BJS Response to OMB Passback

Response to OMB Passback Questions on the 2012/2014 ITS

1. Supporting Statement (SS) A1 says that because of the plan for this supplement, “the household identify theft questions will be removed from the NCVS screener with the implementation of the 2012 ITS.” 

   1. Can BJS confirm that it is its intention that no household should be asked the core and the supplement questions during the same interview?

BJS: Yes, it is BJS intention that no household be asked the core and supplement identity theft questions during the same interview.

2. Is it the case that some households may be asked both during different interview months or will all households be asked one or the other only during the first month in sample?

BJS: Households that rotated into the sample prior to 2012 will have responded previously to the household identity theft questions during 2011.

While the ITS does contain some similar questions as those asked of the household, the 2012 ITS findings will not be linked to the prior household identity theft data so there are no concerns with duplication among those households that responded to the household question in 2011.

3. Shouldn’t OMB therefore be seeing a change request to the main NCVS package in concert with this package?

BJS: BJS is submitting a change request to the main NCVS for the reduction in burden associated with dropping the household identity theft questions in 2012. Because BJS is planning to administer the ITS every other year and regularly collect person-level data, BJS will not be reinstating the household identity theft questions after this year. The reduction in burden hours for the core NCVS because of the removal of these questions will also be reflected in the main NCVS OMB package that will be submitted to OMB in May of 2012.

2. Questionnaire:

   1. The 2008 questionnaire in OMB’s system appears to be corrupted so we are not able to compare the new and old forms itself. Please provide a copy of the 2008 questionnaire. 

BJS: The questionnaire is attached.

2. In addition, please provide a copy of the cognitive testing report(s) and a summary of which questions changed.

BJS: The first attached document is a memo to the Census Bureau summarizing the major changes from 2008 to 2012.

Also attached are the first and second round recommendations from the Census Bureau’s cognitive testing. After the first round of testing, the major changes to the instrument included:

• Ordering of ITS screener questions was changed to ask about the misuse of an existing checking account or debit card prior to asking about the misuse of an existing credit card. This change was made to prevent respondents from thinking they should report the misuse of a debit card when asked about existing credit card accounts.

• Changing of the wording of instructions before 6a and 6b to provide additional clarification for respondents on the meaning of the term ‘incident of identity theft.’ These questions are needed to guide respondents who experienced more than one incident of identity theft to think about only the most recent incident when responding to the remaining questions. Some of the confusion was on the part of the Census Bureau regarding what BJS was trying to correct, but after making adjustments to the instructions there did not seem to be confusion during the second round of testing.

• Adjusting the instructions to respondents before section B that asked those with more than one incident of identity theft to think about the most recent incident when answering questions.

• Changing the wording of Q13 so that respondents would report contact with a credit bureau even if the contact had occurred online or was initiated by the credit bureau.

• Adding a ‘don’t know’ response to the question about whether the fraud alert placed on an account was a ‘seven year fraud alert.’

• Clarifying the question Q32, “Did you seek any kind of professional or medical help for the physical problems you experienced as a result of the misuse of personal information,” to attempt to make it clearer that if the respondent suffered injuries, for example, during a robbery in which identifying information was obtained, the injuries from the robbery should not be reported. The respondent should only report injuries or the need for help resulting from the identify theft itself.

• Adding a sentence to Q38a/Q38b to ensure that respondents do not double count financial loss and report the same loss amount as both direct and indirect loss.

• Consistently using the term ‘identity theft incident’ throughout the instrument rather than switching between ‘identify theft incident’ and ‘identity theft event”

• In the new section on the long-term consequences of identity theft, making sure that all three of the questions have the same wording; and incorporating the concept of long-term consequences into the main question and including past year concept in the sub-items.

• Changing the ordering of the questions in the Commercial and Credit Behaviors section to make the questions flow better and keep all of the questions pertaining to online shopping in a group.

The second round of cognitive testing did not result in substantive changes to the survey instrument. It did reveal several areas in which the field representatives’ (FR) training manual should provide instructions for the FRs on how to handle specific questions that came up during the testing. For example, one respondent had closed down an account that someone had misused and was not sure whether he should report that identity theft because the account was no longer active. Another respondent had discovered an incident of identity theft during the past 12 months, but the identity theft started happening more than one year prior and there was uncertainty as to whether an incident’s inclusion should be based on initial misuse or respondent’s discovery of the misuse. The FR training materials will make note of these potential areas of confusion and advise the FRs on how to code these types of responses.

