Core ICT Indicators

Core ICT Indicators

Partnership on Measuring ICT for Development

Core ICT Indicators

Foreword

Comparable statistics on access to, and use of, infor-

mation and communication technologies (ICTs), are

critical to formulating policies and strategies con-

cerning ICT-enabled growth, for social inclusion and

cohesion, and for monitoring and evaluating the im-

pact of ICTs on economic and social developments.

However, internationally comparable information

society statistics are very limited, in particular in the

developing world.

To close the ICT data gap, a key objective of the

global Partnership on Measuring ICT for Develop-

ment is to set standards and harmonize ICT statistics

at the global level. To this end, it has worked to de-

velop a core list of ICT indicators that could be col-

lected by all countries and serve as a basis for inter-

nationally comparable statistics on the information

society.

The core list of ICT indicators presented in this doc-

ument was the outcome of an intensive consultation

process of the Partnership with national statistics

offices (NSOs). This included a stocktaking exercise

through which metadata on the status of official in-

formation society statistics were obtained from NSOs

worldwide, as well as a series of regional events that

discussed core ICT indicators. From these process-

es, a proposal for a core list of ICT indicators was

presented by the Partnership, and adopted by par-

ticipants, at the WSIS Thematic Meeting on Mea-

suring the Information Society (Geneva, February

2005).

This report provides definitions, model questions

and methodological notes relevant to the core indi-

cators agreed at the February 2005 meeting. By mak-

ing this information available to a wide audience, in

particular in developing countries, it is hoped to be

of use to countries developing their statistical data

collection programmes on the information society.

The report was drafted by Sheridan Roberts of OECD,

based on substantive contributions from ITU,

UNCTAD, UNESCO Institute for Statistics, UN-

ECLAC*, UNESCWA and Eurostat, and draws on dis-

cussions held in the framework of the OECD Work-

ing Party on Indicators for the Information Society

and the ITU World Telecommunication Indicators

meeting. The report was edited by OECD and Euro-

stat, and the layout was contributed by the World

Bank and UNESCWA. The report was printed by

UNESCWA.

* UNECLAC’s Information Society programme thanks for the financial support received from the @LIS programme of theEuropean Commission and the International Development Research Center (IDRC-ICA). The views expressed herein cannotbe taken to reflect the official opinions of these entities.


Table of Contents

INTRODUCTION AND CORE LIST OF ICT INDICATORSCORE INDICATORS ON ICT INFRASTRUCTURE AND ACCESSCORE INDICATORS ON ACCESS TO, AND USE OF, ICT BY HOUSEHOLDSAND INDIVIDUALSCORE INDICATORS ON USE OF ICT BY BUSINESSESCORE INDICATORS ON THE ICT SECTOR AND TRADE IN ICT GOODSREFERENCE MATERIALFOOTNOTES

16

1428394345

1

Core ICT Indicators

Introduction and Core Listof ICT Indicators

Introduction

The core list of ICT1 indicators presented in thisdocument was the outcome of an intensive con-sultation process by the Partnership on Mea-suring ICT for Development2 with national sta-tistics offices (NSOs). This included a stocktak-ing exercise through which metadata on the sta-tus of official information society statistics wereobtained from NSOs worldwide, as well as aseries of regional events on core ICT indicators.From these processes, a proposal for a core listof ICT indicators was presented by the Part-nership, and adopted by participants, at the WSISThematic Meeting on Measuring the Informa-tion Society (Geneva, February 2005).

This document is intended to provide more in-formation on the core list of ICT indicators agreedat the February 2005 meeting. It provides defini-tions, model questions and methodological notesrelevant to the indicators. While much of thismaterial has been debated in developed countryforums, for example, the OECD’s Working Partyon Indicators for the Information Society(WPIIS), the aim of this document is to makethis information available to a wider audience, inparticular, developing countries. Readers shouldnote that this document is not intended to be a

manual for collecting ICT statistics. The refer-ence material at the end of the document pro-vides information on existing guides and manu-als which will be useful for that purpose. It isalso envisaged that additional material will beprepared as future outputs by the members ofthe Partnership.

The core list contains four sets of indicators: (i)ICT infrastructure and access (ii) access to, anduse of, ICT by households and individuals (iii)use of ICT by businesses and (iv) ICT sectorand trade in ICT goods. The principal objectiveof the list is to help countries which are develop-ing ICT surveys, or adding ICT questions to ex-isting collections, to produce internationally com-parable data. The list is not mandatory in anysense - countries will always have to respond tonational policy needs, which may be only par-tially covered by the core list.

It is recognised that not all countries are at thesame level of development or have well devel-oped statistical systems. In respect of the indi-cators for which countries would need to collectdata (generally those on ICT access and use,and the ICT sector), countries with little or noICT infrastructure may not see the need to col-lect those ICT statistics at this stage. Countries


2

with growing use of ICT may want to monitorthis growth by starting to measure ICT usewhile others, with higher levels of ICT use andperhaps a growing ICT sector, may want togo further. To give further guidance, a distinc-tion has been made between ‘basic core’ and‘extended core’ indicators. The latter are gen-erally of lower priority and/or somewhat un-tested. They are therefore more suitable forcountries with relatively advanced ICT statis-tical systems.

An ICT infrastructure - and access to it - areessential prerequisites for benefiting from ICTs.Statistical indicators showing the extent of thatinfrastructure and access are therefore an ob-vious starting point for a core list. These indi-cators are generally already available in theITU World Telecommunication Indicators da-tabase. However, collecting the data is prov-ing more difficult in a world characterised bymultiple operators in liberalised markets. Thusthere is a need to define a core list upon whichregulators and other data providers can focus.

The main actors in a society are people, busi-nesses and governments, therefore the core listshould ideally contain indicators on the accessto, and use of, ICTs by these actors. In the caseof households/individuals and businesses, inter-national standards exist, although indicators arenot universally collected. The collection of datafor these indicators would therefore require ad-ditional efforts by a substantial number of coun-tries. In respect of governments, there are cur-rently no international standards and little com-parable collection work. At this stage, alterna-tives are being considered for so-called ‘e-gov-ernment’ indicators.

The core list also contains indicators relatingto the ICT sector and trade in ICT goods. In-

ternational standards exist for such indicatorsin the form of definitions and classifications.

As indicated above, most of the indicators inthe proposed core list benefit from being basedon internationally agreed standards (especiallythose developed by the ITU, OECD andEurostat). It is one of the objectives of thePartnership to consolidate standards at theglobal level, taking into consideration the de-mands and special circumstances of develop-ing countries.

The current set of ICT indicators is not in-tended to constitute a final list. Rather, the pro-cess will be continuous, and the list will un-dergo periodic review. As countries gain ex-perience with the collection of data for the in-dicators, and as policy needs evolve, indica-tors may be modified, removed or added.

The Partnership also has a continuing role indeveloping and proposing new core indicators.At the WSIS Thematic Meeting on Measur-ing the Information Society (Geneva, Febru-ary 2005), participants called upon the Part-nership to work on the development of newpolicy-relevant indicators in areas such as e-government, education and health and to pur-sue those topics at WSIS Tunis. Additionally,countries and regional organisations might de-velop additional indicators that are relevant andspecific to them, for example on Web contentlanguage, or access to ICT by ethnic and othersocial groups.

The Partnership has two other main objec-tives which are closely related to the develop-ment of a core list of indicators. The first ob-jective is to help build capacity in developingcountries, by activities such as methodologi-cal support, training and regional working

3

Core ICT Indicators

groups. A core list of indicators, and associ-ated statistical metadata, is instrumental in as-sisting these capacity building activities.

The second objective is to develop a global data-base on information society indicators. Actualdata collection is an obvious consequence of de-veloping and adopting a core list of indicators.The Partnership aims to bring these statisticstogether in a database that can be accessed eas-

Basic core

A1 Fixed telephone lines per 100 inhabitants

A2 Mobile cellular subscribers per 100 inhabitants

A3 Computers per 100 inhabitants

A4 Internet subscribers per 100 inhabitants

A5 Broadband Internet subscribers per 100 inhabitants

A6 International Internet bandwidth per inhabitant

A7 Percentage of population covered by mobile cellular telephony

A8 Internet access tariffs (20 hours per month), in US$, and as a percentage of per capita

income

A9 Mobile cellular tariffs (100 minutes of use per month), in US$, and as a percentage of per

capita income

A10 Percentage of localities with public Internet access centres (PIACs) by number of

inhabitants (rural/urban)

Extended core

A11 Radio sets per 100 inhabitants

A12 Television sets per 100 inhabitants

Core indicators on ICT infrastructure and access

ily via the Web.

The list of core indicators is presented below.The subsequent four sections discuss each setof indicators in detail, providing definitions, ex-planatory notes and model questions (where rel-evant), as well as methodological notes con-cerning the collection of these indicators. Thepaper ends by providing information on furtherstatistical resources for measuring the informa-tion society.

Core list of ICT indicators3


4

Core indicators on access to, and use of, ICT by households and individuals

Basic core

HH1 Proportion of households with a radio

HH2 Proportion of households with a TV

HH3 Proportion of households with a fixed line telephone

HH4 Proportion of households with a mobile cellular telephone

HH5 Proportion of households with a computer

HH6 Proportion of individuals who used a computer (from any location) in the last 12 months

HH7 Proportion of households with Internet access at home

HH8 Proportion of individuals who used the Internet (from any location) in the last 12 months

HH9 Location of individual use of the Internet in the last 12 months: (a) at home; (b) at work;

(c) place of education; (d) at another person’s home; (e) community Internet access facility

(specific denomination depends on national practices)4; (f) commercial Internet access facility

(specific denomination depends on national practices)5; and (g) others

HH10 Internet activities undertaken by individuals in the last 12 months·

• Getting information: (a) about goods or services; (b) related to health or health

services; (c) from government organisations/public authorities via websites or email;

and (d) other information or general Web browsing

• Communicating

• Purchasing or ordering goods or services

• Internet banking6

• Education or learning activities

• Dealing with government organisations/public authorities

• Leisure activities: (a) playing/downloading video or computer games;

(b) downloading movies, music or software; (c) reading/downloading electronic books,

newspapers or magazines; and (d) other leisure activities

Extended core

HH11 Proportion of individuals with use of a mobile telephone

HH12 Proportion of households with access to the Internet by type of access: Categories should

allow an aggregation to narrowband and broadband, where broadband excludes slower

speed technologies, such as dial-up modem, ISDN and most 2G mobile phone access.

Broadband will usually have an advertised download speed of at least 256 kbit/s.

HH13 Frequency of individual access to the Internet in the last 12 months (from any location): (a)

at least once a day; (b) at least once a week but not every day; (c) at least once a month but

not every week; and (d) less than once a month.

Reference indicator

HHR17 Proportion of households with electricity

5

Core ICT Indicators

Core indicators on use of ICT by businesses8

Basic core

B1 Proportion of businesses using computers

B2 Proportion of employees using computers

B3 Proportion of businesses using the Internet

B4 Proportion of employees using the Internet

B5 Proportion of businesses with a Web presence

B6 Proportion of businesses with an intranet

B7 Proportion of businesses receiving orders over the Internet

B8 Proportion of businesses placing orders over the Internet

Extended core

B9 Proportion of businesses using the Internet by type of access: Categories should allow an

aggregation to narrowband and broadband, where broadband excludes slower speed

technologies, such as dial-up modem, ISDN and most 2G mobile phone access. Broad-

band will usually have an advertised download speed of at least 256 kbit/s.

B10 Proportion of businesses with a Local Area Network (LAN)

B11 Proportion of businesses with an extranet

B12 Proportion of businesses using the Internet by type of activity

• Sending and receiving email

• Getting information: (a) about goods or services; (b) from government organisations/

public authorities via websites or email; and (c) other information searches or re-

search activities

• Performing Internet banking or accessing other financial services

• Dealing with government organisations/public authorities

• Providing customer services

• Delivering products online

Core indicators on the ICT sector and trade in ICT goods

Basic core

ICT1 Proportion of total business sector workforce involved in the ICT sector

ICT2 Value added in the ICT sector (as a percentage of total business sector value added)

ICT3 ICT goods imports as a percentage of total imports

ICT4 ICT goods exports as a percentage of total exports


6

Core Indicators on ICTInfrastructure and Access9

connection - typically a copper wire - from asubscriber to the telephone company’s switch-ing exchange. Technological change has sinceblurred this definition. For example, in somecountries, telephone service is provided via co-axial cable over pay television networks. Inothers, wireless local loop (WLL) technologysevers the traditional concept of the physicalline represented by a copper wire. The emer-gence of integrated services digital networks(ISDN) has also dramatically affected theconcept of the main line. ISDN converts asingle physical line into virtual channels. Ba-sic rate ISDN provides two channels whileprimary rate provides many more (e.g. 30 inEurope and 23 in North America and Japan).11

This led to the practice, particularly in Europeand Japan, of including ISDN channels in mainline statistics. More recently, the use of fixedline broadband technologies (such as digitalsubscriber line (DSL) or cable modem) hasmeant that a single subscriber connection cansupport much higher data connections, includ-ing data but also, potentially, several simulta-neous telephone conversations (for instance,using voice over IP). In order to enhance com-parability, all countries should provide a break-down of how their main telephone line figureis computed.

