A Panel Study of Income Dynamics: 1985-1999 Marriage History File Documentation (Version 1) Survey Research Center Institute for Social Research The University of Michigan Ann Arbor, Michigan 2001 TABLE OF CONTENTS Section I: Overview of the 1985-1999 Marriage History File Individuals for Whom the Data are Available Background for the Marriage History Files How to Obtain the File and Whom to Contact About Questions Section II: Questionnaire Detail Section III: Structure of the File Number of Records Sort Order of the File Variables on the File Section IV: Idiosyncracies, File Cleaning and Variable Detail How to Identify Individuals Who Were Never Married How to Identify Individuals For Whom Marriage Data Were Not Ascertained Treatment of Individuals Who Become Non-response Idiosyncracies, Data Cleaning and Treatment of Incomplete or Inconsistent Information Who has Cross-Year Information What Cross-Year File to Use for Merging Marriage Order Section V: Linking Records Using the Marriage History File in Conjunction with the Cross-year Individual File Using the Marriage History File with Other Demographic History Files Section VI: Marriage Information Available on the Individual File Section VII: Codebook Section I: Overview of the 1985-1999 Marriage History File The 1985-1999 Marriage History File is designed to facilitate access to information collected in the 1985 through 1999 waves of the Panel Study of Income Dynamics (PSID) regarding retrospective histories of marriages. This file contains details about marriage events of eligible people living in a PSID family at the time of the interview in any wave between 1985 and 1999. The PSID definition of "eligible" is described in detail below. Each set of records for a specified individual contains all known cumulative data about the timing and circumstances of his or her marriages up to and including 1999, or those waves during that period when the individual was in a responding family unit. If an individual has never been married, one record indicates that report. The number of variables are few; they contain relevant identifiers for the individual and his or her spouse, dates of marriages/divorce/widowhood, the order of the specific marriage, total number of marriages and the most recent wave in which data were collected. The 1985-1999 Marriage History File stores information in an efficient manner that allows a high degree of flexibility in linking with the PSID's Individual File. It is designed to be linked to the 1968-1999 Individual File. Linkages can be done from either the individual's or the spouse's standpoint. These linkages are more limited for spouses than for individuals, since some spouses have never been part of the PSID and hence have no record on the 1968-1999 Individual File. Data users who want only some of the detail of the marriage data will find marriage history information summarized on Public Release II versions of the PSID Individual files in several individual-level summary variables. These variables include number of marriages, beginning and ending dates for the first and the most recent marriages, marital status of the individual at the time of the most recent interview, and the most recent wave when marriage data were collected. For particulars, see Section VI below. Individuals for Whom the Data are Available The marriage history data were collected about individuals of marriage-eligible age in responding PSID families, i.e., those with values of 1-20 for the "Sequence Number" variable in a given wave. In waves when individuals were non-response or in an institution, no information was collected about them. The types of individuals asked for marriage history information and the detail gathered about their history differed over the waves from 1985 through 1999; 1985 followed one pattern, and 1986 through 1999 followed another. To keep respondent burden to a minimum and data quality high, different question sequences about these events have been used for PSID individuals depending on their circumstances. Marriage history information was gathered as described below: (a) in the 1985 wave, a complete retrospective marriage history was asked of a Head, Wife, or "Wife" of any age; (b) in all succeeding waves, marriage history was updated for changes since the beginning of the prior calendar year for a Head, Wife, or "Wife" of any age who was also a Head, Wife, or "Wife" in the prior wave's interview; (c) in succeeding waves, details about first and current or most recent marriages were asked for a New Head, New Wife, or New "Wife" of any age; (d) in all waves from 1985 through the present, details about first and current or most recent marriages were asked for an Other Family Unit Member (OFUM) aged 12-44 at the time of the interview. These latter two groups, although initially asked about only first and last marriages, may have additional marriage records on the file if those marriages occurred while the individual was in a responding family. A number of complexities in the overall study design present special challenges for collecting and processing the demographic history data: (a) In any wave of the PSID some family members appear in the study for the first time, whereas most are people who have been participating for years. (b) From one wave to the next a PSID individual can enter or leave eligibility for being asked marital or childbirth histories by