3. Are the response options to Question 12 read to respondents?  All at once prior to a response being accepted?  Were there this many in 2008?  Were items rotated?  Was analysis done to be sure that response patterns suggest that respondents are persisting through the list effectively?

BJS: The response options are not read aloud to the respondent. These response options have not been rotated and are in the same order that they were in 2008. The FR selects the most appropriate response from the list based on what the respondent said, and there is no evidence from 2008 to suggest that FRs primarily selected response options from the beginning of the list. Additionally, none of the respondents from the 2012 cognitive testing had difficulty understanding or answering this question.

There were 11 response options to question 12 in 2008, compared to the 13 response option on the 2012 ITS. The two items that changed from 2008 to 2012 were: 1. in 2008, respondents who became identity theft victims after losing a wallet or checkbook were combined into one response with respondents who experienced identity theft after someone stole their wallet or checkbook. However, we found these to be substantively different ways of having one’s identifying documents obtained and separated the response categories. 2. based on the ‘other’ responses received in 2008, we also added a category for ‘stolen during an online purchase/transaction,’ which respondents seemed to think of separately from other (in-person) purchases/transactions.

4. In question 12, there doesn’t seem to be an obvious response category to data breaches by online retailers, since 12-6 is  for “during” a transaction and 12-12 is written to suggest a physical location.  Does this need to be updated?

BJS: In 2008, three respondents reported that their information had been obtained through a data breach. These three responses were originally coded as ‘other specify’ responses but Census recoded the literal responses to reflect a new category for analysis. With only three unweighted cases, however, this category ultimately had to be combined with other categories for the purpose of analysis and reporting. We based our decision not to have a separate category for “data breach by online retailers” based on the small number of respondents in the 2008 supplement who reported that their information was obtained through such a mechanism. This information can still be obtained and recorded in the ‘other specify’ category.

Additionally, in the commercial and credit section (Q57) we do ask respondents whether they have been notified that their personal information may have been compromised by a data breach. In 2008, about 7% of respondents had received a notification about a breach, but as noted above, a small number of these breaches result in personal information being misused (to the extent that the respondent is aware).

5. Section E: victim impact.  We do not understand the utility of the full set of questions, especially given the low prevalence of non-financial victim impact reported in the 2008 report.

   1. Did BJS examine the correlation between impacts and who the offender was?  We could hypothesize that if the offender was known, it would stand to reason that such personal relationships would be affected.  Otherwise, what is the hypothesis at work?  If most of those reporting such impacts were due to knowing the offender, this set of questions seems unnecessary for the full universe, and may not even be necessary for the subset who knew their offender.  Please discuss.

BJS: These questions were developed through consultation with a panel of experts including Dr. Dean Kilpatrick who has done substantial research in the areas of mental health, Post-Traumatic Stress Disorder (PTSD) and victimization (see http://www.musc.edu/psychiatry/faculty/kilpatrickd.htm) and Dr. Kevin Becker, a psychologist, who has spent over 20 years dealing with psychological trauma and victimization (http://www.realmedicinefoundation.org/member/kevin-becker-psyd). The questions are largely derived from psychological literature on PTSD, which has been found to manifest in some crime victims, ¹ and they are intended to measure the central components of most clinical diagnoses of trauma.

This same set of victim impact questions that is included on the 2012 ITS has been a part of the core NCVS since 2008 and were asked in the 2008 ITS. The primary purpose of these questions is to compare impacts of identity theft with impacts of other crime types (e.g., the range of violent victimizations from simple assault through rape). From these, the Office for Victims of Crimes (OVC), which provided funding for the ITS in 2008 and 2012, will be able to assess some dimensions of the need for and types of victim services among a group of victims that they believe is underserved by victim services agencies. OVC has hypothesized that the emotional tool of victims of ID theft is high. OVC acknowledges that victim services traditionally have focused on violent crime victims, and while the 2008 ITS showed that some identity theft victims also experience emotional and physical distress as a result of their victimization, from OVC’s perspective, these data open up the possibility that there are groups of non-violent crime victims who need services and support but are often not targeted by service providers.

