Drivers of Trust in Public Institutions in Chile

Annex A. Methodological notes on the OECD Trust Survey in Chile

A.1 Data collection and survey implementation in Chile

The 2023 OECD Survey on Drivers of Trust in Public Institutions is the second survey wave after the 2021 inaugural Trust Survey. The survey was fielded between the 25^th of October and the 27^th of November in Chile and in the other 29 OECD countries at a similar time between the 20th of September and the 12th of December 2023. The survey yielded a total number of 2 008 valid responses among the adult population (18+) in Chile and 58 230 in total among 30 OECD countries. The survey data in Chile was collected by the data collection provider Ipsos, coordinated by the OECD Secretariat.

Following the methodology from the 2021 survey wave (Nguyen et al., 2022[1]), in most countries, the 2023 survey relied on a non-probability sampling approach. It consisted of ex ante country-level quotas on the distribution of age, gender, education and regions (hard quotas) and income (soft quota). The country-specific quotas on the distribution of age, gender, education and region, together with the ex-post weighting, ensure national representativeness of the survey data for these characteristics. The quotas were derived from different national and OECD sources. The education levels and income deciles in Chile were compiled from Encuesta Casen 2020 (Encuesta de Caracterización Socioeconómica Nacional) and the regions from OECD.Stat (Regional Demography – Population). In Chile, the online survey was conducted by the survey provider Ipsos and the sample was based on Ipsos' and partners’ online panels, comprised of individuals in the country who willingly signed up to be engaged in market research surveys. Details on the sample are presented in Table A ‎A.1.

Table A ‎A.1. Sample gender, age and education distribution in Chile (unweighted and weighted)

Note: Low education group: <Educación básica; Middle education group: <Educación media; High education group: > Tecnico nivel superior.

Source: OECD Trust Survey 2023.

As indicated in Table A ‎A.1, the education quota was intended to be based on three groups (low, medium and high education). Employing the International Standard Classification of Education scheme (ISCED 2011) (UNESCO, 2012[2]) , respondents were asked to specify their highest level of education attained, corresponding to the equivalent level within their country's educational framework. However, the ‘low’ education demographic group represented a small share of the panels in certain countries, providing challenges to reach the quota target for education even when the survey provider drew on additional panels. For this reason, education levels were combined into two groups (1- low/medium and 2- high) for analysis in Chile, Costa Rica, Colombia, Denmark, Greece, Estonia, Portugal, and South Korea. The survey provider collaborated with additional survey panel providers to ensure a more comprehensive representation of diverse population segments within the survey data; and extended fieldwork in Costa Rica, Chile, Colombia, Greece, Portugal and Slovenia.

The comparative analyses between the weighted samples and the target distribution shows that post-stratification weights efficiently minimised within-country variations across most socio-demographic quota variables. Nevertheless, despite the use of quotas and post-stratification weights, the weighted distribution deviates from the within-country distribution by education and age in a few countries. In Chile and Greece, the combined low and medium education group deviated by 6 and 2 percentage points, respectively, from the actual distribution (Table A ‎A.1). The underrepresentation of these groups of respondents can be attributed to two factors. Firstly, the lower-educated and older 65+ individuals are underrepresented among panel members. Secondly, these groups additionally tend to drop out more frequently during the survey. Detailed information on samples and methodological aspects of the 2023 OECD Trust Survey can be found in the technical annex to the 2024 OECD Trust Survey Report (OECD, 2024[3]).

As the survey was implemented for the first time in Chile, the State Modernisation Secretariat at the Ministry of Finance carried out cognitive interviews to test the Spanish translation of the survey questions, the understandability of the wording in the Chilean context, as well as the wording of the additional survey questions implemented in Chile. For comparability, the survey questions tested in the cognitive interviews were based on the Spanish translations implemented in Mexico and Colombia in the 2021 survey wave. Nonetheless, Chile’s cognitive testing results underscored the need to clarify complex political or economic terms; for example, including definitions of public consultations or greenhouse gases in relevant questions. Additionally, the mentioning of a response scale in the survey questions were streamlined in the English baseline questionnaire. Most of the suggested changes were relevant and included in the Spanish version of the questionnaire implemented in Chile (see Annex B).

Chile added survey questions to the English baseline questionnaire tailored to the countries’ needs (Table A ‎A.2). Chile was curious about the other side of the citizen and government relationship, namely if people perceive the government can trust fellow citizens to provide accurate information; and the perceived ability of regional representatives to voice regional views given recent decentralisation reforms.

