Parental Mediation of Internet Use and Cultural Values Across Europe:Investigating The

Total Page:16

File Type:pdf, Size:1020Kb

Parental Mediation of Internet Use and Cultural Values Across Europe:Investigating The

Parental mediation of internet use and cultural values across Europe: Investigating the predictive power of the Hofstedian paradigm

MERTENS E-Mail: [email protected]

Abstract

The EU Kids Online project aims to enhance knowledge of the experiences and practices of European children and their parents regarding online risks and safety. A crucial research effort by the EU Kids Online network has been a survey in 25 European countries which targeted approximately 1,000 children per country. This article applies a cross-cultural values filter to the data that were gathered on parental mediation and the Internet in this survey. Our intention is to test whether Geert Hofstede’s cross-national research results about national cultural values also apply to the EU Kids Online data on parental mediation. This implies studying collectivism versus individualism, low versus high power distance, masculinity versus femininity and low versus high uncertainty avoidance. We test if differences between nations on these four dimensions correlate with differences between countries in parental mediation of the Internet and we test which European countries form clusters. Key words : parental mediation, children and ICT, cultural values, Hofstede Introduction: Testing Hofstede on EU Kids Online data

Funded by the EC Safer Internet Programme, the EU Kids Online project aims to enhance knowledge of the experiences and practices of European children and their parents regarding online risks and safety, in an effort to make internet use safer for children. The evidence base consists of children’s online experiences, including the coping strategies they resort to in problematic situations. A crucial research effort by the EU Kids Online network has been a survey in 25 European countries which targeted approximately 1,000 children per countryi. In the spring and summer of 2010 the project produced new data on the self-reported attitudes and behaviors of children and parents when faced with online risks and opportunities. The results of this survey have been widely published (e.g., Livingstone, Haddon, and Görzig, 2012). This article applies a cross-cultural values filter to these results, a perspective which had been neglected despite its recognized importance (Lobe and Ólafsson, 2012, p. 275). How did we proceed? The cross-national data identify several internet-related risks and the ways children and parents react to them. With a view to international comparison, it is possible to identify specific country parameters such as the number of parents who adopt a given risk mediation strategy, how the parents’ strategy relates to their offspring’s perceptions of such a strategy, etc. These percentages can then be used as variables per case, that is, per country. This approach has been termed “ecological analysis” (Hofstede, 1995). Geert Hofstede is famous for research based on the data from a survey conducted among IBM workers throughout the world. These workers proved to have different work cultures although the company culture was internationally comparable. Hofstede (1980) found that these differences could be explained by originally four and more recently (Hofstede, Hofstede, and Minkov, 2010) six underlying dimensions, although the two dimensions added later on have been studied less widely. While this paradigm implies classifying national cultures according to a limited set of bipolar dimensions – an approach which has elicited a lot of criticism – it proved useful in explaining national differences within different institutions (Kirkman, Lowe,

1 and Gibson, 2006). Our intention is to test whether Hofstede’s cross-national research results also apply to the EU Kids Online data. In other words, we seek to assess whether the Hofstedian picture of national cultures can explain internet attitudes and behaviors of current- day European parents and children. Research on parenting styles (e.g., Claes et al., 2011) shows that depending on local cultures, parenting strategies may favor emotional bonding (a progressive educational style that promotes independence and negotiation) or behavioral control (stressing notions of duty and respect for parental authority). The parents’ position on the ‘bonding to control’ spectrum may also be linked to the Hofstedian framework. More concretely, we seek to investigate the emotional bonding vs. behavioral control orientations of the parents’ mediation strategies with respect to their offspring’s online communication (based on actual questions in the EU Kids Online survey) and to link such mediation types with the dimensions of national cultures within the Hofstedian paradigm.

Theoretical framework: Four Hofstedian dimensions and their implications for family cultures

Power distance

According to Hofstede (1995), a first, bipolar (high vs. low) dimension on which national cultures differ is ‘power distance’. This expresses the degree to which the less powerful members of a society accept and expect that power is distributed unequally. Power distance in a national culture has an impact on various institutions, including the family ecology. In a higher power distance context, children are expected to be obedient towards their parents, while in a lower power distance context they would be encouraged to develop opinions of their own.

Uncertainty avoidance

Second, the ‘uncertainty avoidance’ dimension expresses the degree to which members of a society feel uncomfortable with uncertainty and ambiguity. Consequently, family life in societies where uncertainty avoidance is strong is inherently more stressful than in societies where it is weaker, as parents in the former tend to be more fearful of the negative consequences of risks.

Masculinity and femininity

A third dimension is the ‘masculinity’ or ‘femininity’ of a national culture. Such a culture is masculine inasmuch it favors ‘achievement’, ‘heroism’, assertiveness and material reward for success, society at large being more competitive. Conversely, it is feminine when evidencing a preference for co-operation, ‘modesty’, caring for the weak, and ‘quality of life’, which boils down to a more consensus-oriented society in which family cultures are expected to set more stock by communication and co-operation.

Individualism and collectivism

The fourth dimension is the country’s individualism or collectivism: Does the organization of the country hinge on individualistic or collectivist values? This position is reflected in its people’s self-image, defined in terms of ‘I’ or ‘we’. This implies significant difference in

2 family culture, from an overpowering family collective to a family culture which gives more importance to each individual.

Interrelatedness among the four dimensions

Although suspicions of interrelatedness among the four dimensions are not entirely unfounded, these dimensions have proven to be differently distributed among countries. It should be added that cross-national studies using European data only evidence strong relations between dimensions. Therefore, in an explorative research effort such as ours, which endeavors to test these dimensions in a new empirical domain, it makes sense to identify the dimension with the highest explanatory value, and investigate whether any cluster of European countries might actually mirror Hofstede’s cultural values typology. Our study looked at 24 of the 25 EU Kids Online countries (Cyprus was not included in Hofstede’s research). Using Hofstede’s index scores (between 1 and 100) and calculating correlations between the dimensions in our sample, ‘power distance’ significantly correlated with ‘uncertainty avoidance’ (r= .78, p<.01) and collectivism (r= 59, p<.01). ‘Collectivism’ and ‘uncertainty avoidance’ were significantly correlated (r=.63, p< .01). ‘Masculinity’ did not correlate significantly with any of the three other dimensions (r= 0.03 between masculinity and power distance, r=-0.17 between masculinity and collectivism, and r=0.29 between masculinity and uncertainty avoidance), but we did include the masculinity factor in our study because it relates to one of the four ‘classic’ Hofstedian dimensions and also because it made it possible to derive expectations about family culture.