passing the threshold ages for these questions. For example, the entry age for eligibility is 12, and for family members other than Head, Wife, or "Wife" the exit threshold for eligibility is 45. (c) A PSID individual can change his or her relationship to Head from one wave to the next, and this can affect whether the demographic event-history information is self-reported or proxy-reported by a parent or by some other relative. (d) From one wave to the next the range of demographic events asked about a given individual can expand or contract--information about adoptions is gathered for Heads, Wives, and "Wives" but not other family members. (e) A PSID individual can become non-response, after which time demographic event history information is not updated. (f) While both Heads and Wives/"Wives" were interviewed in 1985, only one person (usually the Head) has been the respondent in each year since then. In the 1990 wave, a sample of 2,043 Latino households was added to the PSID sample to provide the representative information about Latinos that is now available for blacks and non-Latino whites in the original PSID. The Latino sample size was increased to 2,308 households in 1992. This sample was dropped from the PSID in 1996. The marital history data of the Latino sample are also included in this file. In the 1992 interview, the PSID undertook a pilot effort to recontact former respondents who had attrited from the study and persuade them to rejoin. Additional information regarding the 1992 recontact samples is discussed in the 1992 PSID Documentation, pp. 1-3, but briefly, four types of recontact samples were identified for inclusion in this round of interviewing. They differed in the detail gathered about their marriage history. (a) 1992 Reinterview recontact families were last interviewed during the 1990 wave, but attrited in 1991. (b) 1992 Splitoff recontacts were sample members who moved out of a responding 1991 family but who were not interviewed themselves in 1991. (c) 1992 Sample recontacts were designated as members of original 1968 Panel families who were last interviewed at some time between 1969 and 1989. (d) 1992 Latino recontacts were families in the original Latino National Political Survey sample, but either refused or were never contacted by the PSID during the 1990 wave when the Latino sample was added to the study. Splitoff, Sample, and Latino recontacts (the last three types described above), like all other New Heads/Wives/"Wives", were asked questions detailing their first and most recent marriages. The marriage history of Reinterview recontacts (the first type above) was updated for changes in 1991 and 1992. Since this latter group of people was interviewed in 1990, but not 1991, information was lost for events occurring between the 1990 interview and January of 1991. The unaccounted-for time varies from a month or two up to a maximum of nine to ten months. Reinterview recontacts can be identified on the cross-year Individual Files by data values of 2 or 3 for "1992 Follow Status" (V30799/ER30799). In 1993, 1994, and from 1997 onward the PSID also recontacted attriters, but these individuals were asked the same retrospective marriage histories as persons new to the study. That is, reports of their first and current or most recent marriages were sought. Background for the Marriage History Files The 1985-1999 Marriage History File originated with the 1985 collection of comprehensive, retrospective questions about a number of demographic events, including childbirth, adoption, marriage, separation, divorce and substitute parenting. In each wave from 1986 through the present, these histories, excepting substitute parenting, were updated for eligible individuals. All the retrospective data collected in 1985 on these demographic phenomena are included in the 1985 Ego-Alter File. This file is mostly of interest for the substitute parenting information and for child-specific information on public program participation and health care surrounding a birth. These sets of questions were discontinued after the initial retrospective in 1985. For more details about the 1985 Ego-Alter File, refer to A Panel Study of Income Dynamics: 1985 Ego-Alter File Documentation. Beginning in 1986, we decided to release the demographic history data annually as two separate files: the Childbirth and Adoption History File and the Marriage History File. These files are cumulative, and so their size increases each year as more events happen and additional people become eligible. How to Obtain the File and Whom to Contact About Questions The 1985-1999 Marriage History File is available through the Inter- university Consortium for Political and Social Research (ICPSR). It can be requested by contacting the ICPSR Member Services: e-mail: netmail@icpsr.umich.edu main website: http://www.icpsr.umich.edu/ telephone: (734) 763-5010 This file and documentation are also available on the Internet. The URL (location) of the PSID home page is: http://www.isr.umich.edu/src/psid/ If you have questions that are not answered by this documentation, you can contact PSID staff through our website; click on "Contact Us!" on our home page. Section II: Questionnaire Detail The flow of the PSID questionnaire is complex. As described above in Section I, the types of individuals asked for history information and the detail gathered about their history changed. The