The Correlation Between Crime Rates and Weather Patterns in Northern Brooklyn During 2012

Content

I. Introduction and Case Overview

II. The History of Previous Studies Examining Weather and Crime

III. Hypotheses and Research Questions

IV. Dataset and Methodology
a. The Place: Northern Brooklyn
b. Data
c. Analytical Strategy

V. Results
a. Bivariate Findings
b. Multivariate findings

VI. Discussion and Analysis

VII. Limitations and Recommendations for Future Research

VIII. Conclusions

References

I. Introduction and Case Overview

In March 2014, New York City Police Department (NYPD) Commissioner William J. Bratton declared “ ‘Jack Frost is the best friend of a police officer,’”referring to the factthat “unseasonably cold weather and several snowstorms had contributed to the low crime rates in New York City to begin the year” (Grynbaum, 2014). The decrease in crime Mr. Bratton was referring towaspart of a larger drop that actually beganduring Mr. Bratton’s previous tenure as New York City police commissioner in the early 1990s and one that was continued under former Police Commissioner Raymond W. Kelly between 2002 and 2013.

Unlike the first several months after Mr. Bratton’s reappointment to the NYPD in 2014, early in the penultimate yearofformer Commissioner Kelly’s tenure, on February 9, 2012, the web-basednews serviceDNAinfo reported a near six-percent increase in all major crimes in New York City for thefirst month of 2012 compared to the year before. The cause, they claim, was a result of drastically higher than average temperaturesin January, which averaged 54 degrees as a high temperature, or about 16 degrees above normal for the month, according to statistics obtained from Weather Underground, which culls data from the National Weather Service (Weather Underground, 2012). The article reported, “NYPD insiders say the only logical explanation [for the increase in crime] is the unseasonably warm weather that has New Yorkers, criminals included, spending more time outdoors on snow-free streets. In early 2011, criminals had to contend with snowdrifts and blizzards” (Weiss, 2012).

Later that year, on July 8th,amidst a heat wave that struck the Midwest and East Coastand sent temperatures well into the 90s and even triple digits in some places (Eligon & Santora, 2012), The New York Post reported, “New York City’s homicide rate is rising as fast as its sweltering temperatures” (Harshbarger, 2012). During the mid-summer heat wave the periodical detailed a much higher than normal 16 homicides in only a five day stretch.

The article came only days after high-ranking New York City Police Department (NYPD) precinct commanders were summoned to regular CompStat meeting at NYPD’s headquarters to explain the recent rash of shootings and killings in their precincts over the preceding one month. The uptick accounted for a 46% rise in shootings compared to the same period the previous year and was “blamed on the heat wave and fewer stop-and-frisks” according to several police sources (Celona, 2012).In fact, according to the New York Times, 2012 was the hottest recorded year on the books in the United States, defeating the previous record set in 1998 by about a degree Fahrenheit (Gillis, 2013).

As the days grew cooler in 2012, a late-November day with a temperature of a below-average 42 degrees in New York City saw more than a 24-hour stretch “without a single murder, stabbing, or shooting”(Mathias, 2012). Former Deputy Commissioner of Public Information Paul Browne stated that it was a “Nice way to start the week.” (Parascandola, 2012). It was not only a nice way start the week, as Mr. Browne modestly phrased it, but it was such an astonishing feat that police officials couldn’t actually recall or find statistics to support any such previous occurrence in New York City’s history. The feat was repeated on November 7th, 2013(Parascandola, 2013) and again in March 2014 (Mancuso, 2014). The maximum temperatures those days reached 57 and 39 degrees, respectively (Weather Underground, 2014). The occurrencesmarked a remarkable achievement compared to the average six murders per day the city once saw in 1990 when there were a daunting 2,262 slayings (Lueck, 2007).

As newspapers and media outlets alike have often sensationalized the effects of heat waves on higher rates of crime, researchers have long debated whether or not there is a true statistical correlation between the two and whether the relationship between higher temperatures and higher crime rates is linear. There have even been a number of studies correlating full moons and other lunar events with higher rates of crime (Cahoon, 1977) and one that found a positive relationship between altitude and burglary (Breetzke, 2012). But the relationship between weather patterns and crime has also not been studied extensively in New York City. This paper will identify and analyze the statistically significant relationship, if any, between crime rates and temperature variations in New York City, specifically northern Brooklyn. Weekly crime data aggregated from the New York City Police Department’s (NYPD) CompStat reports from 2012 for the Patrol Borough Brooklyn North (PBBN) area will be analyzed and compared to various temperature variations in New York City. The seven major crime categories will be analyzed—murder, rape, robbery, burglary, felony assault, grand larceny, and grand larceny auto.

The data will be compared to weekly average high temperatures, and weekly precipitation and the control variables for amount of daylight, as reported by the National Weather Service for the Central Park weather station in Manhattan, the closest official weather station to the geographical boundaries set by the NYPD’s Brooklyn North patrol area and the monthly unemployment rate for Kings County in 2012. The statistical analysis is being performed with the intention of testing the much-researched hypothesis that higher temperatures will result in higher rates of both violent crime and property crime.

II. The History of Previous Studies Examining Weather and Crime

The link between increased crime rates, violence and weather patterns has been studied for quite some time resulting in a multitude of both laboratory and field research studies. These studies proliferated in the 1970s and 1980s when scientists began to more widely cite the routine activities theory to help explain the dramatic increase in crime over the previous several decades. Routine activities theory, first introduced by Felson and Cohen, suggested that crime would increase when there was a convergence of three separate ideas: likely offenders, suitable targets, and the absence of responsible guardians (Tewksbury& Mustaine, 2010). In the case of temperature variation, if more people were outside in warmer weather, it would stand to reason there would be a greater number of both offenders and targets. Routine activities theory “helps explain shifts in opportunity as weather changes” (Mares, 2013). There have been a number of studies showing statistically significant correlations between higher temperatures and higher rates of crime. However the discovery of both linear and curvilinear correlations complicated the link.

Studies initially focused on analyzing the link between higher temperatures and human aggression and anger. A field study that heavily cited routine activities theory, conducted by Douglas Kenrick and Steven MacFarlane (1986),found a directly positive relationship between aggressive driving and higher ambient temperatures. The study, conducted in Phoenix, was performed in the spring and summer and specifically studied the “range on the temperature-humidity discomfort index up to 116 degrees Fahrenheit” (Kenrick & MacFarlane, 1986). The researchers identified a direct linear relationship between higher rates of aggression and higher temperatures. The determination of aggression was based on how many times the subjects honked their car horns and made inappropriate gestures or verbal comments from their vehicles. They were observed while they were driving in hot temperatures and had their windows down and, undoubtedly, their vehicle air conditioning turned off. A laboratory study by Anderson (1989) agreed with the linear relationship between higher temperatures and aggression. Ransberger and Baron (1978) also found a positive and linear link between temperature and the potential for civil unrest and riots.

