Content uploaded by Craig A Anderson
Author content
All content in this area was uploaded by Craig A Anderson on Oct 24, 2016
Content may be subject to copyright.
247
16
Implications of Global Climate
Change for Violence Developed
and Developing Countries
CRAIG A. ANDERSON and MATT DELISI
Iowa State University
Rapid global climate change, taking place over decades rather than millen-
nia, is a fact of twenty-rst century life. Human activity, especially the
release of greenhouse gases, has initiated a general warming trend. The
10 warmest years on record between 1880 and 2008 were the last 10. This trend is
expected to continue until the atmospheric composition returns to a preindustrial
era norm.
Climate change effects on specic regions are expected to vary consider-
ably. Though most parts of the globe are warming, a few places may experience
cooler climates as ocean and wind currents shift. Some regions are experiencing
increased rainfall, whereas many others are having prolonged droughts. In 2007,
the Intergovernmental Panel on Climate Change (IPCC) released a report that
included numerous projections of likely effects by the end of this century, under
varying assumptions of how world governments, industries, and people respond.
The best-case scenario assumes huge reductions in net greenhouse gas production,
beginning almost immediately. In this scenario, climate models predict an average
global temperature increase of 1.8°C (3.24°F) and an average sea level increase
of 28 cm (11 inches). The worst-case scenario, which assumes a business-as-usual
approach, predicts increases of 4.0°C (7.2°F) and 43 cm (17 inches). Other pro-
jections, some of which have already become apparent, include increases in heat
waves and heavy precipitation; decreases in precipitation in subtropical areas; and
increases in tropical cyclones. More specic projections include (1) 5–8% increase
in the proportion of Africa that is arid and semiarid; (2) major ooding of heavily
populated areas of Asia from rising sea levels and storms; (3) inundation of low-
Y115381_C016.indd 247 11/2/10 10:38:16 AM
Anderson, C. A., & DeLisi, M.
(2011). Implications of global climate
change for violence in developed and
developing countries. Chapter in J.
Forgas, A. Kruglanski, & K.
Williams (Eds.), The Psychology of
Social Conflict and Aggression. (pp.
249-265). New York: Psychology
Press.
CRAIG A. ANDERSON AND MATT DELISI
lying islands; (4) severe water shortages in Australia and New Zealand; (5) drought
in southern Europe; (6) decreased soil moisture and food crops in Latin America;
and (7) increased winter ooding and summer heat waves in North America. More
recent research being prepared for the next IPCC report suggests that the new
best-case scenario will be worse than the old worst-case scenario, with sea levels
rising a least 1 meter (Vermeer & Rahmstorf, 2009). Because 13% of the world’s
population —hundreds of millions of people —live in low-lying coastal areas
(Engelman, 2009, p. 41), this latter projection is particularly disturbing. Indonesia
may lose as many as 2,000 small islands in the next 20 years to rising sea levels
(Engelman, p. 3).
In addition to the changes in average temperature and rainfall, climate
models also predict an increase in extreme weather events. Recent data suggest
that this increase has already begun, with dramatic increases in oods, wind
storms, and drought disasters in the last 20 years (Engelman, 2009, pp. 16, 30).
Hurricanes, cyclones, and other tropical storms also are increasing in intensity.
The problem with rising sea levels concerns not just the height of high tides but
also storm surge. A once-a-century storm in New York City, for example, will
occur about once every 3 years (Rahmstorf, 2009) if average sea level increases
by 1 meter.
Research from psychology, sociology, political science, economics, history, and
geography suggest that rapid global warming can increase the incidence of violent
behavior in at least three ways. One involves direct effects of uncomfortably warm
temperatures on irritability, aggression, and violence. A second involves indirect
effects of global warming on factors that put children and adolescents at risk for
developing into violence-prone adults. The third involves indirect effects of rapid
climate change on populations whose livelihoods and survival are suddenly at
risk, effects that inuence economic and political stability, migration, and violent
intergroup conict. For example, various governmental and scientic reports have
noted that climate change has exacerbated existing tensions and conicts singled
in the Darfur region of Sudan and in Bangladesh.
Heat and Aggression
Much research has established that uncomfortably warm temperatures can
increase the likelihood of physical aggression and violence (Anderson & Anderson,
1998; Anderson, Anderson, Dorr, DeNeve, & Flanagan, 2000; for a concise review
see Anderson, 2001). Three types of studies have tested and found considerable
support for this heat hypothesis: experimental studies, geographic region studies,
and time period studies.
Experimental Studies of the Heat Effect Early experimental studies of
heat effects yielded considerable inconsistency in outcomes, perhaps because of
participant suspicion and measurement issues. Later studies provided better tests
and cleaner results. For example, Vrij, van der Steen, and Koppelaar (1994) con-
ducted a eld experiment in which Dutch police ofcers were randomly assigned to
perform a training session involving a simulated burglary under hot or comfortable
Y115381_C016.indd 248 11/2/10 10:38:17 AM
250
conditions. Ofcers in the hot condition reported more aggressive and threatening
impressions of the suspect and were more likely to draw their weapon and to shoot
the simulated suspect.
Anderson et al. (2000) reported a series of laboratory experiments on both
hot and cold temperature effects. In separate experiments, uncomfortably warm
temperatures (relative to comfortable temperature) increased participants’ feelings
of anger and hostility, their perceptions of hostility in observed dyadic interac-
tions, and their initial retaliatory aggressive behavior against a person whose prior
harmful behavior was of an ambiguous nature. Recent experiments by Wilkowski,
Meier, Robinson, Carter, & Feltman (2009) linked heat-related imagery to a host
of anger and aggression-related perceptions and judgments.
