Abstract
Recent criminological research has used latent class growth analysis (LCGA), a form of group-based trajectory analysis, to identify distinct terrorism trends and areas of high terrorism activity at the country-level. The current study contributes to the literature by assessing the robustness of recent findings generated by one type of group-based analysis, LCGA, to changes in measurement and statistical methodology. Using data from the Global Terrorism Database (GTD), we consider the challenges and advantages of applying group-based analysis to macro-level terrorism data. We summarize and classify country-level patterns of domestic and transnational terrorism using two types of group-based analyses, LCGA and an alternative yet similar modeling approach, general mixture modeling (GMM). We evaluate the results from each approach using both substantive and empirical criteria, highlighting the similarities and differences provided by both techniques. We conclude that both group-based models have utility for terrorism research, yet for the purposes of identifying hot spots of terrorist activity, LCGA results provide greater policy utility.
Similar content being viewed by others
Notes
Unless otherwise noted, for the sake of parsimony we use interchangeably the terms country, nation, and territory.
This is true even if only countries with non-zero counts of terrorist activities in a given year are considered.
Arguably, some countries may have zero counts because of lack of reporting opportunity or media restrictions.
Counts are used instead of rates for two reasons. It is not clear what the appropriate denominator should be for computing the rates because both people and infrastructure can be the target of attacks. Second, population counts are not available for many of the countries in the dataset for the entire study period and therefore using rates would substantially decrease our sample size.
Countries with less than four years of data (n = 2) were excluded from the analysis. The analysis also excludes data from 1993. Original data for 1993 were lost by Pinkerton Global Intelligence Service (PGIS) and have never been recovered (LaFree and Dugan 2007). Fortunately, both LCGA and GMM can accommodate missing data (Muthén and Muthén 1998–2010; Nagin 2005) and because the 1993 data are missing completely at random, our parameter estimates should not be affected by systematic bias (Allison 2002).
The following countries were coded to reflect their changing geographic boundaries, with the time span in parenthesis representing the period in which the country had valid data values: Soviet Union, Yugoslavia (1970–1991); Armenia, Azerbaijan, Belarus, Estonia, Georgia, Kazakhstan, Kyrgyzstan, Latvia, Lithuania, Moldova, Russia, Tajikistan, Turkmenistan, Ukraine, Uzbekistan (1992–2007); Bosnia-Herzegovina, Croatia, Macedonia, Slovenia, Serbia-Montenegro also known as Federal Republic of Yugoslavia (1992–2007); Czechoslovakia (1970–1992); Czech Republic, Slovak Republic, Eritrea (1993–2007); Namibia (1990–2007); North Yemen, South Yemen (1970–1989); Yemen (1990–2007); and East Germany (1970–1989). Sample sizes for analyses change due to changes in countries over time.
It is important to note that the NB estimates more parameters than the ZIP or Poisson models. This is because NB models estimate a dispersion parameter for each of the 37 outcome variables. Thus it is possible that a ZIP model with fewer constraints might better fit the data. However, exploratory analysis indicated this was not the case. The BIC values for a ZIP model that allowed the threshold values to be freely estimated was still greater than the BIC values obtained by the NB models.
A number of efforts were undertaken to obtain convergence with the Poisson and ZIP models. For example, the data were top-coded at lower values for which convergence could be obtained, and then through an iterative process, parameter estimates produced by the more heavily truncated data were used as start values in the models estimated using less truncated data. In addition, the convergence criteria were relaxed. Still, the Poisson and ZIP models with more than four classes did not converge.
We also estimated zero inflated negative binomial models (ZINB), although these models are not commonly used in the literature. These models follow the same logic as the ZIP, but allow for over dispersion in the count portion of the model. The NB and ZINB produced identical log likelihoods and Mplus fixed the values of the threshold parameters in all classes to −15. This occurs when the probability of being a structural zero is virtually zero and suggests that the inflation parameters are not needed. This was further confirmed by comparing the BICs for the NB and the ZINB models. The NB produced a lower BIC value than the ZINB because the BIC levies a penalty for more complicated models that estimate more parameters. These results are available upon request.
