This is a summery of the article "A Critical Analysis of Current Indexes for Digital Divide Measurement" by Bruno et al. (2011). It also comes with a crude comparative graph at the end.
3. 1
Focusing on technological resources
Individuals who use
computers and the Internet
and individuals who do not
(Mehra et al. 2004)
2
Emphasizing
determining factors
Per capita income,
telecommunications
infrastructure and the quality
of regulation
(Chinn and Fairlie 2006)
Information haves and havenots
(Dewan and Riggins 2005, Ida and
Horiguchi 2008, Belanger and Carter 2009)
Persons who have access to
digital ICTs and those do not
(Dewan and Riggins 2005)
Economic, regulatory, and
sociopolitical characteristics of
countries
(Cuillen and Suarez 2005)
3
Comprehensive
definitions
“unequal patterns of material
access to, usage capabilities
of, and benefits from
computer-based information
and communication
technologies that are caused
by certain stratification
processes that produce
classes of winner and losers
of the information society, and
of participation in institutions
governing ICTs and society.”
(Fuchs 2005: 46)
Techno-centric
Multidimentional
5. Digital Divide Measurement
Composite indexes
•The aggregation of several
indicators into a single
figure
•Representing the relative
position of countries
overtime
11
8
Digital Access
Index
21
Infostate
Index
2003
Digital
Opportunity
Index
10
ICT
Opportunity
Index
2005
11
ICT
Development
Index
2009
6. Critiques of composite indexes
• Emphasize income, education, age, sex, and
ethnicity, while not fully addressing the deeper social,
cultural, and psychological causes behind access
inequalities. … a lack of conceptual elaboration and
definition of the indicators used in composite indexes
(e.g. computer literacy, Internet use)
(Van Dijk 2006)
• Too many indicators make data collection difficult
(Braithwaite 2007)
• Measuring at the national level ignores community
level inequalities
(Barzilai-Nahon 2006)
• The aggregation methodology of individual indicators
is responsible for biases (e.g. the weight) (BarzilaiNahon 2006, James 2007)
7. Critiques of composite indexes
• Emphasize income, education, age, sex, and
ethnicity, while not fully addressing the deeper social,
cultural, and psychological causes behind access
inequalities. … a lack of conceptual elaboration and
definition of the indicators used in composite indexes
(e.g. computer literacy, Internet use)
(Van Dijk 2006)
• Too many indicators make data collection difficult
(Braithwaite 2007)
• Measuring at the national level ignores community
level inequalities
(Barzilai-Nahon 2006)
• The aggregation methodology of individual indicators
is responsible for biases (e.g. the weight) (BarzilaiNahon 2006, James 2007)
A GOOD INDEX?
should be both
efficient and effective
(Jollands et al. 2004)
9. 1
?
To seek the possibility to increase their efficiency by reducing
the number of indicators and using the same technique of
aggregation.
2
To analytically validate the critiques by Van Dijk (2005, 2006)
and Fuchs (2009): current digital divide research is affected by a
“reductionistic” approach to measurement that does not
emphasize the role of factors other than technological access
and use.
10. Geometric average
Geometric average
Main telephone lines per 100 inhabitants
Mobile cellular subscribers per 100
inhabitants
Network
Geometric average
International internet bandwidth
Adult literacy rates
Infodensity
Skills
Gross enrolment rates
ICT-OI
Internet users per 100 inhabitants
Proportion of households with a TV
Uptake
Computers per 100 inhabitants
Total broadband internet subscribers per
100 inhabitants
International outgoing telephone traffic
per capita
Info-use
Intensity
11. Arithmetic average
Fixed telephone lines per 100 inhabitants
Mobile cellular telephone subscriptions
per 100 inhabitants
International Internet bandwidth (bit/s) per
Internet user
Weighted sum
ICT
access
x 40%
Proportion of households with a computer
Proportion of households with Internet
access at home
IDI
Internet users per 100 inhabitants
Fixed broadband Internet subscribers per
100 inhabitants
ICT use
x 40%
ICT skills
x 20%
Mobile broadband subscribers per 100
inhabitants
Adult literacy rate
Secondary gross enrolment ratio
Tertiary gross enrolment ratio
12. Correlation
Matrix
Principal Component
Analysis
Indicator
Selection
Calculate and analyze
the correlation among
each pair of indicators
Detect a set of variables
able to significantly
represent the
phenomenon within a
data set
Correlate each indicator
and each of the p
selected principal
components, then
individuate the
indicators with the
highest values of
correlation for each
principal component
Confirmation of the possibility
to reduce variables
The number of significant
variables (p < n)
Specific indicators to retain
18. Linear Regression Results
Original vs. Reduced DD Indexes
10
4
ICT-OI and ICT-OIreduced
0.946 (R2= 0.896)
11
4
IDI and IDIreduced
0.916 (R2= 0.839)
19. Linear Regression Results
Original vs. Reduced DD Indexes
10
DD Indexes vs. Income Index
4
ICT-OI and ICT-OIreduced
ICT-OI and GDP
0.946 (R2= 0.896)
0.942 (R2= 0.887)
11
4
IDI and IDIreduced
IDI and GDP
0.916 (R2= 0.839)
0.921 (R2= 0.845)
Strong correlation
20. “Redundant”
!
It is possible to increase efficiency
by eliminating less significant
indicators
“Reductionistic”
There is a need to include more
variables to comprehensively
capture the phenomenon
21. Internet Access and Gender Equality by Country
2008-2009, ITU
*
Female Internet
Access %
Male Internet
Access %
50
Corr=0.46
40
30
20
Higher Rank
Female to Male Ratio of Internet Access*
60
10
0
0
10
20
30
40
Higher Rank
Individuals Internet Access
50
60
22. Internet Access and Gender Equality by Country
2008-2009, ITU
*
Female Internet
Access %
Male Internet
Access %
Senegal
Switzerland
50
Korea
Corr=0.46
40
30
Colombia
20
Higher Rank
Female to Male Ratio of Internet Access*
60
Thailand
10
UK
0
0
10
20
30
40
Higher Rank
Individuals Internet Access
50
60