Constructing stable preferences effects of dimensions of experience on movie preference in vietnamese facebook community

1.3 Research Question How can dimensions of experience recognizing, tasting, choice, effort, and consuming affect preference for movie in Vietnamese Facebook community?. In this conditio

VIETNAM- NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

CONSTRUCTING STABLE PREFERENCES: EFFECTS OF DIMENSIONS OF EXPERIENCE

ON MOVIE PREFERENCE

IN VIETNAMESE FACEBOOK COMMUNITY

A thesis submitted in partial fulfilment of the requirements for the degree of

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

Contents

List of Tables ii

List of Figures ii

Abstract iii

" CHAPTER 1 INTRODUCTION 1

1.1 Problem Statement 1

1.2 Research Objectives 2

1.3 Research Question 3

1.4 Scope ofthe Study 3

1.5 Significance of the Study 3

1.6 Structure of thesis 3

1 7 Chapter summary 3

CHAPTER 2 LITERATURE REVIEW 5

2.1 Previous studies 5

2.2 Analytical framework 7

2.2.1 Preference 9

2.2.2 Experience 1 0 2.2.3 Condition ofExperience 11

2.3 Chapter summary 12

CHAPTER 3 RESEARCH METHODOLOGY AND DATA 14

3 1 Empirical Models 14

3.2 Constructing variables 14

3.3 Database 21

3.3.1 Subjects of preference and experience 21

3.3.2 Objects of preference and experience 23

3.4 Chapter summary 24

CHAPTER 4 RESULTS AND DISCUSSION 25

, 4.1 Data Description 25

4.2 Econometric results 26

4.2.1 Unrestricted models 26

4.2.2 Restricted models 29

4.3 Result explanation 32

4.4 Study limitations 34

4 5 Chapter summary 3 5 CHAPTER 5 CONCLUSIONS AND IMPLICATIONS 36

Reference 40

Appendix 42

t

List of Tables

Table 1: Descriptive statistic of dependent variables and five main explanatory

variables 25

Table 2: Correlation between 2 dependent variables and 5 main explanatory variables ··· 26

Table 3: Results from four unrestricted models 27

Table 4: Results from four restricted models 29

Table 5: Objects of preference and experience 45

Table 6: Effects of dimensions of experience on Subjective Preference 49

Table 7: Effect of Recognizing Experience on Subjective Preference 50

Table 8: Effects of dimensions of experience on Objective Preference 51

Table 9: Effects of Tasting and Choice Experiences on Objective Preference 52

Table 10: Effects of dimensions of experience and individual movies on Subjective Preference 54

Table 11: Effects of Recognizing Experience and significant movie on Subjective Preference 56

Table 12: Effects of dimensions of experience and individual movies on Objective Preference 57

Table 13: Effects of Tasting and Choice Experiences and significant movies on Objective Preference 59

List of Figures Figure 1 - Effects of existing preference and experience on constructing preference 7

Figure 2 - Analytical framework 8

- Figure 3 -Asking flow of questionnaires .44

Figure 4 - Share of voice for selected Vietnamese movies from Oct 2011 to Mar 2012 • 60

Figure 5- Vietnam's movies- Sentiment scoring 61

Figure 6- Weekly buzz of Thien M~nh Anh Hung 62

Figure 7-Analysis of Thien M~nh Anh Hung 62

l'

•

Abstract

In market economy, neoclassical economists tell a story of demand side that every consumer compare utility of each unit of every commodity, to decide to consume the unit that provide higher utility In that story, utility is understood as obvious usefulness that any unit of commodity could provide, and preference is just the result

of "utility maximization" process

Nevertheless, real world tells another story A unit of any commodity may be useful, useless, or even harmful for a consumer, depending on specific condition at the time

of consuming the unit Preference may not only be the result of "utility maximization" but also "harm minimization" process Furthermore, complicated adaptation process could also significantly affects preference of consumer

This study investigates the relationship between preference and conditions of experience of Vietnamese movie theater audiences of Facebook community over popular fourteen movies in the time from 3rd to 29th of December 2011 In this study, data was collected by questionnaires over 67 identified Facebook users, and there were 191 raw observations ( 162 pure observations after validating) The results from logistic regression models indicate that subjective preference associates with , recognizing experience (information condition), and objective preference associates

with tasting experience (trial condition) and choice experience (easy-choice condition) The results also suggest that objective preference is more stable than subjective preference This study also have evidence of higher level of preference for some movies excluding effects of experience, imply existing preference or unconsidered dimensions of experience that future works should notice Therefore, this study is consistent with previous studies on "Constructing Stable Preference", and recommends further studies to develop demand-side theories Movie theaters in Viet Nam could also apply this study to affect their customer demand by offering many forms of movie experience

Preference is constructed by the decision maker within the decision task and its context This is an emerging generalization in the behavioral decision theory literature (Payne, Bettman, & Johnson, 1993) The constructive preference perspective argues that people often construct their preferences in a given situation based on information available at the time of preference elicitation (Hoeftler & Ariely, 1999) However, this perspective does not mean that previous experiences do not affect latter decisions of consumers, but consumers decide their choice based on their relevant experiences, including previous relevant decisions

By introducing the concept of "Constructing Stable Preferences", Hoeftler & Ariely (1999) examine some common dimensions of consumer experiences (Effort, Choice and Experience), and effects of these dimensions on formation

of preference Movie is a disposable commodity that defines a domain appropriate for studying the impact of experiences like these on consumer preferences Traditionally, Cinema is the seventh art, which supply an entertainment form for modern life and a center to exchange ideas Different from Architecture, Sculpture, Painting, and even Music, Poetry, Dance, works

of Cinema are more convenient to reach public, because they are of high level

of spreading, they have economies of scale, they maximize viewers' feeling, and they express life questions as whole systems As an art, each work of cinema define a separate concept, which create different levels of audience's preferences, and concepts of different works of cinema also create different levels of preference on the same audience As a center of exchanging ideas, cinema is the place where people go to acquire knowledge, emotion and inspiration from the community as well as from over the world Therefore, in cinematography domain, we have many units (movies) to study relationship between preferences for them and experiences about them

