Study on learner’s satisfaction with training quality at job placement centers, hai duong city

INTRODUCTION

Research Background

Education training activity is a service activity and the quality of training is always a concern of students The students will feel satisfied with training activity if they receive advantages from courses and the training quality of the courses meets your expectations

Researches on customer satisfaction with service quality have been done in many different fields In Vietnam, studies on service quality have also been done in such fields as telecommunication services (Pham Duc Ky, 2007), banking services (Nguyen Thi Phuong Tram, 2008) or retail services, etc Recently there have also been some other researches on student satisfaction with courses at some university of big cities such as the research of Nguyen Dinh Tho in Ho Chi Minh city (2007) However, there have still not any studies done in Hai Duong province, which have students of short term courses at job placement centers as the research objectives Therefore, the research and evaluation of student satisfaction on training quality of courses in Job Centers are really necessary.

In the current training field, there are many training institutions and education centers affiliated with universities for training workers, so the competition to attract students with high-quality education is always a big concern of the centers One of the ways to attract more students to the center is to implement solutions for increasing the satisfaction of the students when taking the courses

Job Placement Center in Hai Duong province is now offering various courses for students, so the research to effectively improve training methods of the center becomes more and more essential To do that, there must be an assessment of factors affecting the quality of training at the center, so that we can propose solutions to improve the training quality and increase more satisfaction of students with their training courses.

Research Motivations

Why research on student satisfaction on the quality of training courses? Quality of course is the key to success and to attract more students to the center, so satisfying students with courses is an urgent need And Job Placement Center in Hai Duong province is also not an exception Competitiveness in the field of personnel training currently tends to increase because there are many more centers affiliated with universities and foreign institutions Therefore, it becomes more necessary to evaluate student satisfaction with courses of the center

Why select Hai Duong Job Placement Center? Hai Duong Job Placement Center is the biggest center of short-term training and job placement in Hai Duong province.Because the center has full characteristics of a typical short-term training center, so the research results can be used to refer to other units and partly reflects a market trend In addition, as an officer working in this center, doing research on evaluating student satisfaction with training quality of job placement center will become more necessary

At the moment, there have been not any studies carried out in Hai Duong province related to the field of evaluating student satisfaction on training quality of education courses In Vietnam, only a few studies done at the universities have the research objectives of undergraduates and graduate students So, a research with workers who have needs of short term training courses will help the testing of differences on the satisfaction with courses between different groups of research objectives Moreover, this study is one of the initial researches in the field of short term training not only in HaiDuong province but also in other provinces besides previous researches in big cities.The study will apply theories of service quality and some other previous researches in the same field to build observed variables and research model to suit research conditions of Vietnam.

Research purposes

This study has three main purposes as follows:

Firstly is to evaluate the satisfaction of students on the short term courses of the center by applying the CEQ (Course Experience Questionnaire) in Australia education (Ramden, 1991) which was adjusted and supplemented by Nguyen Dinh Tho (2007) in the research conditions of Vietnam.

Secondly is to assess and measure the impact intensity of each factor on the satisfaction of students in Hai Duong Job Placement Center

Thirdly is to suggest directions and solutions for improving the quality of training courses and increasing student satisfaction on the courses.

Research Procedures

The research procedures are described by following diagram (figure 1.4.1):

Figure 1.4.1 The research procedure The research procedures in detail are described as follows:

Scale verification Factor analysis Model testing

- Discuss to detect possible factors

Step 1: Define the research issues At this step, the author confirmed the thesis title with advisor and co-advisor Defined the research motivations, research purposes and considered materials related on the study

Step 2: Research theories and models of student satisfaction on courses In this study,

“training activity” was considered as a service activity, so the author researched models on service quality such as Kano’s model, SERVQUAL model, etc and other researches, reports in the research field of student satisfaction with courses such as the report of Australia education applying the CEQ (Course Experience Questionnaire), the research of Nguyen Dinh Tho (2007) using adjusted CEQ After that, the author selected CEQ model as a basic model to build theoretical research model for the study, and the model of Nguyen Dinh Tho (2007) was also selected as the basic model

Step 3: Scale selection After selecting the basic theoretical model, the author selected scales for measuring the research factors In this study, the author decided to select five point Likert scale to measure observed variables This was a qualitative research, so we could not use open questions, because they could not help to conclude which factor affecting student satisfaction on courses The author also could use Stapel scale, however the five point Likert was more popular and applied in many researches because it uses a range of positive numbers and has advantages when using mathematical methods as well as statistical methods for research definitions.

Step 4: Qualitative research To preliminarily evaluate aspects of each factor in the research model, the author referenced questionnaires from other previous researches,adjusted questions to suit the research conditions of a qualitative research At this step, the author used a group discussion with ten students who are studying at the center to assess each aspect of the factors Adjusted words in semantics, deleted meaningless questions and added other meaningful questions The results from this step are the final questionnaires which will be used to investigate with expected sample size

Step 5: Quantitative research and data processing After building and adjusting the final questionnaire, the author tested and processed data Distributed paper questionnaires to students studying at the center The obtained results would be cleaned and analyzed by SPSS 18.0 The author tested research factors by using Cronbach`s Alpha coefficient, and EFA, etc.

Step 6: Model adjustment After tested data by Cronbach`s Alpha and EFA, the theoretical research model would be adjusted based on the results of EFA analysis The author built research hypotheses for the new research model .

Step 7: Model testing and data processing After step 6, the author continued using the method of ordinary Least Squares (OLS) to build regression function and tested the suitability of the model by Scatterplot graph, Histogram graph, tested the linear dependence between independent variables, the residual normal distribution, the collinearity, etc Next, the author tested research hypotheses by coressponding p-value of t-statistics and F-statistics To test the differences between subtotals in the research, the author used ANOVA.

Step 8: Conclusions After analyzing and presenting research results at step 7, the author summed up main results and conclusions, assessed the distributions, limitations of the study as well as proposed directions for further researches.

LITERATURE REVIEW

Service

Service activities have a long history, however, there are many different concepts of service from different researchers This study will present some popular viewpoints as follows:

According to Zeithaml and Britner (2000),” a service is an act or performance offered by one party to another” to create value for customers to use service and satisfy needs and expectations of customers.

According to Kotler and Armstrong (2004), is an activity or benefits that business can offer to customers to build, strengthen and expand the long-term partnerships with them.

In this study, service is understood as the process and methods of organization and implementation to create utilities and using values to satisfy customer expectations. Therefore, training activities are considered as a process of organization to bring satisfaction on courses for the students In other words, training activity is a type of service

Researchers have different points of view on the concept of service, however in general, they all agree with the characteristics of service And service characteristics include following components:

Differ from other physical products (tangible), services are intangible that means, they cannot be touched, gripped, handled, looked at, smelled, tasted or heard. Furthermore, a service cannot be sold or owned by somebody, neither can it be turned over from the service provider to the service consumer nor returned from the service consumer to the service provider Quality of service is only shown in the interactive process between customers and employees of the company (Svensson, 2002)

Heterogeneity is reflected in the difference of level of service performance, which means that service can range from very poor to perfection Service quality has no heterogeneity between the time of service performance and depends on individual perceiveness on service With the same providing process, each customer has different perceiveness of the service quality

Heterogeneity related to high changes in the implementation of a certain service. Service quality may change according to providers, customers and service execution time This characteristic makes the standardization of services more difficult

The service provider is indispensable for service delivery as he must promptly generate and render the service to the requesting service consumer It is reflected in the difficulty of distinguishing the service creation process and service usage A service can not be separated into two separate processes: (1) service creation process and (2) service usage process They are simultaneous.

The creation and usage of a service simultaneously occur This is the big difference with other tangible goods Tangible goods are produced and put in storage or distributed to customers They are two separate processes and can be clearly separated.