Geographic Region Studies of the Heat Effect
Studies dating back to the nineteenth century suggest that hotter regions have higher
violent crime rates than cooler regions (Anderson, 1989). However, even within the
same country regions differ in many ways other than climate. Some of these other
differences (e.g., poverty, unemployment, age distribution, culture) are risk factors
for violence. The best geographic region studies include statistical controls for such
factors. Even when such factors are controlled, temperature predicts violent crime
rates. For example, hotter U.S. cities have higher violence rates than cooler cities,
even after statistically controlling for 14 risk factors including age, education, race,
economic, and culture of violence factors (Anderson & Anderson, 1996). Recent
work by Van de Vliert (2009; in press; under review) further suggests that climate and
economic conditions jointly inuence culture in ways that encourage or discourage
aggression and violence. Particularly vulnerable are populations that live in regions
that are both climatically challenging (hot, cold, or both) and impoverished.
Time Period Studies of the Heat Effect “Time period” studies compare
aggression rates within the same region but across time periods that differ in tem-
perature. Studies vary considerably in terms of the time periods for which violence
and temperature are assessed. Overall, results are remarkably consistent. Hotter
time periods (e.g., days, seasons, years) are associated with higher levels of vio-
lence. For example, riots in the United States are relatively more likely on hotter
than cooler days (Carlsmith & Anderson, 1979). Similarly, violent crimes across a
wide range of countries and measures occur more frequently during hotter seasons
than cooler ones (Anderson, 1989).
Of course, other violence-related factors may differ between hotter versus cooler
time periods, even within the same region. For example, in the United States large
numbers of youth are out of school during the summer months, so one could argue
that the routine activities of the population might account for seasonal differences
in violence. Several studies have addressed this and other alternative explanations
of heat-related time period effects. Although it is clear that routine activities do
inuence aggressive behavior, it is also clear that such alternative explanations do
not parsimoniously account for many observed effects. For example, in two stud-
ies Anderson and Anderson (1984) found signicant day-of-week effects on daily
Y115381_C016.indd 249 11/2/10 10:38:17 AM
IMPLICATIONS OF GLOBAL CLIMATE CHANGE 251
CRAIG A. ANDERSON AND MATT DELISI
violent crime rates, in addition to heat effects. Other time-related routine activities,
such as youth being out of school in the summer, cannot account for the heat effect
found in Study 1 (Chicago), because that study included only the summer months.
Similarly, routine activity theory cannot account for the nding that Major League
Baseball pitchers are more likely to hit batters with a pitched ball on hot days than
on cool days, even after statistically controlling for the possibility of sweat inuenc-
ing the pitcher’s control (Reifman, Larrick, & Fein, 1991).
Differences in violent crime rates for hotter versus cooler days have been
found within cities as varied as Houston, Chicago, and Minneapolis (Anderson &
Anderson, 1984, 1998). Even after controlling for routine activity effects of time of
day and day of week, violent crimes are relatively more frequent in hotter weather
(e.g., Anderson & Anderson, 1998; Bushman, Wang, & Anderson, 2005a, 2005b).
Interestingly, nonviolent crime (burglary, motor vehicle theft) rates are largely
unrelated to heat.
When the time period is years (instead of days or seasons), the kinds of poten-
tially confounded variables change. For example, U.S. youth are out of school in
the summer regardless of whether the year is slightly warmer or cooler. When con-
sidering year-based studies and global warming effects, one might be concerned
about whether aggression-related factors such as age distribution (e.g., proportion
of the population that is in the high-crime age range) and income inequality (e.g.,
LaFree & Drass, 1996) might be confounded with time or systematic temperature
changes. We conducted two new studies to examine the effects of yearly changes
in temperature on violent and nonviolent crime in the United States, beginning
with 1950.
Study 1: Hot Years and Violent Crime
Method
Data This study extends Anderson, Bushman, and Groom’s (1997) Study 1. Major
additions are 13 years of new data and several aggression-related control variables.
Data for the years 1950–2008 were obtained from U.S. government sources. From
the FBI’s Uniform Crime Reports we created two crime measures. Violent crime
was dened as the sum of the homicide and assault rates per 100,000 population.
Nonviolent crime was dened as the sum of the burglary and motor vehicle theft
rates per 100,000 population.1
The primary predictor variable, annual average temperature, was computed
from data from the National Oceanic and Atmospheric Administration. Control
variables were year, age (proportion of the population in the 15–29 high-crime
age range), prison (number of incarcerated state and federal inmates per 100,000
population), poverty (percent of families living below the poverty line), and the
Gini index of income distribution inequality (perfectly equal distribution yields a
1 As in prior studies, robbery and rape were excluded for theoretical reasons. Both appear to have a
greater mixture of aggressive motives (intent to harm) and nonaggressive motives. See Anderson et
al., 1997.
Y115381_C016.indd 250 11/2/10 10:38:17 AM
252
Gini index of 0, perfect inequality = 1.0). Year effects might reect a host of cul-
tural and population changes, such as increased reporting of assaults and improve-
ments in trauma care. The other control variables have obvious theoretical links
to violence.
Correlated Residuals Time-series data often have a problem in which the residu-
als are correlated with time. The most common version is when the residuals at
any given time period (T) are correlated with the residuals at the subsequent time
period (T + 1). Such “autocorrelations” make ordinary least squares (OLS) proce-
dures inappropriate for estimating regression parameters. With a sufciently large
sample of time periods, autoregression (AR) techniques can be used to reduce
or eliminate autocorrelations among residuals and can thus yield more accurate
results. In all regression analyses, chi-square tests (Ljung & Box, 1978) were used
to assess goodness of model t regarding the presence of correlated residuals.
When the chi-square statistic suggested that the model provided a poor t to the
data, autoregressive parameters were added. This process was iteratively repeated
until the chi-square test statistic indicated nonsignicant autocorrelations in the
new residuals.
The present study addresses ve alternative explanations for heat-related time
period effects on violent behavior: (1) seasonal uctuations; (2) correlated residu-
als; (3) coincidental crime, year, and global warming trends; (4) coincidental age
distribution shifts; and (5) coincidental income and poverty shifts. The rst alter-
native explanation is dealt with by using year as the unit of analysis. The remain-
ing alternatives are handled by statistical controls. Nonviolent crime analyses are
included as a point of comparison.