Ensuring universal service and access to infor-mation and communication technology is a topnational objective in many countries, often en-shrined in laws that govern the sector.10 The pro-posed ICT infrastructure and access indicatorscorrespond to individual use and measure ac-cessibility in terms of people. Most of the indi-cators are per capita measures which is the tra-ditional method of illustrating individual accessto ICTs. One reason for this is that virtually allICT service providers compile administrativerecords for operational and billing purposes. It isthen a simple mathematical exercise to dividethe installed base of a particular ICT device orservice by the population to derive a per capitaindicator.

Other indicators, for instance Percentage ofpopulation covered by mobile cellular tele-phony, are useful for tracking universal access,or the percentage of the population that couldtheoretically use an ICT device or service.

Definitions of basic core indi-cators

Fixed telephone lines per 100inhabitantsA fixed line has traditionally referred to the

7

Core ICT Indicators

Mobile cellular subscribers per 100inhabitants12

Mobile density, or the number of mobile sub-scribers per 100 inhabitants, has surpassed 100in some nations. This phenomenon of above100% ownership may arise due to a number offactors, including the practice in some countriesof issuing car phones as standard and becauseof the inactive use of pre-paid accounts. Statis-tics regarding mobile users should include bothpost-paid subscriptions and pre-paid accounts,and should distinguish between the two.

The indicator specifies ‘cellular’ subscribers,which refers to those mobile services which al-low for hand-over of calls between ‘cells’ as auser moves from the area covered by one mo-bile transmitter (or base station) to another.

While it is important to measure the total num-ber of lines in a country, for comparable statis-tics, it is more conventional to express this interms of lines per 100 inhabitants (teledensity).However, in the case of an increasing numberof countries, this is not particularly meaningfulnow that mobile cellular subscribers outnumberfixed lines. An alternative would be to measure‘mobidensity’, or mobile cellular subscribers per

Computers per 100 inhabitantsDespite their importance, only a few countries pub-lish reliable data on the number of personal com-puters (PCs). Unlike television sets, which are ba-sically found in homes or hotel rooms, collectingdata on PCs would involve surveying all the placeswhere there might be a PC: schools, homes, of-fices, libraries, Internet cafés, etc. Therefore, mostestimates regarding the installed base of PCs arederived from data on shipments (e.g. the numberof PCs sold each year) in a given country. Annualshipment data can be multiplied by an estimatedaverage useful lifespan before replacement to ob-tain an approximation of the number of PCs for thecountry. The life of a PC will vary depending onvarious factors such as wear and tear and obsoles-cence. Replacement rates differ between devel-oping and developed nations, with the former keep-

100 inhabitants (see indicator A2), but this woulddisadvantage those economies where the fixed-line network is well established, or others wherethe mobile network is still in its infancy. A pre-ferred alternative is therefore ‘effectiveteledensity’ which may be defined as either fixedlines or mobile subscribers per 100 inhabitants –whichever is higher.

Box A1: Fixed telephone linesper 100 inhabitants

Fixed telephone lines refer to telephone linesconnecting a customer’s terminal equipment(e.g. telephone set, facsimile machine) to thepublic switched telephone network (PSTN)and which have a dedicated port on a tele-phone exchange. Fixed telephone lines per100 inhabitants is calculated by dividing thenumber of fixed telephone lines by the popu-lation and multiplying by 100.

Box A2: Mobile cellular subscribersper 100 inhabitants

Mobile cellular subscribers refer to usersof portable telephones subscribing to an au-tomatic public mobile telephone service us-ing cellular technology, which provides ac-cess to the PSTN. Users of both post-paidsubscriptions and pre-paid accounts are in-cluded. Mobile cellular subscribers per100 inhabitants is obtained by dividing thenumber of mobile cellular subscribers by thepopulation and multiplying by 100.


8

ing PCs for longer.13 Though there is no precisemethodology for determining PC replacementrates, a general rule of thumb is that they arechanged every three to five years.

Apart from wear and tear, computers also becomeobsolete, as software updates require faster chipsand more memory. In light of all these factors, anoverall country figure for the number of PCs couldbe estimated by adding up the last five years’sales.14 It is a major drawback that, as with so manyother statistics, reliable data on the number of PCssold is not available for many developing nations.

A surrogate for PC sales is PC import figures,data that are sometimes available from customsdepartments of national governments. However,use of import data has limitations. Often only valuerather than volume data are available. Also, if PCswere assembled in the country from importedparts, they would not be counted. Customs datawould also not include undeclared imports. Addi-tionally, some of the imported PCs may be laterexported. Given the difficulties associated withestimating the number of PCs in use in a country,

and the importance that this measure carries withregard to the implementation of many ICT na-tional policies, a count of PCs using a census maybe useful to establish a reliable base.

Internet subscribers per 100 inhabit-ants15

Since the early days of Internet use, there havebeen estimates of the number of users. This sta-tistic is represented elsewhere in the core set,and is based on census data, surveys or otherestimates. The number of Internet subscribers- those paying for access to the Internet - is amore precise indicator of access than users.Subscription suggests a certain intensity of usesince it is unlikely one would be paying for Internetaccess unless it is being utilised regularly. It shouldbe noted that the number of subscribers mea-sures all those who are paying for Internet use,including the so-called ‘free Internet’ used bythose who pay via the cost of their telephonecall, those who pay in advance for a given amountof time (pre-paid) and those who pay for a sub-scription (either flat-rate or volume/usage based).

Many Internet users obtain access without pay-ing directly, either as the member of a house-hold, or from work or school. Therefore the num-ber of Internet users will always be much largerthan the number of subscribers, typically by afactor of 2-3 in developed countries or more indeveloping ones. The number of subscribers thussets a minimum threshold for the number of us-ers in a country. Data about the nature of theInternet subscription, such as paid or free, busi-ness or consumer, fixed or mobile, and the ac-cess speed (e.g. dial-up or broadband) and theaccess device (e.g. PC, mobile phone, gamesconsole) are also useful.16

Data about the nature of the Internet subscrip-tion, such as paid or free, business or consumer,fixed or mobile; the access speed (e.g. dial-up

Computers measures the number of comput-ers installed in a country. The statistic includesPCs, laptops, notebooks etc, but excludes ter-minals connected to mainframe and mini-com-puters that are primarily intended for shared use,and devices such as smart-phones and personaldigital assistants (PDAs) that have only some,but not all, of the components of a PC (e.g.they may lack a full-sized keyboard, a largescreen, an Internet connection, drives etc.).Computers per 100 inhabitants is obtainedby dividing the estimated number computers inuse by the population and multiplying by 100.

Box A3: Computers per 100 inhabitants

9

Core ICT Indicators

or broadband); and the access device (e.g. PC,mobile phone, game console) also are useful.17

Broadband Internet subscribers per100 inhabitantsBroadband may be defined as technologies thatprovide speeds of at least 256 kbit/s, where thisspeed is the combined upstream and downstreamcapacity.18 The number of broadband subscrib-ers refers to the number of active registeredInternet accounts including all fixed networkbroadband access technologies: digital subscriberline services (DSL), cable modem services, sat-ellite broadband Internet, fibre-to-the-homeInternet access, Ethernet LANs, wireless localarea network (WLAN) services etc.

International Internet bandwidth perinhabitantBandwidth refers to the range of frequenciesavailable to be occupied by signals. In analoguesystems it is measured in terms of Hertz (Hz)and in digital systems in bits per second (bit/s) -the higher the bandwidth, the greater the amountof information that can be transmitted in a giventime. International Internet bandwidth refersto the capacity which backbone operators provi-sion to carry Internet traffic measured in bits

per second. This indicator is intended to repre-sent the quality of the experience of Internetusers within a country. If the experience of anInternet user in a country is poor, because ofslow speed, then either people will not use ICTs,or they will not be able to use them effectivelyand creatively. In many developing countries, mostInternet access is to sites abroad and thereforethe amount of international bandwidth has a majorimpact on performance.

A useful derivative of this indicator is Interna-tional Internet bandwidth per Internet sub-scriber, which takes out of the equation the dif-fering levels of Internet usage in different coun-tries. It can be used to describe the level ofInternet access in countries, in the same waythat Gross Domestic Product (GDP) per capitais utilised to indicate the level of economic de-velopment in comparison with other countries.

One weakness of this indicator is that countrieswith a large domestic market (like China), whichare major generators of local content (like USA),

Box A4: Internet subscribers per100 inhabitants

An Internet subscriber is someone whopays for access to the public Internet (a TCP/IP connection). The statistic is measured ir-respective of the type or speed of access,the type of device used to access the Internet,or the method of payment. Internet subscrib-ers per 100 inhabitants is obtained by di-viding the number of Internet subscribers bythe population and multiplying by 100.

Box A5: Broadband Internet sub-scribers per 100 inhabitants

A Broadband Internet subscriber is some-one who pays for high-speed access to thepublic Internet (a TCP/IP connection). High-speed access is defined as being equal to, orgreater than 256 kbit/s, as the sum of thecapacity in both directions. The statistic ismeasured irrespective of the type of access,or the type of device used to access theInternet, or the method of payment. Broad-band Internet subscribers per 100 inhab-itants is obtained by dividing the number ofBroadband Internet subscribers by the popu-lation and multiplying by 100.


10

or which have languages that are not widely usedoutside the country (like Japan), may have rela-tively lower than expected levels of internationalInternet bandwidth, in that domestic bandwidthfulfils most of their needs. A further weakness isthat much international bandwidth on leased linesis paid for unilaterally (i.e. the costs are borne byusers at one end of the link) even though users atboth ends use link. Again, this tends to underesti-mate the real level of bandwidth available to us-ers in some countries, e.g. the United States.However, this statistic is valid for all developingcountries as well as for most developed ones.

Internet access tariff (20 hours permonth), in US$, and as a percentage ofper capita incomeThis indicator includes the tariff components ofmonthly line rental, line usage charge and Internetaccess charge but excludes the cost of equip-ment purchase or rental, and connection charges.It may be possible for the consumer to avoid oneor more of the tariff components (for instance,the cost of line rental may be waived for mobileInternet access; or the cost of line charge maybe bundled with the price of Internet access for‘free Internet’ services). The package chosenwould normally be the cheapest, but not neces-sarily the one offering the highest performance(for instance, a dial-up offering may be cheaperthan broadband for 20 hours usage per month).In the longer term, it would be preferable to re-place this indicator with an indicator that includesreference to the quality of the service (for in-stance, price per mbit/s per month). However,until broadband is more widely available, an indi-cator based on simple Internet access is moredirectly comparable.

Percentage of population coveredby mobile cellular telephonyOne of the most useful indicators of universalaccess is the percentage of the populationcovered by a mobile cellular network. Inhab-itants who are covered by a mobile cellular sig-nal have the potential to subscribe to the net-work, irrespective of whether or not they actu-ally do so. Where there is a large gap betweenpopulation coverage and penetration, it suggeststhat bottlenecks in access are more due toaffordability than to infrastructure shortcomings.

International Internet bandwidth refers tothe capacity which backbone operators provi-sion to carry Internet traffic measured in bitsper second. International Internet band-width per inhabitant is obtained by dividingthe amount of bandwidth by the population.

Box A6: International Internetbandwidth per inhabitant

Percentage of population covered by mo-bile cellular telephony refers to the percent-age of a country’s inhabitants that live withinareas served by a mobile cellular signal, irre-spective of whether or not they choose touse it. This should not be confused with thepercentage of the land area covered by amobile cellular signal or the percentage of thepopulation that subscribe to mobile cellularservice. Note that this measures the theoreti-cal ability to use mobile cellular services if onehas a cellular telephone and a subscription.

Box A7: Percentage of populationcovered by mobile cellular telephony

11

Core ICT Indicators

Mobile cellular tariffs (100 minutes ofuse per month), in US$, and as a per-centage of per capita incomeThe chosen indicator, comparing the cost of 100minutes of use per month, is intended to representan average use basket, which is applicable to indi-vidual consumers. Mobile users are split betweenpost-paid subscribers and pre-paid accounts, andfor some countries, and some operators, the formerwould be cheaper for 100 minutes use whereas inothers the latter would be cheaper.

