Waseda Institute for Advanced Study (WIAS)Waseda University


Distorted issue space and people’s choice  SONG Jaehyun, Assistant Professor

Two models that explain issue voting behavior

The importance of research on issue voting behavior has become even more prominent at present, with the increase in the number of independent voters without partisanship.
As for the model to explain voting behavior based on issues, the “proximity model” and “directional model” have been proposed, and debates as to which one is superior are currently taking place. My research is on integrating the two models and on the behavior of political actors (voters and parliamentarians) using the integrated model.

The difference between the proximity model and the directional model

In the proximity model, it is assumed that the voters cast their vote for the political party or candidate that is deemed to be closest to them on the issue space. In the case illustrated in Figure 1, the voter would cast their vote for Candidate A, who is closest to the voter.
On the other hand, in the directional model, the political party, candidates, and voter are positioned in the issue space, but, unlike the proximity model, there is a “Status Quo” point. The status quo clarifies which political parties are “on their side” or “on the opposite side” of the voter in regard to the issue, and the voter would cast their vote for the political parties that are on “their side.” In addition to providing the distinction between “on their side” and “on the opposite side,” the directional model predicts that the voter would cast their vote for the political party with the clearer stance; in other words, the party that is more extreme. In the case illustrated in Figure 1, using 0 as the status quo, the voter would cast their vote for Candidate B, who is deemed to be on the same side as the voter. As we have seen, the two models predict different outcomes under the same circumstances.

Figure 1

A new proposal: the “elastic proximity model”

Following these, I have developed a new “elastic proximity model,” focusing on the issue space at the level of voter perception. The issue space envisaged in the elastic proximity model is a space in which the greater one’s distance from the status quo, the shorter the distance between points becomes. It indicates that when one is near the status quo, the side of it that the voter is on as well as the difference between adjoint points feels large, but when one is far away from the status quo, the difference between adjoint points becomes less salient. For example, in Figure 2, whenω is 1, the distance between all adjoint points is the same; however, if ω were 0.9, the distance between the points would shrink as it moved away from the status quo (i.e., 1, 0.9, 0.81, 0.73…).

Figure 2

The new “elastic proximity model” that I propose predicts that the voter would cast their vote for the political party that is deemed to be closest to the voter, as in the case of the proximity model; however, it argues that the perception of issue space is not objective but elastic.
As this model predicts the voter would choose the political party/candidate that is closest to them (as it proposes that the voter perceives the party/candidate closer if it is on the same side as them), it contains elements of both the proximity and directional models.

Empirical testing via a survey experiment

In order to prove the effectiveness of the new model, a survey experiment is conducted to empirically test whether the issue space is in fact elastic at the voter perception level and whether it has higher explanatory power than the conventional models (for details, see: https://dx.doi.org/10.2139/ssrn.2984430).
If the issue space becomes elastic at the voter perception level, as shown in Figure 2, ω becomes smaller than 1 (if ω=1, the issue space does not become elastic and the outcome is similar to that of the proximity model). The result of analysis of experimental data has shown that for many issues, ω is smaller than 1.
For example, the elasticity of the issue space regarding revision of the constitution is about 0.897. This means that when the distance from 0 to 1 is set at 1, that from 6 to 7 (or from -6 to -7) becomes about 0.521 (in other words, it shrinks to half). However, it is one thing for there to be a distortion in the issue space at the voter perception level, and quite another for the elastic proximity model to explain the reality more appropriately. Thus, I have compared the elastic proximity model not only with the existing proximity and directional models but also with other proposed models that integrate the two. The result has shown that the elastic proximity models performs better than existing models with regards to three goodness-of-fit indicators.

What distortion in the issue space indicates

What outcome does distortion in the issue space bring to the real world? This is my next research agenda.
There are still arguments to be made regarding the rationality of the voter, but my view is that “while the voter is rational, a variety of factors undermine their rationality.” One of these factors is distortion in the issue space. A proximity model is a rational voting model. However, if the issue space is distorted, even if the voter casts their vote in accordance with rational procedures, the outcome is not necessarily rational.
One example is the recent polarization being witnessed across the world. We can see this in the National Front of France and Donald Trump’s election to the United States of America presidential election. Voters’ issue attitudes are normally distributed in reference to most issues. If the proximity model is correct, all political parties would converge at the center in order to maximize the number of votes to win. If the directional model is correct, peoples’ attitudes to issues and ideology taken up by the political parties would be polarized. However, in the real world, it is rare that major parties converge on the same policies or become polarized to opposite ends. One of my future challenges is to explain this phenomenon using the distortion in issue space at the voter perception level.
How Democracies Die (by Steven Levitsky and Daniel Ziblatt), which has recently attracted a lot of attention, argues that political polarization is one of the factors that would lead to the collapse of democracy. There are many approaches and theories to explain such polarization, and “rationality that produces irrational outcomes” could be one of the answers. Developing a method to correct such distortion can contribute to democracy in the future.

Interview and Composition:Ayako Yamamoto
In cooperation with: Waseda University Graduate School of Political Science J-School


Page Top
WASEDA University

The Waseda University official website
<<https://www.waseda.jp/inst/wias/en/>> doesn't support your system.

Please update to the newest version of your browser and try again.


Suporrted Browser