Poll Accuracy

Learning from how the USC/L.A. Times ‘Daybreak Poll’ got it right

In amongst all the post-US-election debate around whether we can trust opinion polls any more, one thing is striking:

The poll that most accurately predicted the outcome (the USC/L.A. Times Daybreak poll) used a different methodology

Recognising that the human brain isn’t digital (and the human heart is even less so), instead of ‘who will you vote for?’, it asked people to indicate their strength of support for each candidate, and their likelihood to vote at all, expressed as a percentage chance

This is exactly the kind of thinking Lucid applies in our insight sessions

We might, for example, ask people to line up on an imaginary scale to show us how cool or warm they feel about a brand, idea or script – and watch what happens. This enables us (literally) to see the strength of feeling, like a spectrum across the room. We also see how quickly (or not) people move to a position, whether they stay firm in their place or waver, and hear what their gut-felt first response is

Combining this with very open, ‘clean’ questioning, we get a richer, fuller, more accurate picture of what people think and feel and thus a more reliable read on how they’re likely to respond in the real world

Maddy Morton


Maddy is Director of Lucid. She has been described as a 'maestra' with groups of people and is appreciated for her passion and commitment as well as the clarity of her strategic thinking, insight and expression. Maddy began her career in advertising at Ogilvy and BBH and then worked at the Arts Council, English National Opera, AEA consulting and Stimulating World Research before setting up Lucid.

Latest posts by Maddy Morton (see all)