Here's a passage from Doe v. Reed, the opinion I criticized below. I don't think the court needed to reach this issue, because I think there's no need for strict scrutiny when state law chooses to disclose the names of people who take legally operative steps such as signing initiative, referendum, or recall petitions. But I want to set that aside here, and focus on the argument the court made -- an argument with a structure that I often hear (especially from law students), and that bears a bit of explicit analysis:

In the alternative, Defendants assert that there exists a second "compelling" interest in favor of disclosure. Defendants argue that the electorate is entitled to know "who is essentially lobbying for their vote, and thus, who likely will benefit from the measure." But this argument is unavailing because neither the Court nor the parties have the ability to identify whether an individual who supports referral of a referendum to the next ensuing general election actually supports the content of the referendum or whether that individual simply agrees that the referendum should be placed before the voting public. In other words, the identity of the person who supports the referral of a referendum is irrelevant to the voter as the voting public must consider the content of the referendum and be entitled to a process by which it can ensure that the petitions are free from fraud.

Note how the court argues here: (1) A person could sign a petition simply because he "agrees that the referendum should be placed before the voting public," and not because he "supports the content of the referendum." (2) "In other words," the identity of the petition signers "is irrelevant" to the voter who seeks to figure out "who is essentially lobbying for" the proposal.

But statement 2 is not at all just "other words" for statement 1. Statement 1 correctly asserts that signing a petition is not dispositive evidence of the signer's views on the merits. Some people do sign petitions because they think the public is entitled to vote on the question, even though they'll vote against the proposal; some people might expect the proposal to lose, and might sign the petition precisely because they want a statewide vote in which the proposal loses; some people might have deeper political reasons for signing, perhaps because they think the proposal's presence on the ballot will bring out voters who will vote the right way on some other measures or in candidate races; some people might just have been buffaloed into signing something; and so on.

But this hardly means that the identity of the signers is irrelevant to determining who is lobbying for the proposal. If, for instance, it turns out that 60% of a petition's signers are registered Democrats and 10% are registered Republicans, in a state that's generally split 50-50, voters can reasonably conclude that the petition is probably a project of Democrats, and would thus be more likely serve the interests of traditionally Democratic interest groups -- relevant information to many voters, especially if there's a dispute on what the law's effects likely will be.

Now if the argument were simply that the connection isn't strong enough to pass muster under strict scrutiny (the test that the court is applying here), that would be plausible. Likewise, if the argument were that the interest in getting such information about the initiative's backers wasn't compelling enough to satisfy strict scrutiny, that would be plausible, too. (My disagreement with the court on the bottom line stems from my view that strict scrutiny shouldn't apply here.) But this leap from (1) the signature being an imperfect proxy for backing the merits of the petition to (2) its being irrelevant to deciding who is backing the merits of the petition is not sound.