Bell and Baron’s laboratory study, published in the Journal of Applied Social Psychology, is perhaps the most widely respected and cited temperature and aggression correlation study. The researchers found a statistically significant curvilinear relationship between uncomfortably warm temperatures and human aggression when tested under artificial conditions (Bell & Baron, 1976). They found that aggression and violence did, in fact, increase as temperature increased. However, quite interestingly they also found that as the temperature rose from around 80 degrees up towards dangerous levels (greater than 95 degrees Fahrenheit), humans in the study actually experienced significantly reduced aggression. The researchers “theorized that extremely high temperatures might serve to suppress aggression because people are more interested in finding an escape from heat” (Mares, 2013). Bell and Baron’s theory, known as the NegativeAffect Escape Model, “posits that a negative affect and associated violence increase as temperatures increase to an inflection point beyond which violence decreases as a person's ‘flight’ motivation (i.e., aversion to heat) overrides aggressive motives” (Gamble & Hess, 2012).

Researchers tried to take it one step further and identify whether higher temperatures would lead to not only aggression but also whether aggression would manifest itself in the committing of crimes. Anderson and Anderson (1984) found that higher temperature was “linearly related to assaults.” The researchers studied daily crime data in Chicago in the summer of 1977. After running a regression model they found that there was no curvilinear trend with assaults and that as temperature continued to rise, assaults continued to rise as well. They named this theory the General Affective Aggression Model, differing from the curvilinear theory proposed in the Negative Affect Escape Model (Lindsay & Anderson, 2000).An Atlanta study by Cook(2012) et al. concurred with Anderson’s linear relationship findings between assault and higher temperature after their regression models showed that “for every 5.6°C (10°F) increase in daily high temperature, there was an increase of 0.10 trauma center admissions for assault.”

Cohn and Rotton’s (1997) study echoed the curvilinear relationship found by Bell and Baron as well. In their study of 1987 assault data in Minneapolis, Minnesota, they found a positive relationship between higher temperatures and higher rates of assault up to about 75 degrees. The relationship then turned negative resulting in a decrease in assaults as temperatures grew uncomfortably hotter.

Gamble and Hess conducted a more recent study published in the Western Journal of Emergency Medicine that studied temperature variations and violent crime rates in Dallas, Texas. After examining violent crime data aggregated from the Dallas Police Department between 1993 and 1999, they also found a curvilinear relationship, echoing the findings ofBell and Baron’s data as well as findings of Cohn and Rotton. By analyzing homicide, assault, and rape, Gamble and Hess (2012) found a positive relationship between an increase in temperature and the studied crimes past 80 degrees Fahrenheit. Above 80 degrees the rates of crime began to level off but the relationship actually turned negative at temperatures above 90 degrees. They offered a potential explanation for the variation, positing “that higher temperatures may encourage people to seek shelter in cooler indoor spaces, and that street crime and other crimes of opportunity are subsequently decreased” (Gamble & Hess, 2012).

However, a study by Cotton (1986)analyzed the frequency of summer-month violent and property crime in two Midwestern cities and disagreed with previous studies finding curvilinear relationships between temperature and violent crime. The results indicated that all violent crimes—not just assault—were linearly related with higher temperatures. The study also found that, unlike Baron and Bell’s findings, aggressive incidents of crime did not decrease when the temperature reached the mid-80s but actually increased as the temperatures reached well into the 90s.

Although robbery is considered a violent crime under the New York State Penal Law, it is widely considered to be a crime of opportunity rather than a violent one and is often referred to, in many research studies, as a variation of property crime. Like New York State, the Federal Bureau of Investigation’s (FBI) Uniform Crime Report (UCR) also classifies robbery as a violent crime. Thus, in this paper, it will be grouped under violent crime due to the geographical area of statistical analysis.Research shows that robberies surge slightly less than other violent crimes in the warmer months of July and August and then increase again in December. The researchers cite the holidays and Christmas time as potential reasons for the second uptick (McDowall et al, 2011).

A field study in the city of Philadelphia, Pennsylvania agreed with McDowall’s research when analyzing robberies finding a linear relationship between higher temperatures and increased frequency of street-level robberies while controlling for socioeconomic status. The researchers found that the numbers of robbery increased in lower socioeconomic areas(Sorg & Taylor, 2011). Contrary to these findings, a study of crime in Beijing, China found no relationship between robbery and temperature but did link burglary with periods of the year that had longer hours of daylight (Chen, 2011). The study found that robberies were linked more with temporal factors such as holidays or days off from work or school.

Rape and homicide have also been studied for their relationship to temperature variations. Anderson and Anderson found that murders and rapes experienced increases as temperature increased and then had a slight decrease between 79 degrees and 91 degrees and then a sharp uptick between 92 and 99 degrees. The slight curvilinear trend generally agreed with Bell and Baron’s study on aggression and temperature (Anderson & Anderson, 1984). Regression models run by Gamble and Hess (2012) in their study of Dallas crime rates showed a positive, yet extremely weak, curvilinear relationship between higher temperatures and both sexual assaultsand homicides that increased as temperature increased. The relationship turned negative after 90 degrees. A Wall Street Journal article found that in a study of monthly homicide tallies in New York City between 2002 and 2009, the months with the most killings in descending order were July, June, September, December, and August. An explanation was offered that there were more domestic homicides in December due to increased family presence during the holidays (Gardiner, 2010).

Other researchers aimed to identify a link between not only violent crime but between non-violent criminal activity, including property crimes, and weather patterns. In a 2009 study that measured the number of daily calls of service for disorderly conduct in the United Kingdom, researchersfound a linear relationship between higher temperatures and levels of humidity and disorderly conduct calls. They also found that precipitation, wind speed, and wind direction had no bearing on the relationship (Brunsdon et al, 2009).

In regard to burglary specifically, researchers cited routine activities theory as a plausible explanation for a positive relationship with the temperature increase. They suggested that the more time that people spend outside, which tends to happen during the warmer months, the greater the likelihood they will be a victim of burglary due to the fact their homes are more likely to be unoccupied (Cohen & Felson, 1979). DeFronzo (1984) found that burglaries are positively and linearly related to temperatures above 90 degrees but a study by Chen (2011) found a curvilinear relationship that showed a positive relationship between rising temperatures and burglaries until approximately 80 degrees Fahrenheit whereupon a negative relationship was observed as temperatures rose further. Grand larceny, petit larceny and other types of theft also have been correlated with weather patterns. Mares (2013) found an increase in property crime as temperatures increased, albeit a relatively weak relationship.

The link between precipitation and crime has also been extensively studied resulting in mixed conclusions. Routine activities theory was referenced again as a reasonwhy precipitationwould bea deterrent to criminals attempting to commit burglaries in periods of precipitation or “bad” weather suggesting that it was due to the increased risks and costs and the increased potential for more residents to be at home during poor weather (Horricks & Menclova, 2011). As previously stated, however, Brunsdon et al (2009) found no correlation between precipitation and higher or lower rates of crime and Field (1992) also failed to find a statistically significant relationship. Jacob et al(2007) found in their study of weekly crime data for over a 100 municipalities throughout the United States between 1995 and 2001 that for every one-inch increase in rainfall there was a concurrent 10% drop in all violent crimes. Their research also found no relationship or statistical significance between precipitation and property crime . The New York Times, on the other hand, found that New York City sees a stark drop in homicides on rainy days and that the effect is especially pronounced in the summer (Lehren & Hauser, 2009).

