The 1990s Homicide Decline

Abstract

Recent studies suggest that the United States was not the only country that experienced a drop in crime in the 1990s. However, this line of research is largely limited to a small number of developed countries and only recently has explored the extent to which homicide rates dropped in developing countries during this time. This study addresses these limitations by using group-based trajectory modeling to explore homicide trends from 1990 to 2005 in 53 developed and developing countries. The results indicate that while most countries experienced downward trends in homicide during this time, this trend was neither universal nor randomly distributed.

Keywords

homicide homicide trends cross-national criminology

Introduction

Since the early 1990s, the United States has experienced an unprecedented drop in levels of crime (Zimring, 2007). While the 1990s crime drop in the United States has been subject to numerous studies, considerably less research has explored whether similar drops occurred throughout the world (Tseloni, Mailley, Farrell, & Tilley, 2010). Thus far, existing cross-national research on the crime drop during this period has primarily focused on a relatively small number of industrialized countries in North America, Europe, Asia, and Oceania, and only one study has explored the extent to which countries experienced a drop in homicide (LaFree, Curtis, & McDowall, 2015). Consequently, it remains questionable as to the extent to which the results of studies that focus on these countries are generalizable to other parts of the world.

This study seeks to contribute to the existing literature on the universality of the crime drop around the turn of the 21st century by exploring cross-national homicide trends in a sample of 53 developed and developing countries. Data for this study come from the World Health Organization (WHO), and homicide trends are explored using group-based trajectory modeling. While there have been several cross-national studies of homicide rates and trends, this study is one of the first ones to explore international homicide trends around the turn of the 21st century (see also LaFree et al., 2015).

Literature Review: The 1990s Crime Decline

Beginning around the 1960s, the United States and several other developed countries experienced crime booms marked by years of near-constant increases in violent crime (LaFree & Drass, 2002). While the rapid increase in violent crime during the 1960s seemed like a historic turn for the worst, the trend appeared to reverse itself in the early 1990s when violent crime in the United States and many other Western countries experienced an unprecedented drop. After approximately 15 years of nearly continuous decline, the U.S. homicide rate in the mid-2000s had reached a level not seen since the 1960s.

Although the bulk of research on the crime drop in the 1990s has focused on the United States, research has found that several other countries also experienced a drop during this time. Crime trends in Canada (Zimring, 2007) and many Western European nations (Aebi, 2004) were remarkably similar to those of the United States throughout the 1990s. Although this was not true across all types of crime, many Western countries did experience a downward trend in homicide during this time (Aebi & Linde, 2010, 2012). Relatively less is known about whether crime dropped outside of Western developed countries due to the lack of available data. One study by Del Frate and Mugellini (2012) found that many non-Western countries experienced a drop in homicide beginning around the 1990s, although several countries had yet to experience such a drop. Recent reports by the United Nations (2013, 2015) also suggest that declines in homicide have occurred in many developing countries in recent years.

While these studies suggest similarities in homicide trends across the Atlantic, this literature is predominantly focused on countries within North America and Europe (for reviews, see Koeppel, Rhineberger-Dunn, & Mack, 2015; Nivette, 2011). This limitation is problematic as it is less clear as to the extent to which the results of cross-national homicide research focused on developed countries are generalizable to countries in less developed regions. Criminological theory (Neuman & Berger, 1988) and research suggests that economic development is associated with crime rates, although the nature of this relationship differs according to offense type with economic development being positively correlated with property offending, yet negatively correlated with violent crime (Bennett, 1991; Shichor, 1990). Furthermore, psychological research also casts doubt on the generalizability of results obtained from research conducted in Western, educated, industrialized, rich, and democratic (WEIRD) to non-WEIRD countries (Henrich, Heine, & Norenzayan, 2010). Thus, it is important that cross-national research considers crime trends in both developed and developing countries.

Recent research attempts to address these limitations by considering a more diverse sample of countries. Tseloni and colleagues (2010) used data from the International Crime Victims Survey (ICVS) to examine crime trends in 26 countries from 1989 to 2004. Although their results indicate that both developed and developing countries experienced a drop in crime, their study did not assess whether homicide, which is widely considered the most reliable indicator of crime in cross-national research (LaFree & Tseloni, 2006), also dropped during this time. Furthermore, the ICVS samples from developing countries are drawn from a single city within that country, which is problematic as crime trends within a particular city may not be representative of national trends (Baumer & Wolff, 2014).

