User Acceptance of a Clinical Trial Materials Management System

Thesis Technology Management Faculty of Management & Organization

University of Groningen

User Acceptance of a Clinical Trial Materials Management System

An analysis of different theoretical models in IS research

Ing. N.M. Sterrenburg

Leiderdorp, July 2005

.2 .

Thesis Technology Management Faculty of Management & Organization

University of Groningen

User Acceptance of a Clinical Trial Materials Management System

An analysis of different theoretical models in IS research

Author: ing. N.M. Sterrenburg

Faculty supervision: dr. W.M.C. van Wezel dr. A. Boonstra

Organization of study: Astellas Europe BV || Yamanouchi Europe BV

Pharmaceutical Development Department

Clinical Trial Materials Section

drs. J. Hilhorst

Date: July 2005

Location: Leiderdorp

.3 . Preface

To finalize my study period, I started my last journey in September 2004. My chemical background inspired me to take a look in the normally non-transparent pharmaceutical business. Astellas former Yamanouchi is one of the few companies performing pharmaceutical research and development work in the Netherlands, and for that reason an interesting company. While, the pharmaceutical business was new for me it was also the first time a business student was doing an internship at this department.

The first two months I learned a lot about the pharmaceutical business, project management and the Japanese company. This was a great experience, but gradually, it was time to focus more on my research and to finalize my academic study. As the assignment was vague in the beginning I was in the position of choosing my own direction for my thesis. A pilot phase of materials management system was initialized and some support was needed from a technology business point of view to the project management.

The suppliers and some important users performed a fit-gap analysis of the operating process. After considering different options I decided to focus more on the acceptance of the system by the users. I believe this was the appropriate choice, to combine theory and practice. Of course, there were moments, especially during the statistical part, that I was pessimistic about the progress. This was a good experience to get to know more about my own good and bad points.

My gratitude goes to all the PDD employees for taking me on a great ten-month journey called Yamanouchi Europe. With special thoughts I will go back to all the personal moments, the laughs and the great conversations I had with all of you. I want to wish you all the luck with your careers. Maybe sometimes you think about that business girl with all the books and big papers. After reading this thesis you finally know what she was doing all day.

My appreciation goes especially to drs. José Hilhorst for taking the time and effort for listening to me and talking with me about her management vision. It was also José who encouraged me to go apply for a consultancy function. This turned out very successfully and I will start as an analyst at Accenture the first of July. I will definitely use a lot of things I learned here on the following journeys I will take.

Next to that, I would like to thank the supervisors, dr. W.M.C. van Wezel and dr. A. Boonstra, from the university of Groningen for taking the time and using their expertise in guiding me on this trip. Maybe, this thesis is a little bit different than the ones you are used to. I tried to make a comprehensive overview of the theory and reflecting this to the practice with barging the scientific boundaries in mind.

Finally, I want to thank all my friends and family for supporting me in writing this thesis. Thanks for listening to my complaining and for reviewing the parts.

Enjoy, reading it!

Nienke Maartje Sterrenburg

.4 . Table of Contents

ABSTRACT

INTRODUCTION...8

1.1 INTRODUCTION...8

1.2 COMPANY DESCRIPTION...8

1.2.1 Production of clinical trial...9

1.3 PROBLEM AREA... 10

1.4 OBJECTIVE OF THIS RESEARCH... 10

1.4.1 Used theoretical models... 10

1.4.2 Purpose of this research ... 11

1.4.3 Methodology steps ... 12

THEORETICAL CONTEXT... 13

2.1 INTRODUCTION... 13

2.2 THE THEORY OF REASONED ACTION... 13

2.2.1 Behavioral intention (BI)... 13

2.2.2 Behavior (B) ... 14

2.2.3 Attitude (A)... 14

2.2.4 Subjective norm (SN) ... 16

2.3 TECHNOLOGY ACCEPTANCE MODEL... 16

2.3.1 Ease of use (EOU) ... 17

2.3.2 Usefulness (U) ... 17

2.3.3 External variables (EV) ... 18

2.4 THEORY OF PLANNED BEHAVIOR... 19

2.4.1 Perceived behavioral control (PBC) ... 20

2.5 TECHNOLOGY ACCEPTANCE MODEL 2... 21

2.5.1 Social influence... 21

2.5.2 Cognitive instrumental process... 22

2.6 TASK-TECHNOLOGY FIT... 23

2.6.1 Task-technology Fit (TTF) ... 24

2.6.2 Tool functionality (TF)... 24

2.6.3 Task characteristics (TC) ... 25

2.6.4 Tool experience (TE) ... 25

2.7 GENDER AND AGE... 26

2.7.1 Gender (G)... 26

2.7.2 Age (A)... 26

2.8 EXPERIENCE... 27

2.9 OTHER RELATIONS... 27

DATA ANALYSIS ... 29

3.1 INTRODUCTION... 29

3.2 SAMPLE AND PROCEDURE SETUP... 29

.5 .