Length of daylight is also positively linked with certain types of crime. Suggesting routine activities as a cause, a study conducted in China found that burglaries increase when it is warm outside due to the longer days occurring during the warmer months (Chen, 2011). Cohn(1993) reports a negative relationship between sexual assault and length of daylight citing the fact that sexual attackers often prefer anonymity and thrive in darkness.Coupe and Blake (2006) found burglary to be more prevalent during the day, again citing routine activities theory as a cause. The findings lead to a reasonable prediction that longer days will lead to an increase in burglaries. Conversely, Sanders and Doleac’s (2012) large-scale analysis found while analyzing crime data collected from the National Incident-Based Reporting System between 2005 and 2008 that the extra hour of daylight given by daylight savings time accounted for a stark reduction in crime during that one hour period. Rotton and Frey (1985) found no correlation between domestic violence and length of day in an analysis of Dayton, Ohio crime data but found a positive and linear relationship with assault. Regardless of how long the daylight is, though, and how many more crimes might occur in the warmer months, it doesn’t change the fact that “the bad guys work around the clock” (Zimring, 2012).

For municipalities that are consistently warm throughout the year, a reasonable hypothesis would be to assume that assaults were predominantly higher in these areas year round.However, a study conducted in the 1970s found no statistically significant correlation between latitude and documented occurrences of assault (Lewis & Alford, 1975). Their study found that crimes generally are much higher during the warmer months regardless of what the climate of the city is. Rotton and Cohn (2003) found no statistically significant relationship between homicide and changing of seasons. The Center for Disease Control and Prevention (CDC) analyzed ten years of data leading up to the 1980s and found that homicides are significantly higher between July and September nationwide (Lehren & Baker, 2009).

Other weather-related variables have been analyzed, such as wind and barometric pressure with a resulting mixed consensus. One study found a significant but weak and negative relationship between daily wind speed and police calls for service for domestic violence complaints (Rotton & Frey, 1985). A different study in Los Angeles found an uptick in homicides on days with high winds (Miller, 1968).Feldman and Jarmon (1979) analyzed barometric pressure and but the results yielded no significant correlation with homicide and assault. Interestingly, Fisher et al (1984) made a valid point that since changes in barometric pressure are often associated with larger weather events, it is possible that it does have a significant correlation with temperature but that the effects are disguised by temperature or precipitation.

The following analysis also comes at the forefront of the global warming and climate change debate. Researchers predict the global warming phenomena will be responsible for a vast amount of social and economic costs before the century is over. Matthew Ranson, a researcher from a firm in Massachusetts “suggests global warming will trigger more crimes including murders and rapes over the next century, with social costs estimated to run as high as $115 billion” (Sahagun, 2014). Ranson (2010) found, in his study, a direct linear relationship between all types of crimes and temperature increase. In his analysis he found that in the presence of climate change, from 2010 to 2099 there would be “an additional 35,000 murders, 216,000 cases of rape, 1.6 million aggravated assaults, 2.4 million simple assaults, 409,000 robberies, 3.1 million burglaries, 3.8 million cases of larceny, and 1.4 million cases of vehicle theft” (Ranson, 2010) compared to the offenses that would occur without climate change.

If Ranson is correct, the correlation between weather and crime is one that must be studied in detail. If the world is in store for forever-increasing average temperatures, it may have dire implications for our economy as well as our future policing strategies and deployments.Rotton and Cohn’s (2003) study and regression analysis on global warming and its effect on crime suggested that a “one-degree increase in temperatures in the United States will increase the rate for… burglaries by 8.16, and larcenies by 10.65 per 100,000 population.”

As a result of this data, law enforcement agencies in several cities have altered their policing strategies based on weather and other factors. A report by Jennifer Bachner of the Johns Hopkins University Center for Advanced Governmental Studies found several departments that have employed predictive policing computer programs. Baltimore County Police, for example, were finding that burglaries were rising in warmer weather as a result of residents tending to stay outside longer and the department adjusted its police patrols in response (Rosen, 2013). The Toronto Police Department in 2012 implemented a mandatory overtime program to combat the anticipated increase in summer violence (Poisson, 2012).

III. Hypotheses and Research Questions

What affect does weather have on crime patterns in Brooklyn North? What effect does precipitation have on crime patterns? The paper is written with the intent of testing the hypothesis that higher temperatures will lead to higher rates of property and violent crime. The second hypothesis being tested is that higher rates of precipitation will be negatively correlated with rates of crime.

IV. Dataset and Methodology

This study adopts a retrospective and longitudinal approach to examine the association between weather patterns and crime rates in northern Brooklyn in 2012. Archival weekly time series from the National Weather Service, the New York City Police Department (NYPD), and New York State Department of Labor will be used in this secondary data analysis.

a. The Place: Northern Brooklyn

Data was collected from the Patrol Borough Brooklyn North area, an area that covers 10 different precincts of varying geographical size, demographics, and socioeconomic statuses. Based on 2010 United States Census data, the population was 904,000 and the median annual income in the study area ranges from a low of $9,001 in a small portion of the Fort Greene section of the borough to a high of $170,481 in the affluent DUMBO neighborhood (Venugopal, 2011). The red boundary line below in the Google Maps map indicates the rough boundaries of Patrol Borough Brooklyn North as set by the NYPD.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1 Google Maps: Brooklyn, New York, USA [Map data © 2015 Google]

b. Data

Weekly crime data for all 52 weeks of 2012 weregathered from the publically accessible NYPD’s CompStat website. The website publishes weekly crime data for all major index crimes as well as shootings, various misdemeanor crimes and total index crimes collected bythe housing and transit bureaus. Data were collected from the Patrol Borough Brooklyn Northarea, an area that covers 10 different precincts of varying geographical coverage, demographics, andsocioeconomic statuses. Based on 2010 United States Census Data, the population was 904,000 and the median annual income in the study area ranges from a low of $9,001 in a small portion of the Fort Greene section of the borough to a high of $170,481 in the affluent DUMBO neighborhood (Venugopal, 2011).

The seven major index crimes that were analyzed as the dependent variables are murder, rape, robbery, burglary, felony assault, grand larceny, and grand larceny auto. Standardized indicator in the form of weekly number of offenses per 100,000 population was obtained for each of these index crimes and each of the 52 weeks of 2012 (see Table 1). Weather data were collected from Weather Underground, a website which culls its data from the National Weather Service, and was analyzed as independent variables.The National Weather Service collects its data from the Central Park weather station and measures its temperature in Fahrenheit and precipitation in inches. Average weekly high temperatures, and weekly precipitation will be used as independent variables.Average weekly length of daylight in hours, and monthly New York State Department of Labor unemployment rate data for Brooklyn (Kings County) for 2012 will be used as control variables. The unemployment rate has been shown in several studies to mostly increase property crime (Fallahi, 2012) and, to a lesser extent, violent crime (Finklea, 2009). Days with longer amounts of sunlight have been linked to higher rates of property crime (Chen, 2011). IBM SPSS Statistics version 22.0 was used for bivariate analysis using Pearson’s r to demonstrate the relationship between the desired variables. Subsequently, multivariate analysis utilizing multiple regression analysis was employed to scan for statistically significant relationships between the variables and explain variance.

