Federal Focus, Inc.

Children's Issue

return to speaker list




Dr. Henry Kelly is the Assistant Director for Technology, The White House Office of Science and Technology Policy, Executive Office of the President.


The thing that is apparent to everyone, and certainly to the administration, is that the technology that is driving most of the growth in the economy is driven, to an extremely heavy degree, by information technology (IT). Technical developments have been responsible for almost half of the economic growth in the last 20 years. IT alone apparently has been responsible for about 30% of the growth in GDP in the last two or 3 years. There is absolutely no question that the momentum in the economy is dependent on innovation, particularly information innovation. However, one of the things that is also clear is that the information jobs are not necessarily jobs of people who think that they are in the information industry, because you find that it is hard to get a job doing anything if you are without a very sound grasp of not only the IT itself, but the fundamental logic in terms of the mathematics and the science that drives it. Have you ever been into a car repair shop lately? You do not find guys who do nothing but replace tires; you find guys who are plugging the local computer into the computer in your car. This car repair shop probably has that computer networked back to Detroit or to Tokyo, or wherever. This guy who is repairing cars is a part of the information economy in a very deep way. An inability to understand that equipment is an inability to do his job. Almost every growth occupation finds itself deeply embedded in technical transformation. Consider the health care industry. We do not find people sitting there counting blood platelets through a microscope; we find people who are each year being forced to learn a new set of medical diagnostic devices.

There is plainly no room for people who do not have an ability to master the basics of mathematical and communication skills. To a growing extent, that also means an ability to feel comfortable with information appliances of all kinds. So, given that there is absolutely no question that we have to find ways of delivering education to people who are entering this workforce, we are also forced to recognize that we will have to improve the training and the skill level of the people who are already in the workforce. If there is anything that is clear about these transformations, it is that we are at the very beginning of this process and not at the end of it.

One of the transformations is the transformation of the education process itself. I would like to spend a little time thinking about that. Plainly, the education enterprise is the largest information business on the planet. It is 8 - 10% of the GNP, $700 billion per year. It is not used to thinking in terms of productivity, in terms of capital investment. It does not know how to buy things with that in mind. It does not know how to obtain economies of scale in purchases. Plainly, it will have to change the way it thinks about its relationship to this new class of technologies. It will also end up having to redefine the skills of the workforce who are actually in the process of delivering instruction. We continually see in areas such as business training an expectation of very rapid growth in the use of IT equipment. Here you see a forecast by the largest training organization that suggests that something like 20% of the hours that most workers spend learning things next year, will be spent learning in some instructional environment intermediated by information appliances of some kind.

Now, is this a good thing or a bad thing? That is something for us to answer. There is no question that in principle, it is a good thing. The optimism about the use of the technology has to grow out of several different dimensions. One is that in principle, one can get back to teaching people things in a way that looks very different than the last 300 years of instruction, which is instruction like I am giving you here, where you are in a chair bolted to the ground, and I am pontificating. No one learned like that in traditional environments. We engaged in apprenticeship experiences. We tried to imitate adults; we played games in which we tried to do things that adults were doing. If you fell on your face or you were trying hit the rhinoceros and kept missing, then you knew that you needed some help. The apprenticeship environment is a tool that we have forgotten how to use, but it is one which we will have at least the opportunity to figure out how to use, which can be useful not only in motivating things, but also can help us cross this terrible barrier of abstraction between the work environment and the adult world in which the skills are relevant and the very abstract representations of ideas that we are forced to grapple with in formal instructional environment. We are also able, in principle, to test a whole range of skills that we are not able to test using conventional paper-and-pencil tests.

My favorite example is a flight simulator, in which if you crash the airplane in the simulation, you do not need to have your test graded. You can probably figure out that you have done something wrong, and you probably have a very powerful incentive to figure out why. This test is useful to you, it is useful to your employer, and it is useful to your instructor. This is the kind of test you would like to be able to see. It is not simply a hazing process, it has the power to tell you why it is that what you do not know is a problem. You can also make sure that what you are learning is current, and you can also make sure that it is available where and when you want it, that it is available at home without having to lug home mountains of text books. Of course, if you are an adult, this is particularly important.