In Mplus the BIC is defined as −2 log likelihood + p log n, where p is the number of parameters and n is the sample size (Kreuter and Muthén 2008). This differs from the computation used by Proc Traj in SAS.
Jeffries (2003) has criticized the LMR-LRT on the basis of its theoretical proof; however, it is unclear to what extent this mistake affects the use of this test in practice (Muthén and Muthén 1998–2010). Although there is some evidence that the computationally intensive Bootstrapped Likelihood Ratio Test (BLRT) provides the best assessment of model fit (Nylund et al. 2007), this test never reached convergence with our data.
We based these decisions on past research on country-level terrorism trends, which also found that altering the polynomial form does little to change the substantive findings (LaFree et al. 2010).
These models relax the assumption that the within class variance for the intercept is zero, but the variances are constrained to be the same across classes. This constraint was put in place for two reasons. First, models that allow growth parameters to vary both within and across classes are computationally demanding and often have problems with convergence. Second, supplemental analyses (not shown) indicated that models that allowed the intercept to vary both across and within groups produced higher BICs and poorer fit than models that constrained the intercept variance to be equal across groups.
The estimation of trajectory models involves an iterative process in which the estimation algorithm attempts to identify the global maximum of the likelihood function. In more complicated models, the algorithm is more likely to converge on a local, versus a global maximum (Feldman et al. 2009). To protect against this possibility, Mplus enables users to estimate models using multiple random start values. The log likelihood values for the best solutions are provided by Mplus and log likelihood values that can be replicated are trust worthy. Here we only present results for models in which the best log likelihood could be replicated.
Group 1’s increase post 2002 may be due to the classification of Iraq within this group and the inclusion of events that followed the United States led invasion of Iraq that began on 20 March 2003. Although the GTD are designed to exclude cases involving open combat between opposing armed forces, it is often difficult to distinguish between acts of terror and warfare. We also conducted the analysis excluding post 2003 Iraq data and the substantive results remain unchanged in terms of general country-level trends and identification of high risk countries; however, the increase in terrorist activity for group 1 is less striking (See also LaFree et al. 2010).
Yugoslavia, which ceased to exist as a political entity in 1991, appears to fit in trajectory Group 1 because they had very low levels of terrorist activity throughout their time series until the very end of their existence (LaFree et al. 2010).
Note, although not shown, all variance parameters are significantly different from zero. This indicates that within each class, the intercepts exhibit significant variability.
Once again, when we reanalyzed the data excluding post 2003 Iraq data the substantive results remain unchanged.
We emphasize Kreuter and Muthén’s (2008) conclusion to be aware that model comparisons will also produce varying conclusions based on the specific data and outcome.