Thematic and operational framework: Types of parental internet mediation

Parental concerns

The ‘parental mediation’ theme can be linked with various questions in the EU Kids Online survey for operationalization purposes. A first group of questions is about parental concerns. These can be seen as the source of all mediation, because they underpin all of the parents’ attempts to influence their children’s media use. The EU Kids Online survey introduces seven concerns related to several life domains: children’s education, falling prey to bullying, alcohol or drug abuse, sexual activities, turning to crime, getting in trouble with the police, and road injuries. In addition, two concerns are specifically linked to children’s internet use: viewing inappropriate material, and offline contact with strangers known through the online sphere only.

Parental mediation types

Four types of parental mediation strategies concerning children’s internet use can be distinguished: Social mediation (type 1) and co-use (type 2) can be seen as stemming from the abovementioned bonding-orientated forms of parenting, while restrictive mediation (type 3) and technical mediation (type 4) are more preoccupied with controlling the child. Technical mediation is an internet-specific strategy that is considered in addition to the three classic strategies, which are more general and apply to all media (Livingstone and Helsper, 2008, p. 4). ‘ Social mediation’ consists in engaging the child in discussions about the online content she or he comes into contact with; it is measured by questions about the parents’ behavior in relation with the child’s online activities: Do they talk with the child about what she or he does online, encourage the child to explore and learn things, help when something

3 or someone has made her or him uncomfortable, explain why some websites are good or bad, help with difficulties, suggest safe behaviors and interactions? ‘Co-use’ simply means that the parent remains present while the child uses the medium. Sitting with the child while he or she uses the internet, staying nearby, and doing shared activities operationalize the ‘co-use’ dimension of parental mediation. Conceptualized as one dimension by Kirwil, Garmendia, Garitaonandia and Martínez Fernández (2009), the social mediation and co-use strategies hinge on a willingness to share in the child’s online experience. The other two, control-based strategies—restrictive and technical mediation— evidence a more hands-off approach. ‘Restrictive mediation’ involves enforcing rules (e.g., time-related) that restrict use of the medium; it is measured by questions on parental permission, supervision or prohibition with respect to various internet applications (instant messaging, downloading music or films, watching video clips, use of social networks, giving out personal information, sharing pictures, videos or music online). ‘Technical mediation’ entails setting up the computer in such a way as to minimize the inherent risks of the child’s internet use; it is investigated based on a list of eight potential strategies aiming to control the children’s internet use (checking on websites visited, supervising e-mail and instant messaging conversations, reading SNS profiles, checking out new ‘friends’ on a SNS, blocking or filtering some types of websites, keeping track of all websites visited and using software to counter spam, junk mail or viruses). The children and parents were made to answer these questions independently for comparison purposes.

Judgment of parental mediation

A similar child/parent analysis is possible for another set of questions. These are questions on the so-called ‘judgment of parental mediation’ (Livingstone, Haddon, Görzig, and Ólafsson, 2011, p. 114), referring to the evaluation of mediation activities. Parents were asked whether they thought their children’s internet experience was improved by their involvement. Another question was whether they thought they should do more. Finally, parents were asked whether they did anything differently after their child had encountered something disturbing on the internet. Children were asked to state whether they thought their parents’ involvement improved or restricted their internet experience, and whether they felt this involvement had taken on a different slant after a disturbing online encounter. Finally, a question asked the children whether they ever ignored their parents’ instructions.

Hypotheses

Hypothesis 1: Parental concerns

The Hofstedian dimension that logically ties in with questions about parental concerns is high uncertainty avoidance. This dimension is the one we most clearly expect to correlate with such questions. High power distance and collectivism—interrelated with high uncertainty avoidance—may also heighten concerns. We also assume that this phenomenon is more prominent in ‘masculine’ countries owing to lower levels of trust, which in turn can be seen as stemming from a lack of communication and co-operation.

Hypothesis 1: Concerns about children in general and their internet use in particular will correlate with the Hofstedian dimensions.

Hypothesis 2: Parental mediation types

4 Bonding-oriented parenting choices could result in social mediation and co-use, and these mediation strategies will correlate with high power distance, high uncertainty avoidance, masculinity and collectivism. Sharing the lifeworld of the children, as is practiced in social mediation and co-use, means that the child is seen as an autonomous individual (hence less focus on collectivism), with her or his own opinions and decisions (hence emphasis on less power distance), and this implies that communicative values are stressed (hence less masculinity) and that children are not as likely to be viewed as needing protection from online dangers (hence less uncertainty avoidance.) A more control-oriented approach is assumed to correlate with a preference for restrictive mediation strategies, in which protection of the children is more important than promoting their internet use. Restrictive mediation is assumed to correlate positively with high power distance, high uncertainty avoidance, masculinity and collectivism. Technical mediation is a special case. Although technical mediation implies control of the children and can be considered a dimension of control-oriented parenthood, it can also be seen as a consequence of the broader societal context and the degree to which digital technologies have been adopted in the countries under study. Earlier research by Gert Jan Hofstede (2000), Geert Hofstede’s son, on the Hofstedian dimensions in connection with the diffusion of communication technology has shown that the degree of adoption of communication technologies itself correlates with the Hofstedian dimensions. He tested this by comparing World Bank indicators on technology adoption with the Hofstedian index scores. Higher GNP per capita logically implies more financial possibilities and hence more communication technology adoption. GNP per capita is also correlated with lower collectivism and lower power distance. Nevertheless, Gert Jan Hofstede found that low collectivism and low power distance independently contribute to more advanced communication technology adoption. Still thinking in economic terms, he labelled this phenomenon ‘Western style affluence’. Furthermore, uncertainty avoidance is correlated with lower communication technology adoption, because fear of the unknown logically translates into fear of novelty (see also de Mooij, 2011, p. 288; Tsatsou, 2012). Somewhat more surprisingly, communication technology adoption correlates negatively with masculinity. Gert Jan Hofstede (2000, p. 5) speculates about the reason: “But why is the number of internet hosts negatively correlated with masculinity? This may be because in feminine countries, public services are deemed important, and internet hosts are being used in these countries as a vehicle for public services.”