3.2.1 Data collection... 29

3.2.2 Questionnaire... 30

3.2.3 Statistical approach... 30

3.3 QUANTITATIVE RESULTS... 30

3.3.1 Theory of reasoned action ... 30

3.3.2 Technology Acceptance Model... 30

3.3.3 Theory of planned behavior ... 31

3.3.4 Task-technology Fit ... 31

3.3.5 Technology acceptance model 2 ... 31

3.3.6 Age... 31

3.3.7 Gender ... 32

REVIEW OF RESULTS ... 33

4.1 INTRODUCTION... 33

4.2 PSYCHOLOGICAL... 33

4.3 INSTRUMENTAL... 34

4.4 CONTROLLABLE... 34

4.5 TASK... 36

4.6 COGNITIVE... 37

4.7 GENDER AND AGE... 37

4.8 EXTERNAL VARIABLES... 38

4.9 PRIORITIZING THE HYPOTHESES... 39

CONCLUSION ... 41

A FINAL MODEL... 43

LIMITATIONS... 44 LITERATURE REVIEWED

APPENDIX 1 PMX CTM FUNCTIONS SETS

APPENDIX 2 RELATED USER ACCEPTANCE MODELS APPENDIX 3 BELIEFS STRUCTURE AND DECOMPOSITION APPENDIX 4 QUESTIONNAIRE

APPENDIX 5 STRUCTURAL EQAUTION MODELLING APPENDIX 6 QUANTITATIVE DATA

.6 . Abstract

Astellas former Yamanouchi, a Japanese pharmaceutical company, has its main therapeutic areas in urology, transplantation and dermatology. In order to finalize the optimization of the clinical trial materials supply chain a new management system is implemented. The PMX system gives the ability to management the master batch record, has the ability of good warehousing and materials management, and weighing according to pharmaceutical operating standards. Before a full implementation a pilot test is conducted that determines an operating fit-gap analysis.

This pilot test also gave the ability to test the user acceptance. Literature addresses several models, which have the capability of testing this user acceptance. The purpose is to investigate factors that are critical to get the best user acceptance. A choice was made to test this user acceptance through the use of intention models. These models give the ability to predict and determine a certain behavior.

The Theory of Reasoned Action tries to determine the behavioral intention by investigating a person’s beliefs by investigating the attitude and subjective norm. This model defined by Ajzen and Fishbein is one of the most popular expectancy-value models.

Davis adopted of this the Technology Acceptance Model that tries to explain the determinants of computer acceptance. To provide a basis for tracing external factors on internal beliefs, attitudes and intentions. The prime aspects are perceived usefulness and perceived ease of use.

Another adoption of the first theory is the Theory of Planned Behavior by Ajzen. This accounts for conditions of variable control, by introducing the perceived behavioral control next to the attitude and subjective norm. This perceived behavioral control can defined as an internal control of self-efficacy and an external controllability.

The Technology Acceptance Model 2 incorporates additional constructs in its predecessor; social influences processes (subjective norm, voluntariness and image), and cognitive instrumental processes (job relevance, output quality and results demonstrability.

The last model addressed is developed by Goodhue. The Task-Technology fit tries to determine the optimal fit between a technology and a task. This is integrated with the Technology Acceptance Model to provide a stronger model. To determine this fit the task characteristics, the tool experience and tool functionality are investigated.

Finally the differences between the age and gender groups are researched.

A questionnaire was submitted among the different users of the system. The used statistical approach is Structural Equation Modeling. This has the purpose to fit a proposed model with empirical evidence to get an applicable model. Due to the smallness of the group, only 15 users, it was hard to completely validate the statistical approach. A qualitative review was, therefore, also applied.

The results show that beliefs, either through attitude or subjective norm, have an influence on the usage. The biggest influence is, however, the usefulness of a system, what is partial related to the ease of use. These are also determined by several external variables such as the design of the system and communication of the project management. There is a high relation between the internal self- efficacy and external controllability that all have an influence on the behavioral intention. The task structure contributes to the task-technology fit, lacking is the tool functionality and tool experience. A final model reflects the issues that were found to be important.

.7 .

I ntroduction

1.1

I

This research is located in the pharmaceutical company of Astellas. This Japanese pharmaceutical company is not that familiar in the business administration field. As one of the top 20 pharmaceutical companies in the world, it has a high focus on research and development to get sustainable growth.

As part of the developing cycles, material is made for clinical research. For optimization of the operating process of the clinical trial materials (CTM) a materials management system is acquired. To get an optimal implementation and user acceptance is the focus of this research.

1.2

C

Astellas, founded on 1 April 2005 of a merger between Fujisawa and Yamanouchi, is a global pharmaceutical company with headquarters in Japan. In Leiderdorp, the European research and development department is situated. Meppel locates one of five European plants. Furthermore each European country has a Sales & Marketing affiliate. The focus is on the main therapeutic areas of:

• Urology – Astellas markets and develops products for the treatment of among others the overactive bladder, prostate cancer and incontinence problems.

• Transplantation – These are very specialist products, manufactured by Fujisawa, to prevent organ rejection in kidney and liver transplants. In the future products will be further developed.

• Dermatology - Astellas brings together former Yamanouchi and Fujisawa products for the treatment of atopic dermatitis (eczema), providing a range of prescribing options for the physician, as well as products for acne and other skin conditions.

In addition, Astellas has a number of trusted and effective products in other therapeutic categories, which are marketed in selected countries across Europe. These include products for conditions in the following therapy areas: Cardiovascular, Respiratory, Gastro-intestinal and CNS, as well as Antibiotics.

