Page 134 - Culture Technology Communication

National Level Culture and Global Diffusion 117

adoption says nothing about the extent of use. In this sense this study is
similar to early diffusion studies of birth control, which were concerned with
adoption rather than actual use.
12. For example, an examination of GDP and PC data from 1990 to
1994 shows the top four and six countries, respectively, are consistent
throughout the period. Changes in ranks for other nations are typically lim-
ited to small increases or decreases, but are also difficult to determine due
to missing data.
13. It must be noted, however, that direct comparison of the
strength of these indicators is not possible from the analysis presented in
Table 5 because each indicator is reported by a different subset of coun-
2
tries. Although the adjusted R takes into account differences in sample
sizes, it does not remedy the situation that the models are being tested on
different groups of countries. This limitation is addressed in subsequent
parts of the analysis.
2
14. A simple interpretation of the R statistical measure is the
amount of variance in the dependent variable explained by the independent
variables. Here it can be described as the amount of variance in the START
variable that can be explained by the variables Teledensity, Call Cost, and
English Language Ability.
15. In a regression equation the standardized betas allow for the
comparison of the predictive strength of the independent variables. The re-
gression equation for model 1a is START .279 NWSPAPR .344
INTCALL .308 SCHLNROL.
16. r START,GEMPWR.TELEDEN .177 (p .1) and r START,TELEDEN.GEMPWR
.5511(p .01) and r START,TELEDEN.GDP_CAP .430 (p .01) and
r START,GEMPWR.GDP_CAP .335 (p .01).
17. This method will inflate the adoption percentage, an input to the
growth measure, for those countries with low teledensity rates. The result is
reduced variance in the growth measure, however, the countries involved in
this part of the analysis have relatively high teledensity rates so the impact
should be minimal.

18. r GROWTH,GEMPWR.PCPTHO95 .237 (n.s.) and r GROWTH,PCPTHO95.GEMPWR
.3746 (p .1).
19. To better understand this measure its bivariate correlation with
the other centrality measures were examined. It was found to have an in-
significant and low correlation with International Call Minutes, and moder-
ate correlations with CENTRALITY and LINKS (r .287, .315,
respectively). Its strongest correlations were with PCs ( .597), Teledensity
( .531), and Gender Empowerment ( .530), all significant.