The radius of Generalized Trust among different educational groups: Are those 'most people' out-groups?

Projektinformationen

Trust Radius

ID Finanzhilfevereinbarung: 750832

DOI

10.3030/750832

Projekt abgeschlossen

EK-Unterschriftsdatum 29 März 2017

Startdatum 1 April 2018

Enddatum 18 Dezember 2020

Finanziert unter

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Gesamtkosten

€ 177 598,80

EU-Beitrag

€ 177 598,80

Koordiniert durch

UNIVERSITEIT VAN AMSTERDAM
Netherlands

Periodic Reporting for period 1 - Trust Radius (The radius of Generalized Trust among different educational groups: Are those 'most people' out-groups?)

Berichtszeitraum: 2018-04-01 bis 2020-03-31

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

Generalized trust (GT) or trust in ‘most people’ is a conspicuous indicator in at least 77 studies on ethnic diversity and intergroup relations. While social scientists have extensively debated the diversity-trust nexus, few have focused on systematic response bias in answers to this widely used questionnaire item. Most extant research relies on a problematic assumption; that GT taps into attitudes towards strangers and even ethnic out-groups. Using think aloud protocols, a study that aimed at unpacking GT has convincingly shed doubt on this assumption and demonstrated that the majority of respondents high in GT think ‘most people’ refers to people they know, whereas a high proportion of those who are low in GT think about strangers. Clearly, this questions the ecological validity of prior research, and the ensuing policy recommendations. Following the homophily principle – the tendency to associate and bond with similar others – we can expect that people known to the respondents are likely to be ethnic in-groups, and strangers are more likely to be ethnic out-groups, but formally we do not know this. It is imperative to corroborate the relation between GT, intergroup attitudes, and behavior, as well as triangulating the measurements with instruments that are less susceptible to social desirability.
This project thus aimed to return to some of the overlooked methodological concerns in this literature: the measurement and operationalization of generalized trust, which is allegedly in decline by ethnic diversity. Furthermore, this project aimed to relate attitudes to behaviour in a controlled setting, which lacked in prior research. Important to note, behaviour related to GT is often measured only by self-reports.
This project envisaged thus three interrelated objectives:
1. To reanalyse a unique data set, which employed cognitive interviews (think alouds) and mapped what respondents think when they are asked about the trustworthiness of ‘most people’ or generalized trust.
2. To investigate the factorial invariance and structural equivalence of generalized trust, trust towards people one has met for the first time (strangers) and out-group trust, comparing groups with different educational backgrounds across the globe.
3. To experimentally investigate relations between GT, out-group trust, implicit prejudice and social distance towards out-groups. Implicit and behavioural indicators will thus be compared with the previously employed explicit attitudes.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

To achieve these objectives, I devised an extensive conceptual framework for each research objective. Training and data (re)analysis served as the backbone of the project resulting in the achievement of most of the milestones. I designed and programmed experiments by creating an unprecedentedly detailed protocol in order to investigate the relations between GT, out-group trust, implicit prejudice, and social distance towards out-groups. Implicit and behavioural indicators were thus compared with the previously employed explicit self-reported attitudes. This project complied fully with open science practices and preregistration. Although some of the work was disrupted by the onset of the COVID-19 pandemic, two lab experiments and an online experiment were completed comprising 668 responses, which is substantially larger than a typical social science project conducted partly in the laboratory. I have drafted three papers during the action and published a research article, while extended my training beyond Advanced Structural Equation Modelling to include programming in Stata, R, Python, and Machine Learning. A pilot has been designed to extend the current project to include other data, strengthening its external validity. Results of the IF action have thus far been disseminated at prestigious international conferences and workshops at specialized centres across Europe. The project has achieved most of its objectives and milestones for the period. Moreover, all deliverables concerning ethics requirements have been provided.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

The experiments conducted for this action allow for the first time to return to some of the overlooked methodological concerns about GT by relying on novel experimental data collected in the laboratory and online. I examined the convergent validity of GT as a proxy for intergroup attitudes and implicit prejudice. Moreover, a few exceptions notwithstanding, little is known about the predictive validity of GT. I examined this research gap by exploring the correlations between (explicit and implicit) intergroup attitudes and behaviour (Trust Game and social distance as measured by Seating Distance). Results suggest that the relation between on the one hand GT and on the other hand, in-group attitudes, out-group attitudes, and behaviour, might be weaker than expected. More specifically, only two correlations run in the expected direction: feeling warmer towards out-groups negatively correlates with transfers to in-groups, while higher Implicit Association Test (IAT) scores negatively correlate to transfers to out-groups. In contrast, relying on a larger sample than typical in social psychology, I could not replicate the negative correlation between the IAT and Seating Distance towards out-groups. Another striking finding is that participants who had completed an IAT before the Trust Game, transferred less money to an out-group, which goes to suggest that the IAT may act as a prime and necessitates further research. All correlations are examined in the larger online sample, randomizing the order of tasks and introducing a novel behavioural measure (willingness to enter a chat room with an ethnic in-group or out-group). Finally, I started data collection before the onset of Covid-19 pandemic and have continued to do so during the pandemic. As such, this allows me to explore whether a natural hazard has the potential to exacerbate negative feelings and behaviour towards out-groups, while inducing greater in-group favouritism, and how generalized trust impacts these relations.
In sum, this project examines the validity of Generalized Trust (GT) as a proxy for out-group attitudes, implicit race bias, and its relation to behaviour in an unprecedentedly broad manner using an innovative experimental protocol. Moreover, I explored the effect of the Covid-19 pandemic on these relations and the impact of taking an IAT on subsequent behaviour. Results are relevant to political methodologists, psychologists, and survey researchers interested in the diversity-trust nexus. When studying the trust radius at the individual level, I therefore suggest analysts should consider alternative mechanisms going beyond intergroup relations. Social psychology has in intergroup relations theory one of its most valuable explanatory tools. Despite its intuitive appeal, scholars and policymakers alike should not lose sight of areas in which it may have little predictive power. The present study articulates this problem for individual-level studies of trust that take self-reported survey questionnaires at face value.