In contrast, service is a process of creating, using at the same time, customers and service providers involved throughout the service creation process

Service has creation process and usage process which are simultaneous, so it can not be owned like other tangible goods In other words, the consumer does not secure ownership of the service and not set up a warehouse to store goods

The simultaneity of service is expressed in the concurrent process of creation and usage Services are rendered and consumed during the same period of time and we can not separate these two processes

Customer satisfaction on service is feelings of customers about services as they expected before using them The satisfaction is a measure of how products and services supplied by a company meet or surpass customer expectations Customer satisfaction is divided into three levels as follows:

 Dissatisfied: The customer feelings about services are not as what they expected before using them The reasons are the gap (distance) between the reality and the expectations of customers (see more in the service quality five gap model)

 Satisfied: Customers feel satisfied with services when their perception and feelings about services are as much as they expected

 Very satisfied: This level is when the values what customers get when using services is more than what they expected In other words, the real quality of services is much better than what customers think and expect, so they feel very satisfied

Perceived quality was defined, realized and applied in many different ways and at different levels including the excellence, values, the suitability with needs, ability to meet the expectations and beliefs of customers, etc (Reeves and Bednar, 1994).

Theories of service quality and customer satisfaction with services

The relationship between service quality and customer satisfaction is a cause and effect relationship, in which service quality is the cause and satisfaction is the effect In this study, the author will introduce some popular theories and models on service quality and customer satisfaction with services in many different fields.

2.1.1 Service quality five gap model

According to Parasuraman (1985), service quality is a function between customer expectations and the service that is provided to customers Customer expectations are built based on word-of-mouth information, individual needs and experiences of customers It includes five gaps as follows (figure 2.1.1):

Figure 2.1.1 The service quality five gap model Source: A.Parasuraman, Valarie A.Zeithaml and Leonard L Berry (1985).

Communications to customers Service Delivery

Change the company’s feeling into quality standard Gap 3

Gap 1: is the distance between what customers expect and what service providers think they expect Basic points of this difference is that the service company does not fully understand the characteristics which create the quality of its service as well as the way to transfer services to customers in order to meet their needs

Gap2: occurs when service providers have difficulty in changing their perception of customer expectations of customer expectations into the feature of service quality In many cases, the providers may be aware of customer expectations but they can not always transfer these expectations to the specific criteria of quality and transfer them right as customers expected

Gap 3: occurs when service staff does not transfer service to customer right according to the determined criteria In the field of service, the staff can directly contact with customers and play a very important role in the process of creating quality. However, the staff can not always fulfill their tasks in accordance with criteria set out.

Gap 4: is the gap between the delivery of the customer experience and what is communicated to customers All too often organizations exaggerate what will be provided to customers, or discuss the best case rather than the likely case, raising customer expectations and harming customer perceptions.

Gap 5: is the gap between a customer's perception of the experience and the customer's expectation of the service Customers' expectations have been shaped by word of mouth, their personal needs and their own past experiences.

Routine transactional surveys after delivering the customer experience are important for the company to measure customer perceptions of service.

SERVQUAL model is one of the most popular models that were applied in measuring customer satisfaction on service quality This model consists of five components as follows (figure 2.2.2):

Tangibles: are facilities' conduciveness to service

Reliability: is the capability to deliver dependable, accurate and consistent service

Responsiveness: can be understood as the willingness or readiness to provide service

Assurance: is likelihood that the service will actually be delivered

Empathy: is the caring, attention to customers, etc.

2.2.3 The model of Nguyen Dinh Tho et al (2007)

In Vietnam, Nguyen Dinh Tho researched the satisfaction of graduate students in the falculty of Business Administration in Ho Chi Minh city through using CEQ (Course Experience Questionaire) (Ramden, 1991, Wilson & Lizzio, 1997), which was adjusted and supplemented to suit the research conditions of Vietnam (figure 2.2.3).

Figure 2.2.3 The research model by Nguyen Dinh Tho

General satisfaction of students The attractiveness of scientific knowledge Objectives and standards

Lecturer team: Lecturer team includes lecturers who are teaching at the university and lecturers from other universities and colleges who are not on the payroll of the university A good lecturer team will guide and help students to know their progress in learning (Nguyen Dinh Tho, 2007) In addition, a good lecturer team can also answer all questions of students, they are dedicated and have ability to make student’s learning become more interesting (Ramsden, 1991)

Generic Skills Development: Generic skills are skills, which students obtain from courses such as: problem solving skills, analytical skills, skills in collaborative group work, planning skills, etc (Nguyen Dinh Tho, 2007) Generic skills are responses to the level that students are trained, developed, and recognized as a valuable outcome of higher education outside the discipline; knowledge and other specific skills (Ramden, 1991).

Graduate Quality: The graduate quality reflects skills, knowledge that the students gained from the course, and also shows assessment possibility of students’ strengths after the course The graduate quality is assessed by ability to apply principles, knowledge obtained from specific situations, ability to explore new issues.

Objectives and Standards: Clear objectives and standards include requirements and standards of learning, which are known easily Students understand and know their expectations about the course and jobs Lecturer team knows the expectations of the course and requires their students to fulfill on the first lesson (Nguyen Dinh Tho, 2007).

They can clearly and plainly use teaching practices according to above criteria of good lecturer (Ramsden, 1991).

Learning Workload: An appropriate learning workload mentions the appropriation of knowledge workload that students obtained from the course, suitability of time for learning and depth studying subjects through references Learning does not create too much pressure on students, and the students can acquire knowledge, skills firmly and fully.

Learning Resources: Learning resources can be understood as resources that provide for student’s learning process, including: learning materials from library, facilities for teaching and learning, update references, etc Learning resources are also university’s ability to meet students’ needs on learning conditions during the course. Learning resources influence the quality of learning and student’s satisfaction with the course.

The attractiveness of scientific knowledge: is a good motivation of students when participating on the course Students feel interested in scientific knowledge that they can obtain from the course, and feel the course is worth.

Related researches on customer satisfaction with services

Researches on the field of evaluating customer satisfaction with services have been done for many years, using different research models In the scope of this study, the author will introduce some researches in education service sector which have reference values for the study (table 1).

Table 1 Summary of research results on student’s satisfaction with courses

Article title Research model Factor Research field

Explaining user satisfaction with academy libraries : strategic implications,

Service quality and satisfaction: A case study at Private higher education institutions,

Enhancing the quanlity of engineering education instittutions (EEIs) through analysis, International quality conference, 5,

The perpormence of Academic Libraries: A case study at research university (Rus) in Malaysia, Global of human Social Sciene,

Teaching evaluations at the introductory finance course at Lund University: A comparison of the Course Experience Questionnaire an a traditional evaluation approach, Lund

1 Good Teaching scale 2.Clear Goals and Standards Scale

Course Experience Questionnaire (CEQ) report, Course

Experience Questionnaire, James Cook University, 1-96

Survey on student satisfaction on Master’s Training Quality at University of Economics, Ho Chi Minh city

7 The attractiveness of scientific knowledge

RESEARCH METHODOLOGY

The background of the research unit

Hai Duong Job Placement Center was established in 1989 and is a public career unit which is responsible for career consultant and short term career training

The center has legal status, seal and may open accounts in accordance with the law, under the leadership and direct management of the Department of Labour, Invalids and Social Affairs In addition, the center is simultaneously under the guidance of speciality, the inspection and supervision by the authorized state agencies The address of the center is at No 106 Hong Quang Street, Quang Trung, Hai Duong city, Hai Duong The organization structure of the center consists of 1 director, 33 vice directors and 7 departments/ bureaus, including (1) Administrative and organizational department, (2) Planning and financial department, (3) Training bureau, (4) Unemployment insurance department, (5) Consulting department, (6) Labor market information department, and (7) Job placement department

Besides the main duty of job placement, the center is also responsible for organizing short term courses for those who have needs of apprenticeship, and partly ensuring costs to the operation of the center.

Research model

The research model selected is the CEQ model of Australia education, which was adjusted to suit the research conditions of Vietnam by Nguyen Dinh Tho et al (2007). The research model includes factors as follows:

Research hypotheses

Lecturer team: Lecturer team includes lecturers who are teaching at the university and lecturers from other universities and colleges who are not on the payroll of the university A good lecturer team includes teachers who provide and help students to know their progress in learning (Nguyen Dinh Tho, 2007) A good lecturer team can also answer all questions of students, they are dedicated and have ability to make student’s learning become more interesting (Ramsden, 1991) It also has influence

General student satisfaction The attractiveness of scientific knowledge

H4 H8 on the feelings of students about the quality of courses and then the students will have positive assessment of trainng quality In other words, a good lecturer team positively affects student satisfaction with the course they are studying This is proven by the research of Nguyen Dinh Tho (2007), Hans NE Bysstrom (2004), Craig Mclnnis et al.

