Perceived sustainable minimum wages – a test of randomization and question scales across countries

We implement a survey experiment in the 2022 wave of the OeNB Euro Survey. The OeNB Euro Survey is a repeated cross-sectional survey conducted annually in ten different countries. The aim of the survey experiment is threefold. First, we test the feasibility of different randomization approaches across countries and whether falsification or errors in implementation regarding randomization might be a concern. Second, we examine data falsification in terms of individual responses. If there is evidence of data falsification, these effects how interviewers or survey institutes behave, would influence and/or blur treatment effects in respondents’ behavior. Third, we test a standard assumption in survey research regarding scale transformation. It is common in surveys when asking respondents to report wages to give them the choice which reporting unit to use, i.e. wage per hour, month or year. If these questions are used in research projects, numbers are then often set to the same scale. In our treatment, we vary on which scale respondents are supposed to report a specific wage. This experiment allows us to analyze if the assumption of, e.g., linear transformation based on full-time working hours, is valid.

External Link(s)

Registration Citation

Citation

Beckmann, Elisabeth and Melanie Koch. 2022. "Perceived sustainable minimum wages – a test of randomization and question scales across countries." AEA RCT Registry. November 17. https://doi.org/10.1257/rct.10398-1.0

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

See pre-analysis plan

Intervention Start Date

2022-09-27

Intervention End Date

2022-11-10

Primary Outcomes

Primary Outcomes (end points)

min_wage_hr: Perceived minimum amount people living in the same town/village should approximately earn per hour after taxes
min_wage_mth: Perceived minimum amount people living in the same town/village should approximately earn per month after taxes

Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)

min_wage_yr: Perceived minimum amount people living in the same town/village should approximately earn per year after taxes

Secondary Outcomes (explanation)

min_wage_yr is only asked in three of ten countries (Bulgaria, Romania and Serbia) if respondents answer 'don't know' on min_wage_hr.

Experimental Design

We randomize at which scale respondents of the survey have to answer what, in their personal opinion, is the minimum amount people living in their town/village should approximately earn. One treatment group is asked about the amount per hour, the other group about the amount per month. The question is asked after several questions about personal finances are answered, right before respondents answer further questions on socio-demographic information. Survey contractors in each country were told to randomize over respondents but to stratify over interviewers, so that we can control for interviewer effects.

Experimental Design Details

Randomization Method

Randomization methods deliberately differ across countries:
Randomization (alternating order of treatments) for all CAPIs is computerized but sometimes require data entry from the interviewer
Randomization (alternating order of treatments) for all PAPIs is done by hand by the survey contractors
The different approaches are outlined in the pre-analysis plan.

Randomization Unit

Randomization by survey respondent, stratified by interviewer

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

The treatment is not clustered but it is conducted in ten different countries and we will mostly look at these countries separately. In each country, 1,000 interviews are conducted, i.e., 500 interviews per treatment group.

Sample size: planned number of observations

In total, around 10,000: at least 1,000 respondents in each country

Sample size (or number of clusters) by treatment arms

In total, around 5,000: at least 500 respondents per treatment in each country

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

See pre-analysis plan

Supporting Documents and Materials

IRB