In Vietnam recent years, unlike in the past, youths got used to watch movies in cinemas, instead of at their homes Therefore, we would choose cinematography domain to investigate relationship between Vietnamese youths' preference for movies and their experience about these movies Vietnamese youths also tend to exchange their perspective, using social networks like Facebook By focusing on young respondents that is target customers of cinemas, in this study, Facebook community was selected to examine as its impact and popularity

1.2 Research Objectives

With objects of movie, and subjects of Vietnamese Facebooker, this study aim to:

• Find evidence of relationship between experience and preference

• Recognize different types of experience in association with different types

of preference

• Find evidence of preference stabilization process

1.3 Research Question

How can dimensions of experience (recognizing, tasting, choice, effort, and consuming) affect preference for movie in Vietnamese Facebook community?

1.4 Scope of the Study

This study aim to clarify relationship between preference and experience in Vietnamese Facebook community, so we ask our friends from our Facebook account friend list (http://www.facebook.com/maiguanghuy?sk=friends) to be respondents of this study

Fourteen movies that are examined in this study were recently shown movies, now showing movies, and coming soon movies in the time from 3rd to 29th of December 2011 (table 4, appendix 3)

1.5 Significance of the Study

This study supports the concept of "Constructing Stable Preferences", and further clarifies the demand side of market economy

1.6 Structure of thesis

There are five chapters in this study The first chapter introduces the problem

of demand side in market economy, and the study itself The second chapter clarifies the concepts of preference and experience, as well as orientates the way this study approach the problem on the base of previous studies The third chapter explains the models being used, and the way these models have been building up The fourth chapter discusses the results of running regression The last chapter concludes the significance of this study

1 7 Chapter summary

This chapter presents evolution of economics theories about preference from existing preference to constructive preference, and the concept of

"Constructing Stable Preferences" in cinematography domain, especially in context of Vietnamese Facebook community In this chapter, research objective is stated as to demonstrate the relationship between preference and experience, to support the concept of "Constructing Stable Preferences", and research question is how five dimensions of experience affect preference for movie The scope of this study is Vietnamese Facebook community in cinematography domain

so that this study could base on

Steve Hoeffler & Dan Ariely (1999) concluded that the condition of experience is related to the type of preference constructed by consumers and the consistency they follow preference They also concluded that the hard and easy conditions disassociate objective and subjective preference Easy conditions of choosing facilitate consumers gain subjective preference, while difficult conditions facilitate objective preference About the consistency, they concluded that, when consumers make a single choice or repeated choices, consistency on the different attributes increased

Jonathan E Alevy, Craig E Landry & John A List (2006) viewed the stability

of preferences as a key argument to understand fundamental concepts of microeconomic theory, which are often used in public policy analysis They investigated further on anchoring a presumed modalities for preferences to

be constructed - in a field environment, and found mixed evidence of the importance of anchoring (with treatment effects), suggesting that market participants are not affected by irrelevant cues However, the behavior of new market participants is still such an interesting topic to study, and the diversity

of anchors in the market environment implies further research in this area to discover basic insights of consumer preferences' nature

Combine results of the two studies, we can understand that inner preference of consumer is constructed by experience in outer condition Later experience interacts with previous constructed preference to form new level of preference,

in a process of evolution In this process, existing preference prevent irrelevant experience to affect constructing preference

Tzahi Neuman, Einat Neuman, Shoshana Neuman (2010) used DCEs (Discrete Choice Experiments) to estimate preference structures for attributes

of maternity-ward services among three groups of women: women expecting their first child that can be considered as group "no experience"; women who had experienced their first delivery that can be considered as group "one experience episode"; and women who had more than one delivery that can be considered as group "more than one experience event" Their purpose is to detect the effect of experience on preference patterns Their main conclusions are experience appears to changes the preference pattern; the amount (intensity) of experience seems to be irrelevant, do not have an additional accumulated effect; socio-economic background variables seem to matter, that less-educated women with a lower household income seem to be more affected by experience

This result implies that intensity of experience may disassociate from preference scale if preference anchoring is already diversified clear enough or there is no mechanism for intensity of experience to diversify preference anchoring Nevertheless, Steve Hoeffler & Dan Ariely ( 1999) imply that intensity of experience may increase preference scale if it can diversify preference anchoring

Leonard Lee, Shane Frederick, and Dan Ariely (2006) focused on two shaping

of preferences: knowledge (top down) and experience (bottom up) From three experiments, results show that the presence of a conceptually offensive ingredient in beer when received before consumption influences preferences more than when received after consumption The regular beer with a few drops of balsamic vinegar (MIT brew) was liked much less when disclosure preceded sampling than succeeded Disclosure of the secret ingredient after

tasting actually did not significantly reduce preferences for MIT brew (no significant differences with the blind condition) These results pointed out that real-time experience is affected by not just people's post hoc characterization

of the experience, but preceded knowledge also

From this result, we can see that between existing preference and relevant experience, there are two ways of interaction: "experience Tasting then experience Recognizing" (that may relate to objective preference) and

"experience Recognizing then experience Tasting" (that may relate to subjective preference) By the first way (bottom up), consumer induces preference for the new concept (MIT brew) first, and then compares with his (or her) existing preferences of the two ingredients (regular beer and balsamic vinegar), so effect of existing preferences is not significant By the second way (top down), consumer recognizes his (or her) existing preferences of the two ingredients first, then combines these preferences into a unique preference, and then compares with Tasting experience, so effect of Tasting experience is not significant