(2000, 2001), Francisco Cano and A,B.G Berben (2009), the report of James Cook university (2011) in Australia education, etc Therefore, this study gives out following hypothesis:

H1: Satisfaction with lecturer team positively affects general student satisfaction on the course.

Generic skills: Generic skills are skills, which students obtain from courses such as: problem solving skills, analytical skills, skills in collaborative group work, planning skills, etc Generic skills are responses to the level that students are trained, developed, and recognized as a valuable outcome of higher education outside the discipline; knowledge and other specific skills (Ramden, 1991) Generic skills development positively affects general student satisfaction on the course A good training environment is a place which has not only professional knowledge provision but also supports for students in generic skill development to adapt to labor market. Generic skills development will help students feel satisfied with the quality of the course This is proven by the research of Nguyen Dinh Tho et al (2007), Nifarta Peingurta Andrew (2010) Jan F.H Nijhuis et al (2005), Nina Clemson (2009), the report of James Cook university(2011), etc Therefore, this study proposes hypotheses as follows:

H2: Satisfaction with generic skills development positively affects general student satisfaction on the course.

Graduate quality: is the assessements and feelings of students about the skills, knowledges that they obtained from the course The graduate quality is also evaluated based on the ability of principle and knowledge application in new situations and new issues Differ from other researches, the concept of “graduate quality” in this study is the expectations of students about the course The students expect to obtain skills and knowledge and other advantages from the course Graduate quality is considered as the signals of the ability that the students will feel satisfied with the quality of the course. Thus, this study hypothesizes that:

H3: Satisfaction with graduate quality positively affects general student satisfaction on the course.

Clear objectives and standards: Clear objectives and standards include requirements and standards of learning, which are known easily Students understand and know their expectations about the course and jobs Lecturer team knows the expectations of the course and require their students to fulfill right on the first lesson(Nguyen Dinh Tho, 2007) They can clearly and plainly use teaching practices according to above criteria of good lecturer (Ramsden, 1991) Clear objectives and standards influence students’ feelings about the course The students will feel more satisfied with the course if the objectives and standards of the course are clear This is proven by the research of Nguyen Dinh Tho et al (2007), Francisco Cano and A,B.G

Berben (2009), Nifarta Peingurta Andrew (2010), Craig Mclnnis et al (2000, 2001), etc. For this reason, it is necessary for the study to launch the hypothesis as follows:

H4: Satisfaction with objectives and standards of the course positively affects general student satisfaction on the course.

Learning workload: An appropriate learning workload is the appropriation of knowledge workload that students obtained from the course, suitability of time for learning and depth studying subjects through references Learning does not create too much pressures on students, and the students can acquire knowledge, skills firmly and fully Appropriate workload will bring students creativeness, help them deeply study subjects and obtained knowledge It does not put pressures on students and make them dissatisfied with the course In other words, appropriate workload has an influence on student’s assessment with the course The research by Biggs (1999) showed that learning process (including learning workload) influences learning results of students. The researches by Hans NE Bysstrom (2004), by Nguyen Dinh Tho et al (2007), etc. also showed that appropriate workload positively affects student satisfaction on the course Thus, this study hypothesizes that:

H5: Satisfaction with learning workload positively affects general student satisfaction on the course.

Learning resources: can be understood as resources that provide for student’s learning process, including: learning materials from library, facilities for teaching and learning, update references, etc Learning resources are also university’s ability to meet students’ needs on learning conditions during the course Learning resources influence the quality of learing and student’s satisfaction with the course The learning resources in CEQ model have similarities with the factor “tangibles” in SERVQUAL model, so it is also one of factors affecting student satisfaction on courses This is proven by the researches of Hishamuddin Fitri Abu Hasan et al (2008), Sri R Chandra Shekhar el al

(2011), Nguyen Dinh Tho et al (2007), etc So, the study to launch the hypothesis as follows:

H6: Satisfaction with learning resources positively affects general student satisfaction on the course.

The attractiveness of scientific knowledge: is good motivations of students when taking the course Students are interested in scientific knowledge that they can obtain from the course, and feel the course is worth A course with activities of the attractiveness of scientific knowledge will be appreciated and also influence student satisfaction with the course (Nguyen Dinh Tho, 2007) Therefore, this study gives out following hypothesis:

H7: Satisfaction with the attractiveness of scientific knowledge positively affects general student satisfaction on the course.

Course organization: consists of activities related on arranging a course, including: information about the course and subjects; subjects are organized orderly and systematically; flexibility in organizing the course Course organization is also reflected in the ability to choose subjects that students want to learn, appropriate subjects, and ability to approach the depth of certain knowledge Course organization is good if it meets students’ needs of expectations The research by Nguyen Dinh Tho et al (2007), research by Craig Mclnnis et al (2000, 2001), research by Nifarta Peingurta Andrew

(2010), the report of James Cook university, etc showed that good course organization has positive impact on student satisfaction with the course Thus, this study hypothesizes that:

H8: Satisfaction with course organization positively affects general student satisfaction on the course.

Variable measurement

The observed variables of this study are referenced from the research of Nguyen Dinh Tho et al (2007), CEQ in Australia education (Ramsden,1991, 1999, Wilson & Lizzio, 1997) an the literature review on service quality They are also collected and adjusted after doing a qualitative research with some students taking the course of the center The contents in detail are as follows:

Table 2 Observed variables for each factor

No Code Contents of questions References

1 GV1 The lecturer team encourages and motivates you to get the best learning outcomes?

2 GV2 The lecturer team has attempted to understand the difficulties you encountered in the learning process?

3 GV3 The lecturer team often gives you useful information about what you should do next?

4 GV4 The lecturer team explains everything very clearly and understandable?

5 GV5 The lecturer team works conscientiously and seriously to make their subjects more interesting?

6 KN1 The course develops your problem solving skills?

(1997), Nguyen Dinh Tho et al. (2007).

7 KN2 The course develops your analysis skills more in- depth?

8 KN3 The course helps you develop skills in collaborative group work?

9 KN4 By taking a course you feel more confident when facing hindering issues or new problems?

10 KN5 The course develops planning skills for yourself?

11 CL1 The course provides you furthur and deep knowledge?

Nguyen Dinh Tho et al (2007)

12 CL2 The course encourages you to assess your strengths and abilities?

13 CL3 You can apply the principles and learned knowledge in new situations?

14 CL4 This course gives you confidence to explore new issues?

15 CL5 You think what you learned and obtained from the course are valuable for your future?

16 MT1 The standards and requirements of studying are known easily?

17 MT2 You know exactly your expectations about the course and what you should do?

18 MT3 You actively explore and discover what people expect from you in the course?

19 MT4 Lecturers clarify what they expect and require the students to do right on the first lesson?

20 KL1 Learning workload is not too much?

(1997), Nguyen Dinh Tho et al (2007)

21 KL2 You have much time to study other materials for subjects?

22 KL3 The course does not creates much pressures in your learning process?

23 KL4 Learning workload is logical to acquire appropriate knowledge and skills?

24 NL1 Resources from library meet your needs? Ramden (1991),

25 NL2 Devices for learning and teaching operate effectively?

Nguyen Dinh Tho et al (2007)

26 NL3 The learning materials are clear and concise?

27 NL4 The learning materials of course are areappropriate and updated?

VII The attractiveness of scientific knowledge

28 TT1 You find it interesting in the attractiveness of scientific knowledge in your learning process?

Nguyen Dinh Tho et al (2007)

29 TT2 You have good motivations when taking the course?

30 TT3 In general, you think your learning process is very worthwhile?

31 TT4 The activities related to the course have been done well?

32 TC1 You get useful information and advices for your learning plan?

Ramden (1991), Nguyen Dinh Tho et al (2007)

33 TC2 The subjects of the course are organized in a systematic way?

34 TC3 The course is very flexible to meet your needs?

35 TC4 The number of subjects in the course is appropriate?

36 TC5 The subjects in the course are built on a satisfactory knowledge?

37 TC6 The course meets all your requirements?

38 HL1 In general, you feel satisfied with the quality of the course? Nguyen Dinh Tho et al (2007)

39 HL2 The course meets your expectations?

40 HL3 You think that the university is a perfect place for you to perfectly take the course?

Research Design

This study uses step of qualitative research through a group discussion with fifteen students taking courses in Hai Duong job placement center and teaching management officials of the center The purposes of this step is to explore and evaluate attitudes, opinions and responses of students to the courses of the center The discussion contents revolves around the factors from CEQ and the research model of Nguyen Dinh Tho (2007).

