The following research was aimed at finding the role in which demography plays for murder rates. The research began by looking into the trends of homicide rates during the pre-1800 era and the post-1800 era to see if the rates are on the change or consistent. Demo data was collected using a few techniques and proxy in order to make the, complete. The data were analyzed using descriptive analysis to determine their trends. A t-test was also carried out to examine the relationship between the dependent variable (homicide rate) and the independent variable p2029, p2034, m2029 and m2034). A regression analysis was then carried out on the homicide rate to find out the relationship between the hr and young share and young male share of 2 age groups. From the regression models, it was found out that age group 2029 shows higher effect in OLS than FE while age group of 20 to 34 age group shows the opposite. Moreover, age group of the 20 to 29 age group has more than twice high impact on homicide rates than that of age group 20-34. On the other hand, male share of age group of 20 to 29 age group has the greatest impact. Finally, the male share of age group 2034 is only statistically significant in FE.
Table of Contents
ABSTRACT
INTRODUCTION
Problem statement
Research objectives
Justification
LITERATURE REVIEW
General demographic variables
Age Distribution and Crime
Gender and crime
Theoretical Framework
DATA DESCRIPTION
Homicide rates
Post-1800
Pre-1800
Demographic variables
RESEARCH DESIGN AND METHODOLOGY
DATA ANALYSIS
Descriptive analysis
Correlation analysis
t-test statistics
Regression analysis
Regression of homicide rate on the youth share 20 to 29
Regression of homicide rate on the youth share in the 20 to 29 age group
Regression of the homicide rate on the young male share of the 20 to29 age group
Regression of the homicide rate on the young male share in the 20 to 34 age group
DISCUSSION
Conclusion
Strength
Weaknesses
Future research
REFERENCES
APPENDIX
ABSTRACT
The following research was aimed at finding the role in which demography plays for murder rates. The research began by looking into the trends of homicide rates during the pre-1800 era and the post-1800 era to see if the rates are on the change or consistent. Demo data was collected using a few techniques and proxy in order to make the, complete. The data were analyzed using descriptive analysis to determine their trends. A t-test was also carried out to examine the relationship between the dependent variable (homicide rate) and the independent variable p2029, p2034, m2029 and m2034). A regression analysis was then carried out on the homicide rate to find out the relationship between the hr and young share and young male share of 2 age groups. From the regression models, it was found out that age group 2029 shows higher effect in OLS than FE while age group of 20 to 34 age group shows the opposite. Moreover, age group of the 20 to 29 age group has more than twice high impact on homicide rates than that of age group 20-34. On the other hand, male share of age group of 20 to 29 age group has the greatest impact. Finally, the male share of age group 2034 is only statistically significant in FE.
INTRODUCTION
According to UNODC, the ultimate crime is the premeditated killing of a human being by another. The UNODC estimated the total number of deaths in relation with homicides in 2010 was 468,000 on a global basis. . Since 1995, the homicide rate has decreased in many countries, mainly in Asia, Europe, and Northern America, to the extent that it can be a relatively rare occurrence. Disparities are present on how homicides are scattered across the world. Moreover, there are also disparities in homicide typologies that show the varying levels of occurrence in different regions. As a result, the disparities make the researcher desire to see the problem from a global and long-term perspective. Homicide can occur through gangs, robbery, sexual motives, fights and family disputes. The UNODC report also takes a look at the demographics of homicide of those people who are a risk. From the report, women make up the bulk of the victims of homicides which are family-related or intimate partner homicide. The bigger picture also shows that males are often the most implicated in specifically accounting for about 80 percent of perpetrators and victims f homicide. Data from the US reveal that the typical homicide pattern is 69 percent of the cases, a man kills another man. On the other hand, less than 3 percent of the cases, a female kills another female which translates to the risk of men being murdered is higher compared to the women with the global homicide rates of 11.9 and 2.6 per 100,000 respectively.
Particularly, youthful males are at the most danger due to their more probable participation in activities prone with prone. The activities include gang membership, street crime, possession of weapons, street fighting, and drug consumption among others. In countries that are characterized by high homicide levels and organized crime, the risk of a man of 20 years is very high at a 2 percent. That indicates that 1 in 50 men in such countries is killed at that age. On the other hand, the danger in countries which have homicide rates which are low is four hundred times lower.
The sex and age comparison of victims of homicide greatly differ between regions. For instance, the distribution of victims of homicides who are female range from 10 percent in the Americas compared to 27 percent in Europe, another apparent indicator of the different homicide typologies rampant in those two regions. Moreover, almost twice that in the Americas, the highest rate of homicide among females globally is in Africa (6.2 per 100,000), where the rate of homicide is not driven by organized crime to the same extent, but street crime, non-specific lethal violence, and intimate partner or family-related homicide which all plays a vital role.
Problem statement
The world is experiencing high homicide rates. Resources need to be mobilized to tackle this global pandemic. The following study objective is to determine the role of democracy in murder rates. By knowing how the factors relate, it will give a picture on how the homicide rates can be predicted and therefore measures to be taken in advance to curb it.
