = 0.01) was used as a requirement. The second criterion is to separate from the analysis those variables with loadings translated into standardized coefficients lower than 0.5. Finally, indicators with a linear relationship *R*<sup>2</sup> lower than 0.3 must be eliminated. For this process, the statistics pack AMOS SPSS version 23 was used. No indicators were eliminated according to any of the three criteria in this analysis. The adjustment indexes of this confirmatory factor model were acceptable: IFI 0.906, CFI 0.905, RMSEA 0.074, Normed χ<sup>2</sup> 5.08.

Once the optimal model was verified, the reliability of each scale was confirmed. Three tests were applied for this: Cronbach's alpha (limit 0.7), composite construct reliability (limit 0.7) [85] and analysis of variance extracted (limit 0.5) [86]. Results show that, in all cases, the minimum values defined (see **Table 4**) by these parameters of reliability are met.


**Table 2.** Multidimensional analysis of hedonic benefit.


**Table 3.** Exploratory factor analysis.


**Table 4.** Reliability of scales.

**Scales Variable Factorial load Variance** 

44 Mobilities, Tourism and Travel Behavior - Contexts and Boundaries

Satisfaction Sat 1 0.82 73.76 3.68

Functional benefit Fube 1 0.86 73.54 2.94

Hedonic benefit visual attractions Hbviat 1 0.91 80.09 2.40

Hedonic benefit escape from routine Hbero 1 0.93 86.99 1.74

Hedonic benefit recreation Hbere 1 0.85 74.51 2.93

Hedonic benefit intrinsic pleasure Hbinple 1 0.89 70.54 1.59

Symbolic benefit Sben 1 0.89 76.4 3.0

Cognitive perception environment Copen 1 0.83 69.34 2.77

Affective evaluation Affev 1 0.90 80.89 2.43

**Table 3.** Exploratory factor analysis.

Fube 2 0.90 Fube 3 0.86 Fube 4 0.80

Hbviat 2 0.91 Hbviat 3 0.85

Hbero 2 0.93

Hbere 2 0.88 Hbere 3 0.87

Hbinple 2 0.89

Sben 2 0.91 Sben 3 0.88 Sben 4 0.82

Copen 2 0.79 Copen 3 0.85 Copen 4 0.86

Affev 2 0.91 Affev 3 0.89

Sat 2 0.87 Sat 3 0.88 Sat 4 0.83 Sat 5 0.88 **explained %**

**Own value**

Validity was confirmed in terms of content and construct. The scales show an adequate level of validity in terms in content, due to the deep analysis of the literature, a critical incident study with tourists and then a clearing process of this scale through different focus groups with tourists and interviews with experts and sales executives from travel agencies. In order to meet the construct validity, it was determined if the scales proposed, already cleared, meet the convergent and discriminant validity. Convergent validity was confirmed by observing that all the standardized coefficients of the confirmatory factor analysis (AFC) were statistically significant to 0.01 and higher than 0.5 [87]. In order to verify the existence of discriminant validity, a confidence interval test was used [88]. It consists in building the confidence intervals resulting from the correlations among the different latent variables that make up the confirmatory factor model (AFC) (see **Table 5**). According to this test, the model has discriminant validity since no confidence interval contained the value 1 [89]. The methodological process developed makes it possible to conclude that the proposed model shows a good level of overall validity.



**Table 5.** Discriminant validity.

Validity was confirmed in terms of content and construct. The scales show an adequate level of validity in terms in content, due to the deep analysis of the literature, a critical incident study with tourists and then a clearing process of this scale through different focus groups with tourists and interviews with experts and sales executives from travel agencies. In order to meet the construct validity, it was determined if the scales proposed, already cleared, meet the convergent and discriminant validity. Convergent validity was confirmed by observing that all the standardized coefficients of the confirmatory factor analysis (AFC) were statistically significant to 0.01 and higher than 0.5 [87]. In order to verify the existence of discriminant validity, a confidence interval test was used [88]. It consists in building the confidence intervals resulting from the correlations among the different latent variables that make up the confirmatory factor model (AFC) (see **Table 5**). According to this test, the model has discriminant validity since no confidence interval contained the value 1 [89]. The methodological process developed makes it possible to conclude that the proposed model shows a good level

 **(df)**

**Bi-variate relationship Confidence intervals Difference** *χ***<sup>2</sup>**

46 Mobilities, Tourism and Travel Behavior - Contexts and Boundaries

Symbolic benefit—satisfaction 0.61–0.64 1956.2(1)

Satisfaction—affective evaluation 0.53–0.55 1972.6 (1) Hedonic benefit—satisfaction 0.82–0.85 1956.4 (1) Satisfaction—functional benefit 0.81–0.85 1966.3 (1) Satisfaction—cognitive perception 0.50–0.54 1955.9 (1) Symbolic benefit—hedonic benefit 0.72–0.75 1959.8 (1)

Symbolic benefit—functional benefit 0.60–0.63 1958.4 (1)

Hedonic benefit—functional benefit 0.84–0.86 1957.7 (1) 1955.9 (385)

0.47–0.50 1961.6 (1)

0.41–0.42 1988.7 (1)

0.47–0.48 1992.7 (1)

0.51–0.53 1971.0 (1)

0.66–0.67 1996.4 (1)

0.61–0.62 1979.3 (1)

0.42–0.45 1962.0 (1)

of overall validity.

**Confidence interval test**

Cognitive perception—functional

Cognitive perception—affective

Symbolic benefit—affective

Hedonic benefit—cognitive

Hedonic benefit—affective

Affective evaluation—functional

Symbolic benefit—cognitive

**Full model**

benefit

evaluation

evaluation

perception

evaluation

benefit

perception