References
Abadie A (2006) Poverty, political freedom, and the roots of terrorism. Am Econ Rev 96:50–56
Allison PD (2002) Missing data series: quantitative applications in the social sciences. Sage Publications, Thousands Oaks
Berk RA, MacDonald J (2008) Overdispersion and Poisson regression. J Quant Criminol 24:269–285
Blomberg SB, Hess GD, Weerapana A (2004) An economic model of terrorism. Confl Manag Peace Sci 21:17–28
Brame R, Mulvey EP, Piquero AR (2001) On the development of different kinds of criminal activity. Sociol Methods Res 29:319–341
Burgoon B (2006) On welfare and terror: social welfare and political-economic roots of terrorism. J Confl Resolut 50:176–203
Bushway SD, Sweeten G, Nieuwbeerta P (2009) Measuring long-term individual trajectories of offending using multiple methods. J Quant Criminol 25:259–286
Eggleston EP, Laub JH, Sampson RJ (2004) Methodological sensitivities to latent class analysis of long term criminal trajectories. J Quant Criminol 20:1–26
Feldman BJ, Masyn KE, Conger RD (2009) New approaches to studying problem behaviors: a comparison of methods for modeling longitudinal, categorical adolescent drinking data. Dev Psychol 45:652–676
Griffiths E, Chavez JM (2004) Communities, street guns, and homicide trajectories in Chicago, 1980–1995: merging methods for examining homicide trends across space and time. Criminology 42:941–978
Hall D (2000) Zero-inflated Poisson and binomial regression with random effects: a case study. Biometrics 56:1030–1039
Hoffman B (1998) Inside terrorism. Columbia University Press, New York
Jeffries NO (2003) A note on ‘testing the number of components in a normal mixture model’. Biometrika 90:991–994
Jung T, Wickrama KAS (2008) An introduction to latent class growth analysis and growth mixture modeling. Soc Pers Psychol Compass 2(1):302–317
Koch MT, Cranmer S (2007) Testing the “Dick Cheney” hypothesis: do governments of the left attract more terrorism than governments of the right? Confl Manag Peace Sci 24:311–326
Kreuter F, Muthén B (2008) Analyzing criminal trajectory profiles: bridging multilevel and group-based approaches using growth mixture modeling. J Quant Criminol 24:1–31
LaFree G (1999) A summary and review of cross-national comparative studies of homicide. In: Smith MD, Zahn MA (eds) Homicide: a sourcebook of social research. Sage Publications, Thousand Oaks, pp 125–145
LaFree G, Dugan L (2007) Introducing the global terrorism database. Terror Polit Violence 19:181–204
LaFree G, Morris NA, Dugan L, Fahey S (2006) Identifying global terrorist hot spots. In: Victoroff J (ed) Tangled roots: social and psychological factors in the genesis of terrorism. IOS Press, Amsterdam, pp 98–114
LaFree G, Yang SM, Crenshaw M (2009) Trajectories of terrorism: attack patterns of foreign groups that have targeted the United States, 1970 to 2004. Criminol Public Policy 8:445–473
LaFree G, Morris NA, Dugan L (2010) Cross-national patterns of terrorism. Brit J Criminol 50:622–649
Li Q (2004) Economic globalization and transnational terrorism: a pooled time-series analysis. J Confl Resolut 48:230–258
Li Q, Schaub D (2005) Does democracy promote or reduce transnational terrorist incidents? J Confl Resolut 49:278–297
Lo Y, Mendell N, Rubin D (2001) Testing the number of components in a normal mixture. Biometrika 88:767–778
Long JS (1997) Regression models for categorical and limited dependent variables. Sage Publishing, Thousands Oaks
