Ask the Experts: Top Companies Chime in on What’s Changing in 2019
By Alex JonesDec 18th
Published October 7th 2016
With those assumptions in mind, we wanted to see just how accurate they are, and whether or not all Clinton supporters drive Subarus (not necessarily) or all Trump supporters drive trucks (actually pretty accurate).
First, a note on methodology. In order to get panels of Clinton and Trump supporters, we used the Brandwatch Audiences tool to search for would-be voters. This search was limited to individuals in the United States with three possible hashtags for either candidate.
For Hillary Clinton supporters, we searched for #imwithher, #hillyes, and #Hillary2016. For supporters of Donald Trump, we looked for #trumptrain, #makeamericagreatagain, and #trump2016. In order to limit the audiences, we then created a random panel of 4,000 authors with between 50 and 50,000 followers.
To find out exactly what they drive, we searched across the general population for mentions in which drivers explicitly mentioned the brands or types of cars they drive. We wanted to look at a variety of car brands, so we compiled a list of 15 leading foreign and domestic brands and searched for personal pronouns used near the brand name (say, Toyota) and the brand’s products (Prius, Tacoma, etc).
Finally, to prevent a few viral mentions dominating the conversation, we excluded retweets at the query level.
First off, we’ll examine the frequency of brand-specific mentions for each panel of supporters. We found that Trump supporters were more likely to discuss their car brand. When considered as part of the full 4,000-author panel, it’s clear that these posts are fairly infrequent, with .21 brand mentions per author among the Trump panel and .14 brand mentions per author for Clinton supporters.
Trump supporters were also more likely to mention their favored candidate within the context of these car brands, suggesting a greater trend toward politicizing among the author panel.
Further to this point, the following chart shows mentions per author broken down by the brands analyzed for our study. We can see that Trump supporters are most likely to discuss their Jeeps, while Clinton supporters are most likely to mention that they drive a Ford.
Reflecting the fact that they are more likely to mention what brand they drive, Trump supporters lead the mentions per author for the majority of brands. Clinton supporters lead this proportion for only Toyota, Mazda, Audi, and Subaru. Among the Clinton supporters tweeting about their Toyota’s, nearly half of all the mentions included ownership of a Prius.
Next, we broke down these branded mentions by unique authors. That way, we were able to account for supporters that may have mentioned their Ford Explorer or Honda Civic more than once in the past year, and only count them once. Once we found these totals, we were able to divide these numbers by the supporter totals (4000) in order to find the percentage of supporters that mentioned their brand.
As shown, the greatest number of Clinton and Trump supporters were Ford owners. Clinton supporters were more likely to drive Toyotas, Hondas, Mazdas, Audis, Subarus, and Lexuses. On the other hand, Trump supporters were more likely to drive more traditionally American brands such as Jeep, Chevrolet, and Dodge but also the strongly German brand BMW.
Taking a more top-down view, we also considered the brands as foreign and domestic, with six domestic and nine foreign brands.
Again, we considered unique authors and not mention volume for this chart, which shows that Trump supporters were more likely (7% of authors) to own American brands. Of the 4,000 Trump supporters, nearly 300 mentioned their American cars (Ford, Jeep, Cadillac, Chevrolet, Buick, and Dodge). Clinton supporters were more likely to mention their foreign brands (Toyota, BMW, Honda, Mazda, Audi, Porsche, Subaru, Lexus, or Kia).
When accounting for the imbalance in brands, with six domestic and nine foreign, these preferences are somewhat negated. Clinton supporters have approximately 31 unique authors per American brand, 23 per foreign. On the other hand, the gulf in ownership for Trump supporters widens, with 47 unique authors per domestic brand, and 19 per foreign.
Finally, we wanted to consider the demographics of non-specific car ownership through the supporter lens. Using a similar methodology as the brand conversation discussed earlier, we searched for personal pronouns near the following: truck, minivan, SUV, or Hybrid.
Again, we found that Trump supporters were more likely to mention their vehicle type than Clinton supporters. However, these distribution totals were far lower than brand-specific mentions, with approximately .045 mentions per author in the Trump panel.
When segmenting by vehicle type, we can see that both Clinton and Trump supporters mention their truck ownership most often. However, a greater portion of the Trump panel discussed their truck ownership. Trump supporters were far more likely to politicize their ownership, with 5% of truck mentions including ‘Trump,’ while not a single mention of trucks among her supporters included Clinton.
Using social data, we can see just how far our preconceived notions go, and where they fall short. To this end, the Audiences tool is especially effective, as it allows panels to be created and compared, moving beyond just bumper stickers and vague ideas of supporter stereotypes. Of course, it remains to be seen which bumper sticker will be the most obsolete in the coming years. We’ll revisit in 2020.