LaFree and colleagues (2015) examined the extent to which homicide victimization rates have declined since 1990 in a sample of 55 developed and developing countries. Their analysis used fixed effects regression modeling to examine the extent of shared variation in homicide between countries at similar levels of wealth as well as those located in the same geographic region. The results suggested that homicide rates declined for WEIRD countries since the 1990s, although many non-WEIRD countries experienced no such decline. Specifically, countries located in Central and South America experienced a slight increase in homicide since the 1990s. The current study aims to further the understanding of the extent to which homicide rates declined cross-nationally throughout the 1990s by using a method which allows countries to be grouped together based on trends found in the data itself rather than the more usual approach of grouping together countries based on criteria such as economic development or geographic region.

The Current Study

This study explores the extent to which homicide victimization rates declined in a sample of 53 developed and developing countries around the turn of the 21st century. We address two primary research questions:

Research Question 1: Did homicide decrease throughout the world beginning in the early 1990s?

Research Question 2: To what extent was the magnitude of the homicide drop similar across countries?

National homicide trends are explored using group-based trajectory modeling (Nagin, 2005). To date, this method has primarily been applied to the study of the development of offending at the individual level, although it has also been used to study homicide trajectories at the macro level (Stults, 2010; Weisburd, Bushway, Lum, & Yang, 2004). Although the merits of this method have been debated as far as its value for theory testing (Nagin & Tremblay, 2005; Sampson & Laub, 2005), there is consensus for its value in exploratory data analysis as is being done in this study.

Data

Homicide data were taken from WHO. The data were originally organized by WHO from medically certified deaths submitted by national authorities of member states annually. These deaths are classified according to their cause following a standardized International Classification of Diseases (ICD), currently in its 10th version (ICD-10), which includes a specific categorization for intentional homicides. WHO homicide data are often regarded as the most reliable and widely used source of cross-national homicide data (LaFree & Tseloni, 2006; Neapolitan, 1997).

As the focus of this study is on homicide trends between 1990 and 2005, countries were selected for the sample based on data availability during this period. Only countries missing fewer than 3 consecutive years of homicide data between 1990 and 2005 were included in the analysis (LaFree & Tseloni, 2006). This criterion dropped the sample size from 75 countries to 53 (see Table 1). Many countries were dropped as a result of geopolitical events, such as the breaking up of the Soviet Union or the reunification of Germany, which continued to affect national borders into the early 1990s and resulted in the dissolution of old countries and formation of new ones.¹ Major geopolitical events such as this often force researchers to drop observations (LaFree & Tseloni, 2006). The final study sample includes both developed and developing countries across five continents.

Table 1.

Countries by Homicide Trajectory Group (n = 53).

Low		Medium-Low		Medium-High		High
Country	Rate	Country	Rate	Country	Rate	Country	Rate
Australia	1.62	Argentina	5.48	Belarus	9.74	Brazil	23.39
Austria	1.02	Bulgaria	3.50	Costa Rica	6.06	Colombia	69.26
Canada	1.67	Chile	3.90	Cuba	6.01	Guatemala	28.85
Czech Republic	1.52	Croatia	2.67	Estonia	15.68	Russia	24.30
Denmark	1.09	Finland	2.65	Kazakhstan	16.56	Venezuela	20.44
France	0.88	Hungary	2.72	Kyrgyzstan	11.46
Germany	0.90	Mauritius	2.58	Latvia	13.22
Greece	1.07	Poland	2.18	Lithuania	9.23
Ireland	0.84	Romania	3.64	Mexico	15.58
Italy	1.47	Slovakia	2.02	Moldova	11.37
Japan	0.56	Uruguay	4.94	Trinidad and Tobago	10.41
Kuwait	1.51	Uzbekistan	4.91	Ukraine	11.11
The Netherlands	1.17			The United States	7.94
New Zealand	1.68
Norway	0.98
Portugal	1.37
Singapore	0.89
Slovenia	1.60
South Korea	1.64
Spain	0.89
Sweden	1.11
Switzerland	1.11
The United Kingdom	0.75
Total average	1.19		3.43		11.11		33.25

Source. World Health Organization.

Analytic Method

The analysis proceeds in two stages. In the first stage, semiparametric group-based trajectory modeling (Nagin, 2005) was used to explore homicide rates over time. Homicide trajectories were estimated using the PROC TRAJ procedure available in SAS (Jones, Nagin, & Roeder, 2001). Trajectory analysis was selected as it offers the advantage of allowing the data to determine groups of nations based on the similarity of their homicide trajectories, rather than using some subjective criteria to classify groups such as geographic region or level of economic development. Due to the skew in the distribution of homicide rates, homicide trajectories were estimated using log-transformed values. The model selected for further analysis was determined based on fit statistics such as the Bayesian Information Criterion (BIC) statistic as well as practical considerations such as parsimony and adequate group size. Once the model for further analysis was determined, sample countries were categorized into the group with which they have the greatest probability of membership. The second stage of the analysis used this categorization to explore the extent to which the magnitude of the homicide drop was similar across countries.

