«13TH INTERNATIONAL PUBLIC RELATIONS RESEARCH CONFERENCE “Ethical Issues for Public Relations Practice in a Multicultural World” Holiday Inn ...»
Data Collection One SNS community from each brand in Mixi and Facebook was selected for the analysis (total of six communities). Among many online communities associated with each of the three car brands on Mixi and Facebook, one was selected by entering each brand’s name on the search box of Mixi and Facebook. The community which had the largest membership was chosen for analysis. As a result, the numbers of members in the three Mixi communities were 13,213 for Lexus, 6,101 for Cadillac, and 18,770 for BMW. For the three Facebook communities, the members were 5,567 for Lexus, 3,108 for Cadillac, and 5,927 for BMW.
Fifty comments were chosen from each community. For Mixi communities, 50 comments were extracted from each of the first 50 threads (one comment for each thread) that appeared on the discussion boards. For Facebook communities, 50 comments were selected from both discussion boards and “walls”. Twenty five comments were chosen from the first 25 threads that appeared on the discussion board, and other 25 comments were chosen from the first 25 comments that appeared on the walls. If a Facebook Group didn’t have more than 25 topics on discussion board, the number of comments collected from the discussion board was the number of the threads that appeared on the discussion board.
The procedure for choosing one comment from each thread is as follows. Firstly, each topic was assigned with a number from 1 to 10 according to the order in which the threads were updated.
Among many comments that appeared in a thread, the ten newest comments were assigned with a number from 1 to 10 according to the order in which the comment were posted. The comment chosen for analysis was the one which had the number corresponding to the number assigned for each 279 thread. If the thread had fewer comments than the number assigned for the thread, the count was repeated until the assigned number was reached. For example, if a thread was assigned with the number of 1 (or 3), the comment which posted most recently (or the third most recent comment) at that time was chosen for analysis. If a thread only had five comments while the thread was assigned the number 9, the fourth most recent comment was chosen for analysis. Same method was used for choosing a comment from a thread that appeared on the “walls” of Facebook communities.
Comments were collected during the last week of December, 2009. Those comments written in languages other than English were excluded from the pool. In addition, the comments that only had pictures were excluded from the pool as well. As a result, 150 comments were collected from both Mixi and Facebook communities, and the total number of comments analyzed in the study was 300.
Coding procedure Coding was conducted by three bilinguals (Japanese and English) including the author herself. Two other coders were hired by the author and given a codebook containing definitions and coding rules. Comments selected from Mixi communities were coded by author and one of the two coders, and comments selected from Facebook communities were coded by the author and another coder. The author conducted the training and answered any questions regarding the definitions of the variables and procedures. Coders worked independently and were not permitted to discuss the SNS communities with each other.
The independent coders examined the presence of visually oriented information and personal information in each comment. In addition, the coders were also asked to identify the most salient comment types for each comment. Eight comment types were constructed for this study.
Three of the eight comment types have subcategories and the coders were asked to identify these subcategories as well. The definitions for each variable are shown in Appendix 1.
Inter-coder reliability analysis was used to test internal consistency. Percentage agreements between the two coders were 100% for the use of visually oriented information, 98% to 100% for the disclosure of personal information in both Mixi and Facebook comments. For type of comments, the percent agreements were 85.3 % for Mixi comments and 87.3% for Facebook comments. These were well within acceptable limits (Kassarjian, 1977). For type of comments, disagreements occurred mainly between “offering advice and information” and “advertising something.” It was found that coders sometimes couldn’t easily discern whether the comments were simply written for providing information to others or if there were advertising intentions behind the comments. The comments that had coding disagreements were rechecked and a final decision was made through discussion between the coders.
Results Functionality The six SNS communities were analyzed regarding functionality. In terms of similarity between Mixi and Facebook communities, all of the six SNS communities had functions to show a general description of the community, who the members of the community are, and who the administrators of the community are. In addition, all the SNS communities provided discussion boards. One of the Mixi communities and two of the Facebook communities had a function to post external links related to the communities. However, for Mixi communities, links that could be added were restricted to links with Mixi communities, while for Facebook communities, there was no restriction. In terms of differences between the two SNS, all of the three Mixi communities had a function to show what other communities the members were in, and a function to list related events.
Moreover, one of the three Mixi communities had a function to create and conduct surveys. These are the functions that couldn’t be found in Facebook. On the other hand, all of the three Facebook 280 communities had a “wall” where members can freely post the comments, and the dedicated places to update pictures and videos.
Disclosure of personal information The first hypothesis predicted that comments from Mixi communities would have more personal information than the comments from Facebook communities. Eight percent of the comments from Mixi and 4% of the comments from Facebook had at least one type of personal information. The most frequently mentioned type of personal information within the Mixi comments was “information about where he/she lives” (3.3%), followed by “family information” (2%), “gender” (0.7%), “age” (0.7%), and “other” (1.3%). Looking at the comments from Facebook, the results were first “information about where he/she lives” (2%), with “family information” (1.2%) next and then “occupation” (0.7%). There was no information about “occupation” mentioned in Mixi comments, and no information about “gender”, “age”, and “other” in Facebook comments. Chisquare test failed to find the significant difference in the disclosure of at least one type of personal information, χ2 (1, N = 300) = 2.13, p.10, and no significant differences was found among subcategories of personal information either. Therefore, H1 was rejected.
Use of visually oriented information The second hypothesis of this study predicted that comments that appeared in Mixi communities would have more visually oriented information than comments that appeared in Facebook communities. Among the 150 comments that were gathered from Mixi communities, 54% of the comments have at least one type of visually oriented information, 12% of the comments have photographs, 48.7% of the comments have emoticons, and no comments have other types of visually oriented information. For the 150 comments that were gathered from Facebook, 23.3 % of the comments have at least one type of visually oriented information, 12.7% of the comments have photographs, 10.0% of the comments have emoticons, and 1.3% of the comments have other types of visually oriented information. Chi-square tests confirmed that Mixi and Facebook differed significantly in the use of at least one type of visually oriented information, χ2 (1, N = 300) = 29.74, p.01, and the use of emoticons, χ2 (1, N = 300) = 54.10, p.01. Therefore, H2 was supported.
Type of comments Comments were further analyzed to examine whether there were differences among types of comments appearing Mixi and Facebook, respectively. Chi-square tests revealed that overall there was a significant difference between the two SNS, χ2 (1, N = 300) = 60.52, p.01. Mixi had significantly more comments than Facebook for “asking for advice and information” (19.3% vs. 8%), χ2 (1, N = 300) = 8.17, p.05, and “offering advice and information” (36.0% vs. 9.3%), χ2 (1, N = 300) = 30.42, p.01. On the other hand, Facebook had significantly more comments than Mixi for “showing positive feeling or support for the brands, companies, and shops” (30.7% vs. 8%), χ2 (1, N = 300) = 24.70, p.01, and “showing neutral feeling about the brands, companies, and shops” (7.3% vs. 1.3%), χ2 (1, N = 300) = 6.51, p.05. No significant differences were found for “showing negative feeling about the brands, companies, and shops”, χ2 (1, N = 300) = 1.31, p.10, “advertising something”, χ2 (1, N = 300) = 2.93, p.05, and “showing opinion to comments wrote by others”, χ2 (1, N = 300) = 0.14, p.10 (Table 1). However, there was a significant difference for “showing appreciation for the comments written by others”, χ2 (1, N = 300) = 7.64, p.01, which is a subcategory of “showing opinion to comments written by others.” In summary, the results of this study suggest that there were differences between Mixi users and Facebook users in terms of the use of visually oriented information and type of comments. Mixi users used significantly more emoticons than Facebook users when they make comments in SNS communities. In addition, Mixi users made more comments about asking information and providing 281 information to other community members, and Facebook users made more comments that show their positive and neutral feeling about brands, companies, and shops.
ConclusionSNS are gaining popularity among various countries in the world, and this makes SNS a valuable medium to communicate effectively with consumers overseas. However, there appears to be a lack of research examining how SNS are used differently among countries and cultures. In this study, two SNS from two different countries were analyzed, and the influence of culture on how people make comments on SNS were examined by adopting Hofstede’s and Hall’s cultural values. It was found that people in high context culture (Mixi users) tended to use more visually oriented information than people in low context culture (Facebook users) when they communicate with others in SNS.
Unexpectedly, no differences were found in terms of the amount of personal information written in the comments, although the study predicted that people in collectivistic culture (Mixi users) would disclose more personal information about themselves in SNS communities. Overall, there were only few comments that contained personal information among both Mixi comments and Facebook comments. This might be due to the fact that all of the three Mixi communities analyzed in this study had more than 6,000 members, and people might not feel comfortable disclosing their personal information in such an environment. An additional argument can be made that although sharing personal information is a way for people to show who they are in a collectivistic culture, it does not translate to the online environment. Future research could analyze different sizes and types of SNS communities in order to further understand people’s behavior of disclosing personal information in different cultures.
The results of this study also revealed that the types of comments people wrote differed among Mixi and Facebook communities. Mixi users tended to write comments in SNS community that involved asking or providing advice and information, while Facebook users made comments because they want to express their feeling towards the brands. These behavioral differences between Mixi users and Facebook users might result from the differences of collectivistic and individualistic cultural values. In a country which has a collectivistic culture like Japan, people tend to be interdependent, and being a good member of a community is regarded as a virtue. The large number of comments asking or providing advice and information that were found in Mixi communities might be a manifestation of this collectivistic cultural value. Meanwhile, the individualistic culture is “I” conscious and there is less mutual obligation among people. This cultural value might influence Facebook users’ behavior on SNS communities, and as a result, Facebook users tend to use SNS communities as a place for expressing their passions and opinions towards the product and brand, rather than sharing advice and useful information.
These findings suggest that when companies utilize SNS for communicating with their international publics, they need to know how the SNS were used differently among countries in order to have better understanding about international publics and build relationships with them effectively. For instance, given the differences in the use of visually oriented information and what people write as comments in SNS communities between Japan and the U.S, companies could position the Japanese SNS communities more as a place to clarify consumers’ inquiries about the products, while providing more space where people can freely express their opinions towards the products in their U.S. SNS communities. In addition, companies need to understand the importance and meaning of using visually oriented information, especially emoticons, when they communicate with consumers from high context cultures.
communities. Future research could analyze low involvement products, such as instant coffee or shampoo, and examine whether people in individualistic/collectivistic and high/low context cultures would comment differently in those products’ SNS communities.
In addition, although Japan is recognized as a typical collectivistic and high contest culture country, and the U.S. is seen as a representative of individualistic and low context culture country, the results of this study might not be applicable to other countries due to the other social and infrastructural differences. Future studies need to examine SNS communities in various countries in order to clarify the effect of individualistic/collectivistic and high/low context cultural values on people’s behavior of using SNS communities.