analyst may find it helpful to actually see the questionnaires. The 1985 sequence is available in the PDF format version of the main 1985 documentation volume on pp. 23-31 and pp. 58-68, located on our website. The 1986-1999 sequences are identical to each other, with the addition, beginning in 1997, of questions about birth location, race and ethnicity. PDF format versions of the 1991-1996 main questionnaires are also available on our website. They contain the questions for Head/Wife/"Wife" marriage history, but do not include the supplementary forms used for details about childbirth and adoptions, or for OFUMs' marriages and childbirths. These supplements are not available on the website as of this writing. The website also includes Computer-Assisted Interview (CAI) Documentation for 1993 forward, and this version of the PSID interview schedule does incorporate the specifics of the questions asked about each spouse and the questions asked for OFUMS. Section III: Structure of the File Number of Records The 1985-1999 Marriage History File contains a total of 43,220 records. This file has a one-record-per-marriage general structure. Each record contains information for a specified marriage for an individual, or information that indicates the individual has no marriage data. Information for an individual is current as of the most recent wave that marriage history was collected for him or her. Multiple records for a given individual result from an individual having more than one marriage. Although the individual remains the same in such circumstances, the spouse differs from one record to the next. The maximum number of marriage records for a given individual is eight on the 1985-1999 Marriage History File. A few multiple records for a given spouse also exist. This occurs when an individual has remarried the same person. In situations where two individuals were married to each other twice and were both present in a responding family unit, the file has four records for this pair, two records for each of the two individuals. Sort Order of the File The 1985-1999 Marriage History File is sorted, in ascending order, by "1968 Interview Number of Individual" (V1), "Person Number of Individual" (V2), and "Order of This Marriage" (V8). Details for an individual's first marriage are followed by those for his or her second, third, etc., or the most recent marriage. As noted in Section I, complete marriage histories were gathered for Heads, Wives and "Wives" in 1985, but information on only first and most recent marriages was initially collected for New Heads, Wives, and "Wives" in 1986 through 1999 and OFUMs in any wave from 1985 forward. All records for the first eligible member of a 1968 PSID family are followed by those for the next eligible member in the same family of origin. When all of the records for all eligible members in the first family are exhausted, records for eligible members in the second family follow. Variables on the File The 1985-1999 Marriage History File contains twenty variables, which can be categorized into those relevant to the individual, those specific to his or her spouse, those in regard to the marriage in question, and some aggregate information about the marriage history for the individual. Information specific to the individual: V1 1968 Interview Number of Individual V2 Person Number of Individual V3 Sex of Individual V4 Month Individual Born V5 Year Individual Born Information specific to the spouse: V6 1968 Interview Number of Spouse V7 Person Number of Spouse Information specific to the marriage: V8 Order of This Marriage V9 Month Married V10 Year Married V11 Status of This Marriage V12 Month Widowed or Divorced V13 Year Widowed or Divorced V14 Month Separated V15 Year Separated Aggregate Information for the marriage history of this individual: V16 Year Most Recently Reported Marriage V17 Number of Marriages of This Individual V18 Last Known Marital Status V19 Number of Marriage Records V20 Release Number Section IV: Idiosyncracies, File Cleaning and Variable Detail Several aspects of the 1985-1999 Marriage History File merit particular attention. This section discusses what they are and how to handle them. How to Identify Individuals Who Were Never Married One caution, particularly relevant to event-history analysis, concerns the records for individuals who have never married. The file has one record for each such person. On these records, the "Number of Marriages" variable (V17) has a value of zero. Codes indicating "Inapplicable" (9s) are padded in the fields for marriage details, with the exception of Spouse ID (V6) and Spouse Person Number (V7); both of these variables have values of zero. How to Identify Individuals For Whom No Marriage Data Were Ascertained Persons who may or may not have married but for whom the PSID has been unable to determine anything relating to his or her marriage situation, not even marital status, also have one record on this file. On this type of record, the "Number of Marriages" variable (V17) has a value of 98, although this value is not unique to such individuals. Missing data codes (8s or 98s) are padded in all the fields for that record, with the exception of Spouse 1968 ID (V6) and Spouse Person Number (V7). These two variables contain values of 9s, which do uniquely identify this sort of record. Treatment of Individuals Who Become Non-response The Marriage History