Table 1. Descriptive Statistics and Frequencies

Abbildung in dieser Leseprobe nicht enthalten

c. Analytical Strategy

Given the continuous nature of the variables, Pearson’s r correlation and regression models were used to assess the proposed relationships at both bivariate and multivariate levels. The presence of autocorrelation in the prediction errors in longitudinal data from least squares regressions violates the assumption of independence of errors. A serious problem of autocorrelation (as shown by a Durbin-Watson statistic considerably lower than 2) affects the reliability of significance test. Durbin Watson test was performed for all seven multivariate regression models, which yielded Durbin Watson statistics ranging from 1.542 to 2.497. This evidence demonstrated that autocorrelation was not a serious problem and that multiple regression modeling can be used without major biases.

V. Results

a. Bivariate Findings

The use of average weekly high temperature and average weekly precipitationas predictors in the Pearson’s r bivariate test found mixed results for the seven dependent variables (see Table 2). Based on the observed p-values for the average weekly precipitation predictor, the correlation between precipitation and all seven dependent variables was found not to be statistically different from zero. While using average weekly high temperature as a predictor, temperature failed to statistically significantly correlate with rape.The p-value for robbery at .035 was also statistically significantly correlated with temperature and the relationship was positive based on the Pearson’s r of .294. The relationship between temperatureand felony assaultwas also statistically significant based on the p-value of .000. The Pearson’s r value of .736 indicates the correlation is positive as well. Temperature was also found to be statistically significantly correlated with burglary and had a positive relationship based on the p-value of .004 and the Pearson’s r of .396. Temperature had a significant and positive correlation with grand larceny with a p-value of .001 and a Pearson’s r of .465. Lastly, the relationship between temperature and grand larceny was also statistically significant with a p-value of .043. The relationship is also shown to be positive based on the Pearson’s r of .282.

In sum, temperature was significantly and positively correlated with robbery, assault, burglary, larceny, and grand larceny auto: the higher the temperature the higher the levels of these criminal activities. But temperature did not significantly associate with the rates of murder and rape. In contrast, precipitation did not significantly predict the fluctuation of any crime rate. While these bivariate findings suggest a strong correlation between temperature and most crime categories, they might have been confounded with some other correlates of crime that share similar patterns of weekly variation. Therefore, multivariate modeling is needed to remove these potential disturbances.

b. Multivariate findings

A series of multiple regression models were built to test the robustness of the influence of temperature and precipitation on the seven crime rates, while holding constant the impact of unemployment and the average length of day. Based on the observed p-values of the models for murder, rape, and grand larceny auto (.458, .569, and .161, respectively), temperature and precipitation were found to be statistically non-significant for the prediction of these three dependent variables (see Tables 3-5 andCharts 1-3). As a result, these models were excluded from any further discussions.

Table 3. Murder Multivariate Coefficientsabc

Abbildung in dieser Leseprobe nicht enthalten

Table 4. Rape Multivariate Coefficientsabc

Chart 2. High Temperature and Rape

Abbildung in dieser Leseprobe nicht enthalten

Table 5. Grand Larceny Auto Multivariate Coefficientsabc

Chart 3. High Temperature and Grand Larceny Auto

Abbildung in dieser Leseprobe nicht enthalten

High temperature was statistically significantly correlated with robbery based on the p-value of .001 (see Table 6 and Chart 4). The R square (.330) suggests that the model explains 33.0% of the variation in temperature’s effect on robbery. As previously stated in the bivariate analysis, robbery was positively correlated with temperature. An increase of one degree Fahrenheit was related to an additional .74 robberies per 100,000 population.Similar to the bivariate test, there was no statistical significance found in the relationship between precipitation and robbery in the multivariate analysis.

Table 6. Robbery Multivariate Coefficientsabc

Chart 4. High Temperature and Robbery

Abbildung in dieser Leseprobe nicht enthalten

High temperature’s correlation with felony assaults was the most statistically significant variable with an observed p-value of .000 (see Table 7 and Chart 5).With the R square obtained for felony assaults of .564, the model explains an impressive 56.4% of the variation in the average high temperatures’ effect on the felony assault rate. The variance was the highest among all of the dependent variables in the study. Based on the unstandardized regression coefficient, a one-degree increase in temperature was related to an additional .41 felony assaults per 100,000 population. The relationship between precipitation and felony assault was found not to have any statistical significance.

Table 7. Felony Assault Multivariate Coefficientsab

Chart 5. High Temperature and Felony Assault

Abbildung in dieser Leseprobe nicht enthalten

High temperature’s effect on burglary was also significant based on the p-value of .007 (see Table 8 and Chart 6). The R square (.253) explains 25.3% of the variance in the average high temperatures’ effect on burglary. The variance was the lowest among all of the dependent variables analyzed. The unstandardized regression coefficient can predict an additional .38 burglaries per 100,000 population for every one-degree increase in temperature. Precipitation and burglary were found to have no significant relationship.

Table 8. Burglary Multivariate Coefficientsabc

Chart 6. High Temperature and Burglary

Abbildung in dieser Leseprobe nicht enthalten

Finally, high temperature was significant correlated with grand larceny with a p-value of .000 (see Table 9 and Chart 7). The R square (.345) explains 34.5% of the variance in the effect temperature has on grand larceny. The unstandardized coefficient for the variable indicates that for every one-degree increase in temperature there will be an additional 1.15 grand larcenies per 100,000 population. Like all of the other dependent variables, precipitation was found to have no significance with grand larceny and was not analyzed further.

Table 9. Grand Larceny Multivariate Coefficientsabc

Chart 7. High Temperature and Grand Larceny

Abbildung in dieser Leseprobe nicht enthalten

VI. Discussion and Analysis

The hypothesesof this study that higher temperatures lead to higher rates of property and violent crime and that higher rates of precipitation are negatively correlated with rates of crimewere only partially proven but theanalyses for several of the variables did yield interesting findings. Whereas some researchers found that rape and homicide were related to temperature, this study found no statistical significance with either dependent variable and unlike the findings made by Anderson and Anderson (1984) and Bell and Baron (1976), there actually appeared to be no significant increase in the rate of these variables whatsoever as temperatures became warmer. Murders, in fact, were pretty well dispersed throughout the seasons in northern Brooklyn in 2012. The week that had the most murders of any other week in the year actually came in March when the average high temperature was a relatively mild 69 degrees. There was an instance in a week in July where there were no murders and the average high temperature was 87.The findings of the study are consistent with previous research conducted by Feldman and Jarmon (1979) in a study of temperature and homicide data in Newark and a study by Michael and Zumpe (1983a, 1983b) that analyzed data from several cities across the nation. Perry and Simpson’s (1987) study in Raleigh, North Carolina found no correlation between temperature and homicide either.

