Assessing the Yield of It Projects in Developing Nations: Aggregated Models Are Not Sufficient

Assessing the Yield of IT Projects in Developing Nations: Aggregated Models Are Not Sufficient

This paper offers an example of a methodology and a process that can enable better policy choices in IT diffusion in developing nations. It updates and refocuses an earlier report by Ruth [19]. There are literally dozens of conceptual frameworks aimed at describing the diffusion of IT in developing nations but no unified approach to determining the micro-level data that underpins the aggregate data. Internet hosts per thousand population, for example, is an interesting summary number in a developing country but gives no clear vision of the individual and group issues--access limits, training approaches, gender roles, etc.-- that cause each host site's inclusion or exclusion. A comprehensive model of the diffusion process for Internet must be able to describe specific approaches to measuring at the micro level--not just aggregations of statistics at the regional or national level. We offer a small step in that direction by demonstrating that these micro level statistics can be obtained with precision, enabling inferences about individual and group performance. The example we use is an Internet course for professors in Romania's capital, Bucharest. Training of this type is an integral step in Internet diffusion in any developing nation and the variables that emerge can be used in many other micro-level studies of other Internet-related issues. The importance of training is reiterated in a study conducted by the World Bank in which connectivity and education are identified as critical factors in ensuring the reduction of the "digital divide". In this research case significant effects are demonstrated in both descriptive statistics and ANOVA results based on gender, academic specialty, computer skills, age, and research productivity. The approach in Romania is robust and replicable and we feel that by using this realistic individual data as an exemplar for other micro level studies, it can be possible to integrate more realism into the rich array of macro-level models that already exist.

The number of aggregate or macro level statistics proposed for measuring of Internet diffusion in developing nations is steadily increasing. A group of experts developed over seventy criteria for studying the effects of Internet in developing countries. Examples of these criteria are:

a. Number (percentage) of chambers of commerce with Internet access

b. Rate of change in the value of a firm's exports since acquiring Internet access

c. Number of ministries/departments with e-mail reply addresses on the Web

d. Percentage of total public information made available through the Internet

e. Number of Web server hits or requests fulfilled per month from domestic versus regional versus international sources

For those who follow this literature the number of macro level frameworks for Internet diffusion in developing nations can be daunting--numbering in the dozens. Each of these approaches is an attempt to establish a series of criteria or predictor variables that can assist in evaluation of the process by which Internet technology deployment takes place. Incidentally, one finding that none of these studies contests is that Internet connectivity in developing nations is tragically low. Internet Hosts in millions in the top five nations, the USA, Japan, Canada, UK, and Germany are 65.9, 4.3, 3.9, 3.2, and 2.7 respectively. The comparable numbers for Africa, China, and India, where nearly half the world population is concentrated, are 0.24, 0.11, and 0.05 respectively. The twenty developed nations of the world use about 95 percent of the Internet capacity.

Typical of the approaches for integrating macro indicators for Internet diffusion in developing nations is a group of periodic reports, many appearing in the Communications of the Association for Computing Machinery, prepared by researchers who make field visits and interview Internet Service Providers (ISP), government officials, Postal, Telephone & Telegraph (PTT) managers, and other Internet policy planners. Typically, they examine either a group of countries like the United Arab Emirates, South America, or individual countries like Bangladesh, Nepal, Israel, Turkey, Cuba or a region like Greater South China. Their reports emphasize descriptive statistics based on Internet hosts, usage ratios among population segments and other summary data.

Another generalized approach to characterizing the dynamics of the Internet diffusion process was developed by John Daly. An essential characteristic of this framework is the interaction among Internet penetration, utilization and results (impacts) for workgroups, organizations and markets. Internet penetration is dependent on such factors as availability of bandwidth, quality and quantity of content transmitted, infrastructure (electrical and telephony), and the policy environment (prices, taxes, etc.). Utilization is about variables that affect the ability to take advantage of Internet: training, quality of use, hours of use, and the like. Impacts are measured as a cluster of results that can be achieved, like improved education, changes in Gross Domestic Product in Internet services, and individual benefits in health, government, commerce, industry, etc. Again, the inputs and outputs in this framework are highly aggregated data, summaries of summaries in many cases.

Perhaps the most detailed and tested aggregate approach to examining Internet diffusion in developing nations is the GDI (Global Diffusion of the Internet) studies. GDI methodology analyzes six characteristics of Internet connectivity at the country level: Pervasiveness, Geographical Dispersion, Sectoral Absorption, Connectivity Infrastructure, Organizational Infrastructure, and Sophistication of Use. Through a standardized approach to these six aggregate indicators, GDI practitioners are able to develop a summary for each country.

The focus of this exploratory study is a rigorous three-day Internet training course offered from 1996-2000 to researchers affiliated with the Romanian Academy of Sciences in Bucharest. The Andrew W. Mellon Foundation, the Soros Foundations and Sun Microsystems, Incorporated, funded the training. The aim of this program was to establish a long term Internet training capacity, staffed by local experts and supporting a major research effort in the country. The study population is a group of professors and lecturers affiliated with the Romanian Academy of Science, a large, multifaceted alliance of academics representing all fields of study, from drama to engineering, and centered in Bucharest, Romania's capital,