The adoption of the web
“A study on web adoption levels established by SMEs”
Danny Kerstholt s1677845
d.kerstholt@student.rug.nl June 2009
University of Groningen, The Netherlands Faculty of economics and business
Msc BA Small Business & Entrepreneurship
First supervisor: Dr. M.J. Brand
Table of contents
1 Introduction 2
2 Literature review: web adoption models 4
2.1 Web adoption models 5
2.2 Review of three web adoption models 6
3 Literature review: main factors of web adoption 14
3.1 Innovation characteristics 16
3.2 Factors on firm level 17
3.3 Factors on market level 19
3.4 Conceptual models 20
4 Research design 21
4.1 Sample 21
4.2 Measurements 22
4.3 Analysis 25
5 Discussion of the results 28
5.1 Analysis step one 28
5.2 Analysis step two 32
5.3 Non-parametric test results 35
6 Conclusions 36
7 Theoretical and practical implications 38
8 Limitations 39
9 References 40
10 Appendix A: Survey questions 43
11 Appendix B: Codebook 44
The adoption and use of the internet has a big impact on the way companies do business, which is expected to increase even more in the upcoming years. E-business W@tch (2006) shows that 96 % of the European companies have an internet connection and 76 % of the European companies are currently connected through broadband internet like cable and DSL. The percentage of companies that place and accept orders online has increased enormously over the last years. The research by e-business W@tch shows that around 37 % of the medium sized companies and 30 % of the small sized companies are currently placing 5 % or more of their orders online. Small and medium sized companies that are accepting 5 % or more of their orders online account for respectively 16 and 17 %. Furthermore, companies are more and more integrating business processes with ICT. Therefore, companies are using software like ERP systems (Enterprise Resource Planning). These systems support and link business processes like administration, supply management and receiving and placing orders. This software can be combined with the use of the web resulting in integration of the web with business processes. Currently, 16 % of the small sized and 25 % of the medium sized European companies are using systems like this (E-business W@tch, 2006).
Our study builds upon the theory of web adoption models. Web adoption can be seen as a term covering the adoption of the internet by companies through the means of e-mail, a simple website offering company or product information, till a highly advanced and fully integrated website with all possible features. Adopting the web enables SMEs to overcome competitive disadvantages of firm size, such as transaction costs, market reach and resources (Santarelli and D’Altri, 2003). In other words, adopting the web can increase a company’s competitiveness.
between SMEs at the various levels of web adoption may lead to new insights for policy makers.
This study’s main research question is: Which kind of companies establish which levels of web adoption? In order to answer this question our first step is to select a web adoption model that is useful for our purpose. Our second step is to identify the main factors that influence a company’s level of web adoption and to develop our propositions. These first two steps will lead to our conceptual model. Subsequently, we will empirically validate the identified relationships by making use of a data-set acquired by a fellow student, who used this data for his master thesis on e-commerce within SMEs. The results of our empirical part will lead to a discussion on the relationship between our main factors of web adoption and the various levels of web adoption. A special emphasis will be placed on the relationship between our main factors of web adoption and the intention to further adopt the web.
This paper is structured as follows. First, our literature review includes a review of various web adoption models in order to determine a model that is useful for answering our main research question. The second part of our literature review includes a discussion on the main factors that influence a company’s level of web adoption which will lead to our propositions. This study’s literature review is concluded with a conceptual model representing the relationship between our main factors of web adoption and the various levels of web adoption and a conceptual model representing the relationships between our main factors of web adoption and the intention to further adopt the web. Our literature review is followed by our research design, describing how we will test our propositions. This study concludes with a discussion of the results, our main conclusions, the implications of our results and the limitations of our study.
In this section we present a review of the relevant literature. First, we describe on which web adoption models our focus will be, followed by a review of three web adoption models. This will result in a web adoption model best suited to the purposes of our study.
2.1 Web adoption models
Literature on web adoption presents several sorts of models for web adoption. Some models are describing the possible types of websites companies might have. For example, Dholakia and Kshetri (2004) differentiate between sites that provide information, interaction and transactions. Other models describe the roles that the web might have within a company. Levy and Powell (2003) for example, have developed a model identifying roles that the use of the web might have within a company. A company may use the web to simply give presence on the web, to utilize opportunities offered by the web, to support business processes and the web can be seen as the key to the development of a company. They have based these roles on the variables perceived value of the web and attitude to business growth, which they discuss to be the main drivers for web adoption. Their case studies discuss the presence of the factors perceived benefits, IT knowledge and competitive pressure between the companies at the various roles that the web might have within a company. Their model can be found in figure 1.
Figure 1: Roles of web adoption (Source: Levy and Powell, 2003)
web adoption models that address these multiple levels of web adoption that companies may establish. The essence of adopting the web at multiple levels is that companies may adopt the web step by step. From the use of e-mail, to adopting a simple web site and finally a fully integrated web site with all possible features/functions. These web adoption models are sequential in nature and describe a route for web adoption that companies usually will follow. Companies will usually move forward in their web adoption process, but may take a step back or remain a specific level of web adoption as well. Such models of web adoption are either descriptive or normative, describing the way that companies are adopting the web or proposing a way that companies are ought to adopt the web. The purpose of these web adoption models is to create an understanding of the way that companies are adopting the web.
2.2 Review of three web adoption models
Today's web adoption literature presents several web adoption models that address the various levels of web adoption that companies may have established. The models we found are proposed by: Daniel, Wilson and Myers (2002), Rao, Metts and Monge (2003) and Teo and Pian (2004). Our search for web adoption models in the current literature started with an article on the impact of the current level of adoption by Brand and Huizingh (2008). This article builds further on research that notices that e-commerce can be adopted at multiple levels. This led to our first two web adoption models, proposed by Daniel, Wilson and Myers (2002) and Teo and Pian (2004). Subsequently, we searched for web adoption models in electronic databases such as EBSCOhost resulting in our third web adoption model proposed by Rao, Metts and Monge (2003).
model on it’s comprehensiveness, i.e. the extent to which the model includes low and high levels of web adoption and the various levels in between.
2.2.1 Daniel, Wilson and Myers (2002)
The model proposed by Daniel, Wilson and Myers (2002) is the result of a cluster analysis. The cluster variables used in their study are a wide range of business activities for which SMEs are using e-commerce, such as providing information on the company, taking orders, ordering and payment, communication etc. The business activities used by Daniel, Wilson and Myers (2002) are discussed in current literature on the use of e-commerce in the specific context of SMEs. The cluster analysis led to a model distinguishing four levels/clusters of web adoption. These clusters and the number of SMEs in each cluster are: Developers (39 SMEs), Communicators (135 SMEs), Web presence (203 SMEs) and Transactors (208 SMEs).
