Consumption occasion

The level of influence that consumption occasion has on consumers’ behavior and purchase decision has been researched for nearly 50 years (Corsi, Cohen & Lockshin, 2015).

First scholars to examine the concept in wine market were Quester and Smart (1998). They conducted a study on Australian wine consumers and found out that although the factor of region did not play a significant role in the wine purchase for a specific occasion, price, grape variety and style did turned out to be significant (Quester & Smart, 1998).

Furthermore, other two research projects looking at the effect of consumption occasion were conducted on Australian consumers in 2000 and 2001 (Hall & Lockshin, 2000; Hall, Lockshin & O’Mahonney, 2001). Concentrating on eight different occasions the studies tested the level of relevance consumers give to products’ quality signals and personal values and the results showed that some differences can be observed in the importance consumers attach to distinct product features.

In the consumption occasion literature there can also be found studies that provide descriptive approach explaining if and when certain purchase occasion influences consumers’

decision to buy wine. One of this studies was done on Chinese consumers and it considered the connection between culture, price and occasion. Chinese turned out to buy more expensive wines for special, public occasions and for a gift in order not to lose “face” (丟臉) and inexpensive ones for home consumption (Liu & Murphy, 2007).

When it comes to Millennials and their wine consumption on different occasions the number of studies is very limited. Latest analysis of young consumers was conducted in 2011

(Fountain &Lamb, 2011; Thach, 2011). Thach (2011) as well as Fountain and Lamb (2011) found that American and New Zealander Millennial consumers are willing to spend more money on wine when purchasing for special occasion and that they, on average, consume wine during formal celebrations and dining in a fancy restaurant. On the other hand, wine consumption with friends and family are being put in second place (Fountain &Lamb, 2011).

Although the studies were conducted only six years ago they could be already outdated, as there has been a shift in wine trends recently. A Vivino survey (2016) studied on American consumers suggests that people are drinking at home more than ever. Some beverage connoisseur calls this a rise of "home-tainment". Nearly fifty percent of Millennials respondents are said to prefer drinking wine at home rather than at restaurants, social gatherings or bars (Vivino, 2016).

In regards to all the merits the above studies exhibit all of them mostly measure the correlation between consumption occasion and variables such as: culture, price, product features like: grape variety, or style. That is why this study aims to measure the relation between occasion and country of origin of wine. It will test if Millennials— consumer group, often overlooked in consumption occasion studies, take into account country of origin, as a quality determinant, when evaluating products for different occasions. If yes, in which of the considered occasions, country of origin exhibit the higher level of influence.

CHAPTER 3 METHODOLOGY

3.1 Research Framework

product knowledge H1

H2

consumer ethnocentrism reliance on country of origin, H3 price, packaging, wine sweetness

and wine color consumption occasion

H4 socio-demographics

Figure 1 Research model

3.1.1 Research hypotheses

The literature review suggests that the level to which consumers rely on country of origin cue when assessing quality of a product depends on their prior knowledge. That is why a consumer with great knowledge about certain product category does not exhibit the same attitude towards intrinsic and extrinsic cues a product provides as a novice consumer who is unable to detect true quality of a product. When it comes to wine, an expert consumer will not depend strongly on country of origin, which serves as an extrinsic cue. He/she is able to assess the alcoholic beverage thoroughly based on all the information signals of a wine such as grape variety, color or alcohol content. On the other hand, a consumer with little

knowledge about wine will be unable to know the true meaning behind these cues and thus will strongly rely on country of origin. For him/her country of origin is a substitute measure of excellence when other information signals cannot be found or/and used. That is why I propose a hypothesis:

H1: Level of Millennials’ product knowledge influences reliance on country of origin cue when evaluating quality of wine in such a way that those who have greater product knowledge will have less reliance on country of origin.

From the literature review there could be derived that consumer ethnocentrism is influential when it comes to the usage of country of origin cue. Consumers who exhibit a low degree of consumer ethnocentrism evaluate foreign products solely based on their merits and do not pay attention to the origin of a product. Consumers who have high level of consumer ethnocentrism, however, are believed to view foreign country image from a negative perspective and accordingly see non-domestic goods as poor in quality. Thus, ethnocentric consumers tend to give greater importance to the country of origin cue and depend on it strongly. Therefore, hypothesis number two is proposed:

H2: Level of consumer ethnocentrism influences Millennials’ reliance on country of origin cue when evaluating quality of wine in such a way that those who exhibit higher level of ethnocentrism will have greater reliance on country of origin.

