Skip to main content

Welcome to a Grasp lesson

Create a free account to keep studying and create your own study path.

Lesson 6

Assessing Partisan Symmetry in Seats-Votes Curves

Recommended resources

Last lesson, the efficiency gap turned district results into a measure of unequal wasted votes. Under its equal-turnout shortcut, a zero gap calls for a larger seat bonus than proportional representation does. Partisan symmetry asks a different question: would either party receive the same share of seats if it received the same share of votes?

This lesson uses a seats-votes curve to make that comparison. The distinction matters when reading an election article: a map can treat the parties symmetrically without awarding seats in proportion to votes, and one election result alone cannot show how a map would behave at other vote levels.


Reading a seats-votes curve

A seats-votes curve plots one party’s statewide vote share on the horizontal axis and its seat share on the vertical axis. Let mean the share of seats party A would win if it received a share of the statewide two-party vote. A point at , for example, means 60% of votes and 70% of seats for A. It does not say whether the map treats A and B equally; that requires a comparison with the outcome if B received 60% of votes.

Redistricting and Representation

Read the American Academy of Arts and Sciences article for its definition of partisan symmetry and its explanation of why a single election result is not a complete seats-votes curve.

In the article’s discussion beginning “Partisan symmetry is a principle for districting plans,” read through the curve comparison. This includes the explanation of the axes, uniform partisan swing, Figure 2, and the paragraph describing ways to spot asymmetry. Focus on the comparison between parties at the same vote share, rather than on the particular election depicted in Figure 2.

In a two-party setting, when B has 60% of votes, A has 40%. At that point A wins of the seats, leaving B with . Thus the two outcomes to compare at a 60% vote share are

A symmetric curve gives them equal seat shares. More generally, the symmetry test is

This is a comparison across two possible elections, not a demand that A and B win equal seats in the same election. The curve is symmetric about the point where each party has half the votes and half the seats. At 50% of votes, symmetry therefore calls for 50% of seats in the idealized setting. But passing that one check does not establish symmetry across the curve.



Three symmetric curves, three different seat allocations

The supplied graph shows three stylized rules. Read each line as the seat share awarded to one party at a given vote share.

The horizontal axis shows a party’s vote share; the vertical axis shows its seat share. The dashed proportional rule, curved quadratic rule, and solid efficiency-gap ideal all treat the two parties symmetrically, but award different seat shares away from a 50–50 vote split.

At 60% of votes, the proportional line awards approximately 60% of seats. The quadratic curve awards about 69%, and the solid efficiency-gap line awards 70%. Now compare each with what the other party would receive at 60% of votes. On this graph, the corresponding seat shares are again about 60%, 69%, and 70%, respectively. Each curve passes the symmetry test, even though the curves disagree about the appropriate reward for a vote majority.

The proportional rule makes seat share equal vote share throughout. The other two give the majority party a seat bonus. For the solid line, the middle portion follows the zero-efficiency-gap benchmark from last lesson, under the equal-turnout, 50%-threshold assumptions:

That gives 70% of seats at 60% of votes and 30% of seats at 40% of votes. At the extremes, the graph flattens at zero or all seats because seat shares cannot fall below 0% or exceed 100%; those flat portions should not be read as maintaining a zero efficiency gap.

Symmetry is about equal treatment, not the size of the seat bonus. All three pictured curves are symmetric. Only the dashed line is proportional.



Detecting asymmetry—even when the center looks fair

Consider a different supplied curve with these readings:

A’s vote share A’s seat share
40% 25%
50% 50%
60% 65%

The 50–50 result looks balanced. At a 60% vote share, however, A would receive 65% of seats, while B would receive of seats: B’s 60% vote share corresponds to the row where A has 40%. The map therefore favors B by 10 percentage points of seats at this vote level, despite its balanced result at the center.

For any supplied curve, the assessment follows the same method: select a vote share , read A’s seat share there, then read A’s seat share at the complementary vote share . Subtract the latter from all seats to find B’s outcome at the same vote share . Repeat at other relevant vote levels; agreement at just one level does not settle whether the curve is symmetric overall. This also keeps partisan bias at a 50–50 vote split distinct from broader symmetry across a range of vote shares.



What a curve can—and cannot—establish

An actual election supplies one point, not the outcomes of elections that did not happen. Analysts often estimate the rest of the curve using uniform partisan swing: they adjust each district’s party vote share by the same number of percentage points, then count which districts would change hands. Seats change in jumps as districts flip, so a curve generated directly this way can have steps rather than the smooth shapes in the supplied illustration.

Exploring the Seats-Votes Curve

Read the methodology in Exploring the Seats-Votes Curve to see how district results can be extended into a hypothetical curve, and why seat shares change in steps.

In “Methodology,” under “How I generate the seats-votes curve,” read steps 1–4, stopping after the sentence about a district changing party hands. Begin with step 1’s actual-results example; in step 4, stop before the discussion of smoothing the curve with random variation. Track the difference between the observed election point and the modeled outcomes at other vote shares.

That distinction sets a limit on the verdict. A supplied curve lets you assess how the map is estimated to treat the parties under the stated voting assumptions. Its counterfactual portions are not observed election results. And even a clear asymmetry describes an outcome; by itself, it does not establish that a map drawer deliberately produced it. The role of residential geography and alternative feasible maps comes later in the course.



Conclusion

To assess partisan symmetry, compare the parties’ seat shares at the same vote share. On a curve drawn for A, B’s seat share at vote share is . Equal results across the relevant range indicate symmetry; a balanced result at 50% alone is insufficient. Symmetry also does not require proportionality: the three supplied ideals treat the parties equally while giving different seat bonuses.

Next, the focus shifts from evaluating maps to who draws and approves them: legislatures and independent commissions, and the constraints each must follow.