Hypothesis 2: Social mediation, co-use and technical mediation will be correlated with the Hofstedian dimensions.

Finally, we seek to correlate the Hofstedian dimensions with data on the child’s and parents’ judgment on mediation. The same dimensions that were supposed to correlate with more concerns could be supposed to correlate with parents feeling the necessity to do more. They also could be correlated with the children’s perception of being limited in their internet use— and possibly seen as a consequence of parents wanting to control their children. This controlling intervention could also translate positively into a better internet experience, looked at it from both a child’s and a parent’s perspective. Yet another dimension of the urge to control, investigated once again through both parental intentions and children’s perceptions, is doing something new because of the increasing pervasiveness of the internet. Answers to the question whether children ever disobey their parents are assumed to correlate with these dimensions, disobedience being a natural product of the presence of rules.

5 Hypothesis 3: Parents’ and children’s judgments about the importance of mediation will be correlated with the Hofstedian dimensions.

Method

Correlation analysis

Hofstede used a database of answers to questions about work-related values among employees in many countries who all worked for the same multinational company: IBM. This database was not only meant to be used at the level of individuals, but also to identify differences between mean scores of countries. Further analyses of these aggregated data proved that the differences between countries could be explained by four latent variables, that is, the aforementioned four cultural dimensions. For each country an index score was calculated for each dimension. These index scores have a value between 0 and 100 for each country on each dimension. Applied to the EU Kids Online data, our analysis follows a similar motion from the individual to the country level. We calculated the mean score for each survey item within each country, testing how these mean country scores correlate with the Hofstedian index scores (to be found in Hofstede, Hofstede, and Minkov, 2011). Tables are presented with the Hofstedian dimensions on the horizontal axis, and the survey items from the EU Kids Online project on the vertical axis. The tables show Pearsons correlation coefficients, representing the degree to which the Hofstedian dimensions and the survey items are intertwined. We calculated significances at the .01, .05 and .10 as cut-off points. Adding a higher cut-off point makes sense, given the statistically small sample of European countries (N=24) (Noymer, 2008). Furthermore, Cronbach’s alpha scores were calculated to check whether different items from scales measure the different dimensions of parental mediation. If the composition of a scale is possible (i.e., when Cronbach’s alpha scores are sufficiently high), mean scores of the items that group together are calculated as well as Pearsons correlations between these mean scores and the Hofstedian dimensions.

Cluster analysis

This first step in the analysis (with EU Kids Online country data as units of analysis) allows us to see those aspects of parental mediation that correlate with the Hofstedian framework. Discussing each country in its own right would be too space-consuming, so we sought to identify groups of countries. For each hypothesis under scrutiny, the 24 countries were divided into four groups. Using four groups creates a sufficient level of analytical complexity without sacrificing too much clarity. To calculate the clusters, hierarchical cluster analysis with the furthest neighbor algorithm was adopted, indicating a preference for four clusters. The input variables for these cluster analyses are the mean scores of all the variables studied in the context of one hypothesis. We also systematically draw a comparison between these clusters and the clusters that can be created with the dimensions from Hofstede as input variables. Depending on the results of the correlation analysis we elaborate on one dimension from Hofstede per hypothesis in our cluster analysis. The dimension we use is the dimension that proved to be the most important one. Analyzing all dimensions for all hypotheses would be too space consuming.

Results

Hypothesis 1: Parental concerns in general and internet-related

6 The mean score of parental concerns about children facing general risks in life was 26.58% (27.49% for internet-related factors) across Europe. Our first hypothesis dealt with the relationships between measurements of parental concerns and the Hofstedian dimensions. The correlation matrix showed a significant relation between the general factors measured and high uncertainty avoidance (r=.42, p<.05), that is, the dimension assumed to be the most important in explaining variations in parental concerns. The two more specifically internet- related factors (concerns about online material and online contacts resulting in encounters in real life) failed to yield significance. The other Hofstedian dimensions also failed to reach significance. Table 1 presents the correlations between two scales measuring parental concerns and the Hofstedian dimensions. All items were strongly interrelated with a Cronbach’s alpha of 0.98, both for general and internet-related items.

[Table 1 inserted here]

Four-group analysis on parental concerns

Looking at the four clusters of countries, the Germanic language countries are found in the lower-ranked categories in Table 2, both for the Hofstedian uncertainty avoidance dimension and the EU Kids Online clusters. In Germanic countries concerns about internet-related risks tend to be somewhat more pronounced than general concerns (with the exception of Austria), as Germanic countries generally appear in the lower-ranked clusters regarding concerns about internet risks.

[Table 2 inserted here]

Romance-language countries, more southerly located, are to be found in the higher clusters in the table, both in the Hofstedian and the EU Kids Online clusters. Belgium, partly Germanic (Dutch-speaking) and partly Romance (French-speaking) behaves as a Romance language country. Rumania is an exception as it is found in the highest Hofstedian cluster but in the lower EU Kids Online parental concern clusters. Turkey and Greece, also part of southern Europe, are in the higher clusters, both from a Hofstedian and an EU Kids Online perspective. The Eastern European countries are generally to be found in the medium low or medium high Hofstedian clusters, but they are in the low or medium low clusters from the EU Kids parental concerns clusters. Concerns were obviously more outspoken for the Hofstedian interviewees and less outspoken for the EU Kids parents. Hungary (medium high on uncertainty avoidance and low on EU Kids general risks and internet risks) is, for instance, a very clear example. In these graphic illustrations we see maps of Europe in which lighter colors represent countries with less worries and darker colors represent countries with more worries.

