SIPP Synthetic Beta v5.1 SSBv51 TBD Virtual RDC Cornell NSF-Census Research Network Cornell NCRN Project June 18, 2014 Cornell Institute for Social and Economic Research (CISER), Cornell University, Ithaca NY CED2AR, Version 1.0 The Comprehensive Extensible Data Documentation and Access Repository 2.5 National Science Foundation (NSF) 1131848 NSF-Census Research Network - Cornell node Cornell Institute for Social and Economic Research Labor Dynamics Institute 2014 2014-06-18 Comprehensive Extensible Data Documentation and Access Repository. Codebook for the SIPP Synthetic Beta 5.1 [Codebook file]. Cornell Institute for Social and Economic Research and Labor Dynamics Institute [distributor]. Cornell University, Ithaca, NY, 2014 SSB_codebook_v5-1-beta SSBv51 Test SSB_codebook_v5-1-beta 5.1 United States Department of Commerce. Bureau of the Census. Social Security Administration. Internal Revenue Service. Cornell University. Labor Dynamics Institute. United States Department of Commerce. Bureau of the Census. none Washington, DC, USA National Science Foundation 1042181 Labor Dynamics Institute United States Department of Commerce. Bureau of the Census. 1 August 2013 5.1 U.S. Census Bureau. SIPP Synthetic Beta: Version 5.1 [Computer file]. Washington DC; Cornell University, Synthetic Data Server [distributor], Ithaca, NY, 2013 SIPP The SIPP Synthetic Beta (SSB) is a Census Bureau product that integrates person-level micro-data from a household survey with administrative tax and benefit data. These data link respondents from the Survey of Income and Program Participation (SIPP) to Social Security Administration (SSA)/Internal Revenue Service (IRS) Form W-2 records and SSA records of retirement and disability benefit receipt, and were produced by Census Bureau staff economists and statisticians in collaboration with researchers at Cornell University, the SSA and the IRS. The purpose of the SSB is to provide access to linked data that are usually not publicly available due to confidentiality concerns. To overcome these concerns, Census has synthesized, or modeled, all the variables in a way that changes the record of each individual in a manner designed to preserve the underlying covariate relationships between the variables. The only variables that were not altered by the synthesis process and still contain their original values are gender and a link to the first reported marital partner in the survey. Seven SIPP panels (1990, 1991, 1992, 1993, 1996, 2001, 2004) form the basis for the SSB, with a large subset of variables available across all the panels selected for inclusion and harmonization across the years. Administrative data were added and some editing was done to correct for logical inconsistencies in the IRS/SSA earnings and benefits data. 1951-2007 1990-2004 United States of America State Level Person Civilian non-institutionalized population living in the United States mixed survey and administrative record data (partially synthetic) U.S. Census Bureau SIPP Panels (1990, 1991, 1992, 1993, 1996, 2001, and 2004) SSA records of retirement and disability benefit receipt (1984-2007) IRS Form W-2 records on respondents (1951-2006)

Creation of the SSB:

The SSB is created from several data sources. The survey data are drawn from multiple panels of the Survey of Income and Program Participation (SIPP): the 1990, 1991, 1992, 1993, 1996, and 2004 panels. The administrative data are drawn from the following SSA files: the Master Earnings File, the Master Beneficiary Records (MBR), the Supplemental Security Records (SSR), the 831 Disability File (F831), and the Payment History Update System (PHUS).

The creation of the SSB begins with the construction of the Gold Standard File (GSF). To construct the GSF, a set of variables from the 1990-2004 SIPP panels are standardized to produce consistent measures across panels. The SIPP respondent identifiers are mapped to Social Security Numbers (SSB) using the Census Bureau's erson Information Validation System (PVS). Using the list of SSN's for the sample is SIPP respondents, SSA creates Summary Earnings Records (SER) and Detailed Earnings Record (DER) extracts from the Master Earnings File. SSA also creates extracts from the four benefit files (MBR, SSR, F831, and PHUS) from the corresponding master files. Using the mapping between the SIPP identifiers and SSN's, Census then links these extracts to the SIPP data. The GSF consists of person-level research variables created from these linked data.

The next step in the creation of the SSB is to impute missing values in the GSF multiple times. This process results in four files (implicates) referred to as the Completed Data implicates. Each of these implicates contains original GSF values where non-missing and imputed values where the original value is missing. The imputations across Completed Data implicates are independent of each other.

The Completed Data implicates from the basis of the data synthesis that produces the SSB files. From each Completed Data file, four synthetic datasets are created by synthesizing variables conditional on the values in the Completed Data file. Thus, the SSB consists of sixteen files (implicates). All but the following data are synthesized in the SSB implicates: gender, OASDI benefit type, and spouse link (specific variables described in the data items section below). Detailed documentation of the process of data synthesis is available in the publication "Final Report to the Social Security Administration on the SIPP/SSA/IRS Public Use File Project" athttp://www2.vrdc.cornell.edu/news/wp-content/papercite-data/pdf/ssafinal.pdf.

The Completed Data and SSB implicates need not all have the same number of records. In order to be included in a Completed Data or SSB implicate, an individual's (possibly imputed or synthesized) age must be at least fifteen years as of January 1 in the first year of his or her SIPP panel. The interaction between this restriction and the variation in imputed and synthesized ages across implicates causes the exclusion of a slightly different set of individuals from each Completed Data and SSB implicate.

Researchers interested in using the SSB can submit an application to the Census Bureau. The application form and instructions can be downloaded fromhttp://www.census.gov/programs-surveys/sipp/methodology/sipp-synthetic-beta-data-product.html. Applications will be judged solely of feasibility of the proposed project (i.e., that the necessary variables are available on the SSB). Once an application has been accepted, the new user will be given an account on a server where the data can be accessed and analyzed. The SSB files have been cleared by the Census Bureau Disclosure Review Board, SSA, and IRS for use by individuals without Census Bureau Special Sworn Status and outside of Census Bureau facilities. The data can only be used on the VirtualRDC Synthetic Data Server at Cornell University. While no SSB data downloads are permitted at this time, users do not have to operate behind the Census Bureau firewall to access this server. sehsd.synthetic.data.use.list@census.gov

We request that researchers who publish results from analyses done using these data cite the SSB as their data source and acknowledge the use of the SDS server at Cornell and the support of Census staff in running any validation programs. These citations will help ensure continued funding for the SDS server and the creation of the Gold Standard File and the SSB.

