Skip to main content

Currently Skimming:

7 Potential Sources of Error in the NCVS: Sampling, Frame, and Processing
Pages 109-126

The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.


From page 109...
... ) showed larger estimates than did the NCVS (or National Crime Survey)
From page 110...
... Biemer (2010) , in the Handbook of Survey Research, outlines the components of nonsampling error in surveys: specification error, frame error, nonresponse error, measurement error, and processing error.
From page 111...
... Table 7-1 shows national-level estimates and their CVs from the NCVS for rape and sexual assault and serious violent crimes for 3 years: 2002, 2010, and 2011. For the category of all serious violent crimes (which includes rape and sexual assault, robbery, and aggravated assault)
From page 112...
... Estimate CVs (%) 2002 Rape and sexual assault 349,810 10.9 1.5 13.3 All serious violent crimes 2,306,710 5.2 10.0 5.0 2010 Rape and sexual assault 268,570 13.4 1.0 10.0 All serious violent crimes 1,694,840 6.4 6.6 6.1 2011 Rape and sexual assault 243,800 14.3 0.9 11.1 All serious violent crim 1,852,650 6.2 7.2 5.6 SOURCE: Data from Criminal Victimization, 2011 (Bureau of Justice Statistics, 2012a)
From page 113...
... Unfortunately, the sampling error for estimates of victimization rates for many subpopulations of interest can become quite large on the NCVS because there are very few affirmative responses to questions about serious violent criminal victimization in the sampled groups. Thus, BJS does not provide estimates for rape and sexual assault for these subpopulations; they only provide estimates for the larger category, serious violent crimes.
From page 114...
... for Selected Subpopulations Estimated Serious Violent Crime CVs (%) Category 2002 2010 2011 2002 2010 2011 Total 10.0 6.6 7.2 5.0 6.1 5.6 Gender Male 10.4 6.4 7.7 6.7 9.4 9.1 Female 9.5 6.8 6.7 7.4 8.8 9.0 Race/Ethnicity White 8.6 5.8 6.5 5.8 8.6 7.7 Black 17.8 10.4 10.8 9.6 12.5 13.0 Hispanic 12.3 6.7 7.2 11.4 13.4 13.9 American Indian/Alaskan Native 14.3 47.3 12.6 47.6 24.1 50.8 Asian 3.4 2.3 2.5 32.4 35.8 36.0 2 or more 17.7 26.2 29.4 24.0 Age 12-17 17.0 11.7 8.8 10.0 12.8 15.9 18-24 24.7 17.0 16.3 8.1 10.0 11.0 25-34 12.3 7.1 9.5 9.8 12.7 11.6 35-49 7.6 5.6 7.0 9.2 12.5 11.4 50-64 4.4 3.7 4.3 13.6 16.2 14.0 65 and older 1.8 0.9 1.7 22.2 33.3 23.5 Marital Status Never married 16.1 11.9 11.7 6.8 7.6 8.5 Married 5.7 2.2 3.7 8.8 13.6 10.8 Widowed 4.4 3.0 0.7 25.0 30.0 57.1 Divorced 10.9 11.2 9.2 13.8 13.4 15.2 Separated 34.8 18.8 26.4 15.2 20.7 18.9 SOURCE: Data from Criminal Victimization, 2011 (Bureau of Justice Statistic, 2012a)
From page 115...
... by Geographic Areas Serious Violent Crime CVs (%) Area 2002 2010 2011 2002 2010 2011 Total 10.0 6.6 7.2 5.0 6.1 5.6 Region Northeast 7.1 6.8 6.4 11.3 13.2 14.0 Midwest 11.5 7.6 7.8 8.7 10.5 11.5 South 10.8 5.4 6.5 7.4 11.1 10.8 West 9.5 7.5 8.4 9.5 10.7 9.5 Location of residence Urban 15.2 9.5 9.7 7.2 8.4 9.3 Suburban 7.8 5.5 5.7 7.7 9.1 8.8 Rural 7.9 4.7 6.7 11.4 14.9 14.9 SOURCE: Data from Criminal Victimization, 2011 (Bureau of Justice Statistic, 2012a)
From page 116...
... (2012) provide analysis of the distribution of series victimizations for all violent crimes, but they are not often able to isolate results for rape and sexual assault.
From page 117...
... 10) reported a maximum value of "750" incidents in series has been reported for serious violent crimes.
From page 118...
... The change ranged from a low of zero percentage change in 2007 (there were no series victimizations reported) to a high of 143 percentage change in 2009.
From page 119...
... CONCLUSION 7-2 Records identified as series victimizations cre ate an outlier problem in the estimation process for the National Crime Victimization Survey. The current method for handling series victimization, although an improvement over the method used until 2011, ­ llows these relatively rare reports to have a large impact on the a n ­ ational estimates of rape and sexual assault and creates large year-to year volatility.
