Words carry immense power. Emphasis, tone, and vernacular all influence how we interpret a message. Simple phrases, often meant to be forgotten, can become embedded in our language forever. Politicians, more than most, understand the weight of words. While norms have been challenged over the last decade, entire careers can still be derailed by a single misstep in phrasing.

Personally, I’ve always been captivated by natural language processing (NLP) for this very reason. The idea that we can extract deeper meaning from our words has fascinated me to the point where I dedicated both my master’s milestones and capstone to exploring it. Through this journey, I’ve learned that there are both complex methods and simple tools that allow us to better understand language—and by extension, the person speaking.

In this article, we’ll take a deep dive into the Harris/Trump debate, using a few R-based NLP packages to gain insight into the candidates’ speech patterns and messaging.

Who’s Really Connecting with the Audience?

When Trump and Harris took the stage, the differences in their communication went beyond just policy and leadership. According to Flesch-Kincaid and SMOG scores, Trump’s speech was at a 4th-grade reading level, while Harris’s required a high school education to fully grasp. But does simpler language create a stronger connection with the audience, or does complexity give Harris the advantage?

The Numbers Behind the Speeches:

When we break down the speeches using readability scores, we see two very different approaches:

• Trump’s Flesch-Kincaid score is 4.50, meaning his speech can be understood by someone with a 4th or 5th-grade education. However, his SMOG score is 8.12, indicating that a middle school education is needed for full comprehension. These two scores highlight different aspects of his communication style: the Flesch-Kincaid score focuses more on sentence length and structure, showing that Trump uses shorter, simpler sentences. The SMOG score, which emphasizes word difficulty (specifically longer or more complex words), suggests that while Trump’s sentences are simple, he does use some words that require more education to fully grasp. Overall, these scores reflect his direct and accessible language, designed to reach a broad audience.
  
• Harris’s Flesch-Kincaid score of 8.36 suggests her speech is more complex, needing an 8th-grade reading level to understand. Meanwhile, her SMOG score of 11.47 shows that high school-level education is required to fully comprehend her language. The difference between the scores can be explained by her use of longer, more intricate sentence structures (as indicated by the Flesch-Kincaid score) and more sophisticated vocabulary (as captured by the SMOG score). This combination points to a more formal and detailed speaking style, which may resonate with a more educated audience.

What Do These Scores Mean?

• Trump’s Strategy: His lower scores indicate a speech understood by many. By keeping his language simple and direct (if we objectively assume it’s intentional), Trump ensures his message can resonate with people from various educational backgrounds. This simplicity might also contribute to clearer, more immediate understanding, which could be crucial for reaching a broad and diverse audience.
• Harris’s Strategy: Harris’s higher scores suggest a more nuanced approach. The complexity in her language could reflect the importance of policy details and leadership depth, appealing to those who are more likely to appreciate a detailed, intellectual discussion. Her use of complex vocabulary and sentence structures positions her speech as more formal and sophisticated.

Here’s the twist: While Trump’s language is simpler, that doesn’t necessarily make it less effective. His approach might foster a sense of clarity and accessibility, allowing his audience to feel more directly connected to his message. Meanwhile, Harris’s more advanced language could signal intelligence and authority, potentially building credibility with a more educated audience. So, does simplicity win out, or does complexity carry the day? As we dig deeper into the structure and content of their speeches, we’ll uncover how each candidate tailored their communication style—and how it might have shaped the perception of their leadership. Stay with us to see what this tells us about their broader strategies—and which approach ultimately connects with the crowd.

Trump Spoke a Lot!

During the debate, Trump spoke for approximately 42 minutes and 52 seconds, while Harris spoke for 37 minutes and 36 seconds (1). The time disparity alone raises questions about the moderators’ control, especially considering Trump had previously accused them of favoring Harris (2). Despite his claims, he actually had more speaking time. Not only that, but he said significantly more as well. When we look at the total word count, Trump spoke 8,118 words, while Harris spoke 5,950. This is surprising since Trump has been criticized for being one of the slowest (recent) US presidents (3).

As a general rule, a 5-minute speech contains roughly 750 words, or about 150 words per minute. Based on this, Trump spoke at approximately 188 words per minute, while Harris spoke at around 165 words per minute. Although neither pace is significantly faster than the average, it’s noteworthy because it suggests both candidates delivered their messages with greater intensity and information density. This difference might reflect contrasting speaking styles—Trump may have used shorter, more direct sentences, while Harris could have taken a more measured approach. These subtle differences in speaking pace, while not drastic, might have influenced how the audience perceived and retained their messages, ultimately shaping how their arguments were received.

What was significant, though, was the total words spoken. I used a chi-square test to determine if the 2,168-word gap was statistically meaningful. Chi-square was chosen because it’s ideal when comparing counts or totals between two or more categories, which is exactly what we’re doing here with total word count. In contrast, a traditional two-tailed test (like a t-test) works better for comparing averages or continuous data. The chi-square test focuses on frequencies and categories, making it perfect for this scenario.

The result was striking. The test statistic (334.11) revealed a huge variance from the expected distribution, and with an incredibly small p-value of 2.2e-16, we can confidently reject the notion that the word counts are statistically equal. In simple terms: the difference in the number of words spoken by Trump and Harris is far from random—it’s significant. This isn’t just about how long they spoke; it’s about the sheer volume of what was said.

Comparing unique word counts gives us a different result. Despite Harris having a slightly higher unique word count (1,257 compared to Trump’s 1,221), the chi-square test shows that this difference could easily be due to random chance, rather than a meaningful variation. Both candidates used a similar range of unique words, indicating that their speeches, in terms of linguistic diversity, were quite comparable.

This suggests that both candidates structured their responses similarly when it comes to complexity and vocabulary. They used a comparable number of distinct words, focusing more on the content and themes of their messages rather than diverging in linguistic style. So, while the volume of words spoken was vastly different, the way they used their vocabulary was strikingly similar.

They Said What?

To compare the content of the words spoken, I first needed to clean up the responses by removing stop words. Stop words are common words like “the,” “and,” “is,” or “but” that appear frequently but don’t carry much meaning. In text analysis, we remove them because they can clutter the data and distract from the more meaningful words. By filtering them out, we can focus on the key terms that actually provide insight into the content. It’s like cutting through the noise to better understand the core message.

Trump’s Top Words:

• People and country are the most frequent, a focus on populist themes, likely appealing to a broad audience by emphasizing national unity and citizens.
• President, world, vote, and plan suggest Trump is speaking about leadership, international topics, and action plans, addressing his achievements or future goals.
• Words like war, millions, border, and billions imply a focus on national security, immigration, and economic matters.

This shows Trump’s speech focused on grand themes of leadership, the people, and international/national concerns, with an emphasis on scale (millions, billions).

Harris’ Top Words:

• President and people are also prominent, indicating her focus on leadership and addressing the populace, which is typical in political discourse.
• The inclusion of Trump and Donald indicates a significant part of her speech was dedicated to discussing the former president.
• Words like plan, understand, care, and time show Harris is focused on empathy, detailed planning, and social issues, such as healthcare or governance.
• Words like American, united, and country suggest a focus on national identity and unity, but more subtly compared to Trump.

Harris’ top words indicate a focus on leadership, social issues, and perhaps a critique or contrast of Trump, while maintaining a tone of unity and care. Both candidates emphasize president, people, and the country, which are key themes in any political speech. However, Trump focuses more on large-scale issues, money, and security, while Harris seems to balance leadership with a focus on empathy, understanding, and social issues. This word frequency highlights how both candidates shaped their messages around leadership but approached their audiences from different thematic perspectives.

Here’s where it gets interesting: While Trump and Harris used a similar range of unique words, their focus was fundamentally different. Trump’s language centered around grandeur and scale—big numbers, big issues—while Harris focused on empathy and care, highlighting personal and social issues. It’s not just what they said, but how they framed their messages. Trump painted with broad strokes, addressing large-scale concerns, while Harris worked on crafting a message of unity and detailed plans. This difference in approach can subtly influence how each candidate is perceived—one as a strong leader tackling big challenges, the other as a compassionate leader focusing on individual concerns.

Sentiment of Spoken Words

When we apply the Bing sentiment lexicon to their responses, we see a stark contrast. Harris’ words are predominantly positive, aligning with our earlier analysis. Her responses appear to be focused on emotional connections with viewers and addressing their personal needs and concerns. Even her negative scores likely reflect moments when she attacked Trump, maintaining a generally hopeful tone.

Trump, on the other hand, shows a nearly equal mix of positive and negative sentiment, with a slight edge toward negativity. This matches his direct and pragmatic speech style. He doesn’t evoke a specific emotional tone like Harris does. Instead, he’s often either on the attack or defense, delivering his points in a straightforward manner.

Trump’s Sentiment:

• Trump used a nearly equal number of positive and negative words, but with a slight tilt toward negative sentiment. The red bar (negative sentiment) is just slightly longer than the blue bar (positive sentiment). This suggests a balanced mix of both positive and negative language in his speech, leaning marginally toward the negative side.