The Pharmaceutical Development Department (PDD) is located in Meppel and Leiderdorp. Former Yamanouchi’s R&D department it is now situated in the new Technology, Supply, Chain and Manufacturing

(TSCM) of Astellas. It serves as an intermediate between both departments. The main task is production of clinical trial materials (CTM), figure 1. This kind of medicine is used for pre-registration and

1

Figure 1 The drug life cycle

.8 .

development work of (new) products. Related activities support this production: R&D support, quality assurance and project management. Further more pharmaceutical studies as: formulation, preliminary stability, scale-up and introduction for production are conducted.

Figure 1 The drug life cycle, shows the supply chain of a pharmaceutical product. PDD main activities are in the first three phases; R&D, Clinical Trials and Scale Up with supporting areas of maintenance and quality assurance of these phases.

1.2.1 Production of clinical trial

The production of clinical trial material consists of several phases. Figure 2 demonstrates the supply chain of the clinical trial material.

Figure 2 The CTM supply chain

The raw materials are purchased from a chemical plant that performs the chemical syntheses, in Japan.

The raw materials are granulated and tablets are stamped or capsulated. The blistering process takes place by putting the tablets into a strip.

Before shipment the material is packed in a center-box and the material labeled. The labeling takes place on basis of a randomization number. This is a unique number and indicates which patient gets what kind of medicine. Most studies are based a ‘double blind’ principle where the patient gets the product under development, a competing product and a placebo.

The last phase is the returning of the used medication. This is to test whether the patient has taken all the medication and leftovers are destructed.

The production of CTM is highly restricted to legislation and quality limitations. All production is based on Good Manufacturing Practice (GMP) that is a European standard for the production of (good) pharmaceutical products. The last few years the operation standards in the production environment were designed to be on a higher level. The last five years a big supply chain project was initialized. The purpose of this project was to rearrange the CTM process so that there is a logical process with clear chain of actions, responsibilities and lead-times, so efficient that the CTM supply will not be on the critical path.

One of the last steps is implementing an integrated system that is capable of supporting the whole clinical trial process. The focus is on both the efficiency of the application and the safety demands given by regulatory authorities. It stands for organization of the complete material and information flow.

ProPack Data from Rockwell Automation has several modules for enterprise production management for the life sciences industry. Their website can be consulted for a full overview of the modules, www.propack-data.com. The PMX-CTM solution is the world class standard and most of the larger pharmaceutical companies have implemented this system. A major CTM supplier in Europe, Fisher, does not use the system but has designed its own. It should be clear that this is a costly and complex process and that it is not for every company a suitable solution.

This PMX system has several models and can be implemented as a stand-alone version or as a fully integrated system. In Appendix 1 the different functions sets of the PMX system are given.

At first a pilot is done where three modules are tested:

- Master batch record

Audit trailed and version controlled management of master batch records and production recipes Raw materials

procurement

Bulk production

Primary packaging

Center box assembly

Shipping Returns handling

.9 .

- Warehouse and materials management

Inventory and material management from goods receiving to goods issuing

- Weighing and dispensing

Integrated weighing & dispensing tailored to the needs of pharmaceutical operations

After a successful pilot test a fit gap analysis is conducted and the system can be adjusted to wishes of the CTM section. After approval of Japan the customized system is ready for implementation. Step-by- step more modules will be implemented so that finally the full electronic system is realized.

1.3

P

Enterprises decide to invest in information systems (IS) for many reasons, among these are: pressures to cut costs, pressures to produce more without increasing costs, and simply to improve the quality of services or products in order to stay in business. The overriding belief is that workers will use the technology to become more productive (in terms of efficiency, effectiveness and/or quality). The availability of the technique does not guarantee increased utilization or job performance. Beside that implementation results are often too optimistic. A 1998 study by the Standish Group revealed that only a quarter of IS projects were on time, half were late or over budget and another quarter was cancelled.

Therefore since the seventies, researchers have concentrated their efforts on identifying the conditions or factors that could facilitate the integration of IS into business. Their research has produced a long list of factors that seem to influence the use of technology. Reviewing literature gives numerous examples of critical success factors, by estimation these factors companies can focus on certain elements so that the implementation of the system is optimal.

After estimating these critical factors actions can be planned so that fewer problems occur and the system is more accepted. Empirical test showed that user satisfaction positively affects the amount of use. These both reflect the interaction of IT with the users.

Several user satisfaction tests are available, for instance: Looking at different performance criteria, as decision quality, consistency, decision quality, and the speed of decision-making.

1.4

O

To successfully implement an information system several actions can be performed. This research tries to determine these actions for management through the use of intention models.

1.4.1 Used theoretical models

A number of intention models try to predict and determine a certain behavior. From the mid eighties researchers tried to use and develop intention models that determine usage.

This research uses some of these models. Before explain the purpose of this research the used intention models are shortly introduced. This contributes to the understanding of the purpose of the research and defining the research scope.

Five models relate to determine the behavioral intention of using an information system. Tough, each has a different underlying theoretical foundation. In chapter 2 these models are further described:

• Theory of reasoned action (TRA)

.10 .

The main aim of the theory is to predict and understand the causes of behavior (Ajzen and Fishbein, 1980). It identifies the behavioral intention to predict the behavior of the human to perform a certain action. It perceives that individual actions are determined by the person’s attitude and the people’s subjective norm.

• Technology acceptance model (TAM)

Davis (1984) developed from the TRA, an acceptance model that determines; the initial usage of an information system, by estimating the perceived usefulness and the perceived ease of use. However these do not fully determine the usage. Sometimes the users’ attitude is used as a mediator for the behavioral intention.