Research objectives
1. To determine the relationship between the homicide trends with the youth share of 20 to 29 years.
2. To determine the relationship between the homicide trends with the youth share of 20 to 34 years.
3. To determine the relationship between the homicide trends with the young male share of 20 to 29 years.
4. To determine the relationship between the homicide trends with the young male share of 20 to 34 years.
Justification
Criminology theory studies give an opportunity for the analysis of crime via the rationalizations for the creation of criminal and criminal behavior. Every individual theory gets to explain for a crime reason while making sense on the causes for the criminal appeal. The ability to make sense on the dilemmas which impact behavior, change, and social structure makes it easier in understanding what needs to be done in the prevention of behavior and criminal actions. Theories of thought, both biological and classical explain crime through two different contemplations which are essential in the deviant behavior rationalization. However, most literature proves that there are no enough researchers on homicides rates. Moreover, there is a lack of explanation regarding the differences in demographics which impact on the homicide rates based on the differences in specific demographic factor. Thus, the researcher found out that it was vital to conduct the following study so as to reveal the effect of specific demographic factors on homicide rates.
The study will also give an insight into the trends of homicide rates during the pre-1800 era and the post-1800 era. The trends will be able to show if the homicide rates are on the change or are the same in the two predetermined eras. Studying the trends of homicide is vital since it provides a guideline on how the research will be conducted. The previous literature will provide a framework in which the research will get to be based on.
The study will adopt t-tests and regression analysis which will be use in determining the relationship between the young share of 20 to 34 and 20 to 29 and the young male share of 20 to 34 and 20 to 29. The relationship will be vital in evaluating some of the causes of the homicide rate based on the two demographics that is age distribution and gender. The relationship will also be linked with the pre-1800 data in showing the significance of the changes in homicide trends in comparison with the post-1800 data.
LITERATURE REVIEW
General demographic variables
There have been various attempts by economists in the application of econometrics and economic theory in understanding the determinant factors of crime. Ehrilch (1973) developed an explicit model aimed at understanding individuals' participation in activities of crime. Both agents had to decide the allotment of its time between work in criminal activities or legal activities and leisure. The formalization saw each agent view criminal activity as labor with a likelihood of being arrested. Moreover, the model foresaw crime rates as being related positively to income inequality and related to arrest prospects.
Generally, more economic factors than demographic factors are examined in crime studies. Factors Such as GDP growth and income inequality are more frequently considered in related topics. However, some economists and criminologists claimed that the factors of demography are vital in affecting individual propensity to participate in criminal activities. Hartung and Pessoa (2000) concluded that economic factors have more influence on property crime, for instance theft and robbery. The violent crimes are related to less financial motivation, but more psychological reasons. Thus demographic factors are necessary to be considered in. An interesting result found by Levitt (2001) is that the legalization of the abortion caused a reduction of both violent and property crimes. This finding was supported by Hartung and Pessoa (2000) that even for crimes with clear economic motivation, the demographic variables still appear to play a role after controlling for economic factors.
The traditional model explained little of the occurrence of violent crimes. Still, some potential demographic factors have been examined. Rasanen (1999) found out that the likelihood of a person born in Finland in 1966 to commit crime till the age of 30 years depends on the level of education of the other, being born to an adolescent mother or being born of a single mother. According to Dagg (1991), a boy born of an unwanted pregnancy has a 60 percent likelihood to commit a crime in his life.
Not surprisingly, using demographic variables alone as a determinant of the cumulative crime level is explored very little by crime literature. Brown and Males (2001) stated that immutable characteristics, such as age, offers rather weak explanation when viewed at the individual scale and age by itself does not predict future criminal behavior. However, immutable characteristics and changeable community contexts considered together, such as age and poverty, seem to better explain the prevalence of youth crime.
Age Distribution and Crime
As Brown and Males (2001) asserted, young age, among all demographic variables, is most recognized and most cited by authorities today. According to Par-melee (1918) , the most accepted and oldest widely criminology concept is that crime involvement diminishes with age. Fox and Piquero (2003), Steinberg (2007), and Reyna and Rivers (2008) all more or less agree on the findings that more male teenagers leads to more crimes. Furthermore, Hartung and Pessoa (2000) suggested that percentage of young people is the only demographic variable widely used in related crime studies.
However, the age-crime relation has experienced great disagreement about its strength and universality.
Mannheim (1965) claimed that the parameters of the age distribution are quite different with the decline in criminality. Moreover, criminality is regular to all distribution by age and crimes across localities or over time. According to the Traditional sociological explanation of crime propensity to traditional sociological explanations, the tendency of crime to decrease with age rests on the Hobbesian assumption that human behavior is not intrinsically compliant. There is a persistent setback on social disorder which faces many societies. Based on age distribution trends on criminality, there needs to be institutional structures which will motivate the youth as they assume adult responsibilities and roles. As a result, the society wll be able to survive.
