SUPPORTING STATEMENT
U.S. Department of Commerce
U.S. Census Bureau
2013 Current Population Survey Annual Social and Economic Supplement Content Test
OMB Control Number 0607-<XXXX>
Part A – Justification
Question 1. Necessity of the Information Collection
The Current Population Survey (CPS) Annual Social and Economic Supplement (ASEC) is used to produce official estimates of income and poverty, and it serves as the most widely-cited source of estimates on health insurance and the uninsured. Health insurance questions have been asked in the CPS since 1980 as a part of a mandate to collect data on noncash benefits. These statistics have far-ranging implications for policy and funding decisions, as described below. The Census Bureau has redesigned the health insurance and income instruments on the CPS, and this necessitates an evaluation of the operational functioning of the CPS instrument with the new health insurance and income battery.
The intention of the CPS ASEC redesign was to obtain an improved calendar-year estimate of health insurance coverage. The existing question series, which asks about coverage at any time in the past year, has been shown repeatedly to be flawed and not to produce estimates of coverage in the past year, but rather a mixture of current and past year coverage. Because of these concerns and evidence of the flawed estimates, a comprehensive research and evaluation program of the existing question series has been underway for the past 12 years. Results indicated that three key design features of the CPS ASEC were associated with measurement error: the reference period, the household-level design and the overall structure of questions asking about plan type. Each of these features was modified to reduce measurement error and tested in several iterative studies until no more evidence of problems was detected.
The redesign also contains new content, including questions on health insurance exchanges and questions on take-up of employer-sponsored insurance (ESI). Exchange questions were requested by the Health and Human Service’s Assistant Secretary for Planning and Evaluation (ASPE), and the questions on ESI take-up were requested by the Council of Economic Advisors.
The Census Bureau needs to act soon because it would like to capture trends associated with certain provisions of the ACA. If the content test is conducted in 2013, it allows the Census Bureau to request approval from the Office of Management and Budget (OMB) to implement the new CPS ASEC questions in the production run for 2014. This will be critical for setting a baseline of estimates captured under the redesign in order to measure the effects of the Affordable Care Act in future years. More specifically, the 2014 CPS ASEC will ask questions about coverage during calendar year 2013. If the production 2014 CPS ASEC instrument uses the redesign, then estimates will represent coverage in calendar 2013 before the Affordable Care Act is fully implemented. Such baseline information is of particular interest to the Centers for Medicare & Medicaid Services, the Health and Human Services (HHS) Assistant Secretary for Planning and Evaluation (ASPE), and the Council of Economic Advisors. These agencies are particularly interested in the new questions included in the content test regarding health insurance exchanges and take-up of employer-sponsored insurance.
The CPS ASEC content test will be conducted by telephone from three of the Census Bureau’s telephone data collection centers in March 2013 with retired CPS sample. The survey is being conducted under the legal authority of Title 13, United States Code, Section 182.
History of the CPS health insurance questions
Congress mandated that the Census Bureau collect data on non-cash benefits in order to show the impact of non-cash benefits on poverty. These new questions, implemented in 1980, dealt with food stamps, free/reduced price school lunches, etc. and health insurance. The time period of coverage for questions on these non-cash benefits matched that of the income questions (i.e.: the last calendar year). Presumably, the type of health insurance coverage one has should be consistent with the income that purchased it. The purpose of the CPS ASEC continues to be to obtain a measure of coverage and non-cash benefits that would match the calendar year income measure (hence, the use of insured over the year).
In 1996, the Children’s Health Insurance Program (CHIP) law was passed that provided federal funds to the states targeted for uninsured low-income children. The Census Bureau was given $10 million annually, starting in 2000, to increase the sample size of the CPS to 100,000 addresses, and to improve estimates of health insurance. The extra sample bought better state estimates (as measured by the coefficients of variation) for uninsured children by income groups. At the time, the CPS’s estimates of uninsured low-income children were used in the federal funding formula for CHIP.
