1. Data used for the composite indexes for human resources management (HRM) are derived from the 2020 OECD (GOV) Survey on Public Service Leadership and Capability and the 2020 OECD (GOV) Survey on the Composition of the Workforce in Central/Federal Governments. Survey respondents were predominantly senior officials in central government HRM departments, and the data refer only to HRM practices at the central government level.
2. Each composite index is based on a theoretical framework representing an agreed upon concept in the area it covers. The theoretical framework for these indicators refers to specific principles of the OECD Recommendation on Public Service Leadership and Capability (PSLC) (OECD, 2019[2]), which represents an international consensus on standards for a fit-for-purpose public service. Each index is constructed in close collaboration with experts and reviewed and validated by the delegates of the Working Party on Public Employment and Management.
3. Three composites indexes have been developed to measure contemporary public sector HRM developments and dilemmas on how best to manage human resources in the public sector in the 21st century, such as the extent of proactive recruitment practices, the management of the senior level public service, and the development of a diverse workforce. The variables comprising the indexes were selected based on their relevance to the concept.
4. When making cross-country comparisons, it is important to consider that definitions of the public service, as well as the organisations governed at the central level of government, may differ across countries.
5. Various statistical analyses were conducted to ensure the validity and reliability of the composite indicators. The survey questions used to create the indexes are the same across countries, ensuring that the indexes are comparable. Missing values were at times an issue for the Public Employment and Management database. Different techniques for estimating missing values were used based on the nature of the missing information, including mean replacement, expert judgment and/or eliminating the country from the calculation of each composite indicator. In order to eliminate scale effects, all the sub-indicators and variables were normalised between “0” and “1” prior to the final computation of the index.
6. After testing several weighting options (including equal weighting and factor weights), and based on expert judgement, the index on the Use of Proactive Recruitment Practices was built on equal weights of the components and the index on Managing the Senior Civil Service was built on equal weights of the variables composing each sub-indicator and then equal weights of the sub-indicators composing the overall index. The index on the Development of a Diverse Central Government Workforce was built with a different weighting structure. To build the composites, all sub-indicators were aggregated using a linear method according to the accepted methodology. Some statistical tools (e.g. Cronbach’s Alpha) were also employed to establish the degree of correlation among a set of variables comprising each index and to check the internal reliability of items in a model or survey. This implies that the variables included in an index each has intrinsic value and they capture the same underlying concept. Finally, sensitivity analysis using Monte Carlo simulations was carried out to establish the robustness of the indicators to different weighting options.