• Theory of planned behavior (TPB)

This model adds an additional construct to the original Theory of reasoned action. This perceived behavioral control gives an internal and external reflection of the control people have to perform an action.

• Technology acceptance model 2 (TAM2)

TAM2 incorporates additional constructs spanning social influence processes and cognitive instrumental processes.

• Task-technology fit (TTF)

Tries to determine the cognitive fit between the technologies and the task, this can be linked to TAM.

These theoretical models all represent a different way in determining the behavioral intention to use an information system. Empirical research (Davis, et al. 1989) found that behavioral intentions to use an information system are significantly correlated with usage, and that behavioral intentions represent a major determinant of users’ behavior.

Additionally, it was found that some constructs are more important in determining usage. A primary attribute has a bigger impact on the usage. While, others have a mediate or minor effect on the intention.

Also, empirical evidence from several studies showed a difference between the models utility in the way these determined the usage.

1.4.2 Purpose of this research

The five models represent the scope of this research. By identifying the behavioral intention the actual usage can be estimated. More than one model is used to get a complete picture of the different determinants that underlie the estimation of this usage.

The objective is:

To identify factors critical to get an optimal user acceptance of the materials management system for Astellas Pharma by reviewing several intention models.

The five models represent three different categories that try to estimate the behavioral intention:

♦ The psychological category identifies the subjective norm and the attitude of the users.

♦ The tangible category represents the instrumental and controllable factors for the users.

♦ The knowledge factors as the cognitive fit and task characteristics are the last category.

.11 .

The conceptual model, figure 3, reflects the factors and how the theoretical models present different factors.

Psychological Tangible Knowledge

Attitude Subjective norm Instrumental Controllable Cognitive Task

TRA TAM TPB TAM2 TTF

Figure 3 Conceptual model

The highlighted fields reflect how the different factors relate to the models. All the theoretical models try to determine the behavioral intention in different ways. This research tries to identify which factors, and in a minority which models, best identify the success of the implementation of the CTM materials management system.

Next to that the differences between the two genders and the various age groups are examined. For men and women and for the three age groups it is determined which elements are more relevant for a successful implementation.

1.4.3 Methodology steps

First the different models are further addressed and a firm theoretical framework is classified. The literature consists of a lot of analyses and comments on the models. A review of some empirical evidence that show the relations and determination the variables are given. This theoretical framework is defined in Chapter 2. After the theory of each model some hypotheses are set that are found literature.

In literature the estimation of the behavioral intention and the elements that are appropriate for this intention is done through a quantitative analysis. Data collection is done through structured interviews, a questionnaire among the 14 users. Chapter 3 gives the results of this analysis.

Mostly used for the statistical analysis of an intention model is Structural Equation Modeling (SEM). If SEM is applied, the purpose is to fit a proposed model with empirical evidence to get a universal theoretical model. However, this is not the purpose of this research; SEM is used to modify the intention models to the current situation, this group and this system.

Due to the small size of the sample and the purpose of this research the quantitative analysis is supplemented with a qualitative part that is collected through informal interviews and taking part as a participant in meetings. Beside that there have been several sessions with the project manager to discuss which gaps are apparent in the models. This qualitative part is described in the final chapter;

Review of Results. Next to the qualitative analysis the hypotheses are discussed.

.12 .

T heoretical C ^ontext

2.1

I

This chapter gives a reflection of the theorem of the five used models, more models are found in Appendix 2. First the models are described in general and the variables that exist in these models are mentioned. Secondly some hypotheses are set, these hypotheses reflect the existing theory. Secondly, some hypotheses are made, that suppose that a relation exists among certain variables. The relations that are applicable for the PMX users are reflected in the final chapter.

2.2

T

A person’s behavior can be predicted from the person’s attitude toward the behavior (A). This is determined by his or her salient beliefs (bi) about consequences of performing the behavior multiplied by the evaluation (ei):

Secondly a person has its subjective norm. That is determined by multiplicative function of the person’s normative beliefs (nbi), i.e. the perceived expectations of specific referent individuals or groups, and the motivation to comply (mci) with these expectations:

The behavioral intention is determined by adding the attitude and subjective norm, each can be multiplied by a weight that simulates the importance of the construct:

2.2.1 Behavioral intention (BI)

An intention is defined as: a person’s location on a subjective probability dimension involving a relation between himself and some action. A behavioral intention, therefore, refers to a person’s subjective probability that he will perform some behavior.

i i

SN=

∑

∑

2

1 2

~

B I=Aw +SNw

Figure 4 Theory of reasoned

.13 .

Intentions involve four different elements: the behavior, the target object at which the behavior is directed, the situation in which the behavior is to be performed, and the time at which the behavior is to be performed.

2.2.2 Behavior (B)

The actual behavior is identical to the actual usage of the system. In IS studies the actual behavior is called the usage. There are three factors that influence the behavior:

1. The outcomes or potential consequences of behavior

2. The expectation a behavior will lead to the attainment or the avoidance of these outcomes 3. An evaluation of these outcomes

There are three major factors that influence the behavior and behavioral intention relation (Fishbein and Ajzen, 75):

1. The specificity of the intentional measure.

2. The moments between the measure of the intention and the behavioral observation. If the timings of the behavior and the intention were nearly at the same time their correlation is high.

If the timings of measurements where allowed to vary the correlation is much lower.