In order to maintain item consistency and skip patterns with the core NCVS, the victim impact questions are only asked of victims who reported that the victimization was moderately or severely distressing. Therefore, in 2008, we were unable to report on the full range of victim impact questions because of the relatively small number of identity theft victims that were asked these questions. Part of the rationale for routinely conducting the ITS is that eventually we will be able to roll up multiple years of data (as is often done with the core NCVS) to be able to analyze these types of rare outcomes with more precision and reliability. It is important to collect these data for agencies like the Office for Victims of Crime, which have no other source of data on the emotional and physical impact of identity theft victimization and no other means for determining the resources needed to provide support and services for identity theft victims.

BJS was unable to examine the correlation between impacts and knowing the offender, because in 2008 there was a problem with the ITS questionnaire skip pattern and offender data could not be used. We do not currently know what impact knowing the offender has on identity theft victims and whether victim-relationship is a more important characteristics than other characteristics, such as the amount of loss, time to resolve the issues, length of victimization, in determining the amount of distress that a victim experiences. With the 2012 ITS,BJS has corrected the problem with the offender section on the instrument and, as a result of the problems in 2008, will be getting a Census Bureau laptop to test the CATI instrument and ensure that it has been properly programmed before it goes out into the field. Until we get this data, we cannot hypothesize about the extent to which knowing the offender is a driving factor in a victim’s emotional and physical response to the victimization and not having other sources of data on the impact of identity theft is part of the justification for including these questions. However, even the Attorney General Guidelines for Victim and Witness Assistance (http://www.justice.gov/olp/pdf/ag_guidelines2011.pdf) acknowledges that there may be harms caused to identity theft victims beyond the harm caused when the offender is a family member or friend, “In extraordinary cases [of identity theft] there may be solely or primarily emotional harm, such as harm to reputation from a false arrest that is a direct result of the misuse, with negligible or no pecuniary harm.  Department personnel may consider extraordinary emotional harm when evaluating whether an individual should be classified as a victim for purposes of victims’ rights and services” (article III.D.3). This being said, as noted above, the reasoning behind these questions is that violent victimization in particular, has been linked to conditions like PTSD.

2. Please provide cites and a summary of the literature to justify each question in section F.

BJS: (we believe you mean section E rather than F) These questions have been asked on the NCVS since 2008 and the justification for including them on the identity theft supplement is to gain an understanding of the full impact (financial, emotional, physical/physiological) of identity theft on victims and to compare the impact of identity theft with impact of violent victimization. There is no other source of data that allows for this type of examination.

The questions are all derived from widely used psychological scales that measure functional impairment, distress and mental health, such as the K6 (http://www.hcp.med.harvard.edu/ncs/ftpdir/k6/IWER%20K6.pdf,² and the Symptom-Checklist-90-Revised (SCL-90-R).³ The questions were developed in consultation with psychologists, such as Dr. Dean Kilpatrick, who has done extensive research examining the emotional and psychological effects of victimization and methodological issues with the measurement of the effects of trauma.⁴

The complete set of questions provides an assessment of victim distress that aligns with Kilpatrick’s and other’s work in the field of psychology on the mental health impact of victimization. The questions on the ITS are identical to the questions that have been asked on the NCVS for the past five years, and this consistency allows for a comparison of the victim impact from identity theft to the victim impact from violent victimization. Again, victim service providers and the agencies that allocate funding for victim services have expressed a need for better understanding the impact of crimes other than UCR index crimes in order to better account for the needs of these victims. While victim services are traditionally focused on violent crime victims, we know from the basic assessment of the victim impact data from the 2008 ITS that some identity theft victims also experience emotional and physical distress as a result of their victimization. These are victims who may also be in need of services and support but without the data from the NCVS there is no way to measure the full impact of identity theft on victims or whether the emotional and physical harm caused by identity theft is changing over time.

3. Response categories to many of the questions to not appear to have been used in the 2008 report (e.g., item 30 and 31).  Were they used elsewhere?  How?  How did BJS decide that these lists was better than open coding or shorter lists?

BJS: In order to maintain consistency with the core NCVS, the majority of the victim impact questions are only asked of victims who reported that the victimization was moderately or severely distressing. Therefore, in 2008, as you note, we were unable to report on the full range of victim impact questions because of the relatively small number of identity theft victims that were asked these questions. Part of the rationale for routinely conducting the ITS is that eventually we will be able to roll up multiple years of data (as is often done with the core NCVS) to be able to analyze these more rare outcomes with more precision and reliability. These questions add little in terms of burden to the survey because they are asked of the relatively small group of victims who report moderate or severe distress, but since data on the full emotional, physical, and financial impact of identity theft victimization is not available from other sources, it is anticipated that the information will increasingly be utilized. While a case may be made to delete these items from the ITS:2012 due to small sample sizes, we believe maintaining consistency between the ITS and NCVS core items is preferred.

