Chapter IV

RESEARCH METHODOLOGY

Target Population

The population of this study is the entrepreneurs in Bangkok employing Electronic Data Interchange (EDI) as a tool to send and receive on-line document among themselves as well as with the Customs Department and the Port Authority of Thailand.  There are several classifications that can categorize the population.  However, TIFFA EDI services Co., Ltd., one of Value Added Network (VAN) service providers, classifies four subgroups, that is:

Group of Transportation consists of freight forwarders, customs brokers, shipping agencies, importers, exporters, maritime, shipping liners, containers, and trailers.  This group is the target group of the EDI policy since it is directly concerned with international trading.  Thus, this study focuses mainly on Group of Transportation.

Groups of Retailer, Supplier, and Health Care are not the significant players at this moment.  They are just supplementary groups being studied for the local EDI awareness.

Sampling Procedures and Results

The sampling scheme takes two stages.  It employs probability sampling.  Since there are several sizes and types of population, stratified sampling will be used in the first stage and take a sample from four groups according to the TIFFA classification.  At the desired 90 percent confidence level with an error not exceeding plus/minus 10 percent, the required sample size for a population of more than 50 cases is 33 cases.

Based on the approximately transportation : 0.2supplier : 1.4retailer : 0.08 health care group ratio, it is expected that the total sample for transportation should be randomly selected at least 33 cases; for retailer at least 7 cases; for supplier at least 46 cases; and for health care at least 3 cases.  However, as shown in Table 4.1, groups of supplier sample and health care are less drawn from the above expectation due to the incompleteness or unsystematic responses to the questionnaires.  Besides, there are not many organizations involving EDI at this stage.  For example, there are only 4 health care organizations.  However, this will not affect the result of the study since these subgroups are not the focus-group.

Table 4.1  Number of Samples Actually Drawn from Population for Each Category

Research Design

This study employs both qualitative and quantitative methods since qualitative methods permit the evaluator to study selected issues in depth and detail of small number of people and cases while quantitative methods measure the reactions of a great many people to a limited set of questions with the use of standard measures (Patton, 1990: 13-14).

Pilot Test

It is necessary to conduct a pretest for a small sample of international trading public organization staff and trading partners.  Pretesting is helpful to discover the respondents’ reactions to the questions and to construct meaningful survey questions.  This analysis collected 30 questionnaires from the population.

Data Collection Instruments

The data collection consists of both primary data and secondary data.  The primary data are guided by the secondary data.  The details are as follows:

Primary Data

Primary data used in this study came from two sources: semi-structured interviews with representatives of state agency and trading partner and a questionnaire conducted with the sample of entrepreneurs using EDI.  The workflow of existing process (EDI) and previous process (non-EDI) will be collected by interview.  The semi-structured interview of key persons of concerned functions will be used to collect data in a large volume in order to prepare questionnaires for measurement.  After constructing and distributing questionnaires for pilot study, questionnaires and results will be revised in order to adjust for better measurement of the variables involved in the analysis.

Secondary Data

This data comes from documentary EDI of concerned organizations, EDI journals & textbooks, and Internet.  The secondary data combined with qualitative analysis from primary data will be used as the basis for questionnaire design, pretest, questionnaire redesign and distribution, and variable measurement for quantitative analysis.

Key-Staff Interview

For the workflow of the process, semi-structured interview will be used to interview key persons in international trade organizations.

Port Authority of Thailand

PAT began to use EDI with Inward Container List to facilitate shipping companies, freight forwarder and etc.  They made an announcement that trading partners should start by September 1, 1996.  At the early stage, users may select to transfer data to PAT by using diskettes or by using EDI via mailbox which connected through CAT/IIN network.  For companies who were not ready may transfer data in the form of paper document according to the old format.  Unfortunately, PAT would charge for recording service for 10 Baht. per container and they had to submit Inward Container List 24 hours prior arrival time except for ships from Singapore, Malaysia, Cambodia, Vietnam, Brunei, Japan, Korea, Hong Kong, Taiwan, and Philippines.  These groups must submit no less than 12 hours.  For EDI users, they may submit Inward Container List no less than 6 hours prior arrival time.  If customers cannot follow the mentioned rule on time, they must be fined for 5,000 Baht. per ship.  Moreover, PAT would not continue processing until the data part had been completed.  If customers submit the incorrect data, they must inform to cancel or correct the previous transaction no less than three hours prior arrival time.  Otherwise, they will be fined for 500 Baht. per container and , PAT would not continue processing until the data part had been completed unless this would not bother surcharge.

The Customs Department

The workflow of non-EDI and EDI are demonstrated as follows:

Non-EDI System

1.      At exporter office, exporter fills out the manifest and brings that form with documents to The Customs Department.

2.      At The Customs Department, the officer will generate the clearance number to the exporter and record to the system.

3.      The officer checks the data and merchandise value listed on the manifest.

4.      The officer orders the responsible function to be ready to check the merchandise.

5.      The exporter brings the manifest form and all documents to registrar office.

6.      The registrar office assigns the officer to check the merchandise.

7.      The checking procedure begins.

EDI System

1.      At exporter office, the exporter enters invoice and manifest data into its own computer and sends electronically to The Customs Department.  All data will automatically be sent to The Customs Department’s computer via communication line.

2.      The Customs Department’s computer will check the data.  If there are no errors, the computer will automatically generate the manifest number and initiate the checking merchandise procedure.  Otherwise, the computer will inform the exporter to retransmit the correct data.  All of these are performed via communication line.

3.      The exporter prints out the manifest.  If there is no custom clearance concerned (Green line), the exporter brings all documents and goes to merchandise checking division.  Otherwise, the exporter goes to custom clearance division.

4.      The Customs Department checks the exporter’s merchandise.  The red line will check 100 percent of merchandise.  The green line will check 30 percent of merchandise.

In sum, importers, exporters, and custom brokers prepare data, invoice, and Customs declaration and send via EDI system using UN/EDIFACT standard through TradeSiam, the official EDI service provider.  If data is correct, the system will automatically generate Customs declaration number.  Then, custom brokers will print Customs declaration and present it to Customs office in order to pay for custom duty and to have the customs check.

For gold card members, their documents and merchandise may not need to be verified except for random checking.

For chain of command, it is certain that state agencies are strongly bureaucracy especially the Customs Department.  This clumsy system causes lack of good cooperation between public and private organizations resulting in less progress in EDI success than it should be.  They criticized further that state agencies which should be the EDI leader lack of readiness and sincerity.  The complaints from trading partners happened right after the Customs Department excessively promoted EDI.

For fairness of the process, it is evident that trading partners e.g. transportation group still have to pay for smoothen procedures at Port Authority of Thailand.  However, they confirmed that EDI is the appropriate technology which we should continue with more collaboration between organizations.

Most of trading partners complained about initial investment which is considered to be high especially for small companies.  Besides this, the operational cost including VAN service provider fee is also high.  They said that this was because state agencies are stick with the only one VAN service provider called Tradesiam which charges too high.

Questionnaire

In developing a questionnaire, earlier researches on EDI in general, EDI in other countries, EDI in organizations, and etc. have formed the contents of questionnaire.  The finalized questionnaire is presented in Appendix A.

The questionnaire consists of twelve parts.  Those are biodata, EDI satisfaction domains, namely, policy satisfaction, inter-organization cooperation, intra-organization cooperation, and individual support; EDI outcomes; EDI impacts; EDI obstructs; EDI success factors; EDI success factors from state agencies; and overall EDI satisfaction.  The details of each part are explained as follows:

Analysis

Variable Measurements

Variables in this research are developed based on underlying concepts in the literature reviewed.  The indexes have been used to measure these variables.  Each set of domain assessments contains multiple dimensions or qualities.  The construct validity and reliability of these measurement scales are employed to certify appropriateness.