Results Table 16.1 presents descriptive statistics and zero-order correlations
among the variables. Average annual temperature has increased during this 59-year
period (r = .54). Note the substantial zero-order correlations among violent crime,
temperature, and year. This suggests that in addition to checking for autocorrelated
residuals a conservative statistical procedure would also control for year effects.
Finally, note that nonviolent crime was not strongly correlated with temperature.
Table 16.2 presents the results of OLS and AR analyses on violent crime (top
section) and nonviolent crime (bottom section). OLS regression revealed a large
effect of temperature on violent crime; each 1°F increase in average annual temper-
ature was associated with 79 more serious and deadly assaults per 100,000 people.
However, the AR test revealed signicant autocorrelations among the residuals (χ2
(6) = 196, p < .05). We added AR parameters to the model until the autocorrelation
test became nonsignicant (three parameters were needed). This greatly reduced
the slope relating temperature to violent crime, but this heat effect remained sta-
tistically and practically signicant. In the next step we controlled for year. The
temperature effect on violent crime remained essentially unchanged. The year
effect also was signicant; each year added 4.90 violent crimes per 100,000 people.
We examined a host of models with the other control variables (age, prison rate,
poverty, Gini). Only prison rate yielded a signicant effect. With three autoregres-
sive parameters—temperature, year, and prison—in the model the temperature
Y115381_C016.indd 251 11/2/10 10:38:17 AM
IMPLICATIONS OF GLOBAL CLIMATE CHANGE 253
CRAIG A. ANDERSON AND MATT DELISI
TABLE 16.2 Destructive Testing Results Using Auto-Regressive
Parameters and Competitor Variables, Study 1, 1950–2008
ViolentaAR Test Temperature Effect
Model χ2df b SE t
OLS 196* 6 79 19.1 4.15* Year Effect
AR-3 3.80 3 4.11 1.25 3.30* b SE t Prison Effect
AR-3 4.07 3 4.16 1.25 3.33* 4.90 1.68 2.91* b SE t
AR-3 3.27 3 4.19 1.21 3.47* 8.34 1.88 4.43* -.42 .203 -2.07*
NonViobAR Test Age Effect
Model χ2df b SE t
OLS 186* 6 16,646 1778 9.37*
AR-3 4.85 3 9645 3321 2.90*
Temp, annual average temperature. Age, proportion of U.S. population in the 15–29 age range. AR,
autoregression. OLS, ordinary least squares.
a Serious and deadly assault: Assault + homicide.
b Nonviolent crime: Burglary + motor vehicle theft.
* p < .05. + p < .10 if t > 1.67.
TABLE 16.1 Correlations Among the Predictor and Outcome
Variables, Study 1, 1950–2008
Year Temp Age Prison Pov Gini Vio NVio
Year 1.00 0.54 0.04 0.90 –0.75 0.88 0.87 0.52
Temp 0.54 1.00 –0.19 0.61 –0.18 0.61 0.48 0.11
Age 0.04 –0.19 1.00 –0.36 –0.44 –0.30 0.25 0.78
Prison 0.90 0.61 –0.36 1.00 –0.43 0.97 0.69 0.13
Pov –0.75 –0.18 –0.44 –0.43 1.00 –0.36 –0.71 –0.79
Gini 0.88 0.61 –0.30 0.97 –0.36 1.00 0.72 0.16
Vio 0.87 0.48 0.25 0.69 –0.71 0.72 1.00 0.75
Nvio 0.52 0.11 0.78 0.13 –0.79 0.16 0.75 1.00
Descriptive Statistics
Mean 1979 57.85 0.227 231 16.1 0.382 239.1 1300
St.Dev. 17.2 0.78 0.025 153 5.6 0.027 127.7 544
Min. 1950 56.60 0.195 93 11.1 0.348 55.7 385
Max. 2008 59.70 0.272 512 32.5 0.432 451.3 2163
Notes: N = 59. If r > .25 then p < .05. Temp, annual average temperature. Age, proportion of U.S.
population in the 15–29 age range. Prison, number of incarcerated state and federal
inmates per 100,000 population. Pov, percent of families living below the poverty line.
Gini, index of income distribution equality (perfectly equal distribution yields a Gini index
of 0; perfect inequality = 1.0). Vi, serious and deadly assaults per 100,000 population. Nvio,
burglaries and motor vehicle thefts per 100,000 population.
Y115381_C016.indd 252 11/2/10 10:38:17 AM
254
effect remained signicant (b = 4.19). Finally, the greater the proportion of the
U.S. population that was imprisoned, the smaller the violent crime rate.
A host of OLS and AR models on nonviolent crime did not yield a single sig-
nicant temperature effect. Indeed, after appropriate AR parameters were in the
model, only age was a signicant predictor of nonviolent crime rates. For every 1%
increase in the proportion of high-crime age individuals in the population, there
was an increase of 96 nonviolent crimes per 100,000 people.
Study 2: Hot Summers and Violent Crime
Method
Data This study extends Anderson et al.’s (1997) Study 2. It examines violent
crime in the summer months in the United States relative to nonsummer months.
Major additions are 9 years of new data and several aggression-related control vari-
ables. Data for the years 1950–2004 were collected from numerous governmental
sources. Seasonal data were unavailable after 2004.
Basically, the dataset is the same as for the prior study, with two major excep-
tions. First, the outcome variable is the difference between the percent of the
year’s crimes that were committed during the summer months (June, July, August)
and the average of the other three seasons, adjusted for number of days in each
season. If violent crimes were equally likely to occur regardless of season, then the
summer months would account for exactly 25% of them, and the summer effect
score used in this study would be zero. If violent crimes were relatively more (less)
likely in the summer months, the summer effect would be greater (less) than zero.
A similar summer effect was computed for nonviolent crimes.
The second major difference from Study 1 was the temperature measure.
Across a sample of cities, we recorded the number of hot days (maximum tempera-
ture was ≥ 90°F) per year. The vast majority of hot days in the continental United
States occur during the summer months, so this measure is a good indicator of the
hotness of each of the 55 summers.