Percentage of localities with publicInternet access centres (PIACs) bynumber of inhabitants (rural/urban)The final indicator in the basic core set of in-frastructure and access indicators is a measureof community access available to individuals.This is distinct from the measures available tohouseholds and to businesses measured else-where in the core set.

The Internet access tariff includes the tar-iff components of monthly line rental, lineusage charge and Internet access charge,plus any tax that may be levied (as this is aservice used by both residential and businessconsumers). The tariff chosen for a particu-lar country would be the package for 20 hoursper month that is the cheapest, that is widelyavailable (or, in the case of regional serviceproviders, is available in the capital city) andis available to the general public without re-striction (e.g. excluding in-company or lim-ited time offers, and excluding offers that arebundled with some other service). The pricecomparison is expressed in a commonly usedcurrency (such as US$), which could be con-verted either at the average exchange rate,or at purchasing power parity (PPP) rates.The indicator should be compared, as far aspossible, for the same date between coun-tries. As a percentage of per capita incomeinvolves dividing the Internet access tariff bythe average monthly gross national incomeper capita of the country.

Box A8: Internet access tariff (20hours per month), in US$, and as apercentage of per capita income

The Mobile cellular tariff includes the tariffcomponents of monthly service rental (if rel-evant), 50 minutes of local peak time callingand 50 minutes of local off-peak calling, plustax. Differences in the distance of calls, whichmay be applicable in some countries, are nottaken into account, nor are international callsor SMS messages. The possible one-timecharge for connection is not taken into account,except where this is bundled into the costs ofa pre-paid account. Countries should calcu-late the tariff either on a post-paid or a pre-paid service, whichever one is more popularlyused. If more than 50% of the mobile cellularsubscribers use pre-paid, then the tariff shouldalso be based on the pre-paid service, and viceversa. The price comparison is expressed in acommonly used currency (such as US$), whichcould be converted either at the average anexchange rate, or at purchasing power parity(PPP) rates. The indicator should be com-pared, as far as possible, for the same datebetween countries. As a percentage of percapita income involves dividing the mobilecellular tariff by the average monthly grossnational income per capita of the country.

Box A9: Mobile cellular tariffs (100minutes of use per month), in US$, andas a percentage of per capita income


12

There are different ways of measuring accessto community ICT facilities, for example,through the inclusion of a specific question inan Internet user survey. A specific questionshould address the location of Internet accessand allow respondents to choose between dif-ferent options, such as at home, at work, ateducational institutions, or at public Internetaccess facilities. This would provide informa-tion on the number or percentage of people us-ing public Internet facilities.

The main approach taken by the ITU is to lookat the percentage of localities with publicInternet access centres (PIACs) by numberof inhabitants (rural/urban). The availabilityof at least one facility in a locality is what isimportant here rather than the total number offacilities. Some countries, such as those in theEuropean Union, have started collecting dataon the number of public Internet access points.19

One major limitation with this indicator is that itdoes not show the distribution of facilities. Nordoes it provide a basis for a recommended valuesince this would be a function of how neces-sary public Internet access centres are in aparticular country (which in turn depends onthe underlying level of ICT ownership). Fur-thermore, localities may be defined differentlyand/or have varying sizes in different countries.In the case of countries where a major portionof population is centred around a metropolitanarea or the capital city, this may result in mis-leading values.

Definitions of extended coreindicators

The indicators for measuring broadcast penetra-tion through conventional ICTs are the numberof radio and television sets.20 Radio and tele-vision are still very dependent and useful ICTs

for many parts of the world where Internet-based ICTs are not yet available or affordable.Few countries collect the number of devicesand thus most data are estimates.21 These arederived from sales of sets or estimates basedon surveys which ask households whether theyhave a television. Some countries with licens-ing regimes collect data on the number of li-cences. This statistic is often used as a proxyfor household availability. However, not allpeople pay the licence fee so the true figure isunderrepresented. This is apparent when li-cence data is contrasted with data from censusor household surveys on the number of homeswith a broadcast reception set. For this reason,and because of the underlying unreliability ofthe data, these indicators are nominated as ‘ex-tended core’ rather than ‘basic core’.22

A public Internet access centre (PIAC) isa site, location, or centre of instruction atwhich Internet access is made available tothe public, on a full-time or part-time basis.This may include digital community centres,Internet cafés, libraries, education centres andother similar establishments, whenever theyoffer Internet access to the general public.All such centres should have at least one publiccomputer for Internet access. Localities re-fer to a country’s villages, towns and cities.The percentage of localities with publicInternet access centres (PIACs) is computedby dividing the number of localities with atleast one PIAC by the total number of thecountry’s localities and multiplying by 100.The indicator should be broken down by rangeof inhabitants.

Box A10: Percentage of localities withpublic Internet access centres (PIACs)by number of inhabitants (rural/urban)

13

Core ICT Indicators

A radio set is a device capable of receivingbroadcast radio signals, using popular frequen-cies, such as FM, AM, LW and SW. A radioset may be a standalone device, or it may beintegrated into another device, such as aWalkman, a car, or an alarm clock. Radiosets per 100 inhabitants is obtained by di-viding the number of radio sets in use by thepopulation and multiplying by 100.

Box A11: Radio sets per 100 inhabitants

A television set is a device capable of re-ceiving broadcast television signals, usingpopular access means such as over-the-air,cable and satellite. A television set may be astandalone device, or it may be integrated intoanother device, such as a computer or a mo-bile phone. It may be useful to distinguishbetween digital and analogue signal deliveryand between TV sets receiving only a lim-ited number of signals (usually over-the-air)and those that have multiple channels avail-able (e.g. by satellite or cable). Televisionsets per 100 inhabitants is obtained by di-viding the number of sets in use by the popu-lation and multiplying by 100.

Box A12: Television sets per 100inhabitants


14

Core Indicators on Access to, and Useof, ICT by Households and Individuals

This section provides definitions and model ques-tions corresponding to the household and individualcore ICT use indicators HH1 to HH13. It also con-tains some methodological notes, the most importantof which for comparability purposes are those onstatistical units, scope and classificatory variables.

It should be noted that this section does not pro-vide general advice on how to conduct or processhousehold surveys. Its aim is to convey thosemethodological points and conceptual issues whichare most relevant to the collection of household/individual ICT use data.

The suggestions are mainly based on developmentwork undertaken by the OECD’s Working Partyon Indicators for the Information Society (WPIIS),work being done by Eurostat and individual OECDmember countries on this topic, and feedback re-ceived on an earlier paper on this topic. Theseand other resources on ICT statistics are referredto at the end of this document.

A radio is a device capable of receivingbroadcast radio signals, using popular frequen-cies, such as FM, AM, LW and SW. Radiosalso include: those combined with other equip-ment such as cassette players/recorders, por-table radios such as transistor radios, and ra-dios in motor vehicles.

The proportion of households with a ra-dio is calculated by dividing the number ofin-scope households with a radio by the totalnumber of in-scope households. Sub-indica-tors may be constructed using the householdclassificatory variables, household compo-sition and household size (for details, see‘Methodological notes’ below), for example,the proportion of households with a radiowhere there are one or more children un-der 16 in the household.

Suggested model question:Does any member of this household/doyoua have access to a radio at home?a. The term ‘do you’ is included to coversingle person households. It does not refer toindividual activities.

Box HH1: Proportion of householdswith a radio

15

Core ICT Indicators

Definitions23 and modelquestions24 for basiccore indicators25

Proportion of households with a radioThis indicator refers to radio access (not use)by in-scope households.26 The equipment shouldbe in working order or expected to be returnedto working order soon.

Proportion of households with a TVThis indicator refers to television access (not use)by in-scope households. The equipment shouldbe in working order or expected to be returnedto working order soon.

Proportion of households with afixed line telephoneThis indicator refers to fixed line telephone ac-cess (not use) by in-scope households. The equip-ment should be in working order or expected tobe returned to working order soon.

A TV (television) is a device capable of re-ceiving broadcast television signals, usingpopular access means such as over-the-air,cable and satellite. A television set may be astandalone device, or it may be integrated intoanother device, such as a computer or a mo-bile phone. The proportion of householdswith a TV is calculated by dividing the num-ber of in-scope households with a TV by thetotal number of in-scope households. Sub-in-dicators may be constructed using the house-hold classificatory variables, household com-position and household size, for example,the proportion of households with a TVwhere there are no children under 16.

Suggested model question:Does any member of this household/doyou have access to a television at home?a

a. The term ‘do you’ is included to coversingle person households. It does not refer toindividual activities.

Box HH2: Proportion of householdswith a TV

Fixed telephone lines refer to telephone linesconnecting a customer’s terminal equipment(e.g. telephone set, facsimile machine) to thepublic switched telephone network (PSTN)and which have a dedicated port on a tele-phone exchange. The proportion of house-holds with a fixed line telephone is calcu-lated by dividing the number of in-scopehouseholds with a fixed line telephone by thetotal number of in-scope households. Sub-in-dicators may be constructed using the house-hold classificatory variables, household com-position and household size.

Suggested model question:Does this household have a fixed linetelephone at home?

Box HH3: Proportion of householdswith a fixed line telephone


16

Proportion of households with a mo-bile cellular telephoneThis indicator refers to access to (not use of) amobile phone by in-scope households. The equip-ment should be in working order or expected tobe returned to working order soon.

Proportion of households with acomputerThis indicator refers to access to (not use of) acomputer by in-scope households. The equipmentshould be in working order or expected to bereturned to working order soon.

Proportion of individuals who useda computer (from any location) inthe last 12 monthsThis indicator refers to use of computers in theprevious 12 months by in-scope individuals.27 Thatuse can be from any location, including work.

Mobile cellular telephones refer to portabletelephones subscribing to an automatic pub-lic mobile telephone service using cellular tech-nology, which provides access to the PSTN.Users of both post-paid subscriptions and pre-paid accounts are included. The proportionof households with a mobile cellular tele-phone is calculated by dividing the numberof in-scope households with a mobile cellulartelephone by the total number of in-scopehouseholds. Sub-indicators may be con-structed using the household classificatoryvariables, household composition andhousehold size.

Suggested model question:Does any member of this household/doyoua have access to a mobile telephoneat home?


Box HH4: Proportion of householdswith a mobile cellular telephone

A computer includes: a desktop, portable orhandheld computer (e.g. a personal digital as-sistant). It does not include equipment withsome embedded computing abilities such asmobile phones or TV sets. Note that this is adifferent definition of a computer than theone used for indicator A3, with the main dif-ference being that personal digital assistants(PDA) are included here but excluded fromA3. There are practical and historical rea-sons for this difference but, for the purposesof indicators HH5 and HH6, the functionalityof PDAs, which may include Internet con-nectivity, is of interest.

The proportion of households with a com-puter is calculated by dividing the number ofin-scope households with a computer by thetotal number of in-scope households. Sub-in-dicators may be constructed using the house-hold classificatory variables, household com-position and household size.

Suggested model question:Does any member of this household/doyoua have access to a computer at home?


Box HH5: Proportion of householdswith a computer

17

Core ICT Indicators

Proportion of households withInternet access at homeThis indicator refers to access to (not use of)the Internet by in-scope households. The con-nection should be functional (that is, any equip-ment or software needed should be in workingorder) or expected to be returned to workingorder soon.

A computer includes: a desktop, portable orhandheld computer (e.g. a personal digital as-sistant). It does not include equipment withsome embedded computing abilities: such asmobile phones or TV sets.

The proportion of individuals who used acomputer etc is calculated by dividing the totalnumber of in-scope individuals who used acomputer from any location in the last 12months by the total number of in-scope indi-viduals. Sub-indicators may be constructedusing the individual classificatory variables,age, gender, highest education level, em-ployment status and occupation (see ‘Meth-odological notes’ below). An example of sucha sub-indicator is the proportion of individu-als with lower secondary education or lesswho used a computer.

Suggested model question:Have you used a computer in the last 12months?

Box HH6: Proportion of individualswho used a computer (from anylocation) in the last 12 months

The Internet is a world-wide public computernetwork. It provides access to a number ofcommunication services including the WorldWide Web and carries email, news, entertain-ment and data files. Access is not assumedto be only via a computer - it may also be bymobile phone, games machine, digital TV etc.The proportion of households with Internetaccess at home is calculated by dividing thenumber of in-scope households with Internetaccess by the total number of in-scope house-holds. Sub-indicators may be constructed us-ing the household classificatory variables,household composition and household size.

Suggested model question:Does any member of this household/doyoua have access to the Internet at homeregardless of whether it is used?


Box HH7: Proportion of householdswith Internet access at home


18

Proportion of individuals who usedthe Internet (from any location) inthe last 12 monthsThis indicator refers to use of the Internet in theprevious 12 months by in-scope individuals. Thatuse can be from any location, including work.

