CoChat
Sign Up
Cited by 135E-Government and Public ServicesPublic Policy and Administration ResearchCorruption and Economic Development

Transparency and Trust in Government. Evidence from a Survey Experiment

Open access
Alessandro, Martin, Cardinale Lagomarsino, Bruno, Scartascini, Carlos, Streb, Jorge, Torrealday, Jerónimo·2021·World Development
DOI 10.1016/j.worlddev.2020.105223
Open this paper in CoChat — free accountOpen PDF

Abstract

Does providing information improve citizens’ perception of government transparency? Does all information matter the same for shaping perceptions about the government? This article addresses these questions in the context of an online randomized survey experiment conducted in Argentina. Results show that providing information to citizens matters for shaping perceptions about transparency, and the content of that information matters for affecting the evaluation people make about the government. Those who received a treatment showing that the government was over-performing on its promises had higher trust in the government than those who received a treatment showing that the government was under-performing. The evidence highlights that the relationship between transparency and trust may be mediated by the performance of the government. Our complementary analytical discussion shows that the rules for disclosing information may be important for transparency to matter.

AI snapshot

Written from the paper’s full text.

TL;DR

An online randomized survey experiment in Buenos Aires finds that actively providing information raises perceived government transparency, and that information content about performance matters more for trust than whether the message is framed as efficiency or empathy. The results suggest transparency’s link to trust is mediated by perceived government performance.

Key findings

  • Providing information about government commitments increased the perception that the government reports what it plans to do by about 8 to 9 percentage points, an increase of more than 10%.
  • Framing did not significantly change trust: there were no relative differences in trust between an “efficiency-based” framing and an “empathy-based” framing.
  • Performance content mattered: those shown that the government was over-performing had higher trust than those shown under-performance, with a difference equivalent to about 0.10 of a standard deviation.
  • Effects were concentrated among those with no previous exposure to the commitments; about 43% of respondents were first exposed, and the treatment × first-exposure coefficient for trust was 0.300 (p < 0.01).

Methods

  • Online randomized survey experiment conducted in the City of Buenos Aires, Argentina, between December 2017 and January 2018, with 1,999 complete interviews stratified by gender, age, socioeconomic status, and commune.
  • Participants were randomized first between control and information treatment, then assigned to one of three vignettes: T1 empathy and positive results, T2 efficiency and positive results, T3 empathy and negative results.
  • Analysis used OLS regressions with controls, principal component analysis to construct trust indices (competence, benevolence, honesty, confidence, and overall trust), and robustness checks including multiple-testing corrections and alternative index definitions.

Limitations

  • The paper states that treatment differences were “not that stark” and respondents had to analyze figures to infer performance, so stronger cues could have larger effects; this remains an open question.
  • The paper calls for future studies on optimal disclosure, rules versus endogenous disclosure, and contexts where performance does or does not matter.
  • Evident from the text: the sample is from one city and relies on self-reported perceptions in an online survey, so findings may not generalize to other populations or settings.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.worlddev.2020.105223

Cites

79 of 79 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

4 of 27 cited sources checked · 1 supported · 0 flagged · accuracy not yet measured

  • Interpretationcitation

    Performing the experiment in a country where overall levels of trust are very low expands the existing evidence and the understanding of the phenomenon.

    “Performing the experiment in a country where overall levels of trust are very low expands the existing evidence and the understanding of the phenomenon (Cucciniello et al. (2017)).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Cucciniello et al. (2017) source

      The sections “Replicate Studies across Different Administrative Contexts” and “Replicate Studies Using Different Methods” say that research in new contexts and experimental replication can improve understanding of transparency’s effects. They do not identify a country with very low overall trust or specifically establish low trust as the condition that makes this experiment an addition to the evidence. Page numbers are unavailable in our copy. (LLM, conf=0.83)

      Experimental — accuracy not yet measured
    • Rests on an assumption: The online survey sample is representative of the adult population of the City of Buenos Aires.
  • Findingown data

    About 43% of respondents were uninformed (first exposure) about the government commitments despite the existence of a dedicated website and public announcement.