3. The degree to which carrying out the intention is completely under the individual’s control.

The reviewed literature addresses several items that influence the reported behavior.

Szajna (1994) mentions several problems that are caused by the use of self-reported measurements.

There is a possibility of the Halo effect; people fill in the questionnaire in such away as what they are expected to answer. This is also related to the consistency motif; respondents have an argument to maintain a consistent line in a series of answers.

Sheppard et. All (1988), made some comments on the model of reasoned action. They propose that a diversification must be made between a goal intention and a behavioral intention. Next to that, there is a difference between estimated future performance and present intentions. A choice or no choice of alternatives is also important to be included.

The issue of captive usage; even when usage is not strictly required as part of the job there may be no alternative but to use that system to effectively complete the job. Such circumstances should lead to an understatement of the relationships between the intention and usage. Therefore captive use makes it less likely to uncover associations.

2.2.3 Attitude (A)

Attitude occupies a crucial position in the mental markup of the individual and serves as a powerful energizer and director of overt behavior.

It is often defined as the study of social psychology and an enormous quantity of research on the subject of attitude has been published. It is not in the scope of this research to do a literal study of attitude. Therefore only some issues related to attitude and the discussed models are addressed here.

An attitude represents a person’s general feeling of favorableness or unfavorableness towards a stimulus object. Even though attitude has been treated as a vague and fragile construct in the IS area, its importance in individual behavior and social influence has been steadily recognized in psychology.

Attitude is contagious and when people work together, they express their own and listen to each other’s attitudes. Changes in attitude occur quickly and require less challenge than changes in non-evaluative

.14 .

beliefs or values. Continuous efforts should also be given to maintain the attitude because it is temporary, unstable and malleable.

Davis found that attitude was at best modest in predicting IS use. Some other studies however suggest that only direct perceived usefulness was best in predicting usage. However this research had a primary focus on the affective aspect of attitude.

The social psychology literature, clearly suggest that attitude has among others affective and cognitive components: The affective component of attitude refers to how much the person likes the object of thought, while the cognitive component refers to an individual’s specific beliefs related to the object.

Most research uses the affective component.

Triandis (in Fishbein and Ajzen, 1975) developed a comprehensive model for analyzing the relationships between the different factors involved in attitude. There is a hierarchical relationship among the different constructs of attitude. Affective attitude is influenced by cognitive attitude, which is affected by non-evaluative beliefs, which is in turn developed by values.

Attitude has a social function, including the individual’s internalization of norms, roles, and values, which are in turn influenced by subjective cultural variables. Several other constructs exert influence on attitude, such as genetic biological factors, but next to social factors, change and perceived consequences influence behavior.

As mentioned earlier, one way to determine attitude is multiplying beliefs with the way people can evaluate them. This represents an information-processing view of attitude formation and changes how it posits that external stimuli influence attitudes only indirectly through changes in the person’s belief structure.

Sometimes there is no diversion between intentions and attitude. This implies that the more favorable a person’s attitude is toward some object, the more he will intend to perform positive behaviors (and the less he will intend to perform negative behaviors) with respect to that object. This proposed positive relations leads to the first hypotheses of this research. If a person has a positive attitude it will give a behavioral intention.

H1. || A – BI

If there is a positive affect toward the behavior an intention is given.

Secondly a person can also give a behavioral intention regarding its feelings if there are some other stimuli.

Enhanced performance is instrumental to achieving various rewards that are extrinsic to the content of the work it self, for example pay increases and promotions. Intentions toward behavior are largely based on cognitive decision rules to improve performance, without each time requiring a reappraisal of how improved performance contributes to purposes and goals higher in one’s goal hierarchy. If an affect is not fully activated when deciding whether to use a particular system, one’s attitude would not be expected to completely capture the impact performance contributions have on one’s intention.

H2. || A – BI

No matter what the feelings are toward the behavior people can have an intention if it will improve their performance.

.15 .

Figure 5 Technology acceptance model 2.2.4 Subjective norm (SN)

Subjective norm is described as a general normative feeling that most people. In theory subjective norm is constructed of the normative beliefs and the motivation people have to comply with these beliefs. On of the models used in the paragraph 2.4 (TAM2) describes this subjective norm extensively. This describes the relation of the subjective norm with other constructs.

2.3

T

Chapter 1 shortly introduced the Technology Acceptance Model (TAM) that is adopted by Davis from the Theory of Reasoned Action (TRA). It tries to give an explanation of the determinants of computer acceptance that is general, capable of explaining user behavior across a broad range of end-user computing technologies and user populations. A key purpose of TAM is to provide a basis for tracing the impact of external factors on internal beliefs, attitudes and intentions.

Technology acceptance is defined as: “An individual’s psychological state with regard to his or her voluntary or intended use of a particular technology”.

Background and origin of TAM

In general terms, satisfaction is defined as the sum of one’s feelings or attitudes toward a variety of factors affecting the situation. These feelings are defined as, the reaction to the factor of an individual and the importance of that factor of the individual.

There is a wide variety of settings that uses user acceptance measurements by TAM. It may be used:

- By system designers to obtain user feedback on different system features of design applications.

- After implementation of a system to diagnose problems in user acceptance.

- In organizations to make selections between contending software packages.

- By examining ratings of different user groups for the same software, to determine problem areas in acceptance or deficiencies in training.