Location of individual use of theInternet in the last 12 monthsThis indicator refers to in-scope individuals whoused the Internet (from any location) in the pre-vious 12 months and specifies the location ofthat use. (See following page.)

The Internet is a world-wide public computernetwork. It provides access to a number ofcommunication services including the WorldWide Web and carries email, news, entertain-ment and data files. Individuals may have ac-cessed the Internet by any means including acomputer, mobile phone, games machine, digi-tal TV etc.

The proportion of individuals who used theInternet etc is calculated by dividing the totalnumber of in-scope individuals who used theInternet (from any location) in the last 12months by the total number of in-scope indi-viduals. Sub-indicators may be constructedusing the individual classificatory variables,age, gender, highest education level, em-ployment status and occupation. An ex-ample of such a sub-indicator is the propor-tion of individuals aged 16-24 who usedthe Internet.

Suggested model question:Have you used the Internet in the last12 months?

Box HH8: Proportion of individualswho used the Internet (from anylocation) in the last 12 months

19

Core ICT Indicators

Location of use includes home, work, place of education, another person’s home, commu-nity Internet access facility, commercial Internet access facility and other places. Individualscan respond in respect of more than one location.

For international comparability, output is most simply presented as the proportion of in-scopeindividuals using the Internet at each location, for instance, the proportion of individuals using theInternet at home, at work etc. An alternative presentation is the proportion of Internet users usingthe Internet from each location (for a discussion of this, see ‘Methodological notes’ below). Sub-indicators may be constructed using the individual classificatory variables, age, gender, highesteducation level, employment status and occupation. An example of such a sub-indicator is theproportion of employed persons who used the Internet at work.

Suggested model questiona:Where did you use the Internet in the last 12 months?b

• Home

• Work: Where a person’s workplace is located at his/her home, then he/she would answer yesto the home category only.

• Place of education

• At another person’s home

• Community Internet access facility: Includes access at community facilities such as publiclibraries, publicly provided Internet kiosks, other government agencies; access is typically freeor low cost. This category would not appear on country questionnaires. It is expected thateach country would tailor categories according to the facilities available.

• Commercial Internet access facility: Includes access at Internet or cyber cafés, hotels, air-ports; even though the venue is commercial, the cost is not necessarily at full market price.This category would not appear on country questionnaires. It is expected that each countrywould tailor categories according to the facilities available.

• Other places

a. The question is asked of all in-scope individuals who used the Internet in the last 12 months.b. Possible country variations to the response categories are to add or split locations according tocountry data requirements.

Box HH9: Location of individual use of the Internet in the last 12 months


20

Internet activities undertaken byindividuals in the last 12 monthsThis indicator refers to the activities undertakenon the Internet by in-scope individuals in the pre-vious 12 months, from any location including work.

Internet activities are: use of the Internetfor getting information (several responsecategories per the model question below), forcommunicating, for purchasing or order-ing goods or services, for Internet bank-ing, for education or learning activities,for dealing with government organisationsand for leisure activities (several responsecategories per the model question below).Note that these activities are restricted toprivate purposes and therefore exclude ac-tivities such as purchasing over the Internetundertaken as part of a person’s job or un-dertaking online courses as part of a job. In-dividuals can respond in respect of more thanone activity and activities are not mutuallyexclusive.

For international comparability, output is mostsimply presented as the proportion of in-scopeindividuals undertaking each activity, for in-stance, the proportion of individuals using theInternet to get information about goods or ser-vices. An alternative presentation is the pro-portion of Internet users undertaking eachactivity. Sub-indicators may be constructedusing the individual classificatory variables,age, gender, highest education level, em-ployment status and occupation.

See Box HH10.B

Box HH10: Internet activities under-taken by individuals in the last 12months

21

Core ICT Indicators

Suggested model questiona:For which of the following activitiesb did you use the Internet for private purposes in the last 12 months?c

For getting information• About goods or services• Related to health or health services: Health information covers injury, disease, nutrition and improv-

ing health generally.• From government organisations/public authorities (from Web sites or via email): Government organisations

public authorities are preferably defined per the SNA93. They include government organisations at local,regional and national level. See http://unstats.un.org/unsd/sna1993/wglossform.asp?getitem=219.

• Other information or general Web browsingFor communicating: Sending or receiving email, using chat rooms/sites, message boards, instant messag-ing, telephoning via Internet etc.For purchasing or ordering goods or services: Includes purchasing and downloading of digitised prod-ucts, such as music, from the Internet.For Internet bankingFor education or learning activities: This refers to formal learning activities such as study associated withschool or tertiary education courses as well as distance education involving online activities. (A morenarrow interpretation is likely to be less meaningful as it could include a range of activities such as usingthe Internet to search for information.)For dealing (interacting) with government organisations/public authorities: Government organisations/public authorities are preferably defined per the SNA93. They include government organisations at local,regional and national level. See http://unstats.un.org/unsd/sna1993/glossform.asp?getitem=219.For leisure activities:• Playing or downloading video games or computer games: Includes file sharing games and playing

games online.• Downloading or watching movies, TV programs, music or software: Includes file sharing and using

Web radio or Web television. For software, includes downloading of patches and upgrades.• Reading or downloading electronic books, newspapers or magazines: Includes accessing news Web

sites.• Other leisure activities: Includes gambling.

a. The question is asked of all in-scope individuals who used the Internet in the last 12 months.b. The activities are not necessarily mutually exclusive, for instance, some activities (such as dealing withgovernment and leisure activities) may also involve purchasing online. In such cases, the individualresponds in respect of more than one response category.c. There are alternative ways of asking activities questions. For instance, each could be rated according toits frequency or intensity of use. This question uses a simplified method of presentation which asksrespondents for all activities. Possible country variations to the response categories are to add or splitcategories according to country data requirements. In particular, countries may wish to add an ‘Other’category to this question.

Box HH10.B:


22

Definitions and modelquestions for extended coreindicators

Proportion of individuals with use ofa mobile telephoneThis indicator refers to in-scope individuals whohad personal use of a mobile phone for some pe-riod within the last 12 months. The question belowis untested in this format so the indicator is includedas an extended rather than basic core indicator.

Proportion of households withaccess to the Internet by type ofaccessThis indicator refers to in-scope households andthe Internet access service/s they use at hometo access the Internet. The response categoriesare designed to enable aggregation to broadbandand narrowband access. Broadband is definedin terms of technologies; more generally, suchtechnologies provide advertised download speedsof at least 256 kbit/s.

Mobile telephones refer to portable tele-phones subscribing to an automatic publicmobile telephone service using cellular tech-nology, which provides access to the PSTN.Use of a mobile telephone does not meanthat the telephone is owned or paid for by theperson but should be reasonably availablethrough work, a friend or family member, etc.It excludes occasional use, for instance, bor-rowing a mobile phone to make a call. Theproportion of individuals with use of a mo-bile telephone is calculated by dividing thetotal number of in-scope individuals with useof a mobile telephone by the total number ofin-scope individuals. Sub-indicators may be con-structed using the individual classificatory vari-ables, age, gender, highest education level,employment status and occupation.

Suggested model question:Did you have personal use of a mobiletelephone during some or all of the last12 monthsa?

a. Countries may wish to consider elaborat-ing this question to ask whether individualsaccessed the Internet using a mobile phone.Such an indicator could be relevant for coun-tries with rapidly growing mobile phone us-age but otherwise relatively poor telecommu-nications infrastructure.

Box HH11: Proportion of individualswith use of a mobile telephone

23

Core ICT Indicators

A major aim of this indicator is to present the proportion of households with broadband access,therefore the response categories chosen allow aggregation to narrowband and broadband. Ashouseholds can use more than one type of access service, multiple responses are possible.Forinternational comparability, output is most simply presented as the proportion of in-scope house-holds using each type of access service, for instance, the proportion of households accessing theInternet by DSL. Additionally, output should be available for the aggregations, the proportion ofhouseholds with broadband and narrowband access to the Internet. Alternatively, outputcould be presented as a proportion of households with Internet access. Sub-indicators may beconstructed using the household classificatory variables, household composition and house-hold size.

Suggested model questiona:What type/s of Internet access services are used for Internet access at home?b

Analog modem (dial-up via standard phone line): An analog modem converts a digital signalinto analog for transmission by traditional (copper) telephone lines. It also converts analog trans-missions back to digital.ISDN (Integrated Services Digital Network): ISDN is a telecommunication service thatturns a traditional (copper) telephone line into a higher speed digital link. It is usually regarded asnarrowband.DSL (ADSL, SDSL, VDSL etc.): Digital subscriber line; it is a high-bandwidth, local looptechnology carrying data at high speeds over traditional (copper) telephone lines.Cable modem: A modem which uses cable TV lines for connection to the Internet.Other narrowband: Including most mobile telephone access and other forms of access with anadvertised download speed of less than 256 Kbps (kilobits per second). This wording would notnormally be used on questionnaires - countries would add appropriate category/ies based onservices available.Other broadband: Including optic fibre cable, some mobile telephone access (e.g. UMTS),powerline, satellite, fixed wireless, with an advertised download speed of >= 256 Kbps. Thiswording would not normally be used on questionnaires - countries would add appropriate cat-egory/ies based on services available.Do not know

a. The question is asked of all in-scope households with access to the Internet at home.b. Possible country variations to the response categories are: remove categories where items arenot feasible, add or split categories according to technologies available and country data require-ments.

Box HH12: Proportion of households with access to the Internet by type of access


24

Frequency of individual access to theInternet in the last 12 months (from anylocation)This indicator refers to the frequency of (typi-cal) use of in-scope individuals who used theInternet, from any location including work, in theprevious 12 months.

Methodological notes

This section provides limited suggestions regard-ing methodology for collection of household andindividual ICT use indicators. For most countries,there will be established procedures in place forhousehold collections (e.g. labour force surveys)which will apply also to the collection of ICT usedata. For this reason, no recommendations aremade on the following:

• use of a particular type of sample frame, sam-pling methodology or sample size, these willvary according to country practices and theavailability of information (for instance, ad-ministrative information on individuals);

• whether collections should be mandatory orvoluntary – both types of collections are usedamong countries; however, where collectionsare voluntary, non-response will tend to behigher;

• how to process collected information, includ-ing editing, imputation and benchmarking ofdata; or,

• a particular type of survey vehicle (for in-stance, labour force surveys, household ex-penditure surveys etc.).

Frequency of use can be: at least once aday, at least once a week but not everyday, at least once a month but not everyweek, or less than once a month. For in-ternational comparability, output is most sim-ply presented as the proportion of in-scopeindividuals using the Internet with each fre-quency, for instance, the proportion of indi-viduals using the Internet at least once a day.An alternative presentation is the proportionof Internet users using the Internet witheach frequency. Sub-indicators may be con-structed using the individual classificatoryvariables, age, gender, highest educationlevel, employment status and occupation.

Suggested model questiona:How often did you typicallyb use theInternet during the last 12 months?c

At least once a dayd

At least once a week but not every dayAt least once a month but not every weekLess than once a month

a. The question is asked of all in-scope indi-viduals who used the Internet in the last 12months.b. ‘Typically’ means on most days (or a typi-cal day). It is included so that respondentswill ignore weekends (if they only accessthe Internet from work) and breaks fromtheir usual routine, such as holidays.c. From any location, including work.d. To be interpreted as at least once eachworking day for respondents who only (ormost frequently) use the Internet from work.

Box HH13: Frequency of individualaccess to the Internet in the last 12months (from any location)

25

Core ICT Indicators

Particular statistical issues associ-ated with household and individualICT use measurementMobile phonesA particular issue concerns mobile phones, forwhich there are both household and individualindicators in the core list. In OECD countries,this information has conventionally been collectedin respect of the household (therefore whetherthe household, through one or more of its mem-bers, has access to a mobile phone). However,there is increasing recognition that it is more im-portant to examine individual use of mobile phonesas they are typically owned and/or used by anindividual rather than a group of people. The corelist therefore includes an indicator on individualmobile phone use. However, it should be notedthat this indicator is based on a relatively un-tested question which may change as experienceincreases.

Recall periodBased on the experience of those OECD coun-tries which use it, it is suggested that a 12-monthrecall period for questions which are less relianton detailed recollection will not introduce signifi-cant recall bias effects. On the other hand, usinga different period (for example, three months)potentially introduces seasonal bias effects.28

A 12-month period is therefore suggested for thecore indicators in order to avoid such seasonaleffects and to also better capture less frequentactivities such as purchasing online or searchingfor health information.

Choice of denominator for indicatorsMost indicators arising from ICT use surveysare presented as proportions data. They include:proportions of the whole population of house-holds/individuals or of sub-populations, such asparticular household types or age groups. Addi-

tionally, countries might present data as a pro-portion of households/individuals which haveaccess to or use computers or the Internet etc.However, this can be potentially confusing tousers so it is important to be clear which de-nominator is used to construct a particular indi-cator and to have a common approach betweencountries.