Discussion results show that there are many different opinions but in the main, most opinions agree that, all factors in the research model of Nguyen Dinh Tho have influence on student satisfaction with the course.

Through this step, the author also adjusts investigation questions for each factor to suit research conditions, sends some questionnaire in advance to avoid difficulties for students before official questionnaire is handed out (see more in the appendix of the group discussion outline).

Scale selection for the study is always necessary Scale used in the study can be open or close questions Open questions will help the author collect more opinions and reject the subjective imposition of the author on respondents Open questions have advantages in exploring inside-information of respondents, however, for a research with large sample size, open questionnaire is not feasible because of the diversity of responses Answers are not focused, so make it difficult to find overall and specific trendancy In contrast, close questions just allow respondents to choose available answers, so it will facilitate the author to summarize, encode and calculate by statistical methods The disadvantages of this question form are that, the answers are dominated by the subjectives of the author, and also depend on the author’s experiences in building investigation questionnaire However, this can be resolved by conducting a sample interview, and adjusted questionnarie to suit interview subjectives Therefore, in this study, the author chooses close questions to design scales for each factor in the research model.

For the close questions, the author can use Stapel or Likert scale to measure. However, Likert scale still has much more advantages than Stapel because it uses a range of positive numbers, then will be more convenient for the calculations by the statistical methods For Likert scale, there are many different levels of agreement, it may be from three to nine levels or even more In principle, the more detail the scale is, the more accurate it is, so in this study, the author decides to select the seven point Likert scale

After selecting measurement criteria (investigation questions) for each factor, the author starts building questionnaire Questionnaire is built based on the principle that, questions must be easy to understand and convenient to data encryption and cleaning To ensure the best quality of official questionnaire, the author builds a draft questionnaire to make a discussion in order to select the most appropriate presentation form After that, the draft questionnaire is handed to respondents to adjust the questionnaire once again to have the final questionnaire.

Due to the limitation of time and cost to implement the research, the author uses the convenience sampling method After defining the expected sample size, the author distributes questionnaire for research onjectives untill getting the valid answers as expected.

To ensure the reliability of the research, selection of an appropriate sample size is necessary In principle, the more number of samples are, the more accurate research results are, however, a too big sample size will affect the cost and time to conduct the research Moreover, the long research time can make the research results become less reliable and tend to change.

For this study, due to the cost limit in implementing the research, the sample size should be determined on the principle of minimum necessary to ensure the reliability of the study The expected number of samples are 200 samples, and to ensure this sample size, 250 samples are distributed.

The determination of how appropriate the sample size is still has many controversies Maccallum et al (1999) sumed up opinions of previous researchers about principle of minimum sample size with Factor Analysis According to Kline (1979), the minimum number of samples is 100, but according to Guiford (1954), the number of samples is 200 and Comrey & Lee (1992) gave the sample size for the respective views with same opinions: 100 = bad, 200 = pretty, 300 = good, 500 = very good, 1000 or more = excellent.

Some researchers did not give specific numbers but the relationship between the number of observed variables with the sample size According to Trong and Ngoc (Vietnam, 2008) using the 5 - Power Rule, i.e the number of samples x 5 = the minimum sample size of the study to ensure reliability According to Tabachnick and Fidell (2007), we can use the following empirical formula to determine the minimumsample size: n >= 50 + 8p, in which: n is minimum sample size, p is factor number (independent variable) in the research model.

In this study, the principles of minimum sampling are based on the principles of Tabachnick and Fidell (2007) According to this, minimum sample size is n >= 50 + 8x8 = 114, thus, the expected sample size 200 is appropriate It is a good sample size according to Comrey and Lee (1992).

The research objectives include students who are taking the short term courses in Hai Duong job placement center

Survey organization method: To collect expected samples, the author uses direct survey by paper questionnaires.

Data analysis method

The obtained data are cleaned and analysized by using SPSS 18.0 through statistical steps, including:

The obtained samples are statistically analyzed according to classified variables based on classification criteria, such as: gender, age, etc At the same time, the author also calculates average value, maximum value, minimum value and standard deviation of the answers in collected questionnaire.

The observed variables are verified by using the Cronbach `s Alpha and the method of Item-total correlation The observed variables which do not ensure about the reliability will be removed from the scale and do not appear in the explore factor analysis (EFA).

In this study, Cronbach `s Alpha coefficient must be at least 0.6 (Hair et al.,

1998) If the total correlation coefficient is less than 0.3, it is considered as a “spam variable” and naturally eliminated from the scale (Nunally and Burstein, 1994)

The factors after Cronbach`s Alpha coefficient tested will continuously be analyzed by EFA method Some standards applied in analyzing the EFA in the study are as follows:

- Testing the suitability of factor analysis through Kaiser-Meyer-Olkin value (KMO) If Kaiser-Meyer-Olkin coefficient (KMO) is greater than 0.5 and p-value (sig.) of Barlett-test is less than 0.05, the factor analysis will be appropriate to data.

In contrast, it will not be appropriate to factor analysis (Garson, 2003).

- The number of factors: The number of factors is determined based on the eigenvalue index which represents the variation explained by each factor According to Kaiser’s standards, the factors with an eigenvalue index less than 1 will be removed from the research model (Garson, 2003).

- Variance explained criteria: The total variance explained criteria must be greater than 50% (Hair et al., 1998).

- Factor loading: To achieve the convergence value, the correlation coefficient and factor loading of each factor must be greater than or equal to 0.5 (Garbing & Anderson, 1988).

- The distinct value: To meet the distinct value, the value of distinction between the factor loadings must be greater than or equal to 0.3 (Jabnoun, 2003).

- Principal components method with Varimax rotation to ensure the number of factors is smallest (Ngoc and Trong, 2008).

3.6.4 Naming and adjusting the research model

After doing the explore factor analysis, the author renames the factors and adjusts the model as well as original research hypotheses to fit the actual data.

After being tested, the scales would be processed by running linear regression by method of ordinary least squares (OLS) with both Enter method and Stepwise method.

3.6.6 Testing the suitability of the research model

After building the regression funtion by method of OLS, to ensure the reliability of the model, the author makes finding the satisfaction with the hypotheses of the OLS model, including:

 Linear dependence between independent variables: Using the scatterplot graph to test the linear dependence hypothesis contact between independent variables in the model.

 Residual autocorrelation: Using the Spearman’s Rank Correlation Coefficient to test the residual autocorrelation hypothesis.

 Testing the residual normal distribution: Using the P-Plot and Histogram graphs to test the residual normal distribution hypothesis.

 Collinearity Diagnostics: When a regressor is nearly a linear combination of other regressors in the model, the affected estimates are unstable and have high standard errors This problem is called collinearity or multicollinearity Using Collinearity Dianostics to find out which variables are nearly collinear with which other variables In the Collinearity Dianostics, the tolerance or the VIF (variance inflation factor) are used According to Hoang Trong & Mong Ngoc (2008), when VIF is less than or equal to 10, it means that the independent variables do not correlate linearly with each other.

 Heteroskedasticity: In statistics, a collection of random variables is heteroscedastic if there are sub-populations that have different variability than others When the variance of the errors changes, the estimates of the regression coefficients are not effective, and the T test and F test are no longer reliable.

 Testing the autocorrelation between variables: This is a basic violation of hypothesis, If skipping the testing of correlation, the predictations and estimates are still unbiased and consistent but not efficient, In this case Durbin-Watson test is the most popular test.