[Figures 1 and 2 inserted here]

Hypothesis 2: Social mediation

Table 3 clearly shows that the social mediation scale, measured from the parents’ perspective, is consistently correlated with the Hofstedian dimensions, resulting in statistical significance in the relation between low collectivism and social mediation (r= -0.39, p<0.1) and between low uncertainty avoidance and social mediation (r=-0.38, p<0.1). When looking into the various items (not represented as such in the table) which constitute the social mediation

7 scale, more statistically significant relations appear. In less collectivist countries, parents tend to help children more in the face of a disturbing online encounter (r=-.45, p<.05). In countries with higher uncertainty avoidance (r=-.41, p<.05) and higher collectivism (r=-.46, p<0.5), parents also talk less with their children about what to do in such circumstances. Talking about internet use in general also occurs less in high uncertainty avoidance countries (r=-.37, p<0.1). Finally, in such countries parents are less likely to help their children ‘find things’ on the internet (r=-0.37, p<0.1).

[Table 3 inserted here]

When looking into the social mediation scale from a child’s perspective in Table 3, the correlations between the social mediation scale and the Hofstedian dimensions are weakly negative (r=-.11 for uncertainty avoidance and r=-.12 for collectivism) or even positive (r=.09 for power distance and r=.04 for masculinity), which contradicts our hypothesis. By taking a closer look at the items which make up the children’s account of social mediation, we see similar results contradicting the hypothesis. The correlations for ‘talking with the children’ (and also ‘encouraging children to explore things on the internet’) are in fact markedly significant in the opposite direction, compared with the hypothesis’ implied direction (Fishers r to z-transformations proved that the difference between parents and children on these two items was significant for each Hofstedian dimension). In more masculine (r=.47, p<.05), high power distance (r=.51, p<.05) and high uncertainty avoidance (r=.52, p<.01) countries, the parents talk more with their children. This remarkable exception can probably be explained by a more restrictive interpretation of their parents talking to them on the part of the children. Parents talking with their children are operationalized as a form of social mediation, however, ‘talk’ can be about what to do and/or what not to do, and can be restrictive as well. The wording ‘helping to find things on the internet’ bears high significance in a negative correlation with high power distance (r=-.52, p<.05), high uncertainty avoidance (r=-.58, p<.05), and collectivism (r=-.65, p<.01). Presumably ‘helping’ is seen as a less unambiguous translation of the underlying dimension of social mediation. When running two factor analyses, the items about talking with children and encouraging them to explore things on the internet form a separate factor in a factor analysis of the children’s data. A second factor includes all the other items, while the factor analysis with the parents’ data yields only one factor that includes all items. Logically, the Cronbach’s alpha testing whether these items could form a scale is higher (alpha = .90) in the parents’ data than in the children’s data (alpha= .79). Children tend to see social mediation in 57.95% of all cases, while parents see more of it (60.83%). This could have to do with the so-called ‘pedagogical reflexivity’, which means that the parents are more aware of mediation than their children, who are the receivers of parental pedagogical efforts. It would seem logical that this gap should be bigger in individualist countries, where children see themselves as more autonomous. This proved to be true, with the difference between parents’ and children’s accounts being correlated with individualism (r=.44, p<.01) in the case of two factors: ‘talking about the internet’ and ‘encouraging children to explore things on the internet’ (r=0.51, p<0.01).

Hypothesis 2: Co-use

Looking at the relations between the co-use scales and the Hofstedian dimensions (Table 3), a similar tendency can be seen, with weaker overall negative correlations on the children’s side except for one, in line with our earlier observation. Nevertheless, when looking at the items composing the scale (not represented as such in the table), the two correlations reaching the

8 level of statistical significance for the items on the children’s co-use account clearly confirm the hypothesis. In less masculine (r=-.45, p<.05) and low uncertainty avoidance (r=-.36, p<0.1) countries, children say their parents stay more often nearby when they use the internet. Although the parents’ account is more consistent, only two items making up the co-use scale reached the level of statistical significance: In less collectivist countries parents report they stay more often near their children when the latter are using the internet (r=-.45, p<.01) and in lower power distance countries (r=-.36, p<0.1) parents more often share online activities with their children. The mean self-reported co-use score is 51.13% from the parents and 49.19% from the children. However, this figure should be considered with some caution as the Cronbach’s alpha is slightly below the 0.7 level deemed necessary for the integration of the items in a scale (alpha=0.68 for both the children’s and parents’ data).

Hypothesis 2: Restrictive mediation

Yet another analysis perspective concerning the relationship between mediation and the internet is that of restrictive mediation. We looked at this relationship based on data about the children being allowed to do several things online. Hofstede (2001, p. 465) suggests that studies using the Hofstedian model as a paradigm should not attempt to apply all the dimensions. Rather, they should search for the most explanatory dimension. This is a difficult task to accomplish with data on social mediation and co-use, because all the Hofstedian dimensions interrelate at some point in a statistically significant way with the EU Kids Online data. The results on allowance (and hence on regulation) are more unambiguous. The allowance mediation scale correlates in a highly significant way with low masculinity (r=-.55, p<.01 for parents and r=-.54, p<.01 for children) as shown in Table 3. Taking a closer look at the items (not represented as such in the table), four applications—Instant messaging (r=-.65, p<.01 for parents and r=-.70, p<.01 for children), watching video clips on the internet (r=-.69, p<.01 for parents and r=-.71, p<.01, for children), children having a social network site profile (r=-.55, p<.01 for parents and r=-.63, p<.01 for children) and uploading pictures, videos or music (r=-.41, p<.05 for parents and r=-.45, p<.05 for children)—are significantly correlated with Hofstede’s ‘masculinity’ dimension. The parents’ and children’s’ perceptions are comparable in these cases. These items prove to be very consistent (parents’ Cronbach’s alpha=0.92 and children’s Cronbach’s alpha=0.93). One in two parents (46.52%) always allows internet use and one in two children (51.03%) think they are always allowed to use the internet. One interpretation for the difference between children and parents might be that parents are more aware of the things they forbid, and that children tend to see these restrictions as mere ‘recommendations’ they may choose to ignore to some extent.

