Facebook’s new Oversight Board is considering whether the company was justified in indefinitely suspending Donald Trump from its platform. The question is important, but it would be a mistake for the board to answer it right now, or on Facebook’s terms. To do so would effectively absolve the company of responsibility for its part in creating the circumstances that made Mr. Trump’s speech—both online and offline—so dangerous.

Facebook announced plans for its board in 2018 in response to concerns from civil society organizations and regulators about the company’s influence over public discourse online. Sometimes described as Facebook’s Supreme Court, the board comprises an impressive group of civic leaders, free speech experts and scholars from around the world. But Facebook narrowly limited the board’s jurisdiction, having it focus almost exclusively on questions relating to the removal of specific pieces of content.

Content moderation decisions can be consequential, of course. But Facebook shapes public discourse more profoundly through its decisions about the design of its platform. Its ranking algorithms determine which content appears at the top of users’ news feeds. Its decisions about what types of content can be shared, and how, help determine which ideas gain traction. Its policies and tools relating to political advertising determine which kinds of users see which political ads, and whether those ads can be countered by ads offering different viewpoints and correcting misinformation.

This creates a problem for the board. It’s not just that the board’s jurisdiction is too narrow. Nor is it merely that the elaborate quasi-judicial structure that Facebook has established for review of its content-moderation decisions draws public attention away from the design decisions that matter more—though this is certainly the case.

The fundamental problem is that many of the content-moderation decisions the board has been charged with reviewing can’t actually be separated from the design decisions that Facebook has placed off limits. Content-moderation decisions are momentous, but they are as momentous as they are because of Facebook’s engineering decisions and other choices that determine which speech proliferates on the platform, how quickly it spreads, which users see it, and in what context they see it. The board has effectively been directed to take the architecture of Facebook’s platform as a given. It shouldn’t accept that framing, and neither should anyone else.

The Trump case starkly highlights the problem with the board’s jurisdiction. Mr. Trump’s statements on and off social media in the days leading up to the Capitol siege on Jan. 6 were certainly inflammatory and dangerous, but part of what made them so dangerous is that, for months before that day, many Facebook users had been exposed to staggering amounts of sensational misinformation about the election, shunted into echo chambers by Facebook’s algorithms, and insulated from counterarguments by Facebook’s architecture.

This is why it would be a mistake for the board to address the question that Facebook has asked it to answer, at least right now. Doing so would draw public attention away from the platform design decisions that warrant the most scrutiny, and from the regulatory interventions that are needed to better align Facebook’s practices with the public interest. It would also let Facebook off the hook for business practices that cause significant harm to democracy.

No doubt the board’s members are thoughtful people who are working in good faith to protect free expression online. (One of the board’s co-chairs was a visiting scholar at the Knight First Amendment Institute, where we work, and other board members have contributed to Knight Institute projects.) At least in this case, though, there’s a real risk that the board is being used as a fig leaf for Facebook’s failures.

Fortunately, the board has another option, as we and our colleagues told the board last week in response to its call for public comments on the case.

Rather than answer the question that Facebook has posed, the board should advise the company to commission an independent investigation into the ways in which the design of its platform may have contributed to the events of Jan. 6. The investigation, which should be conducted by a team that includes engineers, should assess how the platform’s architecture affected what content users encountered on the platform and in what contexts they encountered it. It should also assess the impact of the steps Facebook took to enforce its policies relating to militarized social movements. The board should answer the question about Mr. Trump’s suspension only after Facebook has commissioned this study and published it.

Facebook’s oversight board has been received with extreme skepticism in some quarters and cautious optimism in others. The Trump case presents it with an early and critical test, and a singular opportunity to assert its independence and establish its value. The board should seize this opportunity. It should assess the Trump case not on Facebook’s terms, but on its own.