CHAPTER 4: ANALYSIS OF RESEARCH RESULTS
4.1 Descriptive statistics of the research sample
After the scale was calibrated by interviewing experts, the questionnaire was designed. The total number of survey forms distributed was 378, the number of survey forms collected was 365 and the number of valid surveys was 321. Invalid surveys were eliminated because they did not meet the criteria for screening questions as well as having blank questions. Thus, the number of official research samples used in the study was 321 observations.
Regarding gender: In the descriptive statistics table above, the proportion of men surveyed is 44.5% and the proportion of women surveyed is 55.5%, in which the proportion of women surveyed is higher. Initially, when conducting the survey, the author planned to survey 50% of men and 50% of women. However, because women come to do transactions at travel companies more and are easier to access, the proportion of men and women is different from the original plan. In general, there is not much difference between the proportion of men and women.
In terms of age, the proportion of people aged 35 to 45 years old accounts for the highest proportion of 36.4%, followed by those aged 22 to 35 years old at 30.8%, those over 45 years old at 20.2% and those under 22 years old at 12.5%. The people surveyed are those who come to travel agencies for advice or intend to buy a tour, the age of those who come to book tours is mostly people from 22 to 45 years old, these people are almost able to control their finances and have a high demand for tours. In addition, the demand for foreign travel is also gradually increasing due to stable economic conditions. The age group under 22 years old is mostly dependent on family finances and is quite busy with studying, so the rate of going on tours is low.
The respondents worked in a variety of fields with different occupations. In this study, the main respondents were office workers (45.5%), followed by civil servants (75%), and business owners (75%).
accounting for 16.8%, students accounting for 12.1% and other occupations accounting for 2.2%. When contacted with the survey, office workers said that due to increasing work pressure, they often encounter stress in life, so on weekends or holidays they often book tours to relieve stress. Business owners said that they do not have much time to travel, students are not in control of their finances and have not yet decided on their personal ability to travel.
Regarding education level: Along with the development of the economy and society, the need for learning and improvement, Ho Chi Minh City is one of the places with the most universities and training centers in Vietnam. In addition, people with higher education levels often have higher incomes. In the research sample, 57.6% of the survey participants had university degrees, 24.6% had college degrees and 9% had postgraduate degrees. The proportion of people with high school and intermediate degrees was quite low, 0.6% and 8.1% respectively.
Regarding income: Because they belong to different age groups and occupations, the survey participants also have different income levels. According to the General Statistics Office of Vietnam (2018), the average GDP per capita of Vietnamese people is 2,587 USD/year, equivalent to an average income of about 4,900,000 VND/month. However, in the survey sample, the number of people with an income from 10 million to over 20 million VND is quite high, accounting for about 64.2%, the number of people with an income from 5 million VND to under 10 million VND accounts for 24.0% and the number of people with an income under 5 million VND accounts for 11.8%. Thus, the people surveyed and able to pay for foreign tours are mostly people with high incomes. People with incomes of less than 10 million VND often register for cheap foreign tours such as Thailand, Cambodia, Singapore, etc. People with high incomes often have the ability to register for a variety of tours and register to travel many times a year.
Table 4.1: Demographic descriptive statistics
N= 321
Frequency | % | % accumulated | ||
Sex | Male | 143 | 44.5 | 44.5 |
Female | 178 | 55.5 | 100.0 | |
Age | Under 22 years old | 40 | 12.5 | 12.5 |
22 to under 35 years old | 99 | 30.8 | 43.3 | |
35 to under 45 years old | 117 | 36.4 | 79.8 | |
Over 45 years old | 65 | 20.2 | 100.0 | |
Job | Students | 39 | 12.1 | 12.1 |
Office staff | 146 | 45.5 | 57.6 | |
State official | 75 | 23.4 | 81.0 | |
Business owner | 54 | 16.8 | 97.8 | |
Other | 7 | 2.2 | 100.0 | |
Education | High School | 2 | 0.6 | 0.6 |
Intermediate | 26 | 8.1 | 8.7% | |
College | 79 | 24.6 | 33.3 | |
University | 185 | 57.6 | 91.0 | |
Postgraduate | 29 | 9 | 100.0 | |
Income | Under 5 million VND | 38 | 11.8 | 11.8 |
From 5 million to under 10 million VND | 77 | 24.0 | 35.8 | |
From 10 million to under 20 million VND | 121 | 37.7 | 73.5 | |
Over 20 million VND | 85 | 26.5 | 100.0 |
Maybe you are interested!