Ryder (1965) observed that societies are faced persistently with barbaric invasion. Thus, every adult generation is countered with the task of civilizing these barbarians. The civilization can be attained through processes that are age-related of social control, social integration, and socialization which raise the crime costs and its benefits reduction. Hirschi and Gottfredson (1986) disagreed with social control of the age-crime relationship social control. He instead advocated that crime distribution by age is basically invariant across space and time, in spite of offense. He further argued that the presented relationship between age and crime explanations is flawed. Therefore, age distribution cannot be explained using sociological theories which are presently being employed by criminologists. It can be attributed to the causal variables which are utilized in these theories that are invented to differ across space and time. The sociological theories include strain, different association, and social control. Hirschi and Gottfredson (1986) asserted that the recognition of the crime causes at any age is sufficient to recognize the cause of crime at any age. The identification will be sufficient to spot them at other ages as well. Hirschi also claimed that the form or shape of allotment has continued to be relatively unchanged for the past 150 years. However, his reports do not present any statistical test as their age crime plots are so condensed making it intricate to detect whether any distribution by age differences does exist.
Moreover, the few amounts of studies examining the relationships between age and crime have had self-effacing success in the identification of causes which are social. Tittle (1988) found out that, "Social processes explain much of the age-crime." He concluded that, "Social processes explain the much age-crime relationship when sociological variables are considered simultaneously in an additive model." Sociological variable includes social integration, sanction fears, crime utility and moral commitment. Tittle (1988) data set was intended to overcome statistical flaws in the original investigation. Kercher (1987) found out that, "Tittle data provided strong evidence that age affects crime indirectly through the intervention of sociological variables such as criminal associates and moral commitment."
Gender and crime
According to Denno (1990), the most frequently documented predictor of violent behavior and delinquency is a child's gender. There have been various studies which have shown that the crime rates generally are constantly showing higher rates of offenses for males than for females. Wilson and Hernstein (1985) asserted that gender demands attentions in searching for crime origins since it has been consistent over time and culture. Criminologists have shown concern through the theories application of crime by males to females. However, they recommend for more research which are developmental on males in determining the theories if which could be actually applied on them. On the other hand, other criminologists have projected diverse theories for males. Theories of crime by males emphasize on educational, unemployment, peer influence and social disadvantages. Conversely, the theories of crime by males stress on personal fine-tuning that is accredited to the psychological and biological makeup of males.
Gender behavior can be attributed to the essentiality of social cognitive processing skills. The skills develop in a different way in males and females. As a result, developmental scarcities in social cognitive skills attainment provide a framework in the explanation of differences by gender in violence and crime. According to Ross and Fabiano (1985), females averagely comprise of only 6% of offender populations. The result can be attributed to the acquiring of cognitive skills which are social much earlier in life and subsequently have skills which are superior. Thus, early cognitive skills which are social result from fewer neuro-developmental disorders and a bigger emphasis on role and perception taking, empathy, social problem-solving skills and social reasoning in girls.
According to Moffitt (1993), "Males appear to have a greater likelihood than females of deficits in the frontal lobes of the brain. Moreover, the dysfunction interferes with the normal cognitive processing capabilities and has a higher probability of precipitating a higher incidence of learning antisocial behavior and relate difficulties." Moffitt expounds that, "Attention deficit hyperactivity disorder (ADHD) and antisocial or aggressive behavior are the problems most commonly connected with deficits or damages to the frontal lobes."
The Arizona Sibling Relationship Study by Rowe et al. (1995) examined whether differences by gender in delinquency mirrored the differences in gender risk factors in relation to factors of risk and delinquency. The sample used constituted of almost equal numbers of brother pairs, sister pairs, and mixed-sex sibling pair. The result showed that male delinquency correlates with female delinquency. However, their level of age was greater in males. Thus, the theoretical rationalizations of differences in gender in violence and crime can be considered to be a risk factor which is influential in both females and males though they are frequent in males. For instance, males are linked with delinquent peers more as uncovered by more models of antisocial and had more opportunities for involvement in the crime.
Moffitt et al. (2001), claim that, "Boys and girls differed on cognitive and neurological risk factors for antisocial behavior. The likelihood for boys to show neurocognitive deficits, poor impulse control, hyperactivity and under-controlled temperament, features were higher than those of girls. Thus, these risk factors account for most of the difference in gender found on antisocial behavior."
According to Mears et al. (1998), "Females were differentially affected by exposure to the same criminogenic conditions." They adopted a solitary explanatory structure for the two genders which they experimented the hypothesis that negative moral conduct estimates provide a cushion against the influence of delinquent peer associate. The phenomena can be attributed to females having a greater social cognitive skills compared to the males. By collecting data from the National Youth Survey (NYS) from 1626 participants, Mears et al. (1998) found out that, "Males were to a large extent more likely to have friends who are delinquents." As a result, the disparity exposure added to differences by gender in delinquency.
Theoretical Framework
The following framework was adopted in this study.
Homicide trend = f(demographics)
The dependent variable in the model will be homicide trend while the independent variables will be distribution by age and gender. The following hypothesis was developed to derive the relationship of the variables.
