CN: mention of Covid-19.

What do ‘clap for carers’ and politicians strolling around with masks on have in common?

Both are examples of ‘nudges’: small behavioural signals, which when deployed correctly, can radically alter the behaviour of millions of people.

Since the pandemic began, governments faced the onerous and unenviable task of doing exactly this. In India, a nation of 1.3 billion people, with more than 6 major religions, 28 states, and over 700 dialects, this would not be an easy task.

But Ramit Debnath and his colleagues were determined to find out how India went about this challenge in the early stages of the pandemic. And one of their answers: nudge theory.

A Cambridge-Gates Scholar, Ramit grew up and completed his undergraduate in Mumbai, where he witnessed first hand the extreme poverty in the city’s slums. It was partly this which led him to fuse his background as an electrical engineer specialising in urban design, with a desire to impact public policy - or as he puts it, become a “social justice engineer”.

Historically, India has often made use of what behavioural psychologists call ‘nudge theory’ to modify the behaviour of their entire population at once.

“A nudge is basically a positive reinforcement in behaviour,” Ramit tells me. “The whole idea is that I don’t tell you forcefully to change your behaviour. But I’ll give you certain clues in a more positive way so that you realign yourself towards that.”

If successful, this may trigger a chain reaction, accelerated by peer pressure. “So for example, if I see that six people around me are wearing masks, the higher the chances I will be wearing a mask and maintain the rules of COVID lockdown”.

Ramit suspected the government may be integrating nudges - perhaps unwittingly - into its planning during the first phase of the COVID-19 crisis.

To test this, he employed an AI to conduct ‘topic modelling’, a type of computational analysis. Giving it a desired end-goal, and some basic parameters to work with, the AI groups terms by similarity and how frequently they appear across tests, learning to improve its ability to cluster as it goes (hence ‘machine learning’).

“So, for example, I have a word like ‘health’ in my data corpus, and it is repeated 10,000 times, because it’s a COVID situation. So the algorithm then picks up that word and assigns a higher probability value, because the chances are that you will find that word more.”

He asked it to find similar words from 396 official press releases published by 40 government agencies from January to April, and analyse if they reflected a ‘nudge’ planning strategy. It found exactly this: the more frequently occurring words matched typical ‘nudge’ terms.

“Our Prime Minister, [Narendra] Modi, is quite populist, so his frequent appearance in TV was a very strong nudge element”

“So one thing that was quite very obvious was the feeling of nationalism. The government played a lot on that,” Ramit explains. “Our Prime Minister, [Narendra] Modi, is quite populist, so his frequent appearance in TV was a very strong nudge element, because he always appeared in short clips, asking people to maintain social distancing and wear masks.”

One day before India went into lockdown on the 25th March, for example, Modi declared in a televised speech that “21 days is critical to breaking the infection cycle … or else the country and your family could be set back 21 years”.

“They tried to even convince people through [promoting and showing] nostalgic TV shows so more people stay inside,” Ramit adds.

The results showed the government also cunningly tailored nudges to target specific groups. It encouraged small businesses to produce PPE and donate to charities, and the wellness sector too, which began to adopt labels like ‘#YogaAtHome’ to push its followers to comply with restrictions.

Ramit is quick to point out the research wasn’t meant to be a “confirmatory analysis” evaluating the success of such a strategy, even as some Indian TV stations took it to be a validation of Modi’s politics.

Despite that he cautiously admits “especially when you have so much divide ... people accepting masks and lockdown norms and staying indoors for almost three months shows a great degree of success”.

But nudge theory planning has serious shortfalls too, partly because it breeds ‘hard fatigue’, and requires officials to constantly trial new ‘nudges’. “It fades away over time, so if you don’t do some kind of experimentation with how you nudge it becomes useless after some point in time,” says Ramit.

“They seemed very strong strategies to contain COVID, but what they didn’t tackle is … what will happen to minorities”

“If you want to have a longer effect [of behavioural change], then go with data driven and empirical results and policies,” he stresses.

The particular nudges, especially those rooted in nationalist rhetoric, also do not work for all. “They seemed very strong strategies to contain COVID, but what they didn’t tackle is … what will happen to the minorities,” he argues.

Modi’s centralised approach may have played to Indian nationalists but also meant the four hours notice he gave for the national lockdown left thousands of low-income workers stranded hundreds of kilometres from their home villages.

But Ramit is pleased that his paper may have very practical uses in future crisis planning, “so they can have more informed decision-making [based on] the steps taken by India”. He has already seen the study spread across Sub-Saharan Africa, Southeast Asia and the Middle East on social media.

The UK government too has made use of nudge theory both before and during the coronavirus response, and even has a ’Nudge Unit’, which is home to resident psychologists who advise the authorities. Ramit argues clapping for carers is an archetypal example of nudge theory, which was brought to life here too.

As a second national lockdown seems to loom ever closer, Ramit’s work could be more relevant now than ever. But one thing is clear: combatting the growing angst at the government’s handling of the crisis and a rising number of anti-COVID truthers will need one hell of a ‘nudge’.

You can read Ramit Debnath and Dr Ronita Bardhan’s full paper on PLOS One.