During the most recent reauthorization of CHIP (2009), Congress decided to fund the Census Bureau at $20 million annually (from $10 million). Although the allocation formula for CHIP no longer includes an estimate of the state’s population of uninsured low-income children, the Census Bureau was directed to evaluate the ACS and the CPS. In particular, the bill stated that Census Bureau should “Make appropriate adjustments to the Current Population Survey to develop more accurate State-specific estimates of the number of children enrolled in health coverage.” As part of this evaluation, ASPE contracted with the Committee on National Statistics (CNSTAT) to produce a report: “Databases for Estimating Health Insurance Coverage for Children.” This report evaluated all data sources for health insurance coverage and concluded:
Several factors identified during the workshop appear to drive increased reliance on the ACS as the major survey source of estimates of children’s health insurance coverage, although there will continue to be reliance on the Current Population Survey (CPS) for some purposes. This trend will be based, in part, on the fact that the ACS and the CPS results are close for many characteristics and statuses, particularly in terms of family income and uninsurance. The CPS is generally acknowledged to provide better estimates of poverty due to the income questions. (National Research Council, 2010)
Finally, the production of the new Supplemental Poverty Measure requires data on health insurance coverage and health care costs in the CPS ASEC. Another CNSTAT report, funded by ASPE, is “Medical Care Economic Risk: Measuring Financial Vulnerability from Spending on Medical Care” (National Research Council, 2012). The report recommended that the Census Bureau continue to “use a definition of resources for the Supplemental Poverty Measure (SPM) and estimates of medical care economic burden that incorporates estimates of actual out-of-pocket spending on health insurance premiums and other out-of-pocket expenses for medical care.” The report further recommended “that the Census Bureau and the Agency for Healthcare Research and Quality assess the merits of adding items to both the Current Population Survey Annual Social and Economic Supplement and the Medical Expenditure Panel Survey to at least partially address the most critical data limitations identified for measuring medical care economic risk.”
The quality of the CPS income data is equally as important as the quality of the health insurance data. Furthermore, because income determines what type of health coverage one can afford, it is important to measure income and health insurance together. In the current CPS content test, a new method of collecting some types of income data is being tested as well as improving the calendar year estimate of health insurance.
Flaws in the current CPS questionnaire
Regarding health insurance, the CPS ASEC health insurance questions result in measurement error due to both the reference period and timing of data collection. The flaws in the current CPS ASEC health insurance estimates have been well established, as discussed in the Census Bureau’s annual publication on health insurance. The Census Bureau devotes two-thirds of a page in the Income, Poverty, and Health Insurance Coverage in the United States: 2011 (p. 21) to flaws in the estimate. The quality of health insurance data has long been a concern of HHS. These concerns relate to the full-year coverage/uninsured estimates, the Medicaid undercount, compatibility with other survey estimates and more generally misreporting of the source of insurance. In response to these concerns, HHS has provided support to the Census Bureau over the past 10 to 12 years to improve the health insurance questions in the Current Population Survey. Over this period, HHS has supported several rounds of cognitive testing and field-testing via the “Questionnaire Design Experimental Research Survey” (known as QDERS) -- a survey methods research vehicle carried out periodically from 1999-2006 and sponsored by the Center for Survey Measurement to improve question wording. This support, combined with additional Census Bureau work, led to the redesigned CPS ASEC health insurance questions tested in the 2010 Survey of Health Insurance and Program Participation. Thus, the CPS ASEC redesign is based on numerous studies, including focus groups, cognitive testing, behavior coding, pretests, split-ballot field tests and a record-check study. All evidence indicates the CPS ASEC redesign reduces measurement error compared to the current CPS ASEC. Please refer to the referenced papers in the bibliography for research on the development of the redesigned CPS ASEC (Attachment I). The aim of the March 2013 test is to operationally test these questions.