Messner SF, Rosenfeld R (1999) Social structure and homicide: theory and research. In: Smith MD, Zahn MA (eds) Homicide: a sourcebook of social research. Sage Publications, Thousand Oaks, pp 27–41
Moffitt TE (1993) Adolescence-limited and life-course persistent antisocial behavior: a developmental taxonomy. Psychol Rev 100:674–701
Mullins CW, Young J (2009) Cultures of violence and acts of terror: applying a legitimation-habituation model to terrorism. Crime Delinquency (Online access: doi:10.1177/0011128710364807)
Muthén B (2001) Second-generation structural equation modeling with a combination of categorical and continuous latent variables: new opportunities for latent class/latent growth modeling. In: Collins LM, Sayer A (eds) New methods for the analysis of change. APA, Washington, DC, pp 291–322
Muthén B (2004) Latent variable analysis: growth mixture modeling and related techniques for longitudinal data. In: Kaplan D (ed) Handbook of quantitative methodology for the social sciences. Sage Publications, Newbury Park, pp 345–368
Muthén B (2006) The potential of growth mixture modeling. Infant Child Dev 15:623–625
Muthén B, Asparouhov T (2006) Growth mixture analysis: models with non-Gaussian random effects. In: Fitzmaurice G, Davidian M, Verbeke G, Molenberghs G (eds) Advances in longitudinal data analysis. Chapman & Hall/CRC Press, Boca Raton, pp 143–165
Muthén LK, Muthén B (1998–2010) Mplus user’s guide, 6th edn. Muthén & Muthén, Los Angeles
Muthén LK, Muthén BO (2006) Mplus: statistical analysis with latent variables: user’s guide, 4th edn. Muthén & Muthén, Los Angeles, CA
Nagin DS (2004) Response to ‘‘Methodological sensitivities to latent class analysis of long-term criminal trajectories.’’ J Quant Criminol 20:26–37
Nagin DS (2005) Group-based modeling of development. Harvard University Press, Cambridge
Nagin DS, Land KC (1993) Age, criminal careers, and population heterogeneity: specification and estimation of a nonparametric, mixed Poisson model. Criminology 31:327–362
Nagin DS, Odgers CL (2010) Group-based modeling in clinical research. In: Nelen-Hoekland A, Cannon T, Widger T (eds) Annual review of clinical psychology. Annual Reviews, Palo Alto, pp 109–138
Nagin DS, Farrington DP, Moffitt TE (1995) Life-course trajectories of different types of offenders. Criminology 33:111–139
Neapolitan JL (1996) Cross-national crime data: some unaddressed problems. J Crim Justice 19:95–112
Nylund KL, Asparouhov T, Muthén B (2007) Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equ Modeling 14:535–569
Osgood W (2000) Poisson-based regression analysis of aggregate crime rates. J Quant Criminol 16:21–43
Petras H, Masyn K (2010) General growth mixture models. In: Piquero A, Weisburd D (eds) Handbook of quantitative criminology. Springer, New York, pp 69–100
Piquero AR (2008) Taking stock of developmental trajectories of criminal activity over the life course. In: Liberman AM (ed) The long view of crime: a synthesis of longitudinal research. Springer, Washington, DC, pp 23–78
Piquero NL, Piquero AR (2006) Democracy and intellectual property: examining trajectories of software piracy. Ann Am Acad Political Soc Sci 605:104–127
Piquero AR, Blumstein A, Brame R, Haapanen R, Mulvey EP, Nagin DS (2001) Assessing the impact of exposure time and incapacitation on longitudinal trajectories of criminal offending. J Adolesc Res 16:54–74
