Blog/Polling

When it comes to elections, a party’s leader really matters. But what is it about a leader’s character - or the perception of that character - that makes the difference for voters? Strength? Competence? Empathy? 

In the next election, the British public will choose between two leaders who might seem fairly similar, especially after the newly resigned Johnson and Islington North’s Corbyn. Sunak and Starmer are both leaders characterised by an aura of compromise and technocratic competence - no bedraggled hair or homemade jam in sight. But the similarity between the two is overplayed. What do people see in Sunak and Starmer, and what are they looking for in their next prime minister?

Leader ratings are often measured by standard questions, asking about how ‘favourable’ people feel towards a leader, or how ‘satisfied’ they are with how that leader is doing their job. We’ve used these questions before, writing about Labour’s electoral chances back in February. But this time we ran questions covering a large range of attributes and statements, to get beneath the surface and look at how the various components of leader image interact. 

It turns out that old and young, and progressive and conservative, are not looking for the same things in their leader. Meanwhile, people see Sunak and Starmer very differently, and the attributes that make the difference with voters are not what you might expect.

What do you want in a leader? It depends who you are

Looking at Britain as a whole, what people want in their next prime minister is pretty clear, if slightly unexpected. When we asked “What in your opinion is the single most important characteristic for the next prime minister?” the answer was clear:

1. Honesty,

2. Putting the public interest over their party.

Other options, such as ‘get things done’ (remember the 2019 election?) were much less widely preferred, and ‘standing up to foreign powers’ was bottom of the list. What the British people want most is not competency and delivery - often stalwarts of leader rating discussions - but honesty and integrity.

But that simple picture hides a world of difference. Below, we’ve made a simple ‘values map’, plotting views about immigration along the bottom (as a proxy for liberal-conservative values) and views on economic redistribution up the left (as a proxy for left-right economic values). Then we’ve projected onto this ‘values map’ the average position of respondents who picked each of the options above (size indicates how many picked the option).

Empathy-focused values - fairness and understanding - are strongly associated with progressive economic views, as well as (less strongly) putting the public interest first. Standing up to foreign powers is strongly favoured by social conservatives, as is delivery (‘Get things done’). These two are also noticeably associated with conservative economics. 

There is a progressive-conservative slant to what people are looking for in their leader, aligned along the empathy-delivery axis. But other attributes - honesty and vision, are values-agnostic: they’re considered important by people on both sides (shown by their position in the middle of the plot). 

Vote intention shines a different light on the same pattern. Everyone thinks that honesty is important; many people think that public interest is important - but especially those not currently intending to vote Conservative. Getting things done, along with new solutions, are Conservative voter priorities; understanding and fairness for Labour. 

It’s not all about progressive vs. conservative politics either. Age is very important, especially when it comes to honesty. For all the coverage politicians’ honesty has had, it is disproportionately more important for older respondents: twice as important for over 65s as 18-24s. 

For young people, meanwhile, honesty is no more important than the empathy focused attributes of understanding and fairness. These two seem to form part of a progressive world view: a caring and support-focused society needs an even-handed and understanding leader. These leader attributes are also strongly associated with housing tenure (which is another key factor behind progressive political views). Renters are twice as likely to rank fairness and understanding highly than those who own their home. 

Sunak and Starmer - really quite different

Moving on from what people would like in the abstract - what do people see in specific leaders? Surveying this can be tricky, since what people say about a leader aligns with whether they support them or not. Leaders you like are strong and understanding, leaders you don’t are dishonest and weak. 

To get past this problem, we asked a forced-choice question, presenting people with a list of adjectives and asking people to pick their top 3 for a leader. These were all designed to be positive, to make people look for what they like in a leader even if they did not support them. What this gives us is a rounded view of what aspects of a leader’s character each leader over-indexes on. 

The results showed two adjectives come out top: ‘intelligent’ and ‘persistent’. This is because those were the options most picked by people who did not support a leader - a form of grudging admiration. People viewed Sunak and Starmer as ‘intelligent’, and Johnson as ‘persistent’. The way to interpret this is that these are not truly seen as strengths, but as a way of saying that ‘I don’t support them, but at least they are X’. 

These attributes really show how the leaders contrast with each other. Boris does badly on most attributes, especially honesty. As he resigns from parliament, there is no pattern here to suggest he is still the election winner of old. He does however top out on ‘effective’ and (almost) ‘brave’, albeit with a low percentage.

But what about Starmer and Sunak? Johnson vs Corbyn was a colourful matchup of two very different personalities. Sunak vs Starmer can often seem rather, well, bland by comparison, with two rather technocratic, vanilla, and ultimately similar candidates, high on competence and low on risk. Our survey, though, shows that perceptions of the two have important points of difference.

Starmer stands out from the other leaders on more egalitarian and empathetic attributes such as ‘fair’ and ‘understanding’, as well as the nationally popular value of ‘honesty’. Meanwhile Sunak stands out for being ‘intelligent’ - and more importantly ‘level headed’, but lower performance on attributes such as strength or effectiveness. 

We can get more detail when we compare how respondents describe each leader’s character in their own words. We took the frequency that respondents used each word, and examined which words were used disproportionately to describe one of the two (above the line means more for Sunak, below for Starmer). 

Starmer is more often described as ‘honest’, ‘trustworthy’, ‘confident’, ‘reliable’, and ‘fair’ as well as (before Starmer fans get too happy) ‘smarmy’ and ‘boring’. Sunak outperforms on ‘intelligent’, ‘smart’, ‘hard working’, and ‘friendly’, but also ‘dishonest’ and  ‘untrustworthy’. 