    “Interestingly, despite the existence of a dedicated website and public announcement of commitments, plenty of people are uninformed -first exposure-about the government commitments (about 43%)”
    Quote in the paperNot a substantial claim
  • Findingown data

    Individuals who received information that the government was over-performing on its commitments had higher trust in the government than those who received information that it was under-performing.

    “Those who received a treatment showing that the government was over-performing on its promises had higher trust in the government than those who received a treatment showing that the government was under-performing.”
    • Also stated in: § 1. Introduction · plain
      “The group that received information where the government was over-performing on its goals showed significantly higher trust than those who received information that the government was underperforming on its goals.”
    • Also stated in: § 3.3. Empirical results · plain
      “The first and last bars are statistically different from each other (we don't compare the second and third because both components are changing).”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The information about government commitments and performance presented in the vignettes is accurate and credible.
    • Rests on an assumption: The online survey sample is representative of the adult population of the City of Buenos Aires.
    • Rests on an assumption: Participants in the treatment group read and understood the information in the vignettes as intended.
    • Rests on an assumption: The government's performance relative to its commitments is a valid indicator of government competence, benevolence, and honesty.
  • Findingown data

    Citizens' perceptions about transparency increase regardless of the framing and content of the message.

    “Results in those columns show that the magnitude and the significance is the same for every informative treatment (we cannot reject the hypotheses that the coefficients are the same).”
    • Also stated in: § 3.3. Empirical results · plain
      “That is, citizens' perceptions about transparency increase regardless of the framing and content of the message.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Participants in the treatment group read and understood the information in the vignettes as intended.
  • Findingown data

    The content of the information about government performance affects the evaluation people make about the government.

    “the content of that information matters for affecting the evaluation people make about the government”
    • Also stated in: § 1. Introduction · hedged
      “differences in performance seem to matter”
    • Also stated in: § 1. Introduction · plain
      “The group that received information where the government was over-performing on its goals showed significantly higher trust than those who received information that the government was underperforming on its goals.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The information about government commitments and performance presented in the vignettes is accurate and credible.
    • Rests on an assumption: The online survey sample is representative of the adult population of the City of Buenos Aires.
    • Rests on an assumption: Participants in the treatment group read and understood the information in the vignettes as intended.
    • Rests on an assumption: The government's performance relative to its commitments is a valid indicator of government competence, benevolence, and honesty.
  • Classified as probably not a substantial claim. 27
  • May combine two claims. 27
  • The quote only partly states this claim. 3

  • Classified as probably not a substantial claim.

    Premisecitation

    Trust in institutions and other people is a key factor in social and economic progress as well as democratic stability.

    “Numerous studies have identified trust -both trust in institutions and trust in other people-as a key factor in social and economic progress as well as democratic stability”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites Jacobsen, 1999 source
    • cites Algan & Cahuc, 2014 source

      The 'Future avenues: Trust and Well-Being' section says trust affects economic development and interacts with institutions, while 'Co-evolution of trust and institutions' links social capital to political accountability. These passages support much of the broad claim, but they do not establish that trust in institutions, specifically, is a key driver of democratic stability. Page numbers are unavailable in our copy. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites Bjørnskov, 2015 source

      The paper reports a positive association between social trust (trust in other people) and productivity, with effects linked to the quality of economic-judicial institutions (Conclusion, p. 25). It does not establish that trust in institutions has the same role. Section 2.2 (p. 9) distinguishes democratic institutions and says the evidence for an effect of social trust on them is weaker; the paper does not establish a link to democratic stability. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The experimental setup resembles the setup with two-dimensional asymmetric information in Stein and Streb (2004).

    “This setup resembles the setup with two-dimensional asymmetric information inStein and StrebStreb (2004)”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Stein and StrebStreb (2004) source

  • May combine two claims. The quote only partly states this claim.

    Findingown data

    The distribution of answers for control and treatment groups differs, with more people in the treatment group believing that the government provides information on plans and results.

    “we strongly reject the null hypothesis of equality in distributions (p-value smaller than 1%).”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Participants in the treatment group read and understood the information in the vignettes as intended.

  • May combine two claims.

    Findingown data

    The empathy framing produces slightly better evaluations of government than the efficiency framing, but the differences are not statistically significant for every index.

    “Those who were presented with the empathy framing seem to evaluate the government slightly better than those who received the efficiency one, but the differences are not statistically significant.”
    • Also stated in: § 3.3. Empirical results · plain
      “This is true for every index.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Individuals with no previous exposure trusted the government less but were more influenced by the treatments.

    “Those individuals trusted the government less but were more influenced by the treatments.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Results are quite consistent across the different questions.

    “Results are quite consistent across the different questions.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingreasoning

    For a prior pH=0.5, trust will be larger with two good draws (V_GG) than with two bad draws (V_BB): Pr[CH|V_GG]=0.381 > Pr[CH|V_BB]=0.059.

    “For a prior p H ¼ 0:5, trust will be larger with two good draws, V GG , than with two bad draws, V BB :”
    Quote in the paperSubstantial claim
    • The authors state this as an assumption: Honest types CH and IH always reveal the results of the two random draws, while dishonest types DC manipulate the information to their advantage, revealing only good draws.
    • The authors state this as an assumption: Trustworthy types are identified with types CH that are both competent and honest.

  • May combine two claims.

    Findingown data

    The positive-performance vignette increases ratings of government competence, benevolence, and honesty compared with the negative-performance vignette.

    “Again, those who receive the vignette showing compliance with the promises rate the government as being more competent, more benevolent, and more honest.”
    • Also stated in: § 3.3. Empirical results · plain
      “Similar results exist across the different attributes that evaluate trust in the government (and are part of the composite index).”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    There are no significant effects for the empathy versus efficiency analysis on individual questions.

    “There are no significant effects for the empathy versus efficiency analysis.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The treatment showing overperformance has an effect between 2 and 3 percentage points, representing an increase of about 5 percent for most questions, higher than the one that shows underperformance.

    “Results in the Table show that the treatment showing overperformance has an effect between 2 and 3 percentage points, which represents an increase of about 5 percent for most questions, higher than the one that shows underperformance.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    In Table 4, column (3), first exposure is associated with lower trust in government for the empathy vs. efficiency comparison (coefficient -0.325, p<0.01).

    “First Exposure | À0.145*** | À0.496*** | À0.325***”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    In Latin America, corruption scandals and poor performance of democracies have eroded public confidence in institutions for years.

    “It is particularly true in Latin America, where corruption scandals and poor performance of democracies have eroded public confidence in institutions for years, according to data from the Latinobarometer and the Americas Barometer at LAPOP”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Latinobarómetro, 2018b

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Zechmeister & Lupu, 2019

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Premisereasoning

    Most people tend to remain ignorant and do not pursue information, but they react rationally once they receive information about the workings of government.

    “Therefore, most people tend to remain ignorant and do not pursue information, but they react rationally once they receive information about the workings of government”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Downs, 1957 source
    • cites Brennan & Lomasky, 1997

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Congleton, 2001 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    The rules for disclosing information may be important for transparency to matter.

    “Our complementary analytical discussion shows that the rules for disclosing information may be important for transparency to matter.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Conclusions on statistical inference remain the same with the preferred p-value correction.

    “Conclusions on statistical inference remain the same with our preferred correction.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    In Argentina, less than 20 percent of the population trust other people, and only 24 percent trust the government.

    “In Argentina, less than 20 percent of the population trust other people, and only 24 percent trust the government (data from Latinobarometro (2018a)).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Latinobarometro (2018a)

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    In the last few decades, trust in governments has been in decline, and interpersonal trust is also very low.

    “In the last few decades, trust in governments has been in decline (OECD, 2018;Keefer et al., 2020), and interpersonal trust is also very low.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites OECD, 2018

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Keefer et al., 2020 source

  • May combine two claims.

    Interpretationreasoning

    The content of the vignettes matters because individuals who see the government over-perform on its commitments trust it more than those who see it under-perform.