As literature review suggests a consumption occasion has a significant influence on consumer behavior and product evaluation. Depending on the occasion, consumers rely on different information cues when making a purchase of a certain product. For example, when wine is bought for home consumption consumers choose cheaper options and do not highly determine their decision on information signals other than price. However, when people assess wine to buy for a gift they consider packaging and brand to be most crucial and also are willing to spend more money on the product. As under these circumstances consumers select wine more carefully and look for more details I hypothesize that:

H3: Consumption occasion influences the level of Millennials reliance on country of origin cue when evaluating wine in such a way that those who buy wine for a gift will have greater reliance on country of origin.

In previous studies scholars have found that socio-demographics influence consumer behavior for wine. Lockshin and Corsi (2012) proved that age, gender, as well as education level influence consumers’ wine preferences. When it comes to age, the studies stress that the younger people are the more likely to purchase foreign products. This may suggest that country of origin is a more important purchase criteria to older consumers. Also, according to Helvet and Mullet (2009), younger consumers, because of their limited means, depend on price more when making a purchase. The results of studies done on the influence of socio-demographics revealed as well that for example female, compared to men, prefer sweeter wines with fruity taste (Lockshin & Corsi, 2012). Women also depend on wine color in greater extent as they are more willing to purchase white wine instead of red, however men’s willingness to drink either white wine or red wine is about the same (Bruwer & Johnson, 2011). Further, researchers Barber (2009) and Atkin (2007) in their articles stressed that female consumers look for more information about wine during a purchase, that is why country of origin, and packaging might be a more crucial purchase criteria to them. As for level of education and income, it was found that the higher level of education and income, the less likely consumers are to rely on country of origin (Wall & Heslop, 1986). Following the finding I state that:

H4: Socio- demographics influence Taiwanese Millennial consumers’ reliance on wine purchase criteria, such as country of origin, price, packaging, wine sweetness and wine color.

3.2 Research method

This research employs quantitative research method. After Huysamen (1997) there could be stated that quantitative research method usually “discerns a cycle of successive phases of hypothesis formulation, data collection, analysis and interpretation” (Huysamen, 1997, p.1). It attempts to gather facts, conduct forecasts and examine validity of the hypothesis (Nykiel, 2007).

3.3 Data collection and sampling technique

An online survey was conducted to collect data. The questionnaire was designed and placed on surveymonkey.com. Purposive and snowball sampling techniques were used to reach the qualified respondents. The targeted respondents for this study are Taiwanese wine consumers with various wine knowledge and consumption frequency, born between 1980 and 1995 (McCrindle, 2010). The purposive sampling was employed by posting the link to the survey to Facebook groups which target specific audience, particularly people interested in

wine, such as 微醺 K 姐的 K-Wine, 台南大酒黨~*, and 美酒, 美食, 揪幸福¹. The groups were chosen, as they presented the biggest number of followers among Taiwanese wine Facebook groups. Furthermore, store owners of wine stores, namely: Hungary Wine Taiwan, and Major Cellar were asked to publish the link to the questionnaire on their Facebook pages.

Those stores have also provided incentives in a form of coupons which were offered to the participants of the study to increase response rates. The reason why Hungry Wine Taiwan and Major Cellar stores forwarded the questionnaire is because they were the only wine stores, out of ten, who agreed to help with the research. The snowball sampling was conducted by posting the link to the questionnaire on my personal Facebook account. The link was then shared by twenty users. The reason why Facebook was chosen to send out the survey was that it is the most popular social network in Taiwan, with 77 % penetration rate in 2017 (Newman et al., 2017). Also, young users who are the aim of this research are the most common age demographics to use Facebook worldwide (Greenwood, Perrin, & Duggan, 2016). According to Facebook Audience Insights Tool (2016), 88% of 18–29 year olds and 84% of 30–49 year olds used Facebook. The resulting sample size of this study was 210.

