Blog

Articles about Gorilla, behavioural science, academia, exciting projects and more will be posted here!


March 2018
16/03/2018

How to write a (successful) grant application

At Cauldron Science we see a lot of grant applications written by different researchers. Some get funded, others don’t. Over the years we’ve identified some of the characteristics of grant applications that are more likely to get funded.

The Structure:

The first and biggest factor is the structure and clarity of the research proposal. In our experience, the grants that are easiest to read and most often get funded are those that follow the structure of a research article closely. The methods section tends to be longer in a grant application than in an article. It is this section that will convince the reviewers that you know what you are doing, that it is good science and that you will deliver publishable work.

The introduction should take the reader in a straight line from “Why this is important to society” to the detail of your hypothesis. On route, you will cover what research has been done before and why your hypothesis is the logical and important next question. This section will be shorter than in an article. Make the minimum argument to justify your hypothesis. People will ask you to add stuff that isn’t directly relevant; avoid this.

The methods section should be a detailed description of the operationalisation of the experiment and how you are going to run it. This is the section that is most often overlooked. Reviewers want to go away with a concrete sense of what you plan to do and be able to say that it is good science. Pictures of the trial structure and experimental protocol are useful. They convey more information clearly in limited space. If you can create a prototype of the protocol and provide a link to it, even better. Several of our clients have created prototypes and even collected pilot data in Gorilla.

The results section should cover some detail of the statistical analysis you will run. The anticipated results (if your null hypothesis is rejected) and what that will mean with respect to your hypotheses. If you are testing to competing theories, then cover what results would support each theory and how you will disambiguate between them. If you are going to pre-register your study, then you’ll need this detail anyway, so embrace the opportunity to determine the best statistical analysis. Pilot data can help with this exercise and will often help you identify weaknesses in your experimental design which you can address to make your grant application better!

The discussion section will describe how the results above would contribute to the literature and what that would mean for society. This section will be far shorter than in an article.

Other sections:

Executive Summary: I like an executive summary at the top which states the research questions. This helps me follow the narrative of the introduction and also helps you keep the introduction focused.

Impact: How you will disseminate information to peers and end users of your research.

Steering Committee: Names of three more senior academics that aren't directly involved in the study and that have useful insights and / or contacts to aid dissemination. They are implicitly acting as mentors and sponsors by bringing their support to your project. They would typically meet with you three times over the life of the grant to give advice. Usually once at the beginning to discuss the experimental design, once at the data analysis stage, and once at the write up or dissemination stage.

Supplementary sections: Budget / CV / Publications / Project Schedule (Gantt Chart) / Resourcing

Collaboration

It’s practically impossible to succeed on your own. Identify people whose company you enjoy, whose intellect you admire and whose skill set and knowledge is adjacent to yours. Apply for grants together. You’ll each have your areas of strength; one person may be good at writing methods sections and creating schematics whereas another is better at the razor sharp narrative for the introduction.

Once you have a good draft get feedback from more senior colleagues. When asking for feedback be specific; either ask for input on a specific section or ask something like ‘what are the top three things could I do to make this better’. This avoids using a more senior colleague’s time copy editing your grant application and instead gets their strategic insight which is what you want!

Tailor your research proposal to the funding body

A good success rate with grant applications is around 1/5. Grant applications take a long time to write, so it makes sense to identify several granting bodies to approach. You’ll need to do so sequentially making improvements as you go.

When reviewing grants there are some that are obviously No and others that are obviously Maybe. The Maybes are all good science that are worthy of funding. The granting body would love to fund all of them, but they have limited funds. Consequently, they will have to choose between them: what criteria will they use?

They will choose the research that is most likely to deliver on their aims. To ensure this is your research, spell out the link between your research questions and the granting body’s / the specific call’s aims. At Cauldron we like to champion innovation in behavioural science research, so that is going to be a feature we look for.

Oh gosh, that seems like a lot of work!

When done well, a research proposal can be easily converted into pre-registration and can be the start point for published articles. Don’t let if feel like futile work, you are laying the foundations of numerous other documents that you will write.

How to get started:

No one likes funding an unknown quantity, so start applying for small grants and build your way up. Travel grants can be a good place to start, but also look out for other grants that are under £10k.

Cauldron (www.cauldron.sc) periodically offers small (<2.5k) grants to Early Career Researchers. To make sure you hear about these, sign up for a Gorilla account here (www.gorilla.sc).


December 2017
03/12/2017

The Great Ape Dictionary

Super proud of our grant winner, Kirsty Graham, who launched a huge crowdfunded study using Gorilla and collected data from over 15,000 people.

You can find the full article here: 'Can Humans Understand Chimps?'

Her study was also featured on the BBC Radio 4 programme 'Inside Science'

The team have since done a follow up study on whether humans can tell if two chimpanzees are related just by looking at their photos.

To take part, click here


December 2016
15/12/2016

UCL First Year Students Original Research and Poster Competition

Students were placed at both ends of the microscope, as subjects and scientists, participating in demonstrations and real experiments mid-lecture, reflecting upon their own data, generating their own hypotheses and designing their own experiments in laboratory sessions.

In lectures, lecturers used a system called The Hive, developed by the Eye Think Lab with Cauldron to study collective behaviour in crowds via their mobile devices. We adapted this tool to allow us to perform large scale experiments on students in the middle of the lecture, with their behaviour data sent to R scripts for immediate, online analyses.

In laboratory sessions, students employed Gorilla which was developed in collaboration with UCL. It is an online tool that allows people to create full psychology experiments with little or no expertise, as easily as they might create a survey with a tool like survey monkey.

For example, during the first term students first ran a prepacked experiment (an Implicit Association Test) that had been discussed in lectures. Then in small seminar groups they had a critical discussion of how to interpret results, and posed their own hypotheses.

The next week student created stimuli and adapted the IAT task to test their own ideas. They investigated 26 primary hypotheses from social stereotypes connected to downs’ syndrome to implicit attitudes towards people of different faiths as well as numerous secondary hypothese. Running the experiment online via their social networks, the class was able to collect data from over 1500 participants.

Students presented their work in a poster session modelled on a real conference. Some of their findings are being written up for publication in scientific journals, so that students will make a genuine contribution to the field.


November 2016
11/11/2016

Classic Tasks now on Gorilla

We've been created a series of Classic tasks in Gorilla. These can be found in the Samples folder from your dashboard in the Starter Project: Examples folder and include:

  • Stroop
  • Relational Reasoning
  • Primacy and Recency
  • Implicit Attitudes Test
  • Enumeration Task
  • Change Blindness
  • Big 5 Personality Inventory
  • Flanker
  • Levels of Processing (LOP) with original Craik stimuli
  • Dot Probe

There are also some tasks that can only be provided in code. These can be found in the Code Examples folder and include:

  • Towers of Hanoi (Adaptive and fixed number of rounds)
  • Digit Span (Forward and Backward, adaptive and fixed)