Thank you for your kind letter of March 10, 1983. It was entirely my pleasure to appear before your subcommittee. My response to the Questions posed in your letter are as follows: In regard to Mr. Henry M. Levin's article in the January 30, 1983 issue of The washington Post. As you noted, part of Mr. Levin's article suggests that many new jobs within the data processing industry will not require an "extensive" background in math and science. Mr. Levin is correct. There are indeed many data processing support and spin-off Jots where extensive" math and science is not a mandatory requirement. within my own company, I would estimate that each "high tech" position generates approximately four other less technical positions. While these less technical positions do not require the depth of math and sulence knowledge you would expect of an engineer or a scientist, they do require a sound basis in those disciplines. If we use Mr. Levin's data and logic, it is possible to argue that each new computer programmer position will generate the need for nine other unskilled (non technical, jobs, i.e., janitors, fast-food workers. etc. The economic result then is that one high tech position creates four other less technical jobs which, in turn, generate 36 unskilled jobs. In my opinion, that's a pretty good argument for your Bill. I think that the principal item in Mr. Levin's article that distrubs me is the inference that unskilled labor should be relegated to their fate. Mr. Levin is apparently unconcerned with providing the educational opportunity which would permit them to improve their situation. Having personally worked my way through 17 years of night school to earn a Bachelors and a Masters Degree. I am comforted that Mr. Levin had no control over my destiny. I would not suggest that everyone be trained as a mathematician, scientist, or an engineer, but as those professions advance and influence our daily lives through technology, there is a proportional need for a sound basis in the tools of technology at all other levels, 1.e., science and mathematics. Today's high technology is tomorrow's norm and education must keep pace with technological growth or our society and economy will certainly become stagnant. I have attached an article by Louis A. Girifalco from the Fall 1982 issue of The Wharton Magazine entitled "The Dynamics of Technological Change" which addresses this issue in greater detail. Your second question related to Legislative encouragement to create a mutually beneficial relationship between education and industry. I believe this could occur in several forms ranging from Government sponsored programs to informal contacts between Legislators, educational institutions, state bodies, individual industries, and affiliated associations. Some possible examples of such encouragement might be: a) Encourage industry to seek accreditation of its suitable training programs, at both the high school and college level. b) Recognize that the cost of training in industry is, in some cases, c) Recent industry retirees should be encouraged to enter education -- e) Where appropriate, exchange programs between industry and education f) Finally, I would not overlook the impact of personal appeals on I want to thank you and your subcommittee for the opportunity to offer my testimony and respond to your questions. I am pleased to be of any assistance in your endeavors to improve the quality of education. The Dynamics Of Technological Change Technological change is frequently discontinuous. The major challenge we face as a technological society is to learn how to navigate these discontinuities. By Louis A. Girifalco what is going on and how fast it is happening. For example: What is the time lag between scientific discoveries and their commercial application? What is the frequency of new technological developments, and how does their frequency change over time? What is the rate at which new technological developments spread, and how long does it take for a new technology to supersede an older one? The second kind of question is more profound and concerns the reasons the changes are taking place and the factors controlling their rates of development. Questions at this second, explanatory level would include: What determines the time lags between scientific discovery, invention, and commercial application? What are the forces that inhibit or encourage technological change? What determines the rate of technological innovations and the rate at which they spread? We can say more at the descriptive level than at the explanatory one- where the really critical issues lie. But even at the descriptive level there are serious problems of carefully defining terms, of choosing measures for rates of growth, and of finding or selecting appropriate data. In other words. the conceptual groundwork even for the description of technological dynamics is still uncertain. And the second, explanatory, level is much harder to handle, including a host of technological, economic, regulatory, and social factors that are intertwined in complex and varying ways. Still, some progress can be made by examining the available descriptive information on technological change, and trying to generalize it in the hope that we will then be led to some insight about the basic controlling forces. There are four kinds of data useful to us in this endeavor: 1) the rate of technological diffusion (the spread of a new technology through a given industry). 2) the rate of technological substitution (the replacement of one technology or prod uct by another). 3) the frequency of technological innovation (the number of new developments introduced per year), and 4) the innovation time lag (the time be tween an invention and its commercialization). The work of gathering and analyzing this data can seem prosaic and even dull After all, there is nothing intrinsically exciting about the number of steel companics that adopt the basic oxygen furnace, or the number of television sets purchased, unless you happen to be in the business of selling oxygen furnaces or television sets. However, analysis of the empirical data can lead to far-reaching conclusions with the most profound implications not only for the world's present condition but also for its future. A Look at the Data The seminal work on technology diffusion was done by Edwin Mansfield, a pioneer in the quantitative analysis of technological change. For each of four industries (bituminous coal, iron and steel. brewing, and railroads) he examined the time it took for a new technological development to spread from company to company within its industry. His results indicated that after a new technology was adopted by one or two companies, it spread to other companies at an initially slow rate. This rate increased rapidly for a time, then slowed once more. leveling off as all the companies that would adopt the new technology did so. Studies of this type have been done for a number of new technologies, ranging from catalytic cracking to numerically controlled machine cutting, and also for the international diffusion of technology as a whole. The results are all similar. When the number of firms that have adopted a new technology is plotted as a function of time, we get the S-shaped curve that is shown in Figure 1. Before considering the meaning of this. let us look at the somewhat different but related phenomenon of substitution: the displacement in the marketplace of one product by another of superior technology. Examples would include the substi tution of transistors for radio tubes, of jet engines for piston engines in aircraft, and of synthetic fibers for natural ones. Full 33 widespread and the technology grown First, the growth curves are nonlinear change not inly has a velocity, it has Second, there is a lumut to how much How quickly does a new technology supersede an older one " The takeover time how long before a new technology moves from a 10 to a 90 percent share of the market can vary greatly for instance the takeover time for computer cak ula tors was only eight years, while it took nearly sixty for synthetic rubber to re place natural rubber. For a large number of sex hung a' changes the takeover time has ranged between twenty and féry years Averaging over homoschuld appres ances of various kind for example a takcover time of 27 years was found While the S curve gives a good pen cral picture of how a technology grows the spc, fic values of the lakeover time arc determined by the forces behind the techno These are not clearly under stood although we work has been done on the effects of such factors as the return an investment, the cost of intr sducing a new technology and general ecomC TRATTA It is important to remember that when a new ex fundegy wins an old one loses The Scurve is the result of new tech nology grow ung at the expense of old and indeed often at the expense of com Technological Discontinuity and the Kondratieff Cycle T. ben unes manure. By 1975, for example more than 95 percent of all electrically wond hand in the mused States hat "adem, opti gerators, vacuum clean e washer dryers, and urlevision sets The smart appliance industry, which had hers geared for growth for several dec som omấy a slagnani replace the long range price movements of com As a new cluster of technologies is #magnates. This is in fact, the kind of molupes comes along post about exqual to the average time takes for individual technologies to start grow and go to stagnation as they climb their $ curves But the tramution from onc chụp of technologies to another facem cung Cycle to the next is des contin gogs some there is no preparation for the new wave unto the old one is spent Radi, að une alarms take place during periods of depression. During promper thes fumes they seem to be no reason to þó und ‹ anything new. Investmund in the Current key fino doagies pays well, without the risks gết, nhưng untried ventures. Dur ing depa v however capital socks new kinds of investments because the old ones, arc de menstrably in truubic. Mature industries are not producing near thei not be probable. The validity of thes argument is uncertain but then does ap pear to be a correlation between innova te frequency and the bang tangu huvaness* cycle 21-390 0 - 83 - 34 www |