Now, the point I made earlier about apprenticeship experiences is borne out rather heavily by some statistics that were gathered over the last four or five years . . . The point of the recent research is to compare the outcome of students who were in 30-person classes listening to instruction with people who were taught identical material out of the same text, but with a one-on-one instructor. What you see here is dramatic. The difference between the classroom-based environment and the individual tutor environment is huge - two standard deviations. Equally important is that the width of the distribution is hugely shrunk by a factor of roughly three. Now, this means that not only that everybody did well - we were not leaving a long tail on this distribution of people who did not get it. The reason this is happening, to me, is a productive area of inquiry. One of the things they did do was to take videotapes of people going through these processes. They discovered that in the individual tutor example, the kids asked 36 questions per hour, on the average. In a classroom with 30 people in it, there was an average of 3 questions asked during the entire hour. So, basically you have 36 unanswered questions per student, and most teachers know that you have to move forward in a lesson when most of the kids understand most of the material. Well, that is clearly why we have this big tail in the distribution. There is something the student did not understand three days ago. Perhaps he or she skipped his or her homework or was not paying attention. Suddenly, much of what is going on does not make any sense to him or her. So, at least in principle, we have an opportunity not only to improve the outcome, but also to shrink the distribution by using this technology wisely.

I am showing a goal we have set for ourselves, which is to get at least halfway there. I think our long-term goal ought to be to do better than the individual tutor because in principle, a student ought to be able to get everything the individual tutor offers through mentors, who might be directly available or online. Further, our goal ought to be enable the student to take advantage of some of the things he or she can do with the equipment that he or she cannot do in a regular classroom.

I mentioned one of the other things that is of benefit in this technology, at least for adults, is not so much the expense of the education or the way it is delivered, but the fact that we just do not have time. Certainly, everyone over the age of 30 that was surveyed, time and the difficulty of getting the training was by far the biggest problem. If you are a single mother, or if you are taking care of an aging parent, or if you have some other complicated family situation, there is no way you are going to fly across the country for a three-week course or try to attend night school at seven o'clock . . . Through the internet we can have up-to-date material in a way that is directly accessible to a much broader chunk of the American community.

Having said that the opportunity exists, and I am convinced that "there is gold in them there hills," that the problem is that we are utterly clueless about how to get to it. There are a lot of experiments around; some of them are pretty lousy. Everyone has seen software products that are both good, bad and ugly. This can be a good thing, but it can also create a lot of frustration. It also creates a dynamic in which people like you need to step forward and help us invent a better strategy for using this absolutely unique opportunity.

I will take a fearless forecast and say that something like the structure that I am putting up here will be the way we will be delivering instruction. The reason this is relevant for K-12 is that I think it is the same kind of instruction that will be used to deliver preschool instruction, K-12 instruction, college material and advanced instruction to people in the workforce. One must have at least most of the elements that are on here, which is to begin with the student at the center, where the goal is to try to figure out what it is that this kid wants to do, what it is that he knows and does not know, and deliver it to him. One needs some way of tracking with a fair degree of precision the format transcripts for this person, but also a lot about what his or her strengths or weaknesses are, what his or her learning styles are. Does he or she have a handicap? Does he or she speak Spanish? What are his or her interests? One must be able to follow this continuously by watching with some care how he or she behaves in the instructional environment. Clearly, one must then bring in a series of components that enables one to put together instruction that is suited to the student and the style of the teacher, or the person who is in charge of delivering this instruction. Ideally, one would also want a very open market of materials that are easy to bring in and easy to use. That is to say that one can buy chunks and components which might be a control panel on an aircraft, or it might be simply an electric meter that you want to put into some kind of an experiment for some kind of a simulated physics lab. One would like to be able to buy these things easily, put them together easily without hiring an army of lawyers at a minimum, and have an electronic commerce model that works. And, of course, have a technical basis to have all of this operate.

Crucially, one element of this must be how to get real people into this mix. That is one of the great unknowns. Plainly, there are some things that are better delivered through this artificial system, which allows one very free access to simulations, data, and other material. What role, though, is a teacher uniquely qualified to serve? This requires us to ask some fairly profound questions about what the role is of the person who is in the room with the kids, who is encouraging them, answering questions, giving them hugs at the right moment, whatever is called for. It also requires us to have an ability to call in specialists who have an expertise in some area which might be the Pathfinder, or it might be some aspect of Islamic literature, something for which there is no local expert, but there might be someone who is willing to answer a question online via e-mail.