The reliability of each domain toward EDI satisfaction will employ a Cronbach’s coefficient as an indicator.  The reliability of individual items can be examined using the alpha value of the summative scale when any of the individual items is excluded.  A higher value of the coefficient if the item was deleted indicates that a lower internal consistency would result if the item was included in the scale.

In examining the construct validity of a measurement scale, factor analysis with principal component extraction and varimax rotation is employed to identify the underlying structure of the concept being measured as well as to determine the weight of individual variables included in the scale.  The technique is also used to identify a smaller number of factors representing relationships among sets of many interrelated variables, thus helping to reduce the number of variables in the analysis so that a more meaningful interpretation is possible.

The details of the measurement scales of each domain are described as follows:

Overall Satisfaction Index After Implementing EDI

Overall satisfaction is measured by a composite index which is a summative rating scale of eleven differential scales ranging from completely dissatisfied (0 point) to completely satisfied (10 points).  The respondents are asked to express their feeling about EDI as a whole.  This includes:

The scale is used to measure each of domain assessments as well.  The reliability of the summative scale is quite high as indicated by a Cronbach’s alpha value of .7702 (Table 4.2).

Besides the Overall Satisfaction Index reliability measured by Cronbach’s alpha, domain assessments consisting of Policy Support Assessment (PSA), Inter-organization Cooperation Assessment (SSA), Intra-organization Cooperation Assessment (OSA), Individual Support Assessment (ISA), Output Assessment (OA), Impact Assessment (IA), Obstacle Assessment (OA), General Factors Assessment (GFA), and Factors from State Agency Assessment (FSAA), have been measured.

The calculated Cronbach’s alpha values for domain assessments are also shown in Table 4.2 and ranging from .6511 to .9503, which is very reasonable.

Table 4.2  Reliability Coefficients of Measurement Scales

Policy Support Assessment (PSA)

There are several dimensions to construct policy support assessments.  Each question is about the level of agreement of EDI policy support.  This also asked respondents about the suitability of EDI in this situation and covered laws and regulations questions.  It is known that legal solicitors generally agree that existing legislation and juristic method will apply to electronic commerce.  At present there is neither concrete legislation drafted to govern on-line contracts nor has there been any case raised before a Thai court on issues relating to offer and acceptance made over the Internet or EDI.  At the end of the questions, the future or trend of EDI has been asked.

There are thirteen items constructed to measure policy support assessments.  Each item is measured on a zero-to-ten scale.  The reliability of the summative scale of all the thirteen items is satisfactory as indicated by a Cronbach’s alpha value of .7570.  A factor analysis using principal component extraction and varimax rotation of all items is undertaken.  It yields four factors.  The measure is formed by five items of EDI Policy Implementation Factor (PI), four items of EDI Policy Factor (POLICY), three items of EDI Law Factor (LAW) and one item of Future of EDI Factor (FUTURE) (Table 4.3).

Table 4.3  Factor Analysis of Policy Support Assessments

  Inter-organization Cooperation Assessment (ICA)

There are seventeen items constructed to measure of inter-organization cooperation assessments.  Each item is measured on a zero-to-ten scale.  The reliability of the summative scale is satisfactory as indicated by a Cronbach’s alpha value of .8074.  A factor analysis of all items is undertaken.

This asked respondents about the support from each sector.  It also covers data communication network.  This is the important component in implementing EDI because it includes security system. 

The measure of this domain assessment is formed by five items of Good Cooperation Among Organizations (COOR), five items for Lack of Confidence with Partners (LCON), two items for Resource and Budget (RESBUD), three items for VAN and EDI Software (VANSOFT), and two items for Private Sector Adjust and Support (PRISUP) (Table 4.4).