Suggested acknowledgement:

This analysis was first performed using the SIPP Synthetic Beta (SSB) on the Synthetic Data Server housed at Cornell University which is funded by NSF Grant #SES-1042181. These data are public use and may be accessed by researchers outside secure Census facilities. For more information, visit http://www.census.gov/sipp/synth_data.html. Final results for this paper were obtained from a validation analysis conducted by Census Bureau staff using the SIPP Completed Gold Standard Files and the programs written by this author and originally run on the SSB. The validation analysis does not imply endorsement by the Census Bureau of any methods, results, opinions, or views presented in this paper.
The data synthesis process employed by Census to protect the linked data from the risk of disclosing the identity of individuals is relatively new and substantially changes both the survey and administrative data. The intent of the modeling done as part of the synthesis is to preserve relationships among variables that are of interest to researchers while ensuring that personally identifiable information is not revealed to the data user. It has not been feasible to ensure accuracy by comparing every relationship among SSB variables with the corresponding relationship in the underlying confidential micro-data. Hence, we strongly urge researchers not to publish results produced from the SSB without first requesting that Census validate these results with confidential data housed in a secure environment at the Census Bureau. Census will perform this validation free of charge to researchers, as resources permit and according to the protocol established by the three agencies involved and outlined below. Without validation of results, Census, SSA, and IRS make no guarantee of the validity of the SSB for any research purpose. Seehttp://www.census.gov/programs-surveys/sipp/methodology/sipp-synthetic-beta-data-product.htmlfor validation conditions.

Using SSB:

The GSF and Completed Data implicates contain personally identifiable information protected by Titles 13, 26, and 42 and cannot be accessed without Census Bureau Special Sworn Status nor outside of Census Bureau facilities. The SSB files, however, have been cleared by the Census Bureau Disclosure Review Board, SSA, and IRS for use by individuals without Census Bureau Special Sworn Status and outside of Census Bureau facilities.

Researchers interested in using the SSB can submit an application to the Census Bureau. The application form and instructions can be downloaded fromhttp://www.census.gov/programs-surveys/sipp/methodology/sipp-synthetic-beta-data-product.html. Applications will be judged solely on feasability of the proposed project (i.e., that the necessary variables are available on the SSB). Once an application has been accepted, the new user will be given an account on a server where the data can be accessed and analyzed. While no SSB data downloads are permitted at this time, users do not have to operate behind the Census Bureau firewall to access this server.

The SSB is designed to be analytically valid in that sense that point estimates should be unbiased and estimated variances should lead to inferences similar to those that would be drawn from an identical analysis on the Completed Data implicates. Initial tests of analytic validity of the SSB have been promising. All SSB users are invited to help further test the analytic validity of the SSB by submitting programs used to analyze the SSB to be run on the Completed Data and/or Gold Standard files. Users need only inform Census Bureau staff of the location on the server of such programs and work with Census Bureau staff to ensure that the programs run without error. Census Bureau staff will run the programs on the confidential data and release to the user resulting output that are cleared for release by the Census Bureau Disclosure Review Board. In order to evaluate the effects of the data synthesis separate from the effect of imputing missing data, comparisons should be made between results from the SSB and the Completed Data. To evaluate the effects of missing data imputation, comparisons should be made between results from the Completed Data and the Gold Standard.

Codebook for v5.1 athttp://www.census.gov/content/dam/Census/programs-surveys/sipp/methodology/SSB_v5_1_Codebook.pdf When analyzing the SSB, users should account for the multiple imputation aspect of the SSB by averaging statistics of interests across all sixteen implicates. Variance measures should be created following the appropriate multiple imputation formulae as described in the documentUsing the SIPP Synthetic Beta for Analysis.

Protocol for Validation of Results:

Census will validate results obtained from the SSB on the internal, confidential version of these data (Completed Gold Standard Files). Users who wish to obtain validated results should follow the protocol outlined here. The restricted access site will provide SAS and Stata analysis software and a computing environment similar to the one used to analyze the confidential Completed Gold Standard data on Census Bureau internal computers. Researchers should follow the Census Bureau programming requirements described in SSB Validation Request Guidelines to ensure that the programs will successfully transfer to internal Census computers for validation. Researchers should plan to share their results and programs from the synthetic data analysis with Census, ORES/SSA and SOI/IRS. After programs have successfully run without error on the synthetic data, researchers may request that Census run these programs on the Completed Gold Standard Files. Only programs successfully run without error on the SDS will be eligible to be run on the confidential data by Census staff. Any programs that produce errors on the Completed Gold Standard Files will be returned to users for correction. Once an analysis has been repeated on the Completed Gold Standard File, the results will be reviewed by Census staff for disclosure concerns. Researchers should familiarize themselves with standard Census disclosure rules for outside projects (See theRDC Researcher Handbookhere) and should fill out the appropriate memo documenting the requested output (seeRDC Disclosure Request Memo). Data products and output approved by Census staff will be released to the users, ORES/SSA, and SOI/IRS. The validation process can be accomplished in as little as one week for simple results that are generated by clean code and have no disclosure issues. However if the code does not run properly, the sample sizes are too small, or the researcher does not accurately fill out the disclosure memo, the process can take much longer. Census makes no guarantee on the length of time between submission of programs and the release of results from the confidential data. For more information about the validation process, including advice on how to make the process go smoothly and quickly, please seeSSB Validation Request Guidelines.