From page 120...
... 120 ESTIMATING THE INCIDENCE OF RAPE AND SEXUAL ASSAULT TABLE 7-5  Effect of Series Victimizations on National Crime Victimization Survey Estimates of Rape and Sexual Assault, by Year Number of Victimizations Standard Standard Error Percentage Counting Series Error Ignoring Series Ignoring Change Year Victimization Series Victimization Series in Estimate 1993 898,239 71348 521,223a 51573 72 1994 674,291 47198 433,509a 36858 52 1995 563,249 42418 363,527a 32830 55 1996 437,198 37345 307,100 30430 42 1997 553,523 48035 311,110 34409 78 1998 391,101 39935 332,500 36103 18 1999 591,460 49907 383,170 38340 54 2000 366,747 39485 260,950 32152 41 2001 476,578 46216 248,250 31043 92 2002 349,805 38253 247,730 31352 41 2003 325,311 36759 198,850 27884 64 2004 255,769 33339 209,880 29887 22 2005 207,760 32551 190,592b 31032 9 2006 463,598 50305 260,940 36990 78 2007 248,277 32924 248,277 32924 0 2008 349,691 42837 203,830 31719 72 2009 305,574 39443 125,910 24079 143 2010 268,574 36057 188,380 29399 43 2011 248,803 34800 217,331 32616 14 Average Percentage Change Across Years 52 aThe estimates published in Criminal Victimization, 1993, 1994, 1995 were revised in 1996 to reflect a methodology change to estimate victimizations for the "collection year" rather than the year in which the victimization occurred. bBased on errata issued June 16, 2011.
From page 121...
... POTENTIAL SOURCES OF ERROR IN THE NCVS 121 Incidence Rate per 1,000 People (12+ years) Counting Series Ignoring Series Percentage Change in Victimization Victimization Estimate 4.3 2.5 87 3.2 2.1 60 2.6 1.7 63 2.0 1.4 43 2.5 1.4 79 1.8 1.5 20 2.6 1.7 53 1.6 1.2 33 2.1 1.1 91 1.5 1.1 36 1.4 0.8 75 1.1 0.9 22 0.8 0.8 0 1.9 1.1 73 1.0 1.0 0 1.4 0.8 75 1.2 0.5 140 1.0 0.7 43 55 SOURCES: Bureau of Justice Statistics (1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002a, 2003, 2004, 2005, 2006, 2007, 2008a, 2009, 2010, 2011, n.d.-a)
From page 122...
... . FRAME ERROR A sampling frame is the source material (such as a listing of people, ad dresses, or counties)
From page 123...
... Because the MAF is not comprehensively updated during this time and can become out-of-date, the Census Bureau uses a frame based on new building permits to supplement the existing MAF between censuses. Another secondary frame, the "area" listings, is generated from periodic canvassing of selected census blocks within the PSUs: they consist of identified residential units that are not on the other secondary frames.
From page 124...
... The use of a supplemental frame of new building permits and a periodic "area" canvassing of some primary sampling units are helpful, but not a complete solution. CONCLUSION 7-4 The frame for the ancillary listing of group quar ters, which is an important part of the secondary sample for the Na tional Crime Victimization Survey because their residents may be at higher risk for sexual violence, is seriously flawed in terms of both the building and enumeration of this secondary frame.
From page 125...
... CONCLUSION 7-5 The Bureau of Justice Statistics does not provide public information on the edit process in the National Crime Victimiza tion Survey, although processing and editing errors are an important part of any major survey data collection. The lack of transparency about these processes makes it difficult for data users to fully under stand the survey's estimates.
From page 126...
... attempted? sexual contact without force FIGURE 7-3  A flow chart for crime classification of reported victimizations on the National Crime Victimization Survey for the four most serious victimizations: completed rape, attempted rape, sexual attack with serious assault, and sexual attack with minor assault.


This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More information on Chapter Skim is available.