Harris’ Sentiment:

• Harris, in contrast, used significantly more positive words than negative ones. The blue bar (positive sentiment) is much longer than the red bar (negative sentiment), indicating that her speech was more optimistic and focused on positive messaging.

The difference in sentiment between Trump and Harris likely reflects their rhetorical strategies. Trump’s more balanced sentiment, with a slight lean toward negativity, suggests a more critical or combative tone—perhaps focusing on challenges or opponents. On the other hand, Harris seems to have adopted a more positive tone, likely focusing on hope, solutions, or unity.  Trump’s higher proportion of negative words might resonate with individuals concerned about issues and seeking change, while Harris’s more positive tone could appeal to those looking for optimism and constructive dialogue. Given that these speeches likely took place in the context of a debate or campaign, this sentiment analysis could reflect the nature of their messaging. Trump may have focused more on problems or critiques, while Harris emphasized unity, progress, and solutions.

In essence, the sentiment breakdown highlights the clear contrast in their communication styles: Harris with her positive appeal, aiming to inspire, and Trump with his more balanced but slightly negative approach, reinforcing his direct, pragmatic tone.

This plot shows the overall sentiment scores for Trump and Harris based on the AFINN sentiment analysis. AFINN assigns positive and negative values to words based on their emotional tone, and the overall score is the sum of those values across the entire text.

1. Trump’s Sentiment:
   • The negative sentiment score for Trump (represented in red) is significantly lower than zero, indicating that the overall tone of his speech was strongly negative. This suggests that Trump’s rhetoric focused more on criticisms, problems, or negative framing.
2. Harris’ Sentiment:
   • On the other hand, Harris (in blue) has a positive sentiment score, indicating that her speech had a generally positive tone. This suggests she focused more on hope, solutions, or optimism, delivering her message in a way that resonated with positive language.

Looking at Ngrams

When analyzing political speeches, focusing on individual words can miss the bigger picture. That’s where ngrams—specifically bigrams (two-word phrases) and trigrams (three-word phrases)—come in handy. Ngrams reveal patterns in language that show not just what someone is talking about, but how they frame their ideas. In the case of a speech by Donald Trump, bigrams and trigrams can uncover the core themes, messaging strategies, and even the emotions driving his rhetoric. By looking at the most frequently used phrases, we can get a clearer sense of the key talking points and how they’re packaged for impact.

What Trumps Bigrams and Trigrams Say About Him

Key Examples from Trump’s Speech

Trump’s bigrams and trigrams reveal consistent themes, often centered around economic issues, immigration, and national identity.

• “Millions people” and “billions dollars”: These bigrams highlight Trump’s emphasis on scale, particularly regarding economic policies. By frequently referencing “millions” and “billions,” Trump creates a sense of magnitude, whether he’s talking about jobs, money, or people. This helps him frame his statements in terms of large-scale impact, making his policies seem consequential and wide-reaching.
• “People pouring country”: This trigram points directly to Trump’s focus on immigration. The phrase “pouring” carries a negative connotation, suggesting urgency and a sense of crisis, which fits his broader narrative about controlling the borders. This choice of words subtly creates a visual of overwhelming numbers, reinforcing his policy position on immigration.
• “Billions dollars China”: This trigram indicates a focus on economic relations with China, a frequent topic in Trump’s speeches. It likely ties into discussions around trade deficits or tariffs, emphasizing how Trump frames his policies as protective measures for the U.S. economy against external threats.
• “History country” and “president history country”: These ngrams reflect Trump’s concern with his legacy. By using phrases that connect himself and his actions to the broader history of the nation, Trump positions himself as a pivotal figure in the story of America. This strategy helps bolster his image as a leader who’s making decisions with lasting historical significance.

What This Says About Trump

Trump’s ngrams show a speaker who is keenly aware of the power of repetition and scale. His frequent use of large numbers and expansive terms like “millions” and “billions” suggests a desire to convey the magnitude of his impact, whether through economic policies or national security. By focusing heavily on immigration and trade, particularly in phrases that invoke urgency or threat, Trump positions himself as the protector of American interests.

At the same time, his references to “history” and “country” reveal a preoccupation with his place in the nation’s legacy, framing himself as a leader whose actions will be remembered in the context of America’s historical journey. In sum, Trump’s bigrams and trigrams are designed to reinforce a narrative of large-scale change, national pride, and protection against perceived threats, creating a rhetorical landscape that resonates with themes of power, control, and legacy.

