Identity Theft Victims: Urban, Head of Household, High Education, High Income - Prof. Gali, Study Guides, Projects, Research of School management&administration

Download Identity Theft Victims: Urban, Head of Household, High Education, High Income - Prof. Gali and more Study Guides, Projects, Research School management&administration in PDF only on Docsity! Team Project Characterizing victims of identity theft By Mehmet Hondur Benjama Kounthongkul Brenda Martineau Patcharaporn Makarasara Sophie Shuklin BUDT 733 Prof. Galit Shmueli May 10, 2007 2 Executive Summary Background and Task We acquired an Identity Theft Survey report from the Federal Trade Commission administered in September 2003 by Synovate Research Company1. The data sample included 4,057 observations with 46 variables obtained from four waves of surveys. Among 4,057 observations, about 700 experienced identity theft (17.25%). Since 60% of our group members personally experienced identity theft before, we were interested in two main things: 1. verify our assumptions of what type of individuals are more prone to be identity theft victims, 2. verify the reported results of the FTC report. Some of the questions we wanted to explore were: are men or women more prone to being victims of identity theft? Are there differences in victimization depending on where you live: region and urban vs. rural setting? Does internet use make a difference? Data Since this data comes from a survey, there were many records that had “bad” data. We ended up eliminating about 4% of the data due to missing, invalid or inconsistent occurrences. As a result we went from 4,057 records down to 3,887 with 17.59% victims. Analysis Results From our findings we concluded that the following profile of individual is more prone to identity theft – those that live in an urban area, are head of household, have high education, live in South or West regions as opposed to the Midwest, have Internet at home or work, are between the ages of 31 and 55, are high income and whites are more likely to be victims regardless of how much they earn. We also saw that those who do not report income have a lower incidence of theft. Recommendations Based on our analysis and findings we were able to isolate the victims of identity theft to three primary characteristics: high education, West and South region and high income. From the consumer point of view those who attend college and beyond need to ensure they use their personal information carefully. Also, those individuals that live in the South or West should be more discrete. Those that earn high income also need to ensure they know how to protect themselves. Federal and local governments can use this information to pursue training for those in law enforcement as well as those at risk. Businesses should be aware of these findings and use them to educate employees and customers. These actions could reduce societal costs of identity theft which was at $47.6 billion in 2002, according to the FTC report. Survey administration – there are number of ways to improve the administration of the survey by 1) organizing questions that are not repetitive 2) asking more concrete questions and ensuring that the same questions are being asked across all rounds. Some questions that might help are: how do you dispose of personal papers? Do you use software encrypted sites when on-line? Do you pass personal information via wireless communications? 1 http://www.ftc.gov/os/2003/09/synovatereport.pdf 5 Exhibit 1 – example of pie chart (combine Internet variable) Pie Chart (2) combined internet -0.2 0 0.2 0.4 0.6 0.8 1 1.2 0; 89% 1; 11% 0; 80% 1; 20% Red – Non victims Blue – Victims of identity theft Exhibit 2 – box plot of age variable 10 20 30 40 50 60 70 80 90 100 0 1 Range Mean UAV LAV 81.0 81.0 47.5 46.1 99.0 87.0 18.0 18.0 Combined ID theft a g e Exhibit 3 – Model 1 Exhibit 4 – Exhaustive search The Regression Model Coefficient Std. Error p-value Odds -2.721699 0.18792033 0 * 3878 -0.2861056 0.10188214 0.0049819 0.75118327 3502.971191 0.35316476 0.16072637 0.02799872 1.42356563 17.5971186 0.36140341 0.09561712 0.00015702 1.43534231 8 0.33606958 0.09847201 