America is a land of daunting spatial differences. Per capita income in San Jose, Calif., is more than three times per capita income in McAllen, Texas. The proportion of adults with college degrees is five times larger in the Bethesda, Md., region than it is around Dalton, Ga. According to Pew surveys taken between 1987 and 2003, 56% of adults in Mississippi agree that “AIDS is God’s punishment for immoral sexual behavior.” Fewer than 20% of Connecticut respondents agree. Seventy-three percent of respondents from South Carolina agree that “the best way to ensure peace is through military strength.” Sixty percent of Vermont residents don’t agree.

These differences in beliefs, income, and lifestyles also lead to political heterogeneity — the famous map of red states and blue states, and in “The Big Sort” (Houghton Mifflin, 384 pages, $25), Bill Bishop, a longtime chronicler of American politics, has written an insightful book on the tendency of Americans to live in like-minded political clusters.

The map of red states and blue states requires us to understand why different places have different people and how political parties choose their core issues. Why are coastal Americans generally richer, better-educated, and more socially liberal? Why have our parties increasingly divided on religious issues, like abortion, rather than on economics?

Mr. Bishop’s analysis of spatial sorting elegantly mixes the work of economist Charles Tiebout and the psychologist Solomon Asch. In the 1950s, Tiebout argued that people choose their communities to fit their interests, and Asch showed that peers influence our opinions even on simple things like the length of a line. First we make our communities, then our communities make us. No wonder people who move to Vermont start to doubt the defensive value of carrying a big stick. Mr. Bishop appropriately fears that politically homogeneous communities veer off into extremes of intolerance.

Mr. Bishop then tries to make sense of why Republicans and Democrats have increasingly divided on social rather than economic issues. This divide runs counter to one of the most famous principles in political theory: Two parties should rush to the center to compete for the median voter. Yet this prediction is sharply at odds with the post-1976 tendency of Republican and Democratic platforms to diverge significantly on issues like abortion.

Political extremism caters to a party’s more radical members. Sometimes, as in 1964 or 1972, catering to extremes is the politically costly but inevitable result of the nomination process. In other cases, extremism can be a smart strategy to get out the vote, since people are more willing to go to the polls for a choice than for an echo. But if everyone hears a candidate’s extreme message, then moving away from the middle energizes the opposition’s base as much as it does one’s own. Strategic extremism can only win votes when parties can target their messages. For this reason, party divisions follow the ability to connect quietly to groups, like unions or churches. The economic populism of traditional Democrats reflected their ties to labor unions, and the ability of modern Republicans to connect to church groups pulls them to the right on religion.

Much of Mr. Bishop’s analysis is spot-on, but he has the journalist’s tendency to overemphasize change at the expense of continuity. Yes, America has a lot of sorting, but we have always had a lot of heterogeneity. Segregation by education across areas has risen as highly educated people have moved toward more educated places, but racial segregation has fallen substantially. Any gulf in attitudes between red states and blue states today is dwarfed by the gulf in racial attitudes 50 years ago.

My own reading of the data is that between 40 and 50% of the electoral votes have regularly been in battleground states between 1940 and today. The period between 1896 and 1936 was the real era of “The Big Sort,” when fewer than 30% of electoral votes were in battleground states.

The rise in religious voting isn’t particularly new either. A century ago, the connection between Catholicism and the Democratic Party was every bit as strong as the connection between religion and the Republican Party today. (I published these facts in the Journal of Economic Perspectives.)

Mr. Bishop’s figures seem to show rising political extremism at the state level, but he has selected his states on the basis of current polarization, and Statistics 101 tells you that standard randomness means that extreme places today are likely to have been less extreme in the past. While I don’t think that states have gotten more polarized, Mr. Bishop’s data on counties seem to show a greater increase in polarization at that lower level of aggregation. Even though counties don’t directly pick presidents, the increase in ideological uniformity at the local level gives credence to Mr. Bishop’s fears that too many people are listening only to neighbors who share their own views.

“The Big Sort” provides us with an excellent snapshot of America’s political landscape in the early years of the 21st century, but it is less clear whether the book will seem all that relevant next November. Still, even as parties and states shift, America’s spatial heterogeneity will remain and political parties will continue to take advantage of that heterogeneity.

Mr. Glaeser is the Glimp Professor of Economics at Harvard, director of the Taubman Center for State and Local Government, and a senior fellow at the Manhattan Institute.