COMPUTING CHANGE DISTRIC ELETORAL - A Computation-Intensive...

Info iconThis preview shows pages 1–4. Sign up to view the full content.

View Full Document Right Arrow Icon
A Computation-Intensive Method for Evaluating Intent in Redistricting Micah Altman Harvard University [email protected] Michael McDonald George Mason University [email protected] Prepared for the 2004 Midwest Political Science Association Conference Chicago, IL April 14-18, 2004. Please do not quote without permission
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
A Computation-Intensive Method for Evaluating Intent in Redistricting Shall I count them pure with the wicked balances, and with the bag of deceitful weights? - Micah 6:11 Intent and Effect in Redistricting Legislative intent and effect are intimately intertwined in redistricting lawsuits and academic research. Yet intent is impossible to measure directly and legislative statements of intent are often absent, ambiguous, or misleading. Academics and expert witnesses have offered a number of empirical approaches, such as searching for 'divergent boundary segments', attempting to create 'random' comparison plans and using estimated 'bias' and responsiveness to infer intent. In this article, we show that most of the methods currently in use are statistically biased, and the remainder fail to capture predominant intent. Redistricting is one of the few legal arenas where political scientists and statisticians play starring roles as expert witnesses. These experts analyze voting behavior and election outcomes to determine effect of redistricting plans. As statistical computing has increased, more methodologically rigorous procedures have been devised to estimate effects. Judging by the comments in the recent oral arguments before the Supreme Court in Vieth v. Jubelirer , these experts may have become victims of their own success at constructing complicated methods that cannot be easily grasped by lawyers and judges. The Justices wondered if any definitive standard for partisan gerrymandering could be constructed that would not devolve into a battle of opposing expert witnesses. By putting the problem of determining intent into a formal statistical framework, we also show even the few non-biased methods in use fail as tests of predominant intent
Background image of page 2
2 because they ignore relevant competing explanations. Each relies on examining criteria that are, at best, merely correlated with gerrymandering, without capturing the constraints that a human cartographer must face in creating a particular plan. Are there easy statistical answers to assessing intent of redistricting? An emerging method is to compare a given redistricting plan to hypothetical alternatives and apply classical statistical tests. We find these methods lack valid statistical grounding. The problem posed to these methods is to model the distribution of unobserved redistricting plans. Some approaches simplistically assume a normal distribution when there is no evidence supporting such a claim, while more sophisticated methods devise sampling schemes to construct the distribution of feasible redistricting plans. We show that it is computationally intractable to draw unbiased random samples from the global
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 4
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 05/12/2010 for the course APPLIED ST 2010 taught by Professor Various during the Spring '10 term at Universidad Nacional Agraria La Molina.

Page1 / 38

COMPUTING CHANGE DISTRIC ELETORAL - A Computation-Intensive...

This preview shows document pages 1 - 4. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online