Profiling the non-user: Rethinking policy initiatives stimulating ICT acceptance

ARTICLE IN PRESS Telecommunications Policy 33 (2009) 642–652

Contents lists available at ScienceDirect

Telecommunications Policy URL: www.elsevierbusinessandmanagement.com/locate/telpol

Profiling the non-user: Rethinking policy initiatives stimulating ICT acceptance Pieter Verdegem a,, Pascal Verhoest b a b

Research Group MICT—Ghent University (UGent)—Interdisciplinary Institute for Broadband Technology (IBBT), Korte Meer 7/9/11, B-9000 Ghent, Belgium The Federal Agency for Information and Communication Technology (Fedict)/Free University of Brussels (VUB), Brussels, Belgium

a r t i c l e i n f o

Keywords: Non-users Policy initiatives ICT acceptance E-inclusion User research PC and internet penetration ICT literacy

abstract Business strategies and policies that were successful in increasing internet penetration in the early days may no longer be appropriate. This is most probably so in countries where a bigger proportion of the population is already connected to the internet. As more people are online, it becomes more likely that the remaining fraction of non-users is either hard to convince, under-skilled or simply lacking the financial resources to afford a connection. In view of this, a new policy approach is proposed to increase ICT acceptance. The approach is based on strategies of segmentation and differentiation. This entails that policy initiatives are specifically targeted towards different groups in the population. This article demonstrates that being a non-user can be explained by a combination of access problems, lack of ICT skills or rather negative attitudes towards ICT or by the outweighing effect of one of them. It also provides a framework for setting up new policy measures. & 2009 Elsevier Ltd. All rights reserved.

1. Introduction The pervasiveness of ICT in society and the increasing dependency on ICT in everyday life makes the capacity to use ICT at home a more important condition for social participation (van Dijk, 2005; Warschauer, 2003). The goal of any e-inclusion policy should therefore be to achieve meaningful internet access for all. This will require a continuous effort on behalf of policy makers (Milner, 2006, pp. 178–179). Indeed, inequalities in ICT adoption and usage are not likely to diminish or disappear of their own accord (Selwyn & Facer, 2007, p. 25). In societies that have already reached higher levels of internet adoption,1 increasing internet penetration may require specific measures that differ from those of the early days of the internet. The fraction of remaining non-users may be structurally lacking financial resources to afford a connection (Martin & Robinson, 2007), they may be poorly educated or under-skilled (Livingstone, 2004) or they may be hard to convince to use ICT because of emotional reasons (e.g. technophobia, van Dijk, 2005, pp. 35–44) or simply because they resent using it (Selwyn, 2006). Bearing this in mind, a research track for new policy initiatives concerning ICT acceptance is set up. The approach is born out of a confrontation of theory with political practice. This affects the way in which the research is organized and conducted. The approach is characterized by two main features. Firstly, unlike many e-inclusion policies the approach

 Corresponding author. Tel.: þ32 9 264 84 77; fax: þ32 9 264 69 92.

E-mail addresses: [email protected] (P. Verdegem), [email protected] (P. Verhoest). The research is conducted in Belgium. In 2007 67% of households (with at least one person between 16 and 74 years old) in Belgium owned at least one computer, while 60% of households had internet access at home (FOD Economie, 2007). The information society is thus more developed than on average in the EU; however, Belgium is not one of the frontrunners. 1

0308-5961/$ - see front matter & 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.telpol.2009.08.009

ARTICLE IN PRESS P. Verdegem, P. Verhoest / Telecommunications Policy 33 (2009) 642–652

643

proposed does not only aims at removing barriers but equally, or alternatively, at increasing the value of ICT for end-users. Indeed, a starting assumption is that the (perceived) added value of using ICT products may have a decisive influence on actual usage. Secondly, the input for new policy measures developed within this approach is specifically targeted towards different segments of the population, taking into account the specificities of characteristics of different groups of non-users. Therefore, it may be assumed that policy initiatives based on strategies of group segmentation and differentiation might be more effective and less expensive than generic policy measures. Prerequisites for these kinds of policies, however, are that these groups are relatively homogeneous (especially with regard to determinants that are of major influence on ICT appropriation) and that they can be easily targeted and reached by policy makers (preferably groups with some form of formal organization and representation). In the first part of this article the considerations that have inspired the research approach are explained and contextualized. In the second part, the research outline and its methodological base are explained. The third and main part describes the most important findings of the research and evaluates their significance and illustrates how the findings can be transposed into concrete policy measures. 2. Research outline 2.1. Field experience with stimulating ICT acceptance The proposed approach is mainly inspired by the ‘Internet for All’ campaign of the Belgian government in 2006. This initiative consisted of providing an affordable package deal to potential buyers, comprising a personal computer (PC), an internet connection plus a training session (OECD, 2008, p. 198). The main ‘political’ difficulty seemed to be to convince the industry (PC manufacturers, ISPs and retailers) to participate in this campaign. The main resistance was from the organization of small retailers, who feared the low profit margins would cause an unacceptable loss of income. Nevertheless, three consortia, including well-known PC manufacturers and ISPs, offered a package. On evaluation, the Internet for All campaign proved to be advantageous for the retailers as well as for other parties involved. It has been calculated that the project contributed 16% of the increase of new internet connections over a period of 1 year. The slipstream of the project is estimated to be 50%. The slipstream consists of buyers who were initially interested in the package but eventually opted for another (generally more expensive) commercial offering. The sum total is that the project contributed to almost a quarter of the increase of internet connections between March 2006 and March 2007. A critical evaluation of the Internet for All campaign revealed different elements, two of which inspired the research that is presented within this article. The first was merely confirmation of what could be expected. Not all of the groups in society were equally well served by the offer. Some buyers preferred to buy better-performing and more-expensive equipment whilst for other people the packages were too expensive, either because the up-front entry cost was too high or because of the recurrent costs of an internet connection. The second source of inspiration was an incidental call of a representative of a professional organization for physical therapists who proposed to target the campaign also towards members of his organization. These two apparently banal observations triggered a reflection that inspired the new policy approach and adjoining research. 2.2. Traditional parameters of digital inequalities The concept ‘digital divide’ has become very popular in the last decade, both in scholarly studies as well as in political speeches and in the popular press and media. Gunkel (2003, p. 500) states that this attention appears to be ‘‘an obvious advance over the euphoric ‘cyberbole’ that characterized much of the rhetoric of computer technology since the mid1980s.’’ Overcoming digital inequalities is now considered to be one of the key drivers for social and economic welfare, in order to improve social participation and to increase competitiveness and productivity (Brants & Frissen, 2005, pp. 22–23). Nevertheless, innovation studies have traditionally paid relatively little attention to phenomena such as non-use of ICT (Selwyn, 2004, pp. 342–343; Wyatt, 2005, pp. 77–78). Although the metaphor of the digital divide appeared to be very successful in putting the issue on the agenda of social, political and scholarly discussion, it is a simplification of reality and thus risks introducing several misunderstandings (van Dijk & Hacker, 2003). More specifically, the dichotomous portrayal (the divide between ‘haves’ versus ‘have-nots’ or between ‘technologically rich’ versus ‘poor’) is no longer tenable, as these conceptualizations are too limited and rudimentary in analysis (Barzilai-Nahon, 2006, p. 270; Norris, 2001, pp. 3–4; Selwyn, 2004, pp. 343–346; van Dijk, 2006, pp. 221–222; Warschauer, 2003, pp. 11–12). Research in this context often focuses on ‘traditional’ parameters of digital inequalities. A wealth of scientific studies exists that scrutinizes the most important variables for explaining these traditional inequalities. Income or socio-economic status remains one of the most important factors in explaining differences in ICT adoption and usage (Martin & Robinson, 2007, p. 2). Even in highly industrialized societies, lower levels of income are consistently shown to be associated with ICT inequalities (Fuchs, 2009; Selwyn, 2004, pp. 344–345; Warschauer, 2003, pp. 42–54).