Hypothesis 2: Technical mediation

Technical mediation, the last dimension of parental mediation, was reported in more than one in three families (37.10%) by the parents, once again probably due to higher pedagogical parental reflexivity, compared to 32.17% of the families according to their children. The items measuring technical mediation were also very consistent, resulting in a Cronbach’s alpha of 0.84 for the parents and 0.74 for the children. Table 3 (correlation of mediation strategies) shows that the technical mediation scale correlates mostly negatively with the Hofstedian dimensions, resulting in one significant relationship between the technical mediation scale and the Hofstedian dimensions: less technical mediation in less collectivist countries (both in the parents’ account, r=-.43, p<.05, and the children’s account, r= -.34, p<0.1). However, we do see in the table that masculinity

9 relates positively, albeit not significantly, with the technical mediation scale (r=.17 for parents and r=.33 for children). When looking at the items composing the scale, two different directions can be detected in the results. Some of our results confirm our hypothesis, based upon research by Geert Hofstede (2000), that technical mediation is correlated with low power distance, low uncertainty avoidance, low masculinity and low collectivism. For instance, collectivism is significantly correlated negatively with checking SNS profiles based upon the parents’ views (r=-.44, p<.05) and using parental controls or other means of blocking or filtering according to both the parents (r=.-42, p<.05) and the children (r=-.42, p<.05). We can also see that to a significant extent the use of software to counter spam, junk mail, and viruses is connected negatively with power distance (r=-.48, p<.05 in the parents’ accounts and r=-.49, p<.05 in the children’s account), uncertainty avoidance (r=-.60, p<.01 among parents and r=-.55, p<.01 among the children), and collectivism (r=-.51, p<.05 for parents vs. r=-0.48, p<0.05 for the children). However, other results prove that technical mediation can also be linked to an opposite, restricting digital family culture. Checking on the websites visited is significantly related to masculinity when it comes to both the parents’ (r=.47, p<.05) and the children’s (r=.55, p<.01) views of technical mediation, for instance. Supervising e-mail and instant messaging accounts is also correlated significantly with high power distance (parents: r=.41, p<.05; children: r=.40, p<0.1) and with high masculinity (children: r=.49, p<.05).

Four-group analysis of mediation styles

[Table 4 inserted here]

Because the relation between masculinity and allowance was the most obvious finding in our correlation analysis, we chose to use the masculinity dimension of Hofstede in our four-group analysis of mediation styles. A first clear regional group are the Scandinavian countries along with the Netherlands. These countries have high scores on the bonding types of mediation (social mediation and co-use), but are in lower clusters in the more controlling types of mediation (technical mediation and restrictive mediation). The Scandinavian countries and the Netherlands all have low scores on masculinity. Only Finland has a medium low masculinity. Another group of countries are the countries with a low score on all mediation types: Austria, Hungary, Estonia, Lithuania and Slovenia. These countries are generally in the lower EU Kids clusters, but they have very different degrees of masculinity: low masculinity in Lithuania, Slovenia and Estonia and very high masculinity in Austria and Hungary. A third group of countries are countries which are in the medium clusters of masculinity and score high on all types of mediation. These countries are Poland, France, Italy, Ireland, the UK, Greece and Spain. Clearly these are countries with very different geographical locations. Germany and Turkey are also two countries that are not geographically close to each other, but are similar in the four-group table with higher scores on restrictive social mediation and technical mediation and lower scores on social mediation and co-use. Both countries are also in the same medium high Hofstedian masculinity cluster. Finally, we see yet another group of very different countries that have overall medium scores on both the EU Kids clusters and the Hofstedian masculinity clusters. These countries are Portugal, Rumania, Belgium, Bulgary and the Czech Republic. We illustrate our four-group analysis with graphical maps, whereby higher social mediation, co-use, restrictive mediation and technical mediation are associated with darker

10 colors, and lower social mediation, co-use, restrictive mediation and technical mediation are associated with lighter colors.

[Figures 3, 4, 5 and 6 inserted here]

Hypothesis 3: Judgments on parental mediation

In total, 15.58% of the parents judged parental mediation as very important, while only 11.03% of the children did so. Once again, parents give more importance to mediation. However, these mean scores should be viewed with caution owing to the low Cronbach’s alpha values of the parental mediation judgments scale (parents: 0.55; children: 0.51).

[Table 5 inserted here]

Table 5 almost consistently shows that judgments on parental mediation are positively correlated with collectivism, masculinity, power distance and uncertainty avoidance. This clearly confirms Hypothesis 3. In five (out of 28) cases these correlations reach the level of statistical significance. Parents in collectivist (r=.58, p<.01) and high power distance (r=.39, p<0.1) countries consider it necessary to do more. Furthermore, in masculine countries, both children (r=.59, p<.01) and parents (r=.43, p<.05) consider the children’s internet experience has been improved a lot by parental intervention. Finally, and not surprisingly, in countries with a high power distance, children (r=.49, p<.05) consider their internet experience to be more limited by parental intervention.

Four-group analysis on parental judgments

Our four groups analysis uses the Hofstedian power distance dimension, because this dimension is conceptually and logically linked with the judgment of parental mediation. As a result of our four-group analysis, we logically find in the first cluster countries that are overall low in the mediation types table, that is, Lithuania, Slovenia, Austria and Hungary. Estonia is the only overall low mediation country to be found in the second, medium low cluster. Some of the more social mediation-oriented countries are also found in this first cluster, that is, Sweden, Poland, the Netherlands, Denmark and Norway, which is logical because social mediation is less intrusive than more restrictive forms of mediation. Rather unexpectedly, Belgium stands aside from the other Romance-language countries in the first cluster which view parental intervention as less important, while it is found alongside the Romance language-speaking countries in the parental concerns cluster analysis.

[Table 6 inserted here]

The parental judgment clusters follow the Hofstedian clusters inasmuch as the Romance- language countries appear at the higher end. In fact, the second cluster contains a mix of countries from Southern, Northern, Western and Eastern Europe. It should, however, be noted that Spain, Italy, Bulgaria and Greece give the third medium-high cluster a clear southern outlook, with Rumania—also a southern, Romance language country—being the only country where parental intervention is given the highest importance and also the only country with an outspoken Hofstedian power distance orientation.