File is cumulative. That is, all individuals who have ever been eligible for the marriage history question sequences since it was first begun in 1985 have at least one record on the file. Thus, each new version is current through the most recent wave for individuals in responding families, but the data are up to date only through the last year that non-response individuals were living in a responding family. The marriage history is current through 1999 for those in a responding PSID family at the time of the 1999 interview and who are otherwise eligible for marriage history questions. For those who were non-response in the 1999 wave or who are no longer eligible for the questions, the history is current through the last year they were in an interviewed family unit and eligible. For example, if an individual became non-response for the 1988 wave and has not returned to a PSID family, his or her marriage history is current only through 1987. Similarly, if an OFUM who has remained in a responding family is now 47 years old, his or her marriage history has not been updated in the last few waves, since OFUMs' marriage information is not collected once they achieve 45 years of age. The variable indicating recency of an individual's marriage reports is V16. Idiosyncracies, Data Cleaning and Treatment of Incomplete or Inconsistent Information We have tried very hard to assure access to all available information while also recording occurrences of missing data or unclear identification of spouses. In some situations, the individual was reported to have married, i.e., his or her marital status is known to be divorced, widowed, separated, or currently married, but details about the marriage or the identification of the spouse were not reported. PSID staff can and do assign an identifier to such a spouse, as it is clear that the spouse has never been part of the study. Often the same event (e.g., a divorce) is reported in more than one successive interview. In such cases, the initial report is chosen as the source for the associated date of the event (e.g., month and year of a divorce) unless the date was not ascertained. In that event, the succeeding year's reported date was used on the principle that known dates are better than missing dates. Individuals from whom we have reports of their marriages to each other may disagree on the status. This seeming inconsistency can be legitimate if the currentness of the spouses' reports differs. The variable indicating the wave in which the marriage history was most recently updated is V16. As an example, in 1990 a female sample member marries, and her new husband- head moves into the study. In that wave we receive reports of their mutual marriage, and each of them has a record on the Marriage History File registering the other as spouse. They each receive values of 1990 for V16. In each successive wave their marriage is reconfirmed, and values for V16 are updated until 1994, when they separate and divorce. He, the non-sample husband and former head, leaves the PSID but the sample wife remains response. The record for her marriage to him is updated to indicate the revised status, the dates of separation and divorce are added, and V16 receives a value of 1994. But his record is not updated; his status remains married, and V16 retains a value of 1993. The preparation of the 1985-1999 file involved a great effort to eliminate real inconsistencies. If a couple no longer living together disagreed on their marital status but both were responding at the same time, PSID staff attempted to reconcile the differences. Marriage and birth dates were cross-checked to ensure that marriages do not occur until an individual is at least 13 years old. Eleven marriages remain in which the individual reports a start date before that age. We are unable to resolve these cases. In most of them, marriage at age 11 or 12 is possible. Five of these persons are females from the 1990 Latino sample, and two are from the 1997 immigrant sample. For the latter, subsequent birth date information may clarify matters. On a related note, an individual's birth date is copied from the current wave of the Individual File when his or her marriages first appear on the cumulative Marriage History File. This birth date is not updated in later years for the initial record(s). If a new marriage subsequently occurs for the individual, then his or her birth date from that later wave is used in the construction of the new record. Thus, birth dates for the same individual may disagree across marriages if the reports were not collected in the same wave. These discrepancies were cleaned to some extent, but only as a by-product of other cleaning activities. Attempting birth date consistency is made difficult because of individuals' varying years of participation in the study. In the longer term, we hope to clean these dates, but for the present we advise analysts to use the individual's birth date or age variables from the most recent year of Individual File data for which he or she is present in the study. We checked the internal consistency of marriage dates: termination dates must not precede marriage dates, and an earlier marriage must end by the time a later marriage begins. All cases in which divorce dates preceded separation dates were checked for coding/data entry errors and against corroborating sources (e.g., if the spouse had ever been in the