H0: the variables are all related
H1: the variables are not related
DATA DESCRIPTION
Homicide rates
The following study adopts homicide as an alternate for violent crime. Homicide is one of the most effectively identified and recorded by the police albeit a range of crime being hard to identify and record by the police (UNODC, 2011). It is vital to analyze the homicide trends and patterns to provide a starting point for a more broad research into other forms of crime which are violent. Walby et al. (2014) claimed that violent crime includes violence against other people and also sexual violence. However, according to ONS (2014), violent crime is categorized as violence against a person and not the threats and robbery. In this study, the researcher opted to follow the ONS categorization of violent crimes. The researcher measured interpersonal violence with the rates of homicide which were available for a larger number of countries from the late 19th century onwards. Moreover, there was a small selection of countries from countries I the 14th century onwards.
Post-1800
The sample for the post 1800 was observed on a ten-year interval for 194 countries. The ten year periods were from 1800 to 2010. The nation-time period was used as the unit of analysis. The nations under the study included African nations, European nations, Asian nations, Oceanic nations and American nations. The analysis took macro-level data of nations enough in a worldwide range. It aimed t analyzing the patterns of developing and developed countries. In the United Sates, most macro-level homicide models are referred here. The macro-level homicide models are formulated on the basis of social processes which characterize developed nations in the contemporary times. Developed democracies offer a framework for testing the models' generality. On the other hand, developing nations present relatively few measurement problems.
Figure 3.1: Global crime rate post-1800 trend
illustration not visible in this excerpt
Figure 3.1 shows that the crime rate has been on the rise since 1800 to 2010. The line graph proves this since its trend shows that it has been on the rise with frequent declines over the 220 year period. Crime rate was at its lowest in 1800 at 3.03 but over the years, it had reached at its highest in 200 where it was recorded at 8.79. The trend also shows that despite the control in some of the regions, the overall trend has been on the rise.
Figure 3.2: Overall homicide rate in post-1800 trend based on the region
illustration not visible in this excerpt
The figure 3.2 shows that Oceania recorded some of the highest homicide rates in the world in 1820 and 1830. However, America took over as the highest contributor to homicide rates from 1830 till 2010 with an exception in 1960 and 1970 where Africa was the major contributor to the crime rates. The figure also shows that most of the regions had gaps in their crime rates record. They include all the regions except Europe.
Pre-1800
The data for the rate of homicide for the pre-1800 period was limited. Thus, a proxy technique had to be utilized. Systematically evidence was collected on the kings' biographies while determining whether they were killed or not. Eisner (2014) developed homicide estimates based on trial data and other sourced which had been coiled by crime historian since the high medieval period onwards. Eisner also suggested the use of regicides, which is the killing of kings both in battle and murdered cases. According to Baten (2016), there is a relationship between homicide and regicide. As a result, we can use the killing of kings as a proxy for the overall homicide rates. The world regions which were included in this study were wee, ssa, lac, eeu, india and afr. However, it depended on the availability of data.
Demographic variables
The two main demographic variables used for the study was gender and age distribution. For gender, it entailed the hare of the young population and the share of the young male in the population. Data was collected for two groups which include 10 to 29 and 20 to 34. The two demographics were interlinked to yield the following:
- The share of the young on the age groups of 20-29 and 20-34
- The share of young males in the age group 20-29 and 20-34
Data for the pre-1950 period was collected and calculated based on the historical datasets. The sources of the data include census of death registration, church, immigration, prison, marriage, civil services, and military among others.
Figure 3.3: regicide and youth share trend 1400-1800 Abbildung in dieser Leseprobe nicht enthalten
The figure above shows the regicide and the young population trend between 1400 and 1800. It can be seen that the regicide rate has been moving in unison with the regicide rate in the males and the whole population in the 20 to 34 age group. Moreover, there is a slope pattern between the population and males share in the 20 to 34 with the regicide interval of 20 years from 1600 to 1700.
Figure 3.4: regicide rate trends in the pre-1800
illustration not visible in this excerpt
The figure above shows the trends of the regicide between 1400 and 1800. There were three trends which were grouped in periods of 10, 20 and 50 years. The 10 year and 20 year regicide trends were more volatile compared o the 50 year trend. However, the trend can be seen to follow a similar pattern as they rise and fall with great similarity.
Figure 3.4: demographic trends in the post-1800
illustration not visible in this excerpt
Figure 3.5 shows that the demographic trends which account for the homicide rate has been relatively constant. The data were pooled based on five year age groups. The movement of the data is quite minimal over the years. Moreover, the data seem to follow a similar trend where the homicide rate increased simultaneously for the four categories and fell simultaneously for the four categories.
illustration not visible in this excerpt
Figure 3.5: Homicide and youth share trend 1800 -2010
The figure above shows the trends of the youth share and the homicide rate from 1800 to 2010. Though the age clusters move parallel, the homicide rate has been on the rise. The involvement of the males in both age groups seem to decline as the years go by but the homicide rate is on a constant rise..
illustration not visible in this excerpt
Figure 3.6: Homicide and youth share trend 1950-2010
The figure above takes a closer look from 1950 to 2010. The figure shows that the homicide trend has been on a rise from 1980 till 200 with a slight drop to 2010. However, age and gender clusters seem to be relatively constant. The clusters are on the rise from 1970 to 2010 as the four clusters are seen to move in a parallel way. Thus, the increase in homicide rate can be attributed to the trends of the age and gender clusters.