After checking, if the research results do not violate, we can conlude that, estimates of regression coefficients are unbiased, consistent and effective, and the conclusions obtained from regression analysis are reliable.

The research hypotheses would be tested based on research data from regression function The testing standards used the t statistics and the p-value (sig.) With the reliability of this study which was calibrated by 95%, we would consider the p-value with 0.05 to conclude the posed research hypotheses For testing the difference between subtotals, the study used t-test and ANOVA to test each hypothesis, and the testing standards based on the corresponding p-value for each specific testing step To test the suitability of data and model, the author used F statistics, t statistics, and R-square In addition, to assess the importance of factors, the author considered the Beta coefficient in the regression function.

RESEARCH RESULTS

Descriptive Statistics

The minimum sample size is 200, so to ensure this one, 250 questionnaires were distributed, and from which collected 218 samples that were valid to analyze The sample structure based on demographic signs was classified as follows:

In 211 valid questionnaires from actual data, there are 80 ones from male students(38%), and 131 other ones from female students (62%) This ratio shows that the tend of short term training of females is double higher than males.

Figure 4.1.4 The sample structure by gender

In 211 valid questionnaires, the group under the age of 25 has the biggest percentage of 62% (130 students), next is the group at the age from 26 to 30 with 61 persons (29%), and other groups at older age have quite low ratio This exactly reflects the reality that young persons have often trend to spend much time on learning and training Moreover, this group includes persons who are at the beginning of working age and need a work for life and their own career, so that is the reason why this has a large proportion in the sample structure.

Figure 4.1.2 The sample structure by age

The sample structure by income shows that the majority of students has income less than 3 mil./ month with 149 persons (71%), next is the group of students who have income from more than 3mil to 5 mil./ month makes up 22% with 47 persons The last group of students who have income more than 5 mil./ month just has a small percentage.

Figure 4.1.6 The sample structure by income

Summary of results from questionnaires

The research results show that the answers for questions of factors in the research model aer distributed from level 1 to level 7 The majority of answers has ponits greater than 4 and a quite big standard deviation This proves that the level of assessment of

N Minimum Maximum Mean Std Deviation

Testing the reliability of research scales

As presented in chapter 3, before the step of explore factor analysis, the author will test the reliability of scales in the resarch model to remove inappropriate variables. The test standards are that Cronbach`s Alpha coefficient must be at least 0.6, the total correlation coefficient of variables also must be at least 0.3, and the observed variables which do not ensure these standards will be deleted from the research model The results from scale verification for each factor are as follows:

4.3.1 Testing the reliability of scales for the factor “lecturer team”

The implicit variable “lecturer team” in the research model is measured by five observed variables from GV1 to GV5 From the actual data, results from testing the reliability of scales for this factor show that Cronbach`s Alpha coefficient is 0.860 > 0.6, the total correlation coefficients of observed variables are greater than 0.3 Thus, we can conclude that the scales for the factor “lecturer team” are reliable and appropriate

Table 3 Item (GV)-Total Statistics

Scale Mean if Item Deleted

Scale Variance if Item Deleted

4.3.2 Testing the reliability of scales for the factor “generic skill development”

The independent variable “generic skill development” in the research model is measured by five observed variables from KN1 to KN5 From the actual data, testing the reliability of scales for this factor gives out Cronbach`s Alpha coefficient = 0.887

>0.6, the total correlation coefficients of observed variables are greater than 0.3. Therefore, we can conclude that the scales for the factor “generic skill development” are reliable and appropriate

Table 4 Item (KN)-Total Statistics

Scale Mean if Item Deleted

Scale Variance if Item Deleted

4.3.3 Testing the reliability of scales for the factor “graduate quality”

The implicit variable “graduate quality” is measured by five observed variables from CL1 to CL5 From the actual data, the results show that Cronbach`s Alpha coefficient is 0.882 > 0.6, the total correlation coefficients of observed variables are greater than 0.3 so, we can conclude that the scales for the factor “graduate quality” are reliable and appropriate.

Table 5 Item (CL) -Total Statistics

Scale Mean if Item Deleted

Scale Variance if Item Deleted

Cronbach's Alpha if Item Deleted

4.3.4 Testing the reliability of scales for the factor “Objectives and standards”

The implicit variable “objectives and standards” in the research model is measured by four observed variables from MT1 to MT4 From actual data, the results show that Cronbach`s Alpha coefficient is 0.842 >0.6, the total correlation coefficients of observed variables are greater than 0.3 So, we can conclude that the scales for the factor “objectives and standards” are reliable and appropriate.

Table 6 Item (MT)-Total Statistics

Scale Mean if Item Deleted

Scale Variance if Item Deleted

Cronbach's Alpha if Item Deleted

4.3.5 Testing the reliability of scales for the factor “learning workload”

The implicit variable “learning workload” in the research model is measured by four observed variables from KL1 to KL4 From actual data, the results show that observed variables are greater than 0.3 so, we can concude that the scales for the factor

“learning workload” are reliable and appropriate.

Table 7 Item (KL) -Total Statistics

Scale Mean if Item Deleted

Scale Variance if Item Deleted

4.3.6 Testing the reliability of scales for the factor “learning resources”

The implicit variable “learning resources” in the research model is measured by four observed variables from NL1 to NL4 From actual data, the results show thatCronbach`s Alpha coefficient is 0.814 > 0.6, the total correlation coefficients of observed variables are greater than 0.3 Therefore, we can conlcude here that the scales for the factor “learning resouces” are reliable and appropriate.

Table 8 Item (NL) -Total Statistics

Scale Mean if Item Deleted

Scale Variance if Item Deleted

4.3.7 Testing the reliability of scales for the factor “the attractiveness of scientific knowledge”

The implicit variable “the attractiveness of scientific knowledge” in the research model is measured by four observed variables from TT1 to TT4 From actual data, the results show that Cronbach`s Alpha coefficient is 0.899 > 0.6, the total correlation coefficients of observed variables are greater than 0.3 Thus, we can conclude that the scales for the factor “the attractiveness of scientific knowledge” are reliable and appropriate.

Table 9 Item (TT) -Total Statistics

Scale Mean if Item Deleted

Scale Variance if Item Deleted

4.3.8 Testing the reliability of scales for the factor “course organization”

The implicit variable “course organization” in the research model is measured by six observed variables from TC1 to TC6 From actual data, the results show that Cronbach`s Alpha coefficient is 0.888 >0.6, the total correlation coefficients of observed variables are greater than 0.3 So, we can conclude that the scales for the factor “course organization” are reliable and appropriate.

Table 10 Item (TC) -Total Statistics

Scale Mean if Item Deleted

Scale Variance if Item Deleted

4.3.9 Testing the reliability of scales for the dependent variable

The dependent variable “general satisfaction” in the research model is measured by three observed variables from HL1 to HL3 From actual data, the results show thatCronbach`s Alpha coefficient is 0.899 > 0.6, the total correlation coefficients of observed variables are greater than 0.3 So, we can conclude that the scales for the dependent variable “general satisfaction” are reliable and appropriate.

Table 11 Item (HL) -Total Statistics

Scale Mean if Item Deleted

Scale Variance if Item Deleted

Cronbach's Alpha if Item Deleted

4.3.10 Summary of results from scale verification

To facilitate the viewing of research results, we will summarize obtained results as following table.

Table 12 Summary of results from scale verification for factors

Objectives and standards 0.842 4 Reliability achieved

The attractiveness of scientific knowledge

Explore factor analysis

After observed variables for each factor in the research model are tested, they will continuously be tested by the method of explore factor analysis (EFA) The standards here are: KMO coefficient (factor loading coefficient) must be greater than 0.5 (Hair et al., 2006), Bartlett-test has p-value< 0.05, the variance extracted must equal to at least 50% To analyze factor, Principal components with Varimax rotation will be applied in order to ensure the minimum number of factors (Hoang Trong and Chu Nguyen Mong Ngoc, 2008) The independent variables are analyzed at the same time, and the dependent variable is particularly analyzed Then we get the results as follows:

4.4.1 Explore factor analysis with independent variables

From actual data, we do EFA and delete variables which have factor loading from the research model and get the following results:

Table 13 KMO and Bartlett's Test with independent variables

Kaiser-Meyer-Olkin Measure of Sampling Adequacy .938

Table 14 Rotated Component Matrix with independent variables

Extraction Method: Principal Component Analysis.

