Normalisation options

As discussed in the previous section, the "Towards Responsible Publishing" survey was distributed electronically to researchers across the globe, aiming for wide distribution to capture diverse perspectives from various regions. However, due to differences between regions in the size of the researcher population, the number of public ORCID records, and the response rates, the number of respondents varies significantly, going from 224 in Oceania to 1318 in Western Europe.

Because of that, you can choose to see the survey results on this website according to their absolute numbers - showing the unweighted count of responses and providing a straightforward representation of the data - or choose between two types of normalisation inspired by the bicameral legislative system seen in many countries. In such systems, there are typically two houses of representation:

House of Representatives: The number of representatives is usually proportional to the population of individual states or provinces. The idea is that these representatives will equally represent the voices of each citizen. Our normalisation approach to guarantee a proportional voice for each researcher globally is to weigh the number of responses according to the size of the researcher population in each region, using estimates from UNESCO, as seen in the following table:

Table 4. Researcher population (in full-time equivalent) estimates by region.

Continent	Region	Researcher population
Africa	Northern Africa	189.000
	Sub-Saharan Africa	93.000
Americas	Central America and the Caribbean	53.000
	Northern America	1.705.000
	South America	286.000
Asia	South-eastern Asia	422.000
	Southern Asia	615.000
	Eastern and Central Asia	3.641.000
	Western Asia	272.000
Europe	Northern Europe	567.000
	Southern Europe	454.000
	Western Europe	1.092.000
	Eastern Europe	725.000
Oceania	Oceania	147.000

Source: http://data.uis.unesco.org/index.aspx?queryid=3685, weighted by population

Senate: Each state or province has the same number of senators regardless of population, ensuring their voices can be heard equally. Our approach is to normalise the data per region so that the size of the research population or the response rate of a region does not influence the results, allowing for a view of the results where every region has equal weight.

Users can filter and detail visualisations from different perspectives, whether they select to view results with or without normalisation or the method chosen. This allows them to see the viewpoints of each region or continent as they see fit.