The model Daniel, Wilson and Myers (2002) propose represents a set of sequential levels which SMEs have established. It presents a route that can be followed by other SMEs. At the first level, Developers, SMEs are developing their first e-commerce services. SMEs at the second level of web adoption are using e-mail to communicate with customers, suppliers and employees. The third level, Web presence, is characterized by SMEs having an information based website online and SMEs developing online ordering facilities. At the last level, Transactors, SMEs can be seen as the most advanced adopters, having online ordering in operation and developing online payment features.
were active in an wide variety of industry sectors, across the UK. The size distribution of the responding SMEs can been found in table 1.
Employees
Percentage
Turnover
Percentage
0 – 5
6%
< £ 1.000.000
18%
6 – 50
51%
£ 1m - £ 5m
42%
51 – 250
43%
> £ 5.000.000
40%
Table 1: Size distribution of the respondents (Source: Daniel, Wilson and Meyers: 2002) Since the authors present the model as a route for SMEs considering the adoption of the web and for SMEs at the lower levels of adoption, it suggests the model to be normative. However, due to the cluster analysis, the model corresponds with the actual levels of web adoption established by SMEs in the UK. Therefore, we assume the model to be descriptive.
2.2.2 Rao, Metts and Monge (2003)
Rao, Metts and Monge (2003) recognize the importance of a model describing the logical evolution of e-commerce involving different levels of development. SMEs are considered to be a major component of all economies and are generally considered to be flexible, adaptive and innovative organizations. This makes them a good fit for e-commerce, which has a big impact on the way companies are doing business. The increasing interest in web adoption and the lack of research on web adoption models led Rao, Metts and Monge to propose a theoretical model for web adoption for SMEs. Previous research on web adoption by O’Conner and O’Keefe (1997) and Timmers (1999) underpin the proposed stages by Rao, Metts and Monge (2003). They argue that the use of e-commerce differs on the level of information content, the level of transactions and the level of integration. This led Rao, Metts and Monge (2003) to propose the following levels of web adoption: Presence, Portals, Transactions integration and Enterprises integration.
The model appears to be sequential, but as noted before, an SME may enter at any stage and it is not necessary to move through each subsequent level in order to establish the highest level of web adoption. Furthermore, the authors state that their model is descriptive. However, their intention is to describe the logical evolution of e-commerce. This indicates a normative model. Since they merely describe how the different stages ought to be characterized and given the fact that the model is not empirically validated, we assume the model to be normative.
In order to evaluate the current status of e-commerce development among SMEs, Rao, Metts and Monge (2003) classified 153 SMEs from the KITE inventory. KITE stands for Knowledge and Information Transfer on E-commerce. It contains a data set acquired by a self-selection process, whereby companies engaged in e-commerce can log on and register their particular e-commerce use by filling out a survey form. The data being acquired through a self-selection process, lets us conclude that the SMEs filling out these surveys are already interested in e-commerce. This may lead to a higher level of web adoption among these SMEs. The numbers of firms found to belong at the different levels is as follows: Presence (24 SMEs), Portals (45 SMEs), Transactions integration (84 SMEs) and Enterprises integration (0 SMEs). The SMEs that filled out the KITE survey, were active in a wide variety of industries across the EU. The size of the classified SMEs can been found in table 2.
Employees Percentage
0 – 5 49%
6 – 50 35%
> 50 16%
processes with the web in the classification process of Rao, Metts and Monge (2003) and perhaps the absence of such a category within the Kite inventory as well to be responsible for the fact that not one of the SMEs has been classified as an SME at the highest level of web adoption.
2.2.3 Teo and Pian (2004)
Teo and Pian (2004) found that researchers often distinguish between levels of web adoption according to the functions of web sites. Subsequently, they state that different levels of web adoption (more or less functions) can facilitate different business activities. The role that the web plays within an organization is determined by an organization’s internet strategy. This role determines the business activities that will be supported by the functions offered by the web. Therefore, Teo and Pian take into account an organizations internet strategy and its web site functions in developing their model.
before, an SME may enter at any stage and it is not necessary to move through each subsequent level in order to establish the highest level of web adoption. The adoption of the web can be seen as a continuous process, in which companies may move up or possibly down a level over time.
The model proposed by Teo and Pian (2004) is a theoretical model, which has been validated empirically. In order to validate their model, questionnaires were send to 566 firms, selected from the “Singapore 1000” and the “SME 500” from Singapore. The mailing resulted in 159 usable responses, giving a response rate of 28.8 %. The number of firms found to belong at the different levels is as follows: Level 0 (27 firms), Level 1 (48 firms), Level 2 (52 firms), Level 3 (18 firms) and Level 4 (14 firms). The responding companies are active in a wide variety of industries in Singapore. Unlike the other presented web adoption models, the model of Teo and Pian (2004) is not verified among only SMEs. Teo and Pian found larger firms to be more involved in level 3 and 4 but do not test for further differences between SMEs and large firms according to their levels of web adoption. The size of the companies in numbers of employees and turnover can been found in table 3.
Employees Percentage Turnover Percentage
0 – 100 23.9 % S$ < 100m 34.6 %
101 – 300 27.7 % S$ 100m – 300m 30.8 %
301 – 1000 22% S$ >300m 31.5 %
> 1000 25.1 % Missing 3.1 %
Missing 1.3 % Exchange rate: S$ 1 = € 0,58
Table 3: Size distribution of the respondents (Source: Teo and Pian, 2004)
adoption that actually have been established by firms in Singapore. Therefore, the model can be qualified as descriptive.
2.2.4 Discussion
Since our objective is to select a model for web adoption that is useful for grouping SMEs and determining their levels of web adoption, we have reviewed various models for web adoption. This section will discuss the various models according to the criteria we have developed in order to select the best suited model for our purpose.
A link with the current literature shows the relevance of the model and explains the logic behind a model. Daniel, Wilson and Myers (2002) reviewed the business activities discussed in current literature on e-commerce in the context of SMEs. These business activities are used as their cluster variables. Rao, Mets and Monge (2003) based their model on previous research by O’Conner and O’Keefe (1997) and Timmers (1999). Teo and Pian (2004) found that researchers often distinguish levels of web adoption according to the functions web sites have. This shows that each model has some theoretical foundation. However, an extensive discussion of the literature on web adoption and explanation of their choices is omitted in each study.