2.2 Analytical framework

Existing preferences Previous experiences in

in relevant domains cinematography domain

~

-cinematography domain Existing preference in cinematography domain

~

Movie choices Constructing preference in cinematography domain Figure 1 - Effects of existing preference and experience on constructing preference

-In this study, because of cinematography features, we consider existing preference of audiences as perception of movie categories, which each

•

audience constructs every time they experience anything about any movie A person that has never been to a movie theater cannot define any distinctive category of this domain After watching first movie in a cinema, that person will get his (or her) very first own perception about cinematography domain,

or the unique movie category at that time When this person watches another movie in cinema, he (or she) get another perception of the domain, and immediately relate to the first perception, to compare and determine which his (or her) favorite movie is By this process, under effects of his (or her) existing preference outside the domain, he (or she) defines common categories as well

as distinct categories of the two perceptions (or in other words, to construct preference structure of cinematography domain) For the next movie being watched in cinema, he (or she) begins to compare the third perception with previous defined categories, to increase or decrease preference for these categories, or to define a new category if there is not any relevant one

Experiences on each movie I Cinematography existing preferences

Recognizing

Affect?

Tasting Choice

Preferences for each movie

Consuming

Figure 2 - Analytical framework Furthermore, watching movie in cinema is just a form of expenencmg cinematography, which is full experience One can also have partial experience like getting (hearing and reading) review, learning introduction, receiving advertisement, watching trailer, seeing mock-up, of new movies

By this way, as well as by process of choosing a movie (even new or old) to watch, audiences refer the movie to previous defined categories, to construct

•

preference for it own Preference for a movie, thus, depends on existing movie preference, previous related watched movies and experience of recognizing, tasting, choosing that movie

non-2.2.1 Preference

In economics, the concept of preferences are not familiar as the concept of utility, so we consider its definitions by dictionaries

http://oxforddictionaries.com/) defined preference as "a greater liking for one alternative over another or others" Meanwhile, Cambridge

http://dictionary.cambridge.org/) describe preference as "when you like something or someone more than another person or thing" In behavioral economics, Steve Hoejjler & Dan Ariely (1999) stated that

"Economic theory is built on the assumption that consumers' choices indicate their underlying needs and wants, because it is these underlying preferences which are tapped when consumers make choices" In this study, preference is the inner factor of a consumer, together with scarcity as outer factor, that determine his (or her) demand for a commodity (quantity consumed at each level of price)

By measurement, Steve Hoeffler & Dan Ariely ( 1999) defined two types of preference: subjective preference as "self reports of attribute importance and strength of preference", and objective preference as

"revealed preferences" In this study, Subjective Preference is measured by one's intention to choose commodity, or his wants to consume (contingent valuation by proxy variable of stated preference about the movie) Objective Preference is measured by one's decision

•

to choose commodity, or his needs to consume (dichotomous variable

of action to go see the movie in theaters or not)

2.2.2 Experience

Similarly, Oxford Dictionaries defined concept experience as

"practical contact with and observation of facts or events" or "an event or occurrence which leaves an impression on someone"

Cambridge Dictionaries Online defined it as "(the process of getting)

knowledge or skill from doing, seeing or feeling things" or

"something that happens to you that affects how you feel"

We develop this study based on Steve Hoeftler & Dan Ariely (1999),

so we consider their common dimensions of consumer experiences (Effort, Choice and Experience) However, in such a field environment like movies, concept of "dimension Experience of consumer experiences" is not clear enough In this study, we will split this general dimension into three more specific dimensions (Recognizing, Tasting and Consuming), means that we will consider consumer experiences in five dimensions as following

Choice is experience of choosing one commodity among other

commodities of the same kind In Montgomery (1983) and Beach (1993) (as cited in Steve Hoeftler & Dan Ariely, 1999), it is a process

by which preferences are consolidated in order to arrive at a resolution for a choice task

Effort is experience of investing mental energy in making up one's

minds, according to Alba & Hutchinson (1987) (as cited in Steve Hoeftler & Dan Ariely, 1999)

•

Recognizing is experience of collecting knowledge about the

commodity This is one of the two shaping of preferences in Leonard Lee, Shane Frederick, and Dan Ariely (2006) (knowledge: top down interaction)

Consuming is experiencing the outcome of the choice one makes

(feedback) This is the other one of the two shaping of preferences in Leonard Lee, Shane Frederick, and Dan Ariely (2006) (experience: bottom up interaction)

Tasting is experience of using the commodity in a short period of

time, or trying some of commodity's features It is just the incomplete experience of Consuming

2.2.3 Condition of Experience

From the definitions of "Experience", we can measure above dimensions of Experience by define the following conditions of consumer

Information (proxy of Recognizing): consumer is informed about the

movie (by newspaper, brochure, someone ) or not In this condition consumer (or audience) build up concept of the movie, associate to existing movie preference and non-movie preference, so we can predict that the more consumer experience in this condition, the more

he (or she) prefer the movie (because of his or her expectation about the movie)

Trial (proxy of Tasting): consumer has seen any movie trailer or not,

and consumer has seen any movie mock-up at a cinema or not Experiencing more in this condition, consumer (or audience) defines genres (movie categories) of the movie more clearly, consolidating his

(or her) expectation about the movie, so he (or she) may prefer the movie more

Easy-Choice (proxy of Choice): how many movies consumer

considers at the same time In this condition, consumer (or audience) compare existing preference for the movie and for the others, to determine whether it is deserved to pay for the movie, so we can predict that the more Choice experience (the more movies to compare), the more preference he (or she) construct

