Indicator description

This indicator measures the extent of research on climate change by counting the number of peer-reviewed publications in the Web of Science database utilizing at least one of the following climate-related author keywords, climate change, global warming, climate action, or climate justice and normalizing the counts based on the total enrollment of graduate students in each country. 

Dataset(s) the indicator is based on

Web of Science (WOS) is a database of reference and citation data from academic journals. WOS curates the journals and other sources that it includes in their databases. They have standardized data which they collect about each publication which includes the title, keywords, author(s), affiliation, and year of publication. Clarivate is the owner of the Web of Science database.

Temporal scope

• Time period(s) collected data is available: Data was collected for publications between 2000 to 2020. The indicator was developed using the most recent six years divided into 2 three year periods.
• Frequency of data collection in the past: Daily.
• Frequency of data collection in the future: Daily.
• Change over time: Already included two time periods. More will be added.

Sample size, characteristics, and geographical coverage

All UNFCCC countries are included in this dataset. If a country is not represented then it is assigned a zero which indicates that no authors of articles on climate change were affiliated with an institution in that country during that time period.

Country coverage (out of 197 UNFCCC parties) by SDG grouping. Percentage in brackets corresponds to total number of countries per SDG grouping.

Data Preparations

The database provided by Clarivate included the assignment of a country (based on institution location) for each affiliation of an author of a publication which used any one of the following keywords: climate change, global warming, climate justice, climate action. Credit to each country for publications was assigned in a fractionalized manner based on the author affiliations. To do so, if an author had multiple affiliations these were consolidated (by country) and then assigned proportional credit. Then each author (and therefore their country of affiliation) was assigned a portion of the credit for an article (as a fraction of 1). These fractions of credit were then summed up by country for each year. 

The data was broken up into three year segments, with the most recent two utilized for now.  A country’s publication credit was averaged for each of the three year groups. This yielded articles per country per year for the indicator’s numerator.

In order to normalize the data, a denominator was created based on the number of students enrolled in ISCED 7 and 8 programmes (graduate students).  Reported numbers from the UIS database for each 3 year segment were collected and averaged within each one.  For countries with missing data the enrollment numbers were imputed. 

The articles per year was divided by the average enrollment and multiplied by 100,000 to yield for each country the articles published per 100,000 graduate students per year.

Indicator Calculations

1: No Articles

2: Fewer than 1 article in total

3: Fewer than 100 Articles per 100k students per year

4: Between 100 and 1,000 Articles per 100k students per year

5: More than 1,000 Articles per 100k students per year

Limitations

WOS has built their reputation on carefully vetting their databases and several studies have found error rates to be quite low, including with affiliations.

The underlying assumption of the numerator is that keywords well represent the content of the research in a publication. The denominator assumes that the number of graduate students is a good proxy for the size of the publishing research community who could theoretically choose to publish about climate change in WOS indexed journals.

This data captures all publications in WOS and many authors may be from institutions that are not Higher Education ones. Europe for example has several climate change focused research institutions that produce very large volumes of climate change research. Institutions and countries without graduate programs or pressure to publish in top journals will not be represented in this dataset.

While in theory the database covers the whole world there are several reasons that the less developed world will not be well covered. 1) English language bias 2) WOS systematically has worse coverage of journals in less developed regions 3) More science in developed world (e.g. many scholars from less developed countries will have affiliations in Europe or NA but not vice versa)

Acknowledgements

The MECCE Project would like to thank Clarivate for providing access to their database.