A descriptive model with a strong empirical validation is needed to ensure a good reproduction of levels of web adoption that actually have been established. Daniel, Wilson and Myers (2002) performed a cluster analysis which resulted in an model with a strong empirical validation, which represents a good reproduction of levels of web adoption that actually have been established (i.e. a descriptive model). Rao, Metts and Monge (2003) state that their model is descriptive. However, their intention is to describe the logical evolution of e-commerce, which indicates a normative model. Since they merely describe how the different levels ought to be characterized and given the fact that the model is not empirically validated, we assume the model to be normative. The model Teo and Pian (2004) propose is empirically validated. Their data analysis shows that the implemented web site functions across the various levels of web adoption are significantly different. Furthermore, we have found the model proposed by Teo and Pian (2004) to give a good reproduction of the levels that actually have been established by firms in Singapore, what indicates a descriptive model.
web adoption from each model. The table starts at the top with the lowest levels of web adoption and ends at the bottom with the highest levels of web adoption. Table 4 shows us that Daniel, Wilson and Myers (2002) do not differentiate a level in which integration is taken into account, they mostly differentiate lower levels of web adoption. We expect this to be the result of the fact that they used business activities as cluster variables, whereby integration of business processes is omitted. Based on table 4, Rao, Metts and Monge (2003) seem to give a comprehensive set of web adoption levels. They only omit a level in which SMEs do not adopt the web or only make use of email. Furthermore, they did not find any company to have established their highest level of web adoption. Last but not least, the model proposed by Teo and Pian (2004). They seem to cover all the levels of web adoption that companies have established in a comprehensive way. From firms that have not adopted the web and only make use of an email account, to companies only having an informational site, companies having a more elaborated web site, companies that enable companies to do transactions through their website until companies that have integrated their business processes with the web. Thereby, Teo and Pian seem to give the most comprehensive model of web adoption.
It has been concluded that each model is supported by some literature, what is desirable in order to suit the purpose of this study. In addition, a descriptive model is needed to suit the purpose of this study. Rao, Metts and Monge (2003) do not meet this requirement. Their model is found to be normative. This leaves us with the models, proposed by Daniel, Wilson and Myers (2002) and Teo and Pian (2004). Our overview and discussion of the levels of web adoption and its distribution, reveals that the model proposed by Teo and Pian (2004) gives the most comprehensive distribution of web adoption levels. We conclude this model to be best suited to the purpose of this study.
2.2.5 Adjustments
for their products/services. We rather see the integration of online ordering and online payment possibilities as a way of automating business processes. In fact, by offering online ordering and online payment possibilities, a company transforms their way of doing business. Integration of business processes with the use of the web and ICT is taking place. Therefore, the levels business integration and business transformation will be combined in this study. This leads to a model that consists of three levels: Web presence, Prospecting and Business transformation, which enables us to classify SMEs according to the extent that they have adopted the web (i.e. determining an SME’s level of web adoption). This model describes the levels of web adoption that can be established, which represent the extent to which a company has adopted the web. Companies may enter at any level and can move up a level, remain their level or possibly move down a level within the model.
In the previous section a model for web adoption is selected in order to answer our main research question: which kind of companies establish which levels of web adoption? This model distinguishes various levels of web adoption. In order to describe which kind of companies establish the various levels, we need to identify how companies differ across the various levels. Research on IT related adoptions discusses a wide range of factors that are found to influence the adoption of the web (e.g. Premkumar and Roberts, 1999; Beatty, Shim and Jones, 2001; Mehrtens et al., 2001). We expect these factors to differ across the various levels of web adoption, which will enable us to describe how companies differ between the various levels of web adoption.
In addition we will include a more dynamic aspect. Literature on web adoption mainly focuses on the situation at a certain moment. The aspect of time is disregarded. An exception is the study by Beatty et al. (2001), who incorporates the aspect of time by placing a special emphasis on the differences of several factors of web adoption between early and later adopters. As we have discussed in the previous section on web adoption models, the adoption of the web is a process in which companies may establish a higher or possibly lower level of web adoption over time. We assume that our selected model represents sequential levels of web adoption. In general, companies will remain at their level of web adoption or adopt a higher level of web adoption over time. Therefore, we will place a special emphasis on the concept of companies that may move up a level within our model of web adoption. We will examine the relationship between the main factors of web adoption and the intention to further adopt the web, which we will control for the current level of web adoption.
In this section we will discuss the factors that we have identified to be the main factors of web adoption (i.e. the factors that are repeatedly discussed in research on IT-related adoptions and are found to be significant). Our selection of main factors is the result of extensive literature research and is largely supported by Jeyarai, Rottman and Lacity (2006), who reviewed the factors on IT-based innovations in IT innovation adoption research.
level of web adoption and how we expect that the factor is related to the intention to further adopt the web.
3.1 Innovation characteristics 3.1.1 Perceived benefits
Perceived benefits are efficiency benefits from the relative advantage of adopting the web over traditional methods (Mehrtens et al., 2001 and Beatty et al., 2001). A company has to perceive efficiency benefits from a specific level of web adoption before the decision to adopt this level will be made. Adopting the web to a higher level can lead to higher efficiencies as reduced transaction costs, increased productivity and better customer service (Beatty et al., 2001). Various studies have discussed the relationship between perceived benefits and the adoption of IT related technologies and found perceived benefits to influence the adoption of IT related technologies positively (e.g. Mehrtens et al., 2001 and Premkumar and Roberts, 1999). This leads us to expect that:
1. Companies at the higher levels of web adoption will perceive more benefits from the web than companies at the lower levels of web adoption.
2. Companies that have the intention to further adopt the web will perceive higher benefits from the web than companies that do not have the intention to further adopt the web.
3.1.2 Complexity
Complexity is seen as the degree to which an innovation is perceived as relatively difficult to understand and use (Rogers, 2003). Our model of web adoption distinguishes three levels of web adoption. Each higher level is more elaborated and thereby more difficult to understand and use (i.e. more complex) than a lower level. The degree to which the web or a level of web adoption is perceived as complex will differ among companies. Several studies suggested perceived complexity to influence the adoption of IT related technologies negatively (e.g. Premkumar and Roberts, 1999 and Beatty et al., 2001). This implies that companies that perceive the web to be complex, are less likely to adopt the web to a higher level. This leads us to expect that:
2. Companies that have the intention to further adopt the web will perceive the web as less complex than companies that do not have the intention to further adopt the web.
