Less than two years ago, a Ferguson grand jury decided not to return an indictment in the shooting death of Michael Brown. The grand jury announced their decision on the evening of November 24, 2014. There had been about 70 hours of testimony and over 60 witnesses. (NPR has a helpful blog covering all the action.) Remembering what happened in Ferguson may help us to understand what may not happen in Baton Rouge and St. Anthony: an indictment or conviction of the officers who killed Alton Sterling and Philando Castile.

In line with the adage that a determined prosecutor can indict a ham sandwich, the St. Louis County prosecutor’s office was criticized for its presentation of the evidence in Darren Wilson’s case, especially given its willingness to indict Ferguson protesters. The vote of the grand jury is secret, but transcripts of evidence were released. Commentators observed that the grand jury clearly found Wilson’s testimony reliable, despite troubling inconsistencies within it.

In his testimony, Wilson says that he fired the second shot from the car and missed Brown (page 226). He knows this, he says, because Michael Brown began running away and a “cloud of dust” was kicked up behind him (page 232). The “defense-of-life” standard developed in Tennessee v. Garner (1985) states that you cannot use deadly force except to defend your own life or another’s. Brown was unarmed, although Wilson testifies that he considered even a fleeing Brown to pose a danger to other officers (page 281). Later, Wilson testifies that Brown was approaching the car when he, Wilson, fired the second shot (page 260). Later still (pages 263 and 264), he says that he shot without looking and didn’t know where Brown was. So, according to the officer’s own testimony, Brown was either disengaging, or approaching the vehicle, or Wilson didn’t know what Brown was doing.

This sort of ambiguity is why outcomes in the legal process are so difficult to predict, and why the Sterling and Castile cases may prove inconclusive. In my experience observing criminal cases in open court, the officers who testify come across as relatively persuasive and competent because they are precise. They use compass points, stick to a story, and are experienced at answering questions. Members of the general public do not and are not.

This is where the Department of Justice enters the picture. They investigate intentional violations of civil rights, which can include structural problems with law enforcement and the judicial process as well as the presence of bias in individuals. Their findings in Ferguson, MO, where the criminal and judicial process was used to extract fines from low-income residents, are a promising start, and build on other consent decrees (“settlements”) in other jurisdictions, such as Albuquerque. (The complete list of DOJ investigations into law enforcement agencies can be found here.) But federal convictions of individual officers for civil rights violations require intent — willfully depriving someone of their constitutional rights under cover of law — which is a high standard for prosecutors to meet.

Within the Department of Justice, the Bureau of Justice Statistics also plays a leading role in collecting and analyzing data about police-civilian interactions. To buttress governmental efforts to track police violence, non-governmental agencies such as UCLA’s The Justice Database use data to clarify the who, what, where, and how of police shootings. These agencies, along with crowdsourced databases, play a crucial role in holding law enforcement accountable, but there is still no overarching data collection on police shootings.

We can make some generalizations from what we know. According to a standard textbook on criminal justice, the likely victim of police brutality is male, young, lower class, and from a visible minority. The likely victim in deadly force cases is unarmed, black, male, 17-30 years of age, in public, and detained or encountered in connection with an armed robbery. A 2007 DOJ study found that only .00005 of all arrests end in death. However, it is extremely disturbing that the paradigmatic victim of a deadly police shooting is unarmed and black.

The key problem, as writers at The Atlantic and The New York Times have argued, is that there are too many of these encounters with African-Americans, especially when indignation is perceived as non-compliance, as in this story from St. Paul, MN, and — as in Castile’s case — even apparent cooperation may be life-threatening.

Changing or reducing police-civilian encounters might require major changes to existing statutes, such as decriminalizing marijuana. However, Philando Castile was pulled over by police over 30 times since 2002, a pattern that indicates police profiling.

Increasing federal inquiry into the practices of the nearly 18,000 law enforcement agencies in the US is also important. The aim of federal involvement is not only to improve procedural justice, but to gain insight into effective policing across jurisdictions. Alton Sterling had a long history of troubled relations with the police. Baker Police Department officers recorded in an April 2000 affidavit that Sterling lay down in a parking lot “and told officers to go ahead and beat him down because regardless of the outcome, he was going to have [the officers] fired.” One of the roles of the DOJ and FBI is to investigate whether this type of distrust is the product of police misconduct, or an aggravating factor in deteriorating community-police relations.

In his 1982 book, James Fyfe argues that a single shooting “can unravel years of good police work,” “expose a police agency to great civil liability,” “torment an officer who acted within the vague rules defined for him or her,” and “can be prevented by the simple expedient of directing officers not to use their guns in such nonthreatening circumstances.” President Obama’s task force on 21st century policing concluded in 2015 that trust is essential to ensuring good police-community relations. Not enough was done between 1982 and 2015. Americans of all racial backgrounds and geographic localities should examine why encounters with police so often result in arrest or death in the case of African-Americans. Black lives matter, and different modes of enforcing the laws can lead to black lives mattering more than they do now.