Prior to becoming an independent consultant, advising government and non-governmental agencies and corporations on the use of technology to support developing countries, I was a professor and research scientist for thirty years.
I did a considerable amount of research on the impact of technology on teaching and learning. I also developed educational software for the Macintosh. Consequently, I can attest that empirical data are the sin qua non of both scientific research and engineering design.
Scientists posit their theories as testable hypotheses and conduct experiments to validate their propositions. Engineers design artifacts to achieve goals or solve problems. They test out these designs on a small scale and refine them before implementing on a large scale. Collecting test data is a standard practices in both fields.
But apparently Professor Negroponte doesn't follow these standard engineering or scientific practices, at least when it comes to OLPC. Without the benefit of a single study in support of their efficacy, Professor Negroponte feels that developing nations should spend hundreds of millions of dollars to purchase millions of XO computers to hand out to its students. At a meeting of the Inter-American Development Bank, he claimed that developing nations:
". . . need to do things which isn't futzing around and moving deck chairs. And they can spend the next five years planning. But that's not what they should do. They have to take action. They have to take big action. To do a pilot project is ridiculous!"Now, the suggestion that drew Professor Negroponte's ire at the IADB meeting was a rather modest one. It was made by Dr. Andrew Zucker, a former colleague of mine at SRI's Center for Technology in Learning.
Having done extensive research in on 1-to-1 computing in U.S. schools, he suggested that educators in developing countries would do well to start with a pilot program and test the use of laptops in schools against defined metrics before rolling them out to a larger group. Professor Negroponte disagreed with him. He was even more forceful in his distain of pilot tests at TED 2006 when he claimed about the OLPC:
". . . this is not something you have to test, the days of pilot projects are over. When people say 'Well, we'd like to do three or four thousand in our country to see how it works.' Screw you, go to the back of the line and someone else will do it. Then when you figure out that this works then you can join as well."I am glad to hear that some countries are thinking of conducting pilot tests, despite Negroponte's reproach. But these have to be something more than a press event when XO machines are given to groups of children as the cameras roll. And they have to be something more than just testing the robustness of the hardware and operating system.
The kind of test that Dr. Zucker and I have in mind would involve the introduction of the XO into a small sample of schools and classrooms in a country. Children and teachers would actually use them for at least several months, if not an entire school year. OLPC researchers would observe the teachers and students periodically and collect data on the extent to which the computers are used, the ways they are used, the artifacts that students create with them, and the problems they encounter as they do so.
The observed data would be compared to a pre-specified set of desired behaviors related to student interaction with the machine and with each other and to a set of desired learning outcomes - the kinds of things the OLPC has in mind when they say the XO is an education project and the kinds of results that would make governments feel that their investment paid off.
Given the "openness" of the project, these findings should be publicly shared with the larger development and educational community. If the behaviors and outcomes are on target, it's ready for full implementation. But if they fall short of the desired set or if the students and teachers encounter significant problems, then the hardware, software, or enabling conditions would need to be re-engineered and retested.
This cycle of testing and retesting increases the likelihood of success during a full-scale implementation. Final testing by independent researchers would assure that the results were not unintentionally skewed by the OLPC researchers.
To do otherwise is both contrary to standard practice and irresponsible. Countries that adopt OLPC without pilot testing are in effect conducting a nation-wide experiment. It is a roll of the dice. If the OLPC predictions are correct, the nation and its children win.
Of course, had they conducted a pilot test first, they would have also won. Indeed the probability that they would win increases dramatically; it just would have taken a little longer and cost a little more than without the test.
On the other hand, if the grand national experiment fails, it is developing countries and their children that are least able to manage the consequences of this failure or recover from the expended costs.
This makes Professor Negroponte's dictums not only irresponsible but unethical.