Further Details on Aggregation

Using the normalisation approach, we created standardized indicators for each area, with scales ranging from 0 to 100. For each area, we construct, to the extent possible, both a policy and a practice subindex. Two areas, CE in teacher training and period of compulsory education, do not have a practice subindex and are thus exclusively defined by a policy subindex.

The aggregation process to arrive at the overall EfDI consisted of three stages. In the first stage, the normalized scores of the indicators were combined to form a separate practice and a separate policy subindex for each area. Arithmetic means were employed for this purpose, reflecting the complementary nature of the indicators within each dimension. Each indicator’s relative weight within a subindex was determined inversely by the total count of indicators in that dimension.

The second stage consisted of combining the practice and policy subindexes for each area to develop separate subindexes for the seven areas. For this step we applied an 80% weight to the practice and a 20% weight to the policy subindex for each area, reflecting our understanding that in-school practice contributes more to the development of democratic competence than national laws or guidelines. The seven area subindexes are thus for 80% informed by practice and for 20% informed by policy. For CE in teacher training and period of compulsory education, as areas lacking a practice subindex, we took the policy subindex as the area subindex. In the final stage we combined the subindexes of the seven areas to produce the overall EfDI index. In this process, the areas of (1) CE curriculum, (2) participatory pedagogy, (3) school ethos and (4) period of compulsory education were given a 100% weight.*** CE in teacher training was given a 20% weight as it is only based on a policy subindex. A 50% weight was applied to the areas of unitary system of education and common education as these areas are mostly linked to the distribution, rather than the performance level, of democratic competences. We used the arithmetic mean of the seven area policy subindexes to calculate the EfDI Policy subindex. We used a similar procedure for the five area practice subindexes to develop the EfDI Practice subindex.

The EfDI Performance levels and EfDI Equality are constructed in a similar method as the overall index but with notable variations. Like the overall index, all components are normalized using ratio-best approach to map values on a scale of 0 to 100. We then aggregate these components using an arithmetic mean to create practice and policy indicators. Next, we apply the same weight of 80% for practice indictors and 20% for policy indicators to produce a weighted average for each educational area.

The differences between the overall index, the Performance index, and the equality index arise from the selection of educational areas and indicators included, which, in turn, affects the aggregation process. Additionally, some of the indicators have been constructed differently for the performance levels and equality index as they draw on different survey items. The indicators document explains in detail which indicators go into which index and which items are included in these indicators.

For the Performance index, we incorporate indicators for curriculum, pedagogy, ethos, training, and compulsory education. As in the overall index, curriculum, pedagogy, ethos, and compulsory education each receive a 100% weight, while training is weighted at 20%. However, since the unitary education system and common education were not found to influence performance levels, we exclude both and construct the EfDI performance index using only these five areas.

For the Equality index, we include indicators for curriculum, pedagogy, the unitary system, common education, and compulsory education, while excluding ethos and training. Beyond these exclusions, the equality index differs in its weighting approach. Because this index focuses solely on the distribution of democratic competences, we assign a 100% weight to the unitary system and common education. Additionally, we found that our only measure of curriculum practice was unrelated to distribution, so we removed it. With curriculum policy as the sole remaining curriculum-related component, we applied a 20% weight, mirroring the weighting used for training policy in the overall index.

The aggregation process for the overall EfDI index can be described in the following succinct way:

• Curriculum area = practice (outside activities x 0.8) + policy ((Formal curriculum + different programs + topic + aim / 3) x 0.2)
• Pedagogy area = practice ((student vote + open classroom + open discussion + active participation + participation at school / 5) x 0.8) + policy ((curricula guidelines + learning objectives / 2) x 0.2)
• CE training area = policy ((mandatory training + in-service / 2) x 0.2)
• Ethos area = practice ((student involvement + teacher participation / 2) x 0.8) + policy ((guidelines) x 0.2)
• Autonomy area = (practice ((autonomy of ce activities + official curricula / 2) x 0.8) + policy ((national curricula + autonomy of assessment / 2) x 0.2))
• Tracking area = (practice ((ability grouping) x 0.8) + policy ((age of first selection) x 0.2))
• Length of compulsory education area = years of compulsory education

Overall EfDI = (((Curriculum x 5) + (Pedagogy x 5) + (Training x 1) + (Ethos x 5) + (Autonomy x 2.5) + (Tracking x 2.5) + (Compulsory education x 5)) / 26)

**Even though period of compulsory education is only represented by a policy subindex, we gave this area a 100% weight because laws on compulsory education tend to be rigorously enforced, causing little difference between policy and practice.