In thinking about the differences between children who succeed in school and those who struggle, we tend to focus on two environments: the schools where they learn, and the homes they go to at night. We talk about the course offerings in a classroom and the number of books on the shelf at home. We consider the economic makeup of their classmates and the marital status of their moms at home.

It's likely true, though, that the environments kids navigate between home and school -- their neighborhoods -- matter an awful lot, too. And if we want to address poor academic achievement in high-poverty schools, we probably need to think as well about what happens in the high-poverty neighborhoods that surround them (no, this does not mean that families and teachers don't matter, but that so much more does, too).

This point is driven home by a novel new study from Patrick Sharkey, Amy Ellen Schwartz and Ingrid Gould Ellen at NYU, and Johanna Lacoe at the University of Southern California. They looked at remarkably pinpointed data on the locations of violent crime in New York City, the addresses of public-school children who live on those blocks, and their scores on standardized tests taken within a week of such violence.

The main finding: Neighborhood violence appeared to reduce the passing rates of black children by about 3 percentage points on language arts assessments when that violence occurred on their block in the seven days before the test, compared to children whose streets witnessed violence in the seven days after the test. That result also controls for differences in the children's past test performance.

The effect sounds small. But the design of the study behind it gets us closer to proving an important theory about the significance of where children live for how they learn.

"There are shocks like this that occur on a regular basis in kid’s residential environments," Sharkey says. "It’s hard to demonstrate tangibly that these shocks are brought with them into the school setting when they sit down to take this kind of test. What this [study] does is provide concrete evidence of a causal effect of the exposure to this type of shock."

The study does not imply that these children actually saw such violence with their own eyes, but rather that the ripple effects of violence -- the sound of sirens, the sight of police tape, the worry that accompanies both -- permeate a neighborhood in a way that stays with children for at least a few days.

"When violence is in the air, when the threat of violence is in the air, then it becomes something that spills over to effect not just people who are involved," Sharkey says, "but everyone who lives in the community."

The research, which focused on thousands of elementary- and middle-school students between 2004 and 2010, found no comparable effect of violence on math scores. This reinforces previous findings that violence and stress have a greater effect on the cognitive abilities and performance of children particularly on reading and vocabulary assessments.

To arrive at all of this, Sharkey and his coauthors exploited the randomness of violence itself. The vast majority of students in the study attend high-poverty schools and live in high-poverty neighborhoods, and so the point wasn't to compare children living in violent places to those living in safe ones. It was to look at how bouts of violence might influence children at school who otherwise have much in common other than the random timing of crime.

The study narrowed on violence on a single "blockface," or both sides of a given street bounded by two intersections:

"High Stakes in the Classroom, High Stakes on the Street: The Effects of Community Violence on Students’ Standardized Test Performance," by P. Sharkey et al. in Sociological Science
"High Stakes in the Classroom, High Stakes on the Street: The Effects of Community Violence on Students’ Standardized Test Performance," by P. Sharkey et al. in Sociological Science

That crime data was compared with a longitudinal New York City Department of Education census of public-school children containing their addresses, test scores and more (both of these data sets are of the kinds that you only get your hands on as a very, very fortunate researcher).

At the intersection of those two data sets, it's hard to think of additional forces that might explain the study's results. The timing of violence surely has little to do with the timing of test dates.

"Every aspect of learning is affected by what goes on in a kid’s neighborhood," Sharkey says. "Educators certainly know this is the case. Researchers have presented evidence suggesting this is true. But there hasn’t been much rigorous causal evidence that holds up against criticism."

Now we have some of that evidence -- or at least enough of it to expand the conversation about what types of neighborhood shocks are most important, exactly how they influence kids, and even how much the achievements gaps might be explained by exposure to violence. The study focused on violence during a single week in time. But many of these children no doubt deal with this reality on a regular basis. And so what happens over time when children constantly take the stress and disruption of living with violence into school?

"We’re not arguing that we have to focus our attention on eliminating violent crimes right before kids take a test," Sharkey says. "That’s not the point. The point is more that we should be aware that the impact of things going on in kids' residential environments is not left at home when they enter schools."