study, his or her date of move out was compared to the separation date). In 60 cases, indeed the couples did not separate until after their divorces were finalized. In 31 cases, we were unable to resolve the final status of a person's earlier marriage. These individuals appear to be bigamists, and probably are. The earlier marriage records have values of 7 for the status variable (V11). Yearly fluctuations in status were reconciled as information from each successive wave was incorporated. For example, divorcees have been known to reported themselves as widowed after their ex-spouses' deaths; their statuses were recoded to divorced. Individuals who separate and then resume living together were recoded as married. The fact of their former separation must of necessity disappear from the file. Interruptions of this sort can be found by comparing the couple's data records on the cross- year individual file for co-residence, that is, comparing their yearly family interview numbers and sequence numbers (e.g., for 1990, V30642/ER30642 and V30643/ER30643). And every attempt was made to reconcile differing reports of status between couples if they were present in the same wave. Spouse identifiers were checked against cross-year individual data and against the 1985-1999 Childbirth and Adoption History File to ensure that no spurious cases of intergenerational incest occur. If both spouses had been in the study, their genders from the cross-year individual file were checked against each other. Our final checks assured that all individuals who had ever qualified for marriage history questions had records on the file, and that individuals who had never qualified did not. Who has Cross-Year Information Please keep in mind another PSID intricacy when matching across files: while all individuals were present in a PSID family, many spouses identified in the marriage histories have never been present in a PSID family unit during the years the study has been in progress; these spouses have values for "Person Number of Spouse" (V7) in the range 900-996. Consequently, each individual has been in a PSID family unit and has a record on the 1968-1999 Individual File, but his or her spouse may or may not. What Cross-Year File to Use for Merging The 1985-1999 Marriage History File matches the 1968-1999 Public Release I Individual File exactly. If you attempt to merge the 1985-1999 Marriage History File with earlier individual data releases, some cases on the merged file will NOT match and may result in the assignment of some erroneous spousal relationships! See the next section for details. Marriage Order As described in Section I, in the initial wave of demographic event history collection in 1985, all Heads, Wives and "Wives" were asked to provide details about all of their marriages. But in subsequent waves, the retrospective marriage history questions for new Heads, Wives and "Wives" permitted detailed information about only two marriages, the first and the current or most recent. OFUMS were never asked about all their marriages; even in 1985 we requested reports about only the first and current or most recent. Even so, the Marriage History File contains complete histories for most individuals, since more than two marriages is a relatively rare event. The number of individuals reporting more than two marriages is 1007; 589 of them have reported all their marriages, but 418 have not. The order of each marriage for an individidual is indicated in V8; values are assigned to each marriage in chronological order. The 418 individuals for whom we have received incomplete reports have gaps in the values for marriage order across their records. For example, if a person has been married three times but we have received detailed information about only the first and last marriages, the first marriage is assigned a value of 1 and the last marriage receives a value of 3. No record for marriage number 2 is on the file. These individuals can easily be identified, as values for the number of their marriages (V17) are greater than values for the number of their records (V19), of course excluding cases where the number of marriages contains missing data. The relative order of marriages is always clear because of the way in which the questions are asked and updated. A missing beginning date for a marriage does not cause its order to be unknown. However, 260 individuals have a marriage of unknown order. This has happened in two different circumstances. If the interviewer did not obtain a complete marriage history when the individual entered the study but he or she was known then to be or have been married, we created a record for that current or most recent marriage with the information available. This situation accounts for the overwhelming majority of the cases (257). These persons have a known spouse at V6-V7, a known marital status at V11, but the number of their marriages is not ascertained (V17=98) and the number of records for them (V19) equals 1. The second, much rarer circumstance in which a marriage is known to have taken place but its order is not ascertained occurs when the total number of marriages is not ascertained, but the person reports a first and a last. Only three individuals on the file fit this profile. Their first marriage receives an order value of 1, of course, but the last must of necessity have a value of 98. At least through the 1999 