Rotation Method: Varimax with Kaiser Normalization. a Rotation converged in 6 iterations.

The results from EFA show that KMO coefficient = 0.938 > 0.5, Bartlett-test has p-value = 0.000< 0.05, factor loading coefficients are greater than 0.5, observed variables form six factors, the variance extracted = 68.358% proves that six new formed factors can explain 68.358% of the variability of the data set.

4.4.2 Explore factor analysis with the dependent variable

The results from doing EFA with the dependent variable from actual data are as follows:

Table15 KMO and Bartlett's Test with the dependent variable

Kaiser-Meyer-Olkin Measure of Sampling Adequacy .718

Table16 Total Variance Explained with the dependent variable

Initial Eigenvalues Extraction Sums of Squared Loadings

Extraction Method: Principal Component Analysis.

Table 17 Component Matrix with the dependent variable

Extraction Method: Principal Component Analysis. a 1 components extracted.

The results show that KMO coefficient = 0.718 >0.5, Bartlett-test has p-value 0.000 < 0.05, the variance extracted is 83.477% > 50%, factor loading coefficients are greater than 0.5 Observed variables form only one factor, so the scale for the factor

“general satisfaction” is a unidirectional scale.

Naming new factors and adjusting the research model

After doing Explore factor analysis with factors in the research model, we will rename each formed factor based on the meaning that they reflect The contents in detail are as follows:

The results from EFA show that the first factor is formed by following observed variables:

1 The course provides you furthur and deep knowledge (CL1)

2 The course encourages you to assess your strengths and abilities (CL2)

3 You can apply the principles and learned knowledge in new situations (CL3)

4 This course gives you confidence to explore new issues (CL4)

5 You think what you learned and obtained from the course are valuable for your future (CL5)

6 The course develops your problem solving skills (KN1)

7 The course develops your analysis skills more in-depth (KN2)

8 The course helps you develop skills in collaborative group work (KN3)

9 By taking a course you feel more confident when facing hindering issues or new problems (KN4).

10 The course develops planning skills for yourself (KN5)

If the meaning of these items considered , these observed variables belong to two factors: “graduate quality” and “generic skill development”, so we name the first factor as “ graduate quality and generic skill development ”

The results from doing EFA show that the second factor is measured by following observed variables:

1 You find it interesting in the attractiveness of scientific knowledge in your learning process (TT1)

2 You have good motivations when taking the course (TT2)

3 In general, you think your learning process is very worthwhile (TT3)

4 The activities related to the course have been done well (TT4)

All these factors belong to the factor “the attractiveness of scientific knowledge” in the theoretical model, so we still keep the old name “the attractiveness of scientific knowledge” for the second factor.

The results from doing EFA show that the third factor is formed by following observed variables:

1 The lecturer team encourages and motivates you to get the best learning outcomes (GV1)

2 The lecturer team has attempted to understand the difficulties you encountered in the learning process (GV2)

3 The lecturer team often gives you useful information about what you should do next (GV3)

4 The lecturer team explains everything very clearly and understandable (GV4)

These observed variables belong to the factor “lecturer team” in the theoretical research model, so we still keep this name “lecturer team” for the third factor

The results from doing EFA show that the fourth factor is built by following observed variables:

1 Resources from library meet your needs (NL1)

2 Devices for learning and teaching operate effectively (NL2)

3 The learning materials are clear and concise (NL3)

4 The learning materials of course are areappropriate and updated (NL4)

All these observed variables belong to the factor “learning resources” in the theoretical research model, so we still keep this name “learning resources” for the fourth factor

The research obtained after doing EFA show that the fifth factor is formed by following observed variables:

1 The standards and requirements of studying are known easily (MT1)

2 You know exactly your expectations about the course and what you should do (MT2)

3 You actively explore and discover what people expect from you in the course (MT3).

4 Lecturers clarify what they expect and require the students to do right on the first lesson (MT4).

These observed variables all belong to the factor “objectives and standards” in the theoretical research model, so we still keep the name “objectives and standards” for this factor.

The results from doing EFA show that the factor 6 is formed by following observed variables:

1 Learning workload is not too much (KL1)

2 You have much time to study other materials for subjects (KL2)

3 The course does not creates much pressures in your learning process (KL3).

4 Learning workload is logical to acquire appropriate knowledge and skills (KL4).

These observed variables belong to the factor “learning workload”, so we still keep this name “ learning workload ” for the factor 6.

4.5.7 Adjusting the research model and hypotheses

The results from EFA show that it is needed to adjust the research model and hypotheses to match actual data The contents in detail are as follows:

Figure 4.5.7 The adjusted research model The new research hypotheses are:

Graduate quality and generic skill development

The attractiveness of scientific knowledge

H1: The factor “graduate quality and generic skill development” positively affects general satisfaction on the course.

H2: The factor “the attractiveness of scientific knowledge” positively affects general satisfaction on the course.

H3: The factor “lecturer team” positively affects general satisfaction on the course. H4: The factor “learning resources” positively affects general satisfaction on the course.

H5: The factor “objectives and standards” positively affects general satisfaction on the course.

H6 The factor “learning workload” positively affects general satisfaction on the course.

Correlation analysis

To test the relationship between factors in the research model, we test the Pearson correlation coefficient of them The greater the correlation coefficients of the factors are, the closer the relationship between them is If the correlation coefficient = 0, there is no relationship between factors According to this, we get following results:

** Correlation is significant at the 0.01 level (2-tailed).

F1 is graduate quality and generic skill development

F2 is the attractiveness of scientific knowledge

HL is general satisfaction with the course

The results from correlation analysis show that all correlation coefficients of factors are differ 0 This means the dependent variable and independent variables in the research model have relationship with each other This is also a signal for the possibility of the multi-collinear when doing regression analysis.

Building the regression function

To test the research hypotheses, we need to build the multicollinearity regression function of the dependent variable HL on other independent variables The method used here is the method of Ordinary Least Squares (OLS) and Enter method The contents in detail are as follows:

4.7.1 Estimating the model by Enter method

The results of estimation of the regression function by Enter method are obtained as follows:

Std Error of the Estimate

1 804 a 647 636 79543 2.243 a Predictors: (Constant), F6, F5, F4, F2, F3, F1 b Dependent Variable: HL

Squares df Mean Square F Sig.

Total 365.387 210 a Predictors: (Constant), F6, F5, F4, F2, F3, F1 b Dependent Variable: HL

Table 21 Beta coefficient of the estimated research model

B Std Error Beta Tolerance VIF

Then the regression function is determined as follows:

F1 is graduate quality and generic skill development

F2 is the attractiveness of scientific knowledge

HL is general satisfaction with the course

4.7.2 Testing the violation of hypotheses of OLS

Before testing statistical hypotheses in the regression function, we test the suitability of the model and check if the research model violates the hypotheses of OLS or not to ensure the conclusions The contents in detail are as follows:

- Testing the linear dependence between independent variables : To test the violation of linear dependence between independent variables, we use Scatterplot graph to detect signals of the estimated function if it is appropriate or not The results from Scatterplot graph show that the residuals of independent variables do not distribute according to any rules Thus, the hypothesis of the linear dependence between independent variables does not violate

- Testing the violation of error variance of the residuals in the model : To test the violation of the error variance of residuals, we use Spearman's rank correlation with the corresponding p-value, standard reliability coefficient is 95% The obtained results ofSpearman’s rank analysis are as follows:

** Correlation is significant at the 0.01 level (2-tailed).

From the Spearman table, we can see that there are no variables which have p- value greater than 0.05, so with the reliability coefficient of 95%, we can also conclude that the estimated regression function does not violate the hypothesis that error variance of the residuals does not occur in the estimate model

- Testing the violation of the residual normal distribution: The hypothesis ofOLS method gives out that the residuals must follow the normal distribution rule, if not,the estimate model is not the best model, and its function will be not appropriate To test the residual normal distribution, we use Histogram and P-plot graphs.