ssb_v5_1_synthetic1_1.sas7bdat 1000 299 SAS ssb_v5_1_synthetic1_1.dta 1000 299 Stata Demographic Variables The variables in this section are all drawn from the SIPP and represent demographic information gathered by the survey at a specific point in time. Entries for individual variables describe the exact SIPP source variable and the reference point in time. Economic Variables The variables in this section are all drawn from the SIPP and represent economic information gathered by the survey at a specific point in time. Entries for individual variables describe the exact SIPP source variable and the reference point in time. Marital History Variables Marital history is presented as two arrays of 8 elements describing up to 4 marriages. This history retains at most 4 dates of origin and the associated 4 dates of dissolution (whether due to divorce or death), if applicable, along with the corresponding type of event (marriage, divorce, or widowhood) for each SIPP respondent. The wave 2 Marital History topical module provides the majority of this information for up to 3 marriages. If an individual had more than 3 marriages, no dates for those marriages between the second and most recent are collected during the topical module interview. For individuals who participate in the topical module, we supplement this information by searching for new marriages or the termination of an existing marriage utilizing our knowledge of monthly marital status covering the period of the panel beyond wave 2. We rely exclusively on the complete set of monthly marital status indicators for people who do not participate in the topical module. The two marital history arrays, one comprised of dates and the other of event types, are edited to ensure internal consistency for linked spouses. Missing event dates or reasons are acquired from the spouse who has provided this information during a SIPP interview. This means that in the case of a deceased individual, the surviving spouse's report of widowhood is transferred to their former spouse. Likewise, for respondents who leave the household due to divorce (and are no longer interviewed by the SIPP), the spouse remaining in the household supplies the details regarding marital dissolution for both. It is possible for beginning or ending date information to not identically match for the linked spouses. In this case, we evaluate whether topical module details were supplied. If both spouses participated in the topical module, the data are considered as likely to be reliable from either individual. We examine whether the first spouse's beginning date occurs before the previous marriage end date for either spouse while the second spouse's beginning date occurs after the previous marriage end date of both spouses. If so, then the second spouse's beginning date replaces the first spouse's beginning date. Alternatively, if the reverse is true, then the first spouse's beginning date replaces the second spouse's beginning date. When no obvious conflicts with the date of the start of the current marriage and the date of termination of the previous marriages of either spouse exist, then a random number is used to determine which spouse's information to retain. A similar algorithm is implemented to resolve issues relating to non-matching ending dates of the linked spouses (in this case, each spouse's subsequent marital start date is taken into consideration). When disagreements between the beginning or ending dates occur and only one of the two spouses participated in the topical module, the participating spouse's information is considered to be more reliable (provided that the adoption of the date presents no conflicts with previous or subsequent marital events for either spouse). In the absence of topical module participation for both spouses, a random number is used to determine which spouse's information to retain. Any persisting conflicts between dates of marital events are remedied by utilizing a random number to determine the most reasonable event date for the pair. As a final step, the cleaned file is carefully reviewed once more to ensure internal consistency. Benefits Variables Variables in this section are taken from the 831 Disability File (F831) that tracks all Title II (SSDI) and Title XVI (SSI) applications for disability payments from 1990 onwards. Using the F831 data, we created one record per person that reports number of filings, plus filing and decision dates, and the result of determination for the first, second, and most recent applications listed on the file. We created separate variables for different application types (Title II-SSDI/Title XVI-SSI) so in total there is information on a maximum of 6 applications. We retain records for primary applicants only. Differences may exist between the disability variables on the MBR and the F831 due to differing lengths of the time series, as well as to the fact that research version of the MBR used to create the Gold Standard recorded a maximum of two events (e.g., events occurring between the initial and most recent entitlement may be censored). MBR/PHUS Variables The Master Benefits Records (MBR) is SSA's main file to track who is receiving Old Age Survivor and Disability (OASDI) benefits, the reason for receipt, and the monthly benefit amounts payable to the individual. The Payment History Update System (PHUS) contains actual payments delivered to OASDI beneficiaries. The data from the PHUS may differ from what are contained on the MBR due to discrepancies between the timing of SSA awarded amounts and the actual payments made to participants. This situation would be expected to affect disability cases more than aged cases because it takes more time to establish eligibility to receive disability. Individuals are eligible to receive benefits due to their own earnings history and age, as well as due to a spouse's earnings history and age. In this section retirement and disability are "own" benefits while aged spouse, widowed spouse, and other are "spouse" benefits. The age requirements for receiving each type of benefit are as follows: Retire - minimum age 62 (reduced benefit), full retirement age (full benefit) Disability - under age 65 or full retirement age, whichever is greater; at full retirement age, these benefits convert to retirement. Aged Spouse - minimum age 62(reduced benefit), full retirement age (full benefit), spouse must be retired or disabled Widowed Spouse - minimum age 60(reduced benefit), full retirement age (full benefit), spouse must be deceased Other - no age requirements Until the year 2000, the full retirement age was 65. From 2000 to 2022, the full retirement age is increasing by 2 months each year so that by 2022 the full retirement age will be 67. The benefits reported in this section are total benefits received at a point in time. The MBR research extract provided by SSA to create the Gold Standard contains information about different reasons for receiving benefits but does not always allow the amount due to each reason to be accurately separated from the total. Hence we have elected to report total benefits at a point in time and researchers should be careful to note that when an individual is receiving both own retirement and aged spouse benefits, the amounts listed for each benefit type will be redundant, i.e. there is really only one total amount and two reasons for receiving it. SSA calculates benefits based on an individual's lifetime earnings history following rules which they publish in "Annual Statistical Supplement to the Social Security Bulletin," available for each tax year on the Social Security website, www.ssa.gov. Aged Spouse Benefit These variables pertain to benefits awarded to aged spouses (TOB=3). Disability Benefit These variables pertain to benefits awarded to due to own disability (TOB=2). Other Benefit These variables pertain to benefits awarded to spouses caring for minor children (TOB=4), widows(ers) caring for minor children (TOB=6), disabled widows(ers) (TOB=7), and adults disabled in childhood (TOB=8). Retirement Benefit These variables pertain to benefits awarded at time of own retirement (TOB=1). Widowed Spouse Benefit These variables pertain to benefits awarded to widowed spouses (TOB=5). Supplemental Security Record Variables The Supplemental Security Records (SSR) is SSA's main file to track who is receiving Supplemental Security Income (SSI) benefits and the monthly benefit amounts payable. SSI benefits are paid to elderly, blind, or disabled individuals who fall below certain income threshholds. Eligibility and federal payment standards are uniform across all states but states have the option to supplement federal payments. The payment included here is the total of both federal and state SSI payments. Geographic Variables These variables are included for internal use purposes and describe the geography of the address where the SIPP household was first interviewed. These variables are panel specific, with the year subscript at the end of the variable name representing the source panel. Individuals will have non-missing geography information only for the variables that correspond to their SIPP panels. Identifiers Identifiers for Panel, Person, Spouse, and Panel end date IRS/SSA Variables The Census Bureau sent a list of validated SSNs from the seven included SIPP panels to SSA and extracts from the Master Earnings File (Summary and Detailed Earnings Records), Master Beneficiary Record, Supplemental Security