Stop and Search Data in the Lynfield Local Police Force - Statistics Project Example

 Contents Introduction 1 Analysis 2 List of Questions that were asked during the survey 2 Hypotheses 3 Descriptive Statistics 3 Hypothesis 1 5 Hypothesis 2 5 Hypothesis 3 6 Bivariate Analysis 7 Conclusion 11 Scope for future studies 12 References 13 Introduction The Lynfield local police force has collected the stop and search data for the last year. The force wants an analysis to be done on the data. The police force wants to identify suitable trends that might define stop and search data for the last year. In particular, the police force wants to ascertain if there are any trends that indicate disproportionate and/or discriminatory exercise of stop and search. For this purpose, the police force has collected the stop/search data for last one year. The police force has collected a total of 22 variables for each stop/search incident. Thorough quantitative analysis will be undertaken to observe the descriptive as well as inferential statistics for the data. Research on crime data and the ways to statistically analyze the same has been the focus of many researchers. The first step towards analyzing crime and law enforcement statistics, it is important that the researcher understands the data which forms the basis of statistics. Researchers have identified three primary types of data: crime, calls for service, and arrests (Boba, 2009). One of the key areas of research of crime data has been the analysis of policing and the racial background of the suspect. The debate on policing and race has been one that has been a very long discussion. Reiss (1971) stated that the decision to stop, search or arrest a suspect is influenced by the race of the suspect and that of the neighborhood of the suspect. Smith, Makarios, and Alpert (2006) suggested that the suspicion amongst police officers is influenced by the race of suspect and characteristics of neighborhood. Research has suggested the existence of a perception among minority citizens of police disproportionately stop Africa-American and Hispanic motorists, and that these citizens have a greater probability of being searched or arrested once stopped (Gross and Barnes, 2002). There are many ways that can be utilized for analyzing the relation between the police’s propensity to stop/search and the racial background. One of the popular methods used is survey (Durose, Schmitt, and Langan, 2005). Another way of research is observation of the police behavior, which is however prone to issues related to validity (Alpert et al. 2005). For this particular research, we have used the survey approach. We have collected the data for the stop/search incidents in the past one year and a set of questions related to the same were asked. The data so received was analyzed. Analysis The statistical analysis of the data collected is an important step in understanding the relationship between the racial profile and policing. Before starting the statistical analysis, it is important that we understand the data that was collected. List of Questions that were asked during the survey The survey administered consisted of 22 questions. The following questions were asked for each stop/search incident: 1. Was it a Stop or Search incident? 2. What was the weekday of stop/search? 3. What was the Time of stop/search? 4. What was the time Periods of stop/search (time recoded into 4-hour periods)? 5. What was the location of stop/search? 6. What was the gender of suspect? 7. What was the age of suspect? 8. What is the suspect’s employment? 9. Did suspect have previous record? 10. What was the suspect’s activity prior to stop/search? 11. What was the reason for stop/search 12. What was the authority for stop/search 13. What was the ethnicity defined by officer? 14. What was the suspect’s ethnicity defined by self? 15. Was suspect carrying a weapon? 16. Was suspect arrested? 17. What was the age of Officer involved in stop? 18. How worried was suspect about crime in their area 19. Was any complaint made? 20. If complaint made, how satisfied was suspect with response 21. Can police be trusted to deal fairly with all sections of community? 22. Step1 of Time Recoding to 4-hour periods (not to be used for analysis) Hypotheses A hypothesis can be defined as a statement that we assume to be true for the entire population. Hypothesis testing involves testing the statement for a sample and then testing its significance for population. Descriptive Statistics Age is an ordinal data item that can take any positive value. The table below shows the descriptive statistics for age of the suspects: Descriptive Statistics for age N Minimum Maximum Mean Std. Deviation Age of suspect 10414 0 80 34.38 11.307 Valid N (listwise) 10414 The minimum age of suspect stopped/searched was 5 years while maximum was 80. The average age of the suspect being stopped/searched was 34.38. Standard Deviation which is an indication of the variance of the suspect was 11.307 based on the data. The table below shows the frequency distribution of days of the week on which the stop/search was made: Weekday of stop/search Frequency Percent Valid Percent Cumulative Percent Valid Monday 1633 15.4 15.4 15.4 Tuesday 1784 16.8 16.8 32.2 Wednesday 1779 16.8 16.8 49.0 Thursday 1798 16.9 16.9 65.9 Friday 1310 12.3 12.3 78.3 Saturday 1169 11.0 11.0 89.3 Sunday 1136 10.7 10.7 100.0 Total 10609 100.0 100.0 It can be seen that the maximum percentage of stop/search were