Full year estimates. The purpose of the health insurance questions on the CPS is to produce estimates of full-year health insurance coverage, and the existing instrument is administered in March and asks if the respondent had health insurance “at any time” during the last year. Cognitive testing strongly indicates that this method of asking about retrospective coverage is inadequate to generate reports of past coverage. Some respondents do not focus on the calendar year reference period, but rather report on their current insurance status. Quantitative studies have also shown that those with recent coverage are more likely to report accurately than those with coverage in the more distant past. As a result, it is not clear what the current CPS health insurance questions are measuring but it appears to be a mix of past and current coverage.
The redesign aims to get more accurate reporting of retrospective coverage by first asking if the respondent is currently insured, and then asking if that coverage began before or after January 1 of the past year. This method allows the respondent to anchor their report in the present and more clearly defines the time frame of interest, which helps respondents to provide a more accurate report of past-year coverage. Furthermore, by starting with current status, the possibility of backward telescoping is eliminated precisely because current coverage is in-scope.
A related aspect of improved reporting of retrospective coverage has to do with relatively short spells of coverage. The redesigned CPS is more likely than the existing CPS ASEC to capture such short spells because there are questions that directly ask about any specific months of non-coverage. There is a demonstrated precedent for the effectiveness of this type of verification. In the 2000 CPS ASEC, a verification question on health insurance decreased the uninsured rate by 1.2 percentage points.
Because the redesigned health insurance series begins with questions on current coverage and then asks about when that coverage began, the redesigned CPS will render data at the person-plan type-month level over the past calendar year. While this was a side-effect of the primary goal to improve retrospective reporting, the rendered data will enable analysis of transitions from one plan type to another as well as churning within a plan type over time.
Medicaid. The Medicaid undercount is a persistent problem in surveys. Several studies indicate that Medicaid enrollees are usually reported as having health coverage, but that they misclassify the type of insurance they have. To address this problem, the redesign starts with a general question on coverage status, and then uses a funneling approach to gradually obtain the necessary plan type detail. Respondents are first asked if their coverage is through a job, the government or state, or some other way. Next, respondents who say their coverage is through the government or state are asked if their coverage is through Medicaid, Medicare, military/VA or other. In this way, the redesign helps to ensure that Medicaid and Medicare coverage is correctly classified as government/state-related, and not as employer-sponsored or directly-purchased coverage. Furthermore, by providing both Medicaid and Medicare options at the same time, with on-screen definitions and significant interviewer training, the chance that Medicaid is misreported as another type of coverage is reduced. Another mechanism to prevent misreporting is a soft-check for respondents who report Medicare but who are under 65 and not disabled. The soft-check describes Medicare and Medicaid again and gives the respondent a chance to change their answer. Finally, respondents who answer “don’t know” or “refused” to the series of questions on plan type are asked: “What do you call the program?” At this point, a full list of all known government plans and state-specific program names is displayed.
New content for the CPS ASEC
Exchanges. One of the key features of the ACA is the implementation of state-specific “Health Insurance Exchanges” in 2014. An exchange is a state-level marketplace of private health insurance options for individuals and small businesses. Exchanges are still in development in most states, and states have broad flexibility in designing the programs. Nevertheless, it is essential for the federal government to have a viable methodology in place by 2014 to measure exchange participation and premium subsidization within the exchange.
The redesigned CPS includes questions that capture coverage by the conventional set of sources (employer-sponsored, direct-purchase, Medicaid, etc.) but for relevant conventional sources there is a short set of follow-up questions to determine whether the coverage was obtained through the exchange and, if so, whether the premium was subsidized. The overall approach and the exchange-specific questions were extensively tested through focus groups and cognitive interviewing and found to be unproblematic. As a result, regardless of how each state designs its exchange, the Census Bureau believes that the redesigned health insurance module should be able to accommodate measurement of exchange participation while preserving the meaning of the preexisting type of health coverage plans.
