SIPP USERS’ GUIDE SIPP PUBLIC USE FILES 9. The SIPP Public Use Files This is the first of five chapters in Section II: this chapter provides a general introduction to the public use files; one chapter is devoted to each of the three types of SIPP data files, and a final chapter discusses merging multiple SIPP data files. After reading the current chapter, the user working with just one type of SIPP data file may wish to turn to the chapter on that type of file. For the 2004 and 2008 Panels, most variable names did not change from the 1996+ panels. For the 1996 Panel, most variables names changed from the pre-1996 Panels. To aid users working with files from panels prior to 1996, the chapters in Section II present both variable names for the pre-1996 and the 1996+ panels. In the main body of the text, the pre-1996 Panel names are presented in parentheses following those from the 1996+ Panels. For example, the sample unit ID variable name in the core wave files, which is >SSUID= in the 1996+ Panel, was SUID in pre-1996 Panels. The variable name is written in this chapter as SSUID (SUID). In tables, a variety of methods are used to present both sets of names. Types of SIPP Data Files There are three types of public use files containing SIPP data: core wave files, topical module files, and full panel longitudinal research files (referred to as either longitudinal files or full panel files):  Core wave files are currently issued in person-month format. These files contain up to four records for each primary sample member and each person who lived with a primary sample member at any time during the 4-month reference period covered by the wave.  Each of the records contains data from one of the four reference months covered by the wave.1  Full panel longitudinal research files contain onerecord for each primary sample member and or each person who ever lived with a primary sample member at any time during the SIPP panel – a period of up to 3 years.  Topical module files for the 2004 Panel contain one record for each person who was a sample responding (or Type Z nonresponding) member of a SIPP household during the fourth month of the reference period for the wave. Topical module files from earlier panels contain one record for each primary sample member and each person who lived with a primary sample 1 Prior to the 1990 Panel, core wave files were issued with a single record for each person. Each record contained data for all four of the reference months covered by the wave. The structure of the file was similar to the longitudinal files issued by the Census Bureau. Earlier editions of this Users Guide provide details. 9-1 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES member at the time of the interview for the wave in which the topical module was administered. Understanding the ID Variables in SIPP Because different files contain different information, the capacity to identify people across thosefiles is important. SIPP is a longitudinal survey designed to allow researchers to track people over time; other critical functions include identifying individuals over time and identifying when a person is present in the sample. Finally, because the relationships among people change over time, identification of those relationships at any specific time is important. The key to these tasks lies in understanding how SIPP ID variables are used to identify persons, families, and households.2 The most basic ID variables in SIPP have different variable names in the different types of public use files issued by the Census Bureau. Table 9-1 displays those variables and shows the names they are given in the different files. Sample Unit IDs When initial Wave 1 interviews are conducted, each physical dwelling unit is assigned a unique (random) sample unit ID.3 The sample unit ID assigned to a person never changes: in all subsequent interviews, the Wave 1 primary sample persons carry their sample unit IDs with them. This means that if they move to different addresses, they keep the same sample unit IDs. If new people join those original sample members at their original addresses, they become secondary sample members by virtue of their association with the primary sample person with whom they are living. Secondary sample persons are all assigned the sample unit ID of the primary sample member with whom they are living. At the conclusion of the panel, all people who have ever lived with a member of a given original sample unit share the same sample unit ID. That sample unit ID is their common link to the original sample unit. 2 Other variables are used to identify people who are members of related subfamilies, unrelated subfamilies (also known as secondary families), and transfer program units such as food stamp units. 