ENGAGE - Predicting the Trump Presidency


Behind the smiles: Donald Trump and Hillary Clinton

Many believe JFK won 1960 election thanks, at least in part, to his smooth and marketable debate performance against Richard Nixon.

Previous research suggests that candidates’ emotions can play a role in presidential success: a study published in Political Psychology in 2012 found that voters are influenced by candidates' smiles. Viewers’ levels of happiness while watching a candidate debating has also been proven to predict their voting intentions; but what about the emotions of the candidates themselves?


At CrowdEmotion HQ, we were intrigued. To investigate the potential effects of candidate expressions on their voting public, we employed our ENGAGE facial coding platform to analyse the emotions on display during Trump and Clinton's final debate. The data can be found below.

Screen Shot 2018-02-28 at 12.39.25.png

Trump's Emotion Timeseries

Screen Shot 2018-02-28 at 12.39.52.png

Clinton's Emotion Timeseries



Each candidate’s performance at the debate was isolated from a split-screen recording of the debate. We analysed each video with our facial coding software, and recoded to minute-by-minute differences from the debate average per candidate. Facial coding analyses a participant’s emotions through their facial expressions using machine learning; the six universal expressions of emotion are happiness, surprise, anger, disgust, fear and sadness.