made on Thursday (16.9%). Almost equal percentage of stop/searches were made on Tuesday and Wednesday as well (16.8%). Sunday had the lowest number of stop/searches being made with just 1136 stop/searches, which is about 10.7% of the total count. Hypothesis 1 Generally, it is seen that the age of the suspect is a criteria of being searched/stopped. It can be assumed that the average age of suspects stopped/searched is 25. Ho: Null Hypothesis: The average age of suspects stopped/searched is 25 H1: Alternative Hypothesis: The average age of suspects stopped/searched is not 25. The age here is independent variable. The appropriate test for testing this hypothesis is the two-tailed one sample t-test for mean. The table below shows the output of the same for the data: One-Sample Test Test Value = 25 t df Sig. (2-tailed) Mean Difference 99% Confidence Interval of the Difference Lower Upper Age of suspect 84.647 10413 .000 9.379 9.09 9.66 As can be seen from the table, the sig value is less than the critical value of 0.05. Therefore, we do not accept the null hypothesis. Therefore, the average age of suspects being stopped/searched is not 25. Hypothesis 2 Ho: Null Hypothesis: The percentage of male and female suspects searched/stopped is same H1: Alternate Hypothesis: The percentage of male & female suspects searched/stopped is not same The appropriate test for testing this hypothesis is the two-tailed one sample t-test for proportion. The table below shows the frequency table of males to females in the survey: Gender of suspect Observed N Expected N Residual Male 9811 5229.0 4582.0 Female 647 5229.0 -4582.0 Total 10458 The chi-square test for the same is given below: Gender of suspect Chi-Square 8030.111a df 1 Asymp. Sig. .000 As can be seen from the table, the sig value is less than the critical value of 0.05. Therefore, we do not accept the null hypothesis. This implies that the percentage of male and female suspects searched/stopped is not same. There is a significant difference in the gender of suspects being stopped/searched. From the frequency table, it was seen that there are more number of male suspects as compared to female suspects. Hypothesis 3 Ho: Null Hypothesis: The proportion of stopped suspects is more than that of searched suspects H1: Alternate Hypothesis: The percentage of stopped suspects is not more than that of searched suspects The appropriate test for testing this hypothesis is the two-tailed one sample t-test for proportion. The table below shows the frequency distribution of the data with respect to whether the suspect was stopped or searched: Stop or Search? Observed N Expected N Residual Stop 7296 5304.5 1991.5 Search 3313 5304.5 -1991.5 Total 10609 The Chi-square test on the same will yield if there is a significant difference in the proportion of the suspects being searched and the suspects being stopped. The table below shows the Cho-square test output for the same: Chi-Square tests for Stop or Search? Stop or Search? Chi-Square 1495.361a df 1 Asymp. Sig. .000 a. 0 cells (.0%) have expected frequencies less than 5. The minimum expected cell frequency is 5304.5. As can be seen from the table, the sig value is less than the critical value of 0.05. Therefore, we do not accept the null hypothesis. This implies that the proportion of searches is not equal to the proportion of stops. From the frequency table, it is evident that more percentage of people is stopped as compared to being searched. Bivariate Analysis Employment and arrest/non-arrest Generally, the arrest or non-arrest of a suspect might be dependent on the officer’s knowledge of the employment of the suspect. The table below gives the cross tabulation of the employment of the suspect to whether he/she was arrested/ not arrested. Suspects Employment*Was suspect arrested? Crosstabulation Was suspect arrested? Total No Yes Suspects employment Professional 613 59 672 Skilled 1840 168 2008 Semi Skilled 2604 194 2798 Manual non skilled 2434 211 2645 Unemployed 1741 142 1883 Total 9232 774 10006 It can be seen that the least number of suspects arrested are in the professional category. The proportion of suspects not arrested is almost same in all the cases. Reason for stop/search and officer defined ethnicity It is generally claimed that there is a relationship between the reason for stop/search and the ethnicity of the suspect. Many researchers, as discussed in the introduction have claimed that ethnicity is a major reason behind being stopped/searched. The table below shows the cross-tabulation between the reason for stop/search and officer defined ethnicity. Reason for stop/search * Officer-defined ethnicity Crosstabulation Officer-defined ethnicity Total White Black Asian Other Not Recorded Reason for stop/search Officer Intuition Count 2217 645 362 41 60 3325 % within Reason for stop/search 66.7% 19.4% 10.9% 1.2% 1.8% 100.0% Suspect acting suspiciously Count 1954 578 399 38 36 3005 % within Reason for stop/search 65.0% 19.2% 13.3% 1.3% 1.2% 100.0% Called to Scene Count 952 250 138 23 11 1374 % within Reason for stop/search 69.3% 18.2% 10.0% 1.7% .8% 100.0% Prior Information Count 1004 269 170 17 4 1464 % within Reason for stop/search 68.6% 18.4% 11.6% 1.2% .3% 100.0% Public Complaint Count 976 248 184 22 11 1441 % within Reason for stop/search 67.7% 17.2% 12.8% 1.5% .8% 100.0% Total Count 7103 1990 1253 141 122 10609 % within Reason for stop/search 67.0% 18.8% 11.8% 1.3% 1.1% 100.0% It can be seen that overall 67% of the suspects