3.4 Questionnaire design

To design the questionnaire, items from Shimp and Sharma (1987), Williamson, Lockshin, Francis, Loose (2016), Lindquist (2001), and Ghalandari and Norouzi (2012) were used and further modified to fit this research. The questionnaire consists of five parts, and each question item is measured using five response level Likert scale where the response 1 means strongly disagree or not at all important, and 5 represents strongly agree or very important. Part A asks about demographics; Part B measures subjective and objective knowledge about wine; Part C measures consumer ethnocentrism by employing adjusted to the research CETSCALE where shortened 10-item CETSCALE (Q:1-10) (Lindquist, 2001) is used. The scale has been often used alongside with country of origin variable and has been found more adequate for smaller samples (around 500 respondents) (Jimenez-Guererro, 2014);

Part D tests the importance of purchase criteria when assessing wine, namely price, country of origin, packaging, wine sweetness, and wine color; lastly Part E of the questionnaire examines purchase habits in correlation to consumption occasion. In order to reach the study’s

1 微醺 K 姐的 K-Wine—Tipsy K- sister, K-Wine (eng.) 台南大酒黨~*— Tainan Big Party (eng.)

respondents, the questionnaire has been written in two language versions: in English, and later it has been translated by Taiwanese International Communication students into Chinese. This way I strive to eliminate the possible language barrier and to collect greater number of responses. In addition, to make sure only Taiwanese members of Generation Y completed the survey, preliminary questions asking the year of the birth and nationality were included at the beginning of the questionnaire.

3.5 Measurement of research variables and reliability test

In the literature review I have presented the potential moderating variables of reliance on country of origin. Based on the review independent and dependent variables of the research were formulated. This study adopts four independent variables, namely product knowledge, consumer ethnocentrism, consumption occasion, and socio-demographics.

Reliance on country of origin serves as a dependent variables, together with reliance on price, packaging, wine sweetness, and wine color. In order to analyze the collected data from distributed survey statistical software SPSS was used.

First, factor analysis and reliability test of product knowledge and consumer ethnocentrism were conducted. The resulting factor loadings and Cronbach’s α are presented in Table 1. The factor loading values for three- item subjective and objective knowledge scales were all found to belong to one dimension, and have high validity with scores between 0.7< factor loading>1. When it comes to ten-item consumer ethnocentrism scale, the factor analysis had resulted in items being grouped into two dimensions with three items in the latter one. However, the factor loading of those three item did not have high validity 0.7< factor loading>1. Thus, I did not look into the dimensions of these variables and instead I have used one factor solution to force ethnocentrism to load on one dimension.

The Cronbach’s α for shortened ethnocentrism seven- item scale, and three- item subjective and objective knowledge scales are acceptable to good reliability where all the Alpha values are between 0.5< α >0.8.

Table 1

Measurement of research variables

Measurement items Factor Loading Cronbach's α

Subjective knowledge

feeling knowledgeable about wine .909

having strong interest in wine .872 .880

others ask your advice on wine .934

Objective knowledge

white wine goes well with fish .725

rose wine is made of red grapes .702 .737

chardonnay is a grape variety .734 Ethnocentrism

Taiwanese should always buy Taiwanese products .792 only unavailable in Taiwan products should be imported .776 purchasing foreign made products is un-Taiwanese .826

a real Taiwanese should always buy Taiwanese products .780 .831 It is not right to purchase foreign products .751

we should purchase product manufactured in Taiwan .835 there should be little purchasing from other countries .820

Finally, as the measurement items resulted in exhibiting appropriate reliability, composite measures were constructed by computing the indicators of each research variable, namely subjective knowledge, objective knowledge, and consumer ethnocentrism. Those will be further used in multiple hierarchical regression analysis, testing hypothesis 2 and 3.

CHAPTER 4

DATA ANALYSIS AND RESULTS

A total of 210 responses have been gathered in the period of 10 days in April 2018.

Completed surveys which did not meet the sample requirements were deleted. Also, to ensure the survey was not answered multiple times by the same person, the respondents were asked to leave their e-mail address at the end of the survey. In a situation where the e-mail address has been repeated the survey responses were deleted. After conducting data cleaning, five of the respondents did not meet the research criteria, with the research subjects being foreigners or not belonging to the Generation Y. Thus, those responses were removed, leaving 205 to analyze. Among the respondents 51.7 % were male and 48.3% were female.

The age of the respondents was also spread evenly, with 55.6% being of age between 20-30, and 44.4% ages 31-40. All income, education and consumption frequency groups were represented in the sample. The precise number of the respondents' socio-demographics can be found in Table 2, 3, and 4.

College/ University or higher 197 96.1 100.0

Total 205 100.0

Table 3

Respondent’s monthly income in New Taiwanese Dollar

Frequency Percent data, descriptive analysis was employed. Furthermore, in order to compare the importance of country of origin in relation to other criteria upon wine purchasing, one-way repeated measure ANOVA analysis has been used. The reliance on country of origin was tested in Part D of the questionnaire. Moreover, to test the research H1, H2, and H4, hierarchical multiple regression analysis was run measuring the significance of independent variables, such as socio demographics, product knowledge, consumer ethnocentrism, in connection to reliance on country of origin while purchasing wine. To test H3, SPSS paired-samples t-test was used, comparing reliance on wine quality signals during two different consumption occasions.

