Experimental Design Details
The study uses a between-subjects survey experimental design.
The survey is carried out among a sample of Danish residents (age 18+). The survey data are collected by YouGov Denmark. The respondents are recruited using quota sampling (for gender, age, and geographical location).
The survey respondents are presented with an item asking them to think about the local governments’ delivery of public services within a specified policy area (with random assignment of the policy area as either “schooling”, “health”, or “unemployment”). Next, the respondents are asked to state the extent to which they think that the local governments should prioritize increased attention and money on doing something about particular issues. The respondents are presented with (and asked to provide individual responses to) eight particular issues. The eight issues are:
• Quality of the delivered welfare service
• Efficiency of the delivered welfare service
• Equal treatment irrespective of gender
• Equal treatment irrespective of skin color and ethnicity
• Green transition and climate considerations in the daily operations
• Equal treatment of identical cases/citizens/clients
• Ethical and legitimate behavior among the public employees
• Local [spatial] proximity in the delivery of welfare to citizens outside of the larger cities.
The order of the eight issues is counterbalanced in the survey. For each issue, responses are captured on a 7-point scale (anchored at 1 = “Not at all” and at 7 = “To a very high extent”).
Relative to a control group, a treatment group is treated to an additional textbox asking the respondents to consider that the local governments’ resources are scarce. The local governments function and operate within an economic budget framework. Therefore, prioritization of one or more particular issues will often involve less attention and resources for doing something about other particular issues.
Respondents are randomly assigned to either the control group or the treatment group (i.e., with equal probability; 50/50 chance).
HYPOTHESES
Based on existing theory and research, the study theorizes that citizens’ attitudes about the local governments’ prioritizing of increased attention and money on combating race and gender discrimination and promoting the climate agenda are subjected to so-called “fiscal illusions.” The following hypotheses are derived and empirically tested:
H1: Cost awareness moderates citizens’ support for local governments’ prioritizing of increased attention and money on …
a) combating race discrimination
b) combating gender discrimination
c) promoting the climate agenda
In particular, citizens’ support for a/b/c is expectedly lower when the citizens are made aware (vs. not made aware) that prioritizing of one or more particular issue may involve less attention and resources for doing something about other particular issues; comes at a cost.
H2: Cost awareness moderates citizens’ support for local governments’ prioritizing of increased attention and money on …
a) combating race discrimination relative to promoting the quality of the delivered welfare services
b) combating gender discrimination relative to promoting the quality of the delivered welfare services
c) promoting the climate agenda relative to promoting the quality of the delivered welfare services
In particular, the extent to which citizens’ support for “promoting quality” differs from their support for (a) combating race discrimination, (b) combating gender discrimination, and (c) promoting the climate agenda (i.e., the difference in support for “promoting quality” vs. for a/b/c) is expectedly larger when the citizens are made aware (vs. not made aware) that prioritizing of one or more particular issue may involve less attention and resources for doing something about other particular issues; comes at a cost.
H3: Identical expectations as for H2—with the exception that the text “promoting the quality of the delivered welfare services” is replaced with the text “promoting the efficiency of the delivered welfare services.”
ADDITIONAL DESCRIPTIVE ANALYSES
Moreover, the study uses three sets of additional survey items (non-experimental; from the same survey as the one containing the survey experiment). For all three sets of items, the respondents are asked to think about the local governments’ delivery of public services within the same specified policy area as for the experimental item (i.e., either “schooling”, “health”, or “unemployment”). Similarly, all three sets of items involve the same eight particular issues as for the experimental item.
First, respondents are asked about their perceptions of the extent to which particular issues are subjected to substantial problems and shortcomings (responses for each particular issue; responses captured on a 7-point Likert scale). The responses are used to estimate whether the issues of “Equal treatment irrespective of gender,” “Equal treatment irrespective of skin color and ethnicity,” and “Green transition and climate considerations in the daily operations” are perceived as subjected to substantial problems and shortcomings to the same extent as the other particular issues are.
Second, respondents are asked about their willingness-to-pay for increased local government spending on particular issues (responses for each particular issue; the respondents state how much more they would be willing to pay in taxes (in DKK) in the tax year 2022).
Third, respondents are asked about the perceptions of the most salient particular issues that the local governments should prioritize addressing (respondent state the three particular issues that they think should be given first, second, and third priority, respectively).
The item sets on “willingness-to-pay” and “most salient issues” provide descriptive evidence that may (or may not) support the results of the experimental testing of H1 but especially of H2 and H3. Both sets of items are priming cost awareness to some extent. Thus, lower willingness-to-pay and lower priority for a-c relative to “promoting quality” and to “promoting efficiency,” respectively, is expected.
For the purpose of the survey length, the respondent are exposed only to the “willingness-to-pay” items or the “most salient issues” items. Random assignment with equal probability (i.e., 50/50 chance).
Finally, the survey contains four separate items serving as indicators for descriptive attention and manipulation checks.