People possess cognitive associations between “political” and “apolitical” traits and use these to form judgments about others. I conceptualize this behavior as politicultural linking (PL): the act of drawing inferences about political traits (e.g., ideology, partisanship) from apolitical cues (e.g., sociodemographics, personality, lifestyles) and vice versa. This work answers three main questions: (1) Can PL be observed in Italy? (2) When is PL more likely to occur? Namely, who engages in this behavior, about whom, and in which informational contexts? (3) What are the social consequences of PL? I answer these questions with four survey experiments fielded to original, representative samples. A vignette experiment examines whether ideological and partisan inferences can arise spontaneously from observing someone’s lifestyles, specifically food preferences. A forced-choice textual conjoint investigates the relative roles played by different sociodemographic, personality, and lifestyle traits in engendering these inferences. A visual conjoint experiment investigates whether PL is relevant when directed towards politicians. Finally, a dynamic parallel-design experiment inquires into whether and how PL intervenes in the decisions to engage in political conversation. Together, these designs allow me to map the content and direction of PL, compare the relative weight of different apolitical cues, test robustness when political information is explicit, and quantify social consequences in Italy. Through these empirical endeavors, I find that over half of a representative sample of Italians engage in ideological or partisan inferences from a target food preferences. The people who do are more likely to self-place ideologically, consume more news, and report higher cultural consumption. Moreover, perceiving a target as an out-ideologue decreases willingness to interact and worsens expectations about a political discussion, offering correlational evidence of potential conversational sorting. Second, PL extends far beyond lifestyles. In the textual conjoint, sociodemographic, psychological, and lifestyle attributes independently triggered ideological inferences; profession, religion, and preferred means of transportation produced the strongest shifts in perceived ideology, indicating that everyday, nonpolitical descriptors meaningfully shape political impressions even absent explicit political information. Third, PL is still consequential even when political information is available. In the visual conjoint, respondents viewed sociodemographic and lifestyle cues alongside explicit issue positions. As expected, the weight of apolitical cues shrinks when explicit political information is available, but it does not disappear; sociodemographic signals in particular continued to inflate or deflate ideological perceptions. This demonstrates that PL is not merely an artifact of information scarcity and travels to more realistic contexts. Fourth, PL is not the primary brake on cross-cutting talk. The parallel-design experiment shows that apolitical similarity (e.g., shared sociodemographic, psychological, or lifestyle traits) directly increases the propensity to engage in political conversations, with little evidence that political inferences mediate these effects. The pattern generalizes to other European countries. Overall, the contributions of this work are both theoretical and empirical. Theoretically, PL consolidates disparate literatures on partisan/ideological stereotyping. Moreover, the substantial conclusions of the empirical investigations answer the main “who”, “when”, and “why” about PL. Methodologically, the dissertation advances survey-experimental practice by (i) deploying a visual conjoint that randomizes multiple layered features of social-media-like profiles, and (ii) adapting a parallel-design conjoint to quantify mediation by political inferences in conversational choices.
Le persone possiedono associazioni cognitive tra tratti “politici” e “apolitici” e le utilizzano per formulare giudizi sugli altri. In questo lavoro, concettualizzo questo comportamento come “Politicultural Linking” (PL): l’atto di trarre conclusioni sui tratti politici (ad esempio, ideologia, appartenenza partitica) da indizi apolitici (ad esempio, dati sociodemografici, personalità, stili di vita) e viceversa. Questo lavoro risponde a tre domande principali: (1) Il PL può essere osservato in Italia? (2) Quando è più probabile che si verifichi il PL? Vale a dire, chi si manifesta questo comportamento, su chi e in quali contesti informativi? (3) Quali sono le conseguenze sociali del PL? Queste domande vengono investigate attraverso quattro survey experiments su campioni originali e rappresentativi. Un vignette experiment esamina se le inferenze ideologiche e partitiche possano sorgere spontaneamente dall'osservazione degli stili di vita di qualcuno, in particolare delle preferenze alimentari. Un conjoint experiment testuale a scelta forzata indaga i ruoli relativi svolti da diversi tratti sociodemografici, di personalità e di stile di vita nel generare queste inferenze. Un conjoint experiment visuale indaga se il PL sia rilevante quando è diretto verso i politici. Infine, un esperimento dinamico con disegno parallelo indaga se e come il PL intervenga nelle decisioni di intraprendere conversazioni politiche. Insieme, questi esperimenti mi consentono di mappare il contenuto e la direzione del PL, confrontare il peso relativo di diversi segnali apolitici, testare la robustezza quando le informazioni politiche sono esplicite e quantificare le conseguenze sociali in Italia. Attraverso questi sforzi empirici, ho scoperto che oltre la metà di un campione rappresentativo di italiani formula inferenze ideologiche o partigiane a partire dalle preferenze alimentari delle altre persone. Coloro che lo fanno sono più propensi a collocarsi ideologicamente, consumano più notizie e riferiscono un consumo culturale più elevato. Inoltre, percepire un target come ideologicamente avverso diminuisce la disponibilità a interagire e peggiora le aspettative riguardo a una discussione politica, offrendo prove correlazionali di una potenziale segregazione conversazionale. In secondo luogo, il PL va ben oltre gli stili di vita. Nel conjoint testuale, gli attributi sociodemografici, psicologici e di stile di vita hanno innescato in modo indipendente inferenze ideologiche; la professione, la religione e il mezzo di trasporto preferito hanno prodotto i cambiamenti più significativi nell'ideologia percepita, indicando che i descrittori quotidiani e non politici influenzano in modo significativo le impressioni politiche anche in assenza di informazioni politiche esplicite. In terzo luogo, il PL è ancora rilevante anche quando sono disponibili informazioni politiche. Nel conjoint visuale, gli intervistati hanno visualizzato indizi sociodemografici e di stile di vita insieme a posizioni esplicite su questioni specifiche. Come previsto, il peso degli indizi apolitici diminuisce quando sono disponibili informazioni politiche esplicite, ma non scompare; i segnali sociodemografici, in particolare, hanno continuato a gonfiare o sgonfiare le percezioni ideologiche. Ciò dimostra che il PL non è solo un artefatto della scarsità di informazioni, ma si trasferisce in contesti più realistici. Infine, il PL non è il principale freno al dialogo trasversale. L'esperimento con disegno parallelo mostra che la somiglianza apolitica (ad esempio, tratti sociodemografici, psicologici o di stile di vita condivisi) aumenta direttamente la propensione a impegnarsi in conversazioni politiche, con poche prove che le inferenze politiche possano mediare questi effetti. Il modello si generalizza ad altri paesi europei.