    “This explains why the content of the vignettes also matters, why individuals that see that the government over-performs on its commitments trust it more than those that see it under-performs.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    Differences in performance matter for evaluating the government, while differences in the way the government frames communication do not.

    “The interpretation of Table A5 leads us to the same conclusions as before: differences in performance seem to matter to evaluate the government; this is not the case for differences in the way the government frames the communication.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    If individuals familiar with the program already incorporated most of this information into their priors, the treatment should have little or no effect.

    “If individuals familiar with the program already incorporated most of this information into their priors, the treatment should have little or no effect.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Premisecitation

    The use of an experimental setting could help yield more light into the existence of the link and its determinants and generate usable knowledge for practitioners.

    “The use of an experimental setting could help to yield more light into the existence of the link and its determinants and generate usable knowledge for practitioners”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Levi & Stoker, 2000 source
    • cites Van Walle & Bouckaert, 2014

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Bouwman & Grimmelikhuijsen, 2016 source

  • May combine two claims.

    Interpretationreasoning

    There is less updating with intermediate transparency than with full transparency, but more than with null transparency.

    “While there is less updating with intermediate transparency than with full transparency (the last line of the table), there is more than with null transparency.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisereasoning

    Trust requires both honesty and competence.

    “This is because trust requires both honesty and competence (the government is providing goods and services to the citizens whatever the framing, so benevolence is taken as a given).”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Benevolence is taken as a given in the trust analysis.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Voung tests suggest that the Zero Inflated Poisson model is a better fit than the Poisson estimates.

    “Voung tests suggest ZIP is a better fit than Poisson estimates.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The results of the experiment lie between separating and pooling: vignettes are not a separating signal because not all treated individuals believe the government is transparent, but not a pooling signal because the proportion believing the government is transparent rises significantly.

    “The results of the experiment lie somewhere in between: vignettes are not a separating signal, because not all individuals in the treatment group believe that the government is transparent, but it is not a pooling signal either, because the proportion of individuals that believe the government is transparent rises significantly.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The framework helps explain the results in Table 4, where positive performance has no significant effect on trust in individuals familiar with the program.

    “This framework helps to explain the results in Table 4, where positive performance has no significant effect on trust in individuals familiar with the program.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisereasoning

    Transparency is relevant for trust because it indicates that the government is honest.

    “Furthermore, transparency is relevant for trust because it indicates that the government is honest.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisenone given

    Transparency requires that the information revealed by the government be true, i.e., it requires an honest government.

    “The vignettes by themselves do not assure transparency: transparency requires that the information revealed by the government be true, i.e., it requires an honest government.17”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Trust requires that the government be seen as competent and benevolent.

    “Hence, trust also requires that the government be seen as competent and benevolent.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisecitation

    Trust is a multidimensional concept.

    “Trust is a multidimensional concept.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Grimmelikhuijsen (2012b)

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Study design and conductreasoning

    Aggregating outcomes into indices improves statistical power to detect effects consistent across specific outcomes and reduces the chance of false positives.

    “The aggregation improves statistical power to detect effects that are consistent across specific outcomes when these specific outcomes also have idiosyncratic variation.”
    • Also stated in: § 3.2. Empirical strategy · plain
      “Focusing on the indices helps us to reduce the number of statistical tests performed so as to reduce the chance of false positives”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Evidence about the connection between transparency and outcomes is still scant, particularly regarding the strength of the link between the content of the information (performance) and the framing of the information, and the particular conditions under which transparency fosters trust.

    “Evidence about the connection between transparency and outcomes is still scant (Fukuyama, 2015;Bauhr, Czibik, de Fine Licht, & Fazekas, 2019), particularly regarding the strength of the link between the content of the information (performance) and the framing of the information, and the particular conditions under which transparency fosters trust.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Fukuyama, 2015 source
    • cites Bauhr, Czibik, de Fine Licht, & Fazekas, 2019

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning

    While the posteriors of all individuals improve in the example, only those whose beliefs cross the threshold that makes the binary transparency variable jump from 0 to 1 change their views on transparency.