-
Reliability Testing of Dependent Variable Scale -
Testing the Reliability of the Scale Using Cronbach'S Alpha Coefficient -
A. Results of Testing the Reliability Coefficient of the Factor Scale from the Enterprise Side -
Testing the Reliability of the Scale Through Cronbach'S Alpha Coefficient -
Preliminary Test of Reliability of Scale in Research Model
Source: SPSS analysis In addition to demographic data, to better understand the characteristics of the data table, the author conducted descriptive statistics on the average value of the scale. Each observed variable
Different scales will have different minimum, maximum, mean and standard deviation values. Below are the descriptive statistics of the observed variables in the scale:
Table 4.2: Statistics of average values of the scale
Variable name
Sample size | Minimum value | Great value best | Medium | Deviation standard | |
CL1 | 321 | 1 | 5 | 2.98 | 1,026 |
CL2 | 321 | 1 | 5 | 2.99 | 1,019 |
CL3 | 321 | 1 | 5 | 2.98 | 0.917 |
CL4 | 321 | 1 | 5 | 2.95 | 1,004 |
CL5 | 321 | 1 | 5 | 3.09 | 1,034 |
TC1 | 321 | 1 | 5 | 3.02 | 1,001 |
TC2 | 321 | 1 | 5 | 3.05 | 0.959 |
TC3 | 321 | 1 | 5 | 2.95 | 0.944 |
TC4 | 321 | 1 | 5 | 3.01 | 1,006 |
HI1 | 321 | 1 | 5 | 3.07 | 1,142 |
HI2 | 321 | 1 | 5 | 2.81 | 1,149 |
HI3 | 321 | 1 | 5 | 3.64 | 0.972 |
HI4 | 321 | 1 | 5 | 3.00 | 1,026 |
HI5 | 321 | 1 | 5 | 2.92 | 0.944 |
SL1 | 321 | 1 | 5 | 3.04 | 0.996 |
SL2 | 321 | 1 | 5 | 2.70 | 0.963 |
SL3 | 321 | 1 | 5 | 2.79 | 0.921 |
SL4 | 321 | 1 | 5 | 2.90 | 0.949 |
CM1 | 321 | 1 | 5 | 2.77 | 1,082 |
CM2 | 321 | 1 | 5 | 2.92 | 1,246 |
CM3 | 321 | 1 | 5 | 2.96 | 1,112 |
CM4 | 321 | 1 | 5 | 2.83 | 1,074 |
YD1 | 321 | 1 | 5 | 3.11 | 0.824 |
YD2 | 321 | 1 | 5 | 2.88 | 0.739 |
YD3 | 321 | 1 | 5 | 2.99 | 0.806 |
Valid form | 321 |
Source: Analysis from SPSS
According to the questionnaire design, the respondents will choose from 1 to 5 corresponding to the statements given. In the table above, it can be seen that all variables are rated the highest as 5 and the lowest as 1. At the same time, the mean and standard deviation of the variables in the research model are different. The above data table shows an overview of the research data table.
4.2 Testing the reliability of the scale
Before conducting data analysis and statistics, to evaluate the reliability of the data set and the proposed scale, the author conducted a reliability test. The reliability test of the scale was performed by examining the Cronbach'Alpha coefficient of the variables and the total variable correlation. The variables were included in the reliability test after the qualitative research step. As presented in chapter 3, the best Cronbach' Alpha coefficient must be greater than 0.8. The scale is reliable when the Cronbach'Alpha coefficient is from 0.7 to 0.8 and for scales with Cronbach'Alpha coefficient must be from 0.6 or higher.
The study is based on scales taken from the original scale or scales that have been adjusted to suit the topic. Through interviews with experts to adjust the scale to suit the Vietnamese tourism industry and the survey subjects. In general, the results of the reliability test of the scale with the Cronabach'Alpha coefficient in the study are quite good. In which, the Cronach'Alpha coefficient of the final accepted scale in the study is greater than 0.8.
Only variable HI3 was removed from the research model because the correlation with the total variable was less than 0.3, and when this variable was removed, the Cronbach'Alpha coefficient of eWOm usefulness was also improved. In addition to HI3, the other variables were retained for further research (the number of observed variables retained was 24 observed variables, these 24 variables were used for EFA and subsequent analysis steps). Below is a summary table of Cronbach'Alpha coefficients and the main coefficients in the reliability analysis of the scale.