The other variables, excluding grand larceny auto, were significantly correlated with temperature.Since multiple regression only studies linear relationships, it was not possible to determine if there was a linear or curvilinear relationship between the specific crimes and temperature. Thus, this study cannot refute or prove Bell and Baron’s Negative Affect Escape Model that crime increases with temperature in a curvilinear relationship to a certain point until the oppressive heat overrides human aggression and causes crime to decrease. This study does show linearity between crime and temperature, however, and furthers the explanation of crime and temperature as a linear function as offered by Anderson and Anderson in their General Affective Aggression Model.

There is one possible explanation for why murder and rape are not statistically correlated with temperature and precipitationunlike other violent crimes. In most cities across the nation,routine activities theory could explain the findings of this study in that “temperature…brings victims into contact with strangers who are the primary perpetrators in aggravated assault, but high temperature does not alter interaction patterns in households between family members where most murders occur” (Rotton and Cohn, 2002).It should stand to reason that if many homicides are a result of domestic violence or occur in a situation where the victims knows their attacker that many of these crimes will take places indoors and will not be affected by temperature or precipitation in any sort of way.

However, New York City is a unique study area with a density level that far exceeds other cities. While New York did once yield statistics that provided for more indoor homicides than outdoor, this had generallynot been the case in more recent decades. “The mix changed radically from the late 1980s through the early 1990s, when outdoor killings greatly exceeded those taking place indoors” (Karmen, 2006). Since the 1990s the wide gap between outdoor killings and indoor has narroweddue to the vast increase in police presence on the street but today there are still typically more homicides in New York City committed outside than inside. In fact, homicides attributed to domestic violence only accounted for 68, or 16%, of the 419 citywide homicides in 2012 and presumably at least a few of those murders must have taken place outside (Ruderman, 2012).However, so far in 2014, as of May 8, murders indoors have slightly outpaced murders outdoors 51 to 40, possibly because of the brutal cold weather the city experienced in the beginning months of the year. Time will tell if these numbers stay the same, as the warmer months generally have brought a shift towards outdoor homicide in the city.

As such, in the case of this study and New York City and, Brooklyn in particular, routine activities theory and the opinion by Rotton and Cohn might not offer the best explanation for why temperature shouldn’t be correlated with murder in Brooklyn. Steven F. Messner from the State University of New York offered a similar perspective for why there should be a prevalence of homicide in the warmer months stating,“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…. flooding the city’s streets and neighborhood bars, feed the peak times for murder…”(Lehren & Baker, 2009). We can reasonably infer from Messner’s hypothesis that more of these social gatherings in the summer will lend themselves to an increase in outdoor violent crime and homicide, however that conclusion was not reached by this study.

Similar to homicide, rape was found not to be statistically significantly correlated with temperature or precipitation variations. An increase in suitable targets was cited as one reason for an increase in reports of rape in a study of university students (Franklin et al, 2012). One study found that time of day and the season, as well as finding an isolated area—absent of guardians—were large predictors for rape and that temperature and precipitation generally would not have much of an effect (Belknap, 1987). There is, perhaps, a dark figure of crime that is not being taken into account in New York City. John Jay College of Criminal Justice professor Eugene O’Donnell remarked that certain crimes, like rape and felony assault, are dramatically unreported. Former NYPD Deputy Commissioner of Public Information Paul Browne echoed that statement (Destefano, 2013). Additionally, due to the indoor nature of rape, it is possible it is a crime that is less susceptible to variations in weather patterns. In turn, with a potentially incomplete sample set of the number of rapes in New York City, there could be a relationship with weather patterns that cannotbe determined. Despite these statements, however, most literature agreed with the insignificant relationship and studies that found correlations between rape and temperature variations generally found very weak ones and those results were echoed in this study.

Robberiesand grand larceny were, to no surprise, found to have a statistically significant and positive relationship with temperature. This study agreed with previous studies that these variables go up as the temperature goes up. Routine activities theory is more than a reasonable explanation for this relationship since there are suitable targets outside and more motivated offenders.Also, studies have shown that urban street-level robbery, which could also increase grand larceny considering these crimes are relatively similar in the sense they both involve the theft of property, increase significantly not only with temperature but also in areas of lower socioeconomic status. (Sorg, 2011). Since the study area is home to many neighborhoods that rank the lowest in many measures of socioeconomics, the results of the study are all the more appropriate. In terms of street-level robbery being more prevalent,it would be interesting to see if the same results were yielded in poorer neighborhoods in the less densely populated areas of the city, such as Staten Island or remote areas of Queens, where people are more likely to use cars to travel around as opposed to walking to their destination, which increases their likelihood of being a victim of one of these street-level crimes.

The results for felony assaults indicate that there is a correlation with higher temperatures. The results echo previous findings that supported the link, including one study by Butke & Sheridan (2010) that found the highest rates of aggressive crime are found in the summer months and the lowest are found in the winter months. John et al (2004) agreed with the study stating that violent crimes generally are explained well by both the routine activities theory in that there are a greater number of suitable targetswhen the weather is warmer.

This studyshowsresults similar to previous studies that demonstrated a positive relationship betweenburglary and temperature. While some studies show that lower rates of property crime were linked with higher rates of precipitation, the results of this study found no such link.It makes sense logically when taking into account routine activities theory. With more residents being away from their homes in the warmer months, the number of suitable targets, in this case residences, goes up exponentially.

The other property crime, grand larceny auto, was not found to be significantly linked to temperature. While some previous studies linked carjacking to higher temperature, this type of crime would generally fall under the robbery statute as a violent crime and not be classified as grand larceny auto. Whereas in some cities, such as Newark, carjacking is still a major problem, it is an extremely infrequent crime in New York City (Santora & Schwirtz, 2013) compared to decades ago. The massive decline in car theft could potentially explain the reason why temperature wouldn’t have much of an effect on the crime.

In fact, in 2013 there were only 7,400 vehicles stolen citywide, an astonishing 95% drop when compared to crime statistics from 1990 when there were 146,925. As of 2010, there were approximately 100 NYPD police cars citywide equipped with automatic license plate readers, a further disincentive for car thieves (Doyle, 2010). As of 2013, former NYPD police commissioner Raymond W. Kelly also announced the expansion of the department’s massive camera system and their plan to install license plate readers at every possible bridge and tunnel entrance and exit in Manhattan (Sledge, 2013).With the city’s plan to make it more difficult to steal a vehicle, it should make the link between temperature and grand larceny auto even more insignificant than it was already found to be in this study. What’s more, since temperature’s link to crime is heavily based on the routine activities theory, significantly decreasing the number of suitable targets, in this case, vehicles, to steal should make the link between temperature and grand larceny auto all the more weak. Additionally, the number of motivated offenders should drop significantly knowing that this crime is more difficult to get away with.

The results for precipitation are generally in accordance with previous studiesthat found precipitation to be an insignificant predictorof the crime rate and temperature to be a more compelling predictor of crimes (Feldman & Jarmon, 1979). However, a reasonable expectation would have been for there to be, at the very least, some sort of significant negativerelationship between precipitation and crime for some of the variables in the study. Routine activities theory was cited in many studies as a reason why precipitation would deter criminals from committing burglaries and how suitable targets for robbery or grand larceny would be decreased. However, this study agrees with studies conducted by Brunsdon et al (2009) and Field (1992) that precipitation has no relationship with an increase or decrease in the crime rate and disagrees with Jacob et al (2007).