    “While the posteriors of all individuals improve in this example, only those individuals whose beliefs cross the threshold that makes the binary transparency variable jump from 0 to 1 change their views on transparency.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    Perceptions of trust were evaluated using a multidimensional approach including competence, benevolence, and honesty.

    “For evaluating perceptions of trust we use a multidimensional approach that includes the many components of trust: competence, benevolence, and honesty (Grimmelikhuijsen, 2012b).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Grimmelikhuijsen, 2012b

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    If politicians only report the best results out of a large sample, the value of the signal that good performance provides drops.

    “If politicians only report the best results out of a large sample, the value of the signal that good performance provides drops.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The information about government commitments and performance presented in the vignettes is accurate and credible.
    • Rests on an assumption: The government's performance relative to its commitments is a valid indicator of government competence, benevolence, and honesty.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationnone given

    If citizens learn about deception (fake news), they will not trust government-provided information any longer and the value of any posterior signal disappears.

    “As to lying (providing 'fake news'), if citizens learn about the deception, they will not trust government-provided information any longer and the value of any posterior signal disappears.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The information about government commitments and performance presented in the vignettes is accurate and credible.
    • Rests on an assumption: The government's performance relative to its commitments is a valid indicator of government competence, benevolence, and honesty.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    One would expect the priors to be smaller among the individuals exposed to this information for the first time.

    “One would also expect the priors to be smaller among the individuals exposed to this information for the first time.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    The existence of more information and positive citizen perceptions of government transparency does not necessarily go hand in hand.

    “However, the existence of more information and positive citizen perceptions of government transparency does not necessarily go hand in hand (Grimmelikhuijsen, 2012a;Grimmelikhuijsen, 2012b).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Grimmelikhuijsen, 2012a

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Grimmelikhuijsen, 2012b

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductown data

    The survey experiment randomly assigned participants to treatment/control and to vignettes.

    “Participants were assigned to randomized treatments twice during the survey”
    • Also stated in: § 1. Introduction · plain
      “First, participants were allocated at random with equal probability to receive information about a series of commitments the government had made to the citizens of Buenos Aires.”
    • Also stated in: § 1. Introduction · plain
      “Second, participants were randomly allocated to a vignette that highlighted the government pledge to efficiency and good management or to a vignette that highlighted an empathetic message”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Qualitative results do not change across the different tables.

    “Qualitative results do not change across the different tables.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisereasoning

    Information matters in shaping public perceptions and voting choices, but it cannot sit idle in a website; it must be delivered to citizens.

    “while information matters in shaping public perceptions and voting choices, that information cannot sit idle in a website but rather be delivered to citizens”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Ferraz & Finan, 2008 source
    • cites Mullainathan, Washington, & Azari, 2010 source
    • cites Keefer & Khemani, 2012

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Interpretationreasoning

    Presenting partial information (a few results out of a large set) reduces the value of the signals unless there is a clear obligation to present them without manipulation.

    “First, it shows that presenting partial information (a few results out of a large set) reduces the value of the signals unless there is a clear obligation to present them without manipulation.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    In the context of verifiable reports, manipulating information would eventually reduce trust in the government, and eliminate the informational value of transparency.

    “In the context of verifiable reports, manipulating information would eventually reduce trust in the government, and eliminate the informational value of transparency.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    Citizen's perceptions regarding why and what information the government presents, and its ability to manipulate it matter too.

    “Importantly, citizen's perceptions regarding why and what information the government presents, and its ability to manipulate it matter too.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    To increase trust it is not only necessary to provide information, but also to show good performance.

    “Again, the results show that in order to increase trust it is not only necessary to provide information, but also to show good performance.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    If citizens believe only the best results are shown, or that there may be 'fake news', the value of the signals falls and may even disappear altogether.

    “If citizens believe only the best results are shown, or that there may be 'fake news', the value of the signals falls and may even disappear altogether.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The survey items were adapted from Grimmelikhuijsen (2012b).

    “Items adapted from Grimmelikhuijsen (2012b).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Grimmelikhuijsen (2012b)

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    All regressions include comuna fixed effects and controls for age, gender, education, labor status, socio-economic level, commune, and prior knowledge of the policy.