It has been suggested in this paper that, for in-ternational comparability purposes, the simplestapproach is for countries to provide proportionswhich use the overall population as the de-nominator rather than the active population(being the number of households/individualswhich have access to, or use, computers or theInternet etc.). Note that, for presentation pur-poses, it is possible to calculate indicators withthe active population as the denominator fromother output which is proposed in this paper. Forexample, the proportion of households withInternet access which use broadband is equal tothe proportion of households with broadbandaccess divided by the proportion of householdswith Internet access.

Survey vehicles and collectiontechniquesThere is a variety of survey vehicles which couldbe used to collect data on household and indi-vidual ICT use. A few OECD countries have astand-alone survey for this purpose but most useexisting surveys. Many other countries will alsohave existing survey vehicles, for instance, labourforce surveys, household expenditure surveys,population censuses or general social surveys.

A number of developing countries are likely tostart collecting ICT household indicators by add-ing a few questions to existing surveys. In suchcases, the following suggestions on collectiontechniques may not be practicable to implement.


26

Most OECD countries use personal interviewtechniques for collecting data on household andindividual use of ICT. Face-to-face interviewingmay be better for some situations in that it po-tentially allows the interviewer to explain anytechnical terms and to check the type of Internetconnection or other technical details.

Telephone interviews should generally beavoided where telephone penetration is low or itis not possible to include mobile or unlisted sub-scribers. It is also suggested that postal surveysgenerally not be used. The reasons for this arethat they do not involve interaction with the re-spondent (therefore technical questions on ICTuse may be less well understood) and responserates are likely to be lower, leading to higher sam-pling error and possible non-response bias. Insituations where there are several languages ordialects spoken in a country, personal interviewusing local interviewers is likely to be preferableto postal surveys.

Statistical unitsA key issue concerns the appropriate statisticalunit for measurement. In general, the householdunit is used to elicit information about the facili-ties in place in the household (for example,whether there is a TV, computer or Internet con-nection). The individual unit is used to provideinformation on use of these facilities (both in andaway from the home) and, most importantly, theintensity of that use (for instance, frequency andrange of activities undertaken).

Both households and individuals are proposed asstatistical units. Commonly, a randomly selectedin-scope individual would answer all the ques-tions - the household model questions on behalfof the household and the individual questions inrespect of him/herself. Alternatively, more thanone household member could provide individual

information about themselves. Households, andindividuals within those households, should beselected in an unbiased manner.

Because the sample of households and individu-als selected is unlikely to be perfectly represen-tative of the population, it is important to weightresponses according to independent estimateddistributions of the population.

FrequencyFrequency of surveys is likely to relate to na-tional priorities, available resources and the evo-lution of the ICT market in each country. WhereICTs are being introduced rapidly, an annual sur-vey may be justifiable. Where the level of ac-cess to ICT is low and not growing rapidly, itmight be more appropriate to adopt a longer in-terval between surveys.

Survey scope and coverageIndividualsThe scope of individuals would normally be lim-ited by age. OECD and Eurostat recommend thatall individuals aged 16 to 74 years be included inthe scope of their household model surveys onICT use.29 For consistency, this age range is sug-gested as a minimum individual age scope.Many developing countries may wish to extendthis scope where there are both sufficient re-sources and supporting national data requirements.

In terms of overall coverage, the sample basewould normally be the complete in-scope resi-dent population of a country although countriesmight consider limiting the survey to those livingin private dwellings (therefore excluding individu-als in institutions such as prisons, nursing homesand special dwellings such as hotels).

Generally, countries will want to approach thequestion of scope and coverage with a sense ofwhat can practically be achieved and what ICT

27

Core ICT Indicators

facilities are available to particular social groupsor in particular geographic areas. In the case ofgeographical coverage, it might be unrealistic toask questions about ICT usage in remote ruralareas where it is well known that computers andthe Internet are not generally available. Coun-tries will consider such issues in relation to timeand cost, but they should also consider that theomission of any element of the resident popula-tion should be mentioned in all publications, andthat such an omission may affect the calculationof statistics representative of the whole popula-tion as well as comparison with other countries.

HouseholdsConsistent with the recommended individual agescope of 16-74, it is suggested that householdsconsisting only of members over 74, or under 16,are excluded from the survey. Other scope andcoverage limitations would follow from those ap-plying to individuals, for instance, restrictinghousehold surveys to those households in privatedwellings and omitting remote areas from cov-erage. As with individual scope, countries areable to extend the household scope if they wish.

Classificatory variablesThe metadata associated with the core indica-tors will include a set of classificatory variables,with relevant categories. Data for these vari-ables would usually be collected as part of thesurvey (though note that questions are not sug-gested in this paper). A minimal set of classifica-tory variables is proposed below, though manycountries will decide to use extra variables and/or additional categories. Of possible interest tothose countries where a rural/urban divide ex-ists is a geographical classification. For manycountries, an income variable will also be of in-terest. Because both variables are problematicfrom a collection and international comparabilityviewpoint, they have not been suggested in thispaper. However, countries are encouraged to

include those variables if they are able to, as theoutput can provide critical policy information.

Household characteristics

• Household composition (two-way classifica-tion: households with/without children un-der 16) and

• Household size (number of members, includ-ing those outside the age scope).

Individual characteristics

• Age: to show the differences between agegroupings, reasonably fine and equal-sizedranges are proposed as follows: 16 to 24;25 to 34; 35 to 44; 45 to 54; 55 to 64; 65to 74.30

• Gender.

• Highest education level received: a four-wayclassification is proposed as follows: No for-mal education or primary education(ISCED 0,1); Lower secondary education(ISCED 2); Upper secondary or post-sec-ondary non-tertiary (ISCED 3,4); Tertiary(ISCED 5,6).31

• Employment status (four-way classification:paid employee; self-employed,32 unem-ployed; not in the labour force);33 and,

• Occupation (use ISCO88 major groups wherepossible).34

In terms of output, many countries may want tocross-classify some of these variables. This canproduce information which is very useful foranalytical purposes. However, it should be notedthat cross-classified output is often more detailedand therefore requires higher sample sizes tosupport reliable estimates.


28

Core Indicators on Useof ICT by Businesses

This section provides definitions and model ques-tions corresponding to the business core ICT useindicators B1 to B12. It also provides some meth-odological notes, the most important of whichare those relating to statistical unit, scope andclassificatory variables. It should be noted thatthis section does not provide general advice onhow to conduct or process business surveys. Itsaim is to convey those methodological points andconceptual issues which are most relevant to thecollection of business ICT use information.

The suggestions are mainly based on developmentwork undertaken by the OECD’s Working Partyon Indicators for the Information Society (WPIIS),work being done by Eurostat and individual OECDmember countries on this topic, and feedback re-ceived on an earlier paper on this topic. Theseand other resources on ICT statistics are referredto at the end of this document.

Definitions and modelquestions35 for basic coreindicators

Proportion of businesses usingcomputersThis indicator refers to the use of computers byin-scope businesses36 during the 12-month ref-erence period.

A computer includes: a desktop, portable orhandheld computer (e.g. a personal digital as-sistant), minicomputer, mainframe. A com-puter does not include equipment with someembedded computing abilities: such as mo-bile phones or TV sets, nor does it includecomputer-controlled machinery or electronictills. The proportion of businesses usingcomputers is calculated by dividing the num-ber of in-scope businesses using computersduring the 12-month reference period by thetotal number of in-scope businesses. Sub-in-dicators may be constructed using the clas-sificatory variables, industry and size (fordetails, see ‘Methodological notes’ below). Anexample of a sub-indicator is the proportionof manufacturers using computers.

Suggested model question:Did your business use computer/s dur-ing <period>a?

a. 12-month reference period.

Box B1: Proportion of businessesusing computers

29

Core ICT Indicators

Proportion of businesses using theInternetThis indicator refers to the use of the Internet byin-scope businesses during the 12-month refer-ence period – whether or not the business useda computer (as the Internet may be accessed inother ways).

Proportion of employees usingcomputersThis indicator refers to the proportion of personsemployed by in-scope businesses who routinelyused a computer during the 12-month referenceperiod. The indicator refers to actual use of acomputer by employees rather than simply ac-cess to a computer.

A computer includes: a desktop, portable orhandheld computer (e.g. a PDA), minicom-puter, mainframe. A computer does not in-clude equipment with some embedded com-puting abilities: such as mobile phones or TVsets, nor does it include computer-controlledmachinery or electronic tills.

Employees refer to all persons working forthe business, not only those working in cleri-cal jobs. They include working proprietors andpartners, as well as employees. The propor-tion of employees using computers is cal-culated by dividing the number of employeesusing computers (in all in-scope businesses)by the total number of employees (in all in-scope businesses). Sub-indicators may beconstructed using the classificatory variables,industry and size.

Suggested model questiona:What proportion of persons employed inyour business routinely used a computerat work during <period>b?

a. The question is asked of all in-scope busi-nesses which used computer/s during the ref-erence period.b. 12-month reference period.

Box B2: Proportion of employeesusing computers

The Internet refers to Internet protocol (IP)based networks: WWW (the World WideWeb), an extranet over the Internet, EDI overthe Internet, Internet accessed by mobilephones and Internet email. The proportionof businesses using the Internet is calcu-lated by dividing the number of in-scope busi-nesses using the Internet by the total numberof in-scope businesses. Sub-indicators maybe constructed using the classificatory vari-ables, industry and size.

Suggested model questiona:Did your business use the Internet dur-ing <period>b?

a. The question is asked of all businesses -not only those businesses which used a com-puter (as the Internet may be accessed inother ways).b. 12-month reference period.

Box B3: Proportion of businessesusing the Internet


30

Proportion of employees using theInternetThis indicator refers to the proportion of personsemployed by in-scope businesses who routinelyused the Internet during the 12-month referenceperiod. The indicator refers to actual use of theInternet by employees rather than simply accessto the Internet.

Proportion of businesses with aWeb presenceThis indicator is a measure of the proportion of in-scope businesses with a Web presence as at thereference date. ‘Web presence’ is used rather than‘Web site’, based on the presumption that a pres-ence is more important than a Web site per se.

The Internet refers to Internet protocol (IP)based networks: WWW, an extranet over theInternet, EDI over the Internet, Internet ac-cessed by mobile phones and Internet email.Employees refer to all persons working forthe business, not only those working in cleri-cal jobs. They include working proprietors andpartners, as well as employees. The propor-tion of employees using the Internet is cal-culated by dividing the number of employeesusing the Internet (in all in-scope businesses)by the total number of employees (in all in-scope businesses). Sub-indicators may beconstructed using the classificatory variables,industry and size.

Suggested model questiona:What proportion of persons employed inyour business routinely used the Internetat work during <period>b?

a. The question is asked of all in-scope busi-nesses which used the Internet during the ref-erence period.b. 12-month reference period.

Box B4: Proportion of employeesusing the Internet

A Web presence includes a Web site, homepage or presence on another entity’s Web site(including a related business). It excludes in-clusion in an online directory and any otherWeb pages where the business does not havesubstantial control over the content of thepage. The proportion of businesses with aWeb presence is calculated by dividing thenumber of in-scope businesses with a Webpresence by the total number of in-scope busi-nesses. Sub-indicators may be constructedusing the classificatory variables, industry andsize.

Suggested model questiona:Did your business have a Web presenceas at <reference date>b?

a. The question is asked of all in-scope busi-nesses which used the Internet during thereference period. In theory, a business couldhave a Web presence without having usedthe Internet. For most countries, this is ex-pected to be a rare event. Where it is thoughtto be common, countries could alter the popu-lation to all in-scope businesses which usedcomputer/s during the reference period.b. The reference date would usually be atthe end of the reference period, or shortlyafter.

Box B5: Proportion of businesseswith a Web presence

31

Core ICT Indicators

Proportion of businesses with anintranetThis indicator is a measure of the proportion ofin-scope businesses with an intranet as at thereference date.

Proportion of businesses receivingorders over the InternetThis indicator refers to selling over the Internetby in-scope businesses during the 12-month ref-erence period.

An intranet refers to a network using thesame protocol as the Internet and allowingcommunication within an organisation. It istypically set up behind a firewall to controlaccess. The proportion of businesses withan intranet is calculated by dividing the num-ber of in-scope businesses with an intranetby the total number of in-scope businesses.Sub-indicators may be constructed using theclassificatory variables, industry and size.

Suggested model questiona:Did your business have an intranet as at<reference date>b?a. The question is asked of all in-scope busi-nesses which used computer/s during the ref-erence period.b. The reference date would usually be atthe end of the reference period, or shortlyafter.

