The Justices, the freedom of speech, and ideology

Three noted political scientists (Profs. Lee Epstein, Jeffrey Segal & Christopher Parker, whom I'll call Epstein et al.) recently posted a study (see also here) saying that many Supreme Court Justices' votes in free speech cases are strongly correlated to the ideology of the speech or the speaker. "[T]he votes of both liberal and conservative Justices tend to reflect their preferences toward the ideological grouping of the speaker, and not solely an underlying taste for (or against) the First Amendment."

"We [report] the percentage of votes each Justice cast supporting free speech when the speaker is liberal and when the speaker is conservative." "The probability of Scalia voting in favor of a liberal speaker is about .22; for conservative speakers, it's .62." "The Supreme Court (and its members) can appear more or less supportive of free expression depending on whether it decides cases with speakers left or right of center." (Note that the authors state that their coding of liberal vs. conservative speakers "focuses on the ideological content of the specific speech, rather than the broader partisan or ideological beliefs of the speaker.")

The study got extensive coverage in various places, including in the New York Times. Unsurprisingly, the coverage characterized the study in similar ways. "In cases raising First Amendment claims, a new study found, Justice Scalia voted to uphold the free speech rights of conservative speakers at more than triple the rate of liberal ones."

But looking closely at the study—aided by a detailed response posted by Prof. Todd Pettys and a reply posted by Epstein et al. (including Appendix C)—suggests that the study doesn't really support these assertions. Let me discuss this in a bit of detail.

1. To begin with, how can one measure whether speakers are "left or right of center" when the case involves speech and speakers that aren't particularly ideological, or that are ideological in ways that are hard to see as "liberal" or "conservative"—or involves a bipartisan (or multi-ideological) coalition of claimants?

One possible answer would be to exclude these cases from the analysis, by coding them as "neither." (Indeed, when the study looks at the ideological valence of the government action being challenged, as opposed to the ideological valence of the speech, it has a "no direction" option.) But this is not the approach that the study took; instead, they labeled all cases as involving a "conservative speaker" or a "liberal speaker." Thus, just to give some examples (and this isn't meant as an exhaustive list):

1. Child pornography distributors were coded as "liberal" speakers (e.g., X-Citement Video and Williams).
2. Some speakers engaged in personal injury lawyer advertising were coded as "conservative" speaker (e.g., Ohralik), though others weren't (e.g., Went for It).
3. A newspaper accidentally publishing the name of a rape victim was coded as "conservative," though the fact at issue was hard to see as "liberal" or "conservative" (Florida Star.)
4. Speakers in libel cases involving basically nonpolitical allegations—such as Ulysses Tory, who was picketing Johnnie Cochran's offices, or Philadelphia Newspapers, which had alleged that the owner of a chain of convenience stores had links to organized crime and was corruptly influencing government officials (specifically naming a Democratic state legislator)—were generally labeled "liberal" (Tory, Hepps).
5. Likewise, Dun & Bradstreet, the credit rating agency, was coded as a "liberal speaker" in a case involving a false statement that a small business had declared bankruptcy (Greenmoss Builders).
6. People and organizations challenging the Copyright Term Extension Act, as to works that had no clear overall ideological valence, were labeled as "liberal" speakers (Eldred).
7. A dance club for 18-to-20-year-olds that wanted to allow patrons 21 and above was labeled as a "liberal" speaker (Stanglin.)
8. The violent video game case was labeled as involving a "conservative" speaker (Entertainment Merchants Ass'n).
9. The dogfighting video case was likewise labeled as involving a "conservative" speaker (Stevens).
10. The near-funeral picketing case was labeled as involving a "conservative" speaker/speech, presumably because the Westboro Baptist Church was opposed to homosexuality—but the church was picketing the funeral of a straight soldier, with placards saying (among other things) "God Hates the USA/Thank God for 9/11" and "Thank God for Dead Soldiers"; it's hard to see this sort of absurdity as either conservative or liberal. (Snyder.)
11. Ralph Forbes, a third party candidate who was anti-abortion, and had been a Nazi and a David Duke state campaign manager, was labeled as a "liberal" speaker (Arkansas Educational Television).
12. Turner Broadcasting System, which carries a vast range of speech, and which was run by prominent liberal Ted Turner, and which has CNN as its most prominent politics-related channel, was coded as "conservative" (Turner Broadcasting).
13. Jehovah's Witness door-to-door proselytizers were labeled as "liberal," though the Jehovah's Witnesses are anti-homosexuality, anti-abortion, and generally patriarchal (Watchtower Bible & Tract Society).
14. A company seeking to distribute the identities of arrestees, gathered from police reports, and doing so apparently just for financial purposes, was coded as "liberal," on the theory that the company "releases the information to lawyers, drug counselors and insurance companies. To the extent that they are releasing information to groups that could provide support to arrested individuals, coded as liberal speech" (United Reporting).
15. Keen Umbehr, who was protesting supposed local government waste and abuse (and, for whatever it's worth, has run for office as a Republican and, more recently, Libertarian) was coded as a "liberal" speaker, in a case involving government retaliation against Umbehr in government contracting (Umbehr).
16. Todd Mitchell, a black teenager was prosecuted under a hate crimes law for organizing a mob of teenagers to beat up whites after watching the movie "Mississippi Burning," was coded as conservative (Mitchell).
17. Many election law cases in which Republicans, Democrats, and even other parties were litigating on the same side were coded as "conservative" speaker/speech (e.g., Washington State Grange), though at least one was labeled "liberal" (Jones).

