Boffinry

Setting the terms of the next engagement

The etymology of ‘the boffin’ refers to a Dickensian character, Nicodemus Boffin, a back-office intern, but its use was taken more seriously as science was used in conflict. Watson-Watt uses the term quite specifically for people neither confined to a back office role, nor remote specialists in an ivory tower. They were development scientists, both academic experts and consultants, eliciting the questions in need of a solution and testing new proposals for their effectiveness in the field. This lesson was learned, slowly, repeatedly and grudgingly during the second world war and its insights have not been readily adopted in public services since then, no doubt partly for want of priority on both sides.

Ronald Clark described the rise of the boffins in 1962, citing as motivation that others had not, drawing on interviews with the main protagonists and a few other sources. Indeed the popular accounts appeared as memoir from several of those a few years later, whether for restrictions on secrets or realisation of the insight. But his book is more comprehensive of the history, going back to the 1920s, and covering more fields than one individual was involved in. It can also give more account of the personal conflict between scientists which seems to have been a singular feature, then as now. However, the cadre which developed was well beyond personalities, even as they served to drive key decisions, to delay or proceed, sometimes without cause.

Operations

There are simple parallels with failure to act effectively to plan for events, from the British tanks which overheated in Libya and the deafened soldiers in Ajax tanks now to the list of scientists willing to advise solicited only in August 1939 and the recommendation of the UK pandemic public inquiry that a standing list of emergency expertise be held. But more interesting is the appreciation that science advice still leaves decisions to operational commanders and that it acts at different levels within a hierarchy alien to scientists. So requests can be in earnest but unrealistic at a purely technical function yet complacency prevails at a strategic level with various customs and prejudices or muddling through in tactical operations.

Boffins therefore originated operational research, often initiating the first systematic data collection in longstanding tactics. And the example of the mundane, counting the number of flak impacts to aeroplanes returning from bombing raids, is much more developed than simple survivor bias of their distribution. The vulnerability of different parts of planes was clear, as well as multiple hits, given assumptions about uniform dispersion (more than a mile up in the air anything else would be remarkable). But too many planes were not returning and fires were identified as a problem, not caused by direct impacts from anti-aircraft defences. Like the legend of John Snow and the Broad Street Pump Handle, much is missed in the retelling.

Tactics

Although interpreting negative information is already a profound contribution in a very practical environment, evaluation of different approaches is harder. Where these are constrained within one service, such as the convoying of merchant shipping, presenting data about the characteristics of the options is as much important as comparisons, as leaders are capable of considering the relevant factors. But where trade offs of resources to different functions are involved an evaluative framework confers legitimacy, and even then the emotional value of prioritising attack over defence had greater sway in the deployment of planes to attack civilian infrastructure not defend shipping from submarines.

Indeed RV Jones declared that strategy was just “tactics through a brass hat”. And it is clear that military deployments can be at a grand scale yet still tactical, the strategy is an area where scientists are not welcomed, save by those who think it serves their political advantage. Strategy is something most organisations find hard, instead delivering business as usual and negotiating increases or cuts in budgets as reward or efficiency without making choices. It looks beyond the tactical capacity horizon of doing everything and needs appreciation of objectives, external environment, capabilities and constraints. None of these are things that government is good at appreciating, as clear objectives may fail to have been achieved, politicians have power to raise funds but can be replaced if their plan does not match a manifesto.

Review

Capability is often confused with structures. Labelling something is an indication, and a signpost within a large organisation, but matching the complement and the responsibility is not. In the recent pandemic, advice structures for science evolved because of events but science structures also fell short. The public health agency was oriented towards a growing burden of chronic disease, health inequalities and winter system pressure (including seasonal ‘flu). Civil response mechanisms were designed for short term crises, most often localised extreme weather, and national security. More importantly, being good at one aspect of science did not mean capability to respond to what came next, nor that all of this could be integrated at scale.

In a 1947 valedictory report, science at war, radar, operational research, atomic weapons and naval warfare are highlighted as areas of progress. Similarly the 2022 science review of the pandemic showed success in vaccines, genomics and aspects of new data collection but also weaknesses in infrastructure for data sharing and general science capability. More detailed accounts show that advice was sought from the same source without reference to its expertise or authority. In the war heroic individuals emerge mostly because they prevailed to a hearing through force of personality; yet new advice structures on environmental transmission failed to influence pandemic health guidance without medical authority, and evaluation of gross measures termed non-pharmaceutical interventions was so limited as to leave little basis for future plans.