Hard-Choice (proxy of Effort): consumer chooses to preview the

movie at home or not, and consumer chooses to watch the movie again in cinemas or not In this condition, consumer (or audience) has previously experienced the storyline or the visual-sound effects of the movie, and decide to enjoy again, so Effort experience may have positive significant relationship with subjective (or objective) preference

Repeated-Choice (proxy of Consuming): consumer chooses to watch

the movie of the same series (or the same other movie category like director, actor, actress ) with other previous seen movies or not In this condition, objective preference for other previous seen movies of the same categories construct subjective (or objective) preference for the considering movie, so Consuming experience may have positive significant relationship with subjective (or objective) preference

2.3 Chapter summary

This chapter presents four researches on preference that this study based on, and the analytical framework was formed from them: two types of preference (subjective preference and objective preference) and five dimensions of experience (recognizing experience or information condition of experience

•

tasting experience or trial condition of experience, choice experience or choice condition of experience, effort experience or hard-choice condition of experience, and consuming experience or repeated-choice condition of experience)

According to the "analytical framework" above, we expect that B2, B3, B4, Bs,

B6, Yz, y3, y4, Ys and Y6 are all positive coefficients, express direct relationships between Subjective Preference (or Objective Preference) and Information (or Trial, Easy-Choice, Hard-Choice, Repeated-Choice) condition of experience Meanwhile, we cannot predict sign of coefficients for control variables, because each different movie may have a higher or lower level of preference than the average level

3.2 Constructing variables

It is obviously that consumer's preferences of watching a movie are different from before to after watching it However, movie is rather a disposable commodity, that preference after watching a movie does not generally determine demand of watching that movie in cinema, so we need not to measure consuming effect of watching a movie in cinema on its own

preference Eventually, we measure consuming effect of a movie on following movies of the same genre, and use the action of watching a movie as an evidence (condition) of objective preference (preference that has Jed consumers to decide to buy movie ticket)

The above difference of preference also suggests that there are two groups of respondents: respondents who had never watched the movie, and respondents who had watched the movie With respondents who had never watched the movie, we can identify that they had not got objective preference for the movie yet, and can measure their subjective preference for that movie by a question about their intention of choosing the movie to watch

In contrast, with respondents who had watched the movie, we can identify their objective preference for the movie (by decision of buying movie ticket), but do not measure subjective preference for the movie at the time of answering (because respondents had already got consuming experience at that time) With this group of respondents, we should measure subjective preference at the time of preference elicitation, or the time of buying movie ticket And at the time of buying movie ticket, if we ask respondents "do you intend to watch this movie at a cinema?", the answer must be "yes'' That is the reason why we can assume that if a respondent decide to watch a movie (elicit objective preference), he (or she) also intend to watch it (elicit subjective preference)

Therefore, in this study, we measure preference variables and experience variables at two different point of time: buying ticket (in case respondent had watched the movie in cinema) and answering questions (in case respondent had never watched the movie in cinema) However, to keep respondents clear about meaning of questions, we must limit number of questions in the questionnaire As a result, in this study, the three variables I, T and H were measured by the same questions (from Q03 to Q09) for both groups of

respondent (because audiences usually take information get trial and make hard choice before deciding to buy movie ticket)

About "the diffusion process and the role of promotion", Jonathan Derek Silver (2007) stated that customer's movie preference depends on "the film's marketing communications and also word-of-mouth recommendation from trusted sources" In that thesis, he cited Belch and Belch (200 1) that popular communication tools include mass media advertising, sales, sales promotion public relations, direct marketing and Internet marketing Among those tools, this study does not consider sales and sales promotion, because they make customer's experience after buying ticket

Therefore, Information condition of experience can be constructed by the formula "I= I1(Q03) + I2(Q04) + I3(Q05)", in which Q03 (question number 3)

is "Have you heard about the movie from someone?" (word-of-mouth recommendation), I1 get value of "1" (or "0") when the answer of Q03 is

"Yes" (or "No") Q04 (question number 4) is "Have you read any article about this movie?" (public relations & Internet marketing), I2 get value of "1" (or

"0") when the answer of Q04 is "Yes" (or "No") Q05 (question number 5) is

"Have you seen any leaflet, brochure, poster about this movie?" (mass media advertising & direct marketing), I3 get value of" 1" (or "0") when the answer

of Q05 is "Yes" (or "No")

As cited in Jonathan Derek Silver (2007), Harvard Marketing Professor Howard Lewis observed in 1933: "Trailers constitute a direct consumer appeal

to a class of people easily influenced by such appeals, those already disposed

to attending motion picture theaters The exhibition of trailers can be timed so

as to yield the best possible results, since generally they are only used for pictures exhibited at a particular theater on specific dates Patrons cannot avoid submitting to the appeal, whereas printed advertising may not be read" Hence trailer is the most suitable tool to lead potential consumer taste the film

Besides, mock-ups, as lobby displays or foyer displays at theater, also make consumer visualize and feel the film Therefore, trailer and mock-up can be considered as the two channels of consumer tasting

ofthis movie?", T1 get value of"1" (or "0") when the answer ofQ06 is "Yes" (or "No") Q07 (question number 7) is "Have you seen any mock-up of this movie?", T2 get value of "1" (or "0") when the answer of Q07 is "Yes" (or

"No")

Movie is a disposable commodity, that consumer do not want to watch it again That is the result why if consumer decide to preview the film at home before to go see it in theater, or choose to watch it in theater more than one time, he (or she) have to invest metal energy, or get Effort experience

you watched this movie at home?'', H1 get value of "1" (or "0") when the answer of Q08 is "Yes" (or "No") Q09 (question number 9) is "How many times did you watch this movie in cinemas?", H2 get value of"O" (or "1 ", "2") when the answer of Q08 is "0 (not any)" (or "1 (once)", "> 1 (more than once)") Q02 (question number 2, exclude effect of Consuming experience of the movie on its own preference from effect of Effort experience) is ''Have you ever watched this movie in a cinema?", OP get value of"1" (or "0") when the answer ofQ02 is "Yes" (or "No")