3 The sample unit ID is a random recode of three other variables in the Census Bureau internal files: the respondent=s sampling area, the cluster of housing units within that area (called a segment), and a sequentially assigned serial number. Because the variables in the Census Bureau=s internal files contain detailed information about the location of the dwelling unit, those variables are suppressed in the public use files to protect the confidentiality of survey respondents. 9-2 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES Current Address IDs The current address ID identifies each housing unit occupied by one or more original sample members in any given month.4 Current address IDs are assigned within sample units (they are unique only when combined with the sample unit ID variable), and they have two parts. The first part (one digit for all but the 1992 and 1996 Panels, two digits for the 1992 and 1996 Panels) identifies the wave in which one or more original sample members were first scheduled to be interviewed at the address. The second part of the ID is one digit, and it is used to sequentially number addresses for households that split into two or more households as a result of a move to a different location by original sample persons. All Wave 1 households have a current address ID of 11. Any new addresses that are occupied in Wave 2 are numbered 21, 22, and so on; new addresses occupied during the wave 3 reference period are numbered 31, 32, 33, and so on. The current address ID is a monthly variable, the value of which changes in the month in which an individual moves to a new address. Entry Address IDs The entry address ID is the current address ID that a sample member occupied when he or she first entered the SIPP sample. It is used in conjunction with the person number to uniquely identify persons within the sample unit and does not change even if the person moves. Person Numbers All primary and secondary sample members are assigned a person number when they first enter the SIPP panel. Those numbers are assigned sequentially, within each wave and within each household (current address). The first part of the person number (two digits for the 1992 and 1996 Panels, one digit for all others) indicates the wave in which the person originally entered the sample. Thus, primary sample persons have person numbers in the 100 series, beginning with 101; secondary sample members have person numbers beginning with 201 if they enter the sample in Wave 2, 301 if they enter the sample in Wave 3, 401 if they enter the sample in Wave 4, and so on. Identifying Persons and Their Relationships Each person in SIPP can be uniquely identified by the combination of a sample unit ID, an entry address ID,5 and a person number. These ID variables are useful when linking the records for a single person across multiple SIPP data files. They also contain substantive information that may be useful in some situations. 4 A house, an apartment or other group of rooms, or a single room is regarded as a housing unit if it is occupied or intended for occupancy as separate living quarters. 5 For the 1996 Panels, the entry address is not necessary to uniquely identify individuals in SIPP. Its continued use will not create any problems; it just provides additional information. 9-3 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES Table 9-1. SIPP Variable Names, by File Type Sample Current Person File Type Unit ID Address ID Entry Address ID Number Panels Prior to the 1996 Panel Core Wave Person-Month Files SUID ADDID ENTRY PNUM Topical Module Files ID ADDID ENTRY PNUM Full Panel (and Partial-Panel) Longitudinal Research Files PP-ID HH-ADDID PP-ENTRY PP-PNUM 1996 Panel EENTAID (No longer Core Wave Person-Month Files SSUID SHHADID needed to identify persons) EPPPNUM EENTAID (No longer Topical Module Files SSUID SHHADID needed to identify persons) EPPPNUM 2001 Panel Core Wave Person- EENTAID (No longer Month Files SSUID SHHADID needed to identify persons) EPPPNUM EENTAID (No longer Topical Module Files SSUID SHHADID needed to identify persons) EPPPNUM 2004 Panel Core Wave Person- EENTAID (No longer Month Files SSUID SHHADID needed to identify persons) EPPPNUM EENTAID (No longer Topical Module Files SSUID SHHADID needed to identify persons) EPPPNUM Using the Monthly Interview Status Variable The monthly interview status variable helps determine whether the data for a person in a given month should be used. This variable is labeled PP-MIS in the pre-1996 longitudinal files, in the (older) person-record-format core wave files, and in older topical module files. It is labeled PPMIS in newer pre-1996 topical module files. 