Nothing New

The problem with new things is that they are not developments of old things. There is an adage that sufficiently advanced science is indistinguishable from magic to the uninitiated. The four topics of science at war are all science but may involve distinct scientists who could not usefully drive forward even one of the others. While this can be appreciated in review committees, it makes science strategy more about faith and determination than wisdom. One of the hardest things as a young academic is realising a research proposal is not really judged on the robustness of the methods proposed but on the scientific vision in the proposed activity. Although risk appetite is a concern in the selection process, conservatism in favouring what eminent people want to do and prevailing excitement is more so. Even as basic as saying we need to study language development when everyone else works on reading is unpopular.

In the 1990s, this was recognised in clinical research and a new body was created to manage the evidence assessment of new drugs in particular. NICE did more than set a cost threshold for purchase of treatments (in quality adjusted life years) in that it set a standard of expecting everyone to produce evidence. But most things are not so clear cut as keeping people alive, and even in health care it leads to biases towards fatal conditions but neglecting chronic disease. Some organisations also generate evidence and publish areas of research interest but even the other ‘what works’ evidence centres leaning on the NICE brand struggle to guide practice. This may reflect the amount of time to get health to adopt rigorous testing, which still struggles in public health as interventions are not individualised in impact or allocation.

Government

The arrangement today has a range of specialised scientists employed directly in government bodies and research establishments. Government departments recruit a chief scientific advisor and other bodies generally a chief scientific officer (a nicety, to be sure) who may then convene specific advisory groups. Central coordination is often reinvented, so driven by priorities rather than scientific vision, but the government chief scientific advisor co-convenes a council for science and technology which publishes topical recommendations to the prime minister. However, further engagement of the research community in fellowships and other appointments still looks too limited to be more than development opportunities and small projects.

The big deficit is in scientific approaches to programmes, their design and evaluation, which are steered by senior civil servants with a policy orientation. As these people are often drawn from the fast stream the shift to more science graduates there may make a difference in due course. But science capability of the broad range of policy civil servants ought to get more attention, thinking broadly about large systems, data linkage and behaviour. All of these are the topic of public posts (other platforms are available) but narrative dominates and evidence on each type is rare enough to be difficult or fraught by problems of questionable practice the science. If civil servants rely on stakeholders and synthesis publications, they will meet a great deal of nonsense and need the skills to build plans based on fragmented analyses.

Leadership

In a war many people rally to the cause, as they did in the pandemic, scientists often putting the national response ahead of their own interest in publication. But even in the war they wanted to get back to their familiar publication practices, including holding onto data for their own use, soon afterwards. Crisis leadership from chief advisors convening experts needs to scale appropriately but outside of the crisis the idea of a national mission in science, even as infrastructure from laboratories to computing is acknowledged, is resisted. Indeed organising our scepticism in the Mertonian manner is as honoured in the breach as the observance: disciplines compete for resources and then deploy them without reference to others.

Experience can be relied on too heavily, excluding new specialists when others are out of touch with innovation or overwhelmed by demands on them. But it has many features, including the stress of working urgently, and the changed mode of iterative development and testing, compared to waiting for a publishable discovery. In terms of boffinry, this extends to familiarity with the environment and the people in it, building and sustaining trust in results and the practices which deliver them. Moving from one mode of working to something else is difficult and often unpopular, so leaders need to motivate, support and select individuals and connect between teams.

Faith

The surprising thing is that the public have an enduring faith in the scientific approach, despite criticism. Largely it is a better prospect than anything else, and the concerns about science from inequity, profit motives, political bias, elite interests and lack of transparency are levelled at other prospects much more strongly. And leaders can promote initiatives which ameliorate each of these perceived shortcomings, but as boffins they may be best placed to do that. Because most people take the integrity of particular scientific activity on trust, so the people who want to know about the politics and incentives as well as who is doing the work and how they choose their approach, are exactly those the boffins interact with on a daily basis.

The public has faith that science will improve their prospects, as it has done during their lives, but not because they believe scientists have that in mind. Research produced terrible weapons, and some discoveries were developed by persistence not any strategic support. And the work on what is possible is several steps removed from what is practical, so public support takes for granted that what is discovered will be put into practice. That is the boffinry sometimes reported in the media, yet neglected in the training of scientists and direction of scientific programmes. A more prosperous society is a convivial concept but not one which incentives for individuals promote at this time. Perhaps they should.