ARTICLE IN PRESS 644

P. Verdegem, P. Verhoest / Telecommunications Policy 33 (2009) 642–652

Other dimensions for explaining different levels of engagement with ICT are:

 gender (men having more access and using ICT more than women, although recent research indicates declining gender differences; Selwyn, 2006, pp. 279–283; van Dijk & Hacker, 2003, pp. 319–320);

 age (increased age is associated with lower levels of access, limited modes of use and patterns of connecting; Roe & Broos, 2005, p. 92; van Dijk & Hacker, 2003, pp. 317–318);

 education (lower levels of education seem to be corresponding with divides related to access and use of a range of ICT; Roe & Broos, 2005, p. 94; Servon, 2002, p. 33);

 family structure (the presence of school-age children tends to increase contacts with ICT; Kennedy, Wellman, & Klement, 2003, pp. 87–88; Van Rompaey, Struyf, & Roe, 2002). In addition to these variables there are others such as race (Kvasny, 2005), geography/rural–urban location (Warren, 2007), cultural/social participation (Sassi, 2005), etc. that determine access to and usage of ICT. A thorough understanding of these parameters and their mutual dependencies must be the keystone of any e-inclusion policy. Over the years, these (traditional) socio-demographic parameters have been used, on the basis of some evidence, to explain differences in ICT adoption and use. Many scholarly publications aimed to describe and thoroughly explain digital ¨ inequalities among individuals and groups (Rice & Katz, 2003; Vehovar, Sicherl, Husing, & Dolnicar, 2006; Yu, 2006). In order to set up effective measures in support of e-inclusion, however, these parameters are no longer sufficient. So, new and creative theoretical frames of mind are needed, especially in societies that already have higher levels of computer and internet penetration. Hence, equal attention should be given to non-usage of ICT. ‘‘Analyzing users is important, but by focusing on users and producers we run the risk of accepting a worldview in which adoption of new technology is the norm,’’ as Wyatt (2005, pp. 77–78) states. In addition, insights into these distinctive profiles of non-users should be linked to appropriate policy measures.

2.3. Utility in ICT acceptance: towards an alternative interpretation Utility is a concept that is often used in technology adoption studies. Rogers (2003, p. 265) incorporates ‘relative advantage’ (‘‘the degree to which an innovation is perceived as better than the idea it supersedes’’) as one of the innovation characteristics in his Diffusion of Innovations theory. Davis (1989, p. 320) places ‘perceived usefulness’ (‘‘the degree to which a person believes that using a particular system would enhance his/her job performance’’) at the centre of his TAM—Technology Acceptance Model—in which perceived usefulness and perceived ease of use are the most important determinants of an individual’s adoption decision. According to the Social Cognitive Theory—a theory from social psychology (Bandura, 1986)—‘outcome expectations’ (‘‘the perceived likely consequences of using ICT’’) (Compeau, Higgins, & Huff, 1999, p. 147) is a central concept together with computer/internet self-efficacy. Here the interpretation of the utility concept differs from these definitions. Firstly, utility is defined as all perceived benefits of ICT. Secondly, and most importantly, the perceived utility of using ICT is looked at in relation to the perceived cost for using it. Hence, the notion of ‘relative utility’ is preferred. In other words, the relative utility of a product is the perceived increase of utility obtained by appropriating a product in relation to all emotional, cognitive and material resources available to an individual. Based on these ideas, it becomes possible to determine a hypothetical ‘turning point’ for ICT appropriation, namely the point at which the benefits will outweigh the costs for appropriating an ICT product. The notion of ‘cost’ is extended to any effort needed for ICT adoption, which is not only money but also, for example, the time required to acquire skills. By exploring this turning point for certain categories of non-users, policy makers would be able to focus their efforts to stimulate ICT acceptance. Consequently, the balance between perceived benefits and perceived costs is likely to differ for different categories of individuals. This is based on the assumption that benefits and costs are similar for homogeneous socio-demographic and socio-economic groups. Homogeneity, in this context, means that people share the same characteristics in terms of the most important resources that determine the use of ICT: access, skills and attitudes (ASA).