While the exchange-specific questions were developed and successfully tested in Massachusetts (one of only two states that currently have an operating exchange), most states have yet to finalize the design of their exchange programs and subsidization schemes. The Census Bureau fully expects that additional testing may be necessary as states implement their version of the exchange. The Census Bureau sees the development of the wording for the exchange question as an iterative process and will work with HHS to adapt the wording (within the new module context) as needed over the next few years. In addition to the unknowns about how the exchange will be implemented in each state, it is also unclear how other federal surveys will go about capturing exchange participation and subsidization. However, evidence from Massachusetts strongly suggests that adding the names of exchange plans to an existing list of plan types is flawed.
The 2013 content test provides opportunity to field test the exchange-specific questions in Massachusetts and Utah, the two states that currently have functional exchanges. In these two states, the redesign will include questions about health insurance exchanges.
Employer-sponsored insurance take-up. The ACA may lead to changes in the rates of employer-sponsored health insurance offers and take-up. Current health insurance status information, such as is gathered in the redesigned CPS ASEC, is required to measure these changes. New questions were added to measure employer-sponsored insurance offers and take-up. Specifically, the redesigned series asks questions about whether an employer offers the employee health insurance, whether the employee is eligible (and if not, why not), and if eligible but declined, the reasons for non-take-up. This particular set of questions only works with a current coverage question.
Although this set of questions is new to the CPS ASEC, it has been in CPS production in the Contingent Worker Supplement (CWS), fielded in February of 1995, 1997, 1999, 2001 and 2005 and the Council of Economic Advisors has expressed strong interest in including them in the CPS ASEC.
Please see attachments D, E, F, and H for a description of the redesigned health insurance instrument.
Changes to Income Questions
In 2011, Westat Inc., with assistance from the Urban Institute, was contracted to evaluate the CPS ASEC instrument. In April of that year, they issued a report entitled “Cognitive Testing of Potential Changes to the Annual Social and Economic Supplement of the Current Population Survey” (see Attachment H). They documented potential changes to income questions to correct income errors of varying magnitudes across income types. They also conducted cognitive interviews for the new income questions in particular subpopulations.
Most of the Westat recommendations have been implemented, resulting in several changes to the income portion of the CPS ASEC instrument. The most significant of the changes is the tailoring of the order of income questions to match those sources most likely received by respondents given certain known characteristics of the respondent household. There will be three distinct paths: households with a householder aged 62 and older, low-income households, and a default path for all other households. For example, householders aged 62 and over are presented disability, Social Security, Veteran’s benefits, and retirement account questions up-front (after job earnings questions), while low-income households are asked about public assistance, food stamps, and public housing after job earnings questions.
Some of the other suggestions to be implemented include a dual-pass approach. This means all income sources received are identified first and then the instrument will proceed to ask amounts for only those sources that the respondent indicated receiving. For responses of “Don’t know” or “refuse”, the instrument will present unfolding income range brackets for the respondent to choose. The family income screener for determining which household to ask low-income questions (such as food stamps, public assistance, energy assistance, etc.) has been removed, meaning every household will be asked questions on all possible sources of income. The sections on disability, survivor benefits, and retirement income have been updated to reflect current conditions and to clarify some concepts that were determined to be confusing through cognitive testing. Lastly, for changes to the income section, there are some new questions that deal with: the collection of larger initial checks from certain disability benefits due to processing delays, withdrawals and distributions from retirement accounts, and capital gains received from stocks and mutual funds.
Question 2. Needs and Uses
As discussed above, the primary purpose of the field study is to operationally test the redesigned survey instrument. In particular, we will assess:
1. How does the new instrument perform in the production environment?
2. Does the instrument perform as quickly as before?
3. Is the instrument able to capture and store all necessary information?
4. Does it affect instrument functioning?
In addition, ASPE is supporting an expansion of the sample for the 2013 operations field test of new health insurance and income questions to ensure a sample of sufficient size to characterize the nature of any break in trend. The goal, shared by the Council of Economic Advisors (CEA), is to measure the impact of this break.