3.1.3 Compatibility
Companies are more likely to adopt an IT related technology such as the web if they perceive that it is consistent with their culture, values, work practices and existing IT infrastructure (Beatty et al., 2001). In most cases a simple web site can be easily adopted without perceived incompatibilities. This is in contrast with adopting a higher level of web adoption, which may require a change in existing work practices, existing infrastructure, existing culture and so on. Grandon and Pearson (2004) found compatibility to be a significant determinant for web adoption. Their findings imply that companies that perceive their organization to be compatible with the web are more likely to adopt the web to a higher level. This leads us to expect that:
1. Companies at the higher levels of web adoption will perceive their organization to be more compatible with the web than companies at the lower levels of web adoption.
2. Companies that have the intention to further adopt the web will perceive their current level of web adoption as more compatible with their organization than companies that do not have the intention to further adopt the web.
3.2 Factors on firm level 3.2.1 Firm size
Firms size is widely recognized as a factor of web adoption and is found to influence the level of web adoption positively (e.g. Aguila-Obra and Padilla-Meléndez, 2006; Forman en Goldfarb, 2005; Forman, 2005 and Teo and Pian, 2004). The reason for this positive relationship is twofold.
larger companies are better able to bear adoption risks, making the decision to adopt the web easier.
Second, adopting the web at a higher level can be more attractive for larger companies due to economies of scale and scope. Larger companies are able to spread the fixed costs of implementation and maintenance over a higher number of sales.
One argument suggesting smaller companies to adopt the web faster is that smaller companies are suggested to be more flexible and less complex. This is why smaller companies are also seen as more innovative companies. Smaller and younger companies face less constraints with regard to their existing culture, values and beliefs, and their business processes. Established companies might have to transform their business processes in order to make them applicable for web adoption.
In general, literature on web adoption suggest firm size to influence the adoption of the web positively. A web site at the lower levels of web adoption can be easily established at low costs. Adopting the web at a higher level is more complex and can only take place with substantially higher investments, which is more feasible for larger companies or companies with a higher IT-budget. This leads us to expect that:
1. Companies at the higher levels of web adoption will be larger in numbers of FTE’s and turnover and have a higher IT- budget than companies at the lower levels of web adoption.
2. Companies that have the intention to further adopt the web will be larger in numbers of FTE’s and turnover and have a higher IT-budget than companies that do not have the intention to further adopt the web.
3.2.2 Management support
Roberts, 1999 and Premkumar, 2003). This implies that companies with a supportive management are more likely to adopt a higher level of web adoption and will have a more positive attitude towards the adoption of a higher level. This leads us to expect that:
1. Companies at the higher levels of web adoption will have a more supportive management compared to companies at the lower levels of web adoption.
2. Companies that have the intention further adopt the web will have a more supportive management than companies that do not have the intention to further adopt the web..
3.2.3 IT expertise
IT-expertise can be acquired through desk research, training and experience. Experiences with IT related technologies can be seen as the most important factor for IT-expertise. Experiences with IT related technologies increases the skills and knowledge to use and adopt the web. Huizingh and Brand (2007) found that companies at the advanced level of web adoption have more knowledge with regard to e-commerce compared to companies at the basic level of web adoption. This can be explained by the concept of companies that move from the basic levels of web adoption to the advanced levels of web adoption, where organizational learning takes place and experiences with the web increase. Kuan and Chau (2001) found companies that have adopted EDI (i.e. IT-related technology) to perceive themselves to possess the necessary IT expertise. This leads us to expect that:
1. Companies at the higher levels of web adoption will possess more IT expertise than companies at the lower levels of web adoption.
2. Companies that have the intention to further adopt the web will possess more IT expertise than companies that do not have the intention to further adopt the web.
3.3 Factors on market level 3.3.1 Competitive pressure
higher level. Wymer and Regan (2005) found competitive pressure to be an important factor with regard to the adoption of the web (i.e. Companies that experience competitive pressure are more likely to adopt the web to a higher level). This leads us to expect that:
1. Companies at the higher levels of web adoption will experience more competitive pressure than companies at the lower levels of web adoption.
2. Companies that have the intention to further adopt the web will experience more competitive pressure than companies that do not have the intention to further adopt the web.
3.4 Conceptual models
The discussion on the main factors of web adoption leads us to present two conceptual models. The first model, presented in figure 2, represents the relationships between the levels of web adoption and our main factors on web adoption. This first model will be used for step one of our analysis.
The second model, presented in figure 3, represents the relationships between a company’s intention to further adopt the web and the main factors of web adoption. We will control these relationships for the current level of web adoption. This second model will be used for step two of our analysis.
Innovation characteristics:
-Perceived benefits +
-Complexity
-Compatibility +
Factors on firm level: Level of web adoption
-Firm size +
-Management support +
-IT expertise +
Factors on market level:
+ -Competitive pressure
Level of web adoption
Innovation characteristics:
-Perceived benefits +
-Complexity
-Compatibility +
Factors on firm level: Intention
-Firm size +
-Management support +
-IT expertise +
Factors on market level:
+ -Competitive pressure
Controlled for level of web adoption:
Figure 3: Conceptual model, step two of our analysis
4.1 Sample
For this study we used a data-set that is acquired by a former student, named Ruben Sardjoe. Ruben Sardjoe collected this data-set using an on-line survey in 2005. In total 1213 SMEs were selected at the Dutch chamber of commerce. These companies were active in three sectors. Namely, the manufacturing sector, the logistics and communications sector and the financial and business services sector. The selected SMEs were contacted by telephone in order to ask for their cooperation. The contacted SMEs were divided into the following categories: agree to cooperate (368 SMEs), no decisive answer (277 SMEs) and no cooperation (568 SMEs). Subsequently, a mail with a link to the online survey was sent to the SMEs that agreed to cooperate and those that did not give a decisive answer. This procedure led to 203 usable responses.
three selected sectors in the sample. A chi-square test between both groups confirms this. The calculation and the Chi-squared can be found in table 5. Further analysis of the data shows that a large majority of the responding companies is in business for over ten years (92%). The distribution of the firm size in numbers of employees is as follows: 14% between 1 and 50, 44% between 50 and 100, 36% between 100 and 250 and 6% > 250.