Scaduto, G (2026). Thinking About Politics: An Experimental Inquiry into the Cognitive Associations between Political and Apolitical Traits and Their Social Consequences in Italy. (Tesi di dottorato, , 2026).
Thinking About Politics: An Experimental Inquiry into the Cognitive Associations between Political and Apolitical Traits and Their Social Consequences in Italy
SCADUTO, GAETANO
2026
Abstract
People possess cognitive associations between “political” and “apolitical” traits and use these to form judgments about others. I conceptualize this behavior as politicultural linking (PL): the act of drawing inferences about political traits (e.g., ideology, partisanship) from apolitical cues (e.g., sociodemographics, personality, lifestyles) and vice versa. This work answers three main questions: (1) Can PL be observed in Italy? (2) When is PL more likely to occur? Namely, who engages in this behavior, about whom, and in which informational contexts? (3) What are the social consequences of PL? I answer these questions with four survey experiments fielded to original, representative samples. A vignette experiment examines whether ideological and partisan inferences can arise spontaneously from observing someone’s lifestyles, specifically food preferences. A forced-choice textual conjoint investigates the relative roles played by different sociodemographic, personality, and lifestyle traits in engendering these inferences. A visual conjoint experiment investigates whether PL is relevant when directed towards politicians. Finally, a dynamic parallel-design experiment inquires into whether and how PL intervenes in the decisions to engage in political conversation. Together, these designs allow me to map the content and direction of PL, compare the relative weight of different apolitical cues, test robustness when political information is explicit, and quantify social consequences in Italy. Through these empirical endeavors, I find that over half of a representative sample of Italians engage in ideological or partisan inferences from a target food preferences. The people who do are more likely to self-place ideologically, consume more news, and report higher cultural consumption. Moreover, perceiving a target as an out-ideologue decreases willingness to interact and worsens expectations about a political discussion, offering correlational evidence of potential conversational sorting. Second, PL extends far beyond lifestyles. In the textual conjoint, sociodemographic, psychological, and lifestyle attributes independently triggered ideological inferences; profession, religion, and preferred means of transportation produced the strongest shifts in perceived ideology, indicating that everyday, nonpolitical descriptors meaningfully shape political impressions even absent explicit political information. Third, PL is still consequential even when political information is available. In the visual conjoint, respondents viewed sociodemographic and lifestyle cues alongside explicit issue positions. As expected, the weight of apolitical cues shrinks when explicit political information is available, but it does not disappear; sociodemographic signals in particular continued to inflate or deflate ideological perceptions. This demonstrates that PL is not merely an artifact of information scarcity and travels to more realistic contexts. Fourth, PL is not the primary brake on cross-cutting talk. The parallel-design experiment shows that apolitical similarity (e.g., shared sociodemographic, psychological, or lifestyle traits) directly increases the propensity to engage in political conversations, with little evidence that political inferences mediate these effects. The pattern generalizes to other European countries. Overall, the contributions of this work are both theoretical and empirical. Theoretically, PL consolidates disparate literatures on partisan/ideological stereotyping. Moreover, the substantial conclusions of the empirical investigations answer the main “who”, “when”, and “why” about PL. Methodologically, the dissertation advances survey-experimental practice by (i) deploying a visual conjoint that randomizes multiple layered features of social-media-like profiles, and (ii) adapting a parallel-design conjoint to quantify mediation by political inferences in conversational choices.| File | Dimensione | Formato | |
|---|---|---|---|
|
phd_unimib_896343.pdf
embargo fino al 11/02/2029
Descrizione: Tesi definitiva post revisioni
Tipologia di allegato:
Doctoral thesis
Dimensione
16.56 MB
Formato
Adobe PDF
|
16.56 MB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


