By ANDREW W. LEHREN and AL BAKER
Interactive feature: http://projects.nytimes.com/crime/homicides/map
A young boxer was shot dead outside a Bronx bodega at 3:30 a.m. on a Saturday last August. Weeks later, a 59-year-old woman was beaten to death on a Saturday night on the side of a Queens highway. On the last Sunday in September, violence exploded as five men were killed in a spate of shootings and stabbings between midnight and 6 a.m.
Seven homicides in New York City. None connected in any way but this: They happened during the summer months, when the temperatures rise, people hit the streets, and New York becomes a more lethal place.
There were more homicides in September than in any other month last year: 52. Next highest was August, with 51. Variations, of course, exist. There were 48 homicides last March, for instance.
Still, the prime time for murder is clear: summertime. Indeed, it is close to a constant, one hammered home painfully from June to September across the decades. And the breakdown of deadly brutality can get even more specific. September Saturdays around 10 p.m. were the most likely moments for a murder in the city.
The summer spike in killings is just one of several findings unearthed in an analysis by The New York Times of multiyear homicide trends. The information — detailing homicides during the years 2003 to 2008 — was compiled mainly from open-records requests with the New York Police Department, and a searchable database of details on homicides in the city during those years is available online for readers to explore at nytimes.com/nyregion.
Of course, the dominant and most important trend involving murder in New York has been the enormous decline in killings over the last 15 years, to levels not seen since the early 1960s.
Still, hundreds of people are killed every year in the city, and The Times’s findings provide insights about who is killed in New York, as well as who does the killing, where murders occur and why.
Women, for instance, are less likely to be either victims or killers. Those who were killed — at least 73 women were in 2008 — were almost always murdered by someone they knew — boyfriends, husbands or relatives. From 2003 to 2008, the number of women killed each year by strangers was in the single digits — excluding cases in which the police do not know if the killer knew the victim. Last year, as few as eight women died at the hands of strangers.
Brooklyn — as it has since at least 2003 — led all boroughs in the number of homicides last year, with 213. Last year, the 73rd Precinct, which includes the neighborhoods of Ocean Hill and Brownsville, had the largest death toll, 31. The bloodiest block in Brooklyn was in the 77th Precinct, in Crown Heights, bounded by Schenectady Avenue, Sterling Place, Troy Avenue and St. Johns Place. But the borough with the most homicides per capita was the Bronx.
More often than not, the weapon of choice is a firearm. Each year the percentage of people killed by firearms hovers around 60 percent. Though slightly less than in recent years, at least 56 percent of last year’s homicides were committed with these weapons.
Of all the trends to emerge, the time for killing was among the most enduring.
In New York, the trend goes back well before the years covered in the database — at least as far as 1981, according to an analysis of reports by the city medical examiner’s office done by Steven F. Messner, a criminology professor at the State University of New York at Albany. And he believes it stretches back much further than that.
Nationally, in the early 1980s, scientists at the Centers for Disease Control and Prevention analyzed a decade’s worth of homicide data across the nation, and found that while suicides peak in the spring, homicides swell between July and September.
A prime reason murder peaks during this time has to do with the routines of people’s lives, according to Professor Messner. “Homicides vary with social acting,” he said. “It evolves from interactions.”
Summer is when people get together. More specifically, casual drinkers and drug users are more likely to go to bars or parties on weekends and evenings, as opposed to a Tuesday morning. These people in the social mix, flooding the city’s streets and neighborhood bars, feed the peak times for murder, experts say.
And the trend occurs in other cities, in places like Chicago, Boston and Newark, according to criminologists.
Some of the same trends are on display around Christmastime and are believed to be behind the slight increases in murder that occur then, criminologists say.
Thomas D. Nerney, who retired in 2002 as a detective in the New York Police Department’s Major Case Squad, said the patterns were well known within the department.
Assigned as a detective in Brooklyn from 1972 to 1986, he said that on a hot summer night or in the holiday season, a similar set of factors seemed to be behind the killings: a chance to socialize and to drink or use drugs.
He recalled the late 1970s and early ’80s in Brooklyn, when the heavier homicide caseloads seemed to come as neighborhoods got hotter.
“We had so many of them,” Mr. Nerney said. “They would be on rooftops. There might be somebody who lured someone somewhere; you would have a sex-related killing or a revenge killing. Rooftops or backyards.”
The Times analysis, when compared with Professor Messner’s findings from 1981, shows that increasingly, more victims were killed between midnight and 8 a.m. in recent years than in the past.
According to the professor’s study of homicides in Manhattan, 29 percent of the 1,826 victims in 1981 were killed between midnight and 8 a.m. More recently, from 2006 through 2008, 39 percent of all homicide victims were killed during those hours, the Times analysis shows.
Also, as the number of homicides has shrunk, the data shows that more are occurring on weekends. From 2003 to 2008, 36 percent of all victims were killed on Saturday or Sunday, the analysis shows.
Failing to understand the basic connection between time of year and homicide rates can lead law enforcement agencies to faulty conclusions about what is happening in the streets — and it can affect their strategies.
In St. Louis, a 1990s-era gun buyback program begun each fall was thought by some to be behind a drop in violence. But as Richard Rosenfeld, a professor of criminology at the University of Missouri, St. Louis, studied the program’s impact, he found that the annual crime reductions were more attributable to the normal seasonal ebbing in homicide and assaults.
In New York, Vincent Henry, a retired police sergeant who now teaches criminology and who has studied the department’s Compstat program, in which computerized data is used for more efficient policing, said that time was one of many factors in making decisions about staffing and when and how to deploy officers.
But that was not always the case.
In the early 1990s, police managers altered the working hours for various groups of detectives, including those tracking narcotics cases and those seeking to arrest criminals wanted on open warrants.
It seemed to the top officials at the time that too many officers were keeping bankers’ hours — ending their shifts at dusk and taking weekends off — and not working closely enough with counterparts.
Jack Maple, a former police deputy commissioner who helped develop Compstat, wrote a book, “The Crime Fighter,” in which he detailed the issues of the day. He described the shortfall this way: “Unfortunately, the bad guys work around the clock.”
And in the summer months, the bad guys tend to be deadliest.