ACCESS

ICT APPROPRIATION

SKILLS

ATTITUDES

Fig. 1. The ASA approach.

ARTICLE IN PRESS P. Verdegem, P. Verhoest / Telecommunications Policy 33 (2009) 642–652

645

A specific combination of conditions in terms of access to ICT, skills to master the devices and attitudes towards the technology is called an ‘ASA profile’ (Fig. 1). On a practical level, in order to set up effective e-inclusion measures, the advantage of this method is that groups of individuals with relatively homogeneous ASA profiles can easily be identified and reached by policy makers. Very often they are represented by professional or social organizations that know how to reach them and are willing to collaborate with government. A specific offering can then be proposed to these groups, taking into account the specificities of their ASA profile and their socio-economic background. Additional advantage of the group-based approach is that measures can incorporate the influence of the social network of individuals. The latter is of major importance when it comes to ICT acceptance (Dutta-Bergman, 2005; Frissen, 2000). 2.4. Methodological base The proposed approach is based on a research project comprising three consecutive research stages. Phase I is aimed at refining the assumption that members of homogeneous socio-demographic and socio-economic groups share similar ASA profiles. It consists of a quantitative survey designed to gain insight into the perceptions of ASA by groups of individuals with shared socio-demographic and socio-economic characteristics. Phase II of the project consists of qualitative in-depth and focus group interviews with respondents from each group. The main objective of this phase is to improve the understanding of why people do not use ICT at home and to examine possible incentives to lift people over the turning point between non-usage and usage. Phase III intended to validate the findings of the previous two phases through a confrontation of respondents with policy options inferred from their profiles. The latter is also investigated via a quantitative survey. Individuals are recruited as they are members of groups in society with a certain level of organization and can be reached through a legitimate point of contact. These groups are sampled in a theoretical way, meaning that the selection of individuals is based on a limited number of characteristics, that is, variables for which literature research has shown to be of major importance for (non-)adoption of ICT (see Section 2.2). This results in certain prototypical profiles that are exemplary of the societal diversity without being representative of the overall population. In the theoretical sampling the following groups are selected: (1) single mothers with children; (2) people who just started basic computer and internet training; (3) people who manage a micro-company (in this case butchers); (4) liberal professions (in this case physical therapists); (5) lowly educated people with a technical background (in this case labourers); (6) highly skilled people with a technical education (in this case mostly with an engineering degree); (7) unemployed people; (8) people who work in the social sector (in this case nurses); (9) civil servants and (10) people who are aged 60 years and older. The decision to include these groups is based on two criteria: firstly, the groups correspond with a good (hypothetical) distribution of the main characteristics that may influence non-usage of ICT (e.g. for some groups the financial part could be the main issue while for others the lack of ICT literacy or motivation to use ICT is more of a determinant), and secondly, the individual respondents belong to a group that has some kind of formal organization (through which they can be targeted by policy makers). During Phase I, the quantitative survey, 200 individuals2 completed the questionnaire of which 184 valid questionnaires are retained (after data cleaning). Data collection is organized via personal interviews. For the in-depth (qualitative) personal interviews and focus groups of Phase II, interviews were conducted with 42 respondents in total. Overall 110 respondents returned a validly filled-in questionnaire for the validation study in Phase III. All users except for groups 2 and 6 are self-declared non-users. Nevertheless, a majority of them indicated that they have access to a computer and have an internet connection at their disposal. A preliminary conclusion is that non-use of computer and internet in the household is not only a question of access but also that usage of ICT will not automatically result from access (Brown, Venkatesh, & Bala, 2006, pp. 206–208; Selwyn, 2004, pp. 348–349). 3. Understanding the non-users 3.1. Main findings 3.1.1. ASA statements: perceptions towards ICT at home An important first goal of the research is to map the respondents’ perceptions towards computer and internet at home. To examine their perceptions, each is presented with a list of 37 statements. The list of statements is based on prior research (De Marez, 2006) in which a large-scale meta-analysis of determinants for ICT appropriation has been conducted. Starting from Rogers’ innovation characteristics (Rogers, 2003) the list is further elaborated with determinants from different studies and existing theoretical models in the field of communication, social psychology and marketing. This results in a basic list of 19 determinants, taking into account both innovation-related characteristics (e.g. complexity), and 2 Given the specific recruitment criteria and the data collection method (via personal interviews), it was a difficult task to reach a large sample size. The objective was to assemble at least 10 respondents in each group. In this non-probabilistic way, a sample population of 200 respondents was reached, making it possible to set up a mixed quantitative–qualitative research approach, within which there is a need to be cautious with generalizations.

