Tải bản đầy đủ
2 Research methodology, collecting data and sample size

2 Research methodology, collecting data and sample size

Tải bản đầy đủ

be explore. In this step, 10 respondents who are the managers and owners of veterinary
medicine agencies were interviewed.
3.2.2 Measurement scale
The original measurement scale from previous studies was used in Qualitative stage to
test their validity. After listening to customers’

feedbacks, the scales for study were

modified.
3.2.2.1 Trust
Measurement of Trust was based on the Anderson and Narus (1986) scale. Trust scale
items for this study are symbolized by TRUST with Likert scale ranging from 1
(Totally disagree) to 5 (Totally agree)
Statement

Coding

The agency is concerned about company welfare

TRUST01

The agency considers how its actions will affect customers and

TRUST02

company
Your agency trusts in the service of company

TRUST03

Company is well-known for fair dealings

TRUST04

Company is very dependable, especially about important things

TRUST05

Although times change and the future may be somewhat uncertain,

TRUST06

company will support agency
Company is a faithful business partner

TRUST07

The agency is secure in dealing with new business situations because

TRUST09

company will support agency.

14
 

 

 

3.2.2.2 Commitment
Measurement of Commitment was also based on the Anderson and Narus (1986)
scales. Commitment was measured by four items.Commitment scale items for this study are
symbolized by COMMI
Statement

Coding

The agency is committed to the preservation of good working

COMMI14

relationships with company
The agency believes in company as partners.

COMMI15

The agency expect to have long term relationship with company.

COMMI17

3.2.2.3 Satisfaction
Satisfaction was measured using six items which were adapted from Wrestbrook
(1980) scales. Satisfaction scale items for this study are symbolized by SATIS, such as:
Statement

Coding

This experience of products and services of this company is exactly

SATIS19

what I need.
My choice was a wise one.

SATIS20

If I could do it over again, I would choose products and services of

SATIS21

company.
This experience of products and services from this company has

SATIS22

worked out as well as I thought it would.
I’m sure it was the right thing for me to use products and service of

SATIS23

this company.
3.2.2.4 Customer Loyalty
In the veterinary industry, customer loyalty was measured using the seven items based
on the Luarn and Lin (2013) scales. Customer Loyalty scale items for this study are
symbolized by LOYAL.

15
 

 

 

Statement

Coding

I will transact with company again for future needs.

LOYAL24

I will recommend other people to cooperate with company

LOYAL25

I will say positive things to other people about the products and

LOYAL26

services provided by company.
I will continue to cooperate with company even if the service charges

LOYAL27

are increased moderately.
I have strong preference to company

LOYAL28

I will keep cooperating with company regardless of everything being

LOYAL29

changed somewhat
I am likely to pay a little bit more for using the products and services

LOYAL30

of company

The measurement scale for this study will be based on Likert scale ranging from 1 (totally
disagree) to 5 (totally agree).
(1): Totally disagree
(2): Disagree
(3): Neutral
(4): Agree
(5): Totally Agree
3.2.3 Step 2: Pilot
Pilot before quantitative research was conducted on five Vietnamese customer to finetune
some wordings and ensure statements are easy to understand and use layman language.
After this pilot, there are no changes in the key content of the questionnaire as most
customers feedbacked the questionnaire as easy to understand. After the pilot conducted and
the adjusted questionnaire, the research moved to next step.
16
 

 

 

3.2.4 Step 3: Quantitative study
The quantitative questionnaire in thesis was developed basing on both applying the
theoretical background as well as the modified measurement items from qualitative
researches as stated. The research used five-point Likert scale with 1 mean “Strongly
disagree” to 5 means “Strongly agree”. In addition, before the full fieldwork for
quantitative, a pilot test was conducted on 10 respondents to check if all the questions are
clearly stated. Re-wording has been made accordingly. After adjusting the questionnaire
based on the feedbacks, the final quantitative questionnaire is presented in Appendix 1 (
English version) and Appendix 2 (Vietnamese version).
3.2.5 Sample size of Quantitative research
This research was conducted on a sample of 230 respondents conveniently selected.
Among those, only 207 ones meet the criteria respondents to answer the full questionnaire,
therefore, the final sample size used for the analysis is 207 respondents. A qualified
respondent must currently at the position of owner or manager who serves as operation and
tactical level of a veterinary medicine agency and has been in business relationship with at
least one veterinary medicine company. The sample size must satisfy the rule of thumb by
Hair et all (2010) that the sample size should be at least 5 times the number of observed
variables. It measn that the minimum size must be (=5 x 2). The current sample size
quantity met this critetion.
The critetion for respondents are:
• Males or females
• Currently live in the southern of Vietnam (from Hue to Ca Mau province).