Record, 831 Disability File, and Payment History Update System were created. The variables from these files that are included in the SSB are described below. Not all SIPP respondents have linkages to SSA/IRS administrative data, including: those who refused to provide their SSN; those whose SSNs were not validated; and those with valid SSNs who never worked, and never applied for benefits or received benefits. In the Gold Standard, individuals without a validated SSN or without SSA/IRS administrative records had missing data for all SSA/IRS-derived variables described below. Among these people, those respondents without a validated SSN had all administrative data imputed as part of the data completion process. Hence in the completed Gold Standard and the synthetic data, only individuals with no work or benefit history have zero earnings and missing benefits. Detailed Earnings Record Variables The Detailed Earnings Records (DER) contains historical earnings reports for each person and job held from 1978 onwards. These reports include self-employment income. Earnings are not capped at the taxable maximum. For each tax year, we summed DER information for each person across all jobs and self-employment to create a total earnings amount. Summary Earnings Record Variables The SSA/IRS Summary Earnings Records (SER) contain historical person-level earnings data. In addition to an array of annual FICA-taxed earnings (1951-2006) that are capped at the FICA taxable maximum, the SER provides information regarding quarters of covered work. Quarters of covered work are utilized by SSA to determine eligibility for participation in its old age, survivors, and disability insurance (OASDI) programs. Variables are in array form, i.e., in the form VARNAME_YYYY where YYYY is the year the variable refers to. Fertility Variables Number of children and dates of birth. Lifespan Variables Birth and Death dates. SIPP Arrays This section contains variables that are repeated over time, reflecting the panel nature of the SIPP. For these time series variables, individuals will only have values for the months and years in which they participated in a SIPP panel. We also include only years fully covered by one of the SIPP panels. Hence there are no data for 1995 or 2000 since neither of these years were fully covered by a panel. (VARNAME_YYYYM[M], YYYY=1990-2006, M[M]=1 to 12) Health Insurance Variables Incluces health coverage, by month, and whether that health-insurance was employer-provided. Income Variables Variables that include income. Labor Force Variables Variables about participation in the labor force: hours and weeks worked Education Variables Variables related to education (from SIPP). Disability Variables Variables related to concepts of disability. Flag: Linked marriage ended 0 Linked marriage did not end 1 Linked marriage ended Sysmiss Flag: Reason linked marriage ended 0 Linked marriage ended in divorce 1 Linked marriage ended because of a death Sysmiss There is no linked marriage OR there is a linked marriage but it did not end during panel SIPP Panel Year indicates panel of source record 1990 1991 1992 1993 1996 2001 2004 SIPP Panel End Date The date the SIPP panel ended. SAS Date linked marriage began Date the linked marriage started SAS Date linked marriage ended Date the linked marriage ended Total Number of Children in Family Number of children under the age of 18 that live in a family in the interview month in which marital status is first observed (for those without spouses during the course of the SIPP) or in which the respondent's spouse is assigned. This number is the same for all family members and does not indicate that the children are related to a particular individual (fnkids for 1990-1993 panels, rfnkids for 1996-2004 panels). Flag: Marital History Event 1 First marital history event flag. Indicates whether respondent entered into a first marriage. 0 Never married 1 First marriage occurred Sysmiss Flag: Marital History Event 2 Second marital history event flag. Indicates whether respondent's first marriage ended in widowhood or divorce. 0 First Marriage did not end over course of survey 1 First marriage ended in widowhood 2 First marriage ended in divorce/separation Sysmiss Flag: Marital History Event 3 Third marital history event flag. Indicates whether respondent entered into a second marriage. 0 No second marriage 1 Second marriage occurred Sysmiss Flag: Marital History Event 4 Fourth marital history event flag. Indicates whether respondent's second marriage ended in widowhood or divorce. 0 Second marriage did not end over course of survey 1 Second marriage ended in widowhood 2 Second marriage ended in divorce/separation Sysmiss Flag: Marital History Event 5 Fifth marital history event flag. Indicates whether respondent entered into a third marriage. 0 No third marriage 1 Third marriage occurred Sysmiss Flag: Marital History Event 6 Sixth marital history event flag. Indicates whether respondent's third marriage ended in widowhood or divorce. 0 Third marriage did not end over course of survey 1 Third marriage ended in widowhood 2 Third marriage ended in divorce/separation Sysmiss Flag: Marital History Event 7 Seventh marital history event flag. Indicates whether respondent entered into a fourth marriage. 0 No fourth marriage 1 Fourth marriage occurred Sysmiss Flag: Marital History Event 8 Eighth marital history event flag. Indicates whether respondent's fourth marriage ended in widowhood or divorce. 0 Fourth marriage did not end over course of survey 1 Fourth marriage ended in widowhood 2 Fourth marriage ended in divorce/separation Sysmiss Date of Marital History Event 1 SAS date value Date of marital history event 2 SAS date value Date of marital history event 3 SAS date value Date of marital history event 4 SAS date value Date of marital history event 5 SAS date value Date of marital history event 6 SAS date value Date of marital history event 7 SAS date value Date of marital history event 8 SAS date value Unique person identifier Across the Gold Standard and Completed Data files, personid uniquely identifies SIPP respondents. In the SSB, personid uniquely identifies records within a particular implicate. In order to strengthen confidentiality protection, personid in the SSB does not link records across implicates or to the Gold Standard and Completed Data files. Personid of spouse Personid of linked spouse. Across the Gold Standard and Completed Data files, spouse_personid uniquely identifies spouses of SIPP respondents. In the SSB, spouse_personid uniquely identifies records within a particular implicate. In order to strengthen confidentiality protection, spouse_personid in the SSB does not link records across implicates or to the Gold Standard and Completed Data files. Linked spouse is defined as the first person to whom the SIPP respondent was married during the time period covered by the SIPP panel. Individuals could enter the panel already married and then each would be linked to the other. Individuals could also get married during the course of the panel. If this was the first observed marriage for each member of the couple, they were linked together. Individuals could also get divorced during the course of the panel and then remarry. In many cases, this later marriage caused a new individual to join the panel. This new SIPP respondent would only be linked to his or her spouse if the spouse (and original SIPP sample member) had not already been observed married to someone else. If the original SIPP sample member had been previously linked by marriage to another SIPP sample member, this original link was maintained in spouse_personid. However the marital history reflects the ending of this marriage and the occurrence of the next marriage for the original SIPP sample member. Likewise, the new SIPP sample member who joins through marriage will have that marriage date recorded in his or her marital history but will have a blank spouse_personid. In summary, this variable captures only one marriage partner and does not provide a history of marriage partners even if this history is (partially) observed in the SIPP. The link between SIPP respondents and their spouses has not been perturbed in any way in the SSB. The same individuals will be linked as married partners in the Gold Standard, the Completed Data, and the SSB. Male In the 1990-2004 Census-internal SIPP panels, a value for sex is included on each wave file. Thus, there are actually as many sex variables as there are waves of the survey and some changes occur across waves as a result of data collection error. Sex is selected from the array of variables sex1-sex{max number of waves} in which the wave corresponds either to the month in which marital status is first observed (for those without spouses during the course of the SIPP) or to the month in which the respondent's spouse is assigned instead of from a fixed point in the survey. Thus when a spouse is never assigned, an individual's gender comes from the first wave where they report being not married. For individuals who are assigned a spouse, gender comes from the first wave where they reveal their spouse. Finally, an indicator variable for males was created from the categorical sex variable for analytic convenience. This variable is unsynthesized on the SSB and is never missing so there are no imputed values in the Completed Data. 0 Female 1 Male This variable tells whether a person is male or female Date of Birth