Sector Selected SMEs
Observed frequency (Fi)
Expected
frequency (Ei) (Fi - Ei) ((Fi - Ei) ^ 2) / Ei
Manufacturing 581 99 97,23 1,77 0,032130434
Logistics and communications 181 32 30,29 1,71 0,096419126
Financial and business services 451 72 75,48 -3,48 0,160130412
Total 1213 203 203 Chi-Squared = 0,288679972
Table 5: Chi-squared, Goodness-of-fit Test Statistic (Industries)
4.2 Measurements
The majority of the measurements we have used come from literature on web adoption. An extensive list of items, identified in literature on web adoption was composed by Ruben Sardjoe. The items that are used in the survey are selected from this list. Where applicable, all measurements are standardized to five-point Likert scales.
A companies level of web adoption is measured by making use of the model proposed by Teo and Pian (2004). The respondents are asked to indicate which level they have established according to a description that is given for each level. Since we made two adjustment to the model, we have combined the results for level 3 (business integration) and level 4 (business transformation) and we have disregarded level 0 (e-mail adoption). Disregarding level 0 did not have any consequences, all our respondents had established at least level 1 (web presence).
3.33, are seen as companies that have the intention to further adopt the web. A score higher than 3.33 represents the intention to substantially extend the use of the web for at least one of the core processes in the upcoming two years. A lower score represents the companies that do not make a clear statement with regard to further adoption of the web and represents the companies that do not have the intention to further adopt the web. The companies that score higher than 3.33 account for 42.4 % (86 companies) and the companies that score 3.33 or lower account for 57.6 % (117 companies).
Perceived benefits are measured with the 8-item scale developed by Beatty et al. (2001). For each item the respondents indicated on a five-point Likert scale to which degree the items are perceived as important with regard to the adoption of a higher level. A factor analysis on the items of perceived benefits reveals that there are two underlying components for perceived benefits. The first component is related to serving the market and the second component is related to organizational efficiency. The component matrix of our factor analysis can be found in table 6.
1 2 Operational efficiency ,655 Customer service ,846 Transaction costs ,753 Cash flow ,867 Productivity ,863
Opportunities for competition ,643
Customer reach ,888
Customer relations ,866
Component
Table 6: Rotated component matrix for perceived benefits.
Complexity is measured with the items developed by Venkatraman (1991), Premkumar and Roberts (1999) and Beatty et al. (2001). This resulted in a 4-item scale. The respondents were asked to indicate on a five-point Likert scale to which degree these items were perceived as complex. The scores are combined into a variable for complexity which contains the MEAN score on the 4 items.
Compatibility is measured with the 3-item scale developed by Grandon and Pearson (2004). The respondents were asked to indicate on a five-point Likert scale to which degree the web is perceived to be compatible with these items. The scores are combined into a variable for compatibility which contains the MEAN score on the 3 items.
Firm size is measured in numbers of full time equivalents (FTE’s), turnover and IT budget. These items were used as separate variables in our analysis. Turnover and IT budget are both measured on a ordinal scale and have to be considered as ordinal data. The number of FTE’s are measured in absolute numbers and have to be considered as ratio data. Management support is measured with the items developed by Premkumar and Roberts (1999) and Beatty et al. (2001). This resulted in a 6-item scale. Factor analysis on these items did not reveal any underlying components. The respondents were asked to indicate on a five-point Likert scale to which degree they agree with propositions on these six items of management support with regard to the adoption of the web. The scores are combined into a variable for management support which contains the MEAN score on the 6 items.
IT expertise is measured with the 3-item scale developed by Kuan and Chau (2001). The respondents were asked to indicate the importance of these items with regard to the web on a five-point Likert scale. The scores are combined into a variable for IT expertise which contains the MEAN score on the 3 items.
With regard to competitive pressure no explicit measurements were incorporated in the survey. This leaves us to disregard the factor competitive pressure in the remaining of our study.
Analysis of our variables showed a relatively high number of missings for the variables FTE’s, turnover and IT budget. The missings account for respectively 17, 22 and 23 respondents, on a total of 203 respondents, that did not indicate their firm size in number of FTE’s, turnover or IT budget. Nevertheless, we suppose to have enough respondents that indicated their firm size to perform a reliable analysis. Analysis of the other variables showed no missings.
The reliability of the multi-item scales is measured with Cronbach’s alfa and can be found in table 7. The reliabilities of the scales range from 0,763 to 0,927, which indicates that the internal consistency of the multi-items scales is quite good. The survey questions and the variables we used for our analysis are presented in Appendix A and B. Factor analysis is only conducted for perceived benefits and management support, since these variables are based on a higher number of items. We found two underlying components for perceived benefits and just one component for management support.
Multi-item scales items Cronbach's alfa scales developed by:
Perceived benefits Beatty et al. (2001)
-Serving the market 4 0,843
-Operational efficiency 4 0,809
Complexity 4 0,763 Venkatraman (1991), Premkumar and
Roberts (1999) and Beatty et al. (2001)
Compatibility 3 0,808 Grandon and Pearson (2004)
Management support 6 0,927 Premkumar and Roberts (1999) and
Beatty et al. (2001)
IT expertise 3 0,814 Kuan and Chau (2001)
Table 7: Reliability of the multi-item measurements.
4.3 Analysis
4.3.1 Analysis step one
is conducted to test for differences on the various main factors between pairs of levels. This post-hoc test is used because it is known as a reliable test that gives us a confidence level for the comparison of all pairs combined instead of a confidence level for each comparison on its self.
4.3.2 Analysis step two
To test for differences on the main factors of web adoption between companies that have the intention to further adopt the web and companies that do not have the intention to further adopt the web two-way ANOVA is used. Two-way ANOVA enables us to control the relationship between our main factors of web adoption and the intention to further adopt the web for the effect of the current level of web adoption. For this analysis, we will make use of our grouping variable for the intention to further adopt, which divides the respondents into a group that has the intention to further adopt the web and a group that does not have the intention to further adopt the web.
Subsequently, we will verify the results of our two-way ANOVA by making use of partial correlation. We will test for the correlation between our main factors of web adoption separately and the intention to further adopt the web, controlled for the current level of web adoption. This will strengthen our results. Partial correlation assumes that the variables are interval data. Our main factors for web adoption and the intention to further adopt are interval data. The level of web adoption is ordinal data. Since, we use partial correlation to verify the results of our two-way ANOVA, we assume the variable level of web adoption to be interval data as well.