6 This variable has three possible values: 0, 1, and 2. When using the older person-record-format core wave files, the topical module files for panels prior to 1996, and the longitudinal files, the monthly interview status is the only reliable guide as to whether the data for a given person should be used in a given month. Analysts should use data for only those months in which a person=s interview status is equal to 1. Any data present for months when a person=s interview status is coded either 0 or 2 should be ignored. A code of 0 indicates that the person was not in the sample for that month, and a code of 2 indicates a noninterview for that 6 The person-month-format core wave files contain records only for those months that a person has an interview status code of 1. The monthly interview status variable is not included in those files because it is not needed. The topical module files for the 1996 Panel contain records only for those with an interview status code of 1 in the fourth month of the wave=s core reference period. Although, the interview status variable is included on the topical module files from the 1996 Panel, it need not be used with them. 9-4 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES 7 month. When working with other data sources, analysts often identify which cases will be used in an analysis by examining either the weight variable or the variables used in the analysis itself. In the first case, the rule is generally to use all cases with positive weights and ignore the rest. In the second case, the rule is generally to use all cases with nonmissing data. Each of those rules can lead the SIPP user astray, as illustrated below. The presence of a zero weight is not a reliable guide to whether a person should be excluded from the planned analysis. Although those people will not enter into any weighted tabulations, they may provide important contextual information about people who do enter into those (weighted) tabulations. For example, a person with a calendar year weight of zero who is a member of the same household as a positive-weight person for only 3 months provides information about the positive-weighted person=s household (including, for example, household size, composition, income, and program participation) for the 3-month period that he or she was a household member. It is for this reason that records for zero-weighted persons are retained in the SIPP data files.8 The presence of data in analysis fields for any given month is also not a reliable guide to whether the person should be included in the planned analyses. Data are collected for all months of the reference period for a given wave, even if the interviewed person was in the sample for only part of the reference period. For example, on the topical module and longitudinal files for panels prior to 1996, 4 months= worth of data will generally be present for a person who was a member of a SIPP household for only the last 2 months of the wave. However, only those last 2 months of data should be used.9 7 For those months when a noninterviewed person was both in scope for the survey and had data imputed (this includes the Type Z imputations and the missing wave imputations), the variable is set to 1. In those cases, the data can be used in the same manner as any of the other imputed data in the SIPP public use files. 8 Other important situations also arise. For example, infants are assigned a calendar year weight of zero for the year of their birth even though they have an interview status of 1 from their birth month forward. Also, a person who dies during the year will have a positive calendar year weight even though, past the month of death, he or she will have an interview status of 0 or 2. In neither case does the weight variable reflect the presence or absence of the person, or data associated with the person. 9 The person-month-format core wave files will have only two records for that person. The topical module files for the 1996 Panel will have information only about month 4 of the wave’s core reference period. 9-5 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES Determining Monthly Household Composition A household, as the term is used in Census Bureau publications, consists of all people who occupy a housing unit, regardless of their relationships to each other.10 For many purposes, a household can be thought of as people living at a common address. A person=s current address ID in any given month, together with his or her sample unit ID, identifies the household in which that person is a member for that month. Members of the same household in a given month always have an interview status of 1 and share the same sample unit ID and current address ID. Chapter 2 provides illustrations of changes in household composition. Determining Monthly Family Composition The term family, as used in Census Bureau publications, refers to a group of two or more people related by birth, marriage, or adoption who reside together; all such people are considered members of one family. For example, if the son of the person who maintains the household and the son=s wife are members of the household, they are treated as members of the parent=s family. Every family must include a reference person. Two or more people living in the same household who are related to each other but not to the household reference person form an unrelated subfamily (also referred to as secondary families). The labels primary individual and secondary individual as used by the Census Bureau refer to people