A Roadmap for Disclosure Avoidance in the Survey of Income and Program Participation (2024) / Chapter Skim
Currently Skimming:
9 Maintaining Usability While Preserving Confidentiality: Potential Strategies
9 Maintaining Usability While Preserving Confidentiality: Potential Strategies
Pages 143-170
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 143... ...
Finally, the panel discusses Title 13 and how its interpretation may be modernized to improve data access and usability. As the Census Bureau evaluates potential privacy solutions and develops new modes of access to SIPP data, it will be important to evaluate those solutions in terms of their potential impact on the three dimensions of usability discussed in Chapter 4: accuracy, feasibility, and accessibility.
|
From page 144... ...
, a secure virtual environment where approved users can access stand-alone SIPP data and obtain results following data disclosure review or infused with noise and correct inferences, but with lower start-up costs than when work ing through a Federal Statistical Research Data Center (FSRDC; see Chapter 5)
|
From page 145... ...
A released dataset is effectively unusable if, by comparison with the original dataset, the confidence intervals widen so much and power is so greatly reduced that no reliable inferences can be made. For any particular statistics obtainable from the original dataset, whether proper inferences can be made and validity can be obtained will depend on the mode of access of the data as well as the information in the released data.
|
From page 146... ...
In the sections below, the panel reviews the various ways SIPP has been used, aiming to illustrate the wide range of uses to consider when assessing the different dimensions of statistical usability from data analysis system methods. Once assessments are made of the validity and reliability of SIPP-based estimates for a given data product, it will be important for the Census Bureau to clearly communicate its findings to the SIPP user community.
|
From page 147... ...
Use 1: Analyses Relying on Unique SIPP Content SIPP provides uniquely detailed information about income and public program use, income (including poverty, hardship, and wealth) , detailed relationships between all household members, and life events (including marriage, cohabitation, fertility, schooling, work, and migration)
|
From page 148... ...
148 FIGURE 9-1 Uses of SIPP data in the most cited and recent studies (percentage)
|
From page 149... ...
FIGURE 9-2 Unique SIPP content used in the most cited and recent studies (percentage)
|
From page 150... ...
Use 2: Longitudinal Analysis Another core strength of SIPP is its ability to monitor change. The longitudinal design of SIPP enables users to examine short-term change for individuals, families, and households for the various indicators noted above, including stability and instability in outcomes like employment, income, program participation, household composition, and family structure.
|
From page 151... ...
However, the detail of longitudinal data, particularly dates of events (e.g., births, marriages) , may pose disclosure risks.
|
From page 152... ...
, and it is unclear whether synthesized data could account for complex interactions of multiple variables, particularly if measures are constructed with outside information. These findings are consistent with results from the panel's call for information from SIPP users.
|
From page 153... ...
Use 5: Analyses Relying on Administrative Record Linkages Finally, SIPP users have linked SIPP data with administrative records from the Social Security Administration and other agencies to supplement information collected in SIPP. This research tends to be highly innovative, yielding new findings that would not be possible with SIPP data alone, such as work on income (Meyer et al., 2021; Medalia et al., 2019)
|
From page 154... ...
As a consequence, these codes may be suppressed TABLE 9-1 Matrix for Evaluating Feasibility with the Context of Various Modes of Access Mode of Access Online analysis Public-use Synthetic Uses builder microdata data SODA FSRDC 1. Analyses relying on unique SIPP content 2. Longitudinal analysis 3. Analysis relying on granular data and complex recodes 4. Causal effects of public policies 5. Analysis relying on administrative record linkage
|
From page 155... ...
As accessibility declines, the number of SIPP users will also likely decline, particularly among those with fewer resources, reducing the potential impact of SIPP. An important task facing the Census Bureau is to determine how users would be impacted by making particular
|
From page 156... ...
The survey of SIPP publications described earlier suggests that the bulk of SIPP users probably fall in the second and third categories above -- that is, those with the greatest resources available. Nearly all the first authors were faculty members at universities or researchers in
|
From page 157... ...
Moreover, it is critical to note that proximity to an FSRDC does not fully capture the barriers to working in an FSRDC. Regardless of their proximity to an FSRDC, most published data analyses that the panel surveyed were based on public-use data, and only 4 percent required access to restricted data.
|
From page 158... ...
This group would stand to lose the most if public data were no longer available. At the other end of the scale, high-resourced users could probably access data regardless of access mode, although making SIPP data available only in a highly restrictive environment would slow down research or steer researchers away from SIPP and toward other data sources, such as the Panel Study of Income Dynamics (PSID)
|
From page 159... ...
For researchers who wish to use the restricted SIPP data, obtaining approval through the current FSRDC system requires a multistage process that can take several months. The proposal must demonstrate that the research project's predominant purpose is to benefit Census Bureau programs, and this criterion carries the greatest weight during the approval process over other considerations, including scientific merit, need for restricted data, and feasibility.
|
From page 160... ...
This is illustrated in Figure 9-5, where the barriers to access that would be mitigated if data were provided via SODA FIGURE 9-5 Determinants and barriers to accessing restricted Census Bureau data through the current FSRDC system. NOTE: Green text indicates barriers to access that would be mitigated through a reinterpretation of Title 13, and blue text indicates barriers that would be mitigated through SODA.
|
From page 161... ...
Second, to the degree that the public-use SIPP microdata become less useful due to efforts to increase privacy protections, the Census Bureau should consider providing mid-tiered access to restricted SIPP data through SODA. Current SIPP users tend to be sophisticated data users.
|
From page 162... ...
Title 13 defines the role and functions of the Census Bureau, and it contains provisions designed to both direct and authorize the Census Bureau to protect the confidentiality of its data. The title's provisions are thus very important when considering disclosure avoidance for SIPP.
|
From page 163... ...
The Census Bureau once treated that definition as applied to information such as names and addresses but has been broadening that definition. • It states that the data should be used for the statistical purposes for which they are supplied.
|
From page 164... ...
It sets three criteria for determining if a project qualifies, three individual/organization criteria for determining who may receive Special Sworn Status, and four criteria for determining if a project can take place at a non-Census Bureau facility. In sum, the project must require access to Census Bureau confidential data, benefit the Census Bureau's Title 13 Programs, and be viable.
|
From page 165... ...
. External projects must be carried out at a Census Bureau facility (including FSRDCs)
|
From page 166... ...
The Census Bureau might consider modernizing its interpretation of Title 13 in several ways. • First, Title 13 does not require that data access benefit the Census Bureau specifically, instead focusing on accomplishing the work of the Census Bureau.
|
From page 167... ...
, which applies to all federal statistical agencies. Thus, expanding data access for evi dence building might also be considered as a benefit to the Census Bureau.
|
From page 168... ...
• Most SIPP data users' needs cannot be met by providing stan dardized tables, while accessing SIPP data only through a Federal Statistical Research Data Center would be overly burdensome for many SIPP data users and create inequities in access. • An online tabular/analysis builder for accessing SIPP data has some potential for expanding the SIPP user community to include those who may not have the analytical skills or resources to work with the public-use microdata files.
|
From page 169... ...
data products treated for disclosure protection and communicate the results to SIPP users. Recommendation 9-2: When considering which access modes to pri oritize, the Census Bureau should evaluate feasibility for the most common Survey of Income and Program Participation (SIPP)
|
Key Terms
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.