Table 4.3: Testing the reliability of the scale using Cronbach'Alpha
Observation variable
Average scale if variable type | Scale variance if variable type | Total variable correlation | Cronbach's Alpha if variable type | |
eWOM quality Cronbach'alpha=0.816 | ||||
CL1 | 12.02 | 9,290 | 0.652 | 0.766 |
CL2 | 12.01 | 9,387 | 0.640 | 0.769 |
CL3 | 12.02 | 10,215 | 0.574 | 0.789 |
CL4 | 12.05 | 9,745 | 0.585 | 0.786 |
CL5 | 11.91 | 9,626 | 0.580 | 0.788 |
eWOM Cronbach'Alpha Reliability = 0.817 | ||||
TC1 | 9.00 | 5,834 | 0.630 | 0.774 |
TC2 | 8.98 | 5,965 | 0.641 | 0.769 |
TC3 | 9.07 | 6,013 | 0.644 | 0.768 |
TC4 | 9.02 | 5,787 | 0.638 | 0.771 |
Usefulness of eWOM Cronbach'Apha=0.808 | ||||
HI1 | 8.74 | 6,594 | 0.630 | 0.757 |
HI2 | 8.99 | 6,528 | 0.638 | 0.753 |
HI4 | 8.80 | 7,204 | 0.606 | 0.768 |
HI5 | 8.88 | 7,436 | 0.635 | 0.758 |
Number of eWOM Cronbach'Alpha=0.809 | ||||
SL1 | 8.40 | 5,527 | 0.599 | 0.774 |
SL2 | 8.74 | 5,595 | 0.614 | 0.765 |
SL3 | 8.65 | 5,590 | 0.662 | 0.743 |
SL4 | 8.54 | 5,599 | 0.629 | 0.758 |
eWOM sender expertise Cronbach'Alpha=0.825 | ||||
CM1 | 8.71 | 8,156 | 0.660 | 0.775 |
CM2 | 8.57 | 7,353 | 0.666 | 0.774 |
CM3 | 8.52 | 8,019 | 0.659 | 0.775 |
CM4 | 8.65 | 8,397 | 0.620 | 0.793 |
Intention to purchase foreign tour Cronbach'Alpha=0.868 | ||||
YD1 | 5.87 | 1,993 | 0.764 | 0.800 |
YD2 | 6.10 | 2,200 | 0.777 | 0.792 |
YD3 | 5.99 | 2,131 | 0.709 | 0.851 |
Source: Results from SPSS
4.2.1 Scale of eWOM quality variables
In the eWOM quality variable scale, the Cronbach' Alpha coefficient of the entire scale is 0.816 > 0.8, which means that the eWOM quality variable scale is good. The total correlation coefficient of all variables is greater than 0.3. Besides, if any variable in the eWOM quality scale is removed, it is impossible to achieve a higher Cronbach' Alpha result. Therefore, the eWOM scale is accepted.
4.2.2 eWOM trustworthiness variable scale
For the eWOM reliability variable, the Cronbach' Alpha coefficient is 0.817 > 0.8; the coefficients of the observed variables in the total variable correlation are all greater than 0.3 and when removing any variable in the scale, the Cronbach' Alpha coefficient will decrease. From that, we can see that the scale of the eWOM reliability variable is good and reliable.
4.2.3 The usefulness scale of eWOM
The independent variable usefulness is measured by HI1, HI2, HI3, HI4, HI5; the first analysis result of Cronbach'Alpha coefficient of usefulness variable scale is 0.745. However, variable HI3 is invalid when the correlation coefficient with the total variable is only 0.166 < 0.3; at the same time, if variable HI3 is removed, the Cronbach' Alpha coefficient of the scale will increase significantly (increasing to a value of 0.808). From the above table, it can be seen that variable HI3 is an unreliable variable in the usefulness scale of eWOM. Variable HI3 was removed and the reliability of the usefulness scale was re-tested for the second time. After removing the observed variable HI3, the Cronbach' Alpha coefficient of the usefulness of eWOM increased to 0.808. If any of the variables HI1, HI2, HI4, HI5 were removed, the Cronbach' Alpha coefficient would decrease. In addition, the total correlation of the variables was greater than 0.3. Thus, we removed HI3 for the following analysis and statistics in the study.
4.2.4 Scale of eWOM quantity variables
The correlation coefficient between the total variable of variables SL1, SL2, SL3, SL4 is all greater than 0.3, the Cronbach' Alpha coefficient if the above variables are removed will decrease compared to the Cronbach' Alpha coefficient of 0.809. The scale of eWOM quantity variable is good and reliable.
4.2.5 Scale of eWOM sender expertise variable
The eWOM sender's expertise variable is measured by 4 observed variables including: CM1, CM2, CM3, CM4. The scale of the sender's expertise variable is good and acceptable with Cronbach's Alpha coefficient of 0.825, the correlations with the total variable are all greater than 0.3.
4.2.6 Purchase intention variable scale
Purchase intention is the only dependent variable in the research model. Purchase intention is measured by 3 observed variables: YD1, YD2, YD3 with a large Cronbach' Alpha coefficient of 0.868. If 1 of the 3 variables or 2 of the 3 variables above are eliminated, the Cronbach' Alpha coefficient will decrease, and the total correlation coefficient after analysis is also greater than 0.3. The scale of purchase intention variable is good and acceptable.
4.3 Exploratory factor analysis EFA
4.3.1 Exploratory factor analysis for independent variables
After testing the reliability of the scale, the author eliminated variable HI3 because this variable made the scale measuring the usefulness of eWOM unreliable. The remaining variables will be included in the EFA exploratory factor analysis.
Table 4.4: KMO coefficient test
Kaiser-Meyer-Olkin: Complete Sampling Method
0.864 | ||
Bartlett's Test of Sphericity | Approx. Chi-Square | 2764,145 |
df | 210 | |
Sig. | 0.000 |
Source: Statistics from SPSS KMO index = 0.864 > 0.5 so factor analysis is suitable for the research data. In addition, the significance level sig = 0.000 < 0.05 so the observed variables are correlated with each other in the whole. The KMO test results are completely valid for analysis.
factor