Populists’ Use of Nostalgia: A Supervised Machine Learning Approach

Frischlich L, Clever L, Wulf T, Wildschut T, Sedikides C

Abstract

An emotion that has recently gained traction in the context of populism is nostalgia, a
sentimental longing or wistful affection for the past. Nostalgia can refer to the past of
one’s group or nation, as reflected in populists’ narratives of the heartland—the vision of
a utopian future based on an idealized past in which their country belonged to the “pure
people.” However, research on nostalgia in political communication across the political
aisle is scarce. The current study aimed to fill this gap via supervised machine learning.
First, we used an experimental approach established in psychology to create a groundtruth
data set and trained and evaluated a classifier for detecting nostalgic sentiment in
the German language. We then applied this classifier to a large database (N = 4,022) of
German political parties’ Facebook posts. We demonstrate that (a) populist (vs. nonpopulist)—
especially right-wing—parties employ nostalgia more frequently; (b) nostalgic
narratives differ between parties, and (c) nostalgic (vs. non-nostalgic) posts are
associated with more user engagement

Keywords
Populism; Social Media; Machine Learning; Nostalgia