Key Factors Affecting Crime Rates in California Cities

Title:	Key Factors Effecting Crime Rates in California Cities
Overview:	There are several competing schools of thought regarding the causes of crime, they are the biological, social, and environmental.  The biological school views crime as the result of chemical or structural abnormalities contained within the brain, while the social and environmental schools, instead, view crime as a function of individual’s interactions with other people and the world around them.  This analysis is designed to identify the social and environmental factors that have an effect on crime rates in California cities and to determine how strong of a correlation there is between them.  Crime rates for selected cities will be taken from the Uniform Crime Report (UCR) issued by the FBI, which contains crime statistics for all cities in the United States.  Possible independent variables will be taken from the census, voter participation data, and law enforcement employment statistics, among others.
Research Question:	Is there a strong relation between the crime rates for each city and that city's social and environmental context?
Hypothesis:	Crime rates experienced by cities in California are largely determined by social and environmental factors such as voter participation, poverty levels, single family households, racial make up, income, educational attainment, number of police officers, number of males 18 to 24, and level of segregation, etc. We assume that factors such as income, educational attainment, employment, and voter participation will be negatively correlated with crime rates, that is as levels of these variables increase crime rates will decrease.  Conversely, we assume that factors such as poverty, single family households, minority population, males 18 to 24, and level of segregation will be positively correlated with crime rates, that is when one of these variables increases so to will crime rates in these cities.
Unit of Analysis:	City
Data Required and Sources:	1. 2000 Crime Statistics from the Uniform Crime Report (UCR) (http://www.fbi.gov/ucr/ucr.htm ) 2. 2000 Crime Statistics  for State of California from the Office of the Attorney General, State of California, Department of Justice (http://caag.state.ca.us/cjsc/datatabs.htm ) 3. Census data in selected cities in California from American Factfinder (http://factfinder.census.gov/servlet/BasicFactsServlet ) 4.  Law enforcement data for California Cities in 2000 (http://www.post.ca.gov/employdata/policedata.pdf )
Methodology:	1. Obtain data on variables included above from the sources listed. 2. Enter data in Excel spreadsheet.  Identify a missing cases and inconsistencies in the data 3. Develop an initial statistical summary for each variable, this will consist of the mean, median, and standard deviation for each variable. 4.  Calculate the correlation of crime rate to each other variable and depict results on charts to displace the results. 5. Use regression analysis to locate the key variables, which are associated with crime rates.  Variables with high statistical significance will be highlighted. 6. Write up a data analysis, which will address the initial research question and hypotheses.  Discuss the validity of the results.  Note where the results confirm the initial hypotheses, conflict with the initial hypotheses, and are inconclusive.  Propose improvements to research results and suggest direction for further study.
Anticipated Results:	Crime is a complex topic and its causes are unclear and closely related.  For this reason we believe that while our results will likely show a connection between our variables and crime rates in the included cities, these results will not be definitive or conclusive.  It is impossible to include all the variables effecting crime so the significance of the results will be limited to the variables included in the study.
Selected Bibliography:	1. Roleff, Tamara. Crime and Criminals. Greenhaven Press. 2000. 2. Ward, Carolyn. Community Education and Crime Prevention. Bergin & Garvey. 1998. 3. Weisburd, David. Reorienting Crime Prevention Research and Policy. National Institute of Justice. 1997. 4. Forst, Bryan. The Socio-Economics of Crime and Justice. M.E. Sharpe, 1993.

Regression Analysis

• Dependent Variable: Reported Crimes

• Independent Variables: percentage of African-American and Hispanic populations, percentage of unemployment, percentage of population below poverty level, percentage of single-parent headed families, and percentage of renters

Model Summary

Model	R	R Square	Adjusted R Square	Std. Error of the Estimate
1
0.388	0.151	0.037	9298.106

a Predictors: (Constant), Percent of Other Families, Percentage

of Housing Units Rental Occupied, Percent Population 16 years

and over; In labor force; Civilian; Unemployed, Percent of

Population for whom poverty status is determined

ANOVA

Model		Sum of Squares	df	Mean Square	F	Sig.
1
Regression	4.6E+08	4	1.2E+08	1.331	0.281
Residual	2.6E+09	30	8.6E+07
Total	3.1E+09	34

a Predictors: (Constant), Percent of Other Families, Percentage of Housing Units Rental

Occupied, Percent Population 16 years and over; In labor force; Civilian; Unemployed,

Percent of Population for whom poverty status is determined

b Dependent Variable: FBI Crime Index

Coefficients

		Unstandardized Coefficients		Standardized Coefficients	t	Sig.
Model
	B	Std. Error	Beta
1
(Constant)	-4024.9	10629.428		-0.379	0.708
Percent of Population for whom poverty status is determined	18292.6	47027.945	0.116	0.389	0.7
Percent Population 16 years and over; In labor force; Civilian; Unemployed	-72026	99484.941	-0.21	-0.724	0.475
Percentage of Housing Units Rental Occupied	-4311.3	19106.499	-0.04	-0.226	0.823
Percent of Other Families	65247.8	38849.486	0.439	1.68	0.103

a Dependent Variable: FBI Crime Index

MAP OF CALIFORNIA SHOWING THE LOCATION OF EACH CITY INCLUDED IN THE STUDY

II. Low Crime/Small Population

IV. High Crime/Small Population

III. High Crime/Large Population

POPULATION VS. TOTAL REPORTED CRIME

Comparison Between Groups

Table of Findings

	I. Low Crime, Large Population	II. High Crime, Large Population	III. Low Crime, Small Population	IV. High Crime, Small Population
Average Reported Crime	-	47,579	6,968	31,438
Average Population	-	1,647,474	167,199	411,385
Percent of population that is White	-	51%	60%	43%
Percent of population that is African American	-	7%	7%	20%
Percent of population that is Hispanic or Latino	-	29%	31%	28%
Percentage of Housing Units Rental Occupied	-	54%	50%	53%
Housing units: Median year structure built	-	1961	1968	1964
Percent One-Parent family homes	-	28%	28%	36%
Percent Unemployed population	-	6%	7%	9%
Percent of Population earning under the poverty line	-	14%	15%	21%
Households: Median household income in 1999	-	$51,971	$46,716	$36,447

Preliminary Findings

· While we analyzed the stereotypical variables of crime: percentage of African-American and Hispanic populations, percentage of unemployment, percentage of population below poverty level, percentage of single-parent headed families, and percentage of renters, the model does not render a significant a high R² value nor do any of the variables produce significant t-scores.

· Crime and total population are very highly correlated. This analysis suggests that regardless of dependent variables such as poverty, single parent homes, or percent of renter occupied housing, as total population increases so to does the total reported crime.

· From our analysis, it can be concluded that crime is influenced by multiple variables and can not be solely explained by the variables used in our analysis. In terms of our original research question then there is not a strong relation between crime rates and for each city and that city’s social and environmental context. Our study indicates generally that as the total population of a city increases so to will the number of crimes reported.

· The city of Long Beach presents an interesting case in which it is very close to entering the Low Crime/Large Population class. We will analyze this city to determine whether it possesses any unique characteristics that set it apart from the rest of the cities included in this study.

Limitations

· Cities are very diverse entities; they contain much variation within their boundaries. Census data taken at the city level represents the aggregate and can mask the fact that different sections of a city contain either very high or very low crime numbers of total reported crime. A different level of analysis, perhaps the Census tract may have yielded much more significant and interesting results.

· Second, an analysis done at the Census tract level would have provided a relatively standard population level, roughly 4000 people. This unit of analysis would have controlled for population differences and made the variables explored in out study more appropriate for regression.