4.1. Reliance on country of origin Table 5

One-way repeated measures ANOVA analysis of purchase criteria (I) factor1 (J) factor1 Mean Difference (I-J)

Sig. ^b

3) *. The mean difference is significant at the .05 level.

5) b. Adjustment for multiple comparisons: Bonferroni.

A one- way repeated measures ANOVA explain that wine purchase criteria were not rated equally F(1, 204) = 7594.62, p < 0.001). First, post hoc tests using the Bonferroni correction, presented in Table 5, revealed that the mean of reliance on wine sweetness is statistically different from the means of reliance on price (p<.05), country of origin (p<.05), packaging (p<0.001), and wine color (p<.01). As wine sweetness exhibits the highest mean of reliance (M=3.95), it is found to be the most important wine purchase criteria among Taiwanese Millennials.

Secondly, according to post hoc tests using the Bonferroni correction, mean of reliance on wine color is significantly different from the reliance on wine sweetness (p<.01), and packaging (p<.001). Further, the mean reliance on price is also significantly different from the reliance on wine sweetness (p<.05), and packaging (p<.001). Lastly, the mean reliance on country of origin is as well significantly different from the reliance on wine sweetness (p<.05) and packaging (p<.001). Thus, the results show that wine color (M=3.70), price (M=3.68), and country of origin (M=3.64) together are found to be of secondary importance to Taiwanese Millennials.

Finally, the mean of reliance on packaging is significantly different from the reliance on price (p<.001), country of origin (p<.001), wine sweetness (p<.001), and color (p<.001).

Taking into account that this criteria exhibits the lowest mean (M=2.95), it can be stressed that wine packaging is the least important wine purchase criteria.

4.2. Analysis of research hypothesis 1 and 2

In the described below multiple hierarchical regression analysis hypothesis 1 and hypothesis 2 are measured to test whether reliance on country of origin when purchasing wine depends on product knowledge and also on consumer ethnocentrism (Table 6).

First, product knowledge was measured by testing respondents’ subjective and objective knowledge where both were analyzed using three-item index, with all items measured on a five level Likert scale. A bivariate Pearson Correlation was employed to test whether subjective and objective knowledge should be summated or treated as separate variables in further multiple hierarchical regression analysis. The results indicated that there is a strong correlation between the two variables (r=.66, N=205; p<.001), and thus I constructed composite measures by summating subjective knowledge and objective knowledge into one composite measure—product knowledge. Furthermore, consumer ethnocentrism was tested using ten-item CETSCALE, where all the items were analyzed on a five- level Likert scale.

Table 6

Hierarchical multiple regression analysis with country of origin as a dependent variable Model influence reliance on country of origin is consumer's knowledge about wine (β = -.18 , p <.05).

The provided outcome indicate that the less knowledge an individual has about wine the more

he or she will rely on country of origin. Respectively the more knowledgeable a person is the less likely he or she will be to rely on country of origin. Hypothesis 1 is supported.

Meanwhile, hypothesis 2 suggesting that those who exhibit higher level of ethnocentrism will have greater reliance on country of origin is not supported (β = .04 , p

>.05).

4.3. Reliance on price, packaging, wine sweetness, and color, and test of hypothesis 4 In addition to testing the research hypothesis 1 and 2 hierarchical multiple regression analysis was also used to examine the relative importance of socio- demographics, product knowledge and consumer ethnocentrism in explaining the reliance on other purchase criteria, namely price, packaging, wine sweetness and wine color.

Table 7

Hierarchical multiple regression analysis with price as a dependent variable

Model

Table 7 (continued) 3

R=.406

Adjusted R²=.165 F (6, 198) = 6.190 p=.021

Age -.074 -.849 .397

Gender -.054 -.811 .418

Income -.097 -1.022 .308

Consumption frequency

-.159 -1.792 .075

Knowledge -.179 -2.228 .027

Ethnocentrism .010 .149 .882

Notes:

1) Dependent Variable: price

According to the results presented in Table 7, the only significant factor that may influence reliance on price is consumer's knowledge about wine (β = -.18 , p >.05). The more knowledgeable consumers are the less likely are they to rely on price.