2. Now one possible response—indeed, the one seemingly offered by Epstein et al.—is that these cases were coded based on whether the law stemmed from a generally conservative or liberal ideological policy (or, relatedly, whether the legal challenge to the law was associated with a conservative or liberal view of free speech). Thus, for instance, Epstein et al.'s explanation for the coding of Wisconsin v. Mitchell was, "Hate crime laws are liberal laws, so speaking out against them is coded as conservative speech." (By "speaking out" here, I take it the authors mean "litigating against," since Mitchell was prosecuted for violating the hate crime law, not for speaking out against it.) Likewise, for the rape victim name case, the authors write, "[k]eeping [the identity of a victim of sexual assault] private is more of a liberal interest, so the speech is coded as conservative"; I'm not sure whether that's an accurate summary of political views now, or in 1989, but in any event it focuses on the ideology behind the law, not on the ideology of the speaker.

The coding of the violent video game case as involving a "conservative speaker" was justified on the grounds that "restricting (the business of) violent video games in the name of protecting children … is more frequently supported by liberals and derided by conservatives as an example of the nanny state." The various cases involving bipartisan challenges to election laws were often coded as "conservative," Epstein et al. say, because "a challenge motivated to bring about greater inclusion in the political process is liberal regardless of the challenger's partisan label." And challenges to election laws by third-party speakers, such as Forbes, were coded as liberal because "greater inclusion in the political process is coded as liberal regardless of the partisanship of the speaker."

3. But when the claim is that conservative Justices are more willing to vote to uphold laws that reflect conservative legal principles (or accept free speech challenges that conservatives tend to like), that's a very different claim from the claim that "toward the ideological grouping of the speaker" (emphasis added).

Indeed, this revised claim in part simply reflects what we mean by "conservative" and "liberal." For instance—and I speak here just of general tendencies, from which some on each side dissent—historically pornography bans have been seen as part of a "conservative" policy agenda. Conservative Justice Scalia has long argued in favor of the constitutionality of such bans, and many conservatives (though not all) have agreed with him in some measure. Liberal Justice Douglas had long argued against the constitutionality of such bans, and many liberals (though not all) have agreed with him in some measure. But that simply reflects that, when we say a Justice is "conservative," we generally mean that the Justice tends to endorse a conservative view of the law, of the Constitution, and of the judicial role, and similarly for liberals.

Likewise, liberal Justices tend to favor (again, not always, but generally) restrictions on the spending of money to speak about election campaigns, especially when the spending is by nonmedia business corporations and unions. That too just reflects the fact that such restrictions stem from a view of constitutional rights and government interests that is more likely to be held by liberals.