I put this up here, not to say that I have patented an invention to do this with any perfection. Some system like this will have to be put into place if we are to take advantage of this opportunity. My main point is that we are at the very early stage of inventing this architecture, but the time when it should be invented so that a system like this can encourage creativity of producing. The time is right now, and we have to get moving.

One of the things we need is simply to craft a vocabulary to make all of these pieces fit together. In fact, we are facing some perilous moments; we have a bunch of proprietary systems that are trying to define this universe on their own. I am optimistic that everyone begins to realize that, like the Internet, everyone benefits by adopting a few simple standards on a voluntary basis. Those with you say, "Well, what is the Internet?" There is this huge explosion of creative energy. "What is the Internet?" People may come up to you and say, "Who runs the Internet?" The beauty of it is that no one runs it. The only thing the Internet is an agreement on a set of very simple interoperability standards. That is it. If you could get a vocabulary like that operating for the objects and materials of education and training, I am absolutely convinced that you could unleash an explosion of creative energy because there is a lot of it out there. But these standards need to operate on the level of commerce, and there are a series of technical requirements that need to be developed. Fortunately, there are major organizations - Educause being one of them - which major universities and firms are trying to get together to actually develop this. Unfortunately, it is still at an academic level at the moment. We are struggling to find a way to put it into practice.

Something we must be very leary about is this: While the hardware to make all of this work is coming into place at a dramatic rate (note the access of people to the Internet). You will see in upper income groups, roughly 60% of the people have some access to it, either at home or at work. There really is a problem here concerning the people at the lower end of the spectrum, who are in fact not being connected. There are certainly some racial groups and other demographic groups who are not adequately represented in this. So, as a matter of public policy, we plainly needed to worry about this. Now, I hope that in the next ten years, when you go to K-Mart to buy your $300 information appliance, which is now a TV, that it will also be an appliance that will allow you to take advantage of all of the things I am talking about. In the U.S., at the moment, the percentage of households with color TVs is larger than the percentage of households with indoor plumbing. So, people have put their priorities here. This is something that people are going to want in a ten-year period. I am optimistic that everyone will be able to have some access to it. It is not going to be enough to wait ten years for all of this to sort itself out. Plainly, we have a challenge to make sure that everyone benefits.

Another point is that the scale of the technology itself will change. We are talking about not being at the end of Moore's law, but being somewhere in the middle of the increase in computational power. In fact, we will see factors of 1,000 or 1,000,000 increases in computational speed. We may see even greater increases in the capacity of communications over the next decade. The system we put in place needs to find a way to gracefully grow and accommodate these new powers. In my view, the most difficult part of the program is to figure out how to teach and mentor and help kids think using infinite power. At the moment, I strongly suggest, we are sort of stuck trying to figure out how to put standard textbooks and standard classrooms into an electronic form. I am convinced that we will find that a dumb thing to do, but the better model has yet to be invented. We need to start thinking about that.

I think we should be sensitive, however, to the fact that the things that are currently available are quite extraordinary. The kinds of multimedia computers that are not in most classrooms, and the speed of connections . . . . I think at least 94% of all schools have some Internet connection, and I think 30% of the classrooms do. We can do some pretty slick stuff. E-mail is a killer application]. E-mail allows a student to talk to experts around the world. One can certainly do quite powerful simulations and models of laboratory systems using pretty lowbrow computers. We have not figured out what to do with those two simple tools, let alone figuring out how to expand the margin of these capabilities.

I had said I wanted to start off by establishing why it is that we are intrigued by the possibility that technology has an opportunity to make solid contributions to how people learn at all levels, from preschool up to adult learning and education. The practical question, then, is what do we do about this, particularly since we are in the middle of the process of invention. Well, I am not going into gory detail, but let me rehearse a handful of things we do have in place.

In the first instance, I think we have a pretty good record of moving the hardware into place. We have, at least in the K-12 environment, a series of programs - The Technology Literacy Challenge, The E-rate, and several other programs that are in the process of getting classrooms connected and making sure they have computers at least powerful enough to provide minimal access to the kinds of tools that I described. One thing we have in the computer is something called Computers for Learning. This is a program in which schools can register to receive surplus federal computers.

