Grober and Schram - Neighborhood Information Exchange

Why do people go vote? A common explanation among others is social embededness. This paper focuses on the influence of a voter's social environmen on the decision to votes, using rational choice models. First, the article focuses on social interaction between individuals in small groups (modeling neighborhoods). Second, it considers one specific element in the interaction by focusing on the exchange of information.

there are situations where preferences or vote decisions are explicityly disclosed. People may also implicity reveal their vote intention. People tend to be pretty good about estimating whether or not their neighbors vote and for whom, but people do have a choice whether or not to reveal whether or not to reveal their intention.

Previous studies about embededness seem to demonstrate its importance, but previous studies have a hard time uncovering the structure of social relationships. Lab experiments allow one to focus precisely on the relationship between voters individually that we hypothesize might influence participation.

These authors use experiments to better understand the "neighborhood information exchange."

The NIE game has several stages:

(1) Players have roles as early or late voters. Early voters can either vote or postpone their decision. Late voters can observe what early voters do, but do not know whether a postponement leads to abstention or a late vote. This allows researchers to study strategic use of the turnout decision by late voters.

(2) Players are assigned as allies or adversaires. Whether your adversaries are turning out or not makes a difference on whether you vote.

(3) Players' preferences can stay constant across election or can be varied. With fixed preferences, aggregate behavior is more predictable, which may decrease the value of observing the neighbor's decision as compred to the case where preferences vary.

The NIE Participatoin Game and Experimental Design
The game design and execution is complicated, but basically it's just the previous three points.

Hypotheses

H1 - turnout is higher when neighbors are adversaries than when allies

H2 - When neighbors are allies, senders participate at a higher rate than receivers. (when you know you ally neighbor is going to vote, you don't need to)

H3 - Senders participate at higher rates at stage 2 than at stage 1.

H4 - Receivers participate more after observing abstention than after observing a vote.

H5 - After observing a vote, receivers are more likely to participate if the neighbor is an adversary than in case of an ally.

Experimental Results
There are a number of results and the hypotheses are variously accepted or rejected.

Result 1: Neighborhood information exchange increases turnout.

R2 - The stability of group composition does not affect turnout.

R3 - When information is exchanged, turnout is highest amongst allies and lowest when neighbors do not know each other's preferences.

R4 - Senders participate at a higher rate than receivers.

R5 - Senders attempt to influence their neighbor. If the reciever is an ally, senders mainly vote at stage 1. if the receiver is an adversary, senders participate more at stage 2.

R6 - Receivers participate at a higher rate in parnters than in strangers.

R7 - Recievers reciprocate allied sneders' stage 1 decisions in strangers.

H1 - rejected

H2 - supported

H3 - rejected for allies and supported for adversaries

H4 - rejected

H5 - rejected

Interpreting the Results
NIE substantially increases overall participation. Interaction within neighborhoods hasa strong effect per se.

The Nash Equilibria are often rejected. But, authors say that implicity intragroup coordination at the neighborhood and group level of parnters nad only at the neighborhood level in strangers can explain most of the findings. Senders are important coordinators.

Conclusion
Authors has isolated the environmental element of information about others' behavior. NIE seems to generate higher participation, and even higher turnout when neighbors are allies and can establsih bonds across rounds. There appears to be positive influence of segregation on turnout. Senders strategically use thier first mover position o influence recievers. There isn't much empirical support for the Nash Equilibria.

Authors think an interesting research field could be opened up by systematically varyign the structure of neighborhoods and teh content of information exchanged.

Inormation exchange serves as a coordination device with strong aggregate effects.