If we could find evidence that liberal Justices reject arguments in favor of protecting campaign spending because those arguments tend to be brought by conservative speakers—or that liberal Justices more generally support restricting campaign spending because they think Republicans are more likely to benefit from such spending than Democrats—that would be a pretty serious criticism. Justices' free speech decisions are supposed to generally be viewpoint-neutral, which is to say neutral with regard to the viewpoint of the speakers or of the speech.

But if liberal Justices simply reject arguments in favor of protecting campaign spending because they believe that it's permissible to restrict such speech in the name of equality, or of preventing indirect corruption, that just reflects those Justices' understanding of the proper scope of free speech law. It is an understanding that is generally endorsed by the liberal political movement and also by liberal Justices. Indeed, it's one of the many distinctions that, when aggregated, tell us which Justices are "liberal" and which are "conservative."

To be sure, one could argue for a view of free speech that, for instance, protects all or nearly all speech, or protects no or nearly no speech. In that case, one's positions would be unrelated or largely unrelated to whether a speech restriction is generally seen as part of a liberal legislative agenda or as part of a conservative political agenda.

But all Justices that have considered free speech cases have endorsed some sorts of speech restrictions—whether based on judgments about history and tradition, judgments about what speech is so harmful that it should be restricted, judgments about the government's powers in its managerial roles, such as employer or landowner, or something else. Their views understandably line up (to some degree) with what liberals and conservatives believe counts as worthy tradition, as a sufficiently serious harm, or the proper scope of the government's power as manager. Indeed, sometimes long-serving and prominent Justices' views on such questions may themselves help define what we view as a "liberal" or "conservative" speech-related legislative agenda.

4. So the Epstein et al. study makes a claim about the Justices, and especially the conservative Justices, that could yield a pretty serious condemnation, if it's accurate (and if others build on it by controlling for possible confounding factors, such as various differences in the laws involved in the various challenges). The claim would suggest that the Justices are discriminating among speakers based on the viewpoint of the speakers' speech, something the Justices repeatedly say should not be done. I suspect the study made the news precisely because of the seriousness of this claim, which media accounts repeated.

But actually the study doesn't classify cases just based on the ideology of the speech (or of the speaker). In many cases, the study looks instead to what the authors see as the ideology behind the law, or behind the legal theory of the challenge. That means that the study would yield a very different conclusion: Conservative Justices tend to be persuaded by conservative arguments (not necessarily conservative speakers' arguments) for why a law should be upheld (or struck down), and likewise for liberal Justices.

Conservative Justices are more persuaded than liberals by arguments for restricting pornography (generally seen as conservative arguments). Liberal Justices are more persuaded than conservatives by arguments for restricting spending in election campaigns (generally seen as liberal arguments). An interesting conclusion, but not a terribly surprising one, or one that justifies the criticism of the Justices to which the study has led.

5. Before anything else, Justices are humans. It's perfectly plausible that the human tendency to support those with whom one sympathizes sometimes affects Justices, too, even in ways that violate the requirements of viewpoint neutrality in First Amendment judging.

A well-constructed study can shed light on this question, though of course it would also need to control for principled factors that distinguish different categories of cases. (As I noted, for instance, the fact that liberal Justices tend to support restrictions on spending money for campaign-related speech doesn't mean that they are discriminating based on speaker viewpoint, even though such cases are predominantly brought by conservative speakers. They may have principled viewpoint-neutral reasons for their votes; likewise, in the opposite direction, for conservative Justices.) Such a well-constructed study might thus strongly suggest that, indeed, "the votes of both liberal and conservative Justices tend to reflect their preferences toward the ideological grouping of the speaker," and—if the facts support it—conclude that this tendency is greater for some Justices than for others.

But the Epstein et al. study does not, I think, support its conclusion, precisely because it classifies speakers as "liberal speakers" or "conservative speakers" for reasons other than the "ideological grouping" of the speaker or the speech. And once the study fails to classify based on actual "ideological grouping of the speaker," it can't tell us whether the Justices' votes tend to reflect such an ideological grouping.