ARTICLE IN PRESS 646

P. Verdegem, P. Verhoest / Telecommunications Policy 33 (2009) 642–652

Table 1 Results of cluster analysis (K-Means Cluster Analysis). 1

ENJOYMENT—using PC/internet at home seems funny to me SOCIAL INFLUENCE—when everyone in my social environment is using PC/internet at home, I will consider starting to use it myself COMPLEXITY—PC/internet seems user friendly to me MARKET STRATEGY—if I would buy a PC/internet, it is important to me that it is supplied by a well-known brand COST—using PC/internet at home seems expensive to me TANGIBLES—if I would consider buying a PC, its design would be an important argument for buying to me IMAGE/PRESTIGE—using PC/internet at home has a positive impact on my image and social status PRODUCT KNOWLEDGE—I consider myself well informed about the possibilities and (dis)advantages of using PC/internet COMPLEXITY—I fear that using PC/internet at home is rather complicated for me OPTIMISM—the fast technological developments are a good thing SELF-EFFICACY—I have no problem to sort out how to use PC/internet by myself SOCIAL INFLUENCE—using PC/internet at home is certainly a topic of discussion among my friends and family OBSERVABILITY—I am perfectly able to explain the strengths and weaknesses of PC/internet to others RELATIVE ADVANTAGE—I do not know when I would use PC/internet at home WILLINGNESS-TO-PAY—even if it costs a bit more, using PC/internet at home is something I really want RELIABILITY–I doubt the reliability and proper functioning of using PC/internet at home PERCEIVED RISK—if I would have to use PC/internet at home on my own, I do not think I would manage that MARKET STRATEGY—if I would consider buying a PC/internet, I would first check the ads, brochures and promotions RELATIVE ADVANTAGE—the advantages of using PC/internet at home are more important than the disadvantages for me COMPATIBILITY—if I buy a PC, it has to be a model that fits with my personality COST—using PC/internet at home will probably be too expensive for many people TRIALABILITY—I would like to try out PC/internet before buying it for myself SOCIAL INFLUENCE—I am interested in buying a PC/internet for home, but only if there are enough people in my social environment doing so SOCIAL INFLUENCE—if I use PC/internet at home, it would certainly explain something about me and my personality COMPLEXITY—I fear that using PC/internet offers different possibilities, which makes it too complicated to me INNOVATIVENESS—based on what I already know about PC/internet, I will certainly look for more information in order to buy a PC/internet OPINION LEADERSHIP—people in my environment will certainly come to me for advice concerning the usage of PC/internet RELATIVE ADVANTAGE—using PC/internet at home would make life easier for me OPTIMISM—If you do not want to run behind, adoption of new technologies is necessary VOLUNTARINESS—if I would buy a PC/internet, it would completely be my own decision. Nobody can influence that decision PERCEIVED SOCIAL RISK—if I would use PC/internet at home, people in my environment would look odd at me COMPATIBILITY—using PC/internet at home fits in my lifestyle TRIALABILITY—before buying a PC/internet, I would like the advice of some people PERCEIVED FINANCIAL RISK—I fear that using PC/internet at home would exceed my budget SOCIAL INFLUENCE—even if I am interested, I would not buy a PC/internet if my environment would be negative towards it SOCIAL INFLUENCE—most people in my environment are certainly enthusiastic about using PC/internet at home RELIABILITY—I doubt about the safety of using PC/internet at home (internet transactions and/or privacy) Ac=access, S=skills and At=attitudes, see Table 2.  Statistically significant difference (po0.01).  Statistically significant difference (po0.05).  No statistically significant difference.

Cluster 2 (N=34)

Cluster 3 Cluster 4 (N=13) (N=30)

Cluster 5 (N=68)

5

At 2.95 2.36

3.09 2.56

4.69 4.15

4.50 4.17

4.07 3.87

Atþ

S

2.85 3.21

3.68 3.50

4.23 4.69

4.20 3.90

3.99 3.99

Sþ

Acþ 3.79 2.13

3.68 2.15

3.69 3.00

4.10 2.77

4.10 3.16

Ac

1.56

1.59

3.23

2.00

2.71

2.23

3.41

2.38

4.00

3.00

Sþ

Sþ 3.90 At 3.03 S 2.03 2.41

1.79 2.97 3.00 2.41

3.77 4.23 2.38 4.54

1.30 3.97 3.93 3.20

3.37 3.79 2.85 3.28

S Atþ Sþ

S

1.44

2.56

1.85

3.67

2.46

Sþ

Atþ 2.97 2.38

2.79 2.74

1.85 4.15

1.27 3.93

2.74 3.79

At

Atþ 3.46

2.47

2.92

2.30

3.40

At

Sþ

4.33

2.03

3.38

1.57

3.34

S

3.95

2.65

3.62

2.87

3.71

At 2.72

2.88

4.00

3.57

3.37

Atþ

2.05 Acþ 3.92

1.85 3.18

1.23 4.31

2.73 3.60

3.34 4.04

Ac

4.03 2.51

2.09 1.97

2.31 1.92

2.60 1.80

3.82 3.00

1.97

1.76

2.00

2.53

3.01

3.82

1.79

3.23

1.70

3.37

S

At 2.59

2.47

3.08

3.63

3.85

Atþ

S

1.49

1.82

1.15

3.47

2.79

Sþ

At 2.62 At 3.26

3.29 2.59

4.46 4.00

4.47 3.87

4.22 4.24

Atþ Atþ

3.97

3.74

3.92

4.10

4.01

3.46

2.03

2.62

1.63

2.82

At 1.69 4.56

2.29 3.38

3.23 4.31

3.83 3.63

3.24 4.24

Atþ

Acþ 3.18

2.06

2.31

3.07

3.01

Ac

2.00

1.68

1.08

1.37

2.31

3.92

3.82

4.77

4.23

4.15

Atþ 4.00

3.29

3.00

3.00

4.00

S

Sþ

Cluster 1 (N=39)

At

ARTICLE IN PRESS P. Verdegem, P. Verhoest / Telecommunications Policy 33 (2009) 642–652

