European Innovation Scoreboard: Strategies to Measure Country Progress Over Time

In the following chapters we examine possible alternatives to the current approach used in the European Innovation Scoreboard (EIS) to measure country progress in innovation performance over time. The quantitative approach used to assess country performance is the Summary Innovation Index. The methodology to calculate the SII scores and the SII growth rates is explained in the technical annex of the EIS 2007 report, therefore it is not reported here. We briefly recall the basic steps to calculate the SII growth rate. The SII growth rate is based on the SII values over a 5-year period. Such SII values are calculated using the min/max normalization technique (see below), using the overall minima and maxima scores across the full 5 years and across the EU27 + EFTA countries for each component indicator. Moreover, some identified outliers have been excluded from the calculation of the minima and maxima. Finally, as the EIS report says, >. In the first part of this report we examine whether available re-scaling approaches (i.e. indicization and normalization) are compatible with the formulas for the calculation of SII growth rates. So, we will revisit both min/max normalization and z-scores techniques and analyze their feasibility for the subsequent calculation of SII growth rates. In the second part, we will focus on the different ways to calculate growth rates, and the different meanings of the corresponding outcomes. We provide examples using the data available on the EIS 2007 Excel spreadsheet. We do not recommend specific approaches, yet we highlight which combinations of indicization, normalization and growth rate calculation should be avoided. The focus of the report is to raise discussion among the participants to the workshop of June 16, 2008 upon the relative merits and limitations of these approaches, with the idea to identify potential candidates for further improvements of the SII. The report is an overview of approaches that are in principle applicable to any given dataset. The report is not a feasibility study of a specific technique to the EIS dataset, for which more detailed analyses would be required given the constraints dictated by the quality of the dataset, including the presence of missing values.JRC.G.9-Econometrics and statistical support to antifrau