Strategic Participation in Protests: Evidence from Women´s Day in Chile

2023-03-07

2023-03-08

Given the chosen dimension(s) of heterogeneity, we have the following specific objectives:
1. Determine the degree of substitutability or strategic complementarity at a global level. That is, to study whether a greater total participation decreases or increases incentives to participate in the march.
2. Determine the degree of strategic substitutability or complementarity with one´s own group.
3. Determine the degree of strategic substitutability or complementarity with other different groups.
4. Depending on the number of participants achieved in the study, ordered dimensions of heterogeneity could be defined. In this case, it is intended to understand:
a. What is the distance between two people (on the ideological scale or income level, as appropriate) from which the degree of substitutability or complementarity reverses?
b. How does this distance vary across groups?

The methodology of this investigation revolves around an online survey that will be implemented in 3 waves. This survey will be sent to all female undergraduate students of two Chilean universities: University of Chile and University Diego Portales. The survey consists of: a first wave where student´s basic information and intentions to participate in the march is collected; a second wave where information on expected assistance from different groups is delivered; and a third wave where information on final participation in the march is collected. Participation in the survey will be encouraged with prizes that will be raffled among the participants.

The first wave, implemented between March 1st and 5th, captures general information about the participants, such as sociodemographic characteristics, past participation in the movement and marches, feminism, sexism, ideological position, religion, and beliefs about expected attendance at the upcoming march for International Women´s Day.

In the second wave (March 7th), the previously gathered information is used to inform about the estimated attendance of different groups and/or the total, depending on the assigned treatment. Once the treatment is received, questions are asked about the intention to participate and subsequent beliefs about attending the march.

Finally, in the third and final wave, which occurs after the protest (March 9-12), we measure effective participation in the march and the reasons/reflections of attending.

The most relevant instance within this methodology is the intervention carried out in the second wave. For each dimension of heterogeneity, two groups of relevant information will (potentially) be provided: (i) information on overall expected attendance at the march, (ii) information on attendance relative to one´s own group. In this way, each person can receive a combination of information about global assistance, and assistance from some group to which they belong or not.

The previously described experimental design is directly related to the specific objectives of this investigation. In particular, the information provided within the second wave intervention will be used to measure the effects of changes in the expectations of attendance of different groups on the intention to attend the march. This will allow us to measure and identify the degree of strategic substitutability or complementarity with different groups in the decision to protest.

Randomization will be done automatically by the survey server (Qualtrics) once participants enter the second wave of the survey. The randomization is programmed to assign treatments in a roughly equal manner across all respondents within the randomization unit.

The randomization unit is by groups (defined by the dimension of heterogeneity chosen) and universities. For example, if the dimension chosen is political ideology, then randomization will be done within each group and university: Right and UCHILE, Center and UCHILE, Left and UCHILE, Right and UDP, Center and UDP, Left and UDP.

Yes

As described above, the unit of randomization is by universities and groups. There are two universities that are participating in the experiment and 3 groups for each dimension. Therefore, the methodology involves 6 clusters per dimension.

According to tuition figures of the year 2021, the University of Chile has 34.526 undergraduate students, where 18.010 students (52.2%) declare to be women. On the other hand, the University Diego Portales has 12.651 undergraduate students, where if around 50% declare to be women then there would be 6.325 students that would be part of the target audience of this survey. Considering an average response rate for online student surveys of around 10%, then there would be around 2.433 participants/observations.

Considering the planned number of observations described above and both the randomization units and number of treatments the sample size would be: 2.433(number of participants) ÷ 6 (number of randomization units) ÷ 8 (number of possible treatments) = aprox. 50 people within each treatment arm and cluster.