Spread the Word The role of children as opinion leaders in the adoption of an online application

(1)

Spread the Word

The role of children as opinion leaders in the adoption of an online application

By

Peter van Eck

Rijksuniversiteit Groningen

Faculty of Management and Organization

(2)

EBSCOhost page 1 of 1

Title: The Role of Children as Opinion Leaders in the Adoption of an Online Application

Author: van Eck, Peter S.1 s1365878@student.rug.nl

Source: Students Journal of Marketing (SJM); June2006, Vol. 2 Issue 3, p3-11, 9p

Document Type: Article

Subject Terms: * Opinion leaders

* informational and normative influence * innovators and early adopters

NAICS/Industry Codes: 516110 Internet Publishing and Broadcasting

541910 Marketing Research and Public Opinion Polling

Abstract: It has long been recognized that word of mouth has a great influence on the adoption process of new products. Opinion leaders play an important role in this process. In this article opinion leaders among children are identified. A distinction has been made between informational and normative influence. Furthermore opinion leaders are compared with innovators and early adopters. The results show that opinion leaders differentiate themselves from non-leaders stronger on informational influence than on normative influence. Furthermore opinion leaders are more likely to be early adopters than they are likely to be innovators.

[ABSTRACT FROM AUTHOR]

(3)

3

Students Journal of Marketing, Vol. II (June 2006), 3-11 INTRODUCTION

With the introduction of a new product, like an online application, an important question always needs to be answered: Is it possible to reach the whole target group, without investing too much money. It has long been recognized that Word of Mouth (WOM) has a great influence on the adoption process of new products (Sheth, 1971). WOM is found to be more effective in influencing the intentions of people to start using a product, than mass media (Day, 1971; Still, Barnes and Kooyman, 1984). For children WOM also has an important influence on their decision making process (Hansen and Hansen, 2005; Spungin, 2004; Ward, 1974).

Word of Mouth is defined as the informal communication between customers about a product and it’s characteristics (Westbrook, 1987). It has several advantages above mass media. Besides the fact that it is more effective than mass media, it is also less expensive and more people can be reached (Procter and Richards, 2002). Another advantage of WOM is that it activates people, so people actually act on the information they get from WOM (Gelb and Johnson, 1995). The influence of WOM is found to be strongest for intangible products (Murray, 1991). So, for an online application, a strong WOM influence can be expected.

Katz’s (1957) two-step flow of communication explains an important phenomenon in WOM influence. A small group of individuals is influenced by mass media. People in this group are called the opinion leaders, a term introduced by Lazarsfeld, Berelson and Gaudet in 1948. All other people don’t act on the information from the mass media, but start adopting the product when they hear about the experiences of the opinion leaders. This role of the opinion leader is also confirmed by other authors (e.g. Procter and Richards, 2002; Berelson and Steiner, 1964: 550).

* Peter S. van Eck (e-mail: p.s.van.eck@student.rug.nl) is student Marketing at the Faculty of Management and Organization. The author wishes to thank Dr. W. Jager and Dr. J. Kratzer for their critical reviews in earlier stages. Further thanks go out to everyone who participated in this research and/or made this research possible. EBSCO and the Journal of Marketing Research are thanked for providing layout guidelines.

Hansen and Hansen (2005) identified opinion leaders among children. In the literature, though, not much attention has been paid to this specific subject concerning children. No clear answer has been given on some important questions: What kind of influence do opinion leaders among children have? And what is their position and role in the adoption process?

INFLUENCING OPINIONS Parental Influence vs. Peer Group Influence

Parents influence the decision making process of their children because they are an important source of information for them (Hansen and Hansen, 2005). Parents also have an important influence in how their children make decisions, because children learn this from their parents (Grant and Stephen, 2006). On the other hand, some cases are known in which children influence each other (Spungin, 2004; Ward, 1974). Some great examples are the success of Harry Potter and Pokémon, were children learned most about the subject from each other (Procter and Richards, 2002).

From the viewpoint of the classic literature about WOM, it is not surprising that the influence children have on each other is strong. The strength of WOM influence is affected by several factors, one of which is the similarity between the source and the receiver of the information (Wangenheim and Bayón, 2004; Moschis, 1976; Davis, 1963; Festinger, 1954; Schachter, 1959). Brown and Reingen (1987) defined similarity as the degree to which individuals are similar in terms of some attributes. In different situation, different attributes are important. A higher degree of similarity makes it easier for the receiver to identify with the source (Kelman, 1961), and the chance is higher that attitudes are effected by WOM (Feick and Higie, 1992). Furthermore people (unconsciously) assume that similar people have similar needs and preferences (Festinger, 1954).

This means that the decision of a child to use an online application can be expected to be strongly influenced by other children, like friends or siblings. Those children have similar needs and preferences (e.g. having fun playing with the application), in contrast with the parents who might have totally different preferences (e.g. the application must be more educational).

PETER S. VAN ECK*

It has long been recognized that word of mouth has a great influence on the adoption process of new products. Opinion leaders play an important role in this process. In this article opinion leaders among children are identified. A distinction has been made between informational and normative influence. Furthermore opinion leaders are compared with innovators and early adopters. The results show that opinion leaders differentiate themselves from non-leaders stronger on informational influence than on normative influence. Furthermore opinion leaders are more likely to be early adopters than they are likely to be innovators.

(4)

4 STUDENTS JOURNAL OF MARKETING, JUNE 2006

Opinion Leaders

Although much has been written on the subject of opinion leaders, there is still a variety of definitions. Flynn, Goldsmith and Eastman (1994) defined opinion leaders as people directly influencing other consumers by giving advice and verbal directions for search, purchase and use of a product. Cory (1971) states that opinion leaders act as mediators who are responsible for the information flow from mass media to mass audience, which refers to the two-step communication flow earlier described by Katz (1957). Cory (1971) also found empirical evidence that opinion leaders are more involved in activities related to their subject, more informed about new developments in their subject and read more about it in the media than non-leaders.

Furthermore earlier research showed that opinion leaders are leaders in related subjects (King and Summers, 1970; Myers and Robertson, 1972), but not in multiple product categories (Feick and Price, 1987). So they are very interested in some topics and act as opinion leader only in these topics.

Also some conflicting definitions are mentioned. According to Glock and Nicosia (1964) opinion leaders are also a source of social pressure and social support that influences the decision making process. Hansen and Hansen (2005) define opinion leaders as the persons from whom others take advice and whom others tend to copy. Those definitions are in contrast with Burt’s definition (1999), who states that opinion leaders are no leaders “in the sense of being more attractive such that they are individuals whom others want to imitate”. According to Burt opinion leaders give information about a product to others in their environment. After that, the product is accepted by someone because of the advantage it gives, not because of the opinion leader himself (which would be the case if social pressure was used).

These contradictions can be explained when a distinction is made between different kinds of influence. Deutsch and Gerrard (1955) introduced informational and normative influence. They define informational influence as the tendency to accept information from others as evidence about reality. Burnkrant and Cousineau (1975) define normative influence as the tendency to conform to the expectations of others. The influence described by Glock and Nicosia (1964) and Hansen and Hansen (2005) can be seen as normative influence. This normative influence is identified in relation with opinion leaders by several authors (Procter and Richards, 2002; Bone, 1995). The other authors earlier mentioned describe the influence of the opinion leader mostly as informational influence.

Another important stream in the literature about opinion leaders is based on the diffusion theory Rogers introduced (1962). He made a distinction between five groups of consumers who all adopted the product on different moments. The first 2,5% of the consumers adopting a new product are called innovators. After the innovators, the early adopters (13,5%), early majority (34%) and late majority (34%) follow. The last 16%, called the laggards are the latest adopters and might even never adopt the new product.

The relation between the opinion leader literature and the diffusion theory can be seen in some definitions about opinion leaders: Baumgarten (1975), for example, defines opinion leaders as new product adopters who transmit product information to their peer group. Although some research has been done on this subject, no clear relation

between innovators and opinion leaders has been found (Venkatraman, 1989, Hansen and Hansen, 2005). On the other hand, Baumgarten (1975) refers to earlier research which provides evidence that earlier adopters have more influence on others than later adopters and that opinion leaders are more innovative than non-leaders (Myers and Robertson, 1972; Robertson and Rossiter, 1968; Rogers and Catarno, 1962; Summers, 1971). This indicates that opinion leaders are likely to be early in the adoption process.

For this article a definition will be used introduced by Lyons and Henderson (2005). They combined some widely accepted definitions from King and Summers (1970), Rogers and Cartano (1962), and Goldsmith and Flynn (1994): “Compared with consumers who seek their advice, opinion leaders frequently possess more experience or expertise with the product category, have been exposed to or acquired more information about the product, exhibit more exploratory and innovative behavior and display higher levels of involvement with the product category”.

HYPOTHESES What Kind of Influence?

As mentioned before, two kinds of influence can be distinguished: normative and informational influence (Deutsch and Gerrard, 1955). Several researchers found evidence that both kinds of influence affect purchase decision process or the way customers use the product (Bearden and Etzel, 1982; Burnkrant and Cousineau, 1975; Park and Lessig, 1977). Still, both influences can be expected to be important in different situations. If a product (or the use of it) is visible to others, the product is more sensitive to normative influence than if the product was not visible (Bearden and Etzel, 1982; Batra, Homer and Kahle, 2001). Normative influence is mostly described through the eyes of the one who is influenced: they try to meet the expectations of other people they want to identify with (Kelman, 1961). The opinion leader is setting the norm instead living up to it. Therefore it is expected that: H1a: Non-leaders are more affected by normative

influence than opinion leaders.

When opinion leaders help other people to get more information about a product, this can be defined as a form of informational influence (Lessig and Park, 1978). Since opinion leaders are defined as people who have more knowledge about the product, we can expect that:

H1b: Opinion leaders have more informational influence than non-leaders.

Opinion leaders could be affected by normative influence to a certain degree, if they are aware of their position and don’t want to loose it. Although they still set the norm, they could be afraid that this is only temporary, therefore doing things because they expect others will like it. The informational influence on the other hand is almost part of the definition of opinion leaders. This leads to the following hypothesis:

(5)

The Role of Children as Opinion Leaders in the Adoption of an Online Application 5

Role in Adoption Process

In the WOM literature, many references are made to the diffusion theory introduced by Rogers (1962). Some authors tried to find a relation between opinion leaders and innovators (Hansen and Hansen, 2005), but only some overlap was found. Steenkamp and Baumgartner (1992) suggested that people who engage in exploratory behavior are likely to be early adopters. Rogers (1983:284) related the innovativeness of opinion leaders to the norms of their environment. If change is accepted, opinion leaders are more innovative than when change is not (totally) accepted. All authors seem to agree on the fact that opinion leaders can be found somewhere early in the adoption process. Therefore:

H2a: Opinion leaders stand earlier in the adoption process than non-leaders.

If opinion leaders are indeed more innovative than non-leaders, thereby standing earlier in the diffusion process, important question remains where they stand exactly. Are they more likely to be innovators or early adopters? Baumgarten (1975) found that earlier adopters (innovators and early adopters) were integrated in a social group, with exception of “the first few percent”. The first few percent can be defined as innovators, were the others can be defined as early adopters. Opinion leaders are known for the fact that they are part of a group (Berelson and Steiner, 1964: 550). These facts together lead to the following:

H2b: Opinion leaders are more likely to be early adopters than they are likely to be innovators.

METHOD The Product

For this research it was important to select a product category were WOM was likely to happen. Hansen and Hansen (2005) did research on WOM with respect to several product categories. They found that children talked most about computer games. So it is likely that opinion leaders exist in this product category and that their influence can be measured. Within the category of ‘computer games’, the category of ‘online applications’ was chosen, because the users of these applications could easily be reached through the online application itself.

Three popular online applications were used to reach a large and broad group of respondents: Kijkradio (www.kijkradio.nl), Moovl (www.moovl.nl) and Sketchstudio (www.sketchstudio.nl). All have in common that children can make something and all visitors of the websites can see what they made.

For Kijkradio, this something is that children can build their own radio station. They can tell their own stories, upload pictures (e.g. a live correspondent on location) and invite other children to work on a ‘news item’.

In Moovl, children can make simple games or ‘cartoons’ by drawing objects and making these object move around the screen. Several factors, like gravity and heaviness can be changed.

Sketchstudio is an application in which children can make a sketch, using their favourite characters from a popular television show in the Netherlands. They can provide the characters with their own voices and stories.

The fact that all visitors of the websites can see the work of others, makes the use of the application visible and therefore sensitive to normative influence. For example, Kijkradio has a ‘main radio station’ which is broadcasted on the main website of Kijkradio. The ‘best’ news items from all individual stations are used for this ‘main station’. If you are trying to get your news item broadcasted through the ‘main station’ you are sensitive to normative influence. The fact that all applications have several (sometimes complex) options, makes that children can learn a lot from each other, which makes it sensitive to informational influence.

The Sample

For this research children between 6 and 12 years old who already use the application were asked, because only the children already using the application can say something about how they adopted the product. To only way to reach those children, was the use of an invitation on the login-sites. This was the only option because for the registration of the application no contact information was required. Recent research showed that there is no difference in the response from online surveys or traditional mail surveys (Deutskens, de Ruyter and Wetzels, 2006), so it was not considered a problem to use this method.

The invitation placed on the login-sites stated that the children could win tickets for the Cinekidfestival (a film-festival for children). On the first page of the questionnaire parents had to give permission to proceed with the questionnaire. Without permission, no questions were asked. Only one question was asked at a time to make sure the respondents would focus on the specific question. The questionnaire existed of 14 multiple choice questions (see Appendix I). Only a short questionnaire was used, to prevent that to many respondents would quit before the end of the questionnaire. 142 respondent started the questionnaire, of which 113 respondents completed it. Measuring Opinion Leadership

To identify opinion leaders, a self-report technique was used. The reliability of this technique can be questioned, but since a lot of researchers used this technique it makes comparison with those studies easier (Corey, 1971). Thomas Jr. (2004) reported a research firm using a self reporting technique with children. Considering those facts it is assumed that this method can be used in this research to.

(6)

discussion of new ___, which of the following happens most?’ with the answers ranging from ‘you tell your friends about___’ to ‘your friends tell you about___’. The other questions of the scale were not used because using all items would make the questionnaire (too) long, leaving almost no room for other questions about the influence and innovativeness.

Measuring Influence

Bearden, Netemeyer and Teel (1989) developed a scale to measure interpersonal influence, which was later validated by themselves (1990) and Schroeder (1996). In this scale, the distinction between normative and informational influence is made. To measure the kinds of influence opinion leaders have, two items of this scale were adjusted. Both to make them more appropriate for children and the situation (the original scale is made to measure if someone if influenced, instead of influencing others). Only two were chosen, to keep the questionnaire short.

To find out if someone has a normative influence on others is very difficult especially in a ‘large-world network’ (instead of a small-world network) which is used in this study. Therefore an item of the original scale was used (measuring being influenced), only adjusted to the product used: ‘It is important that others like what I make with ___’ (question 10). Originally (Bearden et al., 1989): ‘It is important that others like the products and brands I buy’). This item is chosen because it has a high factor coefficient, a low item-to-total correlation (making a good distinction between influence and normative influence) and it didn’t need to be adjusted too much.

The item used to measure informational influence is only based on the original scale, using the importance of ‘sharing information’ coming back in all items. This resulted in the following item ‘I help friends with ___’ (question 12). To measure the relative informational influence, also the opposite question was asked: ‘Friends help me with ___’ (question 13). Combining both questions makes it possible to measure relative informational influence. The answers on all items were based on a five-point Likert scale ranging from ‘never’ to ‘always’. Measuring Innovativeness

To measure innovativeness, Goldsmith and Hofacker (1991) developed the DSI scale. Two questions were developed based on the first item of this scale which stated “In general, I am among the first (last) in my circle of friends to buy a new___ when it appears”. This question was adjusted to the situation (using instead of buying). Furthermore absolute numbers were asked to make it easier for the children to answer the question. This resulted in two questions: ‘How many of your friends made something with ___?’ (question 8) and ‘How many of them made something earlier than you did?’ (question 9). The difference between those questions results in an answer to the question: “How many of your friends using the application made something later than you did?”. This indicates if the respondent is among the first or last users of the application within the group.

ANALYSIS Descriptives

A total of 142 respondents started the questionnaire (35,2% male, 64,8% female) of which 113 respondents

finished the questionnaire completely (32,7% male, 67,3% female). The age distribution (shown in table 1) is negatively skewed, but doesn’t differ between the total group of respondents and the group of respondents who totally completed the questionnaire. The average age of the total sample is 10,22 year (SD = 1,630). For those who finished the questionnaire, the average is 10,17 (SD = 1,670).

Table I AGE DISTRIBUTION

Age Started Questionnaire Finished Questionnaire

6 3,5% 3,5% 7 4,2% 5,7% 8 6,3% 6,2% 9 12,7% 13,3% 10 18,3% 17,7% 11 25,4% 23,9% 12 23,9% 23,9% Missing 5,6% 6,2%

30 respondents (21,1%) said friends were a source of information about the application. Only 12 (8,5%) said their parents were a source of information. The 21,1% was compared with the 8,5% in binominal test which showed the difference is significant based on a Z approximation (p = 0,000). This is in line with the expectation that children were more influenced by other children than by their parents concerning the decision to use an online application.

For the further analysis of the hypotheses only complete questionnaires were used, because some important questions for the identification of opinion leaders and the different kinds of influence were asked late in the questionnaire. Since the two groups have similar characteristics concerning age and gender, this was not considered a great problem.

Reliability of the Scales

To verify the reliability of the used scales, an item analysis was performed for the opinion leader scale, the informational influence scale and the innovativeness scale. The normative influence scale only consisted of one item, so no item analyse could be performed.

The Cronbach’s Alpha for the opinion leader scale was 0,099. The reason for the fact that the scale seems to be unreliable is that only two of the original seven items (King and Summers, 1970) were used. When using fewer items, a lower reliability can be expected. To compensate for this lack of reliability some additional analysis were performed to find out if the identified opinion leaders have characteristics of opinion leaders as identified in earlier research. This analysis will be discussed later.

The Cronbach’s Alpha for the informational influence scale and the innovativeness scale are respectively 0,728 and 0,667. So, those scales can be considered reliable. Identifying the Opinion Leaders

(7)

Table II

MANN-WHITNEY U TEST OPINION LEADER SCALE Mean Rank

Non-Leader

Mean Rank

Opinion Leader Z-value p-value How many friends did you tell

about___?** 40,53 90,82 8,105 0,000

If I talk about ___ with friends,…* 61,10 48,58 -2,092 0,018

N = 113

* Significant at α ≤ 0,05 Level ** Significant at α ≤ 0,01 Level

Lazarsfeld (1965: 333) and King and Summers (1970) classified the respondents with the highest scores on the opinion leader scale as opinion leaders (respectively the highest 23,8 to 31,1% and 23 to 30%). In line with their method, the 32,7% of the respondents who scored highest on the combined variables were classified as opinion leaders. This means that 37 of the 113 respondents were classified as opinion leaders. Twelve of the opinion leaders are male (32,4%), 25 are female (67,6%), which are the same proportions as in the total sample. The age doesn’t differ significantly (t=0,844; df=104; p=0,200) between opinion leaders (10,36; SD=1,552) and non-leaders (10,07; SD=1,731).

A Mann-Whitney U test confirmed the fact that opinion leaders differ significantly from non-leaders on the used items (see table II). They talked with significant more people about the application and know significant more about the application than friends.

To verify if this group indeed has some of the characteristics of the opinion leader described earlier in this article, some additional analysis were performed. In the first place, the difference in sources between the two groups was tested, in order to check if the two-step communication theory (Katz, 1957) is valid in this situation and that therefore opinion leaders indeed use more mass media. The respondents were asked from whom they heard about the application. Mass media are defined as television, the internet and teacher. Although this last group is not normally defined as mass media, teachers ‘sent a message’ to many children at the same moment, which makes them function as a mass medium.

The opinion leaders use more mass media than the non-leaders (t = -1,990; df = 80,018; p = 0,025), which is in support of the two-step communication theory and underscores the role of the opinion leader as mediator between mass media and the non-leaders (Cory, 1971). Although the difference is significant, the means do not differ much (non-leaders: 0,7895; SD=0,49842; opinion leaders: 0,9730; SD=0,44011) because only two answers were classified as mass media, so the maximum amount of mass media used was two.

Furthermore the respondents were asked how they played the first time: alone, asking someone else or being asked by someone else. An one-way ANOVA was used to find out if a relation existed between the score on the opinion leader scale and the answer given to this question. Although the ANOVA showed no significant difference between the groups (F=2,197; df = 2; p=0,116), the addition LSD test showed that significant differences were found. Respondent answering ‘I asked a friend’ score significant higher on the opinion leader scale than respondents answering ‘I played alone’ (Mean Difference = 0,67391;

SD = 0,37624; p=0,038) or ‘A friend asked me’ (Mean Difference = 1,02078; SD = 0,52078; p=0,027). This indicates that children asking friends (thereby involving other children) are more likely to be opinion leaders than children who don’t ask friends, which is in line with the expectations.

The results of these analysis indicate that the identified group has great similarities with the opinion leaders described in the literature, and therefore there is no reason to doubt that this group indeed can be defined as opinion leaders.

Testing the Hypotheses about Influence

Hypothesis 1a stated that non-leaders would be more affected by normative influence than opinion leaders. To test this hypothesis an independent-samples t-test was performed, using the ‘It is important that others like what I make with___’ item. There was no significant difference found considering the normative influence (t = 0,658; df = 111; p = 0,256). In contrast with the prediction in hypothesis 1a, that opinion leaders would be less effected by normative influence, they scored slightly higher on this item (see table III). This suggests that opinion leaders are not necessary ‘normative leaders’. Another explanation could be that leaders expressing normative influence are aware of their position (as an example for ‘the others’) and don’t want to loose this position. This means that they still want others to like what they do, thereby being exposed to normative influence. A third explanation could be that they are not aware of their position and (like the others) try to fit in the group.

Hypothesis 1b stated that opinion leaders would have more informational influence than non-leaders. To test this hypothesis an independent-samples t-test was performed, using the ‘I help friends with___’ item. In support of hypothesis 1b, a significant difference was found between opinion leaders and non-leaders in having informational influence (t=3,889; df=111; p=0,000). As expected opinion leaders have more informational influence than non-leaders (see table III). Myers and Robertson (1972) claim that opinion leaders are only relatively more influential than non-leaders, suggesting they also get influenced themselves. The item of the informational scale was also asked the other way around: ‘Do friends help you?’. For this question no significant difference was found (t=1,412; df=111; p=0,081), which supports the statement that opinion leaders are influenced as well (see table III).

(8)

Table III

T-TESTS NORMATIVE AND INFORMATIONAL INFLUENCE Mean (SD)

Non-leader

Mean (SD)

Opinion leader t-value (df) p-value It is important that others like what I make 3,17 (1,350) 3,35 (1,399) 0,658 (111) 0,256

I help friends** 2,11 (1,401) 3,22 (1,475) 3,889 (111) 0,000

Friends help me 1,93 (1,350) 2,32 (1,435) 1,412 (111) 0,081

Relative influence** 0,1711 (1,17062) 0,8919 (1,54171) 2,513 (56,894) 0,008 ** Significant at α ≤ 0,01 Level

influenced yourselves. A t-test showed a significant difference between opinion leaders and non-leaders on this new variable (t=2,513; df=56,894; p=0,008), indicating that opinion leaders influence others more than they get influenced themselves (see table III).

Hypothesis 1c stated opinion leaders differentiated themselves from non-leaders stronger on informational influence than on normative influence. Taking the results for both hypothesis 1a and 1b in consideration, it can be concluded that hypothesis 1c is supported. Opinion leaders don’t distinguish themselves from non-leaders considering normative influence. On the other hand, both groups differ significantly on the informational influence they have. All together opinion leaders indeed distinguish themselves from non-leaders more on informational influence than on normative influence.

Testing the Hypotheses about Adoption

Hypothesis 2a stated that opinion leaders stand earlier in the adoption process, than non-leaders. To test this hypothesis the two items of the innovativeness scale ‘How many of your friends made something with___?’ and ‘How many of them made something earlier than you did?’ were combined, resulting in a new variable: ‘How many of your friends adopted the application later than you’. With a Mann-Whitney U test, the difference between opinion leaders and non-leaders was tested (see table IV). In support of hypothesis 2a, opinion leaders know significant more people who were later with adopting the application, suggesting opinion leaders are earlier in the adoption process than non-leaders.

Table IV

MANN-WHITNEY U TEST ADOPTION Mean rank Non-leader Mean rank Opinion leader Z-value p-value Friends adopting later.** 51,05 69,22 2,939 0,002 N = 113 ** Significant at α ≤ 0,01 Level

Hypothesis 2b stated that opinion leaders are more likely to be early adopters than they are innovators. To test this hypothesis, the characteristics of opinion leaders as found in this research are compared with several characteristics of innovators found in earlier research. As mentioned before, innovators can be expected to be less integrated into a social group (Baumgarten, 1975). If opinion leaders were innovators, it could be expected that

they were likely to play with the application alone. As earlier analysis showed, the opposite is true: respondents who asked a friend scored high on the opinion leader scale. Furthermore, Clark and Goldsmith (2006) concluded that innovators are less sensitive to interpersonal influence than later adopters. The analysis of hypothesis 1a showed that opinion leaders are as sensitive to normative influence as non-leaders (see ‘It is important that other like what I make’, table III). Furthermore, opinion leaders are also as sensitive to informational influence as non-leaders (see ‘Friends help me’, table III). All those outcomes are in support of hypothesis 2b, so it is concluded that opinion leaders are indeed more likely to be early adopters than they are likely to be innovators.

CONCLUSION

A lot of research has been done on the subject of WOM and an important role was identified for the opinion leader. Almost all research was done with adult respondents and only sporadic research was done involving children. With this study, a start is made to fill this gap in the present literature.

The results indicate that a group of opinion leaders can be identified among children. This group has great similarities with adult opinion leaders: they talk with more people about the application, know more about it, hear more often about it from mass media and are more likely than non-leaders to involve others in using the product.

Concerning the influence of the opinion leader, the results show that opinion leaders don’t distinguish themselves from non-leaders in normative influence. On the other hand, opinion leaders do have a lot more informational influence and clearly distinguish themselves on this point from non-leaders.

The results also show that opinion leaders stand earlier in the adoption process than non-leaders, shown by the fact that they know significant more people who adopted the application later. On the other hand, opinion leaders don’t share a lot of characteristics with innovators, indicating that they are more likely to be early adopters than innovators.

(9)

Table V

OUTCOMES OF HYPOTHESES TESTING

H1a: Non-leaders are more affected by normative influence than opinion leaders. Not supported H1b: Opinion leaders have more informational influence than non-leaders. Supported H1c: Opinion leaders differentiate themselves from non-leaders stronger on informational influence

than on normative influence. Supported

H2a: Opinion leaders stand earlier in the adoption process than non-leaders. Supported H2b: Opinion leaders are more likely to be early adopters than they are likely to be innovators. Supported

For research, opinion leaders among children are identified who have great similarities with adult opinion leaders. This research also gives an indication of the relation between opinion leaders and the different kinds of influence they have. Furthermore the suggestion is made that opinion leaders should be compared with early adopters instead of innovators.

DISCUSSION, LIMITATIONS AND INDICATIONS FOR FURTHER RESEARCH

A first point of discussion is the concern that only children with ‘richer’ parents would have access to a computer. Spungin (2004) stated that only the ‘very poor’ were underrepresented on the internet. So, this is not considered a problem for the representativeness of the sample.

Although great care was taken by formulating and constructing the questionnaire and conducting the research, some limitations should be given. In the first place only two items of the originally seven item scale developed by King and Summers (1970) were used. Although the results show that the group identified as opinion leaders based on these items, have a lot of characteristics of adult opinion leaders, in a future research it would be advised to use all items, to improve the reliability of the scale.

Although only online applications were used for this research, similar characteristics of opinion leaders could also be found in other product categories. To test the generalisability to other product categories, the use of more product categories is suggested for future research.

In this research an online questionnaire was used. A great limitation of this method is the length of the questionnaire. Only fourteen questions were asked, but from the 142 who started the questionnaire only 113 completed it totally. Respondents already started quitting after six questions (7 missing), quickly increasing to 18 missing after nine questions. Several techniques were used to prevent this: the first screen indicated that fourteen questions would be asked, every screen showed how far the respondent had proceeded and they would only make a chance to win the cards for the Cinekidfestival if they would completely fill in the questionnaire. Although, with the use of the online applications in this study this was the only method to reach the respondents, in a future research a face to face situation would be advised. In such a situation the respondent would be more focused on the questionnaire, making it less likely that they will start doing ‘something else’ before finishing the questionnaire. In such a situation also more questions could be asked.

An online questionnaire can be a good choice because of the advantages it has above tradition questionnaires: flexible design formats, possibility to use sound and animations etc. (Granello and Wheaton, 2004). Furthermore it is possible to make sure that respondent answer all questions (no missing values) and answer them correctly

(not more answers for questions were only one answer should be selected) (Lazar and Preece, 1999). If an online questionnaire is used, it is important to ask important questions in the beginning of the questionnaire, because of the drop out rate.

APPENDIX I: QUESTIONNAIRE 1. Which of these applications do you like most?

a) Kijkradio b) Moovl c) Sketchstudio 2. I am a … a) Boy b) Girl 3. I am … years old.

4. Who talled you about ___ ? (more answers possible) a) Friends b) Siblings c) Parents d) Teachers e) Someone else f) Internet / TV

5. The first time I made something with ___: a) I did it alone

b) A friend asked me c) I asked a friend

6. How many friends did you tell about ___? a) 0

b) 1 c) 2 d) 3 e) 4 or more

7. If I talk about ___ with friends, I do that: (more answers possible)

a) while I am at school b) using the internet c) using the telephone d) while playing after school e) somewere else

8. How many of your friends made something with___? a) 0

b) 1 c) 2 d) 3 e) 4 or more

9. How many of them made something earlier than you did? a) 0

b) 1 c) 2 d) 3 e) 4 or more

10. It is important that others like what I make with ___. a) Never

(10)

d) Often e) Always

11. I have seen what my friends made with ___. a) Yes

b) No

12. I help friends with ___. a) Never

b) Not often c) Sometimes d) Often e) Always

13. Friends help me with ___. a) Never

b) Not often c) Sometimes d) Often e) Always

14. If I talk about ___ with friends, …: a) I know more

b) I know as much as my friends c) My friends know more

REFERENCES

Batra, R., P. M. Homer, and L. R. Kahle, (2001), “Values, Susceptibility to Normative Influence, and Attribute Importance Weights: A Nomological Analysis”, Journal of Consumer Psychology, 11(2), 115-128.

Baumgarten, S.A., (1975), “The innovative communicator in the diffusion process”, Journal of Marketing Research, 12(1), 12-18.

Bearden, W. O., and M. J. Etzet, (1982), “Reference Group Influence on Product and Brand Purchase Decisions”, Journal of Consumer Research, 9(2), 183-194.

Bearden, W. O., R. G. Netemeyer and J. E. Teel, (1989), “Measurement of Consumer Susceptibility to Interpersonal Influence”, Journal of Consumer Research, 15(4), 473-481. Bearden, W. O., R. G. Netemeyer and J. E. Teel, (1990), “Further Validation of the Consumer Susceptibility to Interpersonal Influence Scale”, Advances in Consumer Research, 17(1), 770-706.

Berelson, B. and G. A. Steiner, (1964), Human Behavior and Inventory of Scientific Findings, New York: Harcourt, Brace & World, Inc.

Bone, P. F., (1995), “Word-of-Mouth Effects on Short-term and Long-term Product Judgements”, Journal of Business Research, 32(3), 213-223.

Brown, J. J. and P. H. Reingen, (1987), “Social ties and word of mouth referral behavior”, Journal of Consumer Research, 14(3), 350-362.

Burnkrant, R. E. and A. Cousineau, (1975), “Informational and Normative Social Influence in Buyer Behavior”, Journal of Consumer Research, 2(3), 206-215.

Burt R. S., (1999), “The Social Capital of Opinion Leaders”, Annals of the American Academy of Political and Social Science, 566, 37-54.

Childers, T. L., (1986), “Assessment if the Psychometric Properties of an Opinion Leadership Scale”, Journal of Marketing Research, 23(2), 184-188.

Clark, R. A., and R. E. Goldsmith, (2006), “Interpersonal influence and consumer innovativeness”, International Journal of Consumer Studies, 30(1), 34-43.

Cory, L. G., (1971), “People Who Claim to be Opinion Leaders: Identifying Their Characteristics by Self-report”, Journal of Marketing, 35(4), 48-53.

Davis, J.A., (1963), “Structural balance, mechanical solidarity and interpersonal relations”, The American Journal of Sociology, 68(4), 444-462.

Day, G. S., (1971), “Attitude Change, Media, and Word of Mouth”, Journal of Advertising Research, 11(6), 31-40. Deutsch, M. and H. B. Gerrard, (1955), “A study of normative and informational social influences upon individual judgement”, Journal of Abnormal Social Psychology, 51, 629-636.

Deutskens, E., K. de Ruyter, and M. Wetzels, (2006), “An assessment of equivalence between online and mail surveys in service research”, Journal of Service Research, 8(4), 346-355.

Feick, L., and R. Higie, (1992), “The effects of preference heterogeneity and source characteristics on ad processing and judgements about endorsers”, Journal of Advertising, 21(2), 9-24.

Feick, L. F. and L. L. Price, (1987), “The Market Maven: A Diffuser of Marketplace Information”, Journal of Marketing, 51(1), 83-97.

Festinger, L., (1954), “A theory of social comparison processes”, Human Relations, 7(1), 117-140.

Flynn, L. R., R. E. Goldsmith and J. K. Eastman, (1994), “The King and Summers Opinion Leadership Scale: Revision and Refinement”, Journal of Business Research, 31(1), 55-64.

Gelb, B. and M. Johnson, (1995), “Word-of-mouth communication: causes and consequences”, Journal of Health Care Marketing, 15(3), 54-58.

Glock, C. Y. and F. M. Nicosia, (1964), “Uses of Sociology in Studying ‘Consumption’ behaviour”, Journal of Marketing, 28(3), 51-54.

Goldsmith, R.E. and L. Flynn, (1994), “Opinion leadership for vacation travel services”, Advances in Business Studies, 4(7-8), 17-29.

Goldsmith, R. E. and C. Hofacker, (1991), “Measuring consumer innovativeness”, Journal of the Academy of Marketing Science, 19(3), 209-221

(11)

Grant, I. J., and G. R. Stephen, (2006), “Communicating culture: An examination of the buying behaviour of ‘tweenage’ girls and the key societal communicating factors influencing the buying process of fashion clothing”, Journal of Targeting, Measurement & Analysis for Marketing, 14 (2), 101-114.

Hansen, F., and M. H. Hansen, (2005), “Children as innovators and opinion leaders”, Young Consumers, 6(2), 44-59.

Katz, E. (1957), “The Two-Step Flow of Communication: An Up-to-Date Report on an Hypothesis”, Public Opinion Quarterly, 21 (1), 61-78.

Katz, E., and P. Lazarsfeld, (1965), Personal Influence, Glencoe: Free Press.

Kelman, H. C., (1961), “Processes of opinion change”, Public Opinion Quarterly, 25(1), 57-78.

King, C. W., and J. O. Summers, (1970), “Overlap of Opinion Leadership Across Consumer Product Categories”, Journal of Marketing Research, 7(1), 43-50.

Lazar, J., and J. Preece, (1999), “Designing and implementing Web-based surveys”, Journal of Computer Information Systems, 39(4), 63-67.

Lazarsfeld, P. F., B. Berelson and H. Gaudet, (1948), The People’s Choice, New York: Columbia University Press. Lessig, V.P., and C.W. Park, (1978), “Promotional Perspectives of Reference Group Influence: Advertising Implications”, Journal of Advertising, 7(2), 41-47.

Lyons, B., and K. Henderson, (2005), “Opinion leadership in a computer-mediated environment”, Journal of Consumer Behavior, 4(5), 319-329.

Moschis, G. P., (1976), “Social Comparison and Informal Group Influence”, Journal of Marketing Research, 13(3), 237-244.

Murray, K. B., (1991), “A test of services marketing theory: consumer information acquisition activities”, Journal of Marketing, 55(1), 10-25.

Myers, J. H., and T. S. Robertson, (1972), “Dimensions of Opinion Leadership”, Journal of Marketing Research, 9(1), 41-46.

Park, C. W. and P. V. Lessig, (1977), “Students and Housewives: Differences in Susceptibility to Reference Group Influence”, Journal of Consumer Research, 4(2), 102-110.

Procter, J., and M. Richards, (2002), “Word-of-mouth marketing: beyond pester power”, International Journal of Advertising & Marketing to Children, 3(3), 3-11.

Robertson, T. S. and J. R. Rossiter, (1968), “Fashion Diffusion: The Interplay of Innovator and Opinion Leader Roles in College Social Systems”, working paper, University of California, Los Angeles.

Rogers, E., (1961), Characteristics of Agricultural Innovators an Other Adopter Categories, Wooster, Ohio: Ohio Agricultural Experiment Station.

Rogers, E. M. , (1962), Diffusion of Innovations. New York: Free Press.

Rogers, E., (1983), Diffusion of innovations, 3rd Ed. New York: Free press.

Rogers, E. M. and D. G. Cartano, (1962), “Methods of measuring opinion leadership”, Public Opinion Quarterly, 26(3), 435-441.

Schachter, S., (1959), The psychology of affiliation, Stanford: Stanford University Press.

Schroeder, J.E., (1996), “An analysis of the Consumer Susceptibility to Interpersonal Influence Scale”, Journal of Social Behavior & Personality, 11(3), 585-599.

Sheth, J., (1971), “Word-of-mouth in low-risk innovations”, Journal of Advertising Research, 11(3), 15-18.

Summers, J. O., (1971), “Generalized Change Agents and Innovativeness”, Journal of Marketing Research, 8(3), 313-316.

Spungin, P., (2004), “Parent power, not pester power”, International Journal of Advertising & Marketing to Children, 5(3), 37-40.

Still, R. R., J. H. Barnes Jr., and M. E. Kooyman, (1984), “Word-of-mouth communication in low-risk product decisions”, International Journal of Advertising, 3(4), 335-345.

Steenkamp, J.E. and H. Baumgartner, (1992), “The role of optimum stimulational level in exploratory consumer behavior”, Journal of Consumer Research, 19(3), 434-448. Thomas Jr., G. M., (2004), “Building the buzz in the hive mind”, Journal of Consumer Behaviour, 4(1), 64-72. Venkatraman, M. P., (1989), “Opinion leaders, adopters, and communicative adopters: A role analysis”, Psychology and Marketing, 6(1), 51-68.

Wangenheim, F. V. and T. Bayón, (2004), “The effect of word of mouth on services switching”, European Journal of Marketing, 38(9-10), 1173-1185.

Ward, S., (1974), “Consumer Socialization”, Journal of Consumer Research, 1(2), 1-17.