This analysis will be important because while the 2014 survey would still contain questions about insurance coverage during the prior calendar year (2013) several aspects of the income and health insurance questions would change. Thus, if the new questions are added to the 2014 production run of the CPS ASEC, it would create a potential break-in-trend for the historical health insurance data from the CPS ASEC, which would in turn complicate HHS’s ability to attribute changes in insurance coverage to the onset of major insurance expansions under the ACA in January 2014. A memo jointly prepared by ASPE and CEA (Attachment K) provides the statistical support for making such a comparison.
Finally, the Census Bureau and HHS will compare the data from the redesigned questions to the 2013 production CPS ASEC in order to inform decisions regarding whether or not to incorporate the new measure into the 2014 production run of the CPS ASEC. Comparisons of these datasets will be carried out by staff in the Social, Economic and Housing Statistics Division (SEHSD) and the Center for Survey Measurement and will involve a range of different methods to address the following research questions:
Reform
Using the new instrument, what are the estimates of the uninsured?
Income
Health Insurance
How do the test estimates of insurance coverage compare to the production estimates?
How do the test estimates of public insurance coverage compare to the production estimates?
How do the test estimates of private insurance coverage compare to the production estimates?
How does the test compare to the production instrument in terms of item non-response rates for types of health insurance?
How does the offer and take-up of employer-provided health benefits compare to the Survey of Income and Program Participation and the Medical Expenditure Panel Survey (MEPS)?
How do the Health Insurance Exchange participation estimates in Massachusetts compare to the administrative totals in Massachusetts?
How do the test estimates of insurance coverage compare to estimates from other surveys, such as the National Health Interview Survey (NHIS), the Medical Expenditure Panel Survey (MEPS), the SIPP and the American Community Survey (ACS).
Information quality is an integral part of the pre-dissemination review of the information disseminated by the Census Bureau (fully described in the Census Bureau's Information Quality Guidelines). Information quality is also integral to the information collections conducted by the Census Bureau and is incorporated into the clearance process required by the Paperwork Reduction Act.
Pending the results of the operational test, the Census Bureau plans to implement the redesigned health insurance and income modules in the 2014 CPS ASEC. Please see Attachments D, E, F, and H for a description of the redesigned health insurance instrument, and Attachment G for a detailed description of the redesigned income module.
Question 3. Use of Information Technology
All interviews will be conducted using a Computer-Assisted Telephone Interviewing (CATI) instrument and data will be transmitted electronically from the telephone facilities to Census Bureau headquarters in Suitland. In general, CATI instruments offer smooth, efficient administration of questionnaires, since the sequencing of questions is handled behind-the-scenes by the program, not by the interviewer. It has been more than 30 years since the last major redesign of the income questions of this questionnaire (1980), and the need to modernize this survey to take advantage of CAI technologies has become more and more apparent. With regard to both income and health insurance, the CPS ASEC was converted to Computer Assisted Interviewing (CAI) in 1994. This conversion essentially took the questions and skip patterns of the paper questionnaire and put them on a computer screen. Automated data collection methods allow for complicated skips, respondent-specific question wording, and utilizing the data gathered about one household member to reduce respondent burden when describing another household member. In other words, it is now possible to carry-over data from one household member to the next.
The test of the new income questions takes advantage of technology because it is exploiting CATI to get better data and reduce burden. It will permit the use of several built-in editing features, including automatic checks for internal consistency and unlikely responses, and verification of answers. It will allow for custom tailoring of income question order to match the demographic make-up of the household and allow for dependent interviewing.
Automation is also heavily exploited in the health insurance section, reducing tedium, respondent fatigue and burden. The hybrid household-person-level design takes full advantage of the fact that in many households all or most members share the same plan type, a feature of household-level designs, but also preserves the advantage of person-level designs by asking about each household member by name. Specifically, as soon as one plan type is identified for a specific household member, questions are asked to determine if other household members share that same plan type. This information is stored and tracked so that when those other household members are asked about, the previously-reported plan can simply be verified and then questions about any additional coverage are asked. A full battery of person-level questions is not needed in many cases, which reduces burden significantly.