The remaining variables are measured by two groups of questions (separated

by Q02) Following is the group of questions measured at the time of buying

OP (Objective Preference) is determined by the answer of Q02 (that is the

value of "1") Because consumer needed to buy ticket (OP = 1), so consumer had wanted to buy ticket earlier, that means SP (Subjective Preference) also get value of "1 " E (Easy-Choice condition of experience) measured by Q 10 (question number 1 0), which is "How many films have you considered at the time ofbuying ticket?" (value ofQ10 is the number of considered movies) R (Repeated-Choice) measured by Q 11 (question number 11 ), which is "Have you watched this movie because of previous movies? State the previous movies and the points of relation." (value of Q11 is the number of related movies)

The group of questions measured at the time of answering questions, in case consumer had never watched the movie (Q02 = 0), is mentioned after this OP (Objective Preference) is determined by the answer ofQ02 (that is the value of

"0") SP (Subjective Preference) measured by Q12 (question number 12), which is "Do you intend to watch this movie at a cinema?", and get value of

"1" (or "0") when the answer of Ql2 is "Yes" (or "No") E (Easy-Choice condition of experience) measured by Q 13 (question number 13 ), which is

"How many films do you consider now?" (value of Q 13 is the number of considered movies) R (Repeated-Choice condition of experience) measured

by Q14 (question number 14), which is "Is there any relation between this movie and one of previous movies? State the previous movies and the relations." (value ofQ14 is the number of related movies)

Appendix 2 illustrates the flow of all asking questions in the questionnaire and their values in the three cases: respondent do not know anything about the movie, respondent know the movie but not watched yet, and respondent has watched the movie

About Objective Preference variable, because of the failure of web-based interactive portal, we cannot measure the real action of buying ticket at

cinema However, we measured variable OP by Q02 to have a proxy measurement, because:

• Right after the classified question (QO 1 ), OP question asked respondents clearly and directly Therefore, there was no orientation for respondents to answer bias

• Respondents attended this study to help or to care about Vietnamese cinematographic, and did not have any personal benefit of answering Therefore, they were not motivated to give an incorrect answer

• We used Q09 as a double check question for OP question, and dropped all observations that made conflict between the two questions (watched but had zero time of watching, or not watched but had at least one time of watching, 5 observations actually)

Like Objective Preference variable, we also have concerns about experience variables:

• Without web-based interactive portal, we cannot create true conditions to measure experiences Therefore, we chose to ask yes-no questions to assume the conditions that respondents had experienced

• Choice experience was determined by a single question about how many movies respondents had considered at the time of buying ticket (QIO, in case respondents already had Objective Preference) or at the time of answering (Q13, in case respondents did not have Objective Preference) The more considered movies, the more respondents experience Choice

• Effort experience was determined by the two questions Q08 and Q09, about hard-choice conditions where respondents chose to watch the movie again This is appropriate, because without eagerness (or expectations for

the new movie), cinema audiences usually do not want to see the movie again

• Recognizing experience was determined by questions about whether respondents received movie information from someone (Q03), some articles (Q04), some advertisings (Q05), or not We chose this type of experience, because this is important channels for a movie to be known and recognized

• Tasting experience was determined by the two questions Q06 and Q07, about feeling of the movie by any trailer or mock-up We chose this type of experience, because this is the way cinema marketers use to do sampling to potential customers

• Consuming experience was determined by a single question about previous watched movie at the time of buying ticket (Q 11, in case respondents already had Objective Preference) or at the time of answering (Q14, in case respondents did not have Objective Preference) We chose this type of experience, because in our opinion, this is the way respondents construct their favorite categories in cinematography domain

• To prevent inappropriate data of experience variables, we used question Q01 as a double check question (about getting to know the movies), and dropped all observations that made conflict between QO 1 and questions from Q03 to Q 14 (except Q 10, Q 12, Q 13 because they are not about experiences of getting to know the movies) as well as Q02 (Objective Preference is also the main experience of the movie) ( 18 observations actually)

Other concerns of the questionnaire:

• Questions from Q01 to Q09 are mandatory Questions from Q10 to Q14 are optional

• If questions Q10, Q11, Q13, and Q14 are without answers or with answer

"no", then they get value of zero

• In case SP was measured at the time of answering questions (Q02 = 0), if question Q12 was without answer, then the observation is eliminated (6 observations actually)

3.3 Database

Data was collected on individual Facebook users Database was established on 14 questionnaires over 14 movies A total of 191 raw observations were collected on 67 users Each questionnaire was designed to enable to measure seven variables (appendix 1 )

We had intended to collect data with a web-based interactive portal for audiences by a cinema However, we could not persuade cinemas

to support us to set up a system like that We had also intended to interview customers at cinemas, but no cinema allowed us to approach their customers at their place That was the reason why we chose to collect answers from Facebook-user respondents, which defined a sample ofyoung, dynamic and open community

We created questionnaires by using application "Facebook Polls'' of Code Rubik incorporation (http://coderubik.com/en/) This application lets us create polls with as many questions as we want, and allow us to choose many ways to publish it on Facebook walls (our wall, our friend's walls, friends of our friend's walls, public account wall

"Phim Vi~t Nam" and wall of page "Margroup") Each questionnaire

has its own link that everyone can access (for example, questionnaire's link of movie "Arthur Christmas" ts http://apps.facebook.com/my-surveys/arthurchristmas) To analyze data and run regression, we used software "Stata!SE 11.1 for Windows"