Results

A series of trajectory models were run with between one and six groups to determine the best fitting model. Although model fit improved with each additional group, a four-group model was selected as the analytic model as it was the most parsimonious model that captured all of the features observable in the statistically better fitting five- and six-group models. Model diagnostics indicate good model fit as the average predicted probability of group membership for the entire model, as well as for each group, was .99.

The trajectories for the four-group model are depicted in Figure 1. The Low group is composed of approximately 42% of the population and is characterized by a steady downward trajectory throughout the observation period. The Medium-Low group is composed of 24% of the population and is also characterized by a steady downward trend with a slight “bump” experienced around 2000. The Medium-High group is similar in size to the Medium-Low group although this group experiences a “bump” in homicide during the early 1990s and declines thereafter. The High group represents approximately 9% of the countries and, unlike the other three groups, is characterized by an increase in homicide throughout the observation period. Thus, it appears that while homicide rates dropped in most of the sample countries between 1990 and 2005, there exists a small number of countries where homicide rates increased during this time.

Figure 1.

Four group trajectory model of log homicide rate (Years 1990-2005).

Table 1 illustrates the breakdown of the groups when countries are classified into the group with which they have the greatest probability of membership. Table 1 shows that the resulting groups are largely composed of countries with homicide rates at relatively similar levels. Low countries have homicide rates below 2 per 100,000 population, Medium-Low countries have homicide rates between 2 and 5, Medium-High countries have rates between 6 and 16, whereas High countries have rates above 20.

Figure 2 displays the extent to which groups are composed of countries from similar geographic regions. For instance, the Low group is largely composed of countries from Western Europe, North America, Oceania, and the Far East, whereas the countries in the High group are primarily found in South and Central America with the exception of Russia. Countries in the two middle groups are primarily found in the Americas and Eastern Europe. This figure also illustrates the lack of homicide data for much of the world, including all of Africa, most of Asia, and some South American countries. Altogether, the results of the first part of the analysis demonstrate (a) while most of the sample countries experienced a drop in homicide between 1990 and 2005, there was a group of countries where homicide rates increased; and (b) trajectories of homicide appear to be clearly distinguished by their average homicide level and geographic region.

Figure 2.

Geographical distribution of homicide trajectory groups.

To assess the magnitude of the homicide drop across each of the trajectory groups, annual homicide rates for each of the four groups were calculated relative to the homicide rate in 1990. Figure 3 shows how the percentage difference in the average homicide rate for each group changed relative to the group’s average rate in 1990. This figure reveals two interesting findings. First, countries in the High group experienced no drop in homicide as the average homicide rate after 1991 is greater than the observed rate in 1990. Second, this figure illustrates the drop in homicide among the other three groups. Among the Low and Medium-Low groups, the homicide decline follows a similar trend from 1990 to 2005. The Medium-High group follows a slightly different trajectory such that homicide rates in these countries initially rose during the early 1990s before declining thereafter. It is important to note that the inclusion of Colombia may distort the homicide trajectory of its group, as presented in Figure 3, as Colombia has maintained a very high homicide rate relative to other countries in the High group and experienced a substantial 50% reduction in its homicide rate during the observation period. When omitting Colombia from the analysis, the High group trajectory changes substantially and follows an increasing trend that equates to a 56% increase in homicide between 1990 and 2005.

Figure 3.

Percent change in homicide by trajectory group since 1990 (n = 53).

To further explore the magnitude of the homicide drop, homicide rates were regressed on their yearly percentage change in homicide using ordinary least squares (OLS) regression. Figure 4 displays the scatterplot and associated regression line, which confirm the relationship between average homicide level and change in homicide from 1990 to 2005. Each point corresponds to a country within the sample and is identified by its group according to the results of the group based trajectory model (GBTM). This figure indicates the extent to which (a) groups are largely clustered according to homicide rate, (b) most countries within the sample have relatively low average homicide rates during the 1990-2005 period, and (c) homicide rates were more likely to increase among those countries with the greatest homicide rates to begin with. Each unit increase in the average homicide rate corresponds to an average 3% increase in homicide between the years 1990 and 2005. In terms of estimated values, this results in a predicted change of −28.7% for Japan, the country with the lowest homicide rate in the sample, which experienced an actual 28% drop, and a 57% increase for Guatemala, the country with the highest homicide rate in the sample, which experienced an actual 64% increase.

