The Climate Change-Temperature-Crime Hypothesis: Evidence from a Sample of 15 Large US Cities, 2002 to 2015

• Author: Paul B. Stretesky,Kimberly L. Barrett,Michael A. Long,Michael J. Lynch
• DOI: 10.1177/0306624X20969934
• Date: 01 March 2022
• Published date: 01 March 2022
• Subject Matter: Articles

https://doi.org/10.1177/0306624X20969934

International Journal of

Offender Therapy and

Comparative Criminology

2022, Vol. 66(4) 430 –450

© The Author(s) 2020

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0306624X20969934

journals.sagepub.com/home/ijo

Article

The Climate Change-

Temperature-Crime

Hypothesis: Evidence

from a Sample of 15 Large

US Cities, 2002 to 2015

Michael J. Lynch1, Paul B. Stretesky2,

Michael A. Long3, and Kimberly L. Barrett4

Abstract

Drawing on prior studies, green criminologists have hypothesized that climate change

will both raise the mean temperature and the level of crime. We call this the “climate

change-temperature-crime hypothesis” (“CC-T-C”). This hypothesis is an extension

of research performed on temperature and crime at the individual level. Other

research explores this relationship by testing for the relationship between seasonality

and crime within a given period of time (i.e., within years). Climate change, however,

produces small changes in temperature over long periods of time, and in this view,

the effect of climate change on crime should be assessed across and not within years.

In addition, prior CC-T-C studies sometimes employ large geographic aggregations

(e.g., the entire whole United States), which masks the CC-T-C association that

appears at lower levels of aggregation. Moreover, globally, crime has declined across

nations since the early 1990s, during a period of rising mean global temperatures,

suggesting that the CC-T-C hypothesis does not fit the general trends in temperature

and crime over time. Addressing these issues, the present study assesses the CC-T-C

relationship for a sample of 15 large (N = 15) US cities over a 14-year period. Given

the CC-T-C hypothesis parameters, we assessed this relationship using correlations

between individual crime and temperature trends for each city. Crime trends were

measured by both the number and rate of eight Uniform Crime Report (UCR)

1University of South Florida, College of Behavioral and Community Sciences, Tampa, USA

2Northumbria University, Newcastle upon Tyne, Tyne and Wear, UK

3Oklahoma State University, Stillwater, USA

4Eastern Michigan University, Ypsilanti, USA

Corresponding Author:

Michael J. Lynch, Department of Criminology, SOC107, Fowler Avenue, University of South Florida,

College of Behavioral and Community Sciences, Tampa, FL 33620-8100, USA.

Email: mjlynch@usf.edu

969934IJOXXX10.1177/0306624X20969934International Journal of Offender Therapy and Comparative CriminologyLynch et al.

research-article2020

Lynch et al. 431

Part I crimes, so that for each city, there are 16 crime-temperature correlations.

Using a liberal p value (p = .10), the temperature-crime correlations were rejected as

insignificant in 220 of the 234 tests (94%). We discuss the Implications of this finding

and suggest that rather than focusing on the temperature-crime relationship, green

criminologists interested in the deleterious effects of climate change draw attention

to its larger social, economic, environmental and ecological justice implications.

Keywords

temperature and crime, climate change and crime, green criminology, climate change,

crime, US cities

The current study was designed to assess the hypothesis proposed in the criminologi-

cal literature that temperature changes that accompany climate change have a posi-

tive association with crime rates (Ratton & Cohn, 2003)—or, in other words, the

argument that climate change (CC), which is associated with increased temperatures

(T), leads to an increase in crime (C). We call this the “climate change-temperature-

crime” (or “CC-T-C”) hypothesis. This proposition has been offered as an extension

of more traditional criminological research which examines the relationship between

cyclical or within-year seasonal changes in temperature and crime (Rotton & Cohn,

2003), and the more specific suggestion by green criminologists that continued cli-

mate change and increasing temperatures will create conditions conducive to the

growth of crime (Agnew, 2012; White, 2016). While criminologists and other

researchers have examined the relationship between temperature and crime, those

studies tend to consider that relationship within years or relative to seasonal tempera-

ture cycles (measured across seasons, months, weeks or days). Whether the observed

season/cyclical relationship between temperature and crime also applies to longer-

term, cross-year temperature changes generated by climate change requires further

empirical attention. While a handful of prior studies state that they specifically mea-

sure a CC-T-C association, we suggest that there are two important limitations in the

designs of those studies. First, some use large geographic areas and are conducted at

too high a level of aggregation. The use of mean temperature to represent large geo-

graphic areas masks significant variability in temperature within those geographic

areas (e.g., the mean temperature of the United States (US) hides more local tempera-

ture variations). Second, these studies, though sometimes done across years, continue

to draw attention to assessing seasonality effects. The relationship between climate

change and temperature appears across longer spans of time (i.e., years), and should

be assessed across years and not seasonally.

To examine these issues, we begin with a summary of climate change research

across disciplines, followed by a review of the criminological research on the associa-

tion between temperature and crime. In assessing criminological studies of the temper-

ature-crime association, we note that the majority of studies test for “seasonality”—or

the variability in crime within or across years by month, weeks, and days, or seasons