Table 1.

Hashtag persistence for each discursive community across the entire temporal period covered by our data set (51 days in total).

$H_t$	M5S	CDX	CSX
100%	dimaio, lega, renzi, berlusconi, m5s, pd, italia	salvini, m5s, centrodestra, pd, lega	renzi, salvini, dimaio, m5s, pd
98%	forzaitalia, salvini	berlusconi, italia, renzi
96%	roma, ottoemezzo	forzaitalia	berlusconi, italia, lega
94%	centrodestra, ricercapubblica		russia
92%	boschi, politica	dimaio	europa, politica, roma
90%	fi, governo	fi, governo
88%	casapound	roma
86%	meloni
84%	fakenews, lavoro, liberieuguali	casapound, politica	forzaitalia, lavoro, usa
82%	8800precari, gentiloni, migranti, senato, voto	governo, lombardia	centrodestra, leu, liberieuguali
80%	bonino, campagnaelettorale, casini, leu, rosatellum	cdx, flattax, sinistra	milano, partitodemocratico, ue
78%	avanti, iovotom5s, movimento5stelle, precari, sinistra	lavoro, ue	campagnaelettorale, fakenews, governo

The first column shows the percentage of days each hashtag is present in the set of tweets of each community. Notice that the hashtags that are always present are those carrying the name of political parties and political leaders, while other relevant themes for the political debate are absent from (at least) some of the discursive communities. These findings suggest that the online political debate is largely focused on single personalities/political entities (as particularly evident upon inspecting the CSX hashtags) and only to a much smaller extent on themes of public interest.