Gauging the effects of transparency on citizen decision-making: Insights from a survey experiment

Using a series of experiments we intend to demonstrate how the structure of government information being publicly disclosed, rather than the actual content, affects citizens’ ability to make decisions that align with their... [ view full abstract ]

Using a series of experiments we intend to demonstrate how the structure of government information being publicly disclosed, rather than the actual content, affects citizens’ ability to make decisions that align with their stated preferences. To operationalize structure, we look to research on attribute partitioning. Attribute partitioning explains how the grouping of attributes pertaining to a certain object (e.g. policy) affects the utility an individual assigns to said object. Higher attribute partitioning divides an object into a greater number of constitutive attributes, and is therefore synonymous with greater transparency, whereas lower attribute partitioning divides an object into a few attributes and can therefore be seen as relatively less transparent.

Using the ideas put forth by attribute partitioning, this study intends to demonstrate how different ways of structuring government information will result in two distinct relationships between transparency and accountability – one where transparency bolsters citizens’ ability to hold their government accountable and a second where transparency detracts from citizens’ ability to hold their government accountable.

For our experiment, we present 1,000 respondents with a hypothetical scenario. Their municipality is considering modifying a controversial policy and must increase taxes. Before raising taxes, city council is asking citizens their opinion first. Respondents will then be presented with a prime to shift them into either an ‘agree’ or ‘disagree’ category. This prime is delivered in order to ensure that the disagree and agree categories are (roughly) comparable. Following this, we break the proposed policy down into constitutive policy attributes that receive funding – one group will be a high partition group and a second a low partition group. The high partition group will represent the high transparency group and the low partition group will represent the low transparency group. We then make clear the amount currently being spent from the policy budget for each attribute. Following this, each respondent is asked the extent to which they wish to increase funding for the policy, decrease funding, or maintain current levels. In keeping with previous attribute partitioning literature, we predict that the high transparency group will consistently allocate greater funding, in aggregate to the proposed policy when compared to the low transparency group.







Agree with the initiative Disagree with the initiative
More funding Less funding More funding Less funding
High partitioning of attributes
Low partitioning of attributed