A guide to random selection

In the democracy of classical Athens, voting was considered a last resort: an unfair and undemocratic popularity contest. Instead, almost all government positions were filled by a random ballot of eligible citizens, meaning that the government was truly representative of the citizenry – although at the time that category was limited to a small hereditary class of men.

We can see the legacy of this idea today in our jury system — a random panel of twelve people are meant to give the fairest hearing of the evidence. In more modern contexts, Citizens’ Assemblies, such as those in Ireland and Denmark, have been formed from randomly chosen citizens to hold in-depth discussions on major constitutional changes, and to propose agendas for future governance.

Old idea put to new uses

In the world of funding schemes, many organisations have experimented with random selection, led by the New Zealand Health Research Council in 2013. These experiments have been intended to increase the variety of work funded, by reducing the need to write ‘safe’ proposals. Organisations like the Swiss National Science Foundation, Innovate UK, the Nigerian government, and the Volkswagen Foundation, have all trialled random selection processes for funding science and innovation. 

In the arts, there have been several schemes from large organisations like Jerwood Arts, as well as smaller, more DIY organisations like The Uncultured. Alongside aiming to increase the variety of projects funded, random selection is seen as a fairer way of distributing funding, which reduces the power differential between artists and funders.

Reducing unpaid labour

For the last year at Watershed in Bristol, we’ve been experimenting with random selection across a range of our programmes, as an alternative to more traditional selection processes for applicants. We were initially inspired by Jerwood Arts and their 1:1 Fund, in which they introduced randomness to increase the fairness of their application process. We’ve found the real benefit of the system comes in reducing the stress and unpaid labour demanded from applicants.

Before we introduced random selection, our annual winter residency programme typically received up to a hundred applications, competing for two or three places on the programme. Each of these applications included a written project proposal, which could take days to produce. By our estimates, adding up the total time spent across all applicants, we were asking for months of unpaid work — far more than the amount of funded time we offered to successful applicants.

In our new model, we ask applicants to complete a simple form to confirm their eligibility. Then we randomly select a pool of twelve applicants who form a shortlist for the residencies. Each of these applicants has time with our team to discuss their idea, after which they write up a full proposal and have a more formal interview. Because of the limited numbers, we’re able to pay a small stipend to everyone on the shortlist to cover their time preparing the full proposal.

Three models: total, early, late

When we talk about random selection, people often assume we’re giving out residencies entirely at random, removing ourselves completely from the process. This is a version of the process that we call ‘total’ random selection.¬¬ It’s the simplest way of doing things, and it has real advantages in terms of speed and transparency. However, it’s a difficult process to control, and may lead to undesirable outcomes.

Our process is a version we call ‘early’ random selection. We put the random element at the first stage and use it to select a shortlist for a more traditional process. This is less chaotic than a total random process, while still greatly cutting down on the amount of work for applicants (and assessors).

Other organisations have used what we call a ‘late’ random selection process. In this version, a traditional selection process produces a shortlist, and successful applicants are chosen randomly from that list. It means that anyone who meets the criteria has an equal chance at success, which promotes fairer outcomes. With this method however, there aren’t the same savings in time and labour possible with other methods.

Part of the toolkit

Random selection has quickly become a standard part of our toolkit when designing selection processes. As well as in our residency programmes, we’ve used it to select participants in research projects, and to identify representatives from our community of resident artists to be involved in our governance structures. In each case it has promoted fairness and a respect for the time and labour of applicants.

Of course, there are times when random selection is inappropriate. In areas such as job applications, or some of our larger funding programmes, we’re sticking with more conventional processes. Nevertheless, the thinking from our random selection work has spread into other areas of our organisation. We’re now much more conscious of the invisible labour that goes into preparing applications, and we weigh every new question on an application form against the additional work that goes into answering it.

Random selection is now firm embedded in our working practices, but we haven’t fully explored its possibilities yet. Organisations like Theatre Deli, New Diorama theatre, and The Uncultured are exploring this too and have created processes that fit their circumstances and the artists they work with. 

We’ve pulled together thinking from across these organisations into a guide to random selection, which we hope will inspire others to experiment. Above all, we hope that this work can lead the arts sector to rethink the purpose of their application processes, and to amplify our focus on how we can best support the artistic community.

Martin O’Leary is Studio Community Lead at Pervasive Media Studio, Watershed
@mewo2 | @PMStudioUK | @wshed