17
 

 

 

• Currently at the owner or manager position as operation and tactical level of a
veterinary medicine agency.
• Must have a business relationship with at least one veterinary medicine company.
3.2.6 Sampling methodology
The target population for this survey includes the 200 to 250 veterinary medicine
agencies and stores in the Southern of Vietnam which spread from Hue to Ca Mau province.
Questionnaire will be distributed to all the agencies and the face to face interview will be
executed to the consisting of owners or directors of these agencies.

Among those

respondents, there is around 90 % of the questionnaire qualified for the analysis.
Those are unqualified including unqualified questionnaire with skipping questions or
illogical answers or conflict answer. Those non-qualified questionnaire were removed out of
the main sample.
3.2.7 Questionnaire explanation
The final questionnaire was structure into 3 main parts
• Screening: This part is to ensure the right qualified respondents selected. As the target
of the survey who currently is the manager or owner of a veterinary medicine agency
and has the business relationship with veterinary medicine companies. Those who are
not a manager or owner of any veterinary medicine agency will not go ahead
answering the questions.
• Main questionnaire: This part is asked to run the model to see the extension to which
respondents agree with different statement in the questionnaire. Those are critical
parts to measure the impact of different factors of relationship quality on customer

18
 

 

 

loyalty to validate out hypothesis. The main part used 5 point scale to capture
customer’s agreement level on variables of all research construct including Trust,
Satisfaction, Commitment and Customer Loyalty.
• Other bio-data question: This part is asked to have more insights for the analysis to see
the different information among each agency which will support the management
decision and actions. In this part, information is to capture the time agency in
operation, the average sales of agency and the number of staffs in each agency.
3.3

Data analysis
The collected data will be analyzed by IBM SPSS software (statistical package for

social science) for statistical purpose, Version 20 which could enable the exploratory
factor Analysis (EFA), Cronbach Alpha Reliability Analysis and standard Multiple
Regression Analysis.
3.3.1 Exploratory Factor Analysis (EFA)
The primary objectives of EFA are to determine the number of factors
impacting to the measures and the strength of the relationship between each factor and each
observed measure (De Coster, 2004). At

first,

Exploratory

Factor

Analysis

was

employed to explore the interrelationship among a set of variables so as to identify the
number of underlying factors. Principal Axis Factoring was used as a method to extract the
factors in this study.
In order to guarantee the suitability of the data for factor analysis, the
following conditions should be met (Pallant, 2005):

19
 

 

 

- The sample size should be appropriate: The sample size should be at least 100 and
there should be a ratio of at least file cases for each of variables. Our valid responses are
207 and therefore our data set was considered valid for this Exploratory Factor Analysis.
- The factorability of the data would be appropriate if:
• The Kaiser-Meyer-Olkin value (KMO) should be equal or greater than 0.5
• The Bartlett’s test of sphericity should be statistically significant: <0.05
- The number of factors were determined when:
• The components have an eigenvalue of 1 or more.
• The total variance explained by these components should be above 50%
• Factor loading criteria should be 0.5 or more to ensure a practical
significance.
3.3.2 Reliability Analysis
Cronbach’s alpha reliability analysis method will be used for examining the reliability
of the measurement scales. The scales are reliable when Cronbach’s alpha coefficient of
each scale is equal to or greater than 0.7 (Pallant, 2005). Besides, those items which have
the value of Corrected Item-to-total Correlation less than 0.5 could be discarded in
order to improve reliability level of the measurement scale.
Cronbach’s alpha is a tool for assessing the reliability of scales. It was provided by
Lee Cronbach in 1951 and is widely used the social science, business, and other fields. The
formulation of Cronbach’s alpha is:
The rule of thumb for evaluating the reliability of scales is as following:


If α >= 0.9 then the reliability of scales is excellent;

20
 

 

 



If 0.8 <= α < 0.9 then the reliability of scales is good



If 0.7 <= α < 0.8 then the reliability of scales is acceptable;



If 0.6 <= α < 0.7 then the reliability of scales is questionable;