Figure 4.

Scatterplot of average homicide rate and percentage change in homicide between 1990 and 2005 (without Colombia).

Table 2 reports the results from an OLS regression assessing the average change in homicide rates over time. Model 1 presents the results reported in Figure 3, highlighting that, when the full sample is considered, there is a positive association between the average homicide rate and the change in homicide between 1990 and 2005. Model 2 demonstrates that when excluding the High group countries, the coefficient is reduced from approximately 3% to a 1.85% change in homicide rates between 1990 and 2005 per additional unit in average homicide during this period. The reduced magnitude of the coefficient estimate is likely due to the high homicide rates among countries in this group as well as the fact that these countries experienced the greatest increase in homicide during this time. Although the magnitude of the coefficient estimate is reduced, there remains a positive association between the average homicide level and the change in homicide between 1990 and 2005.

Table 2.

Regression Coefficients—Dependent Variable: Change in Homicide Rate Between 1990 and 2005 (Std. Errors).

	Model 1Full regression	Model 2Without High group	Model 3High group as control
Average homicide rate (1990-2005)	3.028** (0.741)	1.851 (1.142)	1.668 (1.133)
High group identifier			45.908 (29.257)
Constant	−30.369** (6.820)	−26.572** (7.325)	−25.749** (7.337)
Sample size	51	47	51
R ²	.254	.055	.291

Source. World Health Organization.

p < .001.

Finally, Model 3 includes a dichotomous variable to control for countries in the High group. The results of Model 3 are consistent with those of Model 2 in that the magnitude of the coefficient is reduced, although the positive relationship remains when controlling for countries classified in the High group.²

Discussion

Criminologists have given a great deal of attention to the 1990s crime drop in the United States, although the question of whether a similar drop was experienced throughout the world has received relatively less attention. This study sought to expand upon research in this area by exploring (a) the extent to which homicide rates dropped in various developed and developing countries around the turn of the 21st century, and (b) the magnitude of change in homicide rates in countries from various regions. The results of this study are largely consistent with prior research which finds a general decline in homicide victimization in Western developed nations during this period (Aebi & Linde, 2014; Baumer & Wolff, 2014; LaFree et al., 2015). Moreover, the results are also in line with more recent research which finds that the general homicide decline in the Western world did not extend to certain regions such as Central and South America (LaFree et al., 2015).

The results of this study differ from prior research on cross-national homicide trends in two important ways. First, although there appeared to be an international drop in homicide around the turn of the 21st century, this phenomenon was neither universal nor randomly distributed as there was a small number of countries concentrated in certain parts of the world where homicide rates increased during this time. This result is consistent with LaFree and colleagues’ (2015) finding that homicide rates dropped in countries located in the European, Asian, and North American regions, whereas countries located in Central and South America experienced an increase in homicide rates between 1992 and 2010.

Second, there appeared to be a relationship between the magnitude and direction of change in homicide and the country’s average homicide rate during this time. The relative magnitude of the homicide drop decreased as the country’s average homicide rate during this time increased. Thus, the homicide drop was greatest in countries that already had the lowest homicide rates on average. This result is consistent with Del Frate and Mugellini’s (2012) finding of an increasing security gap between developed and developing countries and also parallels the results of recent neighborhood-level research by Sampson (2012) which finds an enduring effect in regard to levels of crime and violence at the neighborhood level. Specifically, Sampson finds that despite declining violence throughout the 1990s and into the mid-2000s, high rates of violence persisted in neighborhoods that were historically the most violent areas of the city. Similarly, low rates of violence persisted in neighborhoods that were previously less violent. Thus, despite a nearly 50% decline in rates of robbery and homicide citywide from 1995 to 2006,

neighborhoods do not switch places in their relative ranking during this period . . . high violence areas persist and low violence remain so. There is almost no relative change in position, yet violence plummeted and virtually all neighborhoods benefitted . . . (Sampson, 2012, pp. 110-111)

This pattern of results could reflect one of three possible explanations. First, the results may be due to pure chance. However, the geographical distribution of countries according to trajectory group suggests that the results are not due to chance as contiguous countries were often classified in the same trajectory groups. This pattern suggests a second explanation which is a possible spillover effect such that increased safety in one country spills over into neighboring countries. A similar phenomenon regarding spatial proximity has been found in the literature on neighborhoods and homicide (Morenoff, Sampson, & Raudenbush, 2001), although such a phenomenon has not been previously discovered at the cross-national level.