The computerized questionnaire also permits the inclusion of several built-in editing features, including automatic checks for internal consistency and unlikely responses, and verification of answers (such as the soft-check for Medicaid/Medicare described above). These built-in editing features can catch and correct errors during the interview itself, as opposed to relying on post-collection edits.
Question 4. Efforts to Identify Duplication
The CPS ASEC serves as the most widely-cited source of estimates on health insurance and the uninsured, in part because it is helpful to analyze income and health insurance together. There are many other surveys, from both federal and non-federal agencies, that collect data on health insurance coverage. The Department of Health and Human Services’ National Health Interview Survey (NHIS) and Medical Expenditure Panel Survey (MEPS), and three surveys conducted by the Census Bureau, including the CPS ASEC, the Survey of Income and Program Participation (SIPP) and the American Community Survey (ACS), currently provide data on health insurance coverage. Each of these surveys has a different history, purpose, methodology and time frame, and thus produce slightly different estimates. For instance, before the ACS, the CPS was the only source for reliable state estimates of income, poverty, and health insurance. Currently, the NHIS produces state-level estimates for most states and some large metropolitan areas. The surveys differ on what time period the health insurance questions are related to. For example, the ACS collects data on point-in-time coverage but not coverage throughout the previous calendar year.
Furthermore, because these surveys provide richly detailed demographic and economic data as well as the source of insurance coverage estimates, each is useful for explaining different aspects of the variation in coverage within the population. For instance, some analysts find the CPS health insurance estimates are preferable to the ACS estimates because the CPS collects information on income and dependents. Other researchers may prefer the health insurance information collected in the NHIS and the Household Component of the MEPS (conducted by HHS’s Agency for Health Care Quality Research), because these surveys combine information on insurance coverage with detailed data on medical care utilization and health status. Another component of the MEPS surveys employers on the premiums and other features of the insurance coverage they offer employees. Additional information on utilization comes from HHS surveys of health care providers, including office-based physicians, ambulatory care facilities and hospitals.
As stated above, all these surveys employ different methodologies, but many share common general features. For example, the CPS redesign and the NHIS question series both begin with a global question on current coverage status (covered or not), the core wording of which is quite similar. Furthermore, both surveys follow up a “no” response with several questions on plan types typically forgotten or underreported, such as Medicaid and Medicare. The Census Bureau believes the CPS redesign and the NHIS will produce estimates of the uninsured at a point in time that are very similar. Where the surveys depart is in the questions on plan type (discussed more below). Although no head-to-head test has been conducted between the CPS redesign and all the surveys mentioned above (though a head-to-head test was conducted between the CPS, ACS and the CPS redesign), the Census Bureau believes the redesign will likely be an improvement over the existing ACS or NHIS battery. Most of the reasons for thinking so have been spelled out in a report on cognitive testing of the NHIS (Beatty et. al. 20021). The authors of that report concluded: “We recommend simplifying the categories to a more manageable few. The categories listed above [private; military; government; single-service; other; no coverage] seem to constitute the major distinctions people make (although the specific wordings may require tweaking); further differentiation within the categories should probably be relegated to a subsequent question or questions” (Beatty et. al. 2002: 5). Findings from this and other reports guided the development of the CPS redesign in its earliest stages.
In terms of the method of capturing plan type, most existing surveys employ either a “type by type” approach – whereby respondents are asked a long series of yes/no questions on very specific plan types – or a “status/type” approach – whereby respondents are asked the global yes/no question on coverage, followed by a single question on what kind of plan, which includes a “laundry list” of detailed plan types. Both approaches have been found to be problematic in non-trivial ways – prompting double-reporting of the same plan, misreporting of one plan type as a different plan type, under-reporting plan type altogether, and confusing Medicare with Medicaid. Based on prior studies and a literature review, the CPS redesign adopts the global yes/no question on coverage but addresses the plan type issue by using a funneling approach, whereby a very general question on source of coverage (job, government/state, other) is followed by more specific questions, tailored to the general source. The result is that no one question contains too much detail and each individual question is more appropriate to the respondent’s situation. The series as a whole has been found in numerous tests to be comprehensible and easy for the respondent to understand and provide an accurate answer.