Box B6: Proportion of businesseswith an intranet

Orders include orders received via the Internetwhether or not payment was made online. Thisincludes orders received via Web sites, specialisedInternet marketplaces, extranets, EDI over theInternet, Internet-enabled mobile phones andemail. It also includes orders received on behalfof other organisations and orders received byother organisations on behalf of the business. Itexcludes orders which were cancelled or not com-pleted. For international comparability, the pro-portion of businesses receiving orders overthe Internet is most simply calculated by divid-ing the number of in-scope businesses receivingorders over the Internet by the total number ofin-scope businesses. Alternatively, output couldbe presented as the proportion of in-scope busi-nesses using the Internet (for a discussion of this,see ‘Methodological notes’ below). Sub-indica-tors may be constructed using the classificatoryvariables, industry and size.

Suggested model questiona:Did your business receive ordersb for goodsor services (that is, make sales) via theInternet during <period>c?a. The question is asked of all in-scope businesseswhich used the Internet during the reference pe-riod. In theory, businesses without Internet ac-cess could receive Internet orders (for example,via agents). Where this is thought to be common,countries could alter the population to all in-scopebusinesses which used computer/s during thereference period.b. Note that the inclusion of email orders differsfrom the recommendation made to OECD andEurostat countries to exclude orders made overconventional email. Notwithstanding this, a smallnumber of OECD countries do include conven-tional email orders as Internet commerce.c. 12-month reference period.

Box B7: Proportion of businessesreceiving orders over the Internet


32

Proportion of businesses placingorders over the InternetThis indicator refers to purchasing over theInternet by in-scope businesses during the 12-month reference period.

Definitions and modelquestions for extendedcore indicators

Proportion of businesses using theInternet by type of accessThis indicator refers to in-scope businesses andthe Internet access service/s they used to ac-cess the Internet during the reference period.The response categories are designed to enableaggregation to broadband and narrowband.Broadband is defined in terms of technologies;more generally, such technologies provide ad-vertised download speeds of at least 256 kbit/s.

Orders include orders placed via the Internetwhether or not payment was made online.Includes orders placed via Web sites,specialised Internet marketplaces, extranets,EDI over the Internet, Internet-enabled mo-bile phones and email. Excludes orders whichwere cancelled or not completed. For inter-national comparability, the proportion of busi-nesses placing orders over the Internet ismost simply calculated by dividing the num-ber of in-scope businesses placing orders overthe Internet by the total number of in-scopebusinesses. Alternatively, output could be pre-sented as the proportion of in-scope busi-nesses using the Internet. Sub-indicators maybe constructed using the classificatory vari-ables, industry and size.

Suggested model questiona:Did your business place ordersb for goodsor services (that is, make purchases) viathe Internet during <period>c?a. The question is asked of all in-scope busi-nesses which used the Internet during thereference period.b. Note that the inclusion of email orders dif-fers from the recommendation made to OECDand Eurostat countries to exclude orders madeover conventional email. Notwithstanding this,a small number of OECD countries do in-clude conventional email orders as Internetcommerce.c. 12-month reference period.

Box B8: Proportion of businessesplacing orders over the Internet

33

Core ICT Indicators

A major aim of this indicator is to present the proportion of in-scope businesses with broadbandaccess, therefore the response categories chosen allow aggregation to narrowband and broad-band. As businesses can use more than one type of access service, multiple responses arepossible.For international comparability, output is most simply presented as the proportion of in-scope businesses using each type of access service, for instance, the proportion of businessesaccessing the Internet by DSL. Additionally, output should be available for the aggregations, theproportion of businesses with broadband and narrowband access to the Internet. Alterna-tively, output could be presented as a proportion of businesses using the Internet. Sub-indicatorsmay be constructed using the classificatory variables, industry and size.

Suggested model questiona:How did your business connect to the Internet during <period>b?c

Analog modem (dial-up via standard phone line): An analog modem converts a digitalsignal into analog for transmission by traditional (copper) telephone lines. It also converts analogtransmissions back to digital.ISDN (Integrated Services Digital Network): ISDN is a telecommunication service thatturns a traditional (copper) telephone line into a higher speed digital link. It is usually regarded asnarrowband.DSL (ADSL, SDSL, VDSL etc.): Digital subscriber line; it is a high-bandwidth, local looptechnology carrying data at high speeds over traditional (copper) telephone lines.Cable modem:A modem which uses cable TV lines for connection to the Internet.Other narrowband: Including most mobile phone access and other forms of access with anadvertised download speed of less than 256 Kbps (kilobits per second). This wording would notnormally be used on questionnaires - countries would add appropriate category/ies based onservices available.Other broadband: Including high speed leased lines, optic fibre cable, some mobile phoneaccess (e.g. UMTS), powerline, satellite, fixed wireless, with an advertised download speed of>= 256 Kbps. This wording would not normally be used on questionnaires - countries would addappropriate category/ies based on services available.Do not know

a. The question is asked of all in-scope businesses which used the Internet during the referenceperiod.b. 12-month reference period.c. Possible country variations to the response categories are: remove categories where itemsare not feasible; add or split categories according to technologies available and country datarequirements.

Box B9: Proportion of businesses using the Internet by type of access


34

Proportion of businesses with alocal area network (LAN)This indicator is a measure of the proportion ofin-scope businesses with a local area network(LAN) as at the reference date.

Proportion of businesses with anextranetThis indicator is a measure of the proportion ofin-scope businesses with an extranet as at thereference date.

Proportion of businesses using theInternet by type of activityThis indicator refers to the type of activities in-scope businesses undertook using the Internetduring the 12-month reference period.

A local area network (LAN) refers to a net-work connecting computers within a localisedarea such as a single building, department orsite; it may be wireless. The proportion ofbusinesses with a LAN is calculated by di-viding the number of in-scope businesses witha LAN by the total number of in-scope busi-nesses. Sub-indicators may be constructedusing the classificatory variables, industry andsize.

Suggested model questiona:Did your business have a local area net-work (LAN) as at <reference date>b?

a. The question is asked of all in-scope busi-nesses which used computer/s during the ref-erence period.b. The reference date would usually be atthe end of the reference period, or shortlyafter.

Box B10: Proportion of businesseswith a local area network (LAN)

An extranet is a private, secure extension ofan intranet running on Internet protocol. It al-lows selected external users to access someparts of an organisation’s intranet. The pro-portion of businesses with an extranet iscalculated by dividing the number of in-scopebusinesses with an extranet by the total num-ber of in-scope businesses. Sub-indicatorsmay be constructed using the classificatoryvariables, industry and size.

Suggested model questiona:Did your business have an extranet as at<reference date>b?

a. The question is asked of all in-scope busi-nesses which used computer/s during the ref-erence period.b. The reference date would usually be atthe end of the reference period, or shortlyafter.

Box B11: Proportion of businesseswith an extranet

35

Core ICT Indicators

Internet activities are: use of the Internet for getting information (several response categories perthe model question below), for sending or receiving emails, for performing Internet banking oraccessing other financial services, for dealing with government organisations, for providingcustomer services and for delivering products online. Businesses can respond in respect of morethan one activity. For international comparability, output is most simply presented as the proportion ofin-scope businesses undertaking each activity, for instance, the proportion of businesses using theInternet for sending or receiving emails. An alternative presentation is the proportion of businessInternet users undertaking each activity. Sub-indicators may be constructed using the classificatoryvariables, industry and size.

Suggested model questiona:For which of the following activities did your business use the Internet during <period>b?c

For getting information:

• About goods or services

• From government organisations/public authorities (from Web sites or via email): Governmentorganisations/public authorities are preferably defined per the SNA93. They include governmentorganisations at local, regional and national level. See http://unstats.un.org/unsd/sna1993/glossform.asp?getitem=219.

• Other information searches or research activities

For sending or receiving emailsFor performing Internet banking or accessing other financial servicesFor dealing (interacting) with government organisations/public authorities: Includes down-loading/requesting forms, completing/lodging forms online, making online payments and purchasingfrom, or selling to, government organisations. Government organisations/public authorities are prefer-ably defined per the SNA93. They include government organisations at local, regional and nationallevel. See http://unstats.un.org/unsd/sna1993/glossform.asp?getitem=219.For providing customer services: Includes providing online or emailed product catalogues orprice lists, product specification or configuration online, after sales support and order tracking online.For delivering products online: Refers to products delivered over the Internet in digitised forme.g. reports, software, music, videos, computer games and online services, such as computer-relatedservices, information services, travel bookings or financial services.

a. This question is asked of all in-scope businesses which used the Internet during the referenceperiod.b. 12-month reference period.c. Possible country variations to response categories are to add or split categories according to coun-try data requirements. In particular, countries may wish to add an ‘Other’ category to this question.

Box B12: Proportion of businesses using the Internet by type of activity


36

Methodological notes

This section provides limited suggestions regard-ing methodology for collection of business ICTuse indicators. For most countries, there will beestablished procedures in place for conductingbusiness surveys which will apply also to collec-tion of ICT use data. For this reason, no recom-mendations are made on the following:

• use of a particular type of sample frame;

• sampling methodology;

• whether collections should be mandatory ornot;

• how to process collected information, includ-ing editing, imputation (for unit and item non-response) and weighting of data; and,

• survey vehicle (which could be a dedicatedor existing business survey).

Particular statisticalissuesassociated with business ICTuse measurementMeasuring Internet e-commerceAccording to the OECD definition, it is themethod by which the order is placed or received,not the payment nor channel of delivery, whichdetermines whether a transaction is an e-com-merce transaction. Accordingly, an “Internet (e-commerce) transaction is the sale or purchaseof goods or services, whether between busi-nesses, households, individuals, governments, andother public or private organisations, conductedover the Internet. The goods and services areordered over the Internet, but the payment andthe ultimate delivery of the good or service maybe conducted on or offline”.

The Internet e-commerce indicators, B7 and B8,could present some collection challenges, for in-stance, the need to have a larger sample size

where the incidence of Internet e-commerce islow.37 There are also definitional and dataavailability issues. The latter apply especially topurchasing over the Internet given that manybusinesses do not have a centralised purchasingfunction.

In addition, countries might wish to consider howthe collection of such data would affect surveymethods. For instance, if interviewers are used,they may require some technical training.

Choice of denominator for indicatorsMost indicators arising from ICT use surveysare presented as proportions data. They include:proportions of the whole population of businessesor of sub-populations, such as particular indus-tries or size classes. Additionally, countries mightpresent data as a proportion of businesses whichuse the Internet or have a Web site etc. This canbe very confusing to users so it is important tobe quite clear which denominator is used to cal-culate a particular indicator and to have a com-mon approach between countries.

It has been suggested in this paper that, for in-ternational comparability purposes, the simplestapproach is for countries to provide proportionswhich use the overall population as the de-nominator rather than the active population(being the number of businesses which use theInternet or have a Web site etc.). Note that, forpresentation purposes, it is possible to calculateindicators with the active population as thedenominator from other output which is proposedin this paper. For example, the proportion of busi-ness Internet users which have broadband ac-cess is equal to the proportion of businesses withbroadband access divided by the proportion ofbusinesses which use the Internet.

37

Core ICT Indicators

Survey vehicles and collectiontechniquesMost OECD countries conduct dedicated postalsurveys of ICT use by businesses. Countrieswhich do not have such a survey vehicle, forinstance many developing countries, could addquestions to an existing economy-wide surveyor to separate industry surveys. Note that datacan also be collected by means of personal in-terview (face-to-face or telephone) or othermethods such as drop-off/call-back (or postback). In the future, electronic data capture maybe viable for some respondents, though very fewOECD countries are yet using such methods.

Statistical unitThe following discussion refers to the unit aboutwhich data are collected. This may be differentfrom the unit which reports the data (a ‘report-ing unit’). The OECD and Eurostat both specifythe ‘enterprise’ as the statistical unit and this isthe unit used by most OECD countries. Choiceof unit is important as it influences the resultsobtained. As output from ICT use surveys ismainly proportions data, comparability betweencountries is more likely to be attained where theunit chosen is the same. As an example, if coun-try A uses the establishment as a unit and coun-try B uses the enterprise, then it is likely thatcountry B will report higher proportions, espe-cially of more sophisticated uses of ICT, such asbuying and selling over the Internet, or use of anintranet.

Unfortunately, there is no single definition of anenterprise that is used in all countries. The twomain definitions are those of the ISIC (Rev. 3.1)38

and the European Union.39 While they have com-mon characteristics that enterprises exercise acertain degree of autonomy in decision-makingand have full financial accounts, the EU conceptis narrower and it is suggested that this conceptbe used where possible.