A third possible explanation is that similar trends in homicide rates may reflect global factors that affect multiple nations at once such as improvements in the global economy or increased human development. On this point, Baumer and Wolff (2014) noted that

extant literature also points to several other types of shared experiences across nations in recent decades (e.g. a rise in obesity and diabetes, decline in smoking prevalence and alcohol consumption; the mass movement of women into labor markets) tied generally to the diffusion of norms about “civility” and consumption, technologies, government and industry regulations, market prices, and other factors fueled by globalization. (p. 25)

The question remains as to why homicide would be on an upward trend in these countries while declining in most of the other countries examined in this study. It may be that the same countries that have the highest homicide rates are the ones least organized and prepared to deal with violence within their borders. Most often, countries with the highest levels of violence are also the ones with the worst economic conditions and least stable governments (Fund for Peace, 2014; Karstedt, 2015; World Bank, 2014). Although prior research has sought to link government stability and economic conditions to violent crime in a causal relationship, it is possible that the link between these factors and violent crime is spurious such that crime is just another indicator of a socially disorganized country.

This study contains some limitations. First, while this study includes a larger number of countries than has been previously used in the study of the crime drop around the turn of the 21st century, there remains a large number of countries for which homicide data are unavailable. In addition, this study only explored national trends in homicide, and while homicide may be the most reliable indicator of crime, it is not possible to draw any conclusions regarding whether there were drops in other crime types during this period. Finally, the study sample contained no countries from the Africa region due to a lack of data on homicide rates in these countries for the period in question. As homicide rates in African countries tend to be relatively high (United Nations Office on Drugs and Crime [UNODC], 2013), it is possible that the inclusion of countries from this region would change the above results.

In sum, the results of this study suggest that around the turn of the 21st century, wealthier countries not only had some of the lowest homicide rates in the world but also experienced some of the greatest declines in homicide. In contrast, the poorest countries which are also often the most violent and, therefore, have the greatest room for improvement were more likely to experience an escalation in homicide. Accordingly, the results suggest that like economic inequality, global homicide trends follow a similar pattern where the rich get richer and safer, while the poor get poorer and more dangerous.

Footnotes

Acknowledgements

The authors would like to thank James Lynch, Jean McGloin, and the anonymous reviewers for their feedback.

Authors’ Note

All authors contributed equally to this project.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Douglas B. Weiss is an assistant professor in the Department of Criminal Justice at California State University at San Bernardino. His research interests include developmental & life course criminology, comparative criminology, substance use, and corrections.

Mateus Rennó Santos is a PhD student in the Department of Criminology and Criminal Justice at the University of Maryland, College Park. He holds a Master’s in Sociology from the Federal University of Minas Gerais (UFMG, Brazil). He was previously a researcher at the Center for Crime and Public Safety Studies (CRISP / UFMG), and a research advisor for the State Government of Minas Gerais (Brazil).

Alexander Testa is a PhD student in the Department of Criminology and Criminal Justice at the University of Maryland, College Park. His research interests include the consequences of incarceration on life outcomes, macrostructural perspectives of crime and interpersonal violence, and criminal justice policy.

Sumit Kumar is a PhD candidate in the Department of Criminology and Criminal Justice at the University of Maryland, College Park. His research interests include policing and crime prevention, terrorism, comparative criminology and theory testing.

References

Aebi

M. F.

(2004). Crime trends in Western Europe from 1990 to 2000. European Journal on Criminal Policy and Research, 10, 163-186.

Aebi

M. F.

Linde

(2010). Is there a crime drop in Western Europe? European Journal on Criminal Policy and Research, 16, 251-277.

Aebi

M. F.

Linde

(2012). Crime trends in Western Europe according to official statistics from 1990 to 2007. In Van Dijk

Tseloni

Farrell

(Eds.), The international crime drop: New directions in research (pp. 37-75). New York, NY: Palgrave Macmillan.

Aebi

M. F.

Linde

(2014). The persistence of lifestyles: Rates and correlates of homicide in Western Europe from 1960 to 2010. European Journal of Criminology, 11, 552-577.

Baumer

E. P.

Wolff

K. T.

(2014). Evaluating contemporary crime drop(s) in America, New York City, and many other places. Justice Quarterly, 31, 5-38.

