The analysis was performed using the IBM SPSS Statistics software (v. 19).
It includes univariate statistics (Table 16) bivariate correlations (Table 18, 18) and regression analysis. Except of two cases (which were discarded), there were no missing data, so the remaining 122 cases were included in all analyses.
Table 16 - Univariate descriptive analysis
Category Minimum Maximum Mean Std. Deviation Purchase
Valid N (all variables, all categories) 122
All figures are rounded to at most 3 decimal digits
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
28
The univariate analysis results (presented in Table 16), and especially the mean values, reveal interesting findings about the purchase behaviour and attitudes of Taiwanese consumers – the purchase history means are noticeably low (relatively to the medium value) and the strength of this phenomenon seems to be inversely related to the price and lifespan of the products of the respective categories. Expectedly, the brand familiarity means are relatively higher for the car brands and lower for bags and shoes. The COO means are those which approach the medium value the closest of all the variables concerned and its standard deviations are among the lowest, which might suggest, that if more categories were included in the set of possible answers for the question measuring the COO image, it could possibly have measured this variable more accurately. The means for quality perception are quite high for all categories, which suggests that the EU products are perceived by the Taiwanese consumers as being of high quality, while the relatively low standard deviation seems to demonstrate, that the levels of such quality perception are quite stable across the sample.
The means of price perception are high, especially for cars and bags, which is expectable, given that EU is home for many well-known luxurious brands (e.g. Mercedes Benz or BMW for cars, Chanel or Louis Vuitton for bags and shoes); interesting fact is, that the value perception mean is considerably lower for the Cheese category, which seems to reflect the informal everyday consumer observation that cheese and similar products are all rather overpriced in Taiwan, regardless of origin or quality and there is little competition from domestic production of this category. The purchase intention and the purchase intention of EU products are both close to the medium value.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
29
The analysis of these two variables revealed strong correlation between them (see Table 18), and since both of these variables are related to purchase intention, the decision was taken to merge them in an aggregate variable.
This new variable is labelled as “Total purchase intention” and calculated as a sum of the “Purchase intention” (Q6) variable and the “EU product purchase intention” (Q7), where the “EU product purchase intention”
is assigned double value, for it is specifically measuring the attitude towards the purchase of products originating in EU, which is the main object of study in this research. The calculation can be expressed by the following formula:
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝ℎ𝑇𝑇𝑎𝑎𝑎𝑎 𝑖𝑖𝑖𝑖𝑇𝑇𝑎𝑎𝑖𝑖𝑇𝑇𝑖𝑖𝑇𝑇𝑖𝑖 = 𝑃𝑃𝑝𝑝𝑝𝑝𝑝𝑝ℎ𝑇𝑇𝑎𝑎𝑎𝑎 𝑖𝑖𝑖𝑖𝑇𝑇𝑎𝑎𝑖𝑖𝑇𝑇𝑖𝑖𝑇𝑇𝑖𝑖 + 2 ∗ 𝐸𝐸𝐸𝐸 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝ℎ𝑇𝑇𝑎𝑎𝑎𝑎 𝑖𝑖𝑖𝑖𝑇𝑇𝑎𝑎𝑖𝑖𝑇𝑇𝑖𝑖𝑇𝑇𝑖𝑖
This decision is supported by the results of Cronbach’s Alpha tests (Table 17) which were well above the required significance level 0.7 (as recommended in European Social Survey 2013) for all categories. Thus calculated variable will be used as dependent variable in all the relevant analyses.
Table 17 - Relation between "purchase intention" and "purchase intention of EU products" variables
Pearson Correlation Sig. (2-tailed) Cronbach's Alpha
Cars .704
**.000 .826
Bags .578
**.000 .728
Cheese .681
**.000 .802
Shoes .657
**.000 .791
**. Correlation is significant at the 0.01 level (2-tailed)
‧ 4.4.1 Bivariate relationships analysis
Table 18 - Bivariate correlations with the dependent variable
Method Cars Bags Cheese Shoes
** Correlation is significant at the 0.01 level (2-tailed)
* Correlation is significant at the 0.05 level (2-tailed) Dependent variable: Total purchase intention
Table 18 shows the bivariate correlations between the specific variables and the dependent variable (Total purchase intention). There seems to be a moderately strong correlation for the brand familiarity and purchase history in all the relevant categories except for cars – this might possibly be related to the low purchase history and purchase intention means for this category. The COO variable does not seem to be strongly significant for any category except bags. The quality variable seems to be significantly connected to purchase intention, as its correlation is exceptionally strong in all four categories and the quality variable correlation coefficients are some of the highest of the entire model. The value perception variable has significant positive correlation only in the cars and bags categories.