The results are somewhat baffling when taking routine activities theory into account. It would stand to reason that during rainy days or seasons there would be a decrease in motivated offenders and suitable targets. This would seem especially true in a city like New York where due to density and the prevalence of so many buildings there are many places to seek shelter and stay inside. The hypothesis would also be in accordance with a New York Times analysis of six years of homicide data between 2003 and 2008 that found that the homicide rate is much lower when it rains, especially in summertime. Former NYPD commanding officer of the Bronx Homicide squad Vernon Geberth remarked, “ ‘In good weather, there are more people out on the stoops…. Somebody bad-eyeing somebody else…. On bad weather days, people are apt not to run into each other that are carrying grudges from the day before’ ” (Lehren & Hauser, 2009). The study found that there were three fewer homicides per 10-day rate when there was more than an inch of rain. Perhaps in a future study, citywide crime data and data over a longer period of time might yield a statistical significance and a negative relationship between precipitation and crime.

VII. Limitations and Recommendations for Future Research

The most obvious limitation from this research project is the lack of data, as I was not able to obtain additional years of CompStat data from the New York City Police Department (NYPD). With only one year of both weather data and CompStat data, the chances of the examined year being an anomaly in eitheran increase or decrease in crimes or a hotter or colder than average year are higher. As a result, future projects could incorporate additional years of CompStat data to better analyze the relationship between temperature variations and crime occurrences.Additionally since the NYPD only publishes weekly data on its website, it becomes inherently difficult to pinpoint exactly what crimes occurred under what exact temperature, at least for the purposes of statistical analysis in this study.

Additionally, there was a major anomaly in the weather patterns in 2012. Due to the effects of Hurricane Sandy in October 2012, most crime numbers decreased well into the middle of November. As a result of this, there may have been a significant decrease in overall crime or at least in the amount of crime that was reported to police and subsequently recorded into CompStat. 2012 was also a leap year and as a result there was an additional day of crime data. There are also many other longitudinal variables related to crime that are not mentioned in this study such as poverty rate and median household income but these variables are not related to weather and were excluded from analysis.

VIII. Conclusions

The results of this study indicate that climatological patterns can be extremely important and useful tools to use as predictors for many crimes. Despite controlling for other factors temperature appears to be significantly correlated with violent crime—robbery and felony assault in particular—andproperty crime—burglary and grand larceny. Murder, rape, and grand larceny auto were the only dependent variables that were found to have an insignificant relationship with temperature and an in depth explanation for the potential reasons surrounding these results was provided in the discussion section.

Precipitation was not found to have a statistically significant link to any of these crimes, perhaps indicating that New York City’s crime rate generally is not heavily affected by rain. While there was The New York Times study that showed the crime rate in New York City dropped on days it did rain (Lehren & Hauser, 2009), most academic journal articles and longitudinal and archival studies indicate the relationship is likely quite weak or non-existent.

The benefits of theseweather-related studies are numerous. If Walmart can correlate significant weather events with a dramatic uptick in purchases of water, duct tape, and strawberry Pop-Tarts and stock their stores accordingly(Pearsall, 2010), there are ramifications for predictive policing,as well, as “police planners may find such climatological models handy for scheduling patrol deployments and officers vacations” (Heller & Markland, 1970). Armed with knowledge of a greater possibility of violent crimes, cities can also deploy more ambulances in the warmer months to be better prepared for surges in call volume and a greater number of injuries.

Additionally, with regional temperatures “expected to rise between 3.6 and 7.2 degrees Fahrenheit by 2100” (Gamble & Hess, 2012) because of climate change, the rate of several types of violent and property crime nationwide could potentially rise linearly with the rise in temperature creating public safety concerns and economic concerns. As a result of this potential rise, the economic consequences of ignoring the potential challenge—or not conducting further analysis or finding ways to mitigate the correlation—could be significant. According to Ranson (2012), “the present discounted value of the social costs of these climate-related crimes is between 20 and 68 billion dollars” between2010 and 2099. While these predictions assume the rate of crime will remain the same, statistics have indicated that all types of crime are continuing to drop nationwide. New York City, in particular, has seen an astonishing drop in crime that no other city nationwide has come even close to rivaling (MacDonald, 2013).

Even still, researchers from Iowa State University calculated that the aggregate total cost for victims, the criminal justice system and the computedfallout from lost productivity for one single murder amount to $17.25 million, and “…that each rape costs $448,532, each robbery $335,733, each aggravated assault $145,379 and each burglary $41,288” (Blow, 2010).New York City has made an effort to cut down on arrests for minor crimes, such as marijuana possession. These minor crimes aggravate the financial burden on taxpayers (Hamilton, 2013). Beyond direct financial costs, as a committee of the National Research Council concluded in May 2014, the social costs of crime and incarceration are incalculable and sustain generational cycles of poverty (National Research Council, 2014). As temperature continues to rise in the upcoming decades, we would be prudent to keep in mind its obvious effect on many types of crime on a nationwide level, and not just in New York City.

Whatever the statistical relationships may be across the nation, the study reveals that the correlation between many types of violent and property crime and temperature in New York City is positive and significant. The relationshipis complex, however, and still requires further and more intricate study including whether or not the relationship is linear or curvilinear. New York Cityis and will remain a challenging environment to predict crime and we would be remiss to assume the relationship between weather and crime would stay the same as time goes on.

The United States saw its warmest year on record in 2012 (Chuck, 2013) and in the case of the 19 percent decline in murders and 13 percent drop in shootings to begin the first arctic-like two months 2014, The New York Times declared that “…all the cold, snowy weather has undoubtedly helped keep things quiet” (Newman & Correal, 2014).To make things simple, perhaps the answer to the relationship between temperature and crime in New York City is easier than we think and can be summed up in two words: polar vortex. But given the latest (May 2014) warning from a U.S. government's Federal Advisory Committee that climate change is already "causing more frequent or intense heat waves" (Peralta, 2014), counting on the cold is not a practical long-term policing plan. If anything, this study suggests that law enforcement experts need to focus now on further exploring the degree of the causal relationship between warmer temperatures -- including longer and hotter summers -- and criminal behavior, and how best to acclimate their crime-fighting strategy to impending climate change.

References

Anderson, C. A. (1989). Temperature and aggression: Ubiquitous effects of heat on occurrence of human violence (Psychology of aggression). Psychological Bulletin, (1), 74.

Anderson, C. A., & Anderson, D. C. (1984). Ambient temperature and violent crime: Tests of the linear and curvilinear hypotheses. Journal of Personality And Social Psychology, 46 (1), 91-97. doi:10.1037/0022-3514.46.1.91

Baron, R. A., & Bell, P. A. (1976). Aggression and heat: the influence of ambient temperature, negative affect, and a cooling drink on physical aggression. Journal of Personality & Social Psychology, 33 (3), 245-255.

