15 Slides PPT Due within 8 hrs from now With SPSS
Need_Help_Now96
StudentSample2.pptxPredictors of Police-Citizen Interaction
Logistic Regression Meets a Criminological Theory Trifecta
Purpose of the Study
Can data collected in the General Social Survey predict whether U.S. citizens with certain racial or socioeconomic background will be formally questioned by law enforcement? Using Logistic Regression and three theoretical backdrops, this study will attempt to find the best predictor of formal police-citizen encounters.
A Review of the Literature: Urban Inequality & Formal Social Control
“The social transformation of the inner city in recent decades has resulted in an increased concentration of the most disadvantages segments of the urban black population, especially poor, female-headed families” (Sampson & Wilson, 1990/2011, p. 107).
Law enforcement presence and practices in these communities is known as formal social control (Sampson, Raudenbush & Earls, 1997/2011).
We are looking at Formal social control as it relates to collective efficacy/social disorganized neighborhoods. NOT G/H’s belief in society’s use of social control, or the sociological theories behind social control…we are just using a restrictive definition.
3
A Review of the Literature: Labeling Theory
How and Why do racially-based policing practices occur?
“The agents of control, who function on behalf of the powerful in society, impose the labels on the less powerful” (Akers & Sellers, 2013, p. 137).
Racism and inequality are present in the criminal justice system in the over-representation of African-Americans in the criminal justice system who are often labeled and stigmatized by society (Alexander, 2012).
Research has demonstrated that African-Americans are more likely to be watched, approached, and questioned by law enforcement (Brown, 2005).
A Review of the Literature: Inequality and Power-Themes from a Conflict Perspective
In every industrialized society, divisions occur between racial and ethnic groups (Loury, 2000).
The mere idea of crime and deviance is a form of social control meant to protect the interests of the rich. To do this, the powerful target the poor and ethnic minorities (Bryant, 2000).
There is concensus among researchers that the poor and racial groups are marginalized in communities marked by social disorganization.
5
Theory to Practice
Studies have identified AND examined how to combat racially-based policing practices that arise as a result of law enforcement practices (Fridell, L., Lunney, R., Diamond, D., & Kubu, B., 2001)
What drives racially based police practices…that is what we will test
6
Hypotheses
Will a person’s race and/or socioeconomic status predict the likelihood that a citizen is questioned as a murder suspect?
Based on the research, we would predict that law enforcement officials are more likely to formally question a person who is non-white and/or has a low socioeconomic status…
Conceptualization
Based on the previously discussed theories and literature, we can conceptualize “being questioned by police for murder” as a form of formal social control and an indicator of police presence in a neighborhood.
Next, if an urban area is said to have poor, single family ethnic groups, then race and socioeconomic variables should be used to indicate the likelihood (i.e. predict) of police-citizen interaction (as evidenced by police questioning).
Data Source: The GSS
Since 1972, the General Social Survey has attempted to “take the pulse of America”.
The goals of the GSS include:
a) Assemble scientific research on the structure and development of American society.
b) To distribute up-to-date, important, high-quality data to social scientists, students, policy makers, and others (GSS Website, 2014).
Operationalizing: GSS Variables Used in Analysis
Dependent variable: “Polmurdr” Have been questioned by police for murder. This is a dichotomous nominal variable (D/C).
Independent variables:
1. “Race” respondents race, nominal (D/C).
2. “Rincome06” respondent’s total family income, ordinal with 25 categories.
3. “Sei” respondent’s socioeconomic status, a scale-level variable.
D/C=Dummy coded
10
GSS Variables Used in Analysis
Control Variables used to rule out spurious explanations.
1. Respondent’s Sex, nominal (D/C)
2. Respondent’s Age, scale
3. Number of People in Respondent’s Household, scale and transformed due to high skew.
4. Respondent’s highest degree, ordinal with 5 categories.
Statistical Analysis
Due to the fact that we have a categorical dichotomous dependent variable, we can use logistic regression to predict an outcome (the DV) using a set of predictor variables (the IVs).
Results: Univariate Descriptive Statistics
Using the Mode, we can tell what response was selected the most by the respondents…for example, R’s Highest degree, one was selected the most, that represents a high school education. The average respondent age was 25. This chart gives a strong over view of the sample characteristics, but keep in mind, several variables were dummy coded and one was logged due to extreme skew. Here are three graphic visuals of my one of my control and IV, and my DV.