wave, however, no person has more than one marriage with order unknown. Section V: Linking Records Using the Marriage History File in Conjunction with the Cross-year Individual File The Marriage History File is designed to be linked to PSID Individual data for analysis purposes. The Marriage History File has only a modest amount of information about the individual and his or her spouse(s). The analyst will no doubt want to access the much greater volume of data available for these individuals on the Family and Individual Files. Those files can provide abundant information for many individuals dating back to 1968 when the PSID began, although for some the span of available data is more limited or nonexistent. When matching the 1985-1999 Marriage History File to individual data, only the 1968-1999 Individual File should be used. During file merging, a number of unique individual identifiers were corrected. Special care was taken to ensure perfect correspondence in individual identifiers between this file, the 1985-1999 Childbirth and Adoption History File, and Public Release I of the 1968-1999 Individual File. Because of the corrections, do not attempt to match this file with any other PSID data file or a subset derived therefrom using these unique individual identifiers. Non-matches in individual records definitely will occur if any other file besides the above-named is used. Data processing is required to link records between these two files. To achieve linkages, one must match on the unique individual-specific identifier. This unique identifier is a combination of two variables: "1968 Family Interview Number" (V1) and "Person Number" (V2). The corresponding variables for this unique identifier on the cross-year Individual File are V30001 and V30002 (or ER30001 and ER30002). Care must be taken with regard to the proper files to use, the choice of individuals on the Marriage History File for whom matches are attempted, and the years for which data are available. Individuals and spouses vary substantially in terms of which years they have been present in PSID family units over the course of the study. This affects the availability of data for them on the main files because valid information is obtainable on those files only in the years that the individual is present in a PSID family unit ("present" means living in the family unit or having left it to enter an institution). For more details about PSID tracking procedures and classification of people into family units, see the discussion in Section 8 on "Family Composition and Change", p. 55 in A Panel Study of Income Dynamics: A User's Guide by Martha Hill (Sage Publications, 1992). First there is the matter of whether any record exists for an individual. If a record does exist, then the question is in which years of the study are data available for that individual and his or her family. If a person, either individual or spouse, has a record on the 1968-1999 Individual File but less than the full range (31 waves) of data in that record, variables in the waves when he or she was not present in a PSID family unit are, for the most part, filled with zeros. In addition, the annual individual-level variable "Type of Individual Data Record" in those years indicates that he or she is non-response. As noted among the cautions in Section IV, all eligible individuals have records on the 1985-1999 Individual File because they were present in a PSID family unit during that time period. However, many spouses do not have records on that file, since indeed they have never been in a PSID family at any time during the course of the study. Spouses who have been present in a PSID family unit at some time since the study began have records on the 1985-1999 Individual File. Values for their Person Numbers are in the range 001-228. Spouses who have never been present have Person Number values in the range 900-996. Even though all individuals have records on the 1968-1999 Individual File, that file does not necessarily contain data for all of them for all years of the study. Some of them first entered the study in, e.g., 1985, and thus only have data since that year. Others have logged as many as thirty- two years of inclusion in the study. Similarly, there is considerable variation among spouses regarding which years, if any, they were included in the study. For those persons with records on both the Individual File and the Marriage History File, linkages rely on a match of individual-specific identifier variables that appear on both files. As noted earlier, the unique identifier involves two variables: "1968 Family Interview Number" and "Person Number". The corresponding set of variables on the two files must match to properly link an individual's records from the two sources. The variable names for these two variables are V30001 and V30002 (or ER30001 and ER30002) on the Individual File. On the Marriage History File, they are V1 and V2, respectively, for the person designated "individual" and V6 and V7 for the person designated "spouse". Note that such linkages involve a one-to-many-match. One record on the Individual File may have more than one matching record on the Marriage History File because the specified individual has multiple marriages. Using the Marriage History File with Other Demographic History Files Some analysts may be interested in linking information from different records on the Marriage