Frequency and reference period/dateThere is perhaps a greater requirement than forhousehold surveys for the frequency of businesssurveys to be sensitive to the evolution of ICTand its use. OECD countries which conduct sucha survey do so on an annual basis. However, forsome countries, an annual collection will not befeasible, in which case it is important that thosecountries try to align their collection years as faras possible. As much of the information collectedis point-in-time data, it would be preferable tohave alignment of reference dates across coun-tries or regions.

Survey scope and coverageType of organisationThis is generally interpreted as those businessenterprises from the private and public sectors,which are operating in the country. General gov-ernment organisations are excluded. MostOECD countries exclude non-employers.

Industry (activity) scopeOECD collects business ICT use data frommember countries in respect of the following in-dustries: manufacturing (ISIC D), construction(ISIC F), wholesale trade (ISIC 51), retail trade(ISIC 52), hotels and restaurants (ISIC H), trans-port, storage and communications (ISIC I), fi-nance and insurance (ISIC J)40 and real estate,renting and business services (ISIC K). Eurostatreceives data from most countries for NACE41

sections D, F, G, H42, I, K and O43.

It is recommended that the minimum industryscope be: ISIC sections D, F, G, H, I and K.Countries are encouraged to extend this scopewhere there are both sufficient resources andsupporting national data requirements. In par-ticular, some developing countries might wish tocollect data from ISIC sections A, B and C andpart of O (covering agriculture, hunting, forestry,


38

fishing and mining, recreational, cultural andsporting activities, and other service activities).It is likely that a methodology similar to thoseused for household surveys would be appropri-ate for many of the units in these industries.

Size scopeMost OECD countries specify that in-scope busi-nesses are employers and they define size scopein terms of number of employees. Eurostat speci-fies a size cut-off of 10 or more employees. Forcomparability, OECD does likewise even thoughthere is a range of cut-offs used among OECD(including European) countries, with at least twomember countries including enterprises with asingle employee.

This paper proposes a minimum size scope of10 or more employees. It is recognised that thereare important policy issues pertaining to busi-nesses which are smaller than this. Additionally,in most developing countries, very small busi-nesses make up such a large part of the overallbusiness community that their inclusion is highlydesirable. Developing countries are thereforeencouraged to extend the scope to include smallerbusinesses where resources permit.44

Classificatory variablesThe metadata associated with the core questionsneed to include classificatory variables with rel-evant categories. Data for these variables willusually be collected as part of a survey or maybe available from the population frame (for ex-ample, a business register). For business surveys,the classification variables will generally be thesame as, or a subset of, the stratification vari-ables (though the detail may be different). A mini-mal set of classification variables and categorieswhich are consistent with the scope recommen-dations above, is suggested below. This set isconsistent with advice offered by OECD andEurostat for collection of business use of ICT

data. As with the household survey, countriesmay decide to use extra classificatory variablesand/or additional categories. In particular, devel-oping countries with a significant rural/urban di-vide may wish to add a geographic classifica-tion.

Industry (activity)A broad industry output classification consistentwith the suggested industry scope is proposedas follows: manufacturing (ISIC D), construc-tion (ISIC F), wholesale and retail trade (includ-ing repair of motor vehicles, motorcycles andpersonal and household goods) (ISIC G), hotelsand restaurants (ISIC H), transport, storage andcommunications (ISIC I), and real estate, rent-ing and business services (ISIC K).

Size (number of employees)The size classification proposed is based on theminimum size scope and is: 10-49 employees;50-249 employees and 250 or more employ-ees.

39

Core ICT Indicators

Core Indicators on the ICT Sector andTrade in ICT Goods

For manufacturing industries, the products of acandidate industry:

• Must be intended to fulfil the function of in-formation processing and communication in-cluding transmission and display.

• Must use electronic processing to detect,measure and/or record physical phenomenaor control a physical process.

For services industries, the products of a candi-date industry:

• Must be intended to enable the function ofinformation processing and communication byelectronic means.

The definition was initially based on ISIC Rev. 3and updated in 2002 to reflect changes in ISICRev. 3.1. The 2002 definition of the ICT sectoris shown in the box above.

It should be noted that the OECD’s ICT sectordefinition is likely to change somewhat with themajor revision of ISIC in 2007 (to ISIC Rev. 4).

Countries might encounter the following prob-lems in applying the OECD ICT sector defini-tion and collecting data:

• For countries which do not use ISIC Rev. 3.1

This section suggests classifications and defini-tions relevant to the basic core indicators, ICT1to ICT4. The suggestions are based on work ofthe OECD’s Working Party on Indicators for theInformation Society (WPIIS).

ICT sector core indicators

• ICT1. Proportion of total business sectorworkforce involved in the ICT sector

• ICT2. Value added in the ICT sector (as apercentage of total business sector valueadded).

ICT1 and ICT2 would generally come from sur-veys which collect income and expenses datafor national accounts. Metadata characteristicsof these indicators include the definition of theICT sector and the indicator terms. These arediscussed below.

Defining the ICT sectorIt is recommended that countries use the defini-tion of the ICT sector agreed by WPIIS in 1998(and revised in 2002). It consists of manufactur-ing and services industries whose products cap-ture, transmit or display data and informa-tion electronically.


40

(or NACE Rev. 1.1) to classify economicunits, there may be some concordance issueswhich need to be addressed.

• For confidentiality reasons, some countriesmay be unable to report data for telecommu-nications services. Aggregation into total ICTservices as proposed below will often solvethis problem.

For output purposes, the OECD sometimes pre-sents only two categories: ICT manufacturingand ICT services. Individual countries may alsodo this where confidentiality or industry concor-dance is a problem.

Defining variables used in coreindicators ICT1 and ICT2The indicator ICT1 refers to the total workforceinvolved in the ICT sector as a proportion of thetotal business workforce. ICT2 refers to valueadded in the ICT sector as a proportion of total

business value added.

Both are usually calculated using data for theICT and total business sector which are com-patible with national accounts tables. Where ICTsector industries are not present in a country’snational accounts by activity tables, estimates aremade based on business survey results (oftenprovided specifically for the ICT sector byNSOs). OECD is currently reviewing the defi-nition of the total business sector and has morerecently been using ISIC 10-74 (excluding 70)45

as an activity-based business sector definition(rather than a business sector defined on an in-stitutional basis).

The underlying measures used in these indica-tors are workforce and value added.

ICT workforce (or ICT employment) consistsof those persons employed in businesses which

Manufacturing:3000 Office, accounting and computing machinery3130 Insulated wire and cable3210 Electronic valves and tubes and other electronic components3220 Television and radio transmitters and apparatus for line telephony and line teleg-

raphy3230 Television and radio receivers, sound or video recording or reproducing appara-

tus, and associated goods3312 Instruments and appliances for measuring, checking, testing, navigating and other

purposes except industrial process control equipment3313 Industrial process control equipment

Services:5151 Wholesale of computers, computer peripheral equipment and software5152 Wholesale of electronic and telecommunications parts and equipment6420 Telecommunications7123 Renting of office machinery and equipment (including computers)72 Computer and related activities

OECD ICT sector definition, 2002 (ISIC Rev 3.1)

41

Core ICT Indicators

are classified to the ICT sector. Total businessworkforce represents all persons engaged indomestic production in the business sector. In anational accounts framework, employment canbe measured in terms of headcounts, jobs, full-time equivalents or hours worked. Currently, to-tal headcounts or jobs are used for most coun-tries. Further efforts are required to determinebest practices for global ICT sector indicatorsbased on labour inputs.

Value added for a particular industry representsits contribution to national GDP. It is sometimesreferred to as GDP by industry and is not directlymeasured. In general, it is calculated as the dif-ference between production (gross output) andintermediate inputs (the energy, materials and ser-

vices required to produce final output).

The valuation can be made at factor costs, atbasic prices or at producers’ prices. The rela-tionship between these three valuations is shownin the figure below.

ICT goods trade coreindicators

• ICT3. ICT goods imports as a percentage oftotal imports

• ICT4. ICT goods exports as a percentage oftotal exports.

ICT3 and ICT4 would normally come from trade

Value added at factor costs + other taxes, less subsidies, on production(2)= Value added at basic prices + taxes less subsidies, on products(3) (not including imports and VAT)= Value added at producer’s prices + taxes, less subsidies, on imports + Trade and transport costs + Non-deductible VAT= Value added at market prices(4)

Valuation of Value Added(1)

nant). However, many OECD Member coun-tries have now implemented SNA93 (or the EUequivalent, ESA95) which recommends the useof basic prices and producer prices (as well aspurchaser prices for input-output tables).(2). These consist mostly of current taxes (andsubsidies) on the labour or capital employed, suchas payroll taxes or current taxes on vehicles andbuildings.(3). These consist of taxes (and subsidies) pay-able per unit of some good or service produced,such as turnover taxes and excise duties.(4). Market prices are those which purchaserspay for the goods and services they acquire oruse, excluding deductible VAT. The term is usu-ally used in the context of aggregates such asGDP, whereas purchaser prices refer to the indi-vidual transactions.

(1). This table draws on concepts outlined in boththe 1968 and 1993 versions of the System of Na-tional Accounts (SNA68 and SNA93). Until thelate 1990s, most countries adhered to recommen-dations in SNA68 (where the notions of factorcosts, producer prices and market prices were


42

statistics. Metadata characteristics of these in-dicators include the definition of ICT goods andsources and concepts concerning internationaltrade statistics more generally. These are dis-cussed below.

Defining and classifying ICT goodsIn late 2003, the OECD released a definition andclassification of ICT goods, based on work doneby WPIIS. The OECD ICT goods classificationuses the 2002 version of the World CustomsOrganization’s Harmonized System (HS) but hasa concordance to the 1996 version. The full listof six-digit HS categories included in the classi-fication can be found on the OECD Web site asan annex to the paper ICT goods classifica-tion.46 Countries using the HS 1996 versionshould note that there are differences betweenthe 1996 and 2002 versions in 7 HS codes. Thesedifferences are detailed in the above-mentionedpaper.

ICT goods are defined by the OECD as “…..in-tended to fulfil the function of information pro-cessing and communication by electronic means,including transmission and display, or use elec-tronic processing to detect, measure and/orrecord physical phenomena, or to control a physi-cal process”.

It should be noted that this definition results in abroad interpretation of an ICT good and there-fore an extensive classification. For output pur-poses, it is suggested that detailed categories beaggregated into the five broad categories rec-ommended in the OECD paper, that is, telecom-munications equipment, computer and relatedequipment, electronic components, audio andvideo equipment and other ICT goods.

It should also be noted that software products(including packaged software) are not includedin this classification. The OECD will include all

software in its forthcoming classification of ICTservices.

Trade statistics sources and conceptsInternational trade statistics are available for mostcountries from the United Nation’s Comtradedatabase. The Comtrade Web site47 includes ex-tensive metadata by country, including data avail-ability, classifications for which data are available(different versions of HS or SITC) and conceptsused (such as valuation, currency and trade sys-tem). All data are available in current US$. Notethat re-exports48 and re-imports49 are not nettedout from exports and imports respectively. Thismay be problematic for some countries.

The Comtrade Web site also has a glossarywhich explains terms and concepts.

43

Core ICT Indicators

EurostatEurostat’s Methodological Manual is an opera-tional manual containing recommended guidelinesfor the Eurostat model surveys on ICT use byenterprises and households/individuals. It was firstpublished in late 2005 to assist countries conductthe 2006 surveys and will be updated each year.

The following Web site provides methodologicalinformation (including questionnaires) for theEurostat surveys http://europa.eu.int/estatref/info/sdds/en/infosoc/infosoc_base.htm.

ITUDefinitions of key indicators, including ICT in-frastructure and access core indicators, may befound in Key indicators of the telecommuni-cation/ICT sector available from: http://www.itu.int/ITU-D/ict/material/Top50_e-WTIM-2005-8June.doc.

OECDGuide to Measuring the Information Society.The Guide brings together OECD and membercountry work on information society statistics.Of particular relevance for methodologies for thehousehold/individual and business core indicators,the Guide explains OECD work on ICT statisti-cal standards. It also provides:

• metadata in respect of the collection prac-tices of OECD countries (in a Web-basedannex);

• an annex on non-member country work (in-cluding the relevant activities of non-OECDcountries, regional networks and internationalorganisations); and,

• an annex on the application of OECD rec-ommendations to developing countries (thelatter contributed by the UNESCO Institutefor Statistics).

The Guide can be found on the OECD Web siteat: http://www.oecd.org/sti/measuring-infoeconomy. The ICT statistics metadata an-nex is also available on the OECD Web site at:http://www.oecd.org/sti/ictmetadata.

Partnership on MeasuringICT for DevelopmentUNCTAD is the host for material on Partner-ship activities. The home page is http://measur-ing-ict.unctad.org/QuickPlace/measuring-ict/M a i n . n s f / h _ T o c /281E7067B40AD764C1256EE80048DACC/?OpenDocument.