Table A ‎A.2. Additional survey questions in Chile

Survey module	Concept: What is measured?	Survey question	Country
Trust	Trust in electoral system/civil society organizations	On a scale of 0 to 10, where 0 is not at all and 10 is completely, how much do you trust each of the following?	Chile
Trust	Reciprocal trust: Trust of Chilean government in citizens	On a scale of 0 to 10, to what extent would you say the Chilean government can trust citizens to provide accurate information when applying for benefits or paying taxes?	Chile
Openness	Opportunity of regional representatives to voice regional views and needs	If a decision affecting your region is to be made by the national government, how likely or unlikely do you think it is that regional representatives would have an opportunity to voice regional views and needs?	Chile
Openness	Confidence in country's ability to cooperate with stakeholders	On a scale of 0 to 10, how confident are you in the national government’s ability to cooperate with other national stakeholders, such as private sector organisations and trade unions to better tackle long-term challenges?	Chile
Background	Social mobility: Prospects of personal economical situation	Thinking about your current economic situation and future prospects, do you think you will do better, worse or about the same as your parents have?	Chile
Background	Feeling of physical insecurity: Frequency of worrying about crime	How often, if at all, do you worry about becoming a victim of a violent crime in [Country]?	Chile

Note: The table shows the survey questions that were implemented in the 2023 OECD Trust Survey in Chile.

A.2 Logistic regression models to assess the drivers of trust

The econometric results presented below are logistic regression analyses for establishing the main drivers of trust in the national government, the local government and the civil service in Chile. The outcome variables – or dependent variables – in the regression models implemented for this report is trust in the national government, the local government and the national civil service. Those three variables are recoded as a binary variable, with two categories for low or no trust (0-4) and moderately to high trust (6-10). Responses (5) and “don’t know” are excluded.

The analysis operationalizes government competencies (including satisfaction with administrative services) and values through 19 variables, measured on a 0-10 response scale and standardized for the analysis. Political agency is operationalized through two variables measuring internal and external political efficacy, meaning an individual’s confidence in participating in politics and their perception that people like them have a say in what government does. Perceptions of government actions on global and long-term challenges are measured by survey questions on confidence in the country’s success in reducing greenhouse emissions, and confidence that the government balances the interests of current and future generations.

The following explains the technical details about the econometric analysis.

• Model specification:

  All three models control for individuals’ socio-demographic characteristics (age, gender, education), interpersonal trust, experiencing security or financial concerns, perceptions on economic mobility and feeling part of a group discriminated against. The models also control for whether people voted (or would have voted) for one of the parties currently in power. Regression results referring to the control variables are not shown. All models include survey weights. The logistic regression model uses a stepwise deletion process in which only highly significant variables are kept in each deletion step. These regression models also include three additional survey questions implemented in the Chilean survey as independent variables (Table A ‎A.2).

• Technical interpretation:

  The statistically significant drivers are shown as average marginal effects. Statistically significant refers to those public governance variables included in the logistic regression model that resulted in p<0.05. The technical interpretation of the effect of government’s reliability in taking evidence-based decisions on trust, for example, is that a one-standard-deviation increase in perceived reliability is associated with a 8 percentage point increase in trust in the national government. Or – taking into consideration all other variables in the model – all else being constant, moving from the average citizen to one with a typically higher level of confidence in government’s reliability is associated with a 8 percentage point increase in trust in the national government.

• Cautious interpretation:

  Some caution should be taken when interpreting the regression results, as they only represent a correlational not a causal relationship between trust in the three public institutions and the public governance drivers of trust. Additionally, all independent variables in the regression model are correlated and the direction of the relationship between trust and the drivers of trust may point the other way too.

References

[1] Nguyen, D. et al. (2022), “Survey design and technical documentation supporting the 2021 OECD Survey on Drivers of Trust in Government Institutions”, OECD Working Papers on Public Governance, No. 53, OECD Publishing, Paris, https://doi.org/10.1787/6f6093c5-en.

[3] OECD (2024), OECD Survey on Drivers of Trust in Public Institutions – 2024 Results: Building Trust in a Complex Policy Environment, OECD Publishing, Paris, https://doi.org/10.1787/9a20554b-en.

[2] UNESCO (2012), International Standard Classification of Education ISCED 2011.