What is gerrymandering, and how do researchers measure it?

Gerrymandering is the intentional manipulation of district boundaries to unfairly advantage a party or group. Researchers measure it by looking at how efficiently each party's votes turn into seats. Common measures include the efficiency gap, partisan bias, the mean-median difference and declination. No single number is accepted as the final word, and the sources note that some measures work better than others depending on how competitive a state is. [1] [2]

The term was coined in 1812 after Elbridge Gerry, then governor of Massachusetts, signed a bill that created a salamander-shaped district to advantage the Democratic-Republican Party. States set criteria such as contiguity, compactness and preservation of communities of interest to limit the practice. According to the MIT Election Lab, a strong incentive to gerrymander remains where state legislators or political commissions oversee redistricting. Some states also require competitiveness or partisan fairness. [1]

The two most common methods are packing and cracking. Packing crowds as many voters of one party into as few districts as possible, so they win fewer seats overall. Cracking spreads a party's voters across many districts so they form a majority in few or none. Either way, a party can gain an advantage even when all districts hold roughly the same number of voters. [1]

The efficiency gap is built on the idea of wasted votes. A vote is wasted if it is cast for a losing candidate, or if it is an extra vote for a winner beyond the 50 percent threshold needed to win. The gap takes one party's wasted votes, subtracts the other's, and divides by the total votes cast. Packing and cracking both produce wasted votes, which is why the measure aims to capture their combined effect in one number. [3] [4] [2]

Quanta Magazine gives a worked example with ten districts. One party wastes 450 votes and the other 50, a difference of 400, or 40 percent of all voters. That percentage can be read as the share of seats the favored party won beyond what it would receive in a balanced plan with a gap of zero. Quanta also notes that gaps tend to fluctuate over the decade a map is in force, so a legal challenge would need to show the gap is likely to favor the same party throughout. [3]

PlanScore describes three further measures. Partisan bias is the difference between a party's seat share and 50 percent in a hypothetical, perfectly tied statewide election. The mean-median difference compares a party's median and mean vote share across districts; a large gap suggests the distribution is skewed toward one party. Declination treats asymmetry around the win/loss threshold as a sign that lines were drawn with that threshold in mind. [2]

PlanScore says all four metrics are reliable when a state is competitive, but only the efficiency gap and declination should be trusted when one party dominates a state. The Princeton Gerrymandering Project's report card uses further indicators, including packed wins, which compares the average winning percentages of each party's victories. Quanta adds that the Supreme Court's 2006 ruling hinted it might favor measures reflecting partisan symmetry, meaning each party has an equal opportunity to convert votes into seats. [2] [5] [3]