Though the questions on the take-up of employer-sponsored insurance are duplicative of other surveys’ content (the MEPS, SIPP and the CPS Contingent Worker Supplement include them), the CPS CWS has not been fielded since 2005, and the SIPP will no longer ask about this topic after the redesign of SIPP in 2014. The CPS ASEC is the only survey that has produced state-level estimates, and so including ESI take-up questions in the CPS ASEC would fulfill a unique need.
Question 5. Minimizing Burden
Small businesses or other small entities are not asked to report information.
The new CPS health insurance instrument was designed to reduce respondent burden. While it is true that the instrument specifications are complex, the question set when administered to a typical respondent is simple, concise and easily understood (see scripted scenarios in Attachment M for a range of examples). (a) With regard to time frame, first current coverage is established (which is the easiest time frame for respondents), and then only start/end dates are asked – not a lengthy set of questions asking about month-by-month coverage. (b) To identify plan type, first, the general source is established and then only detailed questions relevant to each general source are asked. Thus, not all of the respondents are asked detailed questions about plan types that are clearly irrelevant to them. For example, those with private coverage are asked if the plan is related to military service at all and, if not, the question on Tricare, CHAMPUS, CHAMP-VA, etc. is not asked. (c) With regard to the person/household-level approach, the questionnaire reduces burden significantly compared to the ACS and other person-level surveys by not repeating the entire battery of questions on plan types for each person. In the redesign, once any type of plan is established for a household member, questions are asked to determine what other household members are also covered by that plan type. Therefore, any plan that is shared among household members only needs to be reported once. For any household members who were reported as covered by a plan reported earlier in the series, the questionnaire only asks two questions about any additional coverage they may have. If there is none, no more questions are asked. This serves to reduce burden significantly, especially in large households, and it maintains the advantage of asking about each household member by name.
Question 6. Consequences of Less Frequent Collection
The redesign of the income and health insurance questions is aimed at testing new content and reducing measurement error in the CPS ASEC. It is likely that the Census Bureau will decide that there is a break in series for the health insurance estimates. The Affordable Care Act will go into effect in 2014. The 2014 CPS ASEC will report health insurance status for the calendar year 2013. This is important because the 2014 CPS ASEC will serve as the baseline for health reform, as described above.2 If the changes in the health insurance questions occur after 2014, there will not be a baseline measurement using the new instrument.
Question 7. Special Circumstances
There are no special circumstances.
Question 8. Consultations Outside the Agency
For more than a decade Census Bureau staff have been collaborating and communicating with individuals outside the bureau who have been closely involved in the technical matters of health insurance measurement. These individuals include Don Oellerich and Ben Sommers (at the U.S. Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation), Rob Stewart (formerly at the U.S. Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation and now at the Congressional Budget Office), Michael Davern (of the University of Chicago and the National Opinion Research Center), Kathleen Call (of the State Health Access Data Center), Linda Bilheimer and Diane Makuc (of the National Center for Health Statistics), and Steve Hill (at the Agency for Healthcare Research and Quality). Efforts have also been made to both inform and solicit comments from the research community on research findings and plans for future tests through a number of publications, conferences and seminars.
For the income module, the Census Bureau consulted with Westat Inc. in 2011 as well as Mathematica Policy Research.
The Census Bureau published a notice on July 5, 2012 (77 FR page 39678-9) announcing our intention to submit this test for approval. We received no official comments on this notice.
Question 9. Paying Respondents
This study will not involve any payments to respondents.