From the Histogram graph, we see that the distribution of residuals is the near- normal distribution, the left deviation of average value of residuals is approximately 0, and the normal deviation is approximately 1 In other words, P-plot graph also shows that observed values are quite close to the expected curve, so we can accept the hypothesis supposing that the residual distribution is normal distribution, and there is no violation of the residual normal distribution.

- Testing the multicollinearity : Multicollinearity is a statistical phenomenon in which two or more predictor variables in a multiple regression model are highly correlated In other words, an independent variable can bring information of other independent variables, and it makes the regression function can not exactly predict

(Nguyen Quang Dong, 2003) To test this phenomenon, we use variance inflation factor (VIF) According to Hoang Trong and Chu Nguyen Mong Ngoc (2008), if VIF < 10 we can conclude that the multicollinearity has no impact on regression results Nguyen Quang Dong (2003) supposes that it will be better if VIF dU < d < 4 –dU Therefore, we can conclude that there is no autocorrelation in the research model.

In other hand, F-statistics has p-value =0.000 0 So, with the reliability coefficient 95%, we have not enough basics to reject the hypothesis supposing that the variable F1 positively affects the variable HL In other words, we accept the hypothesis H1.

- Testing the hypothesis H2 : The factor “the attractiveness of scientific knowledge” positively affects general satisfaction with the course This means, Beta coefficient of the variable F2 is positive From actual data, the coressponding t-statistics has p-value

=0.0000 So, we can conclude that with the reliability coefficient = 95%, we have not enough basics to reject the hypothesis supposing that the variable F2 has positive impact on the variable HL In other words, we accept the hypothesis H2.

- Testing the hypothesis H3 : The factor “lecturer team” positively affects general satisfaction with the course This means, Beta coefficient of the variable F3 is positive. From actual data, the coressponding t-statistics has p-value =0.102 > 0.05 Thus, we can conclude that with the reliability coefficient = 95% we have not enough basics to accept the hypothesis supposing that the variable F3 has positive impact on the variable HL In other words, we reject the hypothesis H3

- Testing the hypothesis H4 : The factor “learning resources” positively affects general satisfaction with the course This means Beta coefficient of the variable F4 is positive.From actual data, the coressponding t-statistics has p-value = 0.025 < 0.05, Beta coefficient = 0.118 >0 So, we can conclude that with the reliability coefficient = 95%, we have not enough basics to reject the hypothesis supposing that the variable F4 has positive impact on the variable HL In other words, we accept the hypothesis H4.

The importance of the variables in the model

The final regression function is: HL = -0.494 + 0.292F1 + 0.358F2 + 0.137F4 +0.177 F5 It shows that different variables have different impact level on the dependent variable HL, or in other words, the importance of the independent variables is different.Which variable has greater Beta coefficient will have higher impact on the dependent variable than others Thus, we can see that F2 has the biggest influence on HL, next is F1, F5 and finally is F4.

R-quare coefficient = 0.641 means that the four variables in the model can explain 64.1% of the variability of the dependent variable HL This data is appropriate and acceptable.

Testing the differences between research groups

4.9.1 Testing the differences between groups of different gender

To test the differences between male and female students, we use One Way ANOVA method and get following results:

Table 26 Summary of average values between groups of different gender

Mean Lower Bound Upper Bound

Table 27 Variance analysis by gender

Squares df Mean Square F Sig.

The results show that F-statistics between groups of different gender has p-value

= 0.976 >0.05, so there are no differences between groups of male and female customers on general satisfaction with the course Analyzing the average value between these two groups also shows that the deviation between them is insignificant

4.9.2 Testing the differences between groups of different age

To test the differences between groups of different age , we still use the method of variance analysis and get following results:

Table 28 Summary of average values between groups of different age

Mean Lower Bound Upper Bound

Table 29 Variance analysis by age

Squares df Mean Square F Sig.

We can see here that, F-statistics between groups has p-value = 0.271 >0.05, so we can conclude that there are no meaningful differences between groups of different age on general satisfation with the course Analyzing the average values between these groups also shows that there are no significant deviation between the group under the age of 25 and the group at the age from 26 to 30 (excluding the last groups which have few answers).

4.9.3 Testing the differences between groups of different income

Using the method of variance analysis and get following results and then get results as follows:

Table 30 Summary of average values between groups of different income

95% Confidence Interval for Mean Lower

Table 31 Variance analysis by income

Squares df Mean Square F Sig.

The results show that F-statistics between groups has p-value = 0.411 >0.05, so we can conclude that there are no meaningful differences between groups of different income on general satisfaction with the course In addition, the average values of groups also show an insignificant deviation between groups of different income.

Detections of the research

The research results show that the satisfaction level of students with the course of the center is directly affected by four factors including: (1) graduate quality and generic skill development, (2) the attractiveness of scientific knowledge, (3) learning resources, and (4) objectives and standards.

The results also show that there are no differences between different groups of students The impact level of each factor on general satisfaction with the course is different Below is the summary table of the results and the figure of the impact level of factors in the research model

Table 32 Summary of research hypotheses testing results

H1 The factor “graduate quality and generic skill development” positively affects general satisfaction on the course.

H2 The factor “the attractiveness of scientific knowledge” positively affects general satisfaction on the course.

H3 The factor “lecturer team” positively affects general satisfaction on the course.

H4 The factor “learning resources” positively affects general satisfaction on the course.

H5 The factor “objectives and standards” positively affects general satisfaction on the course.

H6 The factor “learning workload” positively affects general satisfaction on the course.

Figure 4.10.1 The relationship between factors in the research model

Graduate quality and generic skill development

The attractiveness of scientific knowledge

CONCLUSIONS AND RECOMMENDATIONS

Conclusions

The research results answered two questions which are posed in the part of research purposes They are: (1) How is the satisfaction level of students on the course at Hai Duong Job Placement Center?, and (2) How is the impact intensity of factors on student satisfaction with the course?

For the first question, the research results show that the satisfaction level of students on the course is at low average level (the average point >4 in seven point Likert scale) This results is an information signal for the Center to have policies for improvement of service quality in order to increase the satisfaction of the students This information also reflects the results of the center in the customers (students)’ opinions For the second question, the research results show that there are four main factors affecting the satisfaction level of students on the course They include (1) graduate quality and generic skill development, (2) the attractiveness of scientific knowledge, (3) learning resources, and (4) objectives and standards In which, the factor “graduate quality and generic skill development” has biggest influence on student satisfaction, next is the factor “the attractiveness of scientific knowledge” The two left factors:

“learning resources” and “objectives and standards” has lower impact level

The research results obtained from questionnaires also point out that between independent variables and the dependent variable, there is a same-direction-relation, the new formed factors have positive influence on general satisfaction of students on the course Testing the research model shows that the model is reliable and we can generalize the results to the overall Moreover, the research results also show that for short term training courses of the Center, there are no differences between groups of students based on demographic variables such as: gender, age, income.

Recommendations

Basing oneself on the research results, the author gives out some oriented solutions for improving the satisfaction level of student on the courses at Hai Duong Job Placement Center In principles, the solutions which are preferred to implement first are solutions affecting the most influential factors, so the author proposes some oriented solution groups as follows:

1 Solutions for improving “graduate quality and generic skill development”

This is the most important factor for students, so it is needed to improve this factor in advance The Center must identify the purposes of students when taking short term courses are to gain necessary knowledge and get skills for their career The contents of the courses must “practice”-oriented, it means, the knowledge must be practical and can be easily applied and used in working In other hand, the training contents should focus on improving thinking and practice skills of students, help them have solutions for certain situations in their job Especially in current conditions when communication skill, group working skill, presentation skill, etc are preferred and required by employers, the Center needs to train its students soft skills beside general career skills in order to adapt to the rapid changes of the current labor market

2 Solutions for improving the satisfaction level on “the attractiveness of scientific knowledge” This is the second important factor for students To improve this factor, we can apply some solutions such as: Reforming lecture methods: student-centered teaching The contents of learning programs must create interest in acquiring the knowledge for students In addition, the programs must also help students identify the positive motivations when participating in the courses so that the students can actively participate in the subjects and they can see the real value of the courses.

