Doctoral Dissertation Research: A Statistical Analysis of the Location Factors of High-Tech Centers in the United States, 1950-1997: An Evaluation Using Quasi-Experimental Method

决定高科技中心成功的区位因素是什么？之前对高科技产业的研究是否揭示了这个问题？如果不是，我们如何确定合理的因素和原因？本论文将通过对美国高科技中心的统计研究来解决这些问题。高技术产业作为主要的就业创造者和主要的产业创新者，在解释战后美国经济繁荣方面发挥了关键作用。关于如何可能复制硅谷、128号公路或研究三角园区等高科技中心的成功，已经有了相当多的讨论。研究针对主要高科技中心的关键区位因素进行研究，试图帮助希望发展类似创新中心的规划者和政策制定者。以前的研究集中在：(1)联邦支持的作用，(2)大学作为基础研究的提供者和受过教育的工人的提供者，(3)商业条件，如当地产业规模、专利、风险投资和外国直接投资，(4)生活质量，包括气候、生活成本、犯罪、娱乐和教育，以及(5)各种因素的组合被认为是重要的。最近的大多数研究试图找出解释现有中心稳健性的因素，但这些讨论中的一个严重差距是，它们关注的是高科技中心组本身的表现。据我们所知，这些研究得出的结论的可信度是有限的，因为没有使用对照组。这项拟议的研究将填补这一空白，首先确定合适的控制组，然后系统地测试关于高科技中心成功解释的假设。在清楚地了解这些中心的成长和发展之前，很难理解建立一个成功的高科技中心成长政策的过程。还需要其他方法，如偏离份额分析、计量分析和面板数据分析，以全面检验假设。这项研究项目将必然使用统计方法来确定解释成功的高科技中心的可靠区位因素。具体地说，将采用准实验对照组法。这涉及到选择一个与成功的高科技中心具有相似初始条件的控制组，并需要衡量控制组和高科技中心之间的差异。将利用1950年以来的美国县级人口普查数据，以排除大多数内生影响。在适当配对比较的选择过程中，为了防止变量之间的尺度问题，将采用马氏距离；这是人类学、生物统计学、流行病学和其他学科中经常使用的一种衡量标准(Isserman 1999)。回归分析将被用来比较和确定重要的因果因素。将检验相互竞争的假设，并提出结论和政策含义。这项研究的结果应该能更好地理解选定的高科技中心的成功之处。研究结果可以为高技术中心的发展提供指导，而不是现有的高技术中心文献。通过这样做，这项研究将揭示许多地方和地区政府和规划者感兴趣的看似合理的高科技区位政策。

What are the location factors that explain successful high-tech centers? Do previous studies of the high-tech industry shed light on this question? If not, how can we identify plausible factors and causes? These questions will be addressed in this dissertation research project through a statistical study of U.S. high-tech centers. High-tech industries, as a major job generator and as a major industrial innovator, play a key role explaining the postwar U.S. economic prosperity. There has been considerable discussion of how to possibly duplicate the success of high-tech centers such as Silicon Valley, Route 128, or Research Triangle Park. Research has been directed at questions on the pivotal location factors of the major high-tech centers, in an attempt to assist planners and policy makers who desire to develop similar centers of innovation. Previous research has focused on: (1) the role of federal support, (2)universities as providers of basic research and as suppliers of educated workers, (3)business conditions such as local industry size, patent, venture capital, and foreign direct investment, (4) quality of life, which includes climate, cost of living, crime, recreation, and education, and (5) combinations of factors are considered significant. Most recent studies seek to identify the factors that explain the robustness of established centers, but a serious gap in these discussions is that they focus on the performance of the group of high-tech centers by themselves. The plausibility of conclusions from such studies is limited because, to our knowledge, no control groups were used. The proposed research would fill this gap by first identifying suitable control groups and then systematically testing hypotheses on the explanation of high-tech center success. Until the growth and development of these centers is clearly understood, it will be difficult to fathom the process on which a successful high-tech centers growth policy can be built. Other methods such as shift-share, econometric analysis, and panel data analysis will also be required to fully test the hypotheses. This research project will necessarily employ statistical approaches to identify reliable location factors that explain successful high-tech centers. Specifically, a quasi-experimental control group method will be used. This involves the selection of a control group with similar initial conditions as the successful high-tech centers, and requires measuring the differences between the control group and high-tech centers. County-level U.S. census data from 1950 will be utilized in order to exclude most endogenous effects. To prevent a scale problem among variables in the selection process of properly paired comparisons, Mahalanobis distance will be adopted; this is a measure frequently used in anthropology, biostatistics, epidemiology, and other disciplines (Isserman 1999). Regression analysis will be utilized to compare and to identify significant causal factors. Competing hypothesis will be tested, and conclusions and policy implications will be suggested. The results of this study should provide a better understanding of the success of selected high-tech centers. The results can offer guidelines for high-tech center development, more so than the available high-tech centers literature. In so doing, the study will shed light on plausible high-tech location policies that many local and regional governments and planners are interested in.