ALBERT

Notes: Purpose - Based on an Israeli innovation incubators program, aims to describe a procedure to define relevant categories in order to reduce the complexity of representing the population of projects supported by an innovation program in a meaningful way. Design/methodology/approach - Provides a concise description of the Israeli innovation incubators program and presents the characteristics of the data available. Introduces the analysis methodology and describes in detail its application to the specific data set and the classification scheme that it induces. Discusses the information that can be obtained based on this classification, and outlines some policy implications. Concludes with remarks regarding possible applications and extensions of the methodology. Findings - Finds that the classification induced by the iterative category exclusion (ICE) procedure can serve as the basis for more sophisticated analysis using statistical tools such as regression analysis. Once the relevant categorization is obtained, it becomes easier to collect meaningful data about incubator performance for analysis at various levels of aggregation. The ICE procedure has the advantage that it does not give a priori preference to any category or to any relationship between categories, and does not require a priori identification of the categories that are to be kept and those that are to be deleted. Further, it does not need to classify categories under headings such as "scientific and technological" on the one side, and "sector applications" on the other. Both are seen as equal from the beginning to the end and thus avoid any bias in the process. The subjectivity inherent in the selection process is reduced, perhaps even eliminated. The same values provided the basis for the identification and quantification of overlap and proximities between the final categories, throwing some light on their mutual dependencies and interactions and leading to a level of representation from which strategic assessment and thinking can start. Originality/value - The core principle of ICE is to eliminate categories that do not convey sufficient information to justify the additional complexity. This principle is universal and can be applied to a wide variety of situations that suffer from too much data and not enough information.