Predictions We expected the summer effect on violent crime to be signicantly
greater than zero, when averaged across years. Furthermore, we expected years
with more hot days to yield larger summer effects on violent crime than years with
fewer hot days.
Results As expected, the average summer effect on violent crime was signi-
cantly greater than zero (M = 2.57, t(54) = 18.52, p < .05). Violent crimes are
overrepresented in the summer months. In fact, in only 2 of the 55 years was the
summer effect negative.
Concerning the second hypothesis, there was no evidence of autocorrelations
among the residuals in any of the analyses of violent crime, so OLS analyses were
appropriate. The only variable that signicantly predicted the size of the summer
effect on violent crime was the number of hot days (b = .068, t(53) = 3.07, p < .05).
None of the control variables (including year) had a signicant effect, nor did they
Y115381_C016.indd 253 11/2/10 10:38:17 AM
IMPLICATIONS OF GLOBAL CLIMATE CHANGE 255
CRAIG A. ANDERSON AND MATT DELISI
substantially reduce the size of the hot days effect. Nonviolent crime was unaf-
fected by number of hot days.
General Discussion of the Heat Effect on Aggression
In sum, the heat hypothesis has been repeatedly conrmed. Laboratory studies
suggest that this is largely the result of heat induced increases in irritability and
in hostile interpersonal perception biases. There is additional evidence that these
effects can be further traced to thermoregulation and emotion regulation areas
of the brain (Anderson, 1989; Boyanowsky, 1999, 2008; Boyanowsky, Calvert-
Boyanowsky, Young, & Brideau, 1981). The implication for global warming is that
at the level of the individual person increased exposure to uncomfortably hot
temperatures will increase the likelihood of interpersonal conict and violence.
It is difcult to estimate with condence how big an impact global warming will
have on violent crime in modern societies, but Figure 16.1 provides some rough
estimates based on the results of Study 1. If average annual temperature in the
United States increases by 8°F (4.4°C), the best estimate of the effect on the
total murder and assault rate is an increase of about 34 per 100,000 people, or
over 100,000 more such serious and deadly assaults per year in a population of
305 million.
One response to high heat in industrialized countries is increased use of air
conditioning in buildings, cars, buses, and trains. Although such actions might
mitigate heat-induced increases in aggression, they increase the production of
greenhouse gases.
There are no comparable daily, seasonal, or annual data on the heat effect on
violent crime in less developed countries. However, the ndings summarized in
previous sections suggest that uncomfortably hot temperatures can have a fairly
direct effect on aggressive and violent tendencies, perhaps through neuro and hor-
monal pathways that are common to thermoregulation and emotion (see Chapter
9 in this volume).
8
4.4
°F
°C
6
3.3
b =5.40 +1 S.E.
b =4.19 Best Estimate
b =2.98 +1 S.E.
4
2.2
2
1.1
0
Increase in Murder/Assault Rate
Increase in Average Annual Temperature
Increase in # of Murder/Assaults
50
40
30
20
10
0
140,000
120,000
100,000
80,000
60,000
40,000
20,000
0
missing
Y115381_C016.indd 254 11/2/10 10:38:18 AM
256
Figure 16.1. Heat E!ect on Violent Crime in the U.S., 1950-2008.
DEVELOPMENT OF VIOLENCE-PRONE INDIVIDUALS
Global climate change will likely increase the proportion of children and youth
exposed to risk factors known to increase the likelihood of becoming a violence-
prone individual— someone who frequently uses physical aggression or violence
to deal with conict, to get desired resources, and to impulsively and shortsight-
edly satisfy one’s wants (Gottfredson & Hirschi, 1990; Moftt, 1993). Studies of
violent youth and criminals reveal a host of psychological, neuropsychological,
genetic, and environmental risk factors that play a major role in determining who
becomes a violence-prone person. These interrelated risk factors include male gen-
der; strongly heritable antisocial traits including impulsivity, sensation seeking, low
intelligence, and poor self-regulation; poverty; poor prenatal and childhood nutri-
tion; familial dysfunction; growing up in violent neighborhoods; psychopathy; low
education; and disorganized and unstable neighborhood (DeLisi, 2005).
Food, Violence, and Antisocial Behavior
Potentially one of the most catastrophic effects of rapid climate change centers
on food availability. Today, one in eight U.S. households with infants is food inse-
cure— the family has limited or uncertain availability of nutritionally adequate
and safe foods. In many parts of the world, food insecurity is a much larger prob-
lem. This means that a robust proportion of impoverished children (notwithstand-
ing the multifaceted independent effects of poverty on antisocial behavior) face
the specter of poor nutrition or malnutrition—conditions with severe long-term
consequences for crime and violence. A recent study is illustrative. Jianghong Liu
and colleagues examined the longitudinal relationship between malnutrition and
subsequent externalizing and antisocial behaviors using a birth cohort of children
from the island of Mauritius, off the coast of Africa. Children who were malnour-
ished at age 3 were signicantly more aggressive and hyperactive at age 8, more
aggressive and prone to externalizing (acting out) behaviors at age 11, and more
hyperactive and more likely to exhibit symptoms of conduct disorder at age 17 (Liu,
Raine, Venables, & Mednick, 2004).
It is not merely armchair conjecture to assert that food scarcity will result in
increased violence-prone individuals; history has already told such a story. From
October 1944 to May 1945, residents of the western Netherlands experienced mod-
erate to severe food scarcity caused by a German army blockade. Over 100,000
Dutch men born between 1944 and 1946 were studied to examine the effects of
gestational nutritional deciency on subsequent proneness to violence (Neugebauer,
Hoek, & Susser, 1999). Men exposed to severe maternal nutritional deciency dur-
ing the rst and second trimesters were 2.5 times more likely than men not exposed
to severe maternal nutritional deciency to develop antisocial personality disorder,
a psychiatric diagnoses characterized by recurrent use of violence and other antiso-
cial behaviors. Other studies linking poverty to poor developmental outcomes are
reviewed by Huston and Bentley (2009). Similarly, recent work by Chen, Cohen, and
Miller (2010) reveals that poverty effects on children’s stress levels (assessed by corti-
sol) is exacerbated by perceived threat and by chaos in their daily living conditions.