Head to Head

What does this mean for their overall ratings? Using the standard Ipsos wording, “Are you satisfied or dissatisfied with the way X is doing his/her job as Y?” Starmer is 15 points ahead of Sunak. Starmer has 35% satisfied / 34% dissatisfied for a +1% net positive rating (down from +6% in our February poll), while Sunak is 30% satisfied / 46% dissatisfied for a -16% net rating (in line with -15% in our February poll). The results are thus very much in line with last time: Starmer has slipped slightly, but maintains a small but solid lead over Sunak. 

Feeding leadership ratings and voting intention into our historical model (as featured back in Bi Focal #5), this gives us a 2% Labour lead at national level. This is a long way off Labour’s 15-point lead in headline voting intention (Lab 45 / Con 30 in our poll). Our conclusion before - that either Sunak’s ratings will worsen, or the Government will do better than expected - still stands.

Our results are somewhat more positive for Sunak, and especially Starmer, than the latest Ipsos political monitor ratings. However, this 15-point lead at the overall level hides a deeper pattern: Starmer is much more popular among the voters who matter most. Perhaps the most telling lens is occupation.

Sunak’s  relatively high ratings among managers (middle and senior) and retirees is hiding some relatively much poorer ratings among routine and semi-routine occupations, as well as modern professional occupations. By contrast, Starmer does well, even managing a slightly net positive rating among those groups. While Starmer’s ratings aren’t exceptional anywhere, Sunak seems to turn off those routine and professional occupational groups, giving Starmer a big lead in those groups.

When we turn to age we see that Starmer leads Sunak among every age group below 65. While Starmer’s approval roughly follows a conventional young (Labour) to old (Conservative) gradient, Sunak is actually more popular among those under 35 than those aged 35 to 65 (which echoes research on Millennials we carried out recently with Onward). That means that Starmer is leading Sunak even among those in later middle age. 

It’s not surprising that Starmer is leading among modern professional occupations, who these days are Labour’s core vote. But his lead among routine and technical occupations is much more electorally significant. These are the voters in social grade C2 and D who Labour have had trouble holding on to, in 2019 especially. The same goes for Sunak’s low ratings among voters in later middle age: these are voters a Conservative leader would normally expect to lead from their Labour counterpart. Starmer’s overall rating might not be sky-high, but he is popular among exactly the voters who matter most

When it comes to voting, what makes the difference?

We haven’t yet addressed what is perhaps the most crucial question of all. When it comes to attracting people’s votes, and convincing them to switch sides, what particular aspects of a leader’s image are most crucial? What makes the difference?

To look at this, we went behind respondents’ explicit answers, and looked at the latent patterns that link their leader preferences to their vote intention. To do this, we first asked: what do people intending to vote (e.g) Conservative think about (e.g) Sunak? We then compared these people to those who are not intending to vote Conservative, to see what opinions of Sunak are driving people to vote Conservative.

But wait, you say, what about other factors? Many things drive people to vote Conservative: demographic background (e.g old), Conservative social and economic values, or their opinion that certain issues (e.g immigration) are particularly important. In particular, people who voted Conservative at the last election are of course more likely to vote similarly than those who did not vote Conservative. To find out what aspects of Sunak and Starmer’s character are making the difference, you need to take all of these other factors into consideration. 

That’s what we did, with a simple logistic regression model. The model was fitted with a regularising prior, and has pretty good explanatory value (kappa, or accuracy above the baseline, was 0.65-0.7 for the four models shown here). The sample size for these models is 500, large enough for 95% confidence in the effects discussed - confidence intervals are shown below. Here are the results (a full list of control variables can be found at the end of this blog). 

There are a few characteristics that make the difference for Sunak. Some are in line with national priorities we saw earlier, such as honest and effective. But others are more interesting. ‘In control’ comes out top, even higher than honest, showing that the perception that Sunak is in control of his party is very important for his support. Being ‘effective’ is also important (remember how Conservative voters rate ‘get things done’ highly). This reveals the values and belief space occupied by those on the right and centre right, especially as regards authority and effectiveness - something that echoes Jonathan Haidt’s work on Moral Foundations. 

We also ran the same model with a set of statements with which respondents were invited to agree or disagree. “Someone you can trust to do right by you” highlights the perennial importance of trust in politics. A more interesting second place is, ‘would do the right thing even if it was unpopular’ - something we remember the Cameron coalition government polling particularly well on. 

Turning to Starmer (and which of his characteristics drive the Labour vote). ‘In control’ is also important for him, but in this case less so than understanding, and fair, as well as the ubiquitous ‘honest’. ‘Brave’ is also important - someone prepared to confront the hard choices and stick his neck out on occasion.

When it comes to the statements, being 'someone you can trust to do right by you' is as important for Starmer as it is for Sunak. But the other standout points are 'stands up for people like me', and ‘is someone who has a clear vision of the future’. There’s been a fair amount of talk recently about the extent to which Starmer is communicating a distinct ‘vision’ (or not). It seems that convincing people he has one really does make a difference in motivating the British public to support him.

Despite these two leaders appearing very similar at face value, the public perception is, in fact, very different with each having a unique feel - something that could be exacerbated come election time.

Next Time

In next month’s issue, we briefly swerve politics and dive into the world of business - as Focaldata speaks at Nudgestock, the world’s foremost festival of behavioural science and creativity. 

Enjoyed this week's newsletter but haven't subscribed yet? Why not subscribe and get the next one straight to your inbox. You can also follow Focaldata on Twitter and LinkedIn.

You can find tables for the full poll results here.

Model control variables:

Outcome: vote intention is either Labour (Starmer models) or Conservative (Sunak models).

• 2019 vote
• Age
• Gender
• Region
• Education
• Ethnicity
• Tenure
• Occupation
• Top 3 issues
• Immigration values
• Redistribution values