close× Call Us
Title Minimum Wage Increases and Attitudes
Post date 01/10/2019
C1 Background and Explanation of Rationale

We have some preliminary results (based on administrative data) that suggest that people whose wages go up due to minimum-wage increases may also become more likely to vote. There are several possible mechanisms for this: for one, more income could make it feasible for people to vote by removing barriers like housing instability and financial stress. But there could also be “interpretive effects” of such a policy change, where getting a raise because of government policy changes people’s attitudes about the role of government in their lives. We want to use a survey to explore these possible mechanisms, while taking advantage of a planned minimum wage increase in Massachusetts.

On January 1, 2019, the Massachusetts minimum wage will increase from $11/hour to $12/hour for most workers. We plan to use a multi-wave survey (as well as an embedded experiment) to test how this minimum wage increase shapes people's economic well-being as well as their political efficacy. We ran the first (baseline) wave of the survey in late December, and will launch the second wave in early January to collect follow-up measures.

C2 What are the hypotheses to be tested?

We will test the hypotheses that Massachusetts’ 2019 minimum wage increase:
1) improves the economic situation of minimum-wage workers
2) increases political efficacy among minimum-wage workers

Note that we do not have a mediation setup here: we are simply attempting to measure two different pathways by which minimum wage increases could drive turnout and whether there is evidence that each of them is occurring, not trying to distinguish which matters more for turnout. See the attached pre-analysis plan for more detail on our expectations.

There is also a small experimental manipulation embedded in the second wave of the survey: some people will be randomly assigned to see a question that reminds them of the minimum wage increase early in the survey, while others will see it at the end. We will use this treatment to test a secondary hypothesis about how salient the minimum wage increase is for affected workers. We anticipate that:

3) being primed with a question about the minimum wage increase will not change people’s answers to the questions that make up our main outcome measures,

since we anticipate the wage increase will already be relatively salient for minimum wage workers. But the PAP discusses what we will do if the primed group seems to behave differently from the unprimed group (basically, we will check whether our main diff-in-diff results depend on the primed group, and if so, we will suspect that minimum wage increases will have a major political impact only if people are reminded of them right before elections, perhaps by politicians’ statements).

C3 How will these hypotheses be tested? *

We detail four main hypothesis tests in the attached pre-analysis plan.

Briefly, we have two main outcome measures, each an index measure constructed from several survey questions: we label these our “political efficacy” and “economic pessimism” measures, and their construction is described in the PAP. For each of these two outcomes, we run two different tests, each of them a difference-in-differences setup.

For political efficacy, for example, we want to compare the over-time changes (before-after minimum wage increase) in political efficacy among people who are and are not affected by the minimum wage increase. We set up this simple difference-in-differences design (in regression terms, predicting the outcome measure with a dummy for whether the person is part of the group affected by the wage increase, a dummy for whether it is after the wage increase takes place, and the interaction between the two) using two different comparison groups. First, we compare our affected group (people in Massachusetts who earn between $11 and $12/hour before the minimum wage increase) to other people in the same state who are not affected by the minimum wage increase, because they already earned more than the new minimum wage (people in Massachusetts who earn between $12 and $15/hour). But because we worry that this comparison could understate the effect of the policy on minimum-wage workers’ attitudes if above-minimum-wage workers also notice and draw political conclusions from the wage increase, we then do a second test with a different comparison group. This time, we set up the same diff-in-diff, comparing affected workers in MA to low-wage workers in neighboring Connecticut, which did not increase its minimum wage in 2019.

We then run the same two diff-in-diff analyses (comparing across earnings and across geography) for our other outcome measure, economic pessimism. These four tests constitute our main analyses, testing our two main hypotheses.

As noted above, there is also a small experimental manipulation embedded in the second wave of the survey. We will check whether the location of the priming question within the survey changed respondents’ answers by calculating the difference in means between people who saw the question early in the survey and those who saw it at the end of the survey (focusing on people affected by the minimum wage increase, so people in Massachusetts making under $12/hour at baseline). If the priming question affects people’s responses to our main outcome measures, then we will conduct additional analyses (described in detail in the PAP) to see whether our diff-in-diff results seem to depend on the responses of people who received the priming question early in the survey.

C4 Country United States
C5 Scale (# of Units) 1200
C6 Was a power analysis conducted prior to data collection? Yes
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number 1811588507
C9 Date of IRB Approval 11/14/2018
C10 Will the intervention be implemented by the researcher or a third party? Researchers
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research? not provided by authors
C12 If relevant, is there an advance agreement with the implementation group that all results can be published? not provided by authors
C13 JEL Classification(s) not provided by authors