Tourists’ Experiences with Smart Tourism Technology at Smart Destinations and Their Behavior Intentions

Table 1.
Demographic Profile (n=1,010).
Characteristics
Frequency (%)
Gender
Male
Female
500 (49.5)
510 (50.5)
Ethnicity
White
African American
Asian
American Indian or Alaska Native
Native Hawaiian or Pacific Islander
Other
752 (74.5)
126 (12.5)
77 (7.6)
18 (1.8)
13 (1.3)
24 (2.4)
Generation
Baby boomers
Generation X
Generation Y
Generation Z
129 (12.8)
175 (17.3)
590 (58.4)
116 (11.5)
Employment Status
Employed full time
Unemployed
Employed part time
Retired
Self-employed or business owner
552 (54.7)
166 (16.2)
140 (13.9)
82 (8.1)
70 (6.9)
Income Level
Less than $50,000
$50,001 to $70,000
$70,001 to $90,000
$90,001 to $110,000
$110,001 to $130,000
$130,001 to $150,000
More than $150,000
414 (41.0)
181 (17.9)
142 (14.1)
111 (11.0)
66 (6.5)
29 (2.9)
67 (6.6)
Education Level
High school graduate
Associate degree in college
Bachelor’s degree in college
Postgraduate degree
346 (34.3)
225 (22.3)
323 (32.0)
116 (11.5)
Smart Tourism Destination Visited
Boston, MA
Chicago, IL
New York City, NY
San Francisco, CA
Seattle, WA
209 (20.7)
196 (19.4)
225 (22.3)
208 (20.6)
172 (17.0)
Purpose
For business
For pleasure
For both business and pleasure
Other
83 (8.2)
670 (66.3)
240 (23.8)
17 (1.7)
Smart devices used at the destination
Smartphone
Tablet
Camera
Apple watch
965 (95.5)
524 (51.9)
435 (43.1)
179 (17.7)

---

Jeong and Shin 
7
charging stations, and traffic routing apps more than females 
did. Three types of tourists, business, leisure, and bleisure, 
had similar patterns in their use of STTs at the destinations. 
However, both leisure and bleisure travelers used Google 
Maps and ride-sharing program comparatively, while busi-
ness travelers used city guide app and My TSA more often. 
Regardless of age, the respondents used “Google Maps” the 
most. Younger generations (i.e., Generations Y and Z) used 
ride-sharing programs, mobile payment, and charging sta-
tions more frequently than the other generations.
Measurement and Structural Models and 
Hypothesis Testing
CFA indicated that the overall fit of the measurement model 
was satisfactory (
χ
² 
=
853.9, df 
=
186, 
p
=
0.00, CMIN/df 
=
4.591, NFI 
=
0.955, CFI 
=
0.965 IFI 
=
0.965, and 
RMSEA 
=
0.060). As shown in Table 3, the factor loadings 
were greater than 0.73. The factor loadings and average 
amount of variance extracted (AVE) provided evidence for 
the construct validity of the measures at both the item and 
construct levels. All individual item loadings greater than 
0.70 indicated that the constructs were with more measured 
variance than with error variances (Gefen, Straub, and 
Boudreau 2000). An AVE greater than 0.50 indicated that the 
construct shared more variance with its indicators than with 
error variance (Fornell and Larcker 1981). Thus, all mea-
surement items exhibited good convergent validity as sug-
gested. Following Fornell and Larcker (1981), this study 
tested discriminant validity by comparing the correlations 
among constructs to the AVE values. All constructs demon-
strated acceptable discriminant validity because all intracon-
struct correlations were lower than the square root of each 
construct’s AVE, implying that all indicators were better 
explained by their respective construct than the other con-
structs (see Table 4).