Among the personal variables, the only ones, which seem to be correlated with the purchase intention are the age in the bags and cars categories and income in the cars category, the correlation is negative in these cases.
‧
Table 19 - Correlations for hypothesised bivariate relationships
Cars Bags Cheese Shoes H1a Purchase history → Purchase intention
Pearson Correlation
0.053 .398** .417** .234**
Sig. (2-tailed) 0.564 0.000 0.000 0.009 H1b
Brand familiarity → Purchase intention
Pearson Correlation
0.162 .279** .301**
Sig. (2-tailed) 0.075 0.002 0.001 H2a
Purchase history → COO
Pearson Correlation
0.032 0.128 0.118 -0.047 Sig. (2-tailed) 0.728 0.161 0.197 0.605 H2b
Brand familiarity → COO
Pearson Correlation
.236** .216* .307**
Sig. (2-tailed) 0.009 0.017 0.001 H3
COO → Quality perception
Pearson Correlation
0.140 .343** 0.007 0.071 Sig. (2-tailed) 0.125 0.000 0.943 0.434 H4
COO → Value perception
Pearson Correlation
0.129 -0.001 0.012 0.001 Sig. (2-tailed) 0.157 0.991 0.896 0.993 H5
Quality perception → Purchase intention
Value perception → Purchase intention
Pearson Correlation
.295** .268** 0.103 0.022 Sig. (2-tailed) 0.001 0.003 0.259 0.810
** Correlation is significant at the 0.01 level (2-tailed)
* Correlation is significant at the 0.05 level (2-tailed)
Analysis of bivariate correlations relevant to the hypothesised variable relationships (Table 19) seems to support H1a and H1b for all categories except cars (cheese is excluded from H1b); H2a is not supported for any category; H2b is supported strongly for cars and shoes and moderately (on 0.05 level) for bags (cheese is excluded from this hypothesis); H3 is supported only for bags; H4 is not supported at all; H5 is supported for all categories and H6 only for cars and bags (as observed above).
The relevance of the relationships revealed by the correlation analysis will be further tested by the means of regression analysis.
‧ 4.4.2 Regression analysis
4.4.2.1 Cars
Table 20 - Regression analysis results - Cars
Coefficientsa
Model Unstandardized
Coefficients of the Estimate
Change Statistics
R Square Chg. F Chg. df1 df2 Sig. F Chg.
1 .431 .186 .143 2.354 .186 4.373 6 115 .001
2 .436 .190 .140 2.313 .004 .610 1 114 .437
3 .554 .307 .251 2.182 .117 9.419 2 112 .000
a. Dependent Variable: Total Purchase Intention - Cars
For the Cars category, the regression analysis results seem to only support the significance of Income (negative correlation) and the Quality (positive correlation) variables. This implies that the quality of the product is the strongest purchase motivator.
‧
Table 21 - Regression analysis results - Bags
Coefficientsa
Model Unstandardized
Coefficients of the Estimate
Change Statistics
R Square Chg. F Chg. df1 df2 Sig. F Chg.
1 .555 .308 .272 2.354 .308 8.518 6 115 .000
2 .581 .338 .297 2.313 .030 5.148 1 114 .025
3 .649 .421 .374 2.182 .083 8.036 2 112 .001
a. Dependent Variable: Total Purchase Intention - Bags
As in previous case, the regression analysis results show that the Income variable is negatively associated with the purchase intention, while quality has a significant positive correlation. Furthermore, age seems to have an even stronger negative significance than the income variable – this phenomenon might be connected with the fashion aspect of products of this category, which can be expected to be especially important for the younger generations.
The age variable also has a weaker, non-significant negative correlation in Cars (B -0.439, Sig. 0.140) and Shoes (B -0.337, Sig. 0.164) categories.
This partial relevance of age in the other two categories might hint
‧
at the presence of a general trend of consumers of older generations manifesting stronger ethnocentrism and preference of domestic products, which is observed in several works related to the research of consumers’
attitudes towards imported products, like Han (1988) or Chryssochoidis et al. (2007). The plausibility of such explanation seems to be accentuated by the fact that age is irrelevant to the formation of purchase intention in the case of the Cheese category, within which there is virtually no domestic production which could be preferred to the imported products. Another variable, which is important here is the purchase history.
4.4.2.3 Cheese
Table 22 - Regression analysis results - Cheese
Coefficientsa
Model Unstandardized
Coefficients of the Estimate
Change Statistics
R Square Chg. F Chg. df1 df2 Sig. F Chg.
