Chris Hipkins’ leaders debate strategy: Let Luxon talk himself into trouble

Labour's leader has been accused of running a 'small target' campaign, giving doubters as little as possible to dislike. His debate strategy seems to be fitting that mould.For the second head-to-head debate in a row between the two leading candidates for Prime Minister, Chris Hipkins spoke far fewer words than Christopher Luxon.His strategy seems to be to use the debate stage to contrast himself with Luxon, who himself has admitted to a tendency to veer into corporate-speak at times.In last night's NZ Herald Leaders' Debate, Hipkins' strategy was perhaps clearest in a section where moderator Ryan Bridge asked the two leaders to 'rule in or rule out' five different scenarios each.Luxon dealt with the first one - scrapping the Climate Commission - with one word: 'no'. But his answers got longer after that, chewing through 112 words across the remaining four scenarios. Hipkins saw his chance, giving three or four word answers - "off the table", "on the table" or "I'm open to that" - to answer the first four. In 25 words, he'd put a lower voting age, a rainbow affairs ministry, and compulsory te reo on the table while also ruling out making Waitangi Tribunal recommendations binding, or hiking corporate tax.RNZ's data analysis picked up a 'less is more' signal from Hipkins in other areas too.Who talked the most?The result was the same as last week's TVNZ leaders debate, but this time it was even more pronounced. Luxon increased his tally to a whopping 6255 discernable words, up from the 6031 he got out last week. The gap to Hipkins got even wider - he said 3958 words, down from 4270 last week.Moderator Ryan Bridge was more talkative than TVNZ's Jack Tame so Luxon's proportion of airtime actually fell a little to 37 percent of all the words spoken during the evening, slightly down from 39 percent. Hipkins was down to 24 percent from 27 percent last week.Luxon also took more speaking turns than Hipkins. There were 150 instances of six or more words spoken by Luxon compared to 118 by Hipkins.The longest answer of the night was Luxon's on education and mandated time for reading, writing and mathematics. At 218 words it was far longer than Hipkins' longest answer of 127 words about pay equity.Luxon's second longest answer was about Labour's proposed capital gains tax and Hipkins' was related to renewable energy.On average, Luxon said 33 words each time he spoke, compared to Hipkins' 26.Hipkins did again interrupt his rival far more frequently than Luxon interrupted him. But here again, both leaders were notably more restrained than in the TVNZ debate.Our analysis logged an interruption as when one leader cut in and the other stopped to respond. An attempted interruption, or cross-talk, was counted when one leader spoke over the other, but the speaker ignored him and carried on.Hipkins landed eight interruptions and had a further 15 attempts ignored by Luxon - that happened 30 times in the last debate.In one response, Luxon hit back against Hipkins' accusation he tolerates racism with: "You just tolerated your mates going off to a mongrel mob hang out." Three other times he asked Hipkins to let him finish without interrupting.Luxon interrupted Hipkins only once when discussing school uniforms, and talked over him three times. This was also well down on last week when he chimed in over Hipkins 18 times.What was saidAt the very start of the debate, Luxon conceded: "You don't always express yourself clearly or well enough, and sometimes you can do better. I'm a case in point of that. Over the last three years, I think I've illustrated that well."This has been his response to repeated criticisms that he has a weakness for corporatese, or occasionally getting tangled up in a word salad.Over all the main debates Luxon has participated in, the word "actually" has been the one he said the most. The term can be used in a number of ways, including as a way to signal contrast or correction, push back on what somebody else has said, or to pivot to a different point. Sometimes it's used as a filler word that gives the speaker a moment to gather their thoughts mid sentence.Luxon used "actually" 64 times in last night's debate, sometimes multiple in a sentence: "Actually, I think it's actually unfair to the deputy principal in my electorate, as I said the other day, that had a $30,000 inheritance from her mother and thought she was doing the right thing by actually having one rental property."Hipkins' top word for the second debate in a row was "going".For the second debate in a row Luxon used variations of "challenging". He also used terms some consider to the corporate-speak including "going forward", "reaching out" and "outcomes".Hipkins used "not true", once in response to a question from Ryan Bridge about truthfulness and three times to protest points Luxon had raised.The leaders addressed each other by their first names with Luxon doing it far more than Hipkins.Dodging the hard questionsA 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 "the question is" and variations of the term "as I said" twice during the debate. Hipkins did not use any common bridging terms.What was the vibe?According to a language sentiment data dictionary, which rates words as either positive, neutral or negative, Luxon used a higher percentage of positive words and slightly lower percentage of negative words than Hipkins during the debate.How we analysed the data: The debate was transcribed by Otter 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.