This variable was taken from a hierarchy of SSA sources instead of the respondent-provided value in the SIPP. Date of birth was selected from the first non-missing value in the following files: (i) SSA's Master Benefits Record (MBR) file, (ii) the Census Bureau's Person Characteristic File (PCF) whose main input is the SSA Numident file, and (iii) SSA's Supplemental Security Record (SSR) file. Thus, this variable is administrative and sometimes differs from the birth date reported in the SIPP survey itself. When missing due to the lack of a validated SSN for the SIPP respondent, date of birth was imputed using date of birth from the Census-internal version of the SIPP. We chose the administrative source for two reasons. First, the administrative birth date was more often consistent with the other administrative data (benefits and earnings). For example, when age was calculated using the administrative birth date, there were fewer individuals who appeared to retire before age 62. Second, the differences between the administrative birth date and the birth date reported in the survey helped to increase the difficulty of re-identifying a record in the original SIPP public use data from a record in the synthetic data, thus improving the confidentiality protections. This variable is coded as a SAS date variable. This format gives the number of days between the date of birth and January 1, 1960. An individual born on January 1, 1959 would have birthdate=-365 and an individual born on January 1, 1961 would have birthdate = 365.

Race In the 1990-2004 Census-internal SIPP panels, a value for race is included on each wave file. Thus, there are actually as many race variables as there are waves of the survey and some changes occur across waves as a result of data collection error. Race is chosen by creating an array of variables race1-race{max number of waves} and choosing the first non-missing value. Thus race comes from the first wave in which the individual was interviewed instead of from a fixed point in the survey. 1 White 2 Black 3 Other Hispanic In the 1990-1993 SIPP panels, a value for ethnicity is included on each wave file. Thus, there are actually as many ethnicity variables as there are waves of the survey and some changes occur across waves as a result of data collection error. Ethnicity is chosen by creating an array of variables ethncty1-ethncty{max number of waves} and choosing the first non-missing value. Thus, ethnicity comes from the first wave in which the individual was interviewed instead of from a fixed point in the survey. Respondents are coded as Hispanic if they have an ethnicity code between 14 and 20. In the 1996-2004 panels, the longitudinally-edited version contains only one value for ethnicity across all waves (eorigin) and this value is used. Respondents are coded as Hispanic if they have an ethnicity code between 20 and 28 in 1996 and 2001, or if they have an ethnicity code of 1 in 2004. 0 Non-hispanic 1 Hispanic Flag: Existence of Date of Death Flag to indicate that this respondent died after being interviewed and no later than 2006. 0 Death date does not exist, respondent did not die during this interval 1 Death date exists, respondent died during this interval Date of Death Date of death from administrative data. This variable also is obtained using a hierarchy of administrative sources: (i) SSA's MBR file, (ii) the Census PCF with death information coming from the SSA Numident and Master Death Files, and (iii) SSA's SSR file. This variable is coded as a SAS date variable. This format gives the number of days between the date of birth and January 1, 1960. An individual born on January 1, 1959 would have birthdate=-365 and an individual born on January 1, 1961 would have birthdate=365. Ordinal Number of First Obs Marriage Tells which of the marriages described in the marital history (mh) arrays is the linked marriage. 1 Linked marriage is first reported marriage described by mh1, mh2, mh_date1, and mh_date2 2 Linked marriage is first reported marriage described by mh3, mh4, mh_date3, and mh_date4 3 Linked marriage is first reported marriage described by mh5, mh6, mh_date5, and mh_date6 4 Linked marriage is first reported marriage described by mh7, mh8, mh_date7, and mh_date8 Sysmiss When there is no linked marriage Flag: 4 or More Marriages Flag for existence of a marriage for which date is unknown because it was not collected in the SIPP. The marital history topical module asks about a person's first and second marriages and then his or her most recent marriage. If any other marriages occurred after the second but before the most recent, no information about this marriage is collected. However, individuals are categorized as having 1, 2, 3, or more than 3 marriages. We create flag_mar4t to identify individuals who reported more than 3 marriages at the time of the topical module. 0 No additional marriage occurred with unknown date 1 An additional marriage occurred but with unknown date Sysmiss Number of Children Ever Born Number of children ever born. This is taken from the wave 2 Fertility history topical module (TM8752 and TM8754 for 1990-1993 panels; tfrchl and tmomchl for 1996-2004 panels). Year of Birth of First Child This is taken from the wave 2 Fertility history topical module (TM8762 and TM8794 for 1990-1993 panels; tfbrthyr for 1996-2004 panels). Sysmiss - Structurally missing (own_kids_ever=0) Not certain that catgry and sumstat can be combined. Year of Birth of Last Child This is taken from the wave 2 Fertility history topical module (TM8768 and TM8782 for 1990-1993 panels; tlbirtyr for 1996-2004 panels). Sysmiss - Structurally missing (own_kids_ever=0) Not certain that catgry and sumstat can be combined. Education Category (5) Highest level of education attained at the time of the education history topical module. This variable was created from information gathered in the topical module on education history and represents the highest level of education achieved up to the point of the administration of the topical module questions. For individuals who did not answer the topical module education history questions, we imputed values for educ_5cat. 1 No high school degree 2 High school degree 3 Some college 4 College degree 5 Graduate degree Flag currently enrolled in high school This variable indicates whether an individual is still enrolled in high school. It can used to distinguish the difference between people with educ_5cat=1 who have not finished their education and those who have dropped out of high school. 