3 Solutions for improving “learning resources” “ Learning resources” is always an important factor for every learning process To improve the satisfaction level of students on this factor, the Center can implement some suggestions as follows:

- Upgrade the library system on the number of books, reference materials and book borrowing and returning procedure to ensure that students always have access to the necessary resources in a most favorable way

- The curriculum of training books and reference materials for students must be simple, concise, clear according to the purpose of "practice" Training materials should avoid putting a lot of knowledge that do not fit the short-term courses.

- Regularly update documentation system and new knowledge for each training area to ensure the gap between training and practice is narrow After the courses at the center, the students can immediately work at the recruiting units but not take too much time to retrain.

4 Solutions for improving “objectives and standards” The outcomes of the course must be informed to students before they enroll or the information on the learning outcomes for each course must be transmitted broadly so that students can access it Lecturers also need to help students identify specific goals when they participate in the courses and to clarify necessary requirements to implement a high quality course that can meet the requirements of students.

Contributions and the importance of the research

The purpose of this study is to increase understanding of factors affecting the satisfaction of students on the short term courses at the Center Within the scope of this study, the author thinks that the research proved the reliability of the CEQ in the research conditions of Vietnam Through answering questionnaires, the research contributed to the field of research on customer’s motives and behaviors (students) with a certain service

5.3.2 The importance of the research

- In the dormain of learning, the research tested the reliability of the CEQ from Australia education in other new cultural conditions of a developing country like Vietnam This will be one of the initial researches using CEQ in the field of short term educational training and the basic for etablishing furthur researches in education sector

- In the dormain of reality, the research pointed out factors and their impact intensity on the satisfaction of students They are information providing to the finding of solutions for improving the satisfaction of students with the courses, as well as attract more and more students to the training programs of the Center.

Limitations of the research

Like other studies, this research also has certain limitations

First, the research is limited in the scope of Hai Duong Job Placement Center, so it limited its representative and there is no basis to compare the differences between different units which are competing with each other for each indicator.

Second, the R-square coefficient = 0.641 shows that we can add more other factors into the research model to better improve its ability of explanation Therefore,the research might omit other factors affecting the satisfaction of students or did not detect more new factors that also may influence the student’s satisfaction with the course beside the factors in CEQ.

Third, due to the limitations of time and cost, the study has still not made the

“deep interview” with customers after the quantitative research to detect more inside information to find inside-information of students and build more specific solutions Fourth, the study uses convenience sampling method, so it does not assess all risks resulting from system errors, can lead to the uncertainties in the research model and general limitations of the study.

Directions for furthur researches

From above limitations of the research, the author proposes some directions for furthur researches which will be done in the future.

Firstly, to generlize result to the overall, furthur researches should expand the research scale to increase the reliability of the research results and continuously adjust the research scales They can also use the probability sampling method to increase the representative and generality of the research model

Secondly, this study was done in a certain time, so its ability of explanation in the long run is limited Thus, futhur researches should be conducted in many different times and analyze the research model in the long-term in order to assess how the measures impact the factors in the model.

Thirdly, furthur researches can add other important factors into the research model to better increase the ability of explanation of the model and to match the new research conditions.

Fourthly, expand the research objects to customers of different service providers to be able to explain the trendency of all market and have a basis for comparison of the customer satisfaction level with each unit

Craig Mclnnis et al (2000), Development of Course Experience Questionnaire, The University of Melbourne, Australia.

Calvin Smith and Debra Bath (2006), The role of the learning community in the development of discipline knowledge and generic graduate outcomes, Higher

Christian Grửnroos (1984), A Service Quality Model and its Marketing Implications, European Journal of Marketing, 18(4), 36 – 44.

Course Experience Questionnaire (2011), Course Experience Questionnaire (CEQ) report, Course Experience Questionnaire, James Cook University, 1-96.

David P, Curtis and John P.Keeves (2000), The Course Experience Questionnaire Instittutional performance indicator, International Education Journal,1(2), 73 -82. Francisco Cano and A.B.G Berben (2009), University students` achievement goals and approach to learning in mathematics, Bristish Journal of Education

Garson GD (2002), Guide to Writing Empiricel Pappers, Theses, and

Dissertations, New York: Marcel Dekker.

Gerbing WD and Anderson JC (1988), An udate paradigm for scale development incorporating unidimensionality and its assessments, Journal of Marketing Research, 25(2), 186 -192.

Hair et al (1998), Mutivariate data analysis, 5 th ed, Englewood Cliffs, NJ, Prentical – Hall.

Hair et al (2006), Mutilvariate Data Analysis 6 th ed, Upper Saddle River NJ,Prentice –Hall.

Hans NE Bysstrom (2004), Teaching evaluations at the introductory finance course at Lund University: A comparison of the Course Experience Questionnaire an a traditional evaluation approach, Lund University, Sweden.

Hishamuddin Fitri Abu Hasan et al (2008), Service quality and satisfaction: A case study at Private higher education institutions, International business research, 1(3), 163 – 175.

Hoang Trong and Chu Nguyen Mong Ngoc (2008), Analysis of research data with SPSS – Vol 1, Hong Duc Publisher, Ho Chi Minh, Vietnam.

Hoang Trong and Chu Nguyen Mong Ngoc (2008), Analysis of research data with SPSS – Vol 2, Hong Duc Publisher, Ho Chi Minh, Vietnam.

Jan F.H Nijhuis et al (2005), Influence of redesigning a learning environment on student perceptions and learning strategies, Learning Environments Research, 8, 67 -93. Junaida Ismail et al (2011), The perpormence of Academic Libraries: A case study at research university (Rus) in Malaysia, Global of human Social Sciene, 4(8), 19-23

Kano el al (1984), Attractive qulity and must be quality, Hinshitsu, 14, 39 - 48.

Kotler and Armstrong (2004), Principles of Marketing (Vietnamese translation version), Statistics Publisher, Vietnam.

MacCallum et al (1999), Sample size in factor analysis, Psychological Methods,

Nifarta Peingurta Andrew (2010), Applicability of Student Course Experience Questionnaire (SCEQ) in an African context: The case of Nigernian Universities,

Literacy Information and Comouter Education Journal, 1(3), 143 -150.

Nina Clemson (2009), 2009 Course Experience Questionnaire (CEQ)* results, http://www.csu.edu.au/division/plandev/survey/ceq/index.htm.

Nunally and Bernstein (1994), Psychometric Theory, 3th ed, Mc Graw – Hill, New York.

Nguyen Dinh Phan et al (2005), curriculum "Quality management in organizations”, Labor and Social Publisher, Vietnam.

Nguyen Dinh Tho et al (2007), Survey of student satisfactions with the Master training quality at the Ho Chi Minh University of Economics, Ho Chi Minh university of economics, Vietnam

Nguyen Dinh Tho (2011), The scientific method in business research, Labour – Social Publisher, Vietnam.

Nguyen Quang Dong (2003), Econometric lecture, Statistics Publisher, Hanoi,Vietnam.

Nguyen Thi Phuong Tram (2008), Master thesis, E-banking quality: Compare the

SERVQUAL and Gronroos model, Ho Chi Minh unversity of economics, Vietnam.

Patrick Griffin et alt (2003), The development of an Course Experience Questionnaire, Quality in Higher Education, 6(3), 259 -266.

Parasuraman, A V A Zeithaml, & L.L.Berry (1985): A conceptual Model of Service Quality and its Implications for Future Research, Jounal of Marketing,49, 41-50. Parasuraman A., Zeithaml V & Berry L., (1988), Servqual: A Multiple-Item Scale for Measuring Consumer Perceptions of Service Quality, Journal of Retailing, 64(1), 22 – 37.

Pham Duc Ky (2007), Research on the Impact of Mobile Communications

Service Quality for Customer Loyalty in Ho Chi Minh, Master’s thesis, HCMC

Ramsden (1991), A performance indicator of tearching quality in higher education: The Course Experience Questionnaire, Studies in Higher Education, 16(2),

Syed Saad Andaleeb and Patience L Simmonds (1998), Explaining user satisfaction with academy libraries : strategic implications, College and research

Sri R Chandra Shekhar el al (2011), Enhancing the quanlity of engineering education instittutions (EEIs) through analysis, International quality conference, 5, 769-779.