Y115381_C016.indd 255 11/2/10 10:38:18 AM
IMPLICATIONS OF GLOBAL CLIMATE CHANGE 257
CRAIG A. ANDERSON AND MATT DELISI
Environmental–Genetic Interplay
Children in regions of famine, prolonged droughts, civil unrest, and wars (see next
section) are exposed to many known risk factors for the development of violence-
prone adolescents and adults. Longitudinal studies have shown that even fairly
brief exposures (e.g., a few months) to some of these risk factors can put the indi-
vidual child (or fetus) on a high-risk developmental trajectory.
Caspi and colleagues (2002) examined the interaction between monoamine oxi-
dase A (MAOA)—an enzymatic degrader that modulates neurotransmitters—and
childhood maltreatment on later antisocial outcomes. For all antisocial outcomes,
the association between maltreatment and antisocial behavior was conditional on
the MAOA genotype. Just 12% of the sample had both the genetic risk (low-activity
MAOA levels) and maltreatment; they accounted for 44% of the total convictions
for violent crime. Moreover, 85% of those who had both risk factors developed
some form of antisocial behavior. In the absence of maltreatment, the genotypic
risk factor did not manifest itself behaviorally. Similar gene–environment interac-
tions have been found for early life, environmental adversity, and psychiatric out-
comes (Caspi et al., 2003; Uher & McGufn, 2010).
If global warming brings about a world of dramatically increased environ-
mental risk and an unknown number of environmental pathogens, then it is likely
that a proportional proliferation of behavioral risks will result as these pathogenic
environments moderate genetic and neuropsychological risks within individuals.
Recall the pernicious and long-term effects of malnutrition and violent and antiso-
cial behavior. Malnutrition, particularly when it is endured during gestation, causes
a host of neuropsychological decits relating to neuronal reduction, brain toxicity,
altered neurotransmission, and other physiological effects. These neuropsycho-
logical decits also interact with genes to predict antisocial behavior. For example,
Beaver, DeLisi, Vaughn, and Wright (2010) found that neuropsychological decits
(such as those implicated by prenatal nutritionally deciency) interacted with the
low-activity polymorphism in the MAOA gene to predict violent behavior, delin-
quency, and low self-control across two time periods.
Terrorism Susceptibility
Recent research into terrorism and suicide bombers has led to a better under-
standing of the social and environmental conditions that are conducive the
development of individuals willing to use such extreme tactics (Kruglanski,
Chen, Dechesne, Fishman, & Orehek, 2009; see also Chapter 10 in this vol-
ume). Briey, these researchers have shown that such extremely violent tactics
can emerge from a “quest for personal signicance,” triggered by failure to
satisfy basic human motives to belong to a signicant group and to contribute
to its welfare. A variety of events can lead to feelings of failure and exclusion,
events such as personal trauma, loss of family through violence, and social
humiliation. Under the right (or wrong) circumstances, including (1) an avail-
able ideology to justify violence against the perceived perpetrators of trauma,
humiliation, and violent loss; and (2) social pressures to engage in violence
Y115381_C016.indd 256 11/2/10 10:38:18 AM
258
against the perpetrators as a means of gaining or restoring one’s own signi-
cance to one’s group, even suicidal terrorism becomes a viable option to the
individual (see Chapter 3 in this volume for a related discussion of aggression
and ostracism).
CIVIL UNREST, ECOMIGRATION, GENOCIDE, AND WAR
Both of the heat effect and the development of violence-prone individuals focus
on violence at the individual level. This third link between climate change and
violence focuses on larger groups of people—communities, tribes or clans, soci-
eties, and countries. This is a particularly complex set of phenomena. Emerging
research from several elds suggests that rapid climate change (heating or cooling)
often leads to increases in violence. There are several ways this can happen. For
example, in subsistence economies rapid changes in climate lead to a decreased
availability of food, water, and shelter. Depending on the level of social–political
organization, such shortages can lead to civil unrest and civil war, to migration
to adjacent regions and conict with the people who already live in that region,
and even to genocide and war. Although it would be overly simplistic to blame the
bloody conicts in Africa and Asia during the latter twentieth and this rst decade
of the twenty-rst century on climate change and environmental disasters, it also
would be incorrect to ignore the role played by the economic hardships (including
starvation) wrought by the prolonged droughts and resulting resource shortages.
Civil unrest, revolutions, and wars require recruits and leaders who are willing to
risk much to gain valuable resources.
Case Studies
Historical research shows that environmental disasters, many linked to relatively
rapid climate changes, can lead to increases in group-level violence. Of course,
not all environmental disasters are caused by climate change. For example,
earthquakes, tsunamis, and volcanoes can and do cause environmental disasters
but are not directly related to climate change. However, oods due to exces-
sive rainfall or melting glaciers, droughts, hurricanes, and cyclones are climate-
change related.
This section concerns whether environmental disasters increase violence rates
and severity, regardless of whether the environmental disaster was the direct result
of climate change.
In the recent past, evidence of such effects comes from the U.S. Dust Bowl of
the 1930s, clashes in Bangladesh and India since the 1950s, and Hurricane Katrina
in the United Sta tes in 2005 (Reuveny, 2 008). The c ases dif fer i n many ways, inc lud-
ing political organization and strength. But in each case, there is evidence that envi-
ronmental disaster led to increased interpersonal violence, a result of ecomigration
(migration of a large number of people as a result of ecological disaster).