Baron, R. A., & Ransberger, V. M. (1978). Ambient temperature and the occurrence of collective violence: the "long, hot summer" Revisited. Journal of Personality & Social Psychology, 36 (4), 351-360.

Belknap, J. (1987). Routine activity and the risk of rape: analyzing ten years of national crime survey data. Criminal Justice Policy Review, 2 (4), 337. doi:10.1177/088740348700200403

Bell, P.A., & Baron, R.A. (1976). Aggression and heat: the mediating role of Negative Affect. Journal of Applied Psychology, 6 (1), 18. doi:10.1111/j.1559-1816.tb01308.x

Blow, C. (2010, October 8). High Cost of Crime. The New York Times.

Breetzke, G. (2012). The effect of altitude and slope on the spatial patterning of burglary. Applied Geography, 34 66-75.

Brunsdon, C., Corcoran, J., Higgs, G., & Ware, A. (2009). The influence of weather on local geographical patterns of police calls for service. Environment And Planning B: Planning And Design, 36 (5), 906-926. doi:10.1068/b32133

Burke, M., Hsiang, S., & Miguel, E. (2013). Weather and violence. The New York Times, p. SR12.

Butke, P., & Sheridan, S. (2010). An analysis of the relationship between weather and aggressive crime in Cleveland, Ohio. Weather, Climate, And Society, 2 (2), 127-139. doi:10.1175/2010WCAS1043.1

Cahoon, D. D. (1977). Moon phase in behavioral research. Psychological Reports, 41 (1), 803-810.

Celona, L. (2012, July 04). Shoot spree has cop big shots under gun. The New York Post. Retrieved from http://nypost.com/2012/07/04/shoot-spree-has-cop-big-shots-under-gun/

Chen, P., Shu, X., Yuan, H., & Li, D. (2011). Assessing temporal and weather influences on property crime in Beijing, China. Crime, Law & Social Change, 55 (1), 1-13. doi:10.1007/s10611-010-9264-3

Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: a routine activity approach. American Sociological Review, 44 (4), 588-608.

Cohn, E. G. (1993). The prediction of police calls for service: The influence of weather and temporal variables on rape and domestic violence. Journal of Environmental Psychology, 13 (1), 71-83.

Cohn, E. G., & Rotton, J. (1997). Assault as a function of time and temperature: amoderator-variable time-series analysis. Journal of Personality & Social Psychology, 72 (6), 1322-1334.

Cook, M., Emick, D., Foley, C., Beatty, A., McCormick, C., & Shapiro, M. (2012). Daily temperature predicts assault and may allow more efficient policing. The American Surgeon, 78 (4), E194-E195.

Cotton, J. L. (1986). Ambient temperature and violent crime. Journal Of Applied Social Psychology, 16 (9), 786.

Coupe, T., & Blake, L. (2006). Daylight and darkness targeting strategies and risks of being seen at residential burglaries. Criminology, 44 (2), 431-464. doi:10.1111/j.1745-9125.2006.00054.x

DeFronzo, J. (1984). Climate and crime: tests of an FBI assumption. Environment & Behavior, 16 185-210. doi:10.1177/0013916584162003

Destefano, A. (2013, June 13). FBI crime stats mixed for NYC. Newsday.

Doleac, J. L., & Sanders, N. J. (2012). Under the cover of darkness: Using daylight saving time to measure how ambient light influences criminal behavior. Williamsburg, Va.

Doyle, J. (2010, July 19). Big Brother NYPD takes license with car thieves. The New York Post.

Eligon, J., & Santora, M. (2012, July 07). Unrelenting heat wave bakes all in its reach. The New York Times. Retrieved from http://www.nytimes.com/2012/07/08/us/temperatures-soar-as-heat-wave-continues.html

Enten, H. (2012, July 10). Are heatwaves a hotbed for crime? The Guardian.

Fallahi, F., Pourtaghi, H., & Rodríguez, G. (2012). The unemployment rate, unemployment volatility, and crime. International Journal of Social Economics, 39 (6), 440-448. doi:10.1108/03068291211224937

Feldman, H. S., & Jarmon, R. G. (1979). Factors influencing criminal behavior in Newark: A local study in forensic psychiatry. Journal of Forensic Sciences, 24 (1), 234-239.

Field, S (1992). The effect of temperature on crime. British Journal of Criminology, 32(3), 340.

Finklea, K. M. (2009). Economic downturns and crime: R40726. Congressional Research Service: Report, 1-15.

Fisher J.D., Bell, P.A., and Baum, A. (1984). Environmental Psychology, 2nded. New York: Holt, Rinehart, and Winston.

Franklin, C. A., Franklin, T. W., Nobles, M. R., & Kercher, G. A. (2012). Assessing the effect of routine activity theory and self-control on property, personal, and sexual assault victimization. Criminal Justice and Behavior, 39 (10), 1296-1315. doi:10.1177/0093854812453673

Gamble, J., & Hess, J. (2012). Temperature and violent crime in Dallas, Texas: Relationships and implications of climate change. Western Journal of Emergency Medicine, 13 (3), 236-246. doi:10.5811/westjem.2012.3.11746

Gardiner, S. (2010, July 26). Does crime rise in tandem with the mercury?. Wall Street Journal. Retrieved from http://online.wsj.com/news/articles/SB10001424052748703995104575389461974136120?mg=reno64-wsj

Gillis, J. (2013, January 08). Not even close: 2012 was hottest ever in U.S. The New York Times. Retrieved from http://www.nytimes.com/2013/01/09/science/earth/2012-was-hottest-year-ever-in-us.html?r=0

Grynbaum, M. M. (2014, March 11). De Blasio cites drop in crime since taking office. The New York Times. Retrieved from http://www.nytimes.com/2014/03/12/nyregion/de-blasio-cites-drop-in-crime-since-taking-office.html?_r=0

Hamilton, B. (2013, May 12). ‘High’ tolerance: NYC on the road to decriminalizing pot. The New York Post.

Harshbarger, R. (2012, July 08). Slayings soar as city hits boiling point. The New York Post. Retrieved from http://nypost.com/2012/07/08/slayings-soar-as-city-hits-boiling-point/

Heller, N., & Markland, R. (1970). A climatological model for forecasting the demand for police service. Journal of Research in Crime & Delinquency, 7 (2), 167-176.

Horrocks, J., & Menclova, A. (2011). The effects of weather on crime. New Zealand Economic Papers, (3), 231.

Jacob, B., Lefgren, L., & Moretti, E. (2007). The dynamics of criminal behavior: Evidence from weather shocks. Journal of Human Resources, 42 (3), 489-527.

John, R. H., Daniel, J. B., Patrick, J. C., & Kenneth, A. B. (2004). Crimes of opportunity or crimes of emotion? Testing two explanations of seasonal change in crime. Social Forces, 82 (4), 1333-1372.

Karmen, A. (2000). New York murder mystery: the true story behind the crime crash of the 1990s. New York: New York University Press.

Kenrick, D. T., & MacFarlane, S. W. (1986). Ambient temperature and horn honking: a field study of the heat/aggression relationship. Environment & Behavior, 18 179-191.