History File, or linking information from records on different demographic files. For example, access to both marriage and childbirth records for an individual are needed to determine, via comparisons of marriage and childbirth dates, the number of biological children an individual has when he or she remarries. To make links such as these, one must match on the unique individual-specific identifier, which is a combination of two variables -- "1968 Family Interview Number" and "Person Number" (V1 and V2 for the person designated as the individual on the 1985-1999 Marriage History File; V2 and V3 for the person designated as the parent on the 1985-1999 Childbirth and Adoption History File). Section VI: Marriage Information Available on Individual Files Some of the information provided on Marriage History Files is also available on Public Release II versions of the cross-year Individual File. In addition, the Individual File contains some detail relating to marriage issues that the Marriage History File does not. The following listing shows all of the marriage history related variables included in Public Release II versions of the cross-year Individual File. All are individual-level variables. V32033 Year Marital Info Most Recently Updated V32034 # Marriages of This Individual V32035 V32036 Month and Year First/Only Marriage Began V32037 Status of First/Only Marriage V32038 V32039 Month and Year First/Only Marriage Ended V32040 V32041 Month and Year Separated First/Only Marriage V32042 V32043 Month and Year Most Recent Marriage Began V32044 Status of Most Recent Marriage V32045 V32046 Month and Year Most Recent Marriage Ended V32047 V32048 Month and Year Separated Most Recent Marriage V32049 Last Known Marital Status These variables are compiled from marriage history information collected from the 1985 wave through the most current wave of final-release individual information included on the file. Section VII: Codebook The item-by-item descriptions for all of the variables in the Marriage History File follow. Unweighted frequencies were calculated for each variable. The PSID: 1985-1999 Marriage History File Codebook Generated on 22JUN2001 at 14:01:16. _________________________________________________________________________________ Variable List V1 1968 INTERVIEW NUMBER OF INDIVIDUAL V2 PERSON NUMBER OF INDIVIDUAL V3 SEX OF INDIVIDUAL V4 MONTH INDIVIDUAL BORN V5 YEAR INDIVIDUAL BORN V6 1968 INTERVIEW NUMBER OF SPOUSE V7 PERSON NUMBER OF SPOUSE V8 ORDER OF THIS MARRIAGE V9 MONTH MARRIED V10 YEAR MARRIED V11 STATUS OF THIS MARRIAGE V12 MONTH WIDOWED OR DIVORCED V13 YEAR WIDOWED OR DIVORCED V14 MONTH SEPARATED V15 YEAR SEPARATED V16 YEAR MOST RECENTLY REPORTED MARRIAGE V17 NUMBER OF MARRIAGES OF THIS INDIVIDUAL V18 LAST KNOWN MARITAL STATUS V19 NUMBER OF MARRIAGE RECORDS V20 RELEASE NUMBER _________________________________________________________________________________ V1 1968 INTERVIEW NUMBER OF INDIVIDUAL Number: 1 Type: Num Width: 4 Decimals: 0 Location: 1-4 1968 Interview Number (1968 ID Number) of the Individual This variable is the identification number for the 1968 Family Unit with which the individual is associated, whether or not he or she actually resided in that Family Unit. The combination of this variable and the individual's Person Number (V2) provides unique identification for him or her. These two variables are the same identifiers as those on the cross-year Individual File. Ranges of values are also meaningful, as indicated below. 17746 1-2930 Individual is associated with a family from the SRC cross-section sample. 1485 3001-3511 Individual is associated with a family from the New Immigrant sample. 15778 5001-6872 Individual is associated with a family from the Census sample. 8211 7001-9308 Individual is associated with a family from the Latino sample. _________________________________________________________________________________ V2 PERSON NUMBER OF INDIVIDUAL Number: 2 Type: Num Width: 3 Decimals: 0 Location: 5-7 Person Number of the Individual This variable, in combination with 1968 Interview Number (V1), provides a unique identifier for the individual throughout the PSID. Ranges of values are also meaningful, as indicated below. 22035 1-19 Individual who was living in an original core, Latino, or New Immigrant sample family at the time of the initial interview 6 20 Individual who was spouse of the original core, Latino, or New Immigrant sample Head and was living in an institution at the time of the initial interview 336 21-26 Individual who was a child or stepchild of the original core, Latino, or New Immigrant sample Head and was under 25 and in an institution at the time of the initial interview 6099 30-169 Individual who has at least one sample parent and who was born into the sample since the original core, Latino, or New Immigrant interview 14735 170-226 Individual who has no sample parent, or who moved into the sample for the first time after the original core, Latino, or New Immigrant interview 9 227 Individual who was spouse of an original core, Latino, or New Immigrant sample Head and who moved out in the year prior to the initial interview _________________________________________________________________________________ V3 SEX OF INDIVIDUAL Number: 3 Type: Num Width: 1 Decimals: 0 Location: 8 Sex of Individual 20967 1 Male 22253 2 Female _________________________________________________________________________________ V4 MONTH INDIVIDUAL BORN Number: 4 Type: Num Width: 2 Decimals: 0 Location: 9-10 Individual's Month of Birth Note that the individual's birth date on this file may be inconsistent with his or her age and birth date on the PSID main files. See Section IV: Idiosyncracies, Data Cleaning and Variable Details for more information. 