Quetelet A (1984) Research on the propensity for crime at different ages. (Sylvester SF, Trans.) Anderson Publishing Co., Cincinnati (Original work published in 1831)
Rapoport DC (2004) The four waves of modern terrorism. In: Cronin AK, Ludes JM (eds) Attacking terrorism: elements of a grand strategy. Georgetown University Press, Washington, DC, pp 46–73
Raudenbush SW (2005) How do we study what happens next? Ann Am Acad Political Soc Sci 602:131–144
Roeder K, Lynch KG, Nagin DS (1999) Modeling uncertainty in latent class membership: a case study in criminology. J Am Stat Assoc 94:766–776
Sampson RJ, Laub JH (2005) Seductions of method: rejoinder to Nagin and Tremblay’s ‘developmental trajectory groups: fact or fiction’. Criminology 43:905–913
Sampson RJ, Laub J, Eggleston EP (2004) On the robustness and validity of groups. J Quant Criminol 20:37–42
Sandler T, Enders W (2007) Applying analytical methods to study terrorism. Int Stud Perspect 8:287–302
Schmid AP, Jongman AJ (1988) Political terrorism: a new guide to actors, authors, concepts, databases, theories and literature. North-Holland Publishing, Amsterdam
Shaw CR, McKay HD (1942) Juvenile delinquency and urban areas. University of Chicago Press, Chicago [1969]
Shaw CR, Zorbaugh H, McKay HD, Cottrell LD (1929) Delinquency areas. University of Chicago Press, Chicago
Sherman LW, Weisburd D (1995) General deterrent effects of police patrol in crime ‘hot spots’: a randomized, controlled trial. Justice Quart 12:625–648
Sherman LW, Gartin PR, Buerger ME (1989) Hot spots of predatory crime: routine activities and the criminology of place. Criminology 27:27–55
Silke A (2005) Children, terrorism and counterterrorism: lessons in policy and practise. Terror Polit Violence 17:201–213
Stamatel JP (2009) Overcoming methodological challenges in international and comparative criminology: guest editor’s introduction. Int J Comp Appl Crim Justice 33:167–170
Weisburd D, Bushway S, Lum C, Yang SM (2004) Trajectories of crime at places: a longitudinal study of street segments in the city of Seattle. Criminology 42:283–321
Weisburd D, Morris NA, Groff ER (2009) Hot spots of juvenile crime: a longitudinal study of street segments in Seattle, Washington. J Quant Criminol 25:443–467
Yang SM (2010) Assessing the spatial-temporal relationship between disorder and violence. J Quant Criminol 26:139–163
Acknowledgments
We wish to thank the three anonymous reviewers that provided useful comments and suggestions to the current manuscript. We also thank Nicholas Corsaro and participants attending the Working Group Symposium on Quantitative Methods for Studying Terrorism at John Jay College, NY for their suggestions. We appreciate the thoughtfulness and thoroughness of their comments. Finally, we would like to thank Bengt and Linda Muthén for assistance with MPlus.
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
Countries by Trajectory Group Membership, LCGA Five Group Results
Group 1 (n = 29) | Group 2 (n = 49) | Group 3 (n = 64) | Group 4 (n = 31) | Group 5 (n = 28) | ||||
---|---|---|---|---|---|---|---|---|
Afghanistan | Andorra | Maldives | Albania | Liberia | Togo | Australia | Venezuela | Argentina |
Algeria | Antigua & Barbuda | Martinique | Armenia | Libya | Trinidad & Tobago | Austria | Zimbabwe | Chile |
Angola | Bahamas | Mauritania | Bahrain | Lithuania | Tunisia | Azerbaijan | Colombia | |
Bangladesh | Barbados | Mauritius | Bulgaria | Luxembourg | United Arab | Belgium | Corsica France | |
