FAQ

What is Demos.science?

Ideally, questions of broad public interest would be answered by high quality experimental research studies. But research is often confined to very specific remits that funding bodies prioritise, and research agendas can be strongly influenced by commercial interests. Where relevant studies have been carried out, their findings are often hidden behind journal paywalls. And even if you can access them, you might find them to be based on a small number of participants very different to you -- and therefore of questionable relevance.

Demos.science is different. On Demos, questions we all face daily such as what the optimal nutritional choices are for different health conditions can be discussed, alongside potential answers. Those suggestions that garner most interest and support are then put to the test using experimental studies to generate evidence as to how well they really work. The entire process, from brainstorming research questions, through study design, recruitment of participants through social networks and data analysis is conducted openly, collaboratively and transparently.

Who is Demos for?

Demos can be used by anyone with an interest in generating evidence-based answers to everyday questions. User groups include:

High school students. As well as being a platform for research studies, we envisage Demos acting as an educational tool for students, providing an appreciation and understanding for how and why research is needed, and particularly the value of experimental studies in providing high quality, empirical answers. The need for this has never been greater, with ever increasing access to biased, anecdotal and misleading advice people turn to for guidance on everyday questions. We hope Demos will help inoculate students against misinformation through a better understanding of research, and a collaborative and exciting platform to experience this themselves.

Early career researchers. In current research environments, launching randomized trials can be prohibitively costly and potentially high impact interventions that are not on the radar of research donors, or lack commercial viability can be difficult to study. Demos will provide a platform for researchers to run large scale studies on questions that either they have identified as being of importance, or those that Demos users have highlighted as being of broad interest.

Patient groups. The abundance of patient groups and communities online is testament to them being a powerful resource for individuals to turn to for support and advice on how to manage their conditions. These forums also feature large volumes of anecdotal evidence on how conditions can be managed and improved. Demos will provide a resource where patient groups can facilitate randomized studies to generate robust conclusions on the value of different approaches and strategies.

Demos is particularly important in low-income settings, where research funding is scarce and mostly dedicated to specific remits of international donors. Demos facilitates grassroots research into problems viewed locally as high priority and randomized trials with practical solutions suited to local conditions.

What kind of research questions can Demos be used for?

The main focus for many of the research questions we anticipate being addressed relate to healthy life-styles, although we encourage the platform to be used to test the effectiveness of our choices in many other areas, for example parenting techniques, education, or just about anything we're pondering about whether it is worth doing -- if you are, there are probably enough other people out there doing so too, so why not run a trial to find out!

What kind of study designs can be run on Demos?

Demos studies can be of any design, but we're particularly keen on randomized trials, whereby study participants are randomly assigned to different groups, each with a its own intervention, or acting as 'controls'. Such experimental studies can generate high quality evidence, as the participants are assumed to be similar in all ways other than in the use of the intervention in the group they were assigned to (or no intervention for those in a control group).

Aren't big data and AI the way to go??

Big data, and its interpretation through AI, can indeed be useful in searching for patterns and identifying associations between different factors (e.g. a variety of risk factors and health outcomes). But with the advantages of increasingly massive datasets comes the risk of identifying spurious associations that are misinterpreted as causal. Demos is where big data goes experimental, and the risk of falsely concluding that spurious associations are causal is eliminated through large scale randomized trials.