4.3.3 Conditions
Main Factors Levene's
Statistic P-value
Levene's
Statistic P-value Perceived benefits
-Serving the market 0,220 0,803 1,209 0,307
-Organizational efficiency 0,056 0,945 0,381 0,861 Complexity 0,809 0,447 0,739 0,595 Compatibility 0,666 0,515 0,738 0,596 Management support 1,878 0,156 1,646 0,150 IT expertise 0,928 0,397 1,267 0,280 Turnover 2,936 0,056 2,600 0,027* IT Budget 2,047 0,132 2,900 0,015* FTE's 1,319 0,270 0,894 0,486
One-way ANOVA Two-way ANOVA
Table 8: Levene’s test statistics.
The normal distribution of our dependent variables is tested by making use of the statistics for skewness and kurtosis. The statistics for each dependent variable can be found in table 9. We will compare the standard error of both Skewness and Kurtosis with their absolute statistics to check if the Skewness and Kurtosis are significantly non-normal. Twice the value of the standard errors should be higher than the absolute value of the statistics in order to accept that the variables are normally distributed (note: the negative signs have to be disregarded).
The results indicate that the variables organizational efficiency, complexity, compatibility, management support and FTE’s are normally distributed. The rest of the variables are non-normal distributed, of which turnover and IT-budget are found to have unequal group variances for our two-way ANOVA’s as well. In its origin, the variable FTE’s was non-normally distributed. Transformations of our non-normal variables by making use of the function LN (Natural Logarithm) showed a positive outcome for the variable FTE’s. Making use of the Natural Logarithm of FTE’s allows us to consider this variable as normally distributed data. The other non-normally distributed variables still have to be considered as non-normally distributed.
used to test for differences between pairs of levels and to test for differences between companies that have the intention and those that do not have the intention to further adopt the web.
Perceived benefits
Market Organizational Complexity Compatibility
Management support
IT
expertise Turnover IT budget FTE's
Valid 203 203 203 203 203 203 181 180 186 Missing 0 0 0 0 0 0 22 23 17 Skewness -1,136 -,234 -,096 -,153 -,311 -,684 ,962 1,554 ,231 Std. Error of Skewness ,171 ,171 ,171 ,171 ,171 ,171 ,181 ,181 ,178 Std. Error of Skewness * 2 ,341 ,341 ,341 ,341 ,341 ,341 ,361 ,362 ,356 Kurtosis 3,005 ,119 -,512 -,191 -,501 1,006 -,217 2,480 ,493 Std. Error of Kurtosis ,340 3,40 ,340 ,340 ,340 ,340 ,359 ,360 ,355 Std. Error of Kurtosis * 2 ,679 ,679 ,679 ,679 ,679 ,679 ,719 ,720 ,709
Normal or non-normal non normal normal normal normal non non non normal
Table 9: Skewness and Kurtosis of the dependent variables.
5.1 Analysis step one
The purpose of our first step of analysis is to determine whether there are differences on the scores of the main factors on web adoption across the various levels of web adoption and to examine between which pairs of levels these differences exist. In order to determine differences on the main factors of web adoption across the various levels one-way ANOVA is used. In order to examine the differences between pairs of levels we conducted a post-hoc test. The results of our one-way ANOVA and the post-post-hoc tests are presented in table 10.
M ain factors One-w ay Anova Post-hoc: Bonferroni
F P-value Level 1 - 2 Level 1 - 3 Level 2 - 3 Perc eived benefits
-Serving the m arket 1,079 0,342 1,000 0,648 0,476
-Organizational efficiency 5,174 0,006** 0,340 0,395 0,006**
Com plexity 4,697 0,010* 0,014* 0,045* 1,000
Com patibility 4,438 0,013* 0,365 0,010* 0,131
Managem ent support 14,171 0,000** 0,008** 0,000** 0,003**
IT expertis e 8,077 0,000** 0,035* 0,000** 0,066
Turnover 1,677 0,190 0,469 0,267 1,000
IT budget 3,858 0,023* 0,279 0,020* 0,263
FTE's 2,695 0,070 1,000 0,112 0,102
* = sig at 0,05 level, ** = sig at 0,01 level
P-values
The results of our one-way ANOVA confirms our expectation that differences on the various levels of web adoption exist for organizational efficiency benefits, complexity, compatibility management support, IT expertise and IT budget. We will first discuss the results on our innovation characteristics followed by a discussion of the results on our factors on firm level.
5.1.1 Innovation characteristics
No significant differences are found for the relationship between perceived benefits with regard to serving the market and the various levels of web adoptions. A quick look at figure 4, a chart that contains the mean scores on our innovation characteristics, reveals that this factor scores considerably higher than the other innovation characteristics. This implies that companies perceive high benefits from the web with regard to serving the market. However, no significant differences exist across the various levels of web adoption.
Further examination of the relationship between perceived organizational efficiency benefits and the various levels of web adoption shows that the scores on organizational efficiency differ significantly between level 2 and level 3 (p-value: 0,006). Figure 4 shows a positive relationship between these levels. This implies that companies at level 3 perceive more benefits than companies at level 2. A remarkable detail is that the mean score on level 2 is lower than the mean score on level 1. A possible explanation could be that companies expect to gain more efficiencies from level 2 than from level 1, while actual organizational efficiencies are gained at level 3 where integration with the web takes place. This might have had a negative effect on the score for perceived organizational efficiency benefits.
Further examination of the relationship between complexity and the various levels of web adoption shows that the scores on complexity differ significantly between level 1 and level 2 (p-value: 0,014) and between level 1 and level 3 (p-value: 0,045). Figure 4 indicates a negative trend for complexity, which is in line with our expectation. The results imply that companies at level 2 and 3 perceive the web as less complex than companies at level 1. Although, figure 4 also reveals that the scores on complexity are considerably lower that the scores on the other innovation characteristics. This might indicate that companies do not perceive the web as complex at all.
Further examination of the relationship between compatibility and the various levels of web adoption shows that the scores on compatibility differ significantly between level 1 and 3 (p-value: 0,010). Figure 4 indicates a positive relationship for compatibility. This implies that companies at level 3 perceive their organization to be more compatible with the web than companies at level 1.