647

adopter-related characteristics (e.g. innovativeness) as the impact of the marketing strategy (De Marez, Vyncke, Berte, Schuurman, & De Moor, 2007, p. 82). These 19 determinants are translated to the specific context of computer and internet use at home and results in a list of 37 statements (see Table 1) aimed at obtaining information about the respondents’ specific ASA profile. (1) Some of the statements are aimed at examining whether individuals have positive or negative attitudes towards computer and internet at home, measured, for example, via ‘relative advantage’ (‘‘Computer and internet at home will certainly make life easier for me’’) or ‘optimism’ (‘‘Fast technological developments are a good thing’’). (2) Other statements focus on the presence or lack of skills and competences, measured, for example, via ‘complexity’ (‘‘Computer and internet at home seem userfriendly for me’’) or ‘perceived risk’ (‘‘If I would have to use computer and internet at home on my own, I do not think I would manage that’’). (3) Statements also aimed at detecting the presence or absence of barriers to access ICT, measured through ‘cost’ (‘‘Computer and internet at home seem expensive to me’’) or ‘perceived financial risk’ (‘‘I fear that using computer and internet at home would exceed my budget’’). Other statements serve as measurement scales to gain insight into the influence of more generic factors such as, for example, the influence of social networks (‘‘Before buying a PC/internet, I would like the advice of some people’’) or marketing strategies of the ICT industry (‘‘If I would consider adoption of computer and internet at home, I would first check the ads, brochures and promotions’’). Respondents are asked whether or not they could agree with these statements, measured via a five-point Likert scale varying from ‘I do not agree at all’ to ‘I fully agree’. 3.1.2. Costs and attitudes towards computer and internet Based on the mean scores of each statement, the perception of all users alike is that computers and internet are quite expensive. In addition, most respondents believe that ICT may be too expensive for a larger part of the population. The high cost of ICT infrastructure is still perceived as an important barrier for adoption of technology in households (Brown & Venkatesh, 2005). The negative perception of the price factor, however, only weakly relates to people’s attitudes towards ICT. Indeed, even though respondents are selected as non-users (at home), it can be observed that a larger number of them have positive attitudes towards ICT. They think that using computers and internet at home makes their life easier. The respondents indicate that most of the members of their social network are enthusiastic about ICT at home. Social influence plays an important role but for most respondents negative perceptions of members of their social network will not prevent them from adopting computers and internet into the household. 3.1.3. Actual skills to master the devices Responses are much more divided on skills, measured via statements such as ‘complexity’ or ‘self-efficacy’ (‘‘people’s judgments of their capabilities to organize and execute courses of action required to attain designated types of performances’’; Bandura, 1986, p. 391). Some report that they lack basic skills (which prevents them from starting to use a computer at home), whilst others can be considered as sufficiently ICT competent, for example because they (have to) use the computer at work. In addition to the perceptions of the respondents towards complexity and usability of ICT, the respondents’ actual skills are also examined. For this purpose they are shown a list of ICT-related tasks (based on Duimel & De Haan, 2007, pp. 66–78), varying from very basic (for example, sending and receiving e-mail) to very complex (for example, installing a new version of Windows). 3.1.4. Social network analysis Another part of the survey examines the influence of social network of respondents on the use of PCs and internet at home. The number of interactions with family, friends, acquaintances, colleagues and neighbours is investigated and, additionally, the ‘social resources’ that people have at their disposal within their social network are mapped out (social resources are social contacts that people can rely on to ask for advice when purchasing equipment or get assistance from in case of computer problems) (van Dijk, 2005, p. 53). Some scholars who have studied the role of social resources in ICT acceptance call them ‘warm experts’ (Bakardjieva, 2005, p. 99) or ‘local experts’ (Stewart, 2007, p. 551). In addition, attention is also given to the ‘technological culture’ of people’s social network, which is the way people deal with technological artefacts and applications in their social relations and in the everyday culture of their households (Punie, 2000, p. 558). The results of this analysis show that family is still the most important (social) determinant for the appropriation of computers and internet at home. People prefer getting help from family members for commercial advice and for troubleshooting as well as to learn new skills. The presence of these experts is not only important for the domestic use of computers and internet but also for taking full advantage of it, for example to help interpret and to make sense of the new information or services that become available (Wyatt, Henwood, Hart, & Smith, 2005, p. 211). 3.2. Quantification: further analysis An important goal of the first research stage is to test the assumption that socio-demographically and socioeconomically related respondents yield similar profiles in terms of access, skills and attitudes (ASA). Furthermore, of interest is the question—is it possible to draw-up a consistent ASA profile for people who are more connected through affiliation with a representative social (or professional) organisation?

ARTICLE IN PRESS 648

P. Verdegem, P. Verhoest / Telecommunications Policy 33 (2009) 642–652

Table 2 ASA bipolarity. Ac(cess) Ac(cess) S(kills) S(kills) At(titudes) At(titudes)

People People People People People People

þ  þ  þ 

have no problem with access to computer and internet at home have problems with access to computer and internet at home are skilled sufficiently to master the devices lack skills to master the devices have positive attitudes towards the technology have negative attitudes towards the technology

Table 3 ASA profiles.

1 2 3 4 5

LABEL

N

INCAPABLE REFUSERS SELF-CONSCIOUS INDIFFERENTS THE WILLING BUT INCAPABLE SKILLED ICT LOVERS WITH LIMITED ACCESS PRICE-SENSITIVE PRAGMATISTS

39 34 13 30 68

Table 4 Membership of ASA profiles. GROUP MEMBERSHIP

MEMBERSHIP ASA PROFILES

Civil servants Unemployed people Highly skilled technical education Single mothers with children Physical therapists PC and internet students Elderly people Butchers Nurses Lower education with technical diploma