Hurricane Katrina When Katrina hit Louisiana and Mississippi in fall 2005,
it ooded about 80% of New Orleans and destroyed much of the Biloxi–Gulfport
Y115381_C016.indd 257 11/2/10 10:38:18 AM
IMPLICATIONS OF GLOBAL CLIMATE CHANGE 259
CRAIG A. ANDERSON AND MATT DELISI
area. More than a million people left the area. This ecomigration was to at least
30 different states, with Texas (especially Houston) absorbing the most, at least
initially. Texas ofcials ran 20,000 criminal checks and found minimal criminal
data on their Katrina immigrants. Nonetheless, Houston recorded huge increases
in homicides in the following months, relative to the same months in the year
prior to Katrina (Reuveny, 2008). There were other indicators (e.g., polls) of
tension between the long-time residents and the newcomers. However, there
was no outbreak of civil war and no evidence of armed intergroup conict. This
seems to be generally true of ecomigrations in well-organized highly industrial-
ized countries.
U.S. Dust Bowl In the 1930s, poor farming practices combined with a pro-
longed drought and strong winds to produce an environmental disaster in the Great
Plains, particularly Oklahoma. About 2.5 million people left the area, primarily for
adjacent states, but about 300,000 went to California. There are numerous reports
of hostility and violence between the residents and the ecomigrants, including
police efforts to block the migrants or to scatter them from their settlements, beat-
ings, and shack burnings (Reuveny, 2008).
Bangladesh Population pressures from a very high fertility rate combined
with unsustainable farming practices and environmental disasters (possibly related
to climate change) led to large-scale migrations to adjacent regions in Bangladesh
and across the border to India. From 1976 to 2000 about 25 million people were
affected by droughts, 270 million by oods, and another 41 million by rain and
wind storms. Making matters worse, in 1975 the Indian Farakka Barrage began
diverting water from the Ganges River to other parts of India, decreasing the
amount owing into its historic tributaries in Bangladesh. The resulting salt-water
intrusion from the Indian Ocean and increased silting of the riverbed resulted in
additional oods, erosion, and environmental degradation.
An estimated 12 to 17 million Bangladeshis have migrated to adjacent states
in India since the 1950s. Clashes between the residents and the migrants have
occurred along socioeconomic, religious, ethnic, and national lines, resulting in
thousands of deaths, especially after the 1983 elections. Indeed, 1,700 Bengalis
were killed in a 5-hour rampage in 1983.
1967 Arab–Israeli War There is historical evidence of water issues con-
tributing to conict in the Middle East at least as early as the seventh century
B.C. (Gleick, 1993). Since the establishment of Israel in 1948 the region has peri-
odically been at war, for a variety of political and religious reasons. But water
issues also play an important role in the conicts, especially issues concerning the
Jordan River basin. This basin is shared by Israel, Jordan, Lebanon, and Syria.
According to Gleick (p. 85), “one of the factors directly contributing to the 1967
War was the attempt by members of the Arab League in the early 1960s to divert
the headwaters of the Jordan River away from Israel.” (For additional examples of
important water conicts, historical as well as contemporary, see Gleick; Postel &
Wolf, 2001).
Y115381_C016.indd 258 11/2/10 10:38:18 AM
260
Time-Period Studies
Little Ice Age Effects Following the Medieval Warm Period, the Little Ice
Age (roughly 1300–1850) ushered in cooler temperatures, shorter growing seasons,
and a host of other climate-related changes. Scholars from a variety of disciplines
have begun examining the relationships among relatively rapid shifts in climate
and a host of human population events, including war. Fagan (2000) weaves a care-
ful story of climate shifts and their impact on Europeans, linking farming practices
and outcomes, social and cultural changes, civil unrest, and war. Though careful
to avoid extreme claims of environmental determinism, he makes a strong case
for viewing rapid climate change (in this case, cooling) as contributing to war and
other forms of violence. Briey, rapid climate change disrupted food production,
leading to food shortages, famines, civil unrest, and war. This process seems par-
ticularly important in agrarian societies that do not have the political and economic
resources to effectively deal with food shortages and famine. Indeed, according to
Fagan the French revolution was fueled in part by food shortages that were largely
the result of the failure of farming practices to adapt to the changed climate.
Zhang and colleagues (Zhang, Brecke, Lee, He, & Zhang, 2007; Zhang, Zhang,
Lee, & He, 2007) took a more statistical approach to examining the question of
whether rapid shifts in climate from 1000 to 1900 were linked to wars. Using data
from the Northern Hemisphere and from China, they found statistical support for
their model, which is very similar to Fagan’s (2000).
It might seem strange to include studies of rapid cooling in a work that is
focused on global warming and violence. However, the basic model is the same
regardless of whether a rapid shift in climate is warming or cooling, ooding or
drought. A systematic change in climate that threatens basic human resources puts
stress on economic and social systems. That stress can lead to ecomigration and
conict or directly to war over resources.
Civil War in Africa Burke, Miguel, Satyanath, Dykema, and Lobell (2009)
recently analyzed civil wars in Africa from 1981 to 2002. Some models included
per capita income and form of government as well as temperature and precipi-
tation. Overall, the results showed a strong positive relation between tempera-
ture increases and civil war. For a 1°C increase in temperature, there was a 5.9%
increase in civil war. Given the base rate of civil war in this dataset (11%), this rep-
resents a 54% relative increase in the likelihood of civil war for each 1°C increase
in temperature. The authors noted that a 1°C increase is projected by 2030 and
that if future wars are as deadly as past ones an additional 393,000 battle deaths
can be expected in this region.
Additional Ecomigration and War-Related Forms of Violence
A recent report by the United Nations (Engelman, 2009) highlighted a number
of additional ways global climate change can lead to increased violence. Perhaps
the most notable is the likely increase in violent crimes committed against women
and children as a consequence of their increased vulnerability in subsistence
Y115381_C016.indd 259 11/2/10 10:38:18 AM
IMPLICATIONS OF GLOBAL CLIMATE CHANGE 261
CRAIG A. ANDERSON AND MATT DELISI
economies that suffer an ecological disaster. With the breakdown of societal norms
and increased economic stress come increases in rape, assault, and homicide. As
far as we know, there are no studies directly linking global warming to such effects,
but such outcomes have been documented in the aftermath of severe oods, food
shortages, and war (“civil” or otherwise).