TAM proposes that perceived usefulness and perceived ease of use are of prime relevance in explaining the behavioral intention to use systems, figure 5. These instrumental factors are not the only factors that influence a given intention. Also

the usage of the system is determined by a person’s attitude towards the system. Several external factors may also influence the behavioral intention (these are addressed in section 2.2.3).

Due to the uncertain theoretical and psychometric status, subjective norm was not included in TAM. The subjective norm is one of the least understood aspects of the TRA. This is due to the difficulty to disentangle the direct effects of SN on BI from indirect effects via A.

.16 .

2.3.1 Ease of use (EOU)

Ease of use determines the degree of how much the prospective user expects the target system to be free of effort. It is defined as “the degree to which a person believes that using a particular system would be free of effort”.

Determinants of ease of use are; ease to learn, controllable, clear and understandable, flexible, easy to become skillful and easy to use. If these determinants are considered relevant then it is predicted that there is a high behavioral intention.

H3. || EOU - BI

Higher determinants of ease of use will propose a higher behavioral intention.

The technology acceptance model distinguishes between two basic mechanisms that influence attitudes and behavior through the ease of use: self-efficacy and instrumentality. These mechanisms suggest that a system is easy to interact with, the greater the user’s sense of efficacy and personal control is.

H4. || EOU - A

An easy to interact system influences the extent to which people have beliefs in the system will increase.

The self-efficacy construct is one of the major factors underlying intrinsic motivation suggested by Bandura, this is further explained in the following model.

2.3.2 Usefulness (U)

Perceived usefulness is defined as “the degree to which a person believes that using a particular system would enhance his of her job performance”. An analogy of usefulness is relative advantage.

Determinants of ease of use are: work more quickly, job performance, increase productivity, effectiveness, makes job easier and useful.

According to the original TAM, there is only one relation between the ease of use and usefulness. Only the ease of use has influence on the usefulness and not vice versa. This is supported by a number of empirical evidence. Many empirical studies have consistently identified usefulness, as a primary factor that influences IS use, while ease of use plays a less important role, particularly later in the adoption process.

H5. || U – BI

The usefulness of a system has an impact on the behavioral intention

If the design of the system is improved this will contribute to the usefulness of a system and thus improve performance. This proposes a direct influence of the ease of use to the usefulness of the system.

H6. || U – EOU

Design aspects or instrumental aspects on ease of use will influence the perceived usefulness.

Although just assumed that one’s affect toward a behavior does not fully incorporate an affect toward any rewards due to performance outcomes contingent on that behavior. However there is through learning and affective-cognitive consistency mechanisms (Bagozzi, 1982), positively valued outcomes often increase one’s affect toward the means to achieving those outcomes.

.17 .

H7. || U – A

If people experience the usefulness of an affect there is an influence on that person’s attitude.

Attitude is hypothesized to mediate the influences of usefulness and ease of use on use. If attitude is not included in the TAM model then the only constructs that implement the intention are EOU and U.

However if attitude is included it gives a mediation affect.

H8. || A – BI

Attitude serves as a mediator for the influence of ease of use and usefulness on the behavioral intention.

2.3.3 External variables (EV)

These provide the bridge between the internal beliefs, attitudes and intentions represented in TAM and the various individual differences, situational constraints and managerially controllable interventions impinging on behavior.

During literature review different external variables were highlighted. The border between an internal construct and an external variable is vague and not uniform. A short summary is given:

♦ System characteristics have a direct influence on the usefulness and ease of use of the system. TAM defines this as external while other models, for instance TAM2, includes these in a specified model. Later empirical evidence showed that these had a more direct influence on the behavioral intention.

♦ Constructs that are related to time considerations are sometimes also defined as external, such as training and experience during usage. A better approach is to measure these changes during the periods and considering these influences as the whole model is measured again.

♦ Almost every model classifies organizational factors, such as political influences, strong leadership and top management support as an external variable because they are beyond the control of the individual and the project team.

The external system characteristics have a direct influence on ease of use and a direct influence on usefulness. For example, if two systems have the same ease of use but have another output quality there will be a different usefulness but the same EOU. However this usefulness will also influence the EOU, so this has also an indirect relation with the perceived ease of use. System characteristics have an influence on the perceived ease of use.

H9. ||EV – EOU

System characteristics have influence on the ease of use.

Faith and confidence in management are also very important to get a higher user acceptance. If this is absent the users will have a lower attitude towards the usage.

H10. || EOU – EV

A higher faith and confidence in management will have a positive influence on attitude.

Perceived usefulness is affected by various external variables over and above perceived ease of use.

For example, if two systems have the same ease of use but one generates more accurate results people will prefer using this one.

However, what happens if one item has a higher useful system characteristics and another has a higher ease of use. Probably the one with the highest system characteristics is preferred as the usefulness of a system is more important than the ease of use. Although, this is highly depended on the given situation and what are the determinants of the system.

.18 .

Figure 6 The theory of planned behavior H11. ||U – EV

The system characteristics have a bigger influence on the behavioral intention through usefulness than the ease of use.

2.4

T

In 1985, Ajzen developed another adoption of the TRA, next to the TAM, that is defined as the Theory of Planned Behavior (TPB), Figure 6. This accounts for conditions of variable control.

Next to the attitude and the subjective norm, the TPB adds a new construct, perceived behavioral control. This perceived behavioral control refers to the individual’s perception of “…the presence or absence of requisite resources and opportunities”

necessary to perform the behavior.

