Misconceptions about regional inequality and preferences for targeted spending

Respondents complete a questionnaire on support for targeted and non-targeted policy. Each subject is randomly assigned to one of four experimental groups, and completes five consecutive stages. In stage 1, a random group of respondents is asked to estimate how many students in their county leave school without a degree. Another group of respondents is asked to estimate the unemployment rate in their county. Respondents are matched to their county of residence via their postcode, which is elicited earlier in the survey. In stage 2, respondents are asked to guess how many of the other 400 counties in Germany have higher shares of the respective measure than their county. In stage 3, a randomly selected group receives information on the shares in their county and across Germany and then state whether they perceive regional inequality as a problem while another group answers the same question without additional information. In stage 4, respondents record their preferences for whether targeted and non-targeted spending should be increased. In stage 5, respondents rate the effectiveness of targeted and non-targeted spending to reduce the share of school-leavers without a degree and the unemployment rate, respectively. The treatment variation allows us to study to what extent preferences for targeted policy depend on accurate information on regional inequality.

Subjects will conduct the experiment via an online platform. All wordings presented below are translations of the original German question wording. Subjects will be randomly assigned to one of four treatment groups, and each subject takes decisions in five consecutive stages.

TREATMENT 1:

Stage 1: prior beliefs on share of school-leavers without a degree in county

[Question wording:]
“What is your best guess, what is the share of students that leave school without a degree in your county?”

[Answer categories:]
Open field + “share of a cohort”

Stage 2: prior beliefs on the number of counties in Germany with a higher share of school-leavers without a degree

[Question wording]:
“What is your best guess, how many of the 400 other counties in Germany have a higher share of students leaving school without a degree than your county? ‘0’ means that your county has the highest share of students without a degree; ‘400’ means that your county has the smallest share.”

[Answer categories:]
Open field

Stage 3: perception of regional inequality

[Question wording:]
“In your opinion, is inequality in education outcomes across counties in Germany a serious problem?”

[Answer categories:]
“No problem at all/a small problem/somewhat of a problem/a serious problem/a very serious problem”

Stage 4: preferences for increased spending

[Question wording:]
“Do you support or oppose the state markedly increasing spending for schools to reduce the share of school-leavers without a degree? Please remember that the additional spending often has to be financed through taxes.”

[Items:]
Higher spending for schools in all counties
Higher spending for schools in counties with high rates of school-leavers without a degree

[Answer categories:]
“strongly support/somewhat support/somewhat oppose/strongly oppose/neither support or oppose”

Stage 5: perceived effectiveness of spending

[Question wording:]
“And how suitable do you think increased spending for schools is to reduce the share of school-leavers without a degree?”

[Items:]
Higher spending for schools in all counties
Higher spending for schools in counties with high rates of school-leavers without a degree

[Answer categories:]
“very suitable/somewhat suitable/somewhat not suitable/very not suitable/neither suitable nor not suitable”

TREATMENT 2:

Identical to Treatment 1, except that the following was added:

In stage 3: “In your county the [value] percent of students leave school without a degree. In the other 400 counties in Germany [value] have a higher share than your county.” + Map of Germany that shows counties with higher shares in darker colors

In stage 4 and 5: “In your county the [value] percent of students leave school without a degree. In the other 400 counties in Germany [value] have a higher share than your county.”

TREATMENT 3:

Stage 1: prior beliefs on share of unemployed in county

[Question wording:]
“What is your best guess, what is the share of unemployed (the share of unemployed among all persons that are available to the labor market) in your county?”

[Answer categories:]
Open field + “percent”

Stage 2: prior beliefs on the number of counties in Germany with a higher share of unemployment

[Question wording]:
“What is your best guess, how many of the 400 other counties in Germany have a higher rate of unemployment than your county? ‘0’ means that your county has the highest unemployment rate; ‘400’ means that your county has the lowest rate.”

[Answer categories:]
Open field

Stage 3: perception of regional inequality

[Question wording:]
“In your opinion, is inequality in unemployment rates across counties in Germany a serious problem?”

[Answer categories:]
“No problem at all/a small problem/somewhat of a problem/a serious problem/a very serious problem”

Stage 4: preferences for increased spending

[Question wording:]
“Do you support or oppose the state markedly increasing spending to reduce the unemployment rate? Please remember that the additional spending often has to be financed through taxes.”