To collect data, personal Facebook account was used, with a friend list of about 300 persons We invited friends and asked them to invite their friends, and post our invitation on "Phim Vi~t Nam''

("Vietnamese film", a common-attention Facebook account for movement of Vietnamese cinematography) as well as "Margroup" (a Facebook page of Margroup, one of the most popular student club in the University of Economics Ho Chi Minh City) This method of data collection allows us to reach identified respondents Furthermore, Facebook is essentially a community of individuals who are not anonymous (to be accepted in other's Facebook friend list, that is a book of faces, you have to be previously acquainted by that person and you must have a profile picture of your face)

All respondents are Vietnamese, mostly residents of Ho Chi Minh City (44/67 persons, 1321191 observations) and half of them are women (31167 persons, 100/191 observations) This is typical for the scope of this study, because Ho Chi Minh City is the main market (in terms of population and income, Vietnam Pocket Reference Guide

2011 - Nielsen) for the rising cinema industry of Vietnam (leisure time of the young generation is spent mostly on connecting with friends and to enjoy entertainment, Vietnam Pocket Reference Guide

2011 -Nielsen)

•

3.3.2 Objects of preference and experience

Objects of this study are 14 movies that respondents can hear about at the interview time (from 3rd December to 29th December, 20 II), because they were issued at previous time (Hot boy n6i lo1;1n - M02, Khat vQng Thang Long- M03, The Adventures of Tintin- M04, In time - M06), now showing movies (Breaking Dawn - MOO, Puss in Boots- M05, Ghost Protocol- M12, Arthur Christmas- M13), and coming soon movies (Thien m~nh anh hung - MO I, Hmin d6i thfm xac- M07, L~ phi tinh yeu- M08, Hello co Ba- M09, Lai nguy~n

huy€t ngai- M10, Vil di~u duang cong- Mil) However, not any respondent was chosen for any specific movie, but they were completely free to choose which movie to answer about their preferences and experiences, as well as choose how many movies to answer From appendix 3, of the 14 movies, 8 were Vietnamese movies (691162 pure observations) and 6 were foreign ones (931162 pure observations) In term of types, there were 6 comedy movies, 5 action movies, 5 drama movies, 4 adventure movies and 3 animation movies

From appendix 12, the time of collecting data of this study (from 3rd December to 29th December, 2011) is the early stage of "hot movie premiere season" leading up to the Tet (Lunar New Year, or Traditional New Year) holiday in Vietnam (from Oct 2011 to Mar 2012) 23rd January 2012 is the first day of the Tet, so in several days before and after this day, audiences flocked to theaters to watch Vietnamese movies We chose "Thien m~nh anh hung'', "Lai nguy~n

huy€t ngai", "L~ phi tinh yeu", "Hello co Ba", "Vil di~u duang cong",

"Hoan d6i than xac" as objects to examine preference and experience

in this study, because they were Tet movies that attracted the most

attention (while "Co dau dt;ti chi~n" is a Tet movie of last year, "Cot m6c 23" and "T6i nay 8 gia" were screened before Tet) Similarly,

"Hot boy nbi lot;tn" was chosen because it just screened from 14th

October 20 11 and it has become a phenomenon of Vietnamese cinematography (appendix 15)

From appendix 13, "Khat vQng Thang Long" is a collaborative effort

of the state (http://khatvongthanglong.vn) and oversea party (http://theprinceandthepagodaboy.com) to create a compelling cinematic work Because the film satisfied all the content and quality standards as well as meeting the Oscar's criteria, it was selected as Vietnam's representative at the Oscar 2011's Best Foreign Language Film Award, but has been removed from the nominee list Last year, it won a silver prize for Best Film and two golden prizes for Best Director and Best Supporting Actor at Vietnam's Golden Kite Award

3.4 Chapter summary

This chapter introduce empirical models, which is built-up logistic regression models for examining relationship between Subjective Preference (or Objective Preference) and five dimensions of experience, with or without effects of individual movies "Constructing variables" of this chapter explains measurements of explanatory variables and dependent variables in this study Finally, "Database" describes the way of collecting data for this study

• CHAPTER4 RESULTS AND DISCUSSION

of variable T (0.89, close to 1) show us that on average, respondents experience one of the two types of Trial condition Mean of variable E ( 1.68) show us that on average, respondents experience more than one but less than two options of choosing movie Mean of variable H (0.05 of maximum I) show us that on average, there are only 5% of observations in which respondents experience one of the two Hard-Choice of watching movie Mean

of variable R (0.4, close to half) show us that on average, for every two people, there is one person choose to watch movie because of a previously related watched movie

If we consider volatility by the formula "Standard Deviation I (Maximum Minimum)" as in Table 1, then SP fluctuates more than OP Among explanatory variables, T fluctuates the most, and variable I is the next Less volatile than T and I, variables H, R and E are in order from more to less volatile Generally, all the variables fluctuate at high level (20% or higher)

*Correlation is significant at the 0.05 level (2-tailed)

From table 2, at significance level of 0.05, we rank correlation coefficient into three groups (from 0 to under 0.3: weak correlation; from 0.3 to 0.7: moderate correlation; over 0.7: strong correlation) Dependent variable OP correlates with all the five explanatory variables, while dependent variable SP only correlates with variable I and variable T Correlations between SP and L between OP and T, between OP and E are moderate All other correlations are weak Dependent variable OP also correlates with dependent variable SP, at moderate level also Explanatory variable I correlates with explanatory variables T (moderate) and H (weak), which may implies multicollinearity if I and T stand together in a model Explanatory variable T correlates with explanatory variables R (weak) Only correlation between OP and E is negative All other correlations are positive