IMPLICATIONS
Collectively, these three ways global climate change increases human violence
suggest a rather dire future. We prefer to end on a more positive note. Action
can be taken, by individuals, groups, and governments. One obvious action is to
reduce greenhouse gas emissions, thereby reducing the magnitude and speed of
climate change. Many individuals, groups, and governments are taking actions,
albeit somewhat belatedly.
In addition to the technological and lifestyle changes being actively developed,
discussed, and implemented, it also seems worthwhile to consider an infrequently
discussed option, the potential benets of better population control. One thousand
years ago the world population was about 300 million. Currently it is about 7 bil-
lion. Some have estimated that the world population will peak at around 10 billion.
Most of that increase will take place in developing countries, with huge increases
in greenhouse gas emissions as a result of carbon-intensive industrialization and
increasing consumption. Generally speaking, as a country becomes more indus-
trialized and wealthy, the carbon footprint per person increases dramatically, and
population growth eventually slows. The conundrum we face is how to reduce total
greenhouse gas emissions while improving the quality of life of the large propor-
tion of people currently living in poverty. One recent study found that, “dollar-for-
dollar, investments in voluntary family planning and girls’ education would also in
the long run reduce greenhouse-gas emissions at least as much as the same invest-
ments in nuclear or wind energy” (Engelman, 2009, p. 26).
Developed and developing countries will be affected differently by global
warming. In some ways, developed countries will be less affected, in part because
of their locations but more importantly because they have more resources per
capita to deal with the changes. It is unlikely that famines will strike the richest
countries, for example. However, no country will be immune to the violence con-
sequences of global climate change. The heat effect on individual levels of aggres-
sion and violence applies to all countries. Similarly, it seems obvious that even
wealthy countries are likely to see increases in the proportion of children exposed
to known risk factors for the development of violence-prone youth and adults. It
is less obvious how wealthy countries will be affected by the third process, which
leads to increases in civil unrest, ecomigration, genocide, and war. But even if
developed countries do not experience sufcient economic and social stress to
induce war (civil or international), civil unrest and ecomigration within them will
likely lead to increases in violent crime, especially after ecological disasters such
as oods. Furthermore, increased poverty, civil dissolution, and wars in develop-
ing countries have an impact on developed countries. In some cases, the impact
derives from the global economy and the need for resources. Also, differences
Y115381_C016.indd 260 11/2/10 10:38:18 AM
262
between the have and have-not countries create breeding grounds for interna-
tional terrorist groups.
What actions could reduce the likelihood of climate change induced violence?
There is some limited evidence that the heat–aggression effect on individuals can
be reduced by simply making people aware that when they are uncomfortably hot
they tend to react to minor provocations in inappropriately hostile ways. However,
given the immediacy and subtlety of the heat effect on irritability, hostile percep-
tion biases, and aggression, it is doubtful that such an educational intervention will
have a large impact.
On the other hand, the other two ways global warming increases human vio-
lence appear to be good candidates for intervention. If governments began prepar-
ing now to feed, shelter, educate, and move at-risk populations to regions in which
they can maintain their livelihoods and their cultures, we could dramatically
reduce both the development of violence-prone individuals and the civil unrest,
ecomigration, and war problems. This will cost huge amounts of money and will
require more international cooperation than our planet has ever seen. Failure to
do so will result in additional disasters for millions of people.
REFERENCES
Anderson, C. A. (1989). Temperature and aggression: Ubiquitous effects of heat on the
occurrence of human violence. Psychological Bulletin, 106, 74–96.
Anderson, C. A. (2001). Heat and violence. Current Directions in Psychological Science,
10, 33–38.
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, 91–97.
Anderson, C. A., & Anderson, K. B. (1996). Violent crime rate studies in philosophical con-
text: A destructive testing approach to heat and southern culture of violence effects.
Journal of Personality and Social Psychology, 70, 740–756.
Anderson, C. A., & Anderson, K. B. (1998). Temperature and aggression: Paradox, con-
troversy, and a (fairly) clear picture. In R. Geen & E. Donnerstein (Eds.), Human
aggression: Theories, research, and implications for social policy. (pp. 247–298). San
Diego, CA: Academic Press.
Anderson, C. A., Anderson, K. B., Dorr, N., DeNeve, K. M., & Flanagan, M. (2000).
Temperature and aggression. Advances in Experimental Social Psychology, 32,
63–133.
Anderson, C. A., Bushman, B. J., & Groom, R. W. (1997). Hot years and serious and deadly
assault: Empirical tests of the heat hypothesis. Journal of Personality and Social
Psychology, 73, 1213–1223.
Beaver, K. M., DeLisi, M., Vaughn, M. G., & Wright, J. P. (2010). The intersection of
genes and neuropsychological decits in the prediction of adolescent delinquency
and low self-control. International Journal of Offender Therapy and Comparative
Criminology, 54, 22–42.
Boyanowsky, E. O. (1999). Violence and aggression in the heat of passion and in cold blood.
International Journal of Law and Psychiatry, 22, 257–271.
Y115381_C016.indd 261 11/2/10 10:38:18 AM
IMPLICATIONS OF GLOBAL CLIMATE CHANGE 263
CRAIG A. ANDERSON AND MATT DELISI
Boyanowsky, E. O. (2008). Explaining the relationship among environmental tempera-
tures, aggression and violent crime: Emotional-cognitive stress under thermoregula-
tory conict (The ECS-TC Syndrome). Presented at the Biannual World Meeting
of the International Society for Research on Aggression, Budapest, Hungary, July
8–13.
Boyanowsky. E. O., Calvert-Boyanowsky, J., Young, J., & Brideau, L. (1981). Toward a ther-
moregulatory model of violence. Journal of Environmental Systems, 11, 81–87.
Burke, M. B., Miguel, E., Satyanath, S., Dykema, J. A., & Lobell, D. B. (2009). Warming
increases the risk of civil war in Africa. Proceedings of the National Academy of
Sciences, 106(49), 20670–20674.
Bushman, B. J., Wang, M. C., & Anderson, C.A. (2005a). Is the curve relating temperature
to aggression linear or curvilinear? Assaults and temperature in Minneapolis reexam-
ined. Journal of Personality and Social Psychology, 89, 62–66.
