Open hardware provides a better chance for scientific reproducibility
This blog is part of a series on open hardware and key messages for public policy. Read the introduction and access other #OHpolicy blogs here.
By Andrew Hill, co-founder Open Acoustic Devices and Alasdair Davies, Director Arribada Initiative C.I.C
Scientific instrumentation built on open source hardware (OSH) is a compelling infrastructure for science. It not only allows anyone to replicate or reuse hardware design files for free, but also establishes a community framework where physical equipment can be rapidly reproduced, refined and improved on.
Scientific reproducibility has multiple meanings across different disciplines. For example, it might mean that an experiment should be repeatable by the same researcher using the same methodology and instruments, that data is reproducible by different researchers with the same methodology and instruments, or that certain parts of the experiment — which could be the instrumentation itself — can be reused in another experiment to make new observations. This final aspect of reproducible research requires scientific instruments to: (i) be accessible to other researchers, (ii) give accurate and reproducible readings, and (iii) have adequate support for reuse in other experiments. OSH can increase the chances of all three requirements being met.
Increasing access to instruments for researchers
The most apparent benefit of OSH is its free availability to all. This allows scientific equipment to be built by the researcher, saving them up to 90% on cost compared to equivalents provided by private entities. However, OSH is traditionally aimed at technically skilled individuals or those that enjoy coding and building equipment themselves. This makes hardware replication a challenge to less technically trained users. For example, OSH might be inaccessible to a biologist that requires a low-cost open source 3D printed microscope, but lacks the knowledge in mechanical design, software and electronics to build it themselves. One approach to improve accessibility is to provide pre-built OSH and user-centered software. OpenMV, creators of open source computer vision modules, used for smart camera traps, does this by providing rich graphical user interfaces and an extensive range of modules to cater for multiple scientific use-cases. Through the provision of pre-built hardware and well designed supporting software, it creates a user friendly solution aimed at the less technically trained.
Increasing the accuracy of instrument readings
Better evidence for the accuracy of scientific data and the reproducibility of instrument readings is more likely to be achieved if the basic tools on which science is measured are exposed to critical public scrutiny, leading to improved credibility. Hardware designs published in open source journals, such as the Journal of Open Hardware, Hardware X and the Journal of Open Engineering, use open access licenses, giving them a much wider reach and increasing the opportunity for scrutiny by anyone who reads the paper.
Increasing instrument support and the chance of reuse
OSH can be considered part of the collaborative economy, which is inherently transparent and value-driven. This increases trust and the chance for researchers to collaborate together on OSH reuse. Users take on collective accountability and online support communities tend to grow organically from project forums, such as those for Arduino, RepRAP, AudioMoth and OpenFlexure to name just a few. These communities often document and detail new usability and hardware issues, or even host improved documentation and evidence of data accuracy themselves. Forums create openly accessible archives, making it easier for new or existing researchers to overcome reuse barriers, thus helping them achieve their scientific outcomes. Furthermore, if ever the original creators were to stop maintaining the design, communities built on OSH give legacy value to the scientific tools in the future.
A leading question for reuse in OSH is the difficulties in creators obtaining financial resources to continue development after the first prototype is published. This often leads OSH projects to slow down after initial funds are spent and subsequently reduces the reuse of instruments that are yet to achieve critical scrutiny and take traction in a community. We therefore need ways to establish new frameworks and business models to fund creators, while also providing ways to assemble and distribute hardware to researchers that don’t necessarily know how to access OSH. The collaborative economy has shown promise towards this goal in the AudioMoth project, which uses community driven OSH consumption to drive down assembly costs and fulfil distribution of pre-built hardware to end users.
The impact of better reproducibility using OSH
You’re probably wondering why all the wildlife pictures? As authors from the domain of wildlife conservation, we view OSH as relevant and urgently needed to better protect biodiversity. There has been an unprecedented 68% fall in species abundance since 1970, the impact of which is now affecting the global economy, international food security and climate. However, to take effective action it must be better understood. The monitoring of wildlife is severely under resourced with a radical 99.99% of species unmonitored. Monitoring technology is therefore emerging as an essential infrastructure to inform conservation policy. However, technological limitations and high equipment costs presents a barrier to expand coverage. In an urgent push driven by species decline and assisted by OSH, conservationists from universities, governments and NGOs have established collaborative programmes with engineers to create bespoke tools for their specific applications. One example is the Oxford University’s Penguin Watch programme which used open source time lapse cameras to monitor and assess the status of penguin colonies throughout Antarctica. Better scientific reproducibility was achieved because OSH lowered the equipment costs to scale up the number of cameras the programme could deploy. This enabled the overall sample size — number of penguins monitored — to increase, improving data reproducibility, while also enabling other researchers to easily replicate methods through the sharing of hardware designs and configuration settings.
More recently, open source organisations such as Conservify, Smartparks, Rainforest Connection and Open Acoustic Devices have emerged to fill a collaborative engineering role for these types of programmes. An example of open hardware reuse in action was demonstrated by the Horizon biologging and telemetry platform, co-developed by the Arribada Initiative and conservationists at the Zoological Society of London. In this project the hardware was initially designed for low-cost sea turtle telemetry tags. New observations were identified by the National Geographic Society who considered the hardware for alternative uses. Together they iterated on the design with additional collaboration from The University of Exeter to create a repurposed Horizon module to track the movement of plastic waste along the Ganges Delta.
In summary, OSH not only helps researchers become better scientists by providing a heightened chance of data reproducibility, but allows whole communities to rapidly learn, reuse and modify tools to address new research challenges or urgent environmental and humanitarian disasters. More support for open hardware can help mitigate known challenges, and provide the resources required to accelerate professional science while giving communities the tools required to tackle the urgent challenges that matter to them.
