The calculation of the Open Knowledge Index follows closely the conventional literature on composite indicators (OECD 2008). The index is still in an early testing stage and we are still tweaking the dimensions as well as conducting robustness checks – the results should hence be interpreted with care. The data is currently confined to a 2009 cross-section for 38 countries, but we are planning to extend the sample in the future to allow more fine-grained analyses.
To ensure data quality, the initial sample is restricted to the OECD and the BRIC countries, yielding a cross-section based for 2009-2010. In order to avoid missing data, observations in a 2 year window between 2009 were merged to 2009 – this is a standard procedure under the plausible assumption that many of the dimensions measured change only slowly between consecutive years.
THREE DIMENSIONS OF OPEN KNOWLEDGE
The Open Knowledge Indicator measures knowledge in three dimensions:
I) Capability: Following Sen (1999), capability measures whether individuals have the capability to access and process data and knowledge. The availability of data and tools does not necessarily imply that citizens have the knowledge about how to access and understand the information. This problem, known as the Digital Divide, is particularly evident in the stark differences between the high income and lower income countries. As capability is difficult to measure directly, we proxy these dimensions using:
- Fixed broadband Internet subscribers (per 100 people) (OECD)
- Newspaper circulation rate (World Bank)
- Press freedom (Reporters without Borders)
- Tertiary education rate (World Bank)
- Years since first Freedom of Information legislation (OECD “Citizens as Partners”)
- Depth of Freedom of Information legislation (OECD)
- Open Budget Index (Open Budget Partnership)
- Effective access to information (World Bank)
- Number of Wikipedia edits per 100.000 inhabitants (Wikipedia)
- Open Source Index (Red Hat)
- GI Civil Society Index (World Bank)
- (Number of twitter users per 100.000 inhabitants)
- The z-scores of each variable is calculated in order to make them comparable. The intuition follows from the fact that many variables are measured in different units – by standardizing, each country is expressed relative to the other countries, enabling a more intuitive comparison.
- For each of the three sub-dimensions, each variable is averaged to yield a composite sub-indicator for Capability, Legislation and Open Knowledge Society. Since weights are often arbitary (Sen Stiglitz Fitoussi 2009), we have two approaches: In the first approach, we just take the arithmetic average, i.e. assigning equal weight to all dimensions. In the second approach, we open the Index itself and let the public vote (Yourtopia)
- Finally, sub-indices are aggregated to yield the final Open Knowledge Index. Again, we assign equal weights by taking the arithmetic average. After normalization of the index take values between 0 and 1 with higher values indicating better Open Knowledge performance.