in households who are not related to any other household members. For many purposes, they can be thought of as one-person families, and the Census Bureau sometimes refers to them as pseudo-families. Methods for identifying the interrelationships among the household members that define these groups vary, depending on the data file being used. The topical module files do not contain any of the information needed to directly identify the different types of families.11 When it is necessary to identify family membership in an analysis that uses information from a topical module, it is also necessary to merge data from the topical module file with either a core wave file or a longitudinal file. Procedures for merging files are discussed in Chapter 13. Identifying family membership is easiest when working with the person-month-format core wave files. The Census Bureau has two principal methods for distinguishing families.  The first method defines a family as all persons who are related and living together. The family ID variable RFID is used with this definition. RFID groups the household reference 10 The one exception to this definition is people living in group quarters. 11 The one exception is the Wave 2 topical module that collects detailed information about all of the relationships among all of the people who are household members at the time of the Wave 2 interview. 9-6 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES person with all related household members by assigning them the same ID number. This family group corresponds to the Census Bureau=s definition of a primary family. RFID groups members of each unrelated subfamily (and primary and secondary individuals) separately.  The second method is similar to the first in defining a family, but the family excludes members of related subfamilies. The family ID variable RFID2 is used with this definition. RFID2 equals zero for members of related subfamilies. RFID2 groups members of each unrelated subfamily (and primary and secondary individuals) in the same way as RFID each group has a unique number. Analysts who want to analyze multi generational families would use RFID2 (FID2) and the variable RSID (SID). RSID (SID) treats related subfamilies as distinct family units by assigning members of related subfamilies nonzero values. Analysts can easily distinguish unrelated subfamilies from other family units when they use these variables and numbering schemes. Chapter 10 discusses the use of these variables in greater detail. More work is involved when using the longitudinal files or the (older) person-record-format core wave files. When working with those files, analysts must create a unique family ID from several components. Other approaches are described in earlier editions of this Guide. Determining Monthly Transfer Program Unit Composition Some analyses involve summarizing data for units other than households or families. The SIPP core data contain sufficient information to identify program units for participants in a range of transfer programs, including Medicare; Medicaid; Aid to Families with Dependent Children (AFDC); Temporary Assistance for Needy Families (TANF);12 General Assistance (GA); Railroad Retirement; Social Security; Veterans Compensation and Pensions; Food Stamps; and the Women, Infants, and Children’s Nutrition Program (WIC). The SIPP data contain fields for each adult and child, indicating whether the individual received benefits (either directly or by virtue of his or her relationship to another person designated as the principal recipient) from each of these programs in each month. The SIPP data also contain information that permits identification of program units within households. One person in each program unit is identified as a principal recipient, and variables identifying that principal recipient are included on the records of the people who are part of the program unit. People who are members of a common program unit in a given month can then be identified as those who are in the sample in 12 In August 1996, the Personal Responsibility and Work Opportunity Reconciliation Act was signed into law. This legislation replaced the old welfare system, Aid to Families with Dependent Children (AFDC), with a new program, Temporary Assistance for Needy Families (TANF). In the 1996 Panel, the questions for income type 20 referred to the AFDC program prior to Wave 4 and to the TANF program beginning in Wave 4. In Wave 9, the questions were expanded somewhat to capture the larger array of program types that could exist under TANF. 9-7 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES that month (interview status = 1) with common values of: ! The sample unit ID, ! The current address ID, and ! The primary recipient ID. Constructing Household, Family, and Program Unit Level Variables The public use files contain selected characteristics of monthly households and families that can be used directly in planned analyses. Data needs may require analysts