RESEARCH DESIGN AND METHODOLOGY
The following section aims at explaining the methodology that was utilized in the research execution so as to ensure its success. The research's objective was finding the demographic factors which influence homicide rates globally. Some of the chosen variables for consideration that were chosen were the distribution by age and crime and gender and crime. The following study employed a quantitative approach. The advantage of using a quantitative approach is that it provides for phenomena in human behavior and human variables which are explored objectively. Anders (2012) stated that the emphasis of quantitative analysis is that quantitative methods provide for the facilitation of a well-conceptualized research process. Thus, the research methodology presents the researcher with the capacity to gather valid data and analyze them appropriately so as to attain the desired decision.
A descriptive survey research design was used for the research since it has the ability to describe the trends of the variables and thereby give a picture of the factors which affect their trends. The employment of the design is justifiable since because of its exceptionality of making it possible to gather secondary information. In addition, the research design is useful in helping the research in answering his or her research question.
The research design will also incorporate a correlation research design. According to Kumar (2010), correlation research design is helpful in benefiting researchers in the establishment of relationships between two or more phenomena. Since the study aimed at establishing the forces behind income inequality, the chosen design enabled for the determination of the expected associations between the variables and therefore came up with the most appropriate predictions.
Descriptive research designs will be used in order to describe, explain, and validate the data. Moreover, it will serve to organize the findings so as to fit them with explanations. The explanations will then be tested and validated.
Using Stata/SE 9.0 for Windows, paired sample t-tests were chosen in assessing the means of the factors. Calculation of the eta squared statistic was done manually using a calculator.
The study will utilize an econometric model to spell out the statistical relationship which the research seems to grip between the various quantities of which are economical that are pertinent to the observable economic fact in the research. The economic model that was chosen for this research will be a linear regression. The regression best suits the study since it conceptualizes to join methodically two or more variables to have a linear connection.
A population is the entire set of individuals or components which have a common observable characteristic of a particular nature that is distinct from other populations. The target population of the study is the homicide trends in the world.
The nature of the data to be gathered for this study is secondary in nature. Secondary data are acquired from secondary sources. The secondary sources are classified into two groups. That is, from published sources or unpublished sources. Examples of published sources include international, government, semi-government, national, private corporate bodies, expert committees, trade associations, and research reports and commission reports. Unpublished sources, on the other hand, include the government, agencies, private firms, and offices. However, unlike published sources, unpublished sources face various limitations. They do not have a proper procedure for data collection, they may be influenced by the investigator's prejudice, may lack standards of accuracy and may not cover the full period of the investigation. As a result, the researcher chose to use a published source.
The research instruments used in the collection of the data must have undergone checks for reliability. The researcher also adapted the use descriptive statistics with the facilitation of bar graphs and line graphs. The chosen tools were appropriate in analyzing data which were collected for a simple interpretation. A descriptive analysis was made use of to summarize the given data sets of the samples.
The measures of variability were measured through the use of the variance, standard deviations, the minimum and the maximum, the kurtosis and the skewness. Measures of variability aid in analyzing how a distribution of a given data set is spread out. The Pearson's Product Moment Coefficient of Correlation was used in investigating the relationship between the variables. The research employed at least five percent level of probability as the foundation for rejecting a null hypothesis.
The figures to be utilized in the research was easily accessible from the internet. Nevertheless, there was a need for indirect consent for further use and analysis. Hence, the study has to recognize the tenure of the original data. The data were assessed for decisive factors such as the methodology of data collection, accuracy, a period of data, the content of the data and the purpose of collection. The researcher has the mandate of guaranteeing that further scrutiny of the data is suitable.
DATA ANALYSIS
Descriptive analysis
A descriptive analysis was conducted on the five variables, hr, p2029, m2029, p2034 and m2034. The homicide rate had a mean of 6.39 with a standard deviation of 9.07. Moreover, we can see that the homicide rate had a minimum of 0 but with a maximum of 87.8667. On the other hand, it was found out that variable p2029 had a mean of 0.169 with a standard deviation of 0.0395. In addition, it had a minimum of 0.0166 and a maximum o 0.674. The variable m2029 had a mean of 0.86 with a standard deviation of 0.239. The variable has a minimum of 0.009 and a maximum of 0.441. The last variable, m2034, had a mean of 0.122 with a standard deviation of 0.042. Moreover, it has a minimum of 0.003 and a maximum of 0.522.
Correlation analysis
To test the linear relationship between explanatory variables, a correlation matrix is usually adopted. A correlation matrix is a vital indicator in determining the variable strengths of the model. A correlation statistic which will be greater than 0.8 will reflect a high correlation among variables. As a result, the variable will have to be dropped. The study was guided by the research question of the impact of distribution by age and crime and gender and crime on the homicide rate. The correlation of the variables is recorded in table 1 in appendix 3.the table shows that the correlations of some of the variables are strongly correlated. However, the researcher went on to regress the variables by choosing not to regress them all at once. Moreover, the variables depicted positive correlation.