Lehren, A., & Baker, A. (2009, June 19). In New York, number of killings rises with heat. The New York Times, p. A1.

Lehren, A., & Hauser, C. (2009, July 2). In New York City, fewer murders on rainy days. The New York Times, p. A20.

Lewis, L. T., & Alford, J. J. (1975). The influence of season on assault. Professional Geographer, 27 (2), 214. doi:10.1111/j.0033-0124.1975.00214.x

Lindsay, J., & Anderson, C. (2000). From antecedent conditions to violent actions: A general affective aggression model. Personality and Social Psychology Bulletin, 26 (5), 533-547.

Lueck, T. (2007, December 31). Low murder rate brings New York back to ’63. The New York Times. Retrieved from http://www.nytimes.com/2007/12/31/nyregion/31murder.html?_r=0

MacDonald, H. (2013, October 27). Nothing rivals NYC’s crime drop. The New York Post.

Mancuso, V. (2014, March 12). New York went murder (and stabbing and shooting and slashing) free for a full 24 hours read more at http://observer.com/2014/03/new-york-went-murder-free-for-a-full-24-hours/

Mares, D. (2013). Climate change and crime: monthly temperature and precipitation anomalies and crime rates in St. Louis, MO 1990-2009. Crime, Law & Social Change, 59 (2), 185-208.

Mathias, C. (2012, November 28). New York City goes without murders, shootings, stabbings for more than one whole day. Huffington Post. Retrieved from http://www.huffingtonpost.com/2012/11/28/new-york-city-goes-one-day-without-murders-shootings-stabbings-one-whole-day_n_2207088.html

McDowall, D., Loftin, C., & Pate, M. (2012). Seasonal cycles in crime, and their variability. Journal of Quantitative Criminology, 28 (3), 389-410. doi:10.1007/s10940-011-9145-7

Michael, R., & Zumpe, D. (1983a). Annual rhythms in human violence and sexual aggression in the United States and the role of temperature. Social Biology, 30 (3), 263-278.

Michael, R., & Zumpe, D. (1983b). Sexual violence in the United States and the role of season. The American Journal of Psychiatry, 140 (7), 883-886.

Miller, W. H. (1968). Santa Ana winds and crime. Professional Geographer, 20 (1), 23. doi:10.1111/j.0033-0124.1968.00023.x

Monk, K. M., Heinonen, J. A., & Eck, J. E. (2010). Street robbery [electronic resource] / Khadija M. Monk, Justin A. Heinonen, and John E. Eck. Washington, DC: U.S. Dept. of Justice, Office of Community Oriented Policing Services, ^[2010].

National Research Council. The growth of incarceration in the United States: exploring causes and consequences. Washington, DC: The National Academies Press, 2014.

New York State Department of Labor, (2012). Local area unemployment statistics program: Kings County. Retrieved from http://www.labor.ny.gov/stats/laus.asp

Newman, A., & Correal, A. (2014, March 6). New York today: Crime drops. The New York Times.

NYPD. New York City Police Department, Office of the Chief of Department (2012). Compstat. Retrieved from http://www.nyc.gov/html/nypd/downloads/pdf/crime_statistics/cspbbn.pdf

Parascandola, R. (2012, November 28). No one reported shot, stabbed or slashed in New York City on Monday, police sources say. The New York Daily News. Retrieved from http://www.nydailynews.com/new-york/reported-shot-stabbed-slashed-nyc-monday-article-1.1209204

Parascandola, R. (2013, November 09). City goes 24 hours without a shooting, stabbing or slashing for the second time in year. New York Daily News. Retrieved from http://www.nydailynews.com/new-york/city-24-hours-murder-article-1.1511412

Pearsall, B. (2010). Predictive policing: the future of law enforcement? National Institute of Justice Journal, (266), 16-19.

Peralta, E. (2014, May 6). New report finds climate change already having broad impact. NPR.

Perry, J., & Simpson, M. E. (1987). Violent crimes in a city: Environmental determinants. Environment & Behavior, 19 77-90. doi:10.1177/0013916587191004

Poisson, J. (2012, July 26). Chief announces boost to police presence all summer. Toronto Star. Retrieved from http://www.thestar.com/news/gta/2012/07/26/chief_announces_boost_to_police_presence_all_summer.html

Ranson, M. M. (2014). Crime, weather, and climate change. Journal of Environmental Economics and Management, doi:10.1016/j.jeem.2013.11.008

Rosen, J. (2013, November 01). Predictive policing helps stop crime before it starts. Johns Hopkins Magazine. Retrieved from http://hub.jhu.edu/gazette/2013/november/news-roundup-predictive-policing

Rotton, J., & Cohn, E. (2003). Global warming and U.S. crime rates: An application of routine activity theory. Environment and Behavior, 35 (6), 802-825. doi:10.1177/0013916503255565

Rotton, J., & Frey, J. (1985). Air pollution, weather, and violent crimes: concomitant time-series analysis of archival data. Journal of Personality & Social Psychology, 49 (5), 1207-1220.

Ruderman, W. (2012, December 28). 414 homicides in ’12 is a record low for New York City. The New York Times, p. A1.

Sahagun, L. (2014, February 19). Climate change brings more crime. Los Angeles Times. Retrieved from http://www.latimes.com/science/sciencenow/la-sci-sn-climate-change-crime-20140219,0,2765136.story

Santora, M., & Schwirtz, M. (2013, December 28). An epidemic of carjackings afflicts Newark. The New York Times, p. A1.

Sledge, M. (2013, March 16). NYPD license plate readers will be able to track every car entering Manhattan. Huffington Post.

Sorg, E., & Taylor, R. (2011). Community-level impacts of temperature on urban street robbery. Journal of Criminal Justice, 39 (6), 463-470.

Tewksbury, R., & Mustaine, E. (2010). Cohen, Lawrence E., and Marcus K. Felson: Routine Activity Theory. Sage Publications, Inc.

Venugopal, A. (2011, December 08). Census pinpoints city’s wealthiest, poorest neighborhoods. WNYC. Retrieved from http://www.wnyc.org/story/174508-blog-census-locates-citys-wealthiest-and-poorest-neighborhoods/

Weather Underground. (2012, February 01). Weather history for Central Park, NY month of January, 2012. Retrieved from Weather history for Central Park, NY month of January 2012. (2012, February 01). Retrieved from http://www.wunderground.com/history/airport/KNYC/2012/1/26/MonthlyHistory.html

Weather Underground. (2014, April 10). Weather history for Central park, NY month of November, 2013. Retrieved from http://www.wunderground.com/history/airport/KNYC/2013/11/7/DailyHistory.html?req_city=NA&req_state=NA&req_statename=NA

Weiss, M. (2012, February 09). City crime spike blamed on warm winter. DNAinfo. Retrieved from http://www.dnainfo.com/new-york/20120209/manhattan/mother-nature-blame-for-city-crime-spike-cops-say

Zimring, F. E. (2012). The city that became safe: New York's lessons for urban crime and its control . New York: Oxford University Press.