Skilled ICT lovers with limited Skilled ICT lovers with limited Skilled ICT lovers with limited Skilled ICT lovers with limited Self-conscious indifferents Price-sensitive pragmatists Incapable refusers Incapable refusers

access access access access

The 37 statements on ASA in the quantitative survey are coded in the following way. Positive answers were attributed a plus (þ) and negative answers a minus (; see Table 1). For example, a person who fully agreed with the statement ‘‘Computers and internet are user-friendly technologies’’ is considered to provide an indication of a positive attitude and got an Atþ. This way of working (for each of the 37 statements) allowed the ability to distinguish the respondents’ answers in terms of bipolarities between Acþ, Ac, Sþ, S, Atþ and At, as illustrated in Table 2. Adding up these scores for each of the 37 statements allows the profiling of the respondents in terms of their ASA characterization. Subsequently, a K-Means Cluster Analysis (SPSS) is performed, based on the respondents’ answers on this list of 37 statements concerning their perceptions of computer and internet at home (see Table 1). The clusters demonstrate the existence of different typologies in terms of ASA profiles. Each label in this classification (Table 3) represents a specific combination of the factors investigated, in which each factor carries a different weight. Statistical testing is conclusive about the relation between the ASA profile and the group affiliation. The results of Chi-Square Test (Pearson Chi-Square) show a clear-cut relationship (statistical significance po0.01) between the membership of the groups (of the theoretical sampling) and the membership of the ASA profiles. In Table 4 an overview is presented for which groups the hypothesis could be accepted and indicates which ASA profile corresponds to a majority of the group members. As shown in Table 4, there are two cases in which the socio-professional affiliation of people does not correspond with a specific ASA profile: the nurses and the people with a lower technical education. Both groups are distributed across different clusters. The fact that none of the majority of the groups can be classified as ‘willing but incapable’ could easily be explained by the smaller number of respondents of this ASA profile. 3.3. Description of classification The goal of the qualitative research is to refine the insights into the different profiles of non-users. The aim is to understand why certain groups contain a majority of people who belong to a specific ASA profile and why other groups do

ARTICLE IN PRESS P. Verdegem, P. Verhoest / Telecommunications Policy 33 (2009) 642–652

649

not. Another aim is to examine which possible incentives could lift people over the turning point between non-usage and usage. The qualitative research allows the ASA scores of the respondents to be put into perspective, to reflect their specificity and to make suggestions as to how policy can approach these groups.

 Incapable refusers: for these people, computer prices and internet tariffs are not a major obstacle for acceptance, but they 

  

lack the skills to master ICT and they have rather negative attitudes towards ICT. Their ICT illiteracy and their lack of interest reinforce each other, which renders it difficult to persuade this group to start using ICT at home. Self-conscious indifferents: the non-use of individuals of this group can be mainly explained by their negative attitudes. Access is not a major problem and they are sufficiently skilled to deal with ICT. Since they potentially have access and know how to use ICT, these individuals could also be categorised as ‘want nots’ (Selwyn, 2004, p. 348; van Dijk & Hacker, 2003, p. 317). The willing but incapable: members of this group are motivated to use computer and internet at home, but they lack the necessary skills and they have difficulties accessing ICT. Skilled ICT lovers with limited access: these people have positive attitudes towards ICT and they are ICT literate. Consequently, their main problem resides in obtaining access to ICT at home. Price-sensitive pragmatists: individuals of this group have average ICT skills and are moderately motivated. The (perceived) high prize of infrastructure and the connection to internet seems to be the main barrier for technology appropriation in the household for them.

The descriptions still leave the question unanswered of why a generic ASA profile could not be detected for two groups? A sensible explanation could be that these groups are in fact heterogeneous in composition. This could partially account for the ‘lowly skilled with technical education’, but not for the nurses. Closer examination of these two groups during the qualitative interviews revealed that the decisions of these people not to adopt and use ICT at home is particularly strongly motivated by a lack of skills along with low expectations with regard to the added value of ICT usage. In other words, the expected benefits do not outweigh the ‘cost’ of acquiring the necessary skills. Based on the observations about these two low(er) income and education groups it is possible to conclude that the perception of costs is indeed related to the perception of utility. It can be assumed that a relatively low perception of utility will have less negative impact on persons with more resources available. The reason is that the cost of ICT appropriation takes a lower proportion out of the budget that could otherwise be spent on other utilities that may be (rightfully) perceived as more essential. On the contrary, when the perceived added value is more obvious, individuals will be more motivated to acquire that utility despite their limited resources. For policy makers, the observation that attitudes towards ICT differ strongly in groups with relatively less resources suggests that adoption may be stimulated not only by lowering costs, but also by increasing the (real or perceived) utility of ICT for these people, as this would legitimise their expenses for ICT. 3.4. Heuristic reinterpretation In Fig. 2, the 37 ASA scores are condensed, in order to obtain a heuristic instrument that allows a comparison between the different profiles in terms of their ASA characteristics. A four-point scale is used, ranging from very positive (þþ), to positive (þ) over negative () to very negative () for perceptions of problems with respect to ASA. As illustrated in Fig. 2, reflection on the qualitative findings also yields the insight that a negative perception of one of the ASA dimensions may influence the perception of other ASA dimensions. Indeed, the research shows that the three ASA pillars are mutually connected, i.e. they influence each other. More specifically, initiatives to increase skills would allow people to get a more profound understanding of what the benefits of particular applications could be. Policy makers could take advantage of

Ac(cess)

S(kills)

->+

--

-

ASA 2: SELF-CONSCIOUS INDIFFERENTS

+

+>++

--

ASA 3: THE WILLING BUT INCAPABLE

-

-

++

ASA 4: SKILLED ICT LOVERS WITH LIMITED ACCESS

-

+>++

++

->+

->+

->+

ASA 1: INCAPABLE REFUSERS

ASA 5: PRICE-SENSITIVE PRAGMATISTS

At(titudes)

---

ICT UNDER-SKILLED

---

REJECTIONISTS/REFUSNIKS

---

ECONOMICALLY LESS FAVOURED (Legend: + = positive / - = negative / > = varying from X to Y /

= influence

Fig. 2. Heuristic scheme.