0 Not currently enrolled in high school 1 Currently enrolled in high school Sysmiss Year Ended HS (or less) Education The wave 2 Education and Training History topical module provides knowledge of the year that high school was last attended (variables TM8404 and TM8412 for 1984 and 1990-1993 panels; variables tlstschl and thsyr for 1996-2004 panels). Year Began Post-HS Education The wave 2 Education and Training History topical module provides knowledge of the year that post-high school education began (variable TM8420 for 1990-1993 panels; variable tcollstr for 1996-2004 panels). Sysmiss - Structurally missing (educ_5cat=1 or educ_5cat=2) Flag currently enrolled in college Indicates whether an individual is enrolled in college at time of SIPP education history topical module and has not finished his/her education. This variable can be used to differentiate between individuals who completed some college and stopped school and those who have finished some college but not yet stopped attending school. 0 Not currently enrolled in college 1 Currently enrolled in college Sysmiss Year Ended Post-HS Education The wave 2 Education and Training History topical module provides knowledge of the year that post-high school education ended (variables TM8426 and TM8440 for 1990-1993 panels; variables tlastcol, tvocyr, tassocyr, tbachyr, and tadvncyr for 1996-2004 panels). Sysmiss - Structurally missing (educ_5cat=1 or educ_5cat=2) Year of Bachelor Degree year bachelor's degree was finished Field of Bachelor Degree Field in which bachelor's degree was obtained. Taken from topical history module on education history. Categories for 1996-2004 panels and for 1990-1993 panels 0 Unknown 1 Agriculture/Forestry 2 Art/Architecture (1996-2004 panels); Biology (1990-1993 panels) 3 Business/Management 4 Communications (1996-2004 panels); Economics (1990-1993 panels) 5 Computer and Information Sciences (1996-2004 panels); Education (1990-1993 panels) 6 Education (1996-2004 panels); Engineering (1990-1993 panels) 7 Engineering (1996-2004 panels); English/Journalism (1990-1993 panels) 8 English/Literature (1996-2004 panels); Home Economics (1990-1993 panels) 9 Foreign Language (1996-2004 panels); Law (1990-1993 panels) 10 Health Sciences (1996-2004 panels); Liberal Arts/Humanities (including arts, architecture, music, languages, philosophy) (1990-1993 panels) 11 Liberal Arts/Humanities (1996-2004 panels); Mathematics/Statistics (1990-1993 panels) 12 Mathematics/Statistics (1996-2004 panels); Medicine (1990-1993 panels) 13 Nature Sciences (Biological and Physical) (1996-2004 panels); Nursing, Pharmacy, Health Technologies (1990-1993) 14 Philosophy/Religion/Theology (1996-2004 panels); Physical or Earth Sciences (1990-1993 panels) 15 Pre-Professional (1996-2004 panels); Police Science or Law Enforcement (1990-1993 panels) 16 Psychology 17 Social Sciences/History (1996-2004 panels); Religion/Theology (1990-1993 panels) 18 Other (1996-2004 panels); Social Sciences (1990-1993 panels) 19 Vocational or Technical Studies (1990-1993 panels) 20 Other (1990-1993 panels) Sysmiss Foreign Born Immigrant Status, born in country other than U.S. Taken from wave 2 topical module (TM8730, TM8734, TM8709 1990-1993 panels; eprstate, ebrstate and rcitiznt 1996 panel; eprstate, ebrstate and tcitiznt 2001 panel; eprstate, ebrstate, citiz, and ebornus 2004 panel) 0 Born in U.S. 1 Born in country other than U.S. Decade of Arrival to US (Foreign Born) Decade arrive in U.S. (answered when SIPP respondent was foreign_born) The year of arrival to the U.S. is from the Census-internal SIPP files (TM8736 1990-1993 panels; rmoveus 1996 panel; tmoveus 2001-2004 panels) .=Structurally missing, out of scope for question (foreign_born=0) 1 Before 1959 2 1960-1964 3 1965-1969 4 1970-1974 5 1975-1979 6 1980-1981 7 1982-1984 8 1985-1993 9 1994-1999 10 2000-2004 Sysmiss Structurally missing, out of scope for question (foreign_born=0) Own a Home Indicates whether the individual owned a home in the SIPP panel. 0 Do not own a home 1 Own a home Home Equity Self-reported home equity value Non-Housing Financial Wealth Non-housing wealth = total wealth minus home equity Total Net Worth Total net worth Flag: Industry Assigned Does person have valid industry from a job held during survey? 0 No, last worked 1984 or earlier, or no valid industry reported 1 Yes Flag: Occupation Assigned Does person have valid occupation from a job held during survey? 0 No, last worked 1984 or earlier, or no valid industry reported 1 Yes Industry Category (4) Industry is a characteristic of an individual's job and hence varies over time. There are industry values reported for (potentially) two jobs in each wave of the survey. Industry is chosen by summing earnings associated with the array of variables ws1ind1-ws1ind{max number of waves} and ws2ind1-ws2ind{max number of waves} in the 1990-1993 panels, and ejbind1_1-ejbind1_{max number of waves) and ejbind2_1-ejbind2_{max number of waves} in the 1996-2004 panels and choosing the industry associated with the greatest total earnings. Thus industry is the industry from which greatest earnings are derived in the survey. 1 Manufacturing 2 Wholesale/retail trade 3 FIRE, services, public administration, military 4 Agriculture, mining, construction, transportation, communications, and public utilities Sysmiss Occupation Category (3) Occupation is a characteristic of an individual's job and hence varies over time. There are occupation values reported for (potentially) two jobs in each wave of the survey. Occupation is chosen by summing earnings associated with the array of variables ws1occ1-ws1occ{max number of waves} and ws2occ1-ws2occ{max number of waves} in the 1990-1993 panels, and tjbocc1_1-tjbocc1_{max number of waves) and tjbocc2_1-tjbocc2_{max number of waves} in the 1996-2004 panels and choosing the occupation associated with the greatest total earnings. Thus occupation is the occupation from which greatest earnings are derived in the survey. 1 Managerial and professional specialty occupations 2 Technical, sales, and administrative support occupations 3 Other Sysmiss In-Scope for Pension (Level II) Individual must have been employed at time of pension topical module in order to answer the pension questions. 0 Pension was not in scope 1 Pension was in scope Defined Benefit Pension Plan Defined Benefit Pension Plan 0 No defined benefit pension plan 1 Had defined benefit pension plan Sysmiss Defined Contribution Pension Plan Defined Contribution Pension Plan 0 No defined contribution pension plan 1 Had defined contribution pension plan Sysmiss Disability (Sum of Core and TM) Health limits kind or amount of work For the 1996-2004 panels, information on work-limiting disability comes from core question (edisabl) during wave 2 for people ages 15-69, when respondents were asked whether health limited the type or amount of work. This indicator is supplemented with details from the Functional Limitations and Disability topical module (wave 5 for 1996-2004 panels, variable ejobdif) that covers people ages 16-67. For the 1990-1993 panels, disability information comes from the core question (disab) during wave 2 for people ages 15-69, when respondents were asked whether health limited the type or amount of work. This information is supplemented with details from the Functional Limitations and Disability topical module (waves 3, 3, 6, 3 for 1990-1993 panels, variables TM8914, TM8918, and TM8920) that covers people ages 16-67. In order to make the sum_disab consistent for all panels, both responses were taken from the core and topical modules, when available, with any positive indication of health limiting the kind or amount of work flagging the positive response. .