Conclusion and discussion

11 This article sought to investigate whether the Hofstedian dimensions explain differences in parental mediation of internet use in Europe. While many EU Kids Online survey items failed to correlate with the Hofstedian dimensions, a number of strong effects were on display, leading to the plausible conclusion that these cultural value dimensions do play a role in parental mediation strategies. In this article we correlated country percentages from the EU Kids Online project on parental mediation with the Hofstedian index scores on cultural values. We did find that the Hofstedian value scales overall correlated with the mediation data in such a way that differences in collectivism, power distance, masculinity and uncertainty avoidance could be considered important predictors for differences in parental mediation strategies. In particular, less ‘masculine’ countries unsurprisingly proved to be more permissive with respect to their children’s internet use. Three hypotheses were presented: We found some support for all three of them. Our research on the first hypothesis showed that the very source of parental mediation—concern for the children’s wellbeing—can be explained by one Hofstedian dimension only, that is, uncertainty avoidance. Our results on the other hypotheses were less clear; the four Hofstedian dimensions had explanatory value in some cases, especially when considering the relationship between the Hofstedian dimensions and social mediation, co-use, technical mediation, and the judgment of the value of parental mediation. Earlier research (Kirwil, 2009) stressed the importance of the difference between individualist and collectivist countries in explaining the use of different internet parental mediation strategies. Our research confirmed the importance of this dimension, but the integration of the other cultural values also proves that, as stated by Kalmus and Roosalu (2011, p. 56), there are many factors behind cross-national differences in parental strategies. In only one other case, apart from the relationship between uncertainty avoidance and parental concerns, one single dimension of the Hofstedian framework proved to have a more systematic effect: that of ‘masculinity’ on restrictive mediation. In more ‘masculine’ countries, parents tend to impose more restrictions on their children’s online activities. According to Tsatsou (2011, p. 87), one can justifiably say that technological innovation and progress are more likely to happen in ‘masculine’ cultures, due to the importance given in such cultures to such values as determination, improvement, training and success. However, our research points to the opposite trend: Use of technology seems to be better accepted in more ‘feminine’ countries. This reinforces the results of other studies that also show technology adoption being more prevalent in more feminine countries (Hofstede, 2000; Nath and Murthy, 2004). Erumban and de Jong (2005, p. 19; see also de Mooij, 2011, pp. 288-289) provide a useful explanation of this duality in the relationship between online technology and masculine societies: “As is evident from the name itself, ICT can be used to communicate better. If this is the major purpose for which the technology is being adopted, then adoption rates might be higher in feminine cultures. On the other hand, ICT can also be used to compete with others: After all information is power. In that case, the masculine countries might show a higher adoption rate.” At least our data seem to support the first part of Erumban and de Jong’s explanation, namely feminine countries being more online technology friendly. Our data make it difficult to single out one decisive cultural factor in the correlation analysis, but the distribution of cultural values among European countries in the group analyses reproduces a familiar pattern. Although the hypothesis of a North-South divide explaining cultural differences has been called a myth loaded with stereotypes (e.g., O’Donnell, 1994), the evidence presented here, especially the evidence about parental concerns and the judgment of parental mediation, makes it difficult not to come to such a conclusion. The evidence about mediation types is less easily interpreted with geographical regions in mind. Another matter altogether are the reasons behind this cultural divide. Many

12 factors could contribute to the phenomenon, such as socio-economic divides (brought up, among other things, by Tsatsou, Pruulman-Vengerfeldt, and Murru, 2009), differences in welfare regimes (Kalmus and Roosalu, 2011), language differences (explained but also criticized by Zander, 2005), or religious differences (mentioned, among other elements, by Kirwil, 2009). The group analyses show the differences that exist in Europe regarding preferred parental mediation styles. These findings make it important to return to the findings of earlier studies since they lead to the conclusion that country-specific factors and cultural differences cannot be underestimated. Or to put it differently, “there is no easy, one-size-fits-all solution to parental mediation regarding children’s online risks” (Lobe, Segers, and Tsaliki, 2009, p. 181). We have shown in this article that adding a Hofstedian cultural values perspective to the country classification efforts in the EU Kids network (Helsper, Kalmus, Hasebrink, Sagvari, & de Haan, 2013) might be useful. The picture we have shown nevertheless remains a very global picture. Further research might focus on detailed comparisons of fewer countries. This would make it possible to make more detailed assessments of the differences between countries on the Hofstedian dimensions. Such an analysis was developed by Mertens and d’Haenens (2012), where differences between the Netherlands and Belgium were interpreted with EU Kids data and Hofstedian data as a frame of reference. Furthermore, the EU Kids data contain not only data about parental mediation, but also data about other aspects of children’s online behavior. These data could be analyzed from a Hofstedian perspective as well. Finally, questions about cultural values deserve to be integrated in further research projects about internet behavior in general and internet behavior of children and parents in particular.

Note

Research teams from the following countries participated in the EU Kids Online II project (2009-2011): Austria, Belgium, Bulgaria, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Lithuania, Netherlands, Norway, Poland, Portugal, Romania, Slovenia, Spain, Sweden, Turkey, UK.

References

Claes, M., Perchec, C., Miranda, D., Benoit, A., Bariaud, F., Lanz, M., Marta, E., & Lacourse, E. (2011). Adolescents’ perceptions of parental practices. A cross-national comparison of Canada, France and Italy. Journal of Adolescence, 34(2), 225-238. de Mooij, M. (2011). Consumer behavior and culture. Consequences for global marketing and advertising. 2nd edition. Los Angeles, London, New Delhi, Singapore, Washington DC: Sage. Erumban, A. A., & de Jong, S. B. (2005). Cross-country differences in ICT adoption: A consequence of culture? Research report 05C01, University of Groningen, Research Institute SOM (Systems, Organizations and Management). Helsper, E. J., Kalmus, V., Hasebrink, U., Sagvari, B., & de Haan, J. (2013). Country classification: Opportunities, risks, harm and parental mediation. LSE, London: EU Kids Online. Hofstede, G. (1980). Culture’s consequences: International differences in work-related values. Beverly Hills: Sage.