Additional indices and indicators were looked into. Though they offered potential explanations, they were not included in the regression. The indicators and indices include the gender equality index, the gender equality of numeracy, the urban ratio and the income inequality. Table 2 in appendix 1 shows their descriptive analysis. From the correlation analysis of these index and indicators, it was seen that most of them were negatively correlated with the youth share as seen in table 3 in the appendix 1. However, urban ratio and m2034 had a positive relation. Gender equality seemed to have a positive correlation with the demo expect for p2029. Moreover, their relationship with the homicide rate was pretty predictable. All the age groups seemed to have little correlation with the homicide rate. Thus, it was safe to run the regression in a later stage.
t-test statistics
T-test statistics were determined in order to determine the relationship between the independent variables (p2029, m2029, p2034 and m2034) and the dependent variable (hr). The result of the paired sample t-test between the young population share in the 20-29 year range and the Homicide rate can be derived. The paired sample t-test was aimed at evaluating the influence of the population in 20 to 29 years on the homicide rate. There was a statistically significant decrease in the homicide rate statistics from time 1 (mean = 6.53, standard deviation = 9.3) to time 2 (mean = 0.166, standard deviation = 0.03), t(1136) = -23.0346 p < 0.1). The mean decrease in homicide rate statistics scores was 6.36 with a 95% interval of confidence ranging from 6.9 to 5.8. The eta squared statistics (0.32) pointed out that there was a small effect size.
A second paired t-test was run. The paired sample t-test was aimed at evaluating the influence of the young male share population in the 20 to 29 years age group on the homicide rate. There was a statistically significant decrease in the homicide rate statistics from time 1 (mean = 6.53, standard deviation = 9.3) to time 2 (mean = 0.08, standard deviation = 0.0192), t(1136) = -23.256 p < 0.1). The mean decrease in homicide rate statistics scores was 6.49 with a 95% interval of confidence ranging from 7.03 to 5.9. The eta squared statistics (0.32) pointed out that there was a small effect size.
The third paired t-test was run, aimed at evaluating the influence of the population in the 20 to 34 years age group on the homicide rate. There was a statistically significant decrease in the homicide rate statistics from time 1 (mean = 6.53, standard deviation = 9.3) to time 2 (mean = 0.24, standard deviation = 0.057), t(1136) = -23.1783 p < 0.1). The mean decrease in homicide rate statistics scores was 6.29 with a 95% interval of confidence ranging from 6.82 to 5.8. The eta squared statistics (0.32) pointed out that there was a small effect size.
The final paired t-test was run with an aim of evaluating the influence of the male share population in the 20 to 34 years age group on the homicide rate. There was a statistically significant decrease in the homicide rate statistics from time 1 (mean = 6.53, standard deviation = 9.3) to time 2 (mean = 0.119 standard deviation = 0.036), t(1136) = -23.6758 p < 0.1). The mean decrease in homicide rate statistics scores was 6.46 with a 95% interval of confidence ranging from 6.998 to 5.927. The eta squared statistics (0.32) pointed out that there was a small effect size.
Regression analysis
Four regression models were carried out on the datasets. The regression models utilize the Ordinary Least Squares, fixed effect (LSDV) and the random effect. The regression analysis used the homicide rates as the dependent variable while the independent variables were p2029, p2034, m2029 and m2034. Thus, homicide rate (hr) on the youth share d young male share of the two age groups were regressed. The datasets used were in the time period from 1800 to 2010. The first regression aimed at finding the relationship between the homicide trends from the year 1800 to the year 2010 with the youth share of 20 to 29 years. The second regression model was aimed at finding the relationship between the homicide trends from the year 1800 to the year 2010 with the youth share of 20 to 34 years. The third regression model was aimed at finding he relationship between the homicide rate and the young male share of the 20 to 29 age group. The final regression found the relationship between homicide rate an d the young male share in the 20 to 34 age group.
Regression of homicide rate on the youth share 20 to 29
It was found out that the coefficient for age group 20 to 29 is substantial and statistically significant in the first 2 models (OLS and FE). The OLS model show that a1 percent increase of young people aged 20 to 29 will lead to 32 homicide count per hundred thousand people. That is quite significant considering the average global homicide rate is only around 6. However, we can also imagine there are great within country and region effect, so after we control for that, the coefficient is half as before. Still, that's a substantial impact. After controlling for country fixed, the coefficient of the youth share decreases. That means there are less within-country variation explained over time. However, when we control for time fixed effect as well, the coefficient again drops further to around 10. That is, 1 percentage increase of young people aged 20 to 29 will lead to increase of 10 homicide count per hundred thousand people.