The perceived behavioral control is defined as the control beliefs (cbi) and perceived facilitation (pfi). A control belief is (Ajzen and Madden, 1985); a perception of the availability of skills, resources, and opportunities. Perceived facilitation is; the individual’s assessments of the importance of those resources to the achievement of outcomes.

Control beliefs can be either situational (e.g. having access to a terminal) as well as personal (e.g. being able to use the system). In the next paragraph this diversification is further explained.

The used models all try to determine the behavioral intention of an information system. Taylor and Todd made an integration of the theory of planned behavior and the technology acceptance model, Figure 7.

Both extended from the TRA but have different foci, making their integration theoretically compatible and potentially complementary. Hence, the cognitive influences specified by TAM may serve as important precedents of attitudinal beliefs in TPB, which reciprocally may enhance the explanatory power of TAM via its potential for adding dimensions essential to individual technology acceptance.

The in Figure 7 shows the integrated TAM and TPB suggesting that the intention is determined by three factors; attitude, subjective norms, and perceived behavioral control. Furthermore attitude is determined by the perceived usefulness and ease of use.

Attitude (A)

Subjective Norm (SN)

P. Behavioral Control (PBC)

Behavioral

Intention (BI) Behavior (B)

Ease of use (EOU) Usefulness (U)

Figure 7 Integrated TAM and TPB

.19 .

2.4.1 Perceived behavioral control (PBC)

PBC depends on control beliefs and perceived facilitation. A control belief is a perception of the availability of skills, resources, and opportunities. Perceived facilitation is the individual’s assessment of the importance of those resources to the achievement of outcomes. The absence of facilitating resources represents barriers to usage and may inhibit the formation of intention and usage; however the presence of facilitating resources may not, per se, encourage usage.

H12. || PBC – BI

The intention to use a system is determined by the availability of skills, resources and opportunities necessary for it.

If the perceived behavioral control remains stable between the measurement of intention and behavior then this will not have a direct effect on intention. However people can influence them through the perceived facilitation of skills and resources or opportunities.

H13. || PBC – B

There is a direct relationship between perceived behavioral control and behavior due to the fact that people can control and anticipate on the availability of skills, resources and opportunities.

Empirical research provides considerable evidence that a distinction between perceived behavioral control in controllability and self-efficacy items gives a better understanding of the perceived behavioral control. Both are easier to measure and slowly have a higher level of internal consistency.

Self-efficacy

This is defined as a person’s belief in their ability to accomplish a specific task. Bandura (1986) developed the concept of self-efficacy from the Social Cognition literature. Self-efficacy is affected by past experience, by persuasion, by observing others, and affective arousal. Often self-efficacy is measured as an outcome of task performance.

Self-efficacy and perceived behavioral control are quite similar. Both are concerned with the perceived ability to perform a behavior. The distinction is, however, the efficacy expectation (i.e. the perceived ability to perform a behavior) and outcome expectations (i.e., the perceived likelihood that performing the behavior will produce a given outcome. Perceived behavioral control should be read as “perceived control over performance of a behavior”.

Self-efficacy items deal with the ease or difficulty of performing a behavior, with people’s confidence that they can perform it if they want to do so.

H14. ||PBC - SE

Higher level of self-efficacy gives a person more perceived behavioral control Controllability

Self-efficacy beliefs are reflected by internal factors whereas beliefs about the controllability of the behavior are assumed to deal with external factors.

H15. ||PBC - CO

Higher level of controllability gives a person more perceived behavioral control

A hierarchical model best describes the relations among perceived self-efficacy, perceived controllability and perceived behavioral control. In this model self-efficacy and controllability are two separate

components were each is assessed by means of different indicators.

This model proposed by Ajzen states that, although self-efficacy and controllability can be reliably distinguished, they should nevertheless be correlated with each other.

.20 .

H16. ||SE - CO

There is an interaction effect between self-efficacy and controllability

2.5

T

More than a decade after the originally TAM, Davis and Venkatesh (2000), conducted an empirical study using longitudinal data collected through the analysis of four different systems at four organizations. The constructs were measured at three points in time at each organization.

This resulted in Technology Acceptance Model 2 (TAM2), Figure 8. It incorporates additional constructs;

spanning social influences processes (subjective norm, voluntariness and image) and cognitive instrumental processes (job relevance, output quality and result demonstrability).

2.5.1 Social influence

Hartwick and Barki separated the usage in a mandatory and voluntary context. They found that, based on compliance, subjective norm has a significant effect on intention in mandatory settings but not in voluntary settings. The direct compliance effect of subjective norm on intention is theorized to operate whenever an individual perceives that a social actor wants him or her to perform a specific behavior, and the social actors has the ability to reward the behavior or punish misbehavior.

To distinguish between mandatory and voluntary usage, voluntariness is defined as: “the extent to which potential adopters perceive the adoption decisions to be non-mandatory”.

This will lead to the following hypotheses.

H17. || V – BI

If usage is considered voluntary this will have a positive influence on the behavioral intention

TAM2 encompasses two additional mechanisms by which subjective norm can influence intention indirectly through perceived usefulness.

Internalization, that refers to the process of incorporating one’s belief in its own belief structure by thinking that it is important to use a system when an important referent uses it. Mandatory use or not has no effect on this belief incorporation process.