13 Hofstede, G. (1995). Multilevel research of human systems: Flowers, bouquets and gardens. Human Systems Management, 14(3), 207-217. Hofstede, G. J. (2000). The information age across countries. Proceedings of the 5ème Colloque de l’AIM. Systèmes de l’Information et Changement Organisationnel. Montpellier, France. Retrieved August 29, 2012 from http://www.oocities.org/rapportaim/pdf/Hofstede.pdf. Hofstede, G. (2001). Culture’s Consequences: Comparing values, behaviors, institutions, and organizations across nations. 2nd edition. Thousand Oaks, California: Sage Publications, Inc. Hofstede, G., Hofstede, G. J., & Minkov, M. (2010). Cultures and organizations: Software of the mind. Revised & expanded 3rd edition. New York: McGraw-Hill USA. Hofstede, G., Hofstede, G. J., & Minkov, M. (2011). Allemaal Andersdenkenden. Omgaan met cultuurverschillen. Amsterdam: Uitgeverij contact. Kalmus, V., & Roosalu, T. (2011). Parental mediation of EU Kids Online’ internet use revisited: Looking for a complex model of cross-national differences. International Journal of Media and Cultural Politics, 7(1), 55-66. Kirkman, B. L., Lowe, K. B., & Gibson, C. (2006). A quarter century of culture’s consequences: A review of the empirical research incorporating Hofstede’s cultural value framework. Journal of International Business Studies, 36(3), 285-320. Kirwil, L. (2009). Parental mediation of children’s internet use in different European countries. Journal of Children and the Media, 3(4), 394-409. Kirwil, L., Garmendia, M., Garitaonandia, C., & Martínez Fernández, G. (2009). Parental mediation. In S. Livingstone & L. Haddon (Eds.), Kids Online: Opportunities and risks for children (pp. 199-215). Bristol: The Policy Press. Livingstone, S., Haddon, L., & Görzig, A. (Eds.) (2012). Children, risk and safety on the internet: Kids Online in comparative perspective. Bristol: The Policy Press. Livingstone, S., Haddon, L., Görzig, A., & Ólafsson, K. (2011). Risks and safety on the internet: The perspective of European children. Full findings. LSE, London: EU Kids Online. Livingstone, S., & Helsper, E. (2008). Parental mediation and children’s internet use. Journal of Broadcasting & Electronic Media, 52(4), 581-599. Lobe, B., & Ólafsson, K. (2012). Similarities and differences across Europe. In S. Livingstone, L. Haddon, & A. Görzig (Eds.), Children, risk and safety on the internet: Kids Online in comparative perspective (pp. 271-282). Bristol: The Policy Press. Lobe, B., Segers, K., & Tsaliki, L. (2009). The role of parental mediation in explaining cross- national experiences of risk. In S. Livingstone & L. Haddon (Eds.), Kids Online: Opportunities and risks for children (pp. 173-186). Bristol: The Policy Press. Mertens, S., & d’Haenens, L. (2012). Culturele verschillen tussen België en Nederland en hun impact op jongeren en het internet [English translation: Cultural differences between Belgium and The Netherlands and their impact on youngsters and the internet] In L. d’Haenens L. & S. Vandoninck (Eds.), Kids Online. Kansen en risico’s van kinderen en jongeren op het internet [English translation : Kids Online. Opportunities and risks of children and youngsters on the internet.] (pp. 49-67). Ghent: Academia Press. Nath, R. & Murthy, V. (2004). A study of the relationship between internet diffusion and culture. Journal of International Technology & Information Management, 13(1&2), 123-132. Noymer, A. (2008). Alpha, significance level of test. In P. J. Lavrakas (Ed.), Encyclopedia of Survey Research Methods (p. 18). London/Thousand Oaks/New Delhi: Sage. O’Donnell, H. (1994). Mapping the mythical: A geopolitics of national sporting stereotypes. Discourse and Society, 5(3), 345-380.

14 Tsatsou, P. (2011). Digital divides in Europe: Culture, politics and the Western-Southern divide. Berlin: Peter Lang. Tsatsou, P. (2012). The role of social culture in internet adoption in Greece: Unpacking ‘I don’t want to use the internet’ and frequency of use. Information Society, 28(3), 174- 188. Tsatsou, P., Pruulmann-Vengerfeldt, P., & Murru, M. (2009). Digital divides and children in Europe. In S. Livingstone & L. Haddon (Eds.), Kids Online: Opportunities and risks for children (pp. 107-119). Bristol: The Policy Press. Zander, L. (2005). Communication and country clusters: A study of language and leadership preferences. International Studies of Management and Organization, 35(1), 83-108.

15 Figures and tables

Figure 1: General worries in Europe.

 = High general risk perception  = Medium high general risk perception  = Medium low general risk perception  = Low general risk perception

Figure 2: Internet worries in Europe.

 = High internet risk perception  = Medium high internet risk perception  = Medium low internet risk perception  = Low internet risk perception

16 Figure 3: Social mediation of the internet in Europe.

 = High social mediation  = Medium high social mediation  = Medium low social mediation  = Low social mediation

Figure 4: Co-use of the internet in Europe.

Figure 6: Restrictive mediation of the Internet

 = High co-use  = Medium high co-use  = Medium low co-use  = Low co-use

17 Figure 5: Restrictive mediation of the internet in Europe.

 = High restrictive mediation  = Medium high restrictive mediation  = Medium low restrictive mediation  = Low restrictive mediation

Figure 6: Technical mediation of the internet in Europe.

 = High technical mediation  = Medium high technical mediation  = Medium low technical mediation  = Low technical mediation

18 Table 1: Correlations between parental concerns and Hofstedian dimensions.

Dimension of parental concerns Power Uncertainty Masculinity Collectivism distance avoidance Concerns about children facing risks 0.29 0.42** 0.13 0.34 in all aspects of life (Cronbach’s alpha=0.98) Concerns about children facing risks 0.25 0.25 -0.07 0.29 on the internet (Cronbach’s alpha=0.98) (N=24; *=p<0.1 ** = p < .05; *** = p < .01)

Table 2: Clusters of countries on Hofstede’s uncertainty avoidance dimension, concerns about general and internet-related risks (as measured by the EU Kids Online survey).