1 .447 .200 .165 2.628 .200 5.799 5 116 .000
2 .447 .200 .158 2.639 .000 .001 1 115 .971
3 .619 .383 .339 2.338 .183 16.732 2 113 .000
a. Dependent Variable: Total Purchase Intention - Cheese
In the Cheese category, the variables with significant correlation are the Purchase history and Quality variables.
‧
The association between the purchase history and purchase intention may be viewed as indirectly supporting the H1a hypothesis implying that the increased level of product familiarity (hereby represented by purchase history) would have a direct positive effect on the purchase intention. The irrelevance of this variable for the other categories may be explained by the low levels of purchase history for these categories, as its significance is inversely related to its means here (cars - mean: 0.3 Sig.:0.182, shoes - mean: 0.89, Sig.: 0.07).
4.4.2.4 Shoes
Table 23 - Regression analysis results - Shoes
Coefficientsa
Model Unstandardized
Coefficient of the Estimate
Change Statistics
R Square Chg. F Chg. df1 df2 Sig. F Chg.
1 .423 .179 .137 2.696 .179 4.189 6 115 .001
2 .425 .180 .130 2.706 .001 .154 1 114 .696
3 .492 .242 .181 2.625 .061 4.538 2 112 .013
a. Dependent Variable: Total Purchase Intention - Shoes
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
36
In the case of the shoes category, the quality also shows strong correlation, which makes the H5 stating, that “High quality perception will strengthen the purchase intention” the only hypothesis undoubtedly confirmed by the regression analysis for all four categories.
Unlike the other categories, in this category, the Brand familiarity is significant, which is an expectable result, as brand name is an important aspect of product evaluation of shoes, as demonstrated for example in Elliott & Cameron (1994:4).
The brand name importance for shoes is only accentuated by the fact that Europe (and especially Italy, which is worlds’ number-two exporter of leather shoes) is home for many well-known (in many cases, highly luxurious) brands, like Gucci, Prada or Giorgio Armani (as mentioned before).
Furthermore, the European shoes are highly-evaluated commodity, as shown, for example, in Wall and Heslop (1986; as quoted in Al-Sulaiti and Baker 1998:5) – a study of attitudes of Canadian consumers towards domestic and imported products, in which the Italian-made shoes were the only product category, which was rated highly than all the domestic products.
One of the regression analysis results, which appears to be rather paradoxical, is the observation that income is related negatively to the purchase intention for cars and bags (meaning, that the respondents of the lower income groups are more willing to purchase the relatively expensive EU products).
This phenomenon might be related to the positive role of price as discussed in (Lichtenstein, Ridgway, and Netemeyer 1993:2-3) and especially the Prestige sensitivity, where the higher price of the product is supposed to affect positively the perception of the social status of the purchaser. It can be expected, that this effect may be particularly desirable for the respondents from the lower income groups.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
37
This kind of argument seems to be supported by further analysis of the COO image variable – Table 20 shows the relative importance of the “Prestige”
aspect for the different income groups – this importance is calculated as the ratio of the count of the respondents of the specific income groups choosing the “Prestige” aspect of the COO image against the other three aspects – it can be observed that, generally, thus calculated ratio tends to be higher especially in the lower parts of the income scale, indicating the higher proportion of the respondents of these groups choosing the prestige attribute.
Table 24 - The relative importance of brand prestige aspect of the COO image variable
Cars Bags Shoes Cheese
≦ 20 000 .367 .340 .372 .280
20 001 - 40 000 .458 .315 .292 .093
40 001 - 60 000 .270 .200 .171 .267
> 60 001 .410 .297 .225 .071
Total .382 .295 .271 .170
Extract from Income-COO cross-tab analysis - see Appendix 14
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
38
5 Conclusions
The analysis has revealed several important observations. European Union and its brands proved to be relatively well established in the mindset of Taiwanese Consumers, nevertheless there are countries, like Sweden and Germany for cars and Italy for Bags, which are preferred by the consumers.
Nonetheless, the COO image, which has been the core concept of this research does not seem to be very relevant for Taiwanese consumers when deciding on purchasing products originating in EU.
The quality of the EU products is evaluated highly and this fact appears to be the main contributor to the purchase intention (and also the most important aspect of the EU COO image). Generally speaking, the European products seem to be perceived as rather luxurious goods and this fact seems to also reflect on the products’ image and to a certain extent to even contribute to the purchase intention. It is an interesting question, whether an increased trading in European goods of more affordable categories would remove this effect or decrease the high quality perception, however this research does not provide an answer to this question.
The product familiarity, (expressed in this research by purchase history level), has quite strong effect on the purchase intention and its promotion by the importing agents, by activities like product presentations or trials is likely to promote the purchase intention.