The question of what kind of instruction do we want to deliver through this is one of the more vexed questions because there is a lot that is unknown. When you start exploring the boundaries of why it is we do not know how to teach with computers and communications, you rapidly run into the fact that there are a lot of very profound, very basic things about how we teach anyone anything using ancient tools that we do not understand. The public investment in research in education generally is pathetic. It is something like .2% of sales. Any business that spent .2% of its income on R&D would be in deep trouble. We are trying to see what we can do to repair this. We have a small program in the National Science Foundation and in the Department of Education called the Education Research Initiative, which we are hoping to expand. The gentleman who runs the OERI (Office of Educational Research and Improvement) in the Department of Education is absolutely determined that the Department of Education should be a research power, and he has put a lot of thought into how one rebuilds that organization to be the equivalent of the NIH of education research.

Obviously, investment in the use of technology to improve education and training is a key part of our entire research portfolio. We have had several presidential advisory panels now. The President's Council of Advisors on Science and Technology and The President's Information Technology Advisory Committee, if I have my acronyms right, both have written reports in the last year saying that we basically need to significantly increase our support of innovations in instructional technology. We are trying to respond to that in the budget.

So, as a result, one of the things we are doing in this year's budget, in addition to increasing the direct funding for educational research through the National Science Foundation and the Department of Education, we are also making it an inherent part of a very large expansion in software research that was recommended by the PITAC (President's Information Technology Advisory Committee) panel. The PITAC panel's message was, "It's the software, stupid." This was their bumper sticker. All this power means nothing if you do not deliver something that make sense to human brains and is in a form that works for them. Also, it has to be in a form that does not break down all the time, so you are spending most of your time thinking about the computer instead of the subject matter that with which it is supposed to be helping you. We have a big problem there. We have asked for, I believe, a $150 million increase this year in software research, a large part of which is directly applicable to the classic problems in instruction that I just talked to you about.

Another thing we discovered is that the federal government is itself the largest trainer on the planet. We do not even know how much we spend educating people, but we probably spend between $30 and $50 billion per year training people in the military, in the civilian agencies, and among associated contractors. So, as the proprietor of the world's largest training enterprise, the President asked the obvious question, "Well, how are we doing?" There was much shuffling around. His response was to sign an executive order earlier this year that instructs the agencies to make training an integral part of the goals of the organization, and to use technology where it made sense, and to examine those possibilities. So, having been instructed by the President to go off and find the best commercial practice, the best universities in use of technology, they have dutifully have gone scurrying off and called me up saying, "What IS best commercial practice?" This is often a sort of embarrassing question because no one seems to know what best commercial practice is. The real answer is that we will have to help work with universities and the business community to try to invent it. We are hoping that by being intelligent consumers of this stuff ourselves, and by being able to react quickly and efficiently, we can actually help invent an architecture and an approach that will be directly relevant to the instruction at all levels, from K-12 to colleges, or wherever instruction is developed. We ought to be able to react more quickly than 15,000 school boards and be able to experiment on ourselves. So, we have set about doing this under the able leadership of a gentleman named E. B. Blanton. We are trying through this activity to make a number of very simple points.

One point is that in the gipper of plans, every agency must adhere to the Government Performance and Results Act. The President said training your employees is an integral part of your agency's mission. I want to see these plans, as well as the target, timetables and objectives you set. I want to see how you buy the training. If you are training people in house, perhaps it is more efficient to outsource this - certainly in technical developments -to use universities and businesses in other places. I want to see you figure out how you conform to the emerging voluntary standards that are coming in schools and in the business community, and to help that community develop these standards. This turns out to be really difficult because most of the agencies do not have the technical sophistication to do it. So, we have actually asked the Department of Defense (DoD), which has the staff to do this, to attend these meetings and then to help brief the rest of us on making this work.

Finally, we have asked each agency to come forward with a creative use of technology in one of their mainline training activities.

We think we have an interesting portfolio of activity at the federal level to try to help foster this sense of discovery that is essential to taking advantage of certainly technology in this environment, and crucially in figuring out how to tie people and machines together in a way that enhances the learning environment, making it much more of an adventure tied to something that students of every age group find relevant and important. The opportunity is plainly extraordinary. The outcome, I have to say, is uncertain given the small investment we are making compared to the problem we have at hand. The only way this will work is to find some way to exploit that creative energy that is out in the American community. There is a lot of it. If we can provide some way of organizing it around this mission, we will have succeeded brilliantly.


Joint Federal

Risk Assessment
    Toward Common

London Principles

   for Epidemiology

Endocrine Effects

Other Areas



Jazz Band





Home Mission Organization Funding Site Map Contact