Harris’ Most Frequent Bigrams and Trigrams: What They Reveal

Key Bigrams:

• “Donald Trump”: The most frequent bigram in her speech shows just how central Trump is to Harris’s message. By repeatedly mentioning Trump, she’s drawing a sharp contrast between her vision and his leadership. This focus on Trump suggests that much of her speech was dedicated to highlighting the flaws in his administration.
• “American people”: This phrase underscores Harris’s emphasis on inclusivity and populism, directing her speech toward the concerns of everyday citizens. It signals her aim to connect with the electorate, positioning herself as a leader who is in tune with the public’s needs.
• “Affordable care” and “care act”: These bigrams highlight the importance of healthcare in Harris’s speech. By discussing the Affordable Care Act, she’s defending a crucial piece of Democratic legislation and emphasizing healthcare reform as a central pillar of her platform.
• “Middle class”: The mention of the middle class points to Harris’s focus on economic equality and addressing the needs of a key voting demographic. This shows her intent to champion policies that uplift working families and address economic disparities.

Key Trigrams:

• “Affordable care act”: The most frequent trigram in Harris’s speech solidifies healthcare as a major theme. By repeatedly referencing the ACA, she positions herself as a defender of healthcare protections, framing her platform around expanding access and protecting the healthcare system.
• “Donald Trump left” and “understand Donald Trump”: These trigrams reinforce Harris’s critique of Trump’s presidency. By focusing on what Trump “left” behind and helping the audience “understand” his policies, she paints a picture of the damage she believes was done under his leadership. This rhetorical strategy helps her draw a clear distinction between her vision for the future and Trump’s legacy.
• “Protections Roe Wade”: This trigram highlights Harris’s focus on reproductive rights. In a time when Roe v. Wade faces challenges, Harris is clearly signaling her intent to protect abortion rights and stand against any restrictions on women’s autonomy.

What This Says About Harris

Harris’s frequent use of ngrams like “Donald Trump” and “affordable care act” reveals a speech centered on contrast and critique. Her focus on Trump shows she’s positioning herself as the alternative, while her emphasis on healthcare and reproductive rights makes it clear that these issues are central to her platform.

By consistently discussing policies that affect the “American people” and the “middle class,” Harris’s rhetoric is aimed at connecting with the broader public on key concerns like healthcare and economic security. Her repeated references to reproductive rights and the Affordable Care Act also suggest she’s speaking directly to voters concerned about personal freedoms and healthcare access.

In summary, Harris’s bigrams and trigrams show a speech that is equal parts critique of the past administration and advocacy for key Democratic values. Her rhetoric is focused on highlighting Trump’s failures while promoting healthcare reform, reproductive rights, and economic justice—establishing herself as a strong advocate for the people and a stark contrast to her predecessor.

Conclusion: A Tale of Two Rhetorical Styles

By analyzing the ngrams in Trump and Harris’s speeches, we uncover two starkly different approaches to political messaging. Trump’s language revolves around scale and urgency, with phrases like “millions people” and “billions dollars” showcasing his focus on grand, sweeping issues. His repetition of terms related to immigration, economics, and national security highlights a strategy aimed at instilling a sense of protection and strength—often framing himself as the leader who can defend the nation from external threats.

Harris, on the other hand, weaves a narrative that contrasts Trump’s legacy with her own vision for the future. Her frequent mentions of “Donald Trump” indicate that much of her speech is built around directly challenging his policies and leadership. Phrases like “affordable care act” and “protections Roe Wade” reflect her focus on healthcare and reproductive rights—core issues that resonate with voters looking for compassion, equity, and protection of personal freedoms.

In the end, these ngrams tell a story about what each leader values and how they connect with their audience. Trump leans into themes of power, magnitude, and defense, while Harris pushes forward with a message of healing, progress, and inclusion. Both are speaking to the heart of the American people, but in dramatically different ways—one from a place of strength and caution, the other from a vision of hope and unity. The choice in messaging reveals the deeper values and strategies behind their leadership styles, leaving voters to decide which narrative resonates most with their vision for America’s future.

Full Analysis: https://rpubs.com/bseko/harris_trump_debate

Git Repo: https://github.com/BrianS3/harris_trump_debate

References:
1) https://www.northjersey.com/story/news/politics/elections/2024/09/11/kamala-harris-and-trump-debate-time/75143033007/

2) https://www.washingtonpost.com/politics/2024/09/12/trump-blames-abcs-moderators-his-debate-dud-voters-disagree/

3) https://newrepublic.com/article/146864/trump-slowest-speaker-recent-us-presidents