Table 4.4  Factor Analysis of Cooperation Among Organizations Assessments

Intra-organizational Support Assessment (OSA)

There are nine items constructed to measure support within organization.  Each item is measured on a zero to ten scale. Literature review on technological change indicates that it affects organizational change (Castells, 1996: 169).  One of the most direct expressions of systemic change is the transformation of employment and occupational structure, which have been found effects on EDI.  It is very interesting to study organization change resulting in EDI organizational support.

This asked respondents about their organization culture.  Literature on EDI implementation mostly focuses on top management’s support unconditionally.  According to Data Interchange Plc., it is evident that it is essential for top management to strongly support for such a project is gained.  The perception of management in implementing is highly important.  This is because the implementation of EDI crosses so many departmental boundaries.  As a result, it requires some form of central control and backing.  It is also necessary to counter the initial difficulties and low pay back resulting from implementation

(www.dip.co.uk/articles/implemen/imple4.htm). 

For successful implementation it will be necessary to pick a project team from a wide range of departments.  This is so that all areas which will be affected will have a “local expert” to deal with it.  Purely from a psychological point of view this also means that those who will be affected will have felt that they have at least had a say in the change-over in their own backyards.  Better this than an enforced change from the outside, which may not have immediate justifications.  To bring the whole team together it will be necessary to have an enthusiastic and appropriately skilled supporter of EDI as leader.  This individual must be able to understand the needs of the business as much as having an understanding of EDI itself.

Besides, it is essential to educate new users so that the change-over can be conducted as smoothly and effectively as possible.  Information and training can be obtained.  Therefore, support will take place at three levels: top management, project implementation team, and users.  Also, on-time performance must be concerned.

The reliability of the summative scale is satisfactory as indicated by a Cronbach’s alpha value of .6511.  A factor analysis of all items is undertaken.

The measure is formed by five items of EDI Team and Management Support (EDITEAM) and one item of Cross Functionality (CROSSFN) (Table 4.5).

Table 4.5  Factor Analysis of Intra-organization Support Assessments

Individual Support Assessment (ISA)

There are twenty-one items constructed to measure of how each stakeholder support EDI.  Each item is measured on a zero-to-ten scale.  The questions cover job satisfaction, job dissatisfaction, computer-related skills, individual’s attitude toward EDI etc.  The following characteristics have been hypothesized to have effects on their EDI effectiveness.  They are working hours, technology-related work, level of work difficulty, system maintenance, and job satisfaction.  The questions comprise of alertness, stress, job appreciation, opportunity for receiving technological knowledge, job responsiveness, and ability to simultaneously process jobs etc.  Each question is about the level of agreement of EDI individual support.

Table 4.6  Factor Analysis of Individual Support Assessments

The reliability of the summative scale is satisfactory as indicated by a Cronbach’s alpha value of .6857.  A factor analysis of all items is undertaken.

The measure is formed by six items of Job Satisfaction (JOBSAT), six items of Computer-Related (COMREL), three items of Unsecured Feeling (UNSFEEL), two items of Job Dissatisfaction (JOBDIS), one item of Secured Feeling (SECFEEL), and one item of Individual Support (INDSUP) (Table 4.6).

Outcome Evaluation Assessment (OA)

There are eighteen items constructed to measure of outcome evaluation.  Each item is measured on a zero to ten scale and mainly focuses on EDI benefits and so on..  Factor analysis was performed.  Four factors were extracted and labeled as tangible benefits, intangible benefits, EDI data accessibility, and data analysis (Table 4.7).

It is clear for tangible benefit factor that it consists of desired outcome as stated in EDI policy e.g. a reduction in document loss, a reduction in procedures redundancy, etc.  Intangible benefit factor includes the transparent process, the process fairness, etc.  EDI data accessibility factor contains the speed in access the central data, the distribution of central data, etc.  Data analysis factor contains central data utilization, data analysis, etc. 

The reliability of the summative scale is highly satisfactory as indicated by a Cronbach’s alpha value of .9503.  A factor analysis of all items is undertaken.