If 0.5 <= α < 0.6 then the reliability of scales is poor;



If α < 0.5 then the reliability of scales is unacceptable.
The Cronbach’s Alpha was invoked to test the reliability of the answer from

respondents in this survey.
3.3.3 Multi-linear regression analysis
Multi-linear regression is a method used to model the linear relationship between a
single dependent variable and one or more independent variables. The dependent variable is
sometimes also called the predictable, and the independent variables the predictors.
Therefore, Multiple Linear Regression was used to test the research model and hypotheses
in this study. In accordance with Pallant (2005), the assumptions of a multiple regression
are:
-

The sample size is appropriate: n > 50+8m (where m = number of
independent variables)

- No multi collinearity is found.
-

No outliers are found or in case they are, no significant impact of outlier is found.

-

Normality & linearity should exist.

All of these assumptions should be tested before digging into further data
analysis. R-square would tell us how much of the variance in the dependent variable was
explained by the model.

21
 

 

 

3.4

Chapter 3 summary
This chapter provides information related to the method applied to test the hypotheses.

The methodology used in the study is 2 phase approach: qualitative and quantitative.
Qualitative survey was conducted first on 10 managers and owners of veterinary medicine
agencies having the business relationship with veterinary medicine companies to validate
the factors influencing on customer loyalty based on previous study and to explore other
factors applicable in Vietnam veterinary industry. A pilot was conducted before moving to
quantitative survey in some provinces spreading from Hue to Ca Mau. Total size of the
study is 207 respondents. Respondents must currently at the position of managers or owners
of veterinary medicine agencies and have relationship with veterinary medicine companies.

22
 

 

 

CHAPTER 4: DATA ANALYSIS, FINDINGS AND DISCUSSION
Research methodology was presented and the measures were preliminarily tested in
chapter 3. The results from the exploratory study showed that all measures are valid. This
chapter outlines the official assessment of measures and hypotheses testing as well as the
interpretation of the result, in particular, four parts are included in the chapter: part one
briefs the descriptive data analysis; part two present the assessment of measures adopting
Cronbach Alpha reliability analysis and Exploratory Factor Analysis; part three goes
through the hypotheses testing, part four presents discussion of results
4.1

Descriptive data analysis

4.1.1 Descriptive analysis of Sample charateristics
As above mentioned, among 230 respondent conveniently selected, only 207 ones
meet the criteria respondent to answer full questionnaire. The final sample size used for the
analysis is 2017 respondents have below characteristics:
Number of Partners: In this research, around 49.8% agencies have more than 10 partners
in their busisness while 38.2% have relationship with 5 – 10 veterinary medicine companies
and just 12.1% agencies have connected with 1 – 5 companies. The research also looked
into these number to evaluate the size and capabilities of the agencies.
Main Partner: Among veterinary medicine companies in Vietnam, the agencies tend to
choose the large, famous and high reputable in the market to cooperate with. BioPharmachemie is the company which has most chosen to be main partner by agencies with
30%. Second is Vemedimcompany with 23,7% while Bayer VN is the main partner of

23
 

 

 

23.2% respondent. Next are Anova and Hanvetcompany with the percentage is 13.5% and
2.4%, other companies accounted 7.2% selected by respondents.
Reason of Cooperation: In this study, the main reasons of agencies to cooperate with
companies are Good Relationship and Quality Products with 34.3% and 27.5% for each
reason. These numbers show how important of the relationship and product quality to the
customer’s mind to make the decision of cooperation. Beside, Price also the issue for
agencies if they want to do a business with veterinary medicine company with 20.3%
choices from respondent while Brand Name and Origin have the 10.6% and 7.2% .
Time of Cooperation: Most of agencies have a long time operation, times of cooperation
from 5 to 10 years is dominating with nearly 50% and more than 10 years operation have
16.9%. 28% of veterinary medicine agencies have worked from 1 to 5 years while just only
5% agencies have operated under 1 year.
Table 4.1: Descriptive Statistics of Sample
Frequency

Percent

Valid
Percent

Cumulative
Percent

NUMBER OF PARTNERS
1–5

25

12.1

12.1

12.1

5 – 10

79

38.2

38.2

50.2

> 10

103

49.8

49.8

100.0

Total

207

100.0

100.0

Bio-Pharmachemie

62

30.0

30.0

30.0

Vemedim

49

23.7

23.7

53.6

MAIN PARTNER

24