Regression of homicide rate on the youth share in the 20 to 29 age group
For age group 20 to 34, it can be seen that things are on the opposite. First of all, a 1 percent increase of the youth share in the 20 to 29 age group lead to an increase of 13 homicide count per hundred thousand people. After controlling for country fixed, the influence that is the coefficient of the youth share actually increases to 21. That means there are more within-country variation explained over time. But when we control for time fixed effect as well, the coefficient again drops to around 13. That is, 1 percentage increase of young people aged 20 to 34 will lead to increase of 13 homicide count per hundred thousand people. That is a half of the previous age group. The random effect also influences the homicide rate in the 0 to 29 age group in that a 1 percent increase o the young people aged 20 to 34 will lead to an increase of 19 homicide count per hundred thousand people.
Regression of the homicide rate on the young male share of the 20 to29 age group
For age group 20 to 29, it can be seen that things are on the opposite. From the ordinary last squares, a 1 percent increase of the young male share in the 20 to 29 age group lead to an increase of 35 homicide count per hundred thousand people. After controlling for country fixed, the influence young male share decreases to 27. That means there are less within-country variation explained over time. That is, 1 percentage increase of young males aged 20 to 29 will lead to increase of 27 homicide count per hundred thousand people.
Regression of the homicide rate on the young male share in the 20 to 34 age group
The coefficient of young male share of age group 20 to 34 seems only plausible in the model with country fixed effect. That is, a 1 percent increase of young male population in the 20 to 34 age group will lead to around 18 increase of homicide count per hundred thousand people.
DISCUSSION
Some of the data were positively correlated though with a correlation of more than 0.8, which is deemed as too high making the data inappropriate for regression them. However, the researcher chose to regress the data separately.
The t-test statistics showed that the variables showed a decrease in the means. Moreover, the data were statistically significant as they decreased the rate of homicide. All the test statistics had an eta squared of 0.32 thus showing that the change had a small size effect.
Four logistic regression analysis were conducted. The first regression aimed at finding the relationship between the homicide trends from the year 1800 to the year 2010 with the youth share of 20 to 29 years. The second regression model was aimed at finding the relationship between the homicide trends from the year 1800 to the year 2010 with the youth share of 20 to 34 years. The third regression model was aimed at finding he relationship between the homicide rate and the young male share of the 20 to 29 age group. The final regression found the relationship between homicide rate an d the young male share in the 20 to 34 age group.
In appendix 3, table 5 shows the coefficients of the regression models. The coefficient will give a hand in combining the relationship among the four models. It can be seen that the age group of 20 to 29 shows higher coefficient of the Ordinary least squares compared to the fixed effects while the age group of 20 to 34 shows the opposite. That is, the coefficient is 15 which is less than the ordinary least squares coefficient of 32. On the other hand, the young age of the age group of 20 to 29 has more than twice the coefficient of the homicide rates compared to the age group of 20 to 34 in the ordinary least squares coefficients. However, after controlling for the fixed effect, the population of the 20 to 34 age groups has a greater impact. Moreover, it can be seen that the young male share of the age group of 20 to 29 years has the highest coefficient in all the models. Even after controlling for the country's fixed effect, it still is the highest among all the other variables which is pretty impressive. However, the coefficients are not proportionally scaled. It's not like, because the share of male is roughly half of the share of both sex, the coefficient as a result will double. We can see it works quite opposite for age group of 20 to 34 years. In addition, the age group of 20 to 29 years has a coefficient which s no proportionally scaled. Despite the fact that the total share of young male together with the female has a way less impact on the homicide rate after controlling for the country fixed effect, the share of young male will seem to hold the influence even after considering variations within the countries. As a result, it can be used in a great degree to prove the hypothesis that young male share does play a great role in affecting homicide rates globally. On the other hand, the effect of male share of another age group is not as significant as his group at all. It is only statistically significant in the fixed effect model without the time fixed effect which only explains the within country variation over time. However, it still has less impact that the younger age group.
Conclusion
From the results it can be concluded that from a longer term of time and wider range of space, the youth share still shows a positive impact on homicide rates. Thus, demography plays a role for murder rates. Since he share of the young male between the 20 to 29 age group as the greatest impact on the homicide rate, gender roles does not have any influence here. The young individuals of the early 20s show that there is more tendency to contribute in criminal activities. Thus, they are more prone to crime. The findings of the study support the previous researches and literature from a worldwide perspective of view. Thus, the main results remain clear and positive.
Strength
The study used data from reliable data sources. The data, therefore, was accurate. Moreover, reliability tests were done on the datasets which proved that the data were reliable and therefore adopted for the study. The use of secondary data made it possible to save time during data collection thereby making the study feasible with both longitudinal and international comparative studies. Moreover, the findings of the study were also supported by previous researchers and literature. The previous literature made it possible to generate new insights which created an opportunity for the study to discover new unexpected discoveries.
Weaknesses
Moreover, there are some other obvious drawbacks. For example, some of the data used were from unpublished, secondary data sources, without control, may have been affected by some sort of biases or inaccuracies during collection and calculation. In addition, the researcher did not have control over the data due to some quality issues which may be associated with secondary data. The results of regressions on data pre-19th are not as significant as they show in the graph. As a result, they were not included here. Some transformations of variables, were taken on the tests to show the square root of homicide rate. They were the only plausible way to transform them. As a result, the results were almost all statistically significant, but it was hard to interpret them. Due to the lack of time, some tests for data were not shown here. Data of age groups might have had serial correlation. Moreover, there may be many other issues which could not be covered by the scope of this paper.