Partnership project document, http://measur-ing-ict.unctad.org/QuickPlace/measuring-ict/

Reference Material


44

M a i n . n s f / h _ I n d e x /FBFD3BDF8A2AC11EC1256EF5005C2CC5/?OpenDocument.

Measuring ICT: The Global Status of ICT In-dicators (2005).

UNCTADThe Web site on measuring ICT http://measur-ing-ict.unctad.org includes papers and presenta-tions from WSIS thematic meetings on ICT mea-surement and other useful information, for ex-ample, regional initiatives.

UNECLACTowards an Information Society measurementinstrument for Latin America and the Carib-bean: getting started with census, householdand business surveys, http://www.cepal.org/cgi-bin/getProd.asp?xml=/publicaciones/xml/5/2 1 5 9 5 / P 2 1 5 9 5 . x m l & x s l = / d d p e / t p l /p9f.xsl&base=/socinfo/tpl/top-bottom.xsl.

UNESCO Institute for StatisticsMeasuring and Monitoring the Informationand Knowledge Societies: a Statistical Chal-lenge.50

This report was written as one of UNESCO’sinputs to the first phase of the World Summit onthe Information Society (WSIS) in 2003. Its fo-cus is on measurement issues with regard to in-formation and communication technologies(ICTs) and gives an overview of the data thatexist as well as where the gaps lie. Issues relat-ing to data quality are also addressed.

UNESCO Observatory on theInformation Societyhttp://www.unesco.org/webworld/observatory/

As a part of UNESCO’s strategy to build knowl-edge societies and in compliance with

UNESCO’s function as a ‘clearing house’, theObservatory aims at monitoring the developmentof knowledge societies by:

• collecting pertinent information on the evolu-tion of ethical, legal, socio-cultural and policyissues related to the information society, withparticular attention to education, science, cul-ture and communication;

• timely presentation of new trends and con-texts; and

• sharing information about the challenges ofthe information society and advances in ICT.

Monitoring the progress towards an equitable in-formation/knowledge society very much dependsupon the availability of reliable data and key indi-cators on all aspects of ICT, including access,usage and impact. For that reason, the UNESCOObservatory dedicates one of its categories toMeasuring the Information Society.

h t t p : / / w w w. u n e s c o . o r g / w e b w o r l d /p o r t a l _ o b s e r v a t o r y / p a g e s /Measur ing_ the_Informat ion_Soc ie ty /index.shtml.

UN Statistics DivisionComtrade database http://unstats.un.org/unsd/comtrade/default.aspx.

45

Core ICT Indicators

1. Information and communication technol-ogy.2. The Partnership on Measuring ICTfor Development was launched in June 2004,and currently includes the following members:Eurostat, the International TelecommunicationUnion (ITU), the Organisation for Economic Co-operation and Development (OECD), the UnitedNations Conference on Trade and DevelopmentUNCTAD, four United Nations Regional Com-missions (the UN Economic Commission forAfrica (UNECA), the UN Regional Commis-sion for Latin America and the Caribbean(UNECLAC), the UN Economic and SocialCommission for Asia and the Pacific(UNESCAP) and the UN Economic and SocialCommission for Western Asia (UNESCWA)),the United Nations Educational, Scientific andCultural Organization (UNESCO) Institute forStatistics (UIS), the UN ICT Task Force andthe World Bank. For further information on theobjectives and activities of the Partnership, seehttp://measuring-ict.unctad.org.3. For statistical reasons, there are severalminor differences between the wording of theresponse categories suggested in the core list ofindicators presented here and the list agreed uponat the February 2005 WSIS thematic meeting.4. In the list adopted by the February meet-

ing, this response category was entitled FreePublic Internet Access Centre. It is proposedthat it be changed to Community Internet ac-cess facility to reflect the fact that communityaccess may be charged, albeit often at subsidisedrates.5. In the list adopted by the February meet-ing, this response category was entitled ChargedPublic Internet Access Centre. It is proposedthat it be changed to Commercial Internet ac-cess facility to reflect the fact that commercialaccess is not necessarily charged.6. In the list adopted by the February meet-ing, this response category was entitled Internetbanking or other financial services. It is pro-posed that it be changed to Internet bankingfollowing changes made to the OECD andEurostat model questionnaires.7. Since electricity is not specifically anICT commodity, but an important prerequisite forusing many ICTs, it is not included in the corelist, but included as a reference indicator.8. This group of indicators was referred toas ‘Core indicators on access and use of ICTsby businesses’ in the list adopted by the Febru-ary meeting. The title has been changed to bet-ter reflect the nature of indicators (which dealwith usage rather than access). The term ac-cess, in this context, is used for household sur-

Footnotes


46

veys and refers to the availability of a technol-ogy to household members rather than its use.9. These indicators were discussed duringthe World Telecommunication Indicators Febru-ary 2005 meeting, available from: http://www.itu.int/ITU-D/ict/wict05/index.html andcontained in the Key indicators of the telecom-munication/ICT sector, available from: http://www.itu.int/ITU-D/ict/material/Top50_e-WTIM-2005-8June.doc.10. There is no shortage of references touniversal service/access being the main goal oftelecommunication policy. For further informa-tion see ITU (1998), World TelecommunicationDevelopment Report: Universal Access, avail-able from: http://www.itu.int/ITU-D/ict/publica-tions/wtdr_98/index.html; and ITU (2003),Trends in Telecommunication Reform: Promot-ing Universal Access to ICTs – Practical Toolsfor Regulators, available from: http://www.itu.int/publications/docs/trends2003.html.11. Variations on basic and primary ISDNexist in some countries, sometimes referred toas fractional ISDN. For example in Denmark avariant known as Flex-ISDN provides 12 chan-nels per line.12. Indicators A1 and A2 comprise theMillenium Development Goals (MDG) indicator# 47.13. According to some researchers, the PCreplacement rate in the US is as high as 70 percent per year. On the other hand “In more de-veloping regions, PC replacement rates are muchlower.” CyberAtlas (2003), “PC Market headedfor geographic shift”, available from: http://cyberatlas.internet.com/big_picture/hardware/article/0,,5921_988841,00.html.14. Prince and Cooke, December 1998,Mercado Informático, available from: http://www.spkrsbr.com/biblioteca/htm/resultados.htm.15. Indicators A3 and A4 comprise the MDGindicator # 48. Together, they may be used toindicate the level of ICT penetration and utilisation

in a country, and hence its readiness to evolvetowards an Information Society.16. h t t p : / / w w w. i s o c . o rg / i n e t 2 0 0 0 /cdproceedings/8e/8e_1.htm#s6.17. h t t p : / / w w w. i s o c . o rg / i n e t 2 0 0 0 /cdproceedings/8e/8e_1.htm#s6.18. For more on broadband developments,see ITU (2003), Birth of Broadband, availablefrom: www.itu.int/birthofbroadband.19. European Union, eEurope 2005:Benchmarking Indicators, available from: http://europa.eu.int/comm/lisbon_strategy/pdf/655_EN.pdf.20. The broadcast industry uses othermetrics such as ‘universe estimates’ (e.g. po-tential television audience). See “FAQ – AboutRatings” at the Nielsen Media Research website:http://www.nielsenmedia.com.21. In the past, UNESCO had published thenumber of radio and television sets in differentcountries but stopped with its 1999 StatisticalYearbook. UNESCO has conducted a consul-tation exercise which may lead to the revival ofthis data series in 2006.22. This lack of data may be a problem inthe future, as countries shift towards digital ra-dio and television broadcasting. Important policy-decisions on when to turn off analogue broad-cast channels may be delayed due to lack of re-liable data on homes with radios and televisions.23. For historical and practical reasons, somedefinitions of ICTs such as a computer and theInternet differ from those used by the ITU.24. It is not expected that the structure, ques-tion wording or definitions which comprise themodel questions would necessarily be used un-changed (or literally translated) in national sur-veys. However, it is desirable for comparabilitypurposes that their meanings are preserved andthat the logic is preserved to the extent that thespecified populations of households or individu-als are asked each question.25. Note that a question for the reference

47

Core ICT Indicators

indicator on electricity, HHR1 (proportion ofhouseholds with electricity) is not included.26. For household scope, see discussion onscope under ‘Methodological notes’.27. For individual scope, see discussion onscope under ‘Methodological notes’.28. The OECD model questionnaire uses a12-month recall period for all questions. TheEurostat model asks some questions in respectof both 12 months and three months (for in-stance, individual use of a computer and theInternet, whether the individual has purchasedproducts over the Internet) but asks others (e.g.location of use and activities undertaken) in re-spect of the last three months.29. Reasons for the lower age limit includelegal restrictions in some countries on interview-ing children. Regarding the upper age limit, somecountries ask these questions as supplements tolabour force surveys and therefore interviewpeople most likely to be in the labour force.30. This is consistent with recommendationsmade by OECD and Eurostat for their modelsurveys.31. Note that this differs from the Eurostatand OECD levels which combine ISCED 0,1 and2. For more information on ISCED (1997), seeh t t p : / / w w w . u i s . u n e s c o . o r g /ev.php?ID=3813_201&ID2=DO_TOPIC.32. Self employed includes: employers; ownaccount workers; contributing family workers;and members of producers’ cooperatives.33. For more information on the ILO defini-tions, see http://www.ilo.org/public/english/bu-reau/stat/class/icse.htm.34. For more information on ISCO, see http://www.ilo.org/public/english/bureau/stat/class/isco.htm.35. It is not expected that the structure, ques-tion wording or definitions which comprise themodel questions would necessarily be used un-changed (or literally translated) in national sur-veys. However, it is desirable for comparability

purposes that their meanings are preserved andthe logic is preserved to the extent that the speci-fied populations of businesses are asked eachquestion.36. For business scope, see discussion onscope under ‘Methodological notes’.37. This is still the case in most OECD coun-tries.38. ISIC is the International Standard Indus-trial Classification of all Economic Activities. Fordetails of ISIC Rev 3.1, see http://unstats.un.org/unsd/cr/registry/regcst.asp?Cl=17. According toISIC, an enterprise has “autonomy in respect offinancial and investment decision-making, as wellas authority and responsibility for allocating re-sources for the production of goods and services.It may be engaged in one or many productiveactivities. The enterprise is the level at whichfinancial and balance sheet accounts are main-tained and from which international transactions,and international investment position (when ap-plicable) and the consolidated financial positioncan be derived.”39. Defined by the European Commissionas: “……the smallest combination of legal unitsthat is an organisational unit producing goods orservices, which benefits from a certain degreeof autonomy in decision-making, especially forthe allocation of its current resources. An enter-prise carries out one or more activities at one ormore locations. An enterprise may be a sole le-gal unit.”40. The revised (2005) OECD model sur-vey of business ICT use suggests that ISIC sec-tion J be ‘non-core’ for OECD member coun-tries. Eurostat developed a specific module ofthe enterprise survey for a pilot study of the sec-tor in 2004. In 2005, the Eurostat model ques-tionnaire was revised but limited to general ICTvariables. For 2006, the Eurostat model ques-tionnaire was improved and includes variableson e-commerce.41. NACE is the Statistical Classification of


48

Economic Activities in the European Commu-nity, Rev. 1.1 (2002).42. In respect of section H, only about halfthe countries which do the Eurostat survey col-lect data for NACE 55.3 to 55.5 (restaurants,bars etc.).43. Not all countries which do the Eurostatsurvey collect data for all classes of Section O(Other community, social and personal serviceactivities). For collection purposes, divisions 92and 93 are most relevant. The revised (2005)OECD model survey of business ICT use sug-gests that ISIC Division 92 (recreational, cul-tural and sporting activities) be ‘non-core’ forOECD member countries.44. Countries should note that the broaderthe scope, the larger the sample size generallyrequired to obtain adequate aggregate estimates.Extending the scope to businesses with fewerthan 10 employees might increase the samplesize by a factor of two or more.45. The business sector therefore excludes:agriculture, hunting, forestry and fishing (mainlybecause of problems measuring employment insome countries); real estate activities (becausea significant proportion of its value added con-sists of imputed rent of owner-occupied dwell-ings); and, community, social and personal ser-vices (which consists mainly of non-market ac-tivities such as public administration, educationand health services).46. The full URL is http://www.oecd.org/dataoecd/5/61/22343094.pdf.47. The full URL is http://unstats.un.org/unsd/comtrade/default.aspx.48. Re-exports are exports of foreigngoods in the same state as previously imported.49. Re-imports are goods imported in thesame state as previously exported.50. The full URL is http://w w w . u i s . u n e s c o . o r g /ev.php?ID=5504_201&ID2=DO_TOPIC.

49

Core ICT Indicators

Core ICT Indicators

Documents