Bushman, B. J., Wang, M. C., & Anderson, C.A. (2005b). Is the curve relating temperature
to aggression linear or curvilinear? A response to Bell (2005) and to Cohn and Rotton
(2005). Journal of Personality and Social Psychology, 89, 74–77.
Carlsmith, J. M., & Anderson, C. A. (1979). Ambient temperature and the occurrence of
collective violence: A new analysis. Journal of Personality and Social Psychology, 37,
337–344.
Caspi, A., McClay, J, Moftt, T. E., Mill, J., Martin, J., Craig, I. W., Taylor, A., & Poulton,
R. (2002). Role of genotype in the cycle of violence in maltreated children. Science,
297, 851–854.
Caspi, A., Sugden, K., Moftt, T. E., Taylor, A., Craig, I. W., Harrington, H., McClay, J.,
Mill, J., Martin, J., Braithwaite, A., & Poulton, R. (2003). Inuence of life stress
on depression: Moderation by a polymorphism in the 5HTT gene. Science, 301,
386–389.
Chen, E., Cohen, S., & Miller, G. E. (2010). How low socioeconomic status affects 2-year
hormonal trajectories in children. Psychological Science, 21, 31–37.
DeLisi, M. (2005). Career criminals in society. Thousand Oaks, CA: Sage.
Engelman, R. (Ed.) (2009). The state of world population 2009. New York: United Nations
Population Fund.
Fagan, B. (2000). The Little Ice Age: How climate made history 1300–1850. New York:
Basic Books.
Gleick, P. H. (1993). Water and conict. International Security, 18(1), 79–112.
Gottfredson, M. R., & Hirschi, T. (1990). A general theory of crime. Stanford, CA: Stanford
University Press.
Huston, A. C., & Bentley, A. (2009) Human development in societal context. Annual Review
of Psychology, 61, 411–437.
Intergovernmental Panel on Climate Change. (IPCC). (2007). Parry, M.L., Canziani, O.F.,
Palutikof, J.P., van der Linden, P.J. & Hanson, C.E. (Eds.). Intergovernmental Panel
on Climate Change. Cambridge University Press, Cambridge, UK.
Kruglanski, A. W., Chen, X., Dechesne, M., Fishman, S., & Orehek, E. (2009). Fully com-
mitted: Suicide bombers’ motivation and the quest for personal signicance. Political
Psychology, 30, 331–357.
LaFree, G., & Drass, K.A. (1996). The effects of changes in intraracial income inequal-
ity and educational attainment on changes in arrest rates for African Americans and
Whites. American Sociological Review, 61, 614–634.
Ljung, G. M., & Box, G. E. P. (1978). On a measure of lack of t in time series models.
Biometrika, 64, 517–522.
Liu, J., Raine, A., Venables, P. H., & Mednick, S. A. (2004). Malnutrition at age 3 years
and externalizing behavior problems at ages 8, 11, and 17 years. American Journal of
Psychiatry, 161, 2005–2013.
Y115381_C016.indd 262 11/2/10 10:38:18 AM
264
Moftt, T. E. (1993). Adolescence-limited and life-course-persistent antisocial behavior: A
developmental taxonomy. Psychological Review, 100, 674–701.
Neugebauer, R., Hoek, H. W., & Susser, E. (1999). Prenatal exposure to wartime famine
and development of antisocial personality disorder in early adulthood. Journal of the
American Medical Association, 282, 455–462.
Postel, S. L., & Wolf, A. T. (2001). Dehydrating conict. Foreign Policy, 126, 60–67.
Rahmstorf, S. (2009). Climate seminar. Presented at the COWI conference, May 11,
Kongens, Lyngby, Denmark. Downloaded January 4, 2010 from http://www.youtube.
com/v/9bILFEdLy28.
Reifman, A. S., Larrick, R. P., & Fein, S. (1991). Temper and temperature on the diamond:
The heat-aggression relationship in Major League Baseball. Personality and Social
Psychology Bulletin, 17, 580–585.
Reuveny, R. (2008). Ecomigration and violent conict: Case studies and public policy impli-
cations. Human Ecology. 36, 1–13.
Uher, R., & McGufn, P. (2010). The moderation by the serotonin transporter gene of envi-
ronmental adversity in the etiology of depression: 2009 update. Molecular Psychiatry,
15, 18–22.
Van de Vliert, E. (2009). Climate, afuence, and culture. New York: Cambridge University Press.
Van de Vliert, E. (in press). Climato-economic origins of variation in ingroup favoritism.
Journal of Cross-Cultural Psychology.
Van de Vliert, E. (under review). Bullying the media: Cultural and climato-economic read-
ings of press repression versus press freedom.
Vermeer, M., & Rahmstorf, S. (2009). Global sea level linked to global temperature.
Proceedings of the National Academy of Science, downloaded January 4, 2010
from http://www.pnas.org.
Vrij, A., van der Steen, J., & Koppelaar, L. (1994). Aggression of police ofcers as a func-
tion of temperature: An experiment with the Fire Arms Training System. Journal of
Community and Applied Social Psychology, 4, 365–370.
Wilkowski, B. M., Meier, B. P., Robinson, M. D., Carter, M. S., & Feltman, R. (2009). “Hot-
headed” is more than an expression: The embodied representation of anger in terms
of heat. Journal of Personality and Social Psychology, 9, 464–477.
Zhang, D. D., Brecke, P., Lee, H. F., He, Y. O., & Zhang, J. (2007). Global climate change,
war, and population decline in recent human history. Proceedings of the National
Academy of Science, 104, 19214–19219.
Zhang, D. D., Zhang, J., Lee, H. F., & He, Y. O. (2007). Climate change and war frequency
in Eastern China over the last millennium. Human Ecology, 35, 403–414.
Y115381_C016.indd 263 11/2/10 10:38:18 AM
IMPLICATIONS OF GLOBAL CLIMATE CHANGE 265