Future research
The following study was carried out with an expectation of opening up the opportunity for future researchers to have a framework to carry out their own studies. Previous literatures have shown that the topic needs further research due to the dynamic changes in homicide rates and the subsequent factors. On the other hand, gender and distribution by age are very complex topics. There are very many issued which cannot be covered by the scope of this paper. Thus future research will be encouraged to cover some of the shortcomings which may not be identified by this research.
REFERENCES
Anders, l, (2012). Designing and doing research survey. Thousand Oaks, oh: sage.
Bartol, A. M., & Bartol, C. R. (2014). Criminal behavior: A psychological approach. Boston: Pearson, c2014 . xxii, 664 pages: illustrations; 24 cm.
Denno, D. W. (1990). Sociological and human development explanations of crime: Conflict or consequence? Criminology.
Donohue, J. J., & Levitt, S. (2001). The impact of Legalized Abortion on Crime. Quarterly Journal of Economics.
Ehrlich, I. (1973). Participation in illegitimate activities: A theoretical and empirical investigation. Journal of Political Economy.
Eisner, M. (2014). From swords to words: Does macro-level change in self-control predict long-term variation in levels of homicide?. Crime an justice.
George, D. and Maleery, P. (2003). SPSS for Windows step by step: A simple guide and reference. Boston: Allyn and Bacon.
Hirschi, T. & Gottfredson, M. (1986). Age and the Explanation of Crime. American Journal of Sociology.
Kercher, K. (1987). Explaining the relationship between age and crime: The Biological vs. Sociological Model. Presented at the annual meeting of the American Society of Criminology, Montreal.
Kumar, R. (2011). Research Methodology: A Step-by-Step Guide for Beginners. New Delhi: SAGE Publications India Pvt Ltd.
Le Fevre, M., Matheny, J., & Kolt, G. S., (2003). Eustress, distress, and interpretation in the occupational stress. Journal of Managerial psychology.
Mannheim, H. (1965). Comparative Criminology. Boston: Houghton Miffin.
Mears, D. P., Ploeger, M., & Warr, M. (1998). Explaining the gender gap in delinquency: Peer influence and moral evaluations of behavior. Journal of Research in Crime and Deliquency.
Mofitt, T. E. (1993). Adolescence-limited and life-course persistent antisocial behavior: A developmental taxonomy. Psychological Review.
Moffit, T. E., Caspi, A.., Rutter, M., & Silva, P. A. 92001). Sex differences in antisocial behavior. Cambridge, England: Cambridge University Press.
Parmelee, M. (1918). Criminology. New York: Macmillan.
Rasanen, P. (1999). Maternal Smoking during pregnancy and risk of criminal behavior among adult mal offspring in the Northern Finland 1996 Birth Cohort. American Journal of Psychiatry.
Ross, R. R., & Fabiano, E. A. (1985). Delinquency prevention through cognitive training. New Education.
Rowe , D. C., Vazsonyi, A. T., & Flannery, D. J. (1995). Sex differences in crime: Do means and within-sex variation have similar causes? Journal of Research in Crime and Delinquency.
Ryder, N. (1965). The Cohort as a Concept in the Study of Social Change. American Sociological Review.
Tittle, C. R. (1988). Two empirical regularities (maybe) in search of an explanation: Commentary on the age/crime debate. Criminology, 26(1).
UNODC, (2011). Global Study on Homicide. Unite Nations Office on Drugs and Crime.
Walby, S., Towers, J., & Francis, B. (2014). Mainstreaming domestic and gender-based violence into sociology and the criminology of violence. The sociological review..
Wilson. J. Q., & Hermstein, R. J. (1985). Crime and human nature. New York: Simon and Schuster.
APPENDIX
Appendix 1:
Table 1: Descriptive analysis
illustration not visible in this excerpt
Table 2: Description of variablesn
illustration not visible in this excerpt
Table 3: Correlation of variables
illustration not visible in this excerpt
Appendix 2: t-test statistics
Table 1: p2029 and the homicide rate
illustration not visible in this excerpt
Table 2: m2029 and homicide rate
illustration not visible in this excerpt
Table 3: p2034 and homicide rate
illustration not visible in this excerpt
Table 4: m2034 and homicide rate
illustration not visible in this excerpt
Appendix 3: Regression analysis
Table 1: Regression of the 20 to 29 age groups and the homicide rate
illustration not visible in this excerpt
Table 2: Regression of the 20 to 34 age groups and the homicide rate
illustration not visible in this excerpt
Table 3: Regression of young male share population of 20 to 39 age distribution and homicide rate
illustration not visible in this excerpt
Table 4: Regression of young male share of 20 to 34 age distribution and homicide rate
illustration not visible in this excerpt
Table 5: Coefficients of regression models
illustration not visible in this excerpt
-
-
-
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X. -
-
Téléchargez vos propres textes! Gagnez de l'argent et un iPhone X.