ARTICLE IN PRESS 650

P. Verdegem, P. Verhoest / Telecommunications Policy 33 (2009) 642–652

these observations as purposeful measures can be set up especially with regard to enabling people to transfer over the turning point from non-usage to usage. The challenge then is to detect what the incentive to go across this turning point needs to be for a specific group. The original ASA profiles were complemented with three additional profiles of people who have extremely negative scores on one of the ASA dimensions. This addition is inspired by the case of the nurses, who could be found in all profiles, except the ‘skilled ICT lovers’. The nurses’ case thus falls under the category of ‘ICT under-skilled’. The other two categories are poorly represented in the research sample but are known to exist. ‘Rejectionists/refusniks’ are people who consciously refuse to use ICT for cultural or ideological reasons (Selwyn, 2004, p. 348). The category of ‘economically less favoured’ is self-explanatory. As is illustrated in the nurses’ case, an extremely negative score on one of the ASA dimensions may constitute a blockage, regardless of how positive the scores are on the other two categories. A targeted approach based on segmentation techniques is certainly recommend in these cases. 3.5. Exploring input for effective e-inclusion initiatives A final round of in-depth interviews with respondents offers the opportunity to explore possible scenarios that could lift them over the turning point between non-usage and usage. In other words, this research aimed at obtaining a clearer view on e-inclusion measures that are more adapted to specific conditions that apply of non-users. On the basis of these findings, specific sets of measures can be composed for each group. In order to test the validity of these findings, an extra (quantitative) research phase is set up, to which 110 respondents reply.3 This survey allowed a more accurate investigation into how different policy actions are perceived by people belonging to different profiles and whether these findings are consistent with their profiles. These in-depth interviews provided support for the group-based approach in stimulating ICT acceptance. Respondents with similar perceptions concerning computer and internet (ASA profiles) share specific expectations towards policy measures. These initiatives should take into account the insights concerning the three ASA dimensions that characterize each group. For some ASA profiles (ASA profile 4—skilled ICT lovers with limited access, ASA profile 5—price-sensitive pragmatists and ASA profile 3—the willing but incapable) e-inclusion measures should focus in the first place on offering more affordable computer and internet infrastructure. This should be combined, however, with other actions that are specific for each ASA profile: respondents of ASA profile 3 would benefit from initiatives to improve their skills, while respondents of ASA profile 5 would merely take advantage of actions to increase their perception of the utility of ICT products. Other ASA profiles (ASA profile 1—incapable refusers) can hardly detect the added value of ICT use and would therefore benefit from awareness-raising initiatives combined with ICT initiatives to familiarize them with (basic) applications (in which they then perceive added value). Increasing the perception of utility is the major challenge for members of ASA profile 2—self-conscious indifferents. This could probably be obtained through more and better communication about services that are likely to add value to their lives (e.g. more convenience and avoiding loss of time). Based on the specific preferences towards policy measures it is suggested that actions should primarily focus at increasing utility that can be offered by ICT use in everyday life. Perception of the added value, however, is dependent on the perceptions towards ASA by members of each group. As a consequence, policy actions should primarily focus on benefits that could overcome the major blockage of each segment. 4. Conclusion In this article a new policy approach for stimulating ICT acceptance is proposed. It is based on strategies of segmentation and differentiation, demonstrating and emphasizing the viewpoint that non-users may not be seen as one generic group. Empirical foundation is found for the assumption that members of homogeneous groups share similar profiles in terms of conditions to access to ICT (Ac), skills to master the devices (S) and attitudes towards the technology (At). Based on a list of technology acceptance determinants, each group can be profiled in terms of ASA characteristics. The article provides support for an alternative interpretation towards the utility notion. The relative utility of a product is seen as the perceived increase of utility obtained by appropriating a product in relation to all emotional, cognitive and material resources available to an individual. ICT adoption/usage can be explained through the relation of the perceived cost of ICT and the perceived utility. In addition, this benefits versus costs relation is determined by individuals’ perceptions towards ASA and the interplay between them. A relatively low perception of utility may have less negative effect on individuals having more resources available than on individuals with less resources at their disposal, because the costs of ICT use represent a lower proportion of individuals with more resources and therefore takes a lower proportion out of the available resources, which could be spent on other utilities. This approach is born out of a confrontation between theory and political practice. It is inspired by an alternative interpretation of the utility notion. This framework may inspire policy makers to set up creative measures stimulating ICT 3 These respondents are recruited from the sample of the first research stage. The advantage is that information is available about ASA profiles that each respondent belongs to.

ARTICLE IN PRESS P. Verdegem, P. Verhoest / Telecommunications Policy 33 (2009) 642–652

651

acceptance. Firstly, the findings suggest the need for a shift in focus of traditional inclusion away from ‘removing barriers’ towards ‘adding value’. The research indeed recommends that measures that increase the (perceived) utility of ICT usage for users may be as effective as measures that reduce the cost of using ICT. Secondly, the findings suggest that policies might be more effective and relatively less expensive if measures are not generic but targeted to specific audiences. Last but not least, the approach offers practical tools that make it possible to cluster groups into a limited number of non-user profiles comprising and combining several groups with common characteristics in terms of ASA. By compensating for their specific shortcomings, all groups in society can be equally served with a specific offering that will allow them to become ICT users at home.