0.00064289 1.39943635 0.03144305 0.37880364 0.11104752 0.00064681 1.46053624 0.39212537 0.11504573 0.00065338 1.48012328 0.2267748 0.09332977 0.01510621 1.25454736 0.22234415 0.08875898 0.012244 1.24900115 Training Data scoring - Summary Report 0.25 Actual Class 1 0 1 168 516 0 440 2763 Class # Cases # Errors % Error 1 684 516 75.44 0 3203 440 13.74 Overall 3887 956 24.59 Cut off Prob.Val. for Success (Updatable) Classification Confusion Matrix Predicted Class Error Report region_West combined internet income with missing data binning_high age_Bin_Middle Multiple R-squared Constant term rural head of HH High education region_ST Residual df Residual Dev. % Success in training data # Iterations used Input variables 1 2 3 4 5 6 7 8 9 10 11 12 2 3942.123535 60.14414597 0.00000003 Constant High education * * * * * * * * * * 3 3921.002686 41.01782608 0.00002276 Constant High educationmbined internet * * * * * * * * * 4 3912.311768 34.32466125 0.00022865 Constant High educationmbined internetage_Bin_Middle * * * * * * * * 5 3903.959229 27.96995735 0.00191878 Constant rural High educationmbined internetage_Bin_Middle * * * * * * * 6 3897.391846 23.40087509 0.00870676 Constant High education region_ST region_Westmbined internetage_Bin_Middle * * * * * * 7 3889.130127 17.1370163 0.05653239 Constant rural High education region_ST region_Westmbined internetage_Bin_Middle * * * * * 8 3883.302734 13.30811596 0.16722041 Constant rural High education region_ST region_Westmbined internet ta binning_highage_Bin_Middle * * * * 9 3878.590576 10.59473801 0.34391272 Constant rural head of HH High education region_ST region_Westmbined internet ta binning_highage_Bin_Middle * * * 10 3875.092529 9.09578514 0.51956803 Constant rural head of HH High education race_Other region_ST region_Westmbined internet ta binning_highage_Bin_Middle * * 11 3871.366211 7.36850214 0.74536771 Constant rural head of HH High education race_Black race_Other region_ST region_Westmbined internet ta binning_highage_Bin_Middle * 12 3869.122803 7.12451315 0.84976041 Constant rural head of HH High education race_Black race_Other region_NE region_ST region_Westmbined internet ta binning_highage_Bin_Middle 13 3867.59375 7.59506464 0.89910883 Constant rural head of HH High education race_Black race_Other region_NE region_ST region_Westmbined internet ta binning_high binning_middle 14 3866.160156 8.16109943 0.9394502 Constant rural head of HH High education loyment_Retire home owner race_Black race_Other region_NE region_ST region_Westmbined internet #Coeffs RSS Cp Probability Model (Constant present in all models) 6 Exhibit 5 – Pie chart with relationship with income bin Model 2 – The best model with interaction term of race_white and income midpoint The Regression Model Coefficient Std. Error p-value Odds -2.88297915 0.19261804 0 * 3877 -0.2596491 0.10230482 0.01114896 0.77132219 3486.922119 0.3963179 0.16219139 0.01454476 1.48634171 17.5971186 0.32000002 0.09767967 0.00105283 1.37712777 9 0.01073969 0.00227312 0.00000231 1.01079762 0.03588055 0.32555714 0.09886307 0.00099121 1.38480198 0.35945141 0.1115823 0.00127565 1.43254328 0.31102306 0.11866625 0.00876748 1.36482072 0.17830224 0.09010424 0.04783356 1.1951865 -0.00535986 0.00186799 0.00411359 0.99465448 Training Data scoring - Summary Report 0.25 Actual Class 1 0 1 171 513 0 391 2812 Class # Cases # Errors % Error 1 684 513 75.00 0 3203 391 12.21 Overall 3887 904 23.26 Error Report income midpoint*race_white Cut off Prob.Val. for Success (Updatable) Classification Confusion Matrix Predicted Class region_ST region_West combined internet age_Bin_Middle Multiple R-squared Constant term rural head of HH High education income midpoint (k) Residual df Residual Dev. % Success in training data # Iterations used Input variables Chi-square income with missing data binning -0.2 0 0.2 0.4 0.6 0.8 1 1.2 high low middle unknown 0; 78% 1; 22% 0; 88% 1; 12% 0; 83% 1; 17% 0; 85% 1; 15% 0; 69% 1; 31% 0; 83% 1; 17% 0; 82% 1; 18% 0; 90% 1; 10% Chi-square income with missing data binning 0 1 high low middle unknown 0; 69% 1; 31% 0; 86% 1; 14% 0; 78% 1; 22% 0; 91% 1; 9% 0; 79% 1; 21% 0; 88% 1; 12% 0; 84% 1; 16% 0; 85% 1; 15%