to construct characteristics of households, families, or program units that do not already exist on the public use files created by the Census Bureau. Analysts can use the monthly ID variables described in the preceding section to construct monthly characteristics from the public use files. Choosing Appropriate Weight(s) Because SIPP uses a sample design in which different households (and people) are sampled at different rates, weights generally must be used when the user desires (approximately) unbiased estimates of population characteristics. In general, the appropriate weight to use for an analysis can be identified by answering two questions: 1. Which (sub) sample of SIPP is the estimate based on? 2. What population does the sample represent? Weights for each of the calendar months covered by a panel can be found on the core wave files. A single weight appears on the topical module files. Before 1996, the interview month was a frequent reference period for topical module questions, and the weight on the pre-1996 topical module files is the person interview month weight for people who provided data for a topical module. But, as noted earlier, starting with the 1996 Panel the interview month is no longer used as a reference month; the weight on the topical module file for the 1996 Panel is the person cross-sectional weight for the fourth reference month. Weights for estimates that refer to calendar years or, more accurately, the January population as it appears through the balance of the calendar years are on the longitudinal 9-8 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES 13 files, panel files or calendar year files. Chapter 8 provides detailed information about SIPP weights and how to use them. Working with Multiple Files There are a number of reasons that SIPP users commonly use data from more than one file: 1. The overlapping-wave/rotation-group structure of the survey creates many situations in which data for a single calendar reference month are contained on two different core wave files. 2. The overlapping-panel structure of the pre-1996 SIPP created many situations in which data covering a single calendar year could be found on data files from two or sometimes three different panels.14 3. There are many research problems in which reference to a specific calendar date is not crucial and a desire for increased sample size can lead to the use of data from multiple panels (or waves) that do not overlap. 4. Many analyses of data collected in the SIPP topical modules entail merging topical module data with files containing core data (the core wave files or the longitudinal research files). 5. Researchers requiring longitudinal data from more than one wave must create their own linked data files from the available core wave files. As of this writing, longitudinal files are available for all pre-1996 Panels, so this procedure is necessary for only 1996+ Panels. Chapter 13 discusses each of these situations and describes procedures for using data from multiple files to construct estimates. 13 The calendar year weights are based on all sample members who are present in January and interviewed (or imputed) for every month of the year that they were in scope for the survey. In other words, the weights include people who died during the year if they were interviewed until they died, but they do not include people who left the sample during the year. Because they are not members of the population on January 1, infants receive a calendar weight of zero for the year in which they are born. 14 Chapter 2 discusses the overlapping wave and panel structure of SIPP. 9-9 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES Section II Section II is organized as follows: ! Chapter 9 describes how to use public use files ! Chapter 10 describes how to use the core wave files. ! Chapter 11 describes how to use the topical module files. ! Chapter 12 describes how to use the full panel longitudinal research files. ! Chapter 13 describes how to link the different file types. Because many users work with only a single type of file, Chapters 9, 10, 11, and 12 are written so that they stand alone: each chapter can be used independently, without reference to the other two chapters. Differences across the three file types in their structure and in names for common variables make this a natural way to organize the material presented here. The advantage of this organization is that an analyst working with only a single type of file will find a complete discussion of that file type in a single chapter. However, there is substantial overlap in the types of things that analysts will be called upon to do with each of the file types. Thus, many ideas are repeated across the three chapters. Crucial differences do exist among the chapters, however. Those differences are found in the variable names used to accomplish certain common tasks and in the ways of working with data files built around different organizational principles. While the text of a chapter may seem familiar, there are often important differences in the details. Table 9-2 summarizes some of the more important differences among the three file types. Table 9-2 is intended primarily for users who have already worked with at least one type of SIPP data file. Analysts new to SIPP should skip the table and proceed to the chapter that discusses the type of data file with which they are working. When working with a different type of SIPP file, experienced analysts can use Table 9-2 in conjunction with the chapter that discusses that new file type; the table will help to highlight differences that might otherwise be overlooked in the general discussion. 