The 2007 e-readiness rankings: comments and critiques

The Economist Intelligence Unitin co–operation with the IBM Institute for Business Value — has published the The 2007 e-readiness rankings.

One of the caveats the EUI launches is that the ranking methodology has been modified, hence changes in rankings methodology raise the bar of e-readiness leadership, giving more weight to leadership and, thus, strong government role in promotion and adoption of ICT propel Asian countries upward. It is my opinion that stressing leadership role — and, indirectly, the role of the political and legal framework — is a good thing to do, as it is far demonstrated that is one of the most important barriers or catalysts — depending on the sign — when fostering the Information Society.

Another comment on this change in the ranking methodology is that, by doing it, the EUI e-Readiness Raking comes closer to the World Economic Forum Networked Readiness Index. In other words: they seem to be explaining more and more the same thing. Strange as this statement might sound, the following graphic can shed some light on it:

R2 value of NRI vs. EUI regression
R2 value of NRI vs. EUI regression

The figure shows the R2 value to the regression NRI = C0 + C1*EUI + ε. We can read the R2 value for 2006 — even if the report is issued in 2007, the ranking values are 2006’s — as EUI e-Readiness Ranking explains the 90.7% of the NRI Networked Readiness Index. This is far more than 2004’s value of 0.8235 (say, 82.3%). For those concerned in ICT and e-readiness measuring and indices, this is good news, as approaches seem to be getting closer. As can be seen in the next graphic , the value of the independent variable coefficient (X, in the graphic) seems to be (slowly) approaching the value of 1 while the constant is almost unchanged.

Constant and X-coefficient values of NRI vs. EUI regression
Constant and X-coefficient values of NRI vs. EUI regression

So far, the good news and/or comments. But the EUI also highlights the following findings:

  • E-readiness goalposts for countries are shifting.
  • The digital divide continues to narrow, even with the model changes.
  • Broadband is increasingly affordable, and almost everywhere.

I mostly agree with giving more importance to online content and services. Actually, I fully agree: infrastructure makes poor sense if, because of i.e. low digital literacy levels, this infrastructure is underused and no content or no services are provided online. But, again, I cannot agree that the digital divide is narrowing. On one hand, this is something that The Millennium Development Goals Report 2006 and UNCTAD’s Information Economy Report 2006 already put under quarantine. On the other hand, the EUI just ranks 69 economies, which are, of course, the most developed ones. So, even if it is true that the distance between the highest and lowest scoring countries dropped from 6.08 points to 5.80 points this year it is not fair to generalize this statement for the whole world, leaving out of the analysis more than 120 countries, two thirds of the total. And same for broadband.

Summing up: a good tool that comes to its 8th edition, providing good information along years, and that seems to be showing good results. But an information that should be consumed with caution.

More info

If you need to cite this article in a formal way (i.e. for bibliographical purposes) I dare suggest:

Peña-López, I. (2007) “The 2007 e-readiness rankings: comments and critiques” In ICTlogy, #44, May 2007. Barcelona: ICTlogy.
Retrieved month dd, yyyy from http://ictlogy.net/review/?p=537

Previous post: Online Volunteers: Knowledge Managers in Nonprofits

Next post: 3rd IDP Congress on Internet, Law and Politics. Briefings, part I: Jonathan Zittrain keynote speech

3 Comments to “The 2007 e-readiness rankings: comments and critiques” »

  1. I am not sure you are plotting the wrong regression coefficient! The R2 coefficient or COEFFICIENT OF DETERMINATION only ever explains the amount of variation explained by your model or fraction of the variance in yi that is accounted for by a linear fit of xi to yi (see for example http://en.wikipedia.org/wiki/Coefficient_of_determination). What you need is the CORRELATION COEFFICIENT or Beta that explains the degree of relationship between two variables (but does not imply that this relationship is causal). See http://en.wikipedia.org/wiki/Correlation_coefficient

    My best wishes, Phillippa Biggs

  2. Dear Phillippa,

    Thanks for the point, but I actually meant to use the R2 in the first figure and the Beta in the second one. Of course, the explanation is not technically accurate: I just wanted to give a faint idea on how both indices try to explain the same thing.

    In other words: if when doing a regression among the NRI and the EUI we got R2 = 1, we could then (more or less) state that they both explain exactly the same thing.

    I know I should be more “formal” in my explanations, but I try to keep things simple :)

RSS feed RSS feed for comments on this post. TrackBack URI

Leave a Reply

Your comment: