OPINION12 April 2017

Hearts and minds

Behavioural science Brexit Opinion Public Sector UK

What people say can be at odds with what they feel, even when it comes to voting, says ICM’s Martin Boon, who believes this insight could be at the heart of a new way of conducting polling.

Heart and brain crop

You may have noticed that it’s hardly been the easiest of times for opinion pollsters lately. Not content with miserably failing to predict the right outcome at the 2015 General Election, final polls in the EU referendum – with one honourable exception – predicted a Remain triumph, and perhaps the ascendancy of President Trump has sealed polling’s fate among research professionals, pundits and, indeed, the general public.

Some of the critique might be a bit harsh – as many campaign polls predicted a Leave victory as did one for Remain, and actually Presidential polling was more accurate in 2016 than it was in 2012.

But no matter, we all know that the perception is more important than the reality so let’s just accept there’s plenty more work to be done reinventing the polling wheel. Indeed, some of us have already spent many more hours than is healthy trying to re-build a methodological process to accurately predict the state of the political parties at any given point in time.

Some good work has been done, I think, both here and at other polling houses, but we’ll see. But during the course of those long hours it occurred that the act of voting is not rationality encapsulated, but an intrinsically and intensely emotional act – sometimes hinging on lingering party loyalty, on other occasions an act of fury against a flailing government.

So why not assess the emotional certainty between an individual party choice and its impact on headline vote intention numbers? What if ‘intend-ers’ or ‘serious consider-ers’ are actually betrayed by an emotional obstacle that forces their retreat to party safe havens, or more to the point, what if social desirability bias (incorrectly also known as the ‘Shy Tory’ phenomenon) prevents people from admitting who they intend to vote for, pushing them into the unhelpful “I don’t know” or “refusal” boxes?

The advance of neuroscience and associated techniques has been massively helpful to this investigation. The application of Implicit Attitude Testing to the level of seriousness associated with specific party intentions has yielded important outcomes.

For example, we all believed that UKIP was the Marmite party; loved and loathed in equal measure, and intensively on both sides. Labour now holds that precious title under its precarious leader. If the emotional connection with the party really is severed as a result, as it has been in Scotland, its recent decline may only be the start of a perilous journey.

Modelling in emotional certainty is undeniably a problematic and, to some, dangerous evolution, moving stated voting preferences further and further away from their raw component. ICM has some history in this area, and in our case post-data intervention has had, at worst, a neutral impact on headline data, and at best, made some of our final predictions exceptionally accurate.

It’s worth more work, in my view, and we’re looking closely at it. But it’s not just about voting intentions, it’s just as much about policy, and about effectively communicating it to the largely unengaged masses. Perhaps even more than with vote intentions, ICM’s tests have shed previously unforeseen light on the power of public opinion to be potentially misleading, possibly manifestly, and to great detriment to campaigning causes.

Using our new policy and messaging evaluation tool, the Policy Emotional Certainty Score (PECS), we find that emotional reinforcement in a new policy idea or campaign message is often evident, but in other cases there is a real disconnect between explicitly stated support and emotional connectivity.

The EU referendum is a great case in point – arguments about the perceived failure of ‘Project Fear’ have largely been impressionistic and assumptive, but our tests during the referendum campaign revealed that ‘Taking Back Control’ had twice the emotional power of ‘Project Fear’.

I didn’t know it at the time, but this was the single clue I had that Leave really were going to win (although to be fair, there was at least as much suggesting the alternative outcome). What we’re seeing here, at last, is not just artificial wordsmithing, but evidence that what people explicitly say can indeed sometimes be at odds with what they emotionally feel.

For now our concentration is on message and policy development, with the recent Budget being a great opportunity to cross-reference our new technique against conventional polling. For example, headline polling figures might make you think that the Tories should have stuck to their guns on National Insurance Contributions (NICs) for the self-employed. A third of the public supported the NIC measure to raise much needed funds (the same number against), so why not ride out the storm and collect the cash?

Few read party election manifestos these days, except political hacks, and breaking promises is what politicians do. The agenda would soon enough turn back to Labour’s performance. Wrong. Breaking a pledge is seen as a breach of trust, and measuring people’s emotional certainty on this matter provides compelling evidence for politicians never doing so.

That third of the public who didn’t have much of a problem with it? Actually, when accounting for the emotional component the PECS Score indicated that only 5 (Indexed out of 100 ) could be counted on to be unconcerned by pledge-breaking. But there’s much more here. The social care crisis has touched a national nerve, and in confirmation the explicit score ICM found in support of raising a new £1billion for it ( 79%) was matched by the policy emotional commitment score ( 80 ).

The Sugar Tax, too, is as much an injection of energy to the emotional coffers as a post-lunch chocolate bar is to a tired child: 59% support it explicitly but this rises to 77 when we factor in the emotional component. And you know that age-old ropey premise about people being willing to pay extra in tax if the proceeds were ring-fenced for the NHS (but then voting for someone saying the opposite)? Guess what: 42% still say they would support this, but the PECS score plummets to only 15. Look forward to seeing that one on Jeremy Corbyn’s manifesto suicide note.

It gets even worse though for supporters of the Left: if you suggest that income tax would be the funding source and then put a figure of what it’ll cost you next to it, expect to see manifesto pledge-breaking levels of emotional emptiness.

You know what? Journalists, the public and even you dear reader, might not believe pollsters anymore, but all of a sudden we might not be so sure about you either.

Martin Boon is a director at ICM Unlimited