13
Results: Bivariate Correlations
This analysis tells us two things: There are significant correlations (especially with being a white respondent and the level of a respondent’s degree which will be important in the Logistic Regression) and also, that Multicolinarity will not be an issue. Every correlation is below .8 with the largest being .622.
14
Results: Multivariate Logistic Regression
Looking at the Onnibus Test of Model Coefficients is important because it tells us that this model is a good predictor of the variables I used. It is a significant predictor across the variables. Next, looking at the Nagelkerke R Squared, we can see that there is a correlation, but at .145, it is a moderate correlation. This essentially suggests that approximately 14% of the variance in my DV can be accounted for by the IVs that I used in this model.
15
Results: Multivariate Logistic Regression
Beta is weak and negitative, significance…two, Odds Ratios protective vs. risk
16
Discussion
We identified two predictors for the likelihood of being questioned for murder: being white and the level of degree (which was high school for most respondents and also a control variable, not an IV).
The data showed us an inverse, negative relationship between being white and degree.
Odds Ratio: there is a 63.7% DECREASE in the risk of being questioned by police for murder if you are a white respondent, and there is a 34.5% DECREASE in the risk of being questioned by the police for murder if a respondent has a higher degree. These results are compatible with the literature.
Land in the “no category” and be in the higher degree category
17
Conclusions Based on This Model
We can reject the null hypothesis because we found a relationship beyond chance between race and being formally questioned by police for murder. However, there was no correlation with low-socioeconomic status so the research hypothesis is not entirely valid as it currently stands.
GPD Homicide example
18
Study Limitations
Data collection. I cannot be sure what type of neighborhoods are being represented in this particular sample (representative sample?).
Variable Selection. Whether it was a limited data set, or poor conceptualization and analysis, other variables would be stronger predictors than race and socioeconomic status (perhaps prior arrests? Position in a social network?)
The two items above combined with statistical analysis errors ultimately lowers the validity and reliability of this study’s results.
GPD Homicide example
19
Implications and Future Research
Change policing practices and evaluate the outcomes! (Police-citizen encounters may continue to improve and decrease law enforcement bias when community oriented policing models are used in communities (Brown, 2005)).
Increase informal social control in neighborhoods.
Lower the socioeconomic gap and stigmas between the wealthy and the poor (macro-level changes).
References
Akers, R. L., & Sellers, C. S. (2013). Criminological theories: Introduction, evaluation, and
application (6th ed.). New York: Oxford university Press.
Alexander, M. (2012). The new Jim Crow: Mass incarceration in the age of
colorblindness. New York, NY: New Press.
Brown, R. A. (2005). Black, white, and unequal: Examining situational determinants of
arrest decisions from police–suspect encounters. Criminal Justice Studies, 18(1), 51-68. doi: 10.1080/14786010500071121
Bryant, L. (2000). Marxism and Crime. Retrieved May 2, 2014, from
http:// www.historylearningsite.co.uk/marxism_crime.htm
Fridell, L., Lunney, R., Diamond, D., & Kubu, B. (2001). Racially biased policing: A principled
response [Abstract]. Police Executive Research Forum. Retrieved from https :// www.ncjrs.gov/App/abstractdb/AbstractDBDetails.aspx?id=190927
References Continued
GSS General Social Survey. Retrieved April 30, 2014, from http://www3.norc.org/gss+website/
Loury, G. C. (2000). Social exclusion and ethnic groups: The challenge to
economics. The International Bank for Reconstruction an D Development / THE WORLD BANK, 225-252. Retrieved from http :// www.bu.edu/irsd/files/socialethnic.pdf
Sampson, R.J., & Wilson, W.J. (1990/2011). A Theory of Race, Crime, and Urban Inequality. In F.
T. Cullen & R. Agnew (Eds.), Criminological theory: Past to present: Essential readings (4th ed., pp. 105-111). New York: Oxford University Press.
Sampson, R.J., Raudenbush, S.W., & Earls, F. Collective Efficacy and Crime. In F. T. Cullen & R.
Agnew (Eds.), Criminological theory: Past to present: Essential readings (4th ed., pp. 112-117). New York: Oxford University Press.
