MANAGEMENT OF TECHNOLOGY STRATEGIC DECISION BIAS BY ROLE IN FAILED TECHNOLOGY PROJECTS KENNETH ROSSON PENCE Thesis under the direction of David M. Dilts This empirical research examines how personnel, from different levels of an organization, reach decisions to terminate technical projects. Termination decisions are often considered failures and this research investigates strategic decision-making using failure as an outcome measure. Upper management personnel (executives) with the authority to cancel a technical project and the middle management (project managers) who manage the day-to-day operations of the project are surveyed in phase one of this research to determine what, if any, biases may have affected the decision to terminate a project under their control. Information gathering (scanning and interpretation) and environmental factors (importance, resources, complexity), leading to the decision to terminate a project, are also examined from an individualistic, not a monolithic, viewpoint among innovative technology organizations. Termination decisions are shown not to be related to the scale of a project but that different levels of the organization use different decision weighting levels with regard to project termination. Additionally, hindsight and sunk-cost biases are clearly shown by the two levels of decision-makers within an organization. Phase two of the research develops a model from two concurrent case studies portraying how bias affects project outcome perceptions.

This paper examines how the characteristics of a system determine the appropriate method that should be used for its life cycle extension. Such systems are termed legacy systems because they are approaching the end of their useful life, preventing them from maintaining their operational objective. Legacy systems were examined because they are expensive to maintain, can pose environmental hazards, and can hinder an organizations performance if not addressed adequately. My study developed a model to identify characteristics of legacy systems and the methods that should be used to extend the life of a legacy system possessing those particular characteristics. The model was evaluated by examining an example legacy system; a missile system in the defense industry; and noting the characteristics associated with it. These characteristics were then used to compare the method the organization chose to extend the life of its system with the method of life extension our model suggested it use. The study used case study research and pattern matching to evaluate its results; which determined if the model correctly predicted the life extension method the organization should chose. Theses results indicate the model can be used to suggest appropriate methods for life extension of legacy systems.

This thesis focuses on comparing the effects of university technology commercialization practices on their licensing incomes. "High Output" and "High Activity" universities were selected from top U.S. research universities based on the level of resources that are invested in the technology transfer process, as opposed to the level of income generated as a result of licensing activities. We use the multiple case study method,consisting of four "High Outcome" and four "High Activity" universities. The principles of real options theory are used to compare and contrast the technology commercialization processes of the two groups, based on the premise that universities that employ a real options perspective in their technology transfer decisions, will have greater licensing incomes. Based on interviews with technology transfer directors at the universities in the sample, we conclude that the most critical factors affecting licensing income are faculty involvement in the technology marketing process and the level of evaluation of the feasibility of start-up opportunities, prior to engaging in such investments.

This paper analyzes the characteristics of legacy systems and examines how those characteristics can be used to determine the correct method of life cycle extension. A legacy system is a system that is approaching or already at a point where it is no longer of use to an organization, preventing the group from maintaining its operational objective. Legacy systems are important to study because they are expensive to maintain and can hinder an organizations performance if not addressed adequately. This study developed a model that will identify the characteristics of a legacy system and the methods that should be used to extend the life of that legacy system. The model was evaluated using case study research to analyze an example of a legacy system; a guidance system for a missile from the defense industry. The characteristics of the legacy system were analyzed to predict the appropriate method to use for life extension. The method was compared with the organizations life extension method for the guidance system to determine model accuracy. Evaluating the results using simple pattern matching determined if the model could predict the correct life extension strategy. The results indicated that the model can be used to aid in deciding what method or methods could best be used for life extension of a legacy system.

This research will evaluate how changes in information latency during computer deployments affect performance of high risk decision makers. High risk decisions and decision-makers, in this case, are police officers deciding whether or not to arrest an individual based on information provided via computer from a distant source. There are three critical research pillars for this research. First, the impact on performance of naturalistic decision-making (NDM) research described for high-risk decision making through the use of a recognition-primed decision (RPD) model (Klein 1998) will be observed. Next, media richness posits richer media has more impact on equivocal tasks. Media of different richness was tested for impacts on decision performance. The recognition primed decision model, portions of media richness, and information thresholds are integrated to provide a model to relate information latency effects on high-risk decision-making. This model was evaluated using three wireless computer deployments by a city-county police department. Using archival data, subjects in this research, 406 mobile police officers, were followed over a six year period to evaluate the impact of varying media richness (from radio-verbal access to textural computer queries of varying types) using computers of varying access-information retrieval rates. Outcome measures of arrest warrants served and physical arrests based on probable cause (where a reasonable officer would believe that a person had committed a criminal violation of local, state or federal law), were recorded over all interventions. User performance traits are shown to be a major influencing factor in outcomes and processes changes. Finally, this research shows how individual user performance traits impact high-risk decision performance more than latency but only with high performers in the organization. Information latency is shown to have little impact on low performers. User performance traits may explain the productivity paradox 1 with computers. 1 Brynjolfsson, E. and L. Hitt (1998). "Beyond the Productivity Paradox: Computers are the Catalyst for Bigger Changes." Communications of the ACM 41(8): 49-55.