9-10 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES Table 9-2. Differences Among Core Wave, Topical Module, and Longitudinal Files (1990 – 2008 Panels) 1996+ Panels Topic 1996 + Panels Core Pre-1996 Core Wave Topical Modules Pre-1996 Topical Pre-1996 Longitudinal File Wave Files Files Files Module Files File Structure Person-month records Person-month records Person records Person records Person records Table 10-1 Table 10-1 Table 11-1 Table 11-1 Table 12-2 Data Dictionary Size and begin Size and begin position Size and begin Size and begin position 1992-1993 Panels Size, begin, field position Figure 10-1 Figure 10-1 position Figure 11-1 Figure 11-1 length, and number of fields 1990- 1991 Panels Size, begin, index, and length Figure 12-1 Importance of Not needed on the On the person-month files: Not needed. Topical PP-MIS Very important PP-MIS Very important Table Monthly Interview person-month files-- not needed. Person-month module files contain Table 11-2. 12-2. Status Variables they contain records files contain records only records only for only for months in for months in which the people for whom which the respondent respondent=s interview EPPMIS4 = 1. is present and in scope. status equals 1. On the older person-record format files: very important. See earlier editions of this Users= Guide for details. How to Identify a SSUID, EPPPNUM SUID, ENTRY, PNUM SSUID, EPPPNUM ID, ENTRY, PNUM PP-ID, PP-ENTRY, PP- Person Table 10-3 Table 11-6 Table 11-7 PNUM Table 12-6 How to Identify a SSUID, SHHADID SUID, ADDID SSUID, SHHADID ID, ADDID PP-ID, HH-ADDID Table 12-8 Household Table 10-5 Table 11-8 Table 11-9 Identification of Merged households PWSUID, PWENTRY, or Merged households PNUM is between x80 PP-PNUM is between x80 and x99, “Merged cannot be identified in PWPNUM > 0. cannot be identified and x99, inclusively, and inclusively, and x varies from 1 to Households” files from the 1996 + in files from the x varies from 1 to 10. 10. Can identify the person only Panels. 1996+ Panel Can identify the person after the move; need to go to the only after the move; need core wave file to identify the to go to the core wave person before the move. file to identify the person before the move. (table continues) 9-11 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES Table 9-2. Differences Among Core Wave, Topical Module, and Longitudinal Files (1990-2008 Panels) (continued) 1996+ Panels Topic 1996 + Panels Pre-1996 Topical Modules Pre-1996 Topical Pre-1996 Longitudinal Core Wave Files Core Wave Files Files Module Files File Handling of “Merged Not Applicable If the move took place after the Not applicable No matter when the move No matter when the move Households” first reference month, there will be takes place, there will be takes place, there will be two two records for each person whose one record for each person records for each person whose ID information changed. One whose ID information ID information changed. One record reflects what happened changed. That record record reflects what happened before the move and contains the reflects what happened before the move and contains original ID information. The other after the move and contains the original ID information. record reflects what happened after the new ID information. The other record reflects what the move and contains the new ID happened after the move and information. If the move takes contains the new ID place in the reference month, there information. will be only record for each person whose ID information changed. That record reelects what happened after the move and contains the new ID information. How to Identify a SSUID, SHHADID (SUID and ADDID) and [FID or Not in the file Not in the file Create the family ID variables Family and RFID or RFID2 FID2 or SID or (FID2 and SID)] using PP-ID, HH-ADDID, and or RSID or [RFID2 Table 10-7 FAMTYP and RSID)] Table 12-10 Working with Variables for the Variables for the primary family Not applicable Not applicable Variables for the primary Family- primary family include the related subfamily in family include the related Level Income include the related them. Separate variables for the subfamily in them. No Variables subfamily in them. related subfamily. separate variables for the Separate variables Table 10-10 related subfamily. for the related