Hofstedian uncertainty General risks Internet risks avoidance dimension (from low to high) (from low to high) LOW Sweden, Denmark, UK, Sweden, Slovenia, Norway, Hungary, Lithuania, the Ireland The Netherlands, Finland, Czech Republic, Estonia, Austria, Estonia, Lithuania, Austria Hungary, The Czech Republic, Rumania MEDIUM LOW Germany, Austria, Italy, UK, Germany, Poland, Rumania, the Netherlands, Norway, The Netherlands, Belgium, Denmark, France, Finland, Slovenia, Sweden, Finland, Estonia, Lithuania, Ireland, Bulgaria Denmark, Poland, Germany The Czech Republic MEDIUM HIGH France, Poland, Belgium, Italy, Greece Norway, Ireland, Bulgaria, the Spain, Bulgaria, Rumania, UK, France, Greece, Italy, Slovenia, Turkey, Hungary Belgium HIGH Greece, Portugal Spain, Turkey, Portugal Spain, Turkey, Portugal

Table 3: Correlations between mediation strategy scales and the Hofstedian dimensions.

Power distance Uncertainty Masculinity Collectivism avoidance Social mediation, -0.20 -0.38* -0.26 -0.39* parents’ perspective (Cronbach’s alpha=0.90) Social mediation, 0.09 -0.11 0.04 -0.12 Children’s perspective (Cronbach’s alpha= 0.79) Co-use, parents’ -0.23 -0.31 -0.20 -0.45**

19 perspective (Cronbach’s alpha = 0.68) Co-use, children’s 0.10 -0.03 -0.16 -0.27 perspective (Cronbach’s alpha= 0.68) Allowance (mirroring -0.07 -0.15 -0.54*** 0.10 restrictive mediation), parents’ perspective (Cronbach’s alpha=0.92) Allowance (mirroring 0.00 -0.09 -0.55*** 0.097 restrictive mediation), children’s perspective (Cronbach’s alpha=0.93) Technical mediation, -0.17 -0.30 0.17 -0.43** parents’ perspective (Cronbach’s alpha= 0.84) Technical mediation, -0.08 -0.17 0.33 - 0.34* children’s perspective (Cronbach’s alpha= 0.87) (N=24; *=p<0.1 ** = p < .05; *** = p < .01)

Table 4: Country clusters: The Hofstedian masculinity dimension and specific mediation styles.

Hofstede’s Social Co-use (from high Restrictive Technical mediation masculinity mediation to low) mediation (from high to low) dimension (from high to (from high low) to low) LOW HIGH HIGH HIGH HIGH Sweden, Norway, Finland, Norway, The Italy, UK, Ireland, UK, Turkey Denmark, Poland, Netherlands France, The Netherlands, Norway Turkey, Lithuania, Slovenia Germany, Ireland MEDIUM LOW MEDIUM MEDIUM HIGH MEDIUM MEDIUM HIGH Finland, Estonia, HIGH Poland, Spain, HIGH Poland, The Portugal, Bulgaria, Greece, Italy, Sweden, France, Belgium, Netherlands, Norway, Rumania, Turkey, Rumania, Italy, Portugal, Poland, France, Sweden, Spain, Portugal, Ireland, Greece, UK, Finland, Belgium, Spain, the France Denmark, Finland, Denmark Austria, Czech Republic, Italy, Spain, The Greece, Finland, Greece, Netherlands, Spain, Bulgaria, Germany The Czech Portugal

20 Republic, France, Ireland, Sweden, UK, Bulgaria MEDIUM HIGH MEDIUM MEDIUM LOW MEDIUM MEDIUM LOW Italy, The Czech LOW The Czech Republic, LOW Estonia, Hungary, Republic, Poland, Turkey, Hungary, Rumania, Denmark, Portugal, Austria, Belgium, Greece, Austria, Belgium, Bulgaria, The Czech Denmark, Rumania UK, Germany, Lithuania, Germany, Republic, Ireland Hungary, The Estonia, Netherlands, Germany, Sweden, Belgium Norway, Rumania, Hungary, Bulgaria HIGH LOW LOW LOW LOW Austria, Hungary Slovenia Estonia, Turkey, Slovenia, Slovenia, Lithuania Lithuania, Slovenia Estonia, Austria Lithuania

Table 5: Correlations between judgments on parental mediation and the Hofstedian dimensions.

Power distance Masculinity Uncertainty Collectivism avoidance Parents who consider their 0.09 0.43** 0.06 -0.10 children’s internet experience improved a lot due to their interventions Parents who think they 0.39* 0.03 0.29 0.58*** should do a lot more Parents who do a lot of 0.06 0.25 0.05 0.33 different things because their child has been bothered by something on the internet Children who consider 0.29 0.59*** 0.37* 0.17 their internet experience has improved a lot thanks to their parents’ interventions Children who consider 0.49** 0.01 0.33 0.27 their internet experience is being considerably limited by their parents’ interventions Children who consider 0.29 -0.19 -0.06 0.23

21 their parents act very differently after they had a disturbing online experience Children who largely 0.29 0.15 0.13 0.25 ignore their parents’ advice (N=24; *=p<0.1 ** = p < .05; *** = p < .01)

Table 6: Clusters on the Hofstedian power distance dimension and parental judgments.

Hofstedian power distance dimension Parental intervention judged important (from low to high) 1) LOW Denmark, Austria Sweden, Slovenia, Lithuania, Poland, The Netherlands, Austria, Denmark, Norway, Hungary, Belgium 2) MEDIUM LOW UK, Germany, Ireland, Hungary, Italy, UK, Finland, Germany, Portugal, The Czech Sweden, Norway, The Netherlands, Lithuania, Republic, France, Estonia, Turkey Finland, Estonia 3) MEDIUM HIGH France, The Czech Republic, Poland, Belgium, Greece, Bulgaria, Spain, Ireland, Italy Spain, Greece, Bulgaria, Portugal, Slovenia, Turkey 4) HIGH Rumania Rumania

22 i

Recommended publications