New Zealand / Politics

Hipkins vs Luxon: The first leaders’ debate by the numbers

08:09 am on 20 September 2023

Photo: RNZ

Christopher Luxon had the most to say in the TVNZ leaders' debate - but does that make him the winner?

Labour Party leader Chris Hipkins and National Party's Christopher Luxon went head-to-head on Tuesday evening in the first leaders' debate of the campaign.

The two-hour long debate broadcast on TVNZ and hosted by political editor Jessica Mutch-McKay covered a wide range of topics including health, co-governance, gangs and housing.

Some commentators declared the debate a tie, although former MP Tau Henare called it boring and mild and 1News's deputy political editor Maiki Sherman said Hipkins could have brought "a lot more fight, a lot more mongrel" to the debate. Former Labour leader David Cunliffe was also critical of Hipkins' performance, particularly in the first half of the debate, where he felt Hipkins allowed Luxon too much of a free run.

An RNZ data analysis of what was said shows who got the most words in, what topics were the hottest, and what the leaders said when they didn't want to give a straight answer.

Who talked the most?

When it came to getting more words in, Luxon won out. He said 6196 discernible words during the debate compared to Hipkins' 4491 words.

This put Luxon at speaking 43 percent of all the words during the evening.

Luxon's 43 percent is a higher percentage than his predecessor Judith Collins got in any of the three 2020 debates (33%, 35% and 38%). On the other side, Hipkins' 31 per cent share was lower than his predecessor, Jacinda Ardern, in all but one of the 2020 debates (34%, 28%, 37%).

The moderator, Mutch-McKay, had 25 percent of the words spoken. This was fairly typical of previous debates, and well short of the 37% Newshub's Patrick Gower took in one of the 2020 clashes.

Luxon's share aligns with the view of some analysts after the debate, who suggested he was largely untroubled by Hipkins. It also could reflect a common critique of him - that he speaks too fast.

More on the election campaign:

  • National banks 7.5 times more in donations than Labour
  • Latest polling, donations, Facebook advertising and more
  • Guide to party policy
  • What was said

    There were words the leaders said over and over again. Luxon said "actually" 69 times during the debate. Hipkins's top word was "think", followed by "health".

    Terms which have been frequently used by the leaders in the past also made a showing. Hipkins referred to "bread and butter" once. Luxon used "coalition of chaos", "squeezed middle" and "gone backwards" twice each.

    Luxon also referred to "six years" at least six times. "But Chris, you've had six years," he said in relation to crime. Hipkins referred twice to services being "run down" when Labour took power.

    RNZ counted how many times each leader used certain words which relate to key topics. The most mentioned topic was health, which was a topic covered extensively in the debate. Māori ranked highly, as did tax.

    Dodging the hard questions

    A common tactic to avoid answering a question is to use what's referred to as a 'block and bridge'. The politician acknowledges the question using a 'block' and then uses a 'bridge' phrase to move the topic to something they are more comfortable with. Common bridging terms are "I would say", "what I can tell you", "the reality is". The bridge allows them to transition the conversation to their own talking points.

    Luxon used variations of the term "what I say" eight times during the debate. His next most used term was "the reality is", which was also used by Hipkins.

    What was the vibe?

    According to a language sentiment data dictionary, which rates words as either positive, neutral or negative, Luxon used a slightly higher percentage of positive words and lower percentage of negative words than Hipkins during the debate.

    How we analysed the data: The debate was transcribed by Otter.ai and then manually cleaned to improve accuracy. Cross talk - where debate participants spoke over each other was removed where words were unintelligible. Interjections when intelligible were included. Text was analysed using R and the Quanteda package.