Negative Binomial Modeling of Homicides in Sweden
(2021) STAH11 20202Department of Statistics
- Abstract
- It is crucial for a society to keep track of a serious phenomenon such as homicide, but also its underlying process. Having access to high-quality data enables this. Sweden has in the last few years seen an increase in deadly shootings, which has prompted societal debate regarding the matter. But very seldom does the overall context come into question -- how the general homicide tendencies have changed during this same time period? This paper examines how the overall annual homicide rates in Sweden have changed during the years of 1997 to 2019. The development is modeled via negative binomial regression, which reveals that there has been an actual overall increase. It is shown that the expected mean of annual homicides on average has been... (More)
- It is crucial for a society to keep track of a serious phenomenon such as homicide, but also its underlying process. Having access to high-quality data enables this. Sweden has in the last few years seen an increase in deadly shootings, which has prompted societal debate regarding the matter. But very seldom does the overall context come into question -- how the general homicide tendencies have changed during this same time period? This paper examines how the overall annual homicide rates in Sweden have changed during the years of 1997 to 2019. The development is modeled via negative binomial regression, which reveals that there has been an actual overall increase. It is shown that the expected mean of annual homicides on average has been decreasing by 1.45\%\ per year over this time period. But there is a 27\%\ increase for the latter period when comparing 1997 to 2012 and 2013 to 2019. Both of the covariates are shown to be statistically significant at a 95\%\ confidence level. The sharp and non-random shift after 2012 should prompt further research. (Less)
- Popular Abstract
- It is crucial for a society to keep track of a serious phenomenon such as homicide, but also its underlying process. Having access to high-quality data enables this. Sweden has in the last few years seen an increase in deadly shootings, which has prompted societal debate regarding the matter. But very seldom does the overall context come into question -- how the general homicide tendencies have changed during this same time period? This paper examines how the overall annual homicide rates in Sweden have changed during the years of 1997 to 2019. The development is modeled via negative binomial regression, which reveals that there has been an actual overall increase. It is shown that the expected mean of annual homicides on average has been... (More)
- It is crucial for a society to keep track of a serious phenomenon such as homicide, but also its underlying process. Having access to high-quality data enables this. Sweden has in the last few years seen an increase in deadly shootings, which has prompted societal debate regarding the matter. But very seldom does the overall context come into question -- how the general homicide tendencies have changed during this same time period? This paper examines how the overall annual homicide rates in Sweden have changed during the years of 1997 to 2019. The development is modeled via negative binomial regression, which reveals that there has been an actual overall increase. It is shown that the expected mean of annual homicides on average has been decreasing by 1.45\%\ per year over this time period. But there is a 27\%\ increase for the latter period when comparing 1997 to 2012 and 2013 to 2019. Both of the covariates are shown to be statistically significant at a 95\%\ confidence level. The sharp and non-random shift after 2012 should prompt further research. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9039716
- author
- Ryttmarker, Carl LU
- supervisor
- organization
- course
- STAH11 20202
- year
- 2021
- type
- M2 - Bachelor Degree
- subject
- keywords
- Swedish homicides, Cause-of-death registry, Generalized linear model, Negative binomial regression, Poisson-gamma mixture model
- language
- English
- id
- 9039716
- date added to LUP
- 2021-04-08 11:19:50
- date last changed
- 2021-04-09 10:19:53
@misc{9039716,
abstract = {{It is crucial for a society to keep track of a serious phenomenon such as homicide, but also its underlying process. Having access to high-quality data enables this. Sweden has in the last few years seen an increase in deadly shootings, which has prompted societal debate regarding the matter. But very seldom does the overall context come into question -- how the general homicide tendencies have changed during this same time period? This paper examines how the overall annual homicide rates in Sweden have changed during the years of 1997 to 2019. The development is modeled via negative binomial regression, which reveals that there has been an actual overall increase. It is shown that the expected mean of annual homicides on average has been decreasing by 1.45\%\ per year over this time period. But there is a 27\%\ increase for the latter period when comparing 1997 to 2012 and 2013 to 2019. Both of the covariates are shown to be statistically significant at a 95\%\ confidence level. The sharp and non-random shift after 2012 should prompt further research.}},
author = {{Ryttmarker, Carl}},
language = {{eng}},
note = {{Student Paper}},
title = {{Negative Binomial Modeling of Homicides in Sweden}},
year = {{2021}},
}