Bennett

R. R.

(1991). Development and crime: A cross-national, time-series analysis of competing models. The Sociological Quarterly, 32, 343-363.

Del Frate

A. A.

Mugellini

. (2012). The crime drop in “non-Western” countries: A review of homicide data. In Van Dijk

Tseloni

Farrell

(Eds.), The international crime drop: New directions in research (pp. 134-155). New York, NY: Palgrave Macmillan.

Fund for Peace. (2014). Fragile States Index, 10th Edition. Washington, DC. Retrieved from http://fsi.fundforpeace.org/rankings-2014

Henrich

Heine

S. J.

Norenzayan

(2010). The weirdest people in the world? Behavioral and Brain Sciences, 33, 61-83.

10.

Jones

B. L.

Nagin

D. S.

Roeder

(2001). A SAS procedure based on mixture models for estimating developmental trajectories. Sociological Methods & Research, 29, 374-393.

11.

Karstedt

(2015). Does democracy matter? Comparative perspectives on violence and democratic institutions. European Journal of Criminology, 12, 457-481.

12.

Koeppel

M. D.

Rhineberger-Dunn

G. M.

Mack

K. Y.

(2015). Cross-national homicide: A review of the current literature. International Journal of Comparative and Applied Criminal Justice, 39, 47-85.

13.

LaFree

Curtis

McDowall

(2015). How effective are our “better angels”? Assessing country-level declines in homicide since 1950. European Journal of Criminology, 12, 482-504.

14.

LaFree

Drass

K. A.

(2002). Counting crime booms among nations: Evidence for homicide victimization rates, 1956 to 1998. Criminology, 40, 769-800.

15.

LaFree

Tseloni

(2006). Democracy and crime: A multilevel analysis of homicide trends in forty-four countries, 1950-2000. The ANNALS of the American Academy of Political and Social Science, 605, 25-49.

16.

Morenoff

J. D.

Sampson

R. J.

Raudenbush

S. W.

(2001). Neighborhood inequality, collective efficacy, and the spatial dynamics of urban violence. Criminology, 39, 517-558.

17.

Nagin

D. S.

(2005). Group-based modeling of development. Cambridge, MA: Harvard University Press.

18.

Nagin

D. S.

Tremblay

R. E.

(2005). Developmental trajectory groups: Fact or a useful statistical fiction. Criminology, 43, 873-904.

19.

Neapolitan

J. L.

(1997). Cross-national crime: A research review and sourcebook. Westport, CT: Greenwood.

20.

Neuman

W. L.

Berger

R. J.

(1988). Competing perspectives on cross-national crime: An evaluation of theory and evidence. The Sociological Quarterly, 29, 281-313.

21.

Nivette

A. E.

(2011). Cross-national predictors of crime: A meta-analysis. Homicide Studies, 15, 103-131.

22.

Sampson

R. J.

(2012). Great American city: Chicago and the enduring neighborhood effect. Chicago, IL: University of Chicago Press.

23.

Sampson

R. J.

Laub

J. H.

(2005). Seductions of method: Rejoinder to Nagin and Tremblay’s developmental trajectory groups: Fact or fiction. Criminology, 43, 905-913.

24.

Shichor

(1990). Crime patterns and socioeconomic development: A cross-national analysis. Criminal Justice Review, 15, 64-78.

25.

Stults

B. J.

(2010). Determinants of Chicago neighborhood homicide trajectories: 1965-1995. Homicide Studies, 14, 244-267.

26.

Tseloni

Mailley

Farrell

Tilley

(2010). Exploring the international decline in crime rates. European Journal of Criminology, 7, 375-394.

27.

United Nations. (2013). Global study on homicide 2013: Trends, contexts, data. Vienna, Austria: United Nations Office on Drugs and Crime.

28.

United Nations. (2015). State of crime and criminal justice worldwide: Thirteenth United Nations congress on crime prevention and criminal justice. New York, NY: Retrieved from http://www.un.org/en/events/crimecongress2015/pdf/Factsheet_2_The_State_of_Crime_EN.pdf

29.

Weisburd

Bushway

S. D.

Lum

Yang

(2004). Trajectories of crime at places: A longitudinal study of street segments in the city of Seattle. Criminology, 42, 283-322.

30.

World Bank. (2014). World Development Indicators, 2014. Washington, DC: United Nations. Retrieved from http://data.worldbank.org/sites/default/files/wdi-2014-book.pdf

31.

Zimring

F. E.

(2007). The great American crime decline. New York, NY: Oxford University Press.