With open-ended questions it is impossible to ascertain if a respondent did not report an outcome because the outcome did not occur or because the respondent failed to remember the outcome or did not think it was relevant to include in the answer. During the cognitive testing in 2008, respondents had no problems understanding or responding to the close-ended questions. On question 30, a few respondents provided other responses, such as ‘frustrated,’ ‘driven to justice,’ and ‘cautious.’ However, there were no patterns to the other responses to suggest that new categories should be created. Again in the 2012 ITS, respondents will have the opportunity to provide an ‘other’ response to Q30, Q30b, Q31, and Q33 if the close-ended categories do not fit their unique situation.

4. Please provide evidence that individuals are able to meaningfully differentiate between concepts like “mildly distressing” and “moderately distressing” in item 29 or “vulnerable” and “violated” in item 30.

Again, these items are derived from commonly used psychological scales such as the K6 (http://www.hcp.med.harvard.edu/ncs/ftpdir/k6/IWER%20K6.pdf), the PTSD checklist (http://macmhb.org/StateWide%20Trauma%20Seminar/46%20PTSD%20Checklist%20-%20Civilian%20Version%20p.81.pdf), and Symptom-90-Checklist-Revised. These scales are all used widely in psychological research and evaluation⁵ and all utilize response categories that require respondents to differentiate between ‘a little’ and ‘moderately’ or ‘a little’ and ‘some.’ These items have also been included on the NCVS since 2008 and from the NCVS data and the 2008 ITS data we know that respondent do differentiate between these categories.

5. Did anyone say “not at all distressing” to item 29 yet yes to the ”significant problems” questions (27 and 28)?  If not, then this suggests that folks could be screened out of the section after 28. 

BJS: Yes, there were victims who reported yes to at least one of the significant problems and also reported that the identity theft was ‘not at all’ or ‘mildly’ distressing. There were also victims who reported that the experience was severely distressing but did not report problems at work/school or with friends/family.

6. Then, item 29 could read for clarity “Did these significant problems cause…”, and using simplified response categories like “some” and “a lot.”

BJS: A victim’s level of distress may not be tied to problems at work or with family members or friends. Changing item 29 in would lead respondents to think about distress only in terms of these two limited impacts rather than to think about the totality of the experience. This would change the meaning of the question and it would no longer be comparable to the core NCVS data on the distress experienced because of a violent victimization.

1 See for example: Kilpatrick D.G., B.E. Saunders, L.J. Veronen, C.L. Best, and J.M Von. 1987. Criminal victimization: Lifetime prevalence, reporting to police, and psychological impact. Crime & Delinquency 33(4): 479-89.

2 See, for example: Wittchen, Hu. 2010. Screening for serious mental illness: methodological studies of the K6 screening scale. International Journal of Methods in Psychiatric Research 19:1-3.

3 Saunders, B.E., C.M. Arata and D. G. Kilpatrick. 1990. Development of a crime-related post-traumatic stress disorder scale for women within the symptom checklist-90-revised. Journal of Traumatic Stress 3(3) 439-448.

4 See for example: a. Ruggiero, A. A. Rheingold, H.S. Resnick, and D. G. Kilpatrick. 2006. Comparison of Two Widely Used PTSD-Screening Instruments: Implications for Public Mental Health Planning. Journal of Traumatic Stress 19(5): 699-707; b. Kilpatrick, D. G., K.J., R. Acierno, B.E. Saunders, H.S. Resnick, and C.L. Best. 2003. Violence and risk of PTSD, major depression, substance abuse/dependence, and comorbidity: Results from the National Survey of Adolescents. Journal of Consulting and Clinical Psychology 71(4): 692-700.

5 See, Kilpatrick, D. G., K.J., R. Acierno, B.E. Saunders, H.S. Resnick, and C.L. Best. 2003. Violence and risk of PTSD, major depression, substance abuse/dependence, and comorbidity: Results from the National Survey of Adolescents. Journal of Consulting and Clinical Psychology 71(4): 692-700.

Saunders, B.E., C.M. Arata and D. G. Kilpatrick. 1990. Development of a crime-related post-traumatic stress disorder scale for women within the symptom checklist-90-revised. Journal of Traumatic Stress 3(3) 439-448.