Introduction

In the first part of the post, I wanted to find out what are my chances of securing an ESRC New Investigators grant. Now, using open data from the UK Research and Innovation website, I wanted to have a further breakdown of the types of applications that were successful and how similar or dissimilar they are from each other. The ESRC grant reference numbers are found in the historical data from the ESRC website. That allowed me to search the RCUK website and retrieve information on successful applications. Unfortunately, I couldn’t get the data for those grants that were unsuccessful. Anyhow, I did some data mining to extract the data from the RCUK website and managed to obtain additional information on the successful applications, except for 2 that were not found on the website.

A little bit about the data

The list below shows the type of information that I have extracted and consolidated into a simple data table. I guess the column headings are rather self-explanatory. Perhaps only for the url, which is the web link to the project. The abstract is the abstract of the project and the planned impact is the description of the potential impact this project will have on society. The table below shows 3 web links to 3 different project if you are interested in seeing how it looks on the RCUK web page. Research areas are a combination of both research subjects and research topics. One project may have multiple research subjects and/or research topics. I don’t think it is particularly interesting to look at either the research subjects or research topics individually at this stage. Hence, to avoid confusion, I decided to put them all together and rename as research areas.

Quick overview

Okay, given this information let’s do some simple analyses and understand how grants are awarded. Firstly, let’s find out how many Universities were awarded in the past round. In total, 30 universities were awarded the grant for the last 5 months, with a total of 49 grants. Unfortunately, I was not able to extract information for two grants (n = 47). For whatever reason, they were not on the website. Perhaps the ESRC reference number did not match the website parameters. So all analyses are conducted on only 47 grants. If you want to know more about the success rate, please refer to my other blog post. In general, the table below shows the number of successful applicants per round (Please note that the two grants that were excluded from these analyses are from the March 2018 and July 2018 rounds).

Frequency of awards

Which Universities are doing quite well in the grant?

The Universities are shown in the table below and listed in alphabetical order. It would appear that the University College London, University of Exeter, and the University of Cambridge are doing pretty well. Note: We don’t know how many applicants applied from each university. It might be the case that these 3 universities had a lot more applicants than the rest. Hence, they tend to do better than others. To reorder the ranking in the table below you can click on the downward arrow located beside the column headings.

Frequency of awards in different rounds

What are the chances of a University being awarded twice in the same round?

Perhaps the table above does not give sufficient clarity. Researchers might like to see the success rate in each round. The table below shows the distribution of the grants awarded to different universities. We can see that several universities (e.g. University College London, University of Essex University of Cambridge) have been awarded the grant twice in the same round.

Across all 5 rounds, 7 Universities were found to be given the grant twice in a round. So by taking the division of 7 by all universities that applied (n = 30) and then multiplying it by 1/5 (5 rounds), the result of a university getting funded twice in a round is 4.67%.

Frequency of awards at a department level within a University

What are the chances of being awarded twice in the same around if the applications came from the same department?

Your research centre could be in a position where two applications might be submitted in the same round. Is it then possible that both applications are successful? Well, the short answer is Yes! (because it has happened before) But let’s have a look at the likelihood of that happening. With the current data, I can evaluate the information at a departmental level. In this case, only the Epidemiology and Public Health Department from the University of College London has been successfully given the grant twice within the same round. Thus, it is possible for the research department to submit more than one proposal, although the likelihood of such an event happening is really low. By taking the only department to get the grant twice within the same round (n = 1) divided by the number of departments (n = 35) and multiplying it by 1/5 (5 rounds), that will give me a 0.57% chance of such an event happening. Chances are slim, but if the research centre staggered the submission back to back, and it just so happen that the earlier application gets pushed back one round and that the evaluation of both proposals will be happening in the same round, at least you know that there is still a minimum chance of both proposals getting awarded the grant.

Frequency of awards at a department level

Which department has the highest number of successfully funded projects?

If you are just interested in looking at the type of departments achieving a successful application, then the data is presented in the table below. It would appear that Economics is the top with a count of 6, followed by psychology which is 5 (I combined psychology and experimental psychology). The rest are on average about 1 or 2 from each department across the 5 rounds.

Funds allocation

What is the average funding per round, and is that consistent across each round?

According to the call specification, it says that the grant annual budget will be around £7.5 million. Thus, they expect to fund about 35 grants a year, with about 10 to 12 proposals each round. So the expected mean value should be about £214,285 (£7.5m / 35) per proposal. My two questions to that are: 1) if one round uses fewer funds because it has a lower number of funded projects, does it mean that a higher budget will be allowed for the following rounds within the year? 2) Or is it the case that every round has a fixed budget and that cannot be changed even though there are leftovers from the money pot? Looking at the left graph (i.e. Total Funding Value graph) below for the rounds in 2017, I would infer that there is no fixed budget for every round, and the number of projects in each round is variable as long as the total funded projects for that year do not pass the annual budget of £7.5 million. The observed mean funded value for the July and Nov 2017 was above the expected mean funded value, by over £10,000 to £20,000. Such a trend existed probably because only one proposal was funded in March 2017, so given that there was sufficient budget, it was possible to fund more proposals in the following round. So to answer the question above, I would say that the panel tries to be consistent in every round but it is not a set condition. Those that applied in November 2017, lucky you!

Will we expect a similar trend for November 2018?

This becomes rather interesting. While November 2017 has a high number of funded proposals, it should not happen for November 2018 because of 2 plausible reasons. 1) If the expected number of grants to fund is 35, then the maximum number of projects that can be funded is 35 - (10 + 10) is 15. To be exact, including the 2 missing projects that are not in the dataset (one in March 2018, and another in July 2018) then only 13 projects can be awarded the grant. 2) The number of awarded projects are capped by the grant annual budget of £7.5 million. Since about £4.5 million is already spent in the first two rounds of 2018 (see table below), there is only a budget of approximately £3 million left. So if we divide £3 million by 13 projects, we would get an approximate of £230769.2 each, which is slightly above the expected mean funding budget (great!, though I doubt the target of the ESRC for giving out awards are dependent on reaching a maximum number of projects). If it is budget dependent, then a maximum of 14 grant (using the expected mean value) would be approved before it crosses what’s left in the budget. Either way, to have 16 funded proposals at this stage would be a challenging task, unless all successful proposals for November 2018 request for smaller funding compared to other rounds, which is quite impossible really. Given this reasoning, I would say that ESRC is more likely to select a smaller number of proposals but have a bigger allowance in the budget set for each proposal (I could be wrong), let’s see what happens in the Nov 2018 panel. We will have the data soon.