[This article appeared in the June 2020 issue of the ‘Irish Psychologist’]

As we all know by now, contagion is a matter of behaviour as well as biology. What we do, where we go, and how we think can all help to suppress the invisible enemy. Handwashing is back in fashion, experiencing yet another fifteen minutes of fame. But don’t forget the evidence-based guidance: be sure to sing “Happy Birthday” twice. Whatever your tempo, reciting both verses will enhance the antiviral effect.

Even at the best of times, behavioural science is hard. In a public health crisis, the pressure is really on. Wouldn’t it be great to be able to offer some of our “evidence-based” interventions to the world? Psychology’s time has truly come. The problem is that being “evidence-based” is more difficult than it sounds. Much depends on what you mean by “evidence.” And by “based.”

Unfortunately, objectivity demands that psychological advice offered during the current Covid-19 crisis should rightly carry at least six important caveats.

• First, there hasn’t been a similar pandemic for a century. Our so-called “evidence-based” interventions are not based on any evidence that was gathered during an outbreak like Covid-19. In that sense, there is much less “evidence” in this base than we might imagine.
  
• Secondly, when we say an intervention is “evidence-based,” we mean that its effect in a target situation has been found to be statistically significant. However, as with all behavioural research, the size of this effect is rarely emphasised. Many “evidence-based” interventions produce tiny effects in real-world terms.
  
• Third, our interventions are usually tested in an all-other-things-being-equal kind of way: an effect is extracted after holding other variables constant. In most psychology research, the number of extraneous variables is large. In real life, it is not only large, but also unknown. The rolling out of an intervention is itself an experiment, and should be seen as such. Lots of other things, including other preventive actions, are happening at the same time. Real-world results might not conform to our predictions.
  
• A fourth caveat is that no single study is definitive; advice is based on a consensus across multiple studies. However, different studies have different findings, many of them contradictory. It makes little sense to arithmetically subtract a negative effect-size from a positive one, and yet this is commonly done. Meta-analysis provides a misleading facade of sophistication; its garbage-in-garbage-out aspect is frequently ignored. Human judgement is required. Findings should be curated, not just counted.
  
• Fifth, to sensibly apply an intervention in a new context, we need to know more than whether it has “worked” before. We also need to know why it worked. Otherwise we cannot know what will happen when the context changes. Effective science requires robust theory.
  
• And finally: psychology is not the only show in town. Other mechanisms for shaping public behaviour also exist, such as taxation, criminal law, and technology. Whether you find these methods palatable will depend on your politics. But that shouldn’t come into it. When asked to compute the efficacy of an intervention, it is unethical to prioritise your political preferences.

Such caveats hamper our confidence when recommending behavioural interventions. But even if our evidence were perfect, a bigger problem would still remain – one that is philosophical rather than psychological, but all the more profound as a consequence. This is the problem of cost-calibration.

How exactly do we measure the negativity of negative impact?

With Covid-19, the impetus for preventive action has been the likelihood of lethality. “Flattening the curve” is a euphemism for “reducing the death rate.” The calculus of lockdown requires that we put a price on life.

For many people, the loss of even one life will be unacceptable, and everything must be done to avoid it. But for others, human lives are less than priceless; deaths are tolerable if the overall negative impact is offset by a greater gain. (The ethical problem starts when you realise that this “gain” will be felt by people other than those who die.)

The British NHS are directed to fund non-palliative treatment for a seriously ill patient only if said treatment costs no more than £30,000 for each year of “quality” life provided. For a terminally ill patient, this cap is typically lowered to £15,000 per “quality-adjusted life year.” In effect, the NHS has a detailed algorithm that defines the value of each life in financial terms.

It is not for psychology to determine the price, in euros, of a human life. But behavioural scientists should be able to describe how ethics affect judgement. By this I mean more than just such anodyne social-cognitive clichés as “subjective norms” or “perceived behavioural control.” I mean those real-life traits – those prejudices – that influence people’s motivation to protect others: attributes like misanthropy, rejectionism, subversion, selfishness, racism, self-privilege, and apathy. Assholes, after all, are people too.

Most intervention studies grapple with adherence on the assumption that when participants lapse, they do so because of misinformation or poor attention. However, cultural prejudices are also real. People heed and interpret advice in the context of whether it suits their worldview.

Some citizens use a cost-benefit approach, others employ fuzzy logic – but others seek to subvert the mainstream and reject all efforts to nudge them into conformity. They take to the beaches on a sunny bank holiday. They insist on walking the Salthill promenade as they have always done. They claim that Covid-19 is a globalist hoax and go to the High Court to challenge the constitutionality of the lockdown. 

Dare I say it, but even behavioural policy advisors view the world through a political lens. All psychologists and social scientists – including those most avowedly committed to “evidence-based” approaches – have their own value systems.

In the UK, one member of the Scientific Pandemic Influenza group on Behaviour (SPI-B) told the Guardian that “behavioural fatigue” – the idea that citizens would quickly tire of being locked down – “is just not a concept that exists in behavioural science” and so was ignored by their committee. The reality, however, is that there are several large-scale studies on precisely this topic. Even the science on whether people habituate to health prevention measures during pandemics is well advanced, so rejecting the issue as “non-existent” seems a remarkable lapse of expertise. The timing of a lockdown depends on exactly this type of evidence.

Most Irish people have willingly adhered to the cultural changes imposed by government during the Covid-19 crisis. We have had our moments, of course – Cheltenham, Temple Bar, the burning of 5G masts, the turning of blind eyes while Covid-19 ravages Direct Provision centres – and the scale of deaths in our nursing homes has been deeply disturbing. While psychological interventions might be credited for helping in some respects, we need to be able to explain why our “evidence-based” interventions were of such limited effect in others.

Globally, it appears that the countries worst hit by Covid-19 are those most riven by runaway political populism – nations in which embitterment, division, isolationism, and jingoism have become the new political normal. Whenever psychologists seek to advise on behavioural aspects of disease, we would do well to remember that the strongest non-medical predictors of illness are always socioeconomic exclusion and political powerlessness. The poor and the marginalised are always society’s most sick.

We shouldn’t wash our hands of basic ethics, whether or not we sing “Happy Birthday” while doing so.

* * *

The Irish Psychologist is the house magazine for the Psychological Society of Ireland (members-only). Further details about PSI can be found here.