    “All regressions include comuna fixed effects and the following controls: age, gender, education, labor status, socio-economic level, commune, and a factor variable that indicates if the participant knew about the policy before the survey.”
    Quote in the paperNot a substantial claim
48 of 48

  • The paper relies on this without support.

    The information about government commitments and performance presented in the vignettes is accurate and credible.

    “Each vignette describes the policy commitments made by the government, gives specific examples of the commitments, and shows how well the government has been doing in the implementation of the policy according to the benchmarks and objectives it had established at the beginning of the mandate.”
    “These commitments, made by the mayor when he took office, are tracked and results are publicly available on the webpage of the city government.”

  • The paper relies on this without support.

    The online survey sample is representative of the adult population of the City of Buenos Aires.

    “A total of 1,999 complete interviews were carried out by a company that specializes in collecting online survey data.”

  • The paper relies on this without support.

    Participants in the treatment group read and understood the information in the vignettes as intended.

    “The vignette presents a short and simple message to keep the attention of the participant but at the same time provides enough information for understanding the initiative.”

  • The paper relies on this without support.

    The government's performance relative to its commitments is a valid indicator of government competence, benevolence, and honesty.

    “Additionally, individuals received information showing either that the government was over-performing on its commitments or that the government was under-performing.”
    “For evaluating perceptions of trust we use a multidimensional approach that includes the many components of trust: competence, benevolence, and honesty (Grimmelikhuijsen, 2012b).”

  • The authors state this as an assumption.

    Government incumbents differ in competence and honesty.

    “In our setup, government incumbents differ in competence and honesty.18 .”

  • The authors state this as an assumption.

    Governments can either be competent (C) with probability p_C or incompetent (I) with probability p_I = 1 - p_C.

    “In regard to competence, we assume that governments can either be competent, C, with PrGnf ½C ¼ p C , or incompetent, I, with Pr½I ¼ p I ¼ 1 À p C .”

  • The authors state this as an assumption.

    Governments can either be honest (H) with probability p_H or dishonest (D) with probability p_D = 1 - p_H.

    “In regard to honesty, they can either be honest, H, with PrGnf ½H ¼ p H , or dishonest, D, with Pr½D ¼ p D ¼ 1 À p H .”

  • The authors state this as an assumption.

    Responses to the binary transparency variable can be associated with priors on honesty p_H if an individual switches from believing the government is not transparent to believing it is when priors exceed a certain threshold.

    “The responses regarding the binary variable transparency in Table 2, which takes values 0 and 1, can be associated with the priors on honesty p H in Table 5 if an individual switches from believing that the government is not transparent to believing that it is when priors exceed a certain threshold.”

  • The authors state this as an assumption.

    Priors p_H vary by individual.

    “Priors p H , which can take any value between 0 and 1, must thus vary by individual (we omit the subindices i for individuals),”

  • The authors state this as an assumption.

    Vignettes can be sent not only by honest types H with probability q_H=1, but also by some dishonest types D with probability q_D, 0<q_D<1.

    “Suppose that vignettes can be sent not only by honest types H with probability Pr½VjH ¼ q H ¼ 1, but also by some of the dishonest types D with probability Pr½VjD ¼ q D ; 0 < q D < 1.”

  • The authors state this as an assumption.

    All dishonest but competent types CD can provide vignettes (whether true or false), but dishonest and incompetent types ID cannot.

    “By reference to Table 5, we will specifically assume that all dishonest but competent types CD can provide vignettes (whether true or false), but that dishonest and incompetent types ID can't.”

  • The authors state this as an assumption.

    Benevolence is taken as a given in the trust analysis.

    “(the government is providing goods and services to the citizens whatever the framing, so benevolence is taken as a given)”

  • The authors state this as an assumption.

    Under full transparency, there are two types of policymakers in terms of competence: type C with probability p_C and type I with probability p_I=1-p_C, and p_C=p_I=0.5.

    “Under full transparency, let there be two types of policymakers (senders) in terms of competence: type C, with Pr½C ¼ p C ,Gnf and type I, with Pr½I ¼ p I ¼ 1 À p C , and let p C ¼ p I ¼ 0:5 as before.”

