Demographic and socio-economic factors

國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

26

Chapter 4:

Data Analysis

The key objective of the research is to understand the factors that are driving customers towards Luxury goods and services in Taiwan. Out of all these, demographic data, shopping trends and purchase factors are measures to better understand Taiwanese behavior towards luxury consumption.

Since the demographic factors and the consumption of luxury goods may not be inter linked, qualitative analysis was performed to understand the impact of the demographic factors over the consumption of luxurious products and services factor by assessing the descriptive statistics and the count of responses towards the luxury categories.

All analyses were conducted using MS Excel and Stata.

4.1. Demographic and socio-economic factors

On 71 surveys, 64 were valid for data analysis. Table 1 shows the composition of the survey sample. The sample is relatively balanced by gender as it is composed of 31 men (44.44%) with mean age 38.5 and 33 women (55.56%) with mean age 32.16.

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

27

Table 2:

Sample composition

Gender Frequency Mean age

Male 31 38.5

Female 33 32.16

Total 64 35.17

Before assessing the determinants of luxury consumption, a new binary dependent variable was created from the survey responses. This new variable takes value 1 if the respondent consumes luxury products and services, and 0 otherwise. From the 64 valid responses, 46 respondents (71.9%) do effectively consume luxury products and services in Taïwan.

This high proportion of positive responses demonstrates the relevance of further analyses.

However, in order to minimize errors, some observations what do not represent enough information for data analysis (such as Corporte executive and not employed) have been dropped automatically by the Sata program representing a smaller sample of 55.

In an attempt to define the profile of luxury consumers in Taïwan, the association between the dependent variable and demographic and socio-economic factors was tested using appropriate statistical tests. The results are presented hereafter for each association.

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

28

LUXURY CONSUMPTION VERSUS AGE

Figure 1 and Table 2 below show the descriptive statistics for the age variable in each category of the binary response variable of luxury consumption.

Figure 5:

Boxplot age versus luxury consumption

Table 3:

Descriptive statistics of age versus luxury consumption

Lux Binary N Mean Std. Deviation Std. Error Mean Age

0 18 37.67 15.423 3.635

1 46 34.48 14.776 2.179

The mean age is about 34.48 years in the series of 46 respondents who consume luxury products and about 37.67 years in the series of 18

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

29

respondents who do not. We therefore notice a mean age difference of about 3.188 in absolute value between the 2 categories of the binary response variable, with a similar standard deviation from the mean in the two groups. A Levene’s test for the equality of variances between the two groups was performed before conducting the appropriate t-test. A Student t-test for the equality of means was then implemented to see if the mean age difference reported above is statistically significant from 0. The associated results are presented in Table A.1 in the Appendix. The Levene’s test fails to reject the null hypothesis of variance equality at the usual significance level of 5% (p-value > 5%). Therefore, the appropriate t-test assumes equality of variances between the two groups and the results show that we cannot reject the null hypothesis that the mean between the two groups are statistically different. Hence, there is no statistically significant difference in terms of age between those who consumes luxury products and those who do not.

LUXURY CONSUMPTION VERSUS CATEGORICAL DEMOGRAPHIC FACTORS

Table 3 presents the counts or absolute frequencies for each association between the categories luxury consumption (binary) and the other categorical demographic factors.

‧

Demographic factors - Frequencies and chi-square tests

Variables Categories Lux Binary Pearson chi2 test p-value

Significance level: *1%, **5%,***10%

From this table, it is clear that there is no absolute difference between women and men among those who consume luxury products, while the difference remains marginal among those who do not consume. We notice a difference in terms of absolute counts towards luxury consumption for those who are only-child. Indeed, on 11 only child basis, 10 claim consuming luxury good. In a larger sample, the P-value of 0.123 may be significant at 10% level. However, Pearson chi-square tests fail to reject the null hypothesis of independence between luxury consumption and the each of the three categorical demographic factors, which suggests that there is no statistical difference in terms of gender, only-child and children between those who consume and those who do not consume luxury products in Taïwan. Hence, hypothesis 1, only child may consume more luxury goods and services, may not be statistically verified in this research.

‧

LUXURY CONSUMPTION VERSUS CATEGORICAL SOCIO-ECONOMIC FACTORS

The results of counts and Pearson chi-square independence tests between the dependent, binary response variable (luxury consumption) and the other categorical socio-economic variables in the sample are presented in Table 4.

Table 5:

Socio-economic factors - Frequencies and chi-square tests

Variables Categories Lux Binary Pearson chi2 test (2-sided p-value)

‧

Significance level: *1%,**5% ,***10%

From these results, we observe some differences in terms of counts but the associated Pearson’s chi-square tests all fail to reject the null hypothesis of independence between luxury consumption and the other categorical socio-economic variables. One potential reason is the small sample size leading to statistically non-significant results.

Therefore, there is no statistical evidence of a clear profile of luxury consumption among the interviewed respondents. However, based on the count results presented in Table 4, below is a list of potential characteristics that could be associated with luxury consumption and that could eventually lead to significant differences in a larger sample: