![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
Abstract
In Japan, the mobile phone has become an increasingly popular medium of accessing the Internet and e-mail, especially among those who may not have affordable access to a computer. The lower startup costs and the easier overall accessibility of mobile phones relative to personal computers suggest that the mobile phone may be a possible alternative for overcoming unequal access to information communication technology in Japan and elsewhere. Using microdata from 2001, we compare and contrast the determinants of mobile versus computer-based Internet in Japan. Our findings show that mobile Internet access, as compared to computer access, is determined less by demographics, socioeconomic status, and technological readiness, suggesting that barriers to Internet access are lower for mobile phones than is the case for computers.
An earlier version of this article was presented at the miniconference of the American Sociological Association, section on Communication and Information Technology in August 2005. We thank Juro Toda and Kaga Yanagisawa for assistance with the Nomura Research Institute dataset. We are grateful to Gerald Lombardi and Madeline Zavodny for their comments.
Notes
∗p < .05
∗∗ p < .01.
∗p < .05
∗∗ p < .01.
1. See CitationOno and Zavodny (2007) for a study of the determinants of ICT access and use in the United States, Japan, Sweden, South Korea, and Singapore, and CitationDiMaggio et al. (2004) for a comprehensive review of the empirical literature in the U.S. and elsewhere.
2. Ninety percent of total mobile phone users in Japan are mobile Internet service subscribers. The corresponding figures for South Korea, Finland, and the United States are 87%, 18%, and 12%, respectively (CitationMinistry of Internal Affairs and Communications, 2004).
3. For example, in Bangladesh, Grameen Telecom provides women with mobile phones in villages where landline telephones are not available (CitationNorris, 2001).
4. These include computer and Internet training programs targeted to nonusers, promotional activities to publicize the benefits of Internet use, etc. For example, the Ministry of Internal Affairs and Communications held an “Internet Expo” in 2000 and has subsidized computer training to promote Internet access (CitationMinistry of Internal Affairs and Communication, 2004).
5. In a survey conducted in 1999, 47% of younger Japanese nonusers (18 to 39 years old) and 19% of older Japanese (40 years old or older) mentioned cost as a reason for not using the Internet (CitationOffice of Research, 1999).
6. According to CitationUNESCO (2005) statistics, the percentage of web pages in English declined from 75% in 1998 to 45% in 2003. But English web pages still have the largest share among all languages. English speakers also make up the largest share of all Internet users (at 35%).
7. The percentage of observations with missing household income was 28%. Income was imputed based on the other control variables included in the regressions. To examine the robustness of the results reported here, we used income categories that included a category for missing income. The results showed that ownership and usage patterns were for the most part monotonically increasing with respect to household income categories, and that including imputed income did not significantly affect the coefficients or standard errors of the other covariates.
8. We use the expression computer keyboard skills to distinguish this from the familiarity of using the key pad functions on the mobile phone.
9. Predictions were generated using the following equation: Y = 
where ∩ β
i
are the coefficients from and arx
i
are the corresponding means, and age ranges from 18 to 59.
10. There is already some evidence of the diminishing cohort effect in the United States. For example, CitationDiMaggio et al. (2004) explain that the gap in Internet access between the young and the old is diminishing in the case of the United States.
11. Since keyboard skills and education are both measures of human capital, we conducted a diagnostic to check for multicollinearity. Results reveal that the variance inflation factor (VIF) was less than 2 for all covariates, confirming that multicollinearity did not significantly influence our estimations.
12. We can see this by testing the differences in the log-likelihood ratios: Δ L = –2(L 2 – L 1) where L 2 is the log-likelihood for the full model and L 1is the log-likelihood for the simpler model. For mobile Internet access, model A from is nested in model B from and Δ L = 42. For computer Internet access, model B from is nested in model F from and Δ L = 134. While Δ L for both access methods is statistically significant (at Δ it df = 6), the larger difference in the log-likelihood for computer Internet access indicates that the technological readiness variables improve the model predictions more than they do for mobile Internet access.
13. See for example, BBC News World Edition, “Mobiles to leapfrog into the future,” September 30, 2002.