=Structurally missing, out of scope for question (sum_disab_in_scope=0) 0 No (sum_disab_in_scope=1) 1 Yes (sum_disab_in_scope=1) Sysmiss Structurally missing, out of scope for question (sum_disab_in_scope=0) Disability Prevents Work (Sum of Core and TM) Health prevents work For the 1996-2004 panels, information on work-preventing disability comes from core question edisprev during wave 2 for people ages 15-69, when respondents were asked whether health prevented work. This indicator is supplemented with details from the Functional Limitations and Disability topical module (wave 5 for 1996-2004 panels, variable ejobcant) that covers people ages 16-67. For the 1990-1993 panels, the core questionnaire does not ask respondents whether health prevents work. This information is solely obtained from the Functional Limitations and Disability topical module (waves 3, 3, 6, 3 for 1990-1993 panels, variables TM8922 and TM8924) that covers people ages 16-67. When available, the core and topical modules were used in conjunction to construct summary measures of disability, with any positive indication of health preventing work flagging the positive response. .=Structurally missing, out of scope for question (sum_disab_in_scope=0 or { sum_disab_in_scope=1, sum_disab=0}) 0=No (sum_disab=1) 1=Yes (sum_disab=1) 0 No (sum_disab=1) 1 Yes (sum_disab=1) Sysmiss Structurally missing, out of scope for question (sum_disab_in_scope=0 or { sum_disab_in_scope=1, sum_disab=0}) SER: Total Earnings Annual earnings taxed by FICA; these variables include earnings only up to the FICA taxable maximum, i.e., these earnings measures are capped. YYYY=1951-2006 SER: Annual Total Covered Quarters of Work Indicates the total number of quarters of FICA-covered work in year YYYY, where YYYY=1951-2006. 0 1 2 3 4 DER: Non-Deferred FICA Non-deferred earnings paid to an individual from jobs covered by FICA tax; summed across all employers in the DER to give a person-level total for each year YYYY, where YYYY=1978-2006. DER: Non-Deferred Non-FICA Non-deferred earnings paid to an individual from jobs NOT covered by FICA tax; summed across all employers in the DER to give a person-level total for each year YYYY, where YYYY=1978-2006. DER: Deferred FICA deferred earnings from jobs covered by FICA tax; summed across all employers in the DER to give a person-level total for each year. While the variable exists on the Gold Standard for the years 1978-1986, it is always missing in this time period. The year 1987 is the first year with positive deferred wages. On the synthetic and completed gold standard files, we only keep 1990-2006 because so few people had deferred wages between 1987 and 1989 that we could not reliably synthesize these variables. DER: Deferred Non-FICA deferred earnings from jobs NOT covered by FICA tax; summed across all employers in the DER to give a person-level total for each year. While the variable exists on the Gold Standard for the years 1978-1986, it is always missing in this time period. The year 1987 is the first year with positive deferred wages. On the synthetic and completed gold standard files, we only keep 1990-2006 because so few people had deferred wages between 1987 and 1989 that we could not reliably synthesize these variables. Flag: In MBR This flag indicates that a person matched to the SSA Master Beneficiary File (MBR). The person's SSN showed up in the MBR because they received benefits of some kind. 0 Respondent was not matched to MBR 1 Respondent was matched to MBR MBR: receive retire benefit This variable indicates that a person received retirement benefits at some point during the time period covered by the MBR extract (1964-2007). These benefits were the result of the individual's own earnings history. This variable is only in scope if flag_in_mbr=1. 0 Does not receive monthly retire benefit 1 Receives monthly retire benefit Sysmiss MBR: startdate of benefit Date when the person first began receiving own retirement benefits, conditional on having ever received this type of benefit. MBR: total monthly benefit Total monthly amount of benefits received at beginning of own retirement benefit entitlement. In most cases this amount is from the same month as in MBR_retire_benefit_stdate. However, if data for that month were missing in the MBR extract, we searched through the monthly benefit array to find the first positive value. This amount can be a combination of payments due to multiple entitlement reasons (i.e. dual entitlement). PHUS: startdate of benefit Date retirement benefits began being paid, as recorded in the PHUS. This date must be greater than or equal to the MBR retirement benefit start date. It also must be 1984 or later because PHUS data began in 1984. PHUS: total monthly benefit Total amount of benefits as recorded in the PHUS in the first month of receiving own retirement benefits. This amount can be a sum of benefits received for different reasons (i.e. dual entitlement). MBR: receive disab benefit Indicates that individual received disability benefits at some point over the time period covered by the MBR. This variable is only in scope if flag_in_mbr=1. 0 Does not receive monthly disability benefit 1 Receives monthly disability benefit Sysmiss MBR: startdate of benefit Date at which individual began receiving own disability benefits. This date must be before individual reaches the full retirement age (FRA). FRA depends on the year the person reaches age 62. Any individual who turned 62 before 2000 had FRA=65 years old. Beginning in 2000, any individual turning 62 had full retirement age of 65 years 2*(year_age_62 - 1999) months. MBR: total monthly benefit Total monthly amount of benefits received at beginning of disability benefit entitlement. In most cases this amount is from the same month as in MBR_disab_benefit_stdate. However, if data for that month were missing in the MBR extract, we searched through the monthly benefit array to find the first positive value. This amount can be a combination of payments due to multiple entitlement reason (i.e. dual entitlement). PHUS: startdate of benefit Date disability benefits began being paid, as recorded in the PHUS. This date must be greater than or equal to the MBR disability benefit start date. It also must be 1984 or later because PHUS data began in 1984. PHUS: total monthly benefit Total amount of benefits as recorded in the PHUS in the first month of receiving own disability benefits. This amount can be a sum of benefits received for different reasons (i.e. dual entitlement). MBR: receive agedsp benefit This indicator reports receipt of aged spouse Social Security benefit. This variable is only in scope if flag_in_mbr=1. 