It can also be suggested that it would be beneficial for Taiwanese international trading ties, if steps were taken towards restructuring of the tariff policies with greater focus on establishing better relationships with strong trading partners (instead of using preferential trade agreements to promote political goals), or even general diminishing of trading barriers (which are especially significant for the imports of cars) for all partners.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
39
6 Limitations and potential future research
It was necessary in this work to gather as many responses as fast as possible, however, if there had been enough space for a more comprehensive research employing more extensive questionnaires, it is probable that the country-of-origin image could be measured more precisely either by redesigning the variable representation in the questionnaire, or by refining the means of the measurement.
It is probable, that more accurate results would have been achieved, if the luxurious goods and more affordable products could have been distinguished in the research, however such procedure would probably have required for the research sample to be controlled much more precisely, as it would probably be necessary to address especially the higher-income respondents in order to get relevant answers to research questions concerning highly luxurious goods.
Furthermore, as there are several, differently perceived, importers present in the market, a more comprehensive study of country-of-origin image should encompass comparison with other large importers – in the case of the Taiwanese market, USA and Japan could probably constitute especially suitable candidates for an interesting comparison with EU.
‧
Trade defence measures taken by EU against Taiwan
Product Type Status Type
of measures
Level (min - max)
Bicycles Anti-Dumping Expired ad-valorem 2,4%-18,2%
Bicycles forks Anti-Dumping Term.
Bicycles frames Anti-Dumping Term.
Capacitors Anti-Dumping Term. ad-valorem 10,7%-75,8%
Compact disks - recordable Anti-Dumping Expired ad-valorem 17,7% - 38,5%
DVD+/-R Anti-Dumping Term.
Electronic weighing scales Anti-Dumping Expired ad-valorem 5,5%-13,4%
Glyphosate Anti-Dumping Repealed ad-valorem 29.9%
Hair brushes Anti-Dumping Term.
Hot-rolled coils Anti-Dumping Term. ad-valorem 2,1%-24,9%
Hot-rolled coils Anti-Subsidy Expired ad-valorem 0%-4,4%
Laser optical reading systems Anti-Dumping Term.
Lighters Anti-Dumping Expired specific 0,065 EUR/lighter
Magnetic disks Anti-Dumping Term. ad-valorem 19,8%-32,7%
Monosodium glutamate Anti-Dumping Expired ad-valorem 20.40%
Monosodium glutamate Anti-Dumping Term. undertaking 20.40%
Open mesh fabrics of glass fibres Anti-Dumping Definitive ad-valorem 48,4%-62,9%
Peroxosulphates (persulphates) Anti-Dumping Expired ad-valorem 22.60%
Personal fax machines Anti-Dumping Term. ad-valorem 6,0%-36,6%
Polyester high tenacity filament yarn Anti-Dumping Term.
Polyester high tenacity filament yarn Anti-Dumping Term.
Polyester staple fibres Anti-Dumping Term. ad-valorem 5,9%-13,0%
Polyester staple fibres Anti-Dumping Term. ad-valorem 14,7%-29,5%
Polyester staple fibres Anti-Subsidy Term.
Polyester textured filament yarn
(PTY) Anti-Dumping Term. ad-valorem 0%-16,1%
Polyethylene terephthalate (PET) Anti-Dumping Expired specific 36,3-143,4 €/tonne Polyethylene terephthalate (PET) Anti-Subsidy Term.
Polyvinyl alcohol (PVA) Anti-Dumping Term.
SBS thermoplastic rubbers Anti-Dumping Repealed ad-valorem 5,3%-20,0%
SBS thermoplastic rubbers Anti-Subsidy Expired ad-valorem 1,0%-8,2%
Silicon metal (silicon) Anti-Dumping Definitive Stainless steel cold-rolled flat
products Anti-Dumping Term.
Stainless steel fasteners and parts
thereof Anti-Dumping Def/IR ad-valorem 8,8%-23,6%
Stainless steel fasteners and parts
thereof Anti-Dumping Expired ad-valorem 5,3%-23,1%
Stainless steel tube and pipe
butt-welding fittings Anti-Dumping Term.
Tube and pipe fittings, of iron or steel Anti-Dumping Term.
Tube and pipe fittings, of iron or steel Anti-Dumping Definitive ad-valorem 58.60%
Tube and pipe fittings, of iron or steel Anti-Dumping Term.
Woven glass fibre fabrics Anti-Subsidy Term.
Woven glass fibre fabrics Anti-Subsidy Term.
Source: European Commission (European Commission n.d.)