Question 10. Assurance of Confidentiality
Respondents are informed through an advance letter and in the survey introduction that the survey: (1) is being conducted under the authority of Title 13, Unites States Code, Section 182; (2) has been approved by the OMB under project number 0607-xxxx; (3) takes an average of 40 minutes per household to complete; and (4) is voluntary. Respondents are also informed that the Census Bureau is required under Section 9 of Title 13 to keep their information confidential and use it for statistical purposes only. The advance letter also solicits comments from respondents and provides an address and email address for sending these comments. It also states that the OMB number legally certifies the information collection. Please see attachments B and C.
Question 11. Justification for Sensitive Questions
No sensitive questions are asked in this study.
Question 12. Estimate of Hour Burden
The 2013 CPS ASEC content test will be conducted only one time, by telephone, with retired CPS sample, with a goal of 15,000 completed household interviews. Typically, a single respondent reports for the entire household. The interview is expected to take 40 minutes per household on average, resulting in 10,000 total annual burden hours.
Compared to the production CPS, the 2010 Survey of Health Insurance and Program Participation battery (i.e., the point-in-time as well as calendar year) takes 66 seconds longer (median) and 91 seconds longer (mean). However, this difference is expected to decrease after interviewers become familiar with the new instrument.3
Question 13. Estimate of Cost Burden
There are no costs to respondents other than that of their time to respond.
Question 14. Cost to Federal Government
The total cost estimate for the production of the CATI instrument, testing, and all phases of preparation and administration of data collection is $1,474960. The Census Bureau will cover the cost of $1,399,960, and the Assistant Secretary for Planning and Evaluation will cover the remaining $75,000 for the larger sample size.
Question 15. Reason for Change in Burden
The increase in burden is attributable to the information collection being submitted as new.
Question 16. Project Schedule
General research leading up to this study has been ongoing. Preliminary planning for this particular field test began in December 2011. Below is a general schedule.
2012
January-May: (1) Prepare general project plan
(2) Develop detailed specifications for CATI instrument.
(3) Develop detailed post-processing specifications.
May: Hold kick-off meeting among inter-divisional Census Bureau staff; develop schedule; assign tasks.
June - December: Work with staff from the Demographic Surveys Division and the Technology Management Office, who will author the instrument. Conduct first formal test of CATI instrument (“user’s test”).
2013
January: (1) Develop specifications for sample selection.
(2) Develop advance letter and procedures
(3) Conduct formal test of CATI instrument (“systems test”)
(4) Develop training materials
February: (1) Select and prepare sample.
(2) Conduct third formal test of CATI instrument (“verifications test”)
March: (1) Mail advance letters
(2) Conduct training
March - May (3) Conduct training and data collection
June - August Analysis and preparation of reports for internal use, research conferences and publications.
Question 17. Request to Not Display Expiration Date
The expiration date will be contained in the advance letter sent to respondents.
Question 18. Exceptions to the Certification
There are no exceptions to the Certification for Paperwork Reduction Act Submissions.
1 Beatty, Paul, B. Wilson, K. Miller, A. Calvillo, and K. Whitaker. 2002. “NHIS Insurance Module and additional questions on Work, Income, and Health Care Utilization” Unpublished report, April, 2002
2 Explicitly, the uninsured rate will be used as a factor in calculating payments made by Medicare to hospitals that treat a disproportionate share of low-income patients (H. R. 4872—27). ACA does not specify which Census Bureau data to use.
3 With regard to length, from the SHIPP 2010 paper: “The average SHIPP interviewer had 8.39 years of field interviewing experience at the Census Bureau, and 5.41 years of experience on surveys that include questions about health insurance (the CPS being among the most common of these). The EXP field period, by contrast, lasted only ten days and training lasted two days.” With a more extended training and learning curve, it is likely that the gap between the current and redesigned CPS would shorten.
| File Type | application/vnd.openxmlformats-officedocument.wordprocessingml.document |
| Author | pasca001 |
| File Modified | 0000-00-00 |
| File Created | 2021-01-30 |