4.2 Econometric results

4.2.1 Unrestricted models

To examine relationships between different dimensions of experience and subjective preference (or objective preference), we include all five experience variables in each relation models with the two types of preference In addition, we also consider private effect of each movie,

so we have four unrestricted models all together, as in table 3

In the two models without private effects of movies, Subjective Preference is associated with Recognizing Experience, and Objective Preference is associated with Tasting Experience and Choice Experience At the 99 percent confidence level, the more Recognizing Experience the more moviegoer has Subjective Preference (coefficient

= 0.608 > 0), the more Tasting Experience means the more Objective Preference (coefficient = 1.054 > 0), and the more Choice Experience leads to the less Objective Preference (coefficient= -1.04 7 < 0)

Table 3 - Results from four unrestricted models

Marginal effect Marginal effect Marginal effect Marginal effect (std err.) (std err.) (std err.) (std err.) Recognizing 0.608*** (0.187) 0.327 (0.271) 0.653*** (0.228) 0.345 (0.367) Tasting 0.248 (0.26) 1.054*** (0.337) 0.378 (0.296) 0.973** (0.39) Choice -0.121 (0.088) -1.047*** (0.266 -0.193* (0.101) -1.353*** (0.331 Effort 0.203 (0.874) 0.958 (0.904) 0.2 (0.934) 0.932 (0.98) Consuming 0.239 (0.282) 0.493 (0.34) 0.125 (0.345) 0 726* (0.43 7)

If we also examine effects of individual movies, beside effects of the five dimensions of experience, on Subjective Preference, at the 99 percent confidence level, only "Thien m~nh anh hung" (MO 1) has a significant impact There are three movies that have significant impacts at the 95 percent confidence level ("Hot boy n6i lo~m"- M02,

"Lai nguy~n huyet ngai"- MIO, "Arthur Christmas''- Ml3) and two movies that have significant impacts at the 90 percent confidence level ("Khat VQng Thang Long" - M03, "Puss in boots" - M05) These six movies (79/162 observations) are of the higher level of Subjective Preference than the rest eight movies Exclude private effects of these six movies, Subjective Preference is still associated with Recognizing Experience at the 99 percent confidence level, but also associated with Choice Experience at the 90 percent confidence level

If we also examine effects of individual movies, beside effects of the five dimensions of experience, on Objective Preference at the 99 percent confidence level, there are two movies that have significant impacts ("Hot boy n6i lo~m"- M02, "Ghost Protocol"- Ml2) There

is one movie that has a significant impact at the 95 percent confidence level ("Puss in boots" - M05) and no movie that has significant impact at the 90 percent confidence level These three movies (521162 observations) are of the higher level of Objective Preference than the rest eleven movies Beside private effects of these three movies Objective Preference is still associated with Choice Experience at the

99 percent confidence level, but associated with Tasting Experience at lower confidence level (95%), and also associated with Consuming Experience at the 90 percent confidence level

Experience effects only With effects of specific movies

Marginal effect Marginal Marginal effect Marginal effect

is significantly associated with Subjective Preference at the 99.9 percent confidence level The effect coefficient does not differ very much from the coefficient in unrestricted model (0.6986 versus 0.608), confirms the only effect of Recognizing Experience on Subjective Preference Holding all the other relevant factors constant, when respondents experience one more channel of Recognizing the

(Pi+I is the probability of Subjective Preference to be I after respondents experience one more channel of Recognizing the movie, and Pi is the probability of Subjective Preference to be I before respondents experience one more channel ofRecognizing the movie)

So we can say for a one-unit increase in channel of Recognizing the movie, we expect to see about 101.09% increase in the odds of having Subjective Preference

In another sencondary model, if we only examine the effects of Tasting Experience and Choice Experience on Objective Preference (table 4, second model), we can see that these two dimensions of experience are significantly associated with Objective Preference at the 99.9 percent confidence level The effect coefficients suggest that: when respondents experience watching trailer of the movie or seeing mock-up of the movie, the log odds of their Objective Preference for the movie increases 1.2803 approximately (or the odds increases: exp(1.2803) - 1 = 259.77%); when respondents experience easy Choice of one more movie, the log odds of their Objective Preference for the movie decreases 1.0517 approximately (or the odds decreases: 1- exp(-1.0517) = 65.07%)

'

If we drop all (both main and auxiliary) explanatory variables that cause insignificant associations with Subjective Preference, then we have the third model in table 4 At the 90 percent confidence leveL dimension of Recognizing Experience and movie of "Thien m~nh anh hung" are significantly associated with Subjective Preference The effect coefficient of Information condition suggests that when respondents experience one more channel of Recognizing the movie, the log odds of their Subjective Preference for the movie increases 0.7249 approximately (or the odds increases: exp(0.7249) - I =

106.45%) The effect coefficient of movie "Thien m~nh anh hung'' suggests that the odds of Subjective Preference for this movie is 239.36% (exp(1.2219)- 1) higher than the average odds of Subjective Preference for other movies

If we drop all (both main and auxiliary) explanatory variables that cause insignificant associations with Objective Preference, then we have the fourth model in table 4 At the 95 percent confidence level, dimensions of Tasting Experience and Choice Experience are significantly associated with Objective Preference At that level, movies of "Hot boy n6i lo~m", "Puss in boots", "Ghost Protocol" are also significantly associated with Objective Preference

The effect coefficient of Trial condition suggests that when respondents experience watching trailer of the movie or seeing mock-

up of the movie, the log odds of their Objective Preference for the movie increases 1.1824 approximately (or the odds increases: exp(1.1824)- 1 = 226.22%)

The effect coefficient of Easy-Choice condition suggests that when respondents experience easy Choice of one more movie, the log odds

of their Objective Preference for the movie decreases 1.2534 approximately (or the odds decreases: 1 - exp( -1.2534) = 71.45% )