[Items:]
Higher spending in all counties
Higher spending in counties with high unemployment rates

[Answer categories:]
“strongly support/somewhat support/somewhat oppose/strongly oppose/neither support or oppose”

Stage 5: perceived effectiveness of spending

[Question wording:]
“And how suitable do you think increased spending is to reduce the unemployment rate?”

[Items:]
Higher spending in all counties
Higher spending in counties with high unemployment rates

[Answer categories:]
“very suitable/somewhat suitable/somewhat not suitable/very not suitable/neither suitable nor not suitable”

TREATMENT 4:

Identical to Treatment 3, except that the following was added:

In stage 3: “In your county the unemployment rate is [value] percent. In the other 400 counties in Germany [value] have a higher unemployment rate than your county.” + Map of Germany that shows counties with higher unemployment rates in darker colors

In stage 4 and 5: “In your county the unemployment rate is [value] percent. In the other 400 counties in Germany [value] have a higher unemployment rate than your county.”

2020-06-03

2020-06-17

Our primary outcomes of interest are respondents’ perception of regional inequality as a problem (elicited in stage 3) and preferences for increased spending (elicited in stage 4).

The experimental design will allow us to test whether information on regional inequality across counties in Germany changes respondents’ perception of regional inequality as a problem and preferences for targeted and non-targeted spending to benefit school-leavers without a degree and the unemployed.

Misconceptions of respondents related to the share of school-leavers without a degree/unemployment rates in their county (elicited in stage 1) as well as their counties rank across counties in Germany (elicited in stage 2). Also, the perceived effectiveness of increased spending to reduce the share of school-leavers without a degree/unemployed (elicited in stage 5). In addition, heterogeneities by prior beliefs elicited in stage 1 and stage 2.

The effects on perceptions of regional inequality as a problem and preferences for increased spending (see Primary Outcomes) might be mediated through respondents’ prior beliefs on regional inequality in Germany and their perception of the effectiveness of higher spending.

We conduct the experiment in a sample of 10,000 adults aged between 18 and 69 years. The survey is conducted in cooperation with respondi. The recruitment and polling is managed by respondi, who collect the data via an online platform. That is, our participants answer the survey questions autonomously on their own digital devices. Randomization is carried out by respondi at the individual level, using a computer.

Our experiment is structured as follows:
Treatment 1:
Stage 1: prior beliefs on share of school-leavers without a degree in county
Stage 2: prior beliefs on the number of counties in Germany with a higher share of school-leavers without a degree
Stage 3: perception of regional inequality with respect to school-leavers without a degree
Stage 4: preferences for increased spending for all schools/schools in counties with high shares of school-leavers without a degree
Stage 5: perceived effectiveness of spending for all schools/schools in counties with high shares of school-leavers without a degree

Treatment 2:
Stage 1: prior beliefs on share of school-leavers without a degree in county
Stage 2: prior beliefs on the number of counties in Germany with a higher share of school-leavers without a degree
Stage 3: information on share of school-leavers without a degree + perception of regional inequality with respect to school-leavers without a degree
Stage 4: information on share of school-leavers without a degree + preferences for increased spending for all schools/schools in counties with high shares of school-leavers without a degree
Stage 5: information on share of school-leavers without a degree + perceived effectiveness of spending for all schools/schools in counties with high shares of school-leavers without a degree

Treatment 3:
Stage 1: prior beliefs on unemployment rate in county
Stage 2: prior beliefs on the number of counties in Germany with a higher unemployment rate
Stage 3: perception of regional inequality with respect to unemployment rates
Stage 4: preferences for increased spending for all counties/counties with high unemployment rates
Stage 5: perceived effectiveness of spending for all counties/counties with high unemployment rates

Treatment 4:
Stage 1: prior beliefs on unemployment rate in county
Stage 2: prior beliefs on the number of counties in Germany with a higher unemployment rate
Stage 3: information on unemployment rates + perception of regional inequality with respect to unemployment rates
Stage 4: information on unemployment rates + preferences for increased spending for all counties/counties with high unemployment rates
Stage 5: information on unemployment rates + perceived effectiveness of spending for all counties/counties with high unemployment rates

Randomization is carried out by the survey company respondi, using a computer.

at the individual level

No

10,000

10,000 adults aged 18-69 years

approx. 2,500 will be assigned to each of the Treatment groups