0 Does not receive monthly agedsp benefit 1 Receives monthly agedsp benefit Sysmiss MBR: startdate of benefit Date when the person first began receiving aged spouse benefits, conditional on having ever received this type of benefit. MBR: total monthly benefit Total monthly amount of benefits received at beginning of aged spouse benefit entitlement. In most cases this amount is from the same month as in MBR_agedsp_benefit_stdate. However, if data for that month were missing in the MBR extract, we searched through the monthly benefit array to find the first positive value. This amount can be a combination of payments due to multiple entitlement reasons (i.e. dual entitlement). PHUS: startdate of benefit Date aged spouse benefits began being paid, as recorded in the PHUS. This date must be greater than or equal to the MBR aged spouse benefit start date. It also must be 1984 or later because PHUS data began in 1984. PHUS: total monthly benefit Total amount of benefits as recorded in the PHUS in the first month of receiving aged spouse benefits. This amount can be a sum of benefits received for different reasons (i.e. dual entitlement). MBR: receive widowsp benefit This variable indicates receipt of widowed spouse Social Security benefits at some point in the time period covered by the MBR (1962-2007). This variable is only in scope if flag_in_mbr=1. 0 Does not receive monthly widowsp benefit 1 Receives monthly widowsp benefit Sysmiss MBR: startdate of benefit Date when the person first began receiving widowed spouse benefits, conditional on having ever received this type of benefit. MBR: total monthly benefit Total monthly amount of benefits received at beginning of widowed spouse benefit entitlement. In most cases this amount is from the same month as in MBR_agedsp_benefit_stdate. However, if data for that month were missing in the MBR extract, we searched through the monthly benefit array to find the first positive value. This amount can be a combination of payments due to multiple entitlement reasons (i.e. dual entitlement). PHUS: startdate of benefit Date widowed spouse benefits began being paid, as recorde in the PHUS. This date must be greater than or equal to the MBR widowed spouse benefit start date. It also must be 1984 or later because PHUS data began in 1984. PHUS: total monthly benefit Total amount of benefits as recorded in the PHUS in the first month of receiving widowed spouse benefits. This amount can be a sum of benefits received for different reasons (i.e. dual entitlement). MBR: receive other benefit This variable indicates receipt of Social Security benefits due to one of four reasons: young widow caring for minor children, young spouse caring for minor children, disabled widow, and adult disabled in childhood. 0 Does not receive other monthly benefits 1 Receives other monthly benefits Sysmiss MBR: startdate of benefit Date when the person first began receiving other benefits, conditional on having ever received this type of benefit. MBR: total monthly benefit Total monthly amount of benefits received at beginning of other benefit entitlement. In most cases this amount is from the same month as in MBR_other_benefit_stdate. However, if data for that month were missing in the MBR extract, we searched through the monthly benefit array to find the first positive value. This amount can be a combination of payments due to multiple entitlement reasons (i.e. dual entitlement). PHUS: startdate of benefit Date other benefit began being paid, as recorded in the PHUS. This date must be greater than or equal to the MBR other benefit start date. It also must be 1984 or later because PHUS data began in 1984. PHUS: total monthly benefit Total amount of benefits as recorded in the PHUS in the first month of receiving other benefits. This amount can be a sum of benefits received for different reasons (i.e. dual entitlement). Flag: in SSR This flag indicates that a person's SSN was found in the SSA Supplemental Security Records (SSR). This database tracks people who received SSI. 0 SSN not found in SSA Supplemental Security Records (SSR) 1 SSN found in SSR SSR: SSI Amount - Initial ($2000) amount of monthly SSI payment at time of initial receipt SAS Date - SSR: SSI Date of Initial Entitlement Date of initial entitlement to SSI benefits Health Insurance Coverage 0 Respondent did not have health insurance coverage during this month 1 Respondent had health insurance coverage during this month Sysmiss Health Insurance Coverage from Employer 0 Respondent did not have employer-provided health insurance 1 Respondent had employer-provided health insurance Sysmiss Total Personal Income monthly personal income summed from all sources Sysmiss Total Hours Worked at All Jobs Total number of hours worked at all jobs in a given month Sysmiss Weeks at a Job Total number of weeks worked at a job in a given month Sysmiss Weeks With Pay Total number of weeks worked with pay in a month. Weeks worked with pay = weeks worked - weeks worked without pay; Sysmiss Total SIPP Earnings This is variable is taken from the recoded public-use variable totearn - total person monthly earnings from all sources. Sysmiss State of Residence: FIPS code (modified) State of residence. FIPS State Code for state of residence first recorded in the SIPP. For married couples, we take the state value for both partners at the same point in the survey when we first observed the marriage. For individuals who never have an observed marriage during the panel, we take their first ever reported state value. *All panels prior to 2004 group some states together and give only one code for the group. For these panels, the individual FIPS code will not appear for states contained in a group. 1 Alabama 2 Alaska *see description 4 Arizona 5 Arkansas 6 California 8 Colorado 9 Connecticut 10 Delaware 11 DC 12 Florida 13 Georgia 15 Hawaii 16 Idaho *see description 17 Illinois 18 Indiana 19 Iowa *see description 20 Kansas 21 Kentucky 22 Louisiana 23 Maine *see description 24 Maryland 25 Massachusetts 26 Michigan 27 Minnesota 28 Mississippi *see description 29 Missouri 30 Montana *see description 31 Nebraska 32 Nevada 33 New Hampshire 34 New Jersey 35 New Mexico *see description 36 New York 37 North Carolina 38 North Dakota *see description 39 Ohio 40 Oklahoma 41 Oregon 42 Pennsylvania 44 Rhode Island 45 South Carolina 46 South Dakota *see description 47 Tennessee 48 Texas 49 Utah 50 Vermont *see description 51 Virginia 53 Washington 54 West Virginia *see description 55 Wisconsin 56 Wyoming *see description 61 (1990, 1991, 1992, 1993, 1996, 2001 panels) Maine, Vermont *see description 62 (1990, 1991, 1992, 1993 panels) Iowa, North Dakota, South Dakota (1996, 2001 panels) North Dakota, South Dakota, Wyoming *see description 63 (1990, 1991, 1992, 1993 panels) Alaska, Idaho, Montana, Wyoming *see description 90 (1984 panel only) Idaho, New Mexico, South Dakota, Wyoming *see description 91 (1984 panel only) Mississippi, West Virginia *see description