The effect coefficient of movie "Hot boy n6i lo~m" suggests that the log odds of Objective Preference for this movie is 842.91% (exp(2.2438) - 1) higher than the average log odds of Objective Preference for the rest movies (other than the three significant movies)

The effect coefficient of movie "Puss in boots" suggests that the log odds of Objective Preference for this movie is 371.57% (exp(l.5509)

- 1) higher than the average log odds of Objective Preference for the rest movies (other than the three significant movies)

The effect coefficient of movie "Ghost Protocol" suggests that the log odds of Objective Preference for this movie is 769.89% (exp(2.1632)

- 1) higher than the average log odds of Objective Preference for the rest movies (other than the three significant movies)

of commodity that has more utility in constraint of budget It means that there

is a negative relationship between Consuming Experience on a homogeneous commodity and Preference for it

Meanwhile, "Constructing Stable Preference" suggests that Preference associate with not only Consuming Experience, but also other dimensions of

experience, and by the above results, we can see that dimensions of

experience can also increase or decrease preference, depending on specific cases Because of different economic conditions of respondents, budget constraint can also be considered as easy-choice or hard-choice condition of experience, so budget constraint may imply Choice Experience or Effort Experience that construct Preference Therefore, we can consider

"Diminishing Marginal Utility" as a special case of "Constructing Stable Preference"

Although Preference is the inner factor that determine consumers' demand, but the above logistic regression results also show that this inner factor depends on the different types of influence of outer factor Subjective Preference reflects conscious mind, or theoretical understanding about features of a commodity, constructed by experience in Information condition Meanwhile, Objective Preference reflects unconscious mind, or practical feeling about utilities of the commodity, constructed by experience in Trial condition and Easy-Choice condition

From the classification result in appendix 4, we can see that the significant relationship between Subjective Preference and Experience is determined by both Sensitivity (rate of number of observations that predict probability of Subjective Preference greater than 0.5, over number of observations that have value of 1) and Specificity (rate of number of observations that predict probability of Subjective Preference less than 0.5, over number of observations that have value of 0) In this model, Sensitivity (76.67%) is rather higher than Specificity (56.94%), so we can predict that the more audience experience in Information condition of a movie, the more he (or she) intend to watch it

From the classification result in appendix 6, we can see that the significant relationship between Objective Preference and Experience is determined by a

•

'

very high rate of Specificity (95.28%, close to 100%) This high rate offsets the rather low rate of Sensitivity (45.71 %, slightly less than 50%), so we can predict that the more audience not experience in Trial condition of a movie, the more he (or she) not decide to watch it In other words, the less he (or she) experience in Trial condition of that movie, the less he (or she) decide to watch it

Higher rate of Sensitivity in relationship between Subjective Preference and Experience, and higher rate of Specificity in relationship between Objective Preference and Experience, demonstrate that having Recognizing Experience leads to constructing Subjective Preference and not having Tasting Experience (or not having Choice Experience) leads to not constructing Objective Preference (or not destructing Objective Preference) This explain why Objective Preference do not significantly associate with Recognizing Experience, despite that Objective Preference goes along with Subjective Preference (set of observations that have Objective Preference is a subset of observations that have Subjective Preference) and Subjective Preference significantly associate with Recognizing Experience

4.4 Study limitations

• In this study, data was collected by questionnaires that might cause some incorrect answers To control the accuracy of answers, we used double check questions to detect and exclude disqualified observations

• 29/67 respondents (109/191 raw observations) that participate in this study are author's friends This may raise concerns about sample selection For your reference, their characteristics are from 17 to 36 years old, of university degree, most residing in H6 Chi Minh City, more male than female

'

• Respondents were free to select which movie-questionnaires to answer, do not depend on whether they watched that movie or not This may raise concerns about control group for each movie, but this way satisfy the rule

of randomness for all 14 movies in general

• This study do not examine the process of preference elicitation of each respondent after every experience Further studies could use audience interactive data to clarify this process, and to understand the relationship between Subjective Preference and Objective Preference

4.5 Chapter summary

After describing statistics and correlations of dependent variables and five main explanatory variables, this chapter presents the results of different logistic regression models Then it explains extensive meaning of those results, and notify limitations of this study as the problems of incorrect answers, sample selection, and control group

CHAPTERS CONCLUSIONS AND IMPLICATIONS

In general, results from this study is consistent with Steve Hoeffler & Dan Ariely ( 1999) The condition of experience is related to the type of preference constructed

by consumers and the consistency they follow preference Structure of experience disassociates objective preference and subjective preference, but the hard and easy conditions are not the points Subjective Preference is significantly associated with Recognizing Experience, which is over double the odds of Subjective Preference without effects of individual movies Meanwhile, Objective Preference is significantly associated with Tasting Experience and Choice Experience, which respectively increases 260% the odds and decreases 65% the odds of Objective Preference without effects of individual movies These demonstrate the two ways of interaction between existing preference and relevant experience, which are "top down" (Recognizing) and "bottom up" (Tasting and Choice), and confirm relations

of these two ways to Subjective Preference and Objective Preference respectively Although there is a moderate correlation between the explanatory variables (Recognizing Experience) and T (Tasting Experience), there is no multicollinearity

in this study as I and T do not stand together in any model above However, if this correlation is considered in context of the similar moderate correlation between dependent variable SP and dependent variable OP, these correlations will demonstrate the common root of the two types of preference, as well as of their structures Weak correlations between I and H, and between T and R show that Effort may be an experience dimension that affects Subjective Preference (together with Recognizing Experience), and Consuming may be an experience dimension that affects Objective Preference (together with Tasting Experience), which further studies can find out

Beside Recognizing, at confidence level of 70%, Subjective Preference is also associated with Choice Experience only (appendix 4) Meanwhile, at the same confidence level, Objective Preference is also associated with all the three other