Cost as an Independent Variable (CAIV) is a way to develop systems in which the cost for the system is fixed and all other system attributes are traded off to meet the fixed cost. The Government has required CAIV to be used on systems that have billions of dollars committed to them. Research up to this point in time reveals that CAIV is effective in saving money on the front end up system development. However, there is no research regarding other system attributes that might be affected by CAIV. This paper reviewed the current literature to generate a survey to explore how CAIV could affect other system attributes. The survey was given to professionals who have used CAIV to develop systems, and from the responses received, propositions are generated. These propositions identify areas that need to be researched to determine the usefulness and limitations of using CAIV in developing systems.

This thesis presents a pilot study conducted to determine the effects of life-cycle costing on completed system properties (performance, cost and time). Life-cycle costing is a costing technique used to determine the long term costs of a system, from the initial research and development through product disposal. Cost categories such as operations, maintenance and repair are included in life-cycle costing estimates. A literature review of life-cycle costing revealed that little research has been done on the long term effects of life-cycle costing, specifically from the point of view of experts in the field who use it in their systems. Therefore, this was the topic of the exploratory research done in this study and the basis for the research model. The pilot study consisted of developing a survey instrument, testing it on a sample population, and analyzing the results to determine how accurately the survey questions answered the research question. The results of the pilot study indicated that there is a large degree of inconsistency in the perceived successes of life-cycle costing amongst experts in the field. Suggestions for additional areas to be explored in future research models are suggested in the conclusion of this pilot study.

Majority of Americans are paying moderate or a great deal of attention to medical health news. The ability to clearly determine patients past and potential medical information sources can help both physicians and patients make more efficient medical decisions, but limited study has been done to identify health care providers attitude about their patients use of this data. A survey was implemented at the Vanderbilt Ingram Cancer Center, Nashville, TN. Of the 110 individuals approached, 66 (60%) completed the questionnaires. Almost 90% health care providers expected patients to gather information from human sources, even though the quality of information obtained from some of those sources, particularly friends, relatives and acquaintances was poor (2.68 of 7). Patients were recommended to limit using friends, relatives and acquaintances as sources of information. The quality of information obtained from various sources was consistently rated lower by health care providers than patients. Health care providers felt that certain patient demographics are good predictors for cancer information source preferences.

Today patients and their companions use various sources to gain knowledge of cancer. The ability to clearly determine their current and future information sources can help both physicians and patients to better communicate and make more efficient decisions together, but limited studies have done for this. A survey was implemented in the oncology clinic of the Vanderbilt Ingram Cancer Center, Nashville, TN. Of the 468 individuals approached, 424 (91%; 257 patients and 167 companions) completed the questionnaires, with 166 patients paired with companions. Over 95% believed information enhanced their involvement in decision making and abilities to cope with cancer; 77% reported that information reduced anxiety. 85% will continue to search for cancer information in the future. Physicians, nurses, and medical pamphlets are still the most trusted sources. Internet and email were not used as much as expected but showed more potential uses and better quality in the future. Demographics including education, income, gender, age and working status are found as good predictors for cancer information source preferences. Future research is needed on physicians¡¯ views on cancer patients¡¯ medical information sources.

A Supply Network (SN) is a collection of firms that maintain local autonomy but who interact together to fulfill customer requirements. SN researchers have focused on the need to understand the reasons behind the diversity in the number and types of supply networks, as well as how these diverse networks interact, change and adapt over time. My research focuses on the dynamic growth aspect of SNs and addresses two fundamental questions: 1) how do Supply Networks grow and emerge and 2) are there simple rules and conditions that control the growth and emergence process? The dissertation presents an investigative framework for studying these and other questions concerning the growth and evolution of complex supply networks. An inductive approach is used to answer these questions. It starts by creating a new theory-based unified model of supply network (called UMSN) that incorporates four theoretical lenses, namely 1) industrial growth theory, 2) network growth theory, 3) game & market structure theory and 4) complex adaptive systems theory. The UMSN provides a holistic view for modeling growth and emergence in SNs. A generic rule-based modeling framework and a simulation-based computational framework using software agent technology was developed to operationalize and implement the UMSN. For investigating the growth phenomenon data and parameters from the US automobile industry over the last 80 years were utilized. The results and analysis show similar growth trends as empirically published data of well-structured industries such as the US automobile industry. The SN system grew and emerged as a complex adaptive system. The research also presents statistically significant results that supply networks grow and emerge based on interactive effects of local decision-making rules and environmental conditions, and that there is an underlying order to the emergence process. This research also develops chaos theory analysis techniques for predicting the SN system behavior over time; showing how such techniques can generate insights for policy makers and managers. My research contributes by extending the current state of knowledge of SNs as a dynamic system (network) and developing novel classification and analysis techniques. The insights drawn can aid managers/decision makers towards a better understanding of how SN emerge and grow.