Table 4.7  Factor Analysis of Outcome Evaluation Assessments

  Impact Evaluation Assessment (IA)

There are twenty-three items constructed to measure of impact evaluation.  Each item is measured on a zero-to-ten scale.  The questions cover multi-dimensions.  Besides review of literature, stakeholders took part in describing the impact of EDI on themselves and others.  As a result, the reliability of the summative scale is very satisfactory as indicated by a Cronbach’s alpha value of .9427.  A factor analysis using principal component extraction and varimax rotation of all items is undertaken showing that the measure is formed by five items of Direct Value of EDI Factor (VALUE), Organization Concerned Factor (ORG), Staff Concerned Factor (STAFF), Initial Cost and Investment Factor (INVEST), and Future Trend Factor (TREND) (Table 4.8).

Direct Value of EDI Factor means when EDI existed, it has been valuable supporting from operation work to international trading.  Organization Concerned Factor means the impacts of EDI to organization e.g. common data utilization within organization, inventory control, information searching within organization, intra-organization communication etc.  Staff Concerned Factor means impacts to staff in the organization e.g. knowledge on technology gaining, Initial Cost and Investment Factor means the impact to the organization in terms of financial and accounting issues.  The impact to the organization in the long run is the last factor called Future Trend Factor.  This is aimed toward the goal of EDI to increase potential in international competition.  It may not be easy to measure an export issue in term of international competition directly.  However, EDI can be a tool to support export scenario.

It is evident that EDI may imply changes not only in skills and organization structure but also in business strategy (Cash & Konsynski, 1985).  The impact of EDI can thus ripple through the entire organization.  The factor analysis done in this study reveals the compatibility with literature reviewed.  The impacts spread through organization and staff.  Barber (1991) classifies the effects of EDI in four major categories which are:

1.      People (skills, leadership style),

2.      Management practices (tasks, operating procedures),

3.      Support structures (information systems, organizational structure), and

4.      Corporate culture (shared beliefs and assumptions).

Holland, Lockett & Blackman (1992) describe the internal effects of EDI on organizational structure and process (individual roles and tasks, size and productivity of staff, new organizational units).  MIT’s “Management in the 1990s” impacts along the following areas: management, technology, strategy, structure, individuals and roles.

Finally, the use of EDI should lead to the desired organizational changes which, ultimately, should improve business performance.

Table 4.8  Factor Analysis of Impact Evaluation Assessments

Obstacle Assessment (OBA)

There are thirty items constructed to measure of obstacle assessment.  Each item contributed from the pretest given advice directly by stakeholders is measured on a zero-to-ten scale.  The reliability of the summative scale is satisfactory as indicated by a Cronbach’s alpha value of .8187.  A factor analysis of all items is undertaken.

A measure is formed into eight factors as shown in Table 4.9.  Organization Connection Factor (ORGCON) consisting of 9 items means cross-organization constraints e.g. government organizations, VAN provider organizations, and other concerned organizations.  This constraint mostly focuses support and cooperation among organizations.  The second factor is EDI Characteristics Factor (EDICHAR).  It involves EDI itself including steps of working, the whole cycle EDI expenditure, and lacking of precise conclusions.  EDI Users and Budget Factor (USERBUD) contains four items.  The constraints are from EDI users who have no skill or have less skill.  They may not have understanding about EDI knowledge and concepts.  In conclusion, they lack of readiness.  When they do not have a clear picture of EDI, they may resist it.  The fourth factor is, therefore, Staff’s Resistance Factor (RESIST).  Individual staff or the whole units of organization may resist and do not accept EDI.  The fifth factor is Management Concerned Factor (MGT).  This is the very important factor.  All EDI literatures reviewed state the same things about management support sincerely.  It also means investment and value of EDI.  Other factors are Initial Investment Factor (INIT), Law Factor (LAW) which no law to support EDI for now, and Government Agencies Lack of EDI Knowledge Factor (KNOWLEDGE).

Table 4.9  Factor Analysis of Obstacle Assessments