  • The authors state this as an assumption.

    A commitment has an outcome G or B with probabilities depending on sender type; type C has q_C and type I has q_I, where q_C > q_I.

    “Furthermore, let a 'commitment' have an outcome that can take the values G (good) and B (bad) with probabilities that depend on the type of sender.Say type C has PrGnf ½GjC ¼ q C while type I has Pr½GjI ¼ q I , where q C > q I .”

  • The authors state this as an assumption.

    Each policy commitment can be thought of as an independent draw.

    “We can think of each policy commitment as an independent draw.”

  • The authors state this as an assumption.

    Honest types CH and IH always reveal the results of the two random draws, while dishonest types DC manipulate the information to their advantage, revealing only good draws.

    “Honest types CH and IH always reveal the results of the two random draws, while dishonest types DC manipulate the information to their advantage, which in the examples below means revealing only good draws.20”

  • The authors state this as an assumption.

    Trustworthy types are identified with types CH that are both competent and honest.

    “Trustworthy types are identified with types CH that are both competent and honest.”

  • The authors state this as an assumption.

    A transparent government always shows a vignette (Pr[V|H]=1), while a non-transparent one shows it half the time (Pr[V|D]=1/2).

    “Notes: Assumption is that a transparent government always shows a vignette (Pr½VjH ¼ 1), while a non-transparent one shows it half the time (Pr½VjD ¼ 1=2).”
18 of 18

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Reported trust comparison is called significant despite a p-value of 0.150

    Major

    The first and last groups are the positive-empathy and negative-empathy treatments. Table A4 reports p = 0.150 for their trust-in-government comparison, which does not support the statement that those bars are statistically different at conventional significance levels.

  • Several reported medians are not above the midpoint

    Minor

    The perception variables are standardized from zero to one, so their midpoint is 0.50. Table A2 reports medians of exactly 0.50 for several variables, including Competence4, contradicting the claim that every median is above the midpoint.

2 of 2

Cites

Flagged first, then by citations

Research works printed but not linked: 3

  1. Chong, 2015, Does corruption information inspire the fight or quash the hope? A field experiment in mexico on voter turnout, choice, and party identification, The Journal of Politics, 77, 55, 10.1086/678766 · DOI
  2. Keefer, P., Rojas M, A. M., Scartascini, C., & Valle L, J. (2020). Trust to Advance Inclusive Growth. In: Nuguer, V., Powell, A. (Eds.), Inclusion in Times of Covid-19. Inter-American Development Bank. pp. 41–52.
  3. Nock, 2010, Survey experiments, Handbook of Survey Research, 2, 837

Cited by

Sorted by citations

Related

Authors
Alessandro, M., Cardinale Lagomarsino, B., Scartascini, C., Streb, J., & Torrealday, J.
Affiliations
  • Inter-American Development Bank (US)
  • Universidad del Centro de Estudios Macroeconómicos de Argentina (AR)
  • Universidad de Buenos Aires (AR)
Corresponding author
Scartascini, Carlos
ORCID
Scartascini, Carlos
Venue
World Development
Publisher
Elsevier BV
Published
2021-02
Volume
138
Pages
105223
ISSN
0305-750X (ISSN-L)
DOI
10.1016/j.worlddev.2020.105223
OpenAlex
W2915464736
Cited by
135
References
78
Citations by year
2026: 38 · 2025: 30 · 2024: 22 · 2023: 23 · 2022: 11 · 2021: 8 · 2020: 2
FWCI
39.18
Type
journal article
Language
en
Open access
Licence
Locations
  • World Development (publishedVersion, cc-by-nc-nd)
  • RePEc: Research Papers in Economics (submittedVersion)
Topics
  • Social Sciences › Political Science and International Relations › E-Government and Public Services
  • Social Sciences › Public Administration › Public Policy and Administration Research
  • Social Sciences › Sociology and Political Science › Corruption and Economic Development
Keywords
government transparency, public trust in government, information disclosure, public perception, government performance, promise fulfillment, disclosure requirements, Argentina

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-06.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-06.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review