that were stopped/searched were White. The appropriate test for testing if there is a significant difference in the mean value of reasons of stopping for different officer defined ethnicity is one-way ANOVA. It can be used to test if there is significant difference in the variance of reason of stops as per the officer defined ethnicity. Ho: Null Hypothesis: There is no significant difference in the variance of reasons for stop/search for different officer defined ethnicity groups H1: Alternate Hypothesis: There a significant difference in the variance of reasons for stop/search for different officer defined ethnicity groups The table below provides the result of one-way ANOVA between the two variables: ANOVA Reason for stop/search Sum of Squares df Mean Square F Sig. Between Groups 49.207 4 12.302 6.263 .000 Within Groups 20827.041 10604 1.964 Total 20876.248 10608 As can be seen from the table, the sig value (p) is less than the critical value of 0.005. Therefore, we do not accept the null hypothesis. Therefore, there is a significant difference in the mean value of the mean reason for stopping/searching for different groups of officer defined ethnicity. Recoding of Age Age is a continuous variable which can take any positive value. For our data, the minimum age is 0 years and maximum is 80. We recode the age variable into categories of 10 years each and name the transformed variable as RecAge. We utilize the transformation function available in SPSS. The age of the suspect may be related to the reason for stop/search. The transformed variable (RecAge) shall also be related to the reasons for stop/search. Ho: Null Hypothesis: There is no significant relation between the recage and reason for stop/search H1: Alternate Hypothesis: There is a significant relation between the recage and reason for stop/search Both the variables are categorical in nature. In order to test the significance of relationship between the two variables, the appropriate test is the Chi-Square test. The p-value is 0.002 is less than the critical value of 0.005. Hence, we do not accept the null hypothesis. Therefore, this indicates that there is a significant relationship between the recage and the reason for stop/search. Chi-Square Tests on RecAge and Reason for stop/search Value df Asymp. Sig. (2-sided) Pearson Chi-Square 55.377a 28 .002 Likelihood Ratio 57.487 28 .001 Linear-by-Linear Association 1.282 1 .258 N of Valid Cases 10414 a. 8 cells (20.0%) have expected count less than 5. The minimum expected count is .65. Conclusion Analysis of the stop/search data for the last one year has yielded some very useful insights in the area of policing, racial profiling, age and other statistics. It was seen from the data that most of the stop/searches were done on males as compared to females. The average age of the sample data that was collected was 34.1 years. Maximum number of searches/stopped were made on Thursday in the last year. Also, there were more number of stops being made as compared to the number of searches. From the cross tabulation between “suspects employment” and “was suspect arrested?”, it was evident that maximum number of suspects arrested were in the category of manual non-skilled. This category also has the maximum number of stops/searches being made. Professionals were least arrested. In addition, it was found that there is a significant difference in the mean reason for stop/search for different groups of officer defined ethnicity. This is in line with previous studies that we have discussed in the introduction. We also transformed age into a categorical value. Chi-square test suggested that there is a significant relationship between the age of the suspect and the reason for which s/he was stopped/searched. Based on the survey, it was also revealed that majority of the respondents believe that police can be trusted to deal fairly with all sections of community. Scope for future studies While the present study various key aspects related to stop/search, there are still areas which need to be explored more. There can be more focus on the kind of complaints being made to the police department. In addition, the demographics of the crime data can be studied to identify potential hotspots for criminal activities. In addition, very few of the respondents had responded to the question “21. Can police be trusted to deal fairly with all sections of community?” Future studies should aim at getting this view from the public at large which will give the complete picture. References Alpert, G., MacDonald, J. H., and Dunham, R. G. (2005), “Police Suspicion and Discretionary Decision Making During Citizen Stops,” Criminology, 43, 407–434 Boba, R. (2009). Interpreting Crime Data and Statistics. [Online]. Retrieved from: http://www.iaca.net/ExploringCA/2Ed/exploringca_chapter8.pdf. Last Accessed on Jaunary 13th, 2013. Durose,M. R., Schmitt, E. L., and Langan, P. A. (2005), “Contacts Between Police and the Public: Findings From the 2002 National Survey,” NCJ 207845, Bureau of Justice Statistics, U.S. Department of Justice. Gross, S. R., and Barnes, K. Y. (2002), “Road Work: Racial Profiling and Drug Interdiction on the Highway,” Michigan Law Review, 101, 653–754. Reiss, A. (1971), The Police and the Public, New Haven, CT: Yale University Press. Smith, M. R., Makarios, M., and Alpert, G. P. (2006), “Differential Suspicion: Theory Specification and Gender Effects in the Traffic Stop Context,” Justice Quarterly, 23, 271–295. Read More