5.1.2 Factors on firm level
differ across the various levels of web adoption with a confidence level of 95 % (p-value: 0,023). Firm size measured in numbers of FTE’s (p-value: 0,070) and firm size measured in terms of turnover (p-value: 0,190) are not found to differ across the various levels significantly. Figure 5 will help us to examine if the relationships are positive or negative.
1,0000 2,0000 3,0000 4,0000 5,0000 1 2 3 Level S c o re Management support IT expertise Turnover IT budget FTE's
Figure 5: Mean scores on the factors on market level.
Further examination of the factor management support shows that this factor differs between each pair of levels significantly at a confidence level of 99 %. Figure 5 indicates a positive relationship, which is in line with our expectation. This implies that companies at the higher levels of web adoption have a more supportive management than companies at the lower levels of web adoption.
Further examination of the factor IT expertise shows that this factor differs between level 1 and 2 (p-value: 0,035) and between level 1 and 3 (p-value: 0,000). Figure 5 indicates a positive relationship for IT expertise, which is in line with our expectation. These results imply that companies at level 3 and level 2 possess significantly more knowledge than companies at level 1.
Further examination of the factors turnover and FTE’s shows that these factors do not differ significantly between any pair of levels, which is in line with the results on our one-way ANOVA. Tough, figure 5 shows a positive trend for turnover and FTE’s, which indicates that companies at the higher levels might have a higher turnover and more FTE’s.
5.2 Analysis step two
The purpose of our second step of analysis is to determine whether there are differences on the scores of our main factors between companies that have the intention to further adopt the web and those that do not have the intention to further adopt the web. This relationship will be controlled for the current level of web adoption. In order to determine differences on the main factors of web adoption we have conducted a two-way ANOVA with the fixed factors (i.e. grouping variables) intention to further adopt and level of web adoption. In addition, we have used partial correlation to determine if there is correlation between the main factors of web adoption and the intention to further adopt the web to verify our findings. The correlations are also controlled for the current level of web adoption. The results on our two-way ANOVA’s and the correlations can be found in table 11.
Main factors Two-way ANOVA Intention * Level Partial Correlation
F P-value F P-value Correlation P-value
Perceived benefits
-Serving the market 8,054 0,005** 1,330 0,267 0,376 0,000**
-Organizational efficiency 5,194 0,024* 1,690 0,187 0,302 0,000** Complexity 0,249 0,618 0,041 0,959 -0,014 0,420 Compatibility 5,211 0,024* 2,523 0,083 0,305 0,000** Management support 17,780 0,000** 0,637 0,511 0,371 0,000** IT expertise 8,846 0,003** 0,718 0,489 0,344 0,000** Turnover 3,053 0,082 0,058 0,944 0,197 0,004** IT budget 8,311 0,004** 1,296 0,276 0,174 0,011* FTE's 0,058 0,810 0,006 0,994 0,036 0,318
* = sig at 0,05 level, ** = sig at 0,01 level
I II III
Table 11: Results on the two-way ANOVA’s and the partial correlations.
Our two-way ANOVA results do show significant differences on the mean scores between companies that have the intention to further adopt the web and those that do not have the intention to further adopt the web for the main factors: perceived benefits with regard to serving the market, perceived organizational efficiency benefits, compatibility, management support, IT expertise and IT-budget (table 11, column I). First, we will discuss the results on the main innovation characteristics followed by a discussion on the results on the main factors on firm level.
5.2.1 Innovation characteristics
Both our two-way ANOVA results and the results for partial correlation show that perceived benefits with regard to serving the market are positively related to the intention to further adopt the web with a confidence level of 99 % (p-values: 0,005 and 0,000). This implies that companies that have the intention to further adopt the web perceive more benefits from the web than those that do not have the intention to further adopt the web.
Both our two-way ANOVA results and the results for partial correlation show that perceived organizational efficiency benefits are positively related to the intention to further adopt the web (p-values: 0,024 and 0,000). This implies that companies that have the intention to further adopt the web perceive more benefits from the web than those that do not have the intention to further adopt the web.
Analysis of our results for complexity reveals that complexity is no significant factor for further adoption of the web. The p-value for our two-way ANOVA is 0,618 and the p-value for partial correlation with the intention to further adopt is 0,420. These are far from significant. Considering the mean scores on web adoption on both the intention to further adopt the web and the various levels of web adoption reveals that these scores are relatively low compared to our other main factors of web adoption. The mean scores are smaller than 3 which indicates that companies do not perceive the web to be complex.
5.2.2 Factors on firm level
Both our two-way ANOVA results and the results for partial correlation show that management support and IT expertise are positively related to the intention to further adopt the web with a confidence level of 99% (p-values < 0,003). This implies that companies that have the intention to further adopt the web have a more supportive management and possess significantly more IT expertise than companies that do not have the intention to further adopt the web.
Our two-way ANOVA results indicate no significant differences for turnover but the results on partial correlation do show significant differences with regard to turnover. The p-values for turnover on the two-way ANOVA and the test for correlation are respectively 0,082 and 0,004. The p-value on our two way ANOVA indicates that there are some, but no significant differences with regard to turnover at a confidence level of at least 95 %. In this case we will rely on the results of our two-way ANOVA test because our data with regard to turnover are ordinal and partial correlation assumes its data to be interval. This implies that there are no significant differences with regard to turnover between companies that have the intention to further adopt the web and those that do not have the intention to further adopt the web.
Both our two-way ANOVA results and the results for partial correlation show that budget is positively related to the intention to further adopt the web. The p-values for IT-budget on our two-way ANOVA test and our test for correlation are respectively 0,004 and 0,011. This implies that companies that have the intention to further adopt the web have assigned a higher IT-budget than those that do not have the intention to further adopt the web.
5.3 Non-parametric test results
The non-parametric tests are conducted for the main factors of web adoption that are non-normally distributed and/or have unequal group variances. These factors are perceived benefits, IT expertise and firm size in numbers of turnover, FTE’s and IT-budget. The results on our non-parametric tests can be found in appendix E. A summary of these results can be found in table 12.
Level Wilcoxon rank sum Intention
Main factors Kruskal Wallis Level 1 - 2 Level 1 - 3 Level 2 - 3 Wilcoxon rank sum
Chi-square P-value W P-value W P-value W P-value W P-value
Perceived benefits
-Serving the market 2,975 0,226 3972,000 0,640 1964,000 0,091 8104,000 0,149 10657,000 0,002** IT expertise 16,660 0,000** 3421,000 0,013* 1699,000 0,000** 7895,500 0,016* 10792,500 0,005**
turnover 3,562 0,168 3221,000 0,243 1521,500 0,052 6750,000 0,279 8609,000 0,044*
IT budget 6,742 0,034* 3092,500 0,082 1486,500 0,012* 6581,000 0,143 8799,000 0,011* * = sig at 0,05 level, ** = sig at 0,01 level
I II III
Table 12: Results on the non-parametric tests.