Bosnia-Herzegovina | Belize | North Korea | Cameroon | Macao | Emirates | Bolivia | El Salvador | |
Burundi | Benin | Qatar | Cen.African Rep | Madagascar | Uzbekistan | Brazil | France | |
Cambodia | Bermuda | Romania | Chad | Malaysia | Zambia | Canada | Germany | |
China | Bhutan | Seychelles | Congo | Mali | Costa Rica | Greece | ||
Egypt | Botswana | Singapore | Cuba | Malta | Croatia | Guatemala | ||
Georgia | Brunei | Solomon Islands | Czech Republic | Moldova | Cyprus | India | ||
Haiti | Burkina Faso | South Vietnam | Czechoslovakia | Morocco | Dominican Republic | Iran | ||
Indonesia | Belarus | South Yemen | Denmark | Namibia | East Germany | Israel | ||
Iraq | Cayman Islands | St Kitts and Nevis | Djibouti | New Caledonia | Ecuador | Italy | ||
Kenya | Comoros | Tanzania | East Timor | New Zealand | Ethiopia | Lebanon | ||
Macedonia | Dominica | Vanuatu | Estonia | Niger | Honduras | Mexico | ||
Myanmar | Equatorial Guinea | Vietnam | Fiji | North Yemen | Ireland | Nicaragua | ||
Nepal | Eritrea | Virgin Islands Us | Ghana | Norway | Japan | Northern Ireland | ||
Nigeria | Falkland Islands | Wallis & Futuna | Guadeloupe | Papua New Guinea | Jordan | Pakistan | ||
Rwanda | Finland | Western Samoa | Guinea | Paraguay | Kuwait | Palestine | ||
Serbia-Montenegro | French Guiana | Western Sahara | Guyana | Poland | Mozambique | Peru | ||
Somalia | French Polynesia | Hong Kong | Saudi Arabia | Netherlands | Philippines | |||
Soviet Union | Gabon | Hungary | Senegal | Panama | Russia | |||
Sudan | Gambia | Ivory Coast | Sierra Leone | Portugal | South Africa | |||
Tajikistan | Gibraltar | Jamaica | Slovak Republic | Puerto Rico | Spain | |||
Thailand | Grenada | Kazakhstan | Slovenia | Sweden | Sri Lanka | |||
Uganda | Guinea-Bissau | Kyrgyzstan | South Korea | Switzerland | Turkey | |||
Yemen | Iceland | Laos | Suriname | Syria | United Kingdom | |||
Zaire | Isle of Man | Latvia | Swaziland | Ukraine | United States | |||
Yugoslavia | Malawi | Lesotho | Taiwan | Uruguay |
Countries by Trajectory Group Membership, LCGA Four Group Results
Group 1 (n = 50) | Group 2 (n = 51) | Group 3 (n = 69) | Group 4 (n = 31) | |||||
---|---|---|---|---|---|---|---|---|
Afghanistan | Nepal | Andorra | Maldives | Albania | Laos | Suriname | Algeria | United Kingdom |
Angola | Nigeria | Antigua and Barbuda | Isle of Man | Armenia | Latvia | Sweden | Argentina | United States |
Australia | Panama | Bahamas | Martinique | Azerbaijan | Lesotho | Taiwan | Bangladesh | |
Austria | Portugal | Barbados | Mauritania | Bahrain | Liberia | Togo | Chile | |
Belgium | Puerto Rico | Belize | Mauritius | Bulgaria | Libya | Trinidad and Tobago | Colombia | |
Bolivia | Rwanda | Benin | North Korea | Cameroon | Lithuania | Tunisia | Corsica France | |
Bosnia-Herzegovina | Serbia-Montenegro | Bermuda | Qatar | Canada | Luxembourg | Ukraine | Egypt | |
Brazil | Somalia | Bhutan | Romania | Central African Republic | Macao | United Arab Emirates | El Salvador | |
Burundi | Sudan | Botswana | Seychelles | Chad | Madagascar | Uzbekistan | France | |
Cambodia | Switzerland | Brunei | Singapore | Congo | Malaysia | Zambia | Germany | |
China | Syria | Burkina Faso | Solomon Islands | Croatia | Mali | Yugoslavia | Greece | |
Costa Rica | Tajikistan | Belarus | St Kitts and Nevis | Cuba | Malta | Guatemala | ||
Cyprus | Thailand | Cayman Islands | Swaziland | Czech Republic | Moldova | India | ||
Dominican Republic | Uganda | Comoros | Tanzania | Czechoslovakia | Morocco | Iran | ||
East Germany | Uruguay | Dominica | Vanuatu | Denmark | Namibia | Israel | ||
Ecuador | Venezuela | Equatorial Guinea | Vietnam | Djibouti | New Caledonia | Italy | ||
Ethiopia | Yemen | Eritrea | Virgin Islands US | East Timor | New Zealand | Lebanon | ||
Georgia | Zaire | Falkland Islands | Wallis and Futuna | Estonia | Niger | Mexico | ||
Haiti | Zimbabwe | Finland | Western Samoa | Fiji | Norway | Nicaragua | ||
Honduras | Soviet Union | French Guiana | Western Sahara | Ghana | North Yemen | Northern Ireland | ||
Indonesia | French Polynesia | South Yemen | Guadeloupe | Papua New Guinea | Pakistan | |||
Iraq | Gabon | South Vietnam | Guyana | Paraguay | Palestine | |||
Ireland | Gambia | Hong Kong | Poland | Peru | ||||
Japan | Gibraltar | Hungary | Saudi Arabia | Philippines | ||||
Jordan | Grenada | Ivory Coast | Senegal | Russia | ||||
Kenya | Guinea | Jamaica | Sierra Leone | South Africa | ||||
Macedonia | Guinea-Bissau | Kazakhstan | Slovak Republic | Spain | ||||
Mozambique | Iceland | Kuwait | Slovenia | Sri Lanka | ||||
Myanmar | Malawi | Kyrgyzstan | South Korea | Turkey |
Countries by Trajectory Group Membership, Random Intercept GMM Four Group Results
Group 1 (n = 51) | Group 2 (n = 36) | Group 3 (n = 72) | Group 4 (n = 42) | |||||
---|---|---|---|---|---|---|---|---|
Afghanistan | Maldives | Albania | Senegal | Andorra | Greece | South Vietnam | Azerbaijan | Man, Isle of |
Algeria | Myanmar | Antigua and Barbuda | Sierra Leone | Argentina | Grenada | South Yemen | Belize | Martinique |
Angola | Namibia | Benin | Slovenia | Armenia | Guyana | Spain | Bolivia | Mauritius |
Bahrain | Nepal | Bermuda | Soviet Union | Australia | Iran | Sweden | Botswana | Moldova |
Bangladesh | Niger | Brunei | St Kitts and Nevis | Austria | Ireland | Switzerland | Burkina Faso | Mozambique |
Bhutan | Nigeria | Burundi | Togo | Bahamas | Israel | Syria | Chile | New Caledonia |
Bulgaria | Pakistan | Cameroon | Vietnam | Barbados | Italy | Tanzania | Croatia | Nicaragua |
Cambodia | Palestine | Cayman Islands | Virgin Islands Us | Belgium | Jamaica | Trinidad And Tobago | Cyprus | Panama |
Central African Rep | Poland | Czechoslovakia | Wallis and Futuna | Bosnia-Herzegovina | Jordan | Tunisia | Denmark | Peru |
Chad | Qatar | Czech Republic | Western Samoa | Brazil | Kuwait | Turkey | Djibouti | Romania |
Congo | Russia | China | Belarus | Latvia | Ukraine | Ecuador | Singapore | |
Cuba | Saudi Arabia | Comoros | Canada | Lebanon | United Arab Emirates | El Salvador | Slovak Republic | |
East Timor | Serbia-Montenegro | Equatorial Guinea | Colombia | Macedonia | United Kingdom | Estonia | South Africa | |
Egypt | Solomon Islands | French Polynesia | Corsica France | Mauritania | United States | Gibraltar | South Korea | |
Fiji | Somalia | Gambia | Costa Rica | Mexico | Uruguay | Guadeloupe | Suriname | |
Finland | Sri Lanka | Ghana | Dominica | Morocco | Vanuatu | Guatemala | Tajikistan | |
Guinea | Sudan | Hungary | Dominican Republic | Netherlands | Venezuela | Honduras | ||
Guinea-Bissau | Swaziland | Laos | East Germany | New Zealand | Western Sahara | Iceland | ||
Haiti | Taiwan | Liberia | Eritrea | Northern Ireland | Zambia | Japan | ||
Hong Kong | Thailand | Macao | Ethiopia | North Yemen | Zimbabwe | Kazakhstan | ||