Table 12-2 subfamily. Table 10-9 (table continues) 9-12 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES Table 9-2. Differences Among Core Wave, Topical Module, and Longitudinal Files (1990-2008 Panels) (continued) Topic 1996+ Panels Core Pre-1996 Core 1996+ Panels Module Pre-1996 Topical Pre-1996 Longitudinal Files Wave Files Wave Files Files Module Files Variables Describing RHNF HNF Household and RHNFAM HNFAM Family Composition RHNSF HNSF EHREFPER HREFPER EHHNUMPP HNP RHTYPE HTYPE EFREFPER FREFPER EFTYPE FTYPE EFKIND FKIND ESFT ESFRFPER FAMTYP FAMTYP FAMREL FAMREL ERRP RRP ERRP RRP RRP RRPU ENTID-PNSP EPNSPOUS PNSP EPNSPOUS PNSP EPNS ENTID-PNPT PNPT PNPT PNPT EPNMOM EPNMOM EPNDAD EPNDAD Table 11-12 Table 12-11 EPNGUARD PNGDU EPNGUARD Table 10-8 Table 10-8 Table 11-12 (table continues) 9-13 SIPP USERS’ GUIDE SIPP PUBLIC USE FILES Table 9-2. Differences Among Core Wave, Topical Module, and Longitudinal Files (1990-2008 Panels) (continued) 1996+ Panels Core Wave Files Pre-1996 Core Wave Files 1996+ Panels Pre-1996 Pre-1996 Full Panel Files Authorized Person - Level Authorized Person -Level Topical Topical Coverage Authorized Person - Level Topic Coverage Recipient Amount Coverage Recipient Amount Module File Module Files Recipient Amount Identifying ProgramUnits Sourcesare identifiedin Social RCUTYP01 RCUOWN01 T01AMTA SOCSEC SSPNUM S01AMTA SOCSEC SS-PIDX G1SRC1- G1SRC10. Security T01AMTK S01AMTK Amountsarelocatedin Railroad NA T02AMT RAILRD RRPNUM SO2AMTA Notintopical Notintopical RAILRD RR-PIDX themonthlyarrays S02AMTK modulefiles modulefiles G1MT1– GIAM10 Fed SSI RCUTYP03 RCUOWN03 T03AMTA SSICOVRG S03AMT T03AMTK Veteran’s Admin. RCUTYP08 RCUOWN08 T08AMT VETS VETNUM S08AMT VETS VA-PIDX AFDC/TAIF RCUTYP20 RCUOWN20 T20AMT AFDC AFDCPNUM S20AMT AFDC AFDCPDIX General Assistance RCUTYP21 RCUOWN21 T21AMT GENASST GAPNUM S21AMT GENASST GA-PIDX FosterChild Care RCUTYP23 RCUOWN23 T23AMT FOSTKID FKPNUM S23AMT FOSTKID FOSTPIDX OtherWelfare RCUTYP24 CUOWN24 T24MT OTHWELF OWPNUM S24AMT OTH-WELF OTH-PIDX WIC RCUTYP25 RCUOWN25 T25MT WICCOV WICPNUM WICVAL WICCOV WIC-PIDX Food Stamps RCUTYP27 RCUOWN27 T27AMT FOODSTMP FSPNUM S27AMT FOODSTMP FS-PIDX Medicare ECRMTH CARECOV MCDPNUM CARECOV Medicaid RCUTYP57 RCUOWN57 CAIDCOV CAIDCOV CHAMPUSor RCHAPM CHAMP CHPNUM CHAMP CHAMPVA RCUTYP58 HIND Tables Tables12-19 Health Insurance Table10-16 RCUOWN58 10-17and10-18 HIPNUM and12-20 (table continues) 9-14 SIPP USERS’ GUIDE DATA EDITING AND IMPUTATIO N Table 9-2. Differences Among Core Wave, Topical Module, and Longitudinal Files (1990-2008 Panels) (continued) Topic 1996+ Panels Topical Pre-1996 Core Wave 1996+ Panels Topical Pre-1996 Topical Pre-1996 Longitudinal Wave Files Files Module Files Module Files Files Imputed Data: The whole record If no prior wave data and If MIS5 = 2 and MISj = 1 for j If EPPMISA = 2 or EPPINTVW If PP-MIS5 = 2 and PP-MISj If WAVFLG > 0 or INTVW = is imputed EPPINTVW = 3, 4 = 1, 2, 3, 4 or INTVW = 3, 4 = 3, 4 = 1 for j = 1, 2, 3, 4 or 3, 4 INTVW = 3, 4 The corresponding If the corresponding If the corresponding If the corresponding imputation If the corresponding If the corresponding imputation wave of imputation flag indicates imputation flag indicates flag and calculation flags imputation flag and calculation flag and calculation flags information is imputation imputation. indicate imputation. flags indicate imputation. indicate imputation imputed The variable’s Almost all person-level Almost all person-level Most person-level variables Most person-level variables Limited set of imputation flags. value is imputed. variables have imputation variables have imputation have imputation flags. There are have imputation flags. There There are no household a nd flags. There are no imputation flags. There are no imputation no household and family are no household and family family aggregates. Use the flags on household and family flags on household and family aggregates. Use the person-level aggregates. Use the person-level imputation aggregates. Use the person- aggregates. Use the imputation flags of household person-level imputation flags flags of household and level imputation flags of person-level imputation flags and family members to identify of household and family household and family of household and family aggregate amounts that include members to identify aggregate family members to identify members to identify aggregate members to identify aggregate imputed values. amounts that include imputed aggregate amounts that amounts that include imputed amounts that include imputed values. include imputed values values. values. Topcoding Yes Yes Yes Yes Yes How to Identify TFIPSST HSTATE TFIPSST STATE GEO-STE Sates Weight Variables Household WHFNWGT HWGT, H5WGT Family WFFINWGT FWGT Subfamily WSFINWGT SWGT FNLWGTyy, where yy is the Person WPFINWGT FNLWGT, P5WGT WPFINWGT FINALWGT calendar year PNLWGT Metropolitan Areas TMETRO (for 1996 - 2008) HMETRO Not on the file Not on the file Not on the file TMSA for (for 1996) 9-15