Acknowledgements The authors gratefully acknowledge the support and input of the members of the advisory group of the research project of which this article is the result: Jan Deprest (Chairman of Fedict), Mila Druwe (Fedict), Bregt Bourgeus (Fluor), Patrick Slaets (Agoria), Laurence Hauttekeete (Ghent University) and Eric Van Heesvelde (Ghent University). References Bakardjieva, M. (2005). Internet society: The internet in everyday life. London: Sage. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs: Prentice-Hall. Barzilai-Nahon, K. (2006). Gaps and bits: Conceptualizing measurements for digital divide/s. The Information Society, 22(5), 267–278. Brants, K., & Frissen, V. (2005). Inclusion and exclusion in the information society. In R. Silverstone (Ed.), Media, technology and everyday life in Europe. London: Ashgate Publishing. Brown, S., & Venkatesh, V. (2005). Model of adoption of technology in households: A baseline model test and extension incorporating household life cycle. MIS Quarterly, 29(3), 399–426. Brown, S., Venkatesh, V., & Bala, H. (2006). Household technology use: Integrating life cycle and the model of adoption of technology in households. The Information Society, 22(4), 205–218. Compeau, D., Higgins, C., & Huff, S. (1999). Social cognitive theory and individual reactions to computing technology: A longitudinal study. MIS Quarterly, 23(2), 145–158. Davis, F. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340. ¨ Ph.D. dissertation, Department of De Marez, L. (2006). Diffusie van ICT-innovaties: Accurater gebruikersinzicht voor betere introductiestrategieen. Communication Studies, Ghent University, unpublished. De Marez, L., Vyncke, P., Berte, K., Schuurman, D., & De Moor, K. (2007). Adopter segments, adoption determinants and mobile marketing. Journal of Targeting, Measurement and Analysis for Marketing, 16(1), 78–95. Duimel, M., & De Haan, J. (2007). Nieuwe links in het gezin. De digitale leefwereld van tieners en de rol van hun ouders. Sociaal en Cultureel Planbureau (SCP), Den Haag. Dutta-Bergman, M. (2005). Access to the internet in the context of community participation and community satisfaction. New Media and Society, 7(1), 89–109. FOD Economie, M. (2007). Statistieken over de informatie- en communicatietechnologie (ICT), Enquˆete bij de Belgische bevolking. Brussel: FOD Economie, Algemene Directie Statistiek en Economische Informatie. Frissen, V. (2000). ICTs in the rush hour of life. The Information Society, 16(1), 65–75. Fuchs, C. (2009). The role of income inequality in a multivariate cross-national analysis of the digital divide. Social Science Computer Review, 27(1), 41–58. Gunkel, D. J. (2003). Second thoughts: Toward a critique of the digital divide. New Media and Society, 5(4), 499–522. Kennedy, T., Wellman, B., & Klement, K. (2003). Gendering the digital divide. IT and Society, 1(5), 72–96. Kvasny, L. (2005). The role of the habitus in shaping discourses about the digital divide. Journal of Computer-Mediated Communication, 10(2). Livingstone, S. (2004). Media literacy and the challenge of new information and communication technologies. The Communication Review, 7(1), 3–14. Martin, S., & Robinson, J. (2007). The income digital divide: Trends and predictions for levels of internet use. Social Problems, 54(1), 1–22. Milner, H. (2006). The digital divide: The role of political institutions in technology diffusion. Comparative Political Studies, 39(2), 176–199. Norris, P. (2001). Digital divide, civic engagement, information poverty and the internet worldwide. Cambridge: University Press. OECD, P. (2008). Belgium. OECD e-government studies. Paris: OECD Publishing. Punie, Y. (2000). Domesticatie van informatie- en communicatietechnologie. Adoptie, gebruik en betekenis van media in het dagelijks leven: continue beperking of discontinue bevrijding? Ph.D. dissertation, Free University Brussels (VUB), Brussels, unpublished. Rice, R. E., & Katz, J. E. (2003). Comparing internet and mobile phone usage: Digital divides of usage, adoption, and dropouts. Telecommunications Policy, 27(8–9), 597–623. Roe, K., & Broos, A. (2005). Marginality in the information society: The socio-demographics of computer disquietude. Communications: The European Journal of Communication, 30(1), 91–96. Rogers, E. M. (2003). Diffusion of innovations (5th ed). New York: The Free Press. Sassi, S. (2005). Cultural differentiation or social segregation? Four approaches to the digital divide. New Media and Society, 7(5), 684–700. Selwyn, N. (2004). Reconsidering political and popular understandings of the digital divide. New Media and Society, 6(3), 341–362. Selwyn, N. (2006). Digital division or digital decision? A study of non-users and low-users of computers. Poetics, 34(4–5), 273–292. Selwyn, N., & Facer, K. (2007). Beyond the digital divide: Rethinking digital inclusion for the 21st century. Opening Education series, Futurelab. Servon, L. J. (2002). Bridging the digital divide: Technology, community and public policy. Oxford: Blackwell. Stewart, J. (2007). Local experts in the domestication of information and communication technologies. Information, Communication and Society, 10(4), 547–569. van Dijk, J., & Hacker, K. (2003). The digital divide as a complex and dynamic phenomenon. The Information Society, 19(4), 315–326. van Dijk, J. (2005). The deepening divide: Inequality in the information society. Thousand Oaks, CA: Sage. van Dijk, J. (2006). Digital divide research, achievements and shortcomings. Poetics, 34(4–5), 221–235. Van Rompaey, V., Struyf, K., & Roe, K. (2002). Children’s influence on internet access at home: Adoption and use in the family context. Information, Communication and Society, 5(2), 189–206. ¨ Vehovar, V., Sicherl, P., Husing, T., & Dolnicar, V. (2006). Methodological challenges of digital divide measurements. The Information Society, 22(5), 279–290. Warren, M. (2007). The digital vicious cycle: Links between social disadvantage and digital exclusion in rural areas. Telecommunications Policy, 31(6–7), 374–388. Warschauer, M. (2003). Technology and social inclusion: Rethinking the digital divide. Cambridge: MIT Press.

ARTICLE IN PRESS 652

P. Verdegem, P. Verhoest / Telecommunications Policy 33 (2009) 642–652

Wyatt, S. (2005). Non-users also matter: The construction of users and non-users of the internet. In N. Oudshoorn, & T. Pinch (Eds.), How users matter. The co-construction of users and technology (pp. 67–79). Cambridge: MIT Press. Wyatt, S., Henwood, F., Hart, A., & Smith, J. (2005). The digital divide, health information and everyday life. New Media and Society, 7(2), 199–218. Yu, L. (2006). Understanding information inequality: Making sense of the literature of the information and digital divides. Journal of Librarianship and Information Science, 38(4), 229–252.