India | Uganda | Madagascar | Falkland Islands | Norway | Lesotho | |||
Indonesia | Uzbekistan | Malawi | France | Paraguay | Libya | |||
Iraq | Yemen | Mali | French Guiana | Philippines | Lithuania | |||
Ivory Coast | Zaire | North Korea | Gabon | Portugal | Luxembourg | |||
Kenya | Yugoslavia | Papua New Guinea | Georgia | Puerto Rico | Malaysia | |||
Kyrgyzstan | Rwanda | Germany | Seychelles | Malta |
Countries by Trajectory Group Membership, Random Intercept GMM Five Group Results
Group 1 (n = 22) | Group 2 (n = 38) | Group 3 (n = 45) | Group 4 (n = 59) | Group 5 (n = 37) | |||||
---|---|---|---|---|---|---|---|---|---|
Afghanistan | Albania | Senegal | Angola | Norway | Andorra | Kuwait | South Vietnam | Azerbaijan | Nicaragua |
Algeria | Antigua & Barbuda | Sierra Leone | Bahrain | Palestine | Argentina | Latvia | Belize | Panama | |
Bangladesh | Benin | Slovenia | Bhutan | Paraguay | Armenia | Lebanon | Bolivia | Peru | |
Chad | Bermuda | Solomon Islands | Bulgaria | Philippines | Australia | Lesotho | Botswana | Romania | |
Guinea-Bissau | Brunei | St Kitts & Nevis | Cambodia | Poland | Austria | Mauritania | Burkina Faso | Singapore | |
East Timor | Burundi | Togo | Central African Republic | Qatar | Bahamas | Mexico | Chile | Slovak Republic | |
Indonesia | Cameroon | Vietnam | Colombia | South Korea | Barbados | Morocco | Croatia | South Africa | |
Iraq | Cayman Islands | Virgin Islands US | Congo | Sri Lanka | Belgium | Netherlands | Cyprus | Suriname | |
Ivory Coast | China | Wallis & Futuna | Corsica France | Swaziland | Bosnia-Herzegovina | Portugal | Denmark | Tajikistan | |
Kyrgyzstan | Comoros | Western Samoa | Cuba | Taiwan | Brazil | Puerto Rico | Ecuador | ||
Maldives | Czech Republic | Djibouti | Trinidad & Tobago | Belarus | Seychelles | El Salvador | |||
Nepal | Czechoslavkia | Egypt | Tunisia | Canada | Spain | Estonia | |||
Nigeria | Equatorial Guinea | Fiji | Uganda | Costa Rica | Sweden | Guadeloupe | |||
Pakistan | French Polynesia | Finland | Uzbekistan | Dominica | Switzerland | Guatemala | |||
Russia | Gambia | French Guiana | Vanuatu | Dominican Republic | Syria | Honduras | |||
Saudi Arabia | Ghana | Gabon | Yemen | East Germany | Tanzania | Iceland | |||
Serbia-Montenegro | Hungary | Gibraltar | Eritrea | Turkey | Japan | ||||
Somalia | Laos | Guinea | Ethiopia | Ukraine | Kazakhstan | ||||
Sudan | Liberia | Guyana | Falkland Islands | United Arab Emirates | Libya | ||||
Thailand | Macao | Haiti | France | United Kingdom | Lithuania | ||||
Zaire | Madagascar | Hong Kong | Georgia | United States | Luxembourg | ||||
Yugoslavia | Malawi | India | Germany | Uruguay | Malaysia | ||||
Mali | Israel | Greece | Venezuela | Malta | |||||
Namibia | Kenya | Grenada | Zambia | Man, Isle of | |||||
North Korea | Macedonia | Iran | Zimbabwe | Martinique | |||||
Papua New Guinea | Mauritius | Ireland | Northern Ireland | Moldova | |||||
Rwanda | Myanmar | Italy | Western Sahara | Mozambique | |||||
New Zealand | Jamaica | North Yemen | New Caledonia | ||||||
Niger | Jordan | South Yemen | Nicaragua |
Rights and permissions
About this article
Cite this article
Morris, N.A., Slocum, L.A. Estimating Country-Level Terrorism Trends Using Group-Based Trajectory Analyses: Latent Class Growth Analysis and General Mixture Modeling. J Quant Criminol 28, 103–139 (2012). https://doi.org/10.1007/s10940-011-9158-2
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10940-011-9158-2