This exploration examines the relationship between the level of funding received by biomedical research grants and the number of citations received by their scientific publications. The study focuses on {{dataset.length.toLocaleString('en-US')}} biomedical research grants supported by the Swiss National Science Foundation (SNSF), with funds granted between 2012 and 2022.
The dataset is consolidated from the SNSF data portal, and the number of citations for each publication associated with the grants has been obtained using the CrossRef cited-by API (as of May 15, 2023). To address the potential influence of grant duration on the total number of publications and funding received, the analysis focuses primarily on two factors: the grant's average annual funding amount and the average number of citations received by their publications.
You can follow along with the numbered sections below, which will guide you through some key findings. Click on the blue text or the section titles themselves to see changes in the chart. The fifth section allows you to visually explore the entire dataset independently.
For more information about the dataset, please refer to the footer.
The grants are categorized based on the funding scheme through which they were awarded.
Each grant is also associated with a specific research area within the broader field of medicine.
While each grant's call decision year was not taken into account for the analysis, it is evident that the number of completed grants decreases as the years approach the year of the exploration (2023).
Next, let's examine the financial aspect.
The chart displays the average funding amount per year compared to the average number of citations received per publication for each grant. Due to heavy skewness in the data, we apply log-transformation to both the x-axis and the y-axis. However, even after log-transformation, no evident relationship emerges between the funding amount and the number of citations per grant.
Next, let's examine a key factor that significantly influences the number of citations.
The only noticeable relationship is observed between the average number of papers published per year and the total number of citations received. It is evident that as the number of papers per year increases, the total citations also tend to increase, assuming the papers are cited.
The linear model used in this analysis explains only 39% of the variability observed in the data. The model shows further improvement when incorporating the log-transformed amount per year and the categorical variable of research area. However, there is evidence for heteroscedasticity, which suggests a violation of the underlying assumptions of the regression model. In this context, it may be more appropriate to consider regression modeling for grants that are more comparable within the same category. E.g. grants from the "Programmes" funding scheme.
By examining grants with the highest citation counts (above 8 on the log-transformed y-axis), it becomes evident that a majority of them fall under the categories of "Projects". Alternatively, when looking at the research areas, the prominent ones are "Experimental Medicine" and "Basic Medical Science".
In the next section, we will delve deeper into an analysis of the categories in general.
The chart shows the average citations per publication, categorized by the research area of the grant. Variances across the groups were found to be homogeneous. An analysis of variance confirmed that grants in "Experimental Medicine" and "Basic Medical Science" receive significantly more citations per publication. The difference between these groups is also statistically significant.
Upon closer examination of the funding received, it becomes apparent that grants categorized under "Experimental Medicine" receive significantly higher funding compared to other groups.
The main driving factor for citations appears to be the area of the research. The relationship between the higher number of average citations per publication in "Experimental Medicine" and the higher amount of funding received in this research area remains uncertain. It is possible that conducting research in this field is simply more expensive. Further analysis is needed.
Feel free to customize the settings according to your preferences. You have access to all columns of the underlying dataset for further exploration. If applicable, you can utilize log-transformation of the data.
However, it is important to keep in mind that relying solely on data visualization without conducting a thorough analysis of the underlying numerical data may not provide a complete representation of the statistical truth.
Tip: Click on any grant dot to open its corresponding page on the SNSF data portal.