In the case of internalization, subjective norm has an indirect effect on intention through perceived usefulness. An example of this indirect effect is that a co-worker suggests that a particular system might be useful; a person may come to believe that it actually is useful. This is applicable even when use is mandatory or not.

H18. || SN – A

If usefulness is perceived by another person this may have effect on the person’s attitude.

Identification, that refers to the image one has in a group. The degree of using an innovation is perceived to enhance one’s status in a social system. If important members of a person’s social

Figure 8 Technology acceptance model 2

.21 .

working environment believe that he or she should perform a behavior (e.g. use a system) then performing it will tend to evaluate his of her standing within the group. Related to image are job performance and power. The concept of identification has no influence on mandatory use.

When people are sensible for identification, people are willing to use an innovation when this use can enhance a person’s social status in a group, image.

H19. || SN – A

If using the system would enhance a person’s status in a group than they are willing to use it.

Additionally, the subjective norm has a direct relationship to the behavioral intention because people may choose to perform a behavior, even if they are not themselves favorable toward the system or its consequence, if they believe one or more important referents themselves think they should, and they are sufficiently motivated to comply with the referents.

H20. || SN – BI

If an important referent has certain feelings toward the system, one incorporates the referent’s belief into one’s own belief structure.

Subjective norm determines people’s willingness to perform the behavior or its consequences even if they are themselves not favorable toward this behavior. This is due to notion of inferential beliefs that are derived through the process of inference from descriptive, informational, or other inferential beliefs.

This creates the possibility that attitudinal beliefs may be formed from normative beliefs, and vice versa.

Attitude and subjective norm can have a separate influence on the usage but it is unlikely that they are completely interdependent. Other theorists think that the attitude and subjective norm are highly related. Yet Fishbein and Ajzen constantly show that there is utility in separating both beliefs, despite the possibility that they are highly correlated.

In general there are four different empirical views on the relation of the influences belief have on behavior trough attitude.

- Indirect (Fishbein and Azjen) - Co-determinants (Triandis, 1977)

- Attitudes are antecedents of beliefs (Weiner, 1986)

- Beliefs (Attitude) do not fully mediate the effects of usefulness and ease of use (Davis, 1989)

If the user gets some experiences then the proposed effects will weaken. The direct effect on subjective norms will be stronger during implementation and early use but will weaken over time as experiences grow. Similarly the effect on perceived usefulness (internalization) will weaken over time. The issue of experience is later further addressed.

2.5.2 Cognitive instrumental process

The cognitive part has three theoretical underpinnings: work motivation theory (Vroom, 1964), action theory from social psychology (e.g. Fishbein and Ajzen, 1975) and task-contingent decision making from behavioral decision theory (e.g. Beach and Mitchell, 1978). Recent work in these areas has converged on a view of behavior being driven by a mental representation that links higher-level goals to specific actions that are instrumental for achieving those goals.

Job relevance refers to the task that one is capable of supporting and the importance these tasks are in the corporate performance. It is known that users can possess distinct knowledge about their job situation, in which they can use as the basis for determining what tasks can be performed with a given

.22 .

system. Job relevance is a cognitive judgment that exerts a direct effect on perceived usefulness, distinct form social influence processes.

H21. ||JR – U

If a system is more relevant for a person’s job this will have a positive impact on perceived usefulness.

A second item mentioned in this cognitive process is output quality. Over and above if the tasks are relevant for the people’s jobs, people will take into consideration how well the system performs those tasks. Output quality is considered to be distinctly different from perceived usefulness. This has high relations with several end-user satisfaction studies (for instance performed by Doll).

H22. || OQ – U

When the output quality is considered low it will have a negative effect on the perceived usefulness

A situation can occur when the system produces job-relevant results, but does so in an obscure way, that the users of the system are unlikely to understand the real usefulness of the system. This so-called result demonstrability (Moore and Benbasat, 1991) is important to get a “tangibility of the results of using the innovation”.

The relation between results demonstrability and perceived usefulness is also consistent with the job characteristics model, which emphasizes knowledge of the actual results of work activities as a key psychological state underlying work motivation (Hackman and Oldham 1976, Loher et al. 1985).

H23. || RD – U

When the result are considered optimal it will have a positive effect on perceived usefulness.

The effects in the cognitive process are not related through time and will remain the same.

2.6

T

The emphasis on business productivity and efficiency suggests an additional perspective, namely the task-technology fit (TTF). A model developed by Goodhue in 1988, matches the capabilities of the technology to the demands of the task. TTF was designed to evaluate an organization’s overall information technology system rather than an individual application. Technologies are viewed as tools used by individuals in carrying out their tasks. Tasks are more broadly defined as the actions carried out by individuals in turning inputs into outputs. Rational, experienced users will choose the tools and methods that enable them to complete the task with the greatest benefit. A higher TTF would result in a higher performance.

The TTF can also be identified as the cognitive fit. There is no universal definition but according to Vessey and Galletta (1991), quoted by Goodhue.

“Cognitive fit is a cost benefit characteristic that suggest that, for most effective and efficient problem solving to occur, the problem representation and any tools or aids should all support the strategies (methods and processes) required to perform the task”.

This is similar to the definition of “task-system” fit, by Goodhue and Thompson. This is the degree to which an information system environment assists and individual in performing his or her portfolio or tasks.

TAM and the TTF overlap in a significant way and if integrated they could provide an even stronger model than either standing alone.