5.3.1 Analysis step one
The results on our Kruskal-Wallis test (table 12, column I) show that the scores on IT expertise (p-value: 0,000) and IT-budget (p-value: 0,034) differ significantly across the various levels of web adoption. Perceived benefits with regard to serving the market (p-value: 0,226) and firm size measured in terms of turnover (p-(p-value: 0,168) do not differ significantly across the various levels of web adoption These findings are in line with the results on our one-way ANOVA.
5.3.2 Analysis step two
The results on our Wilcoxon rank sum test (table 12, column III) show that the scores on perceived benefits with regard to serving the market value: 0,002), IT expertise (p-value: 0,005) and firm size measured in terms of turnover (p-(p-value: 0,044) and IT-budget (p-value: 0,011) differ significantly between companies that have the intention to further adopt the web and those that do not have the intention to further adopt the web. These results are in line with the results on our two-way ANOVA, except for the score on turnover. The results on our two-way ANOVA show that turnover does not differ significantly between companies that have the intention to further adopt the web and those that do not have the intention to further adopt the web. This dissimilarity can be caused by the fact that our two-way ANOVA controls for the effect of the current level of web adoption and the fact that our two-way ANOVA assumes the factor turnover to be normally distributed, while it is non-normally distributed. One way or the other, we do not consider turnover to differ significantly since we do not have the evidence to state with certainty that turnover differs significantly between companies that have the intention to further adopt the web and those that do not have the intention to further adopt the web. Although, the results indicate that turnover seems to be related to the intention to further adopt to some extent.
The main research question of our study is: which kind of companies establish which levels of web adoption. In order to answer this question we have selected a model for web adoption and we have identified and discussed the main factors of web adoption. This discussion led to several expectations with regard to the relationship between the main factors and the various levels of web adoption and the relationship between the main factors and the intention to further adopt the web, which we have tested by making use of a data-set acquired by a former student.
perceived organizational efficiency benefits, does differ between the various levels of web adoption as well as between companies that have the intention to further adopt the web and those that do not have the intention to further adopt the web. At last, the scores on the main factor complexity are rather low, what indicates that the web is not perceived as complex.
The results of our study are summarized in figure 6 and figure 7. The + and - signs indicate whether the scores on the main factors increase or decrease. The model presented in figure 6 presents the main factors that significantly differ between the various levels of web adoption. All main factors, presented in figure 6, except for complexity, show a positive trend which means that the scores on these factors are significantly higher at the higher level(s) of web adoption. This lets us conclude that companies differ across the various levels of web adoption on the degree to which they perceive the web to be complex, the degree to which they perceive organizational efficiency benefits, the supportiveness of their management, the degree to which they posses IT expertise and the size of their IT budget. Level 1 - Complexity + Management support + IT expertise + Perceived organizational efficiency benefits
+ Management support Level 2
- Complexity
+ Management support + IT expertise
+ IT budget
Level 3
No intention
Perceived benefits:
+ organizational efficiency + serving the market + Compatibility
+ Management support + IT expertise
+ IT budget
Intention
Figure 7: The main factors that are found to differ between intention and no intention.
The adoption of the web is discussed widely in current literature. However, the larger part of research focuses on the situation at a certain moment and only distinguishes between adopters and non-adopters or companies at the basic level of web adoption and companies at the advanced level of web adoption. Our study incorporates a model for web adoption that distinguishes between three distinctive levels of web adoption, which correspond with the notion that the web can be seen as an innovation that can be adopted at multiple levels. Furthermore, our study incorporates a more dynamic concept, the concept of companies that have the intention to further adopt the web over time. Further research should incorporate our model or a refined model for web adoption to address the multiple levels of web adoption that can be established and should incorporate a more dynamic concept. After all, the extent to which companies are using the web is still increasing and new opportunities offered through the web will keep evolving.
A more explicit recommendation for further research concerns the way that the main factors of web adoption influence each other. We expect that analysis of the relationships between the main factors will reveal that most factors are correlated. It would be interesting to examine how these factors influence each other and to find out which factor or combination of factors actually lead to further adoption of the web. Such a study could propose a more effective way to stimulate the adoption of the web.
companies. Adopting the web to a lower level might already increase a companies market reach while efficiency benefits seem to be achieved at our highest level of web adoption. Increasing a companies market share or increasing a companies efficiency benefits may secure or improve a companies competitiveness.
Third parties that promote the adoption of the web should consider the various levels of web adoption that can be established. The adoption of the web should be seen as a process that takes place over time. A company can adopt the web to a higher extent over time or possibly take a step down. Some companies will only benefit from adopting the web by simply giving presence while other companies might benefit from adopting the web to the highest level where full integration takes place.
The factors that are found to differ between the various levels of web adoption are also found to influence a companies intention to further adopt the web. The most important factors are IT expertise and management support. Promoting further adoption of the web starts with informing a company’s management about the opportunities that are offered by the web. A companies management has to get familiar with the web which means that IT expertise with regard to web adoption needs to be available. A management that is familiar with the web will be more supportive towards the adoption of a higher level which will positively influence the adoption of the web.
The first limitation concerns our data-set. Our data-set dates from 2005, companies might have established higher levels of web adoption over the last four years. Nevertheless, the differences on the scores on the main factors of web adoption across the various levels and on the intention to further adopt are expected to be comparable to the current situation.
The second limitation concerns our external factor competitive pressure. Our data set did not contain any data with regard to competitive pressure. Therefore, we had to omit the factor competitive pressure in the remaining of our study.
The third limitation concerns our multi-item scales and is more of a suggestion. We have used the SPSS function MEAN to construct our variables. This function simply calculates the mean score on the various items for each of the respondents and is used more over. Factor analysis may be considered next time to construct a (weighted) factor scale for our variables. It would be interesting to find out if this will lead to differences on our test results.
Our fourth limitation concerns the measurements for perceived benefits and IT expertise in the survey. These measurements were slightly different from the way we would have constructed these measurements as we based these measurements on the discussion on these factors in the literature review. Therefore, we had to make some assumptions which we have discussed in our research design.