3798 1 January; winter 3295 2 February 3593 3 March 3258 4 April; spring 3416 5 May 3401 6 June 3649 7 July; summer 3771 8 August 3718 9 September 3680 10 October; fall; autumn 3521 11 November 3622 12 December 498 98 NA; DK Missing Data: 98 _________________________________________________________________________________ V5 YEAR INDIVIDUAL BORN Number: 5 Type: Num Width: 4 Decimals: 0 Location: 11-14 Individual's Year of Birth Note that the individual's birth date on this file may be inconsistent with his or her age and birth date on the PSID main files. See Section IV: Idiosyncracies, Data Cleaning and Variable Details for more information. 42738 1884-1987 Year born 482 9998 NA; DK Missing Data: 9998 _________________________________________________________________________________ V6 1968 INTERVIEW NUMBER OF SPOUSE Number: 6 Type: Num Width: 4 Decimals: 0 Location: 15-18 1968 Interview Number (1968 ID Number) of the Spouse This variable is the identification number for the 1968 Family Unit with which the spouse is associated, whether or not he or she actually resided in that Family Unit during the course of the PSID. The combination of this variable and the spouse's Person Number (V7) provides unique identification for him or her. These two variables are the same identifiers as those on the cross-year Individual File. Ranges of values are also meaningful, as indicated below. 14544 0 Inap.: no spouse; never married (V17=0) 13128 1-2930 Spouse was associated with a family from the SRC cross-section sample. 900 3001-3511 Spouse was associated with a family from the New Immigrant sample. 9381 5001-6872 Spouse was associated with a family from the Census sample. 5139 7001-9308 Spouse was associated witha family from the Latino sample. 128 9999 NA or DK whether ever married Missing Data: 0,9999 _________________________________________________________________________________ V7 PERSON NUMBER OF SPOUSE Number: 7 Type: Num Width: 3 Decimals: 0 Location: 19-21 Person Number of the Spouse This variable, in combination with the spouse's 1968 Interview Number (V6), provides a unique identifier for him or her. Ranges of values are also meaningful, as indicated below. 14544 0 Inap.: never married (V17=0) 12570 1-19 Spouse who was living in an original core, Latino or New Immigrant sample family at the time of the initial interview 10 20 Spouse who was spouse of the original core, Latino, or New Immigrant sample Head and was living in an institution at the time of the initial interview 188 21-26 Spouse who was a child or stepchild of the original core, Latino, or New Immigrant sample Head and was under age 25 and in an institution at the time of initial interview 784 30-169 Spouse who has at least one sample parent and who was born into the sample since the original core, Latino, or New Immigrant interview 8207 170-226 Spouse who has no sample parent, or who moved into the sample for the first time after the original core, Latino, or New Immigrant interview 38 227 Spouse who was spouse of an original core, Latino, or New Immigrant sample Head and who moved out in the year prior to the initial interview 21 228 Spouse who was spouse of original core, Latino, or New Immigrant sample Head and who died in the year prior to the initial interview 6730 900-990 Spouse who has never been in any sample family 128 999 NA or DK whether ever married Missing Data: 0,999 _________________________________________________________________________________ V8 ORDER OF THIS MARRIAGE Number: 8 Type: Num Width: 2 Decimals: 0 Location: 22-23 Order of This Marriage This variable indicates the order of this marriage relative to the individual's other marriages, from earliest to most recent. Within records for an individual, gaps in the values for this variable may exist because detailed marriage histories are incomplete for some individuals. See "Marriage Order" in Section IV of the front matter for a larger discussion. 22229 1 First marriage 4914 2 Second marriage 937 3 Third marriage 148 4 Fourth marriage 40 5 Fifth marriage 10 6 Sixth marriage 9 7 Seventh marriage 1 8 Eighth marriage 387 98 NA; DK 14545 99 Inap.: never married (V17=0) Missing Data: 98,99 _________________________________________________________________________________ V9 MONTH MARRIED Number: 9 Type: Num Width: 2 Decimals: 0 Location: 24-25 Month in which this Marriage Occurred 1638 1 January 1654 2 February 1646 3 March 1882 4 April 2183 5 May 3077 6 June 2286 7 July 2591 8 August 2246 9 September 1974 10 October 1993 11 November 2209 12 December 7 21 Winter 16 22 Spring 35 23 Summer 9 24 Fall; autumn 3229 98 NA; DK 14545 99 Inap.: never married (V17=0) Missing Data: 98,99 _________________________________________________________________________________ V10 YEAR MARRIED Number: 10 Type: Num Width: 4 Decimals: 0 Location: 26-29 Year in which this Marriage Occurred 27017 1901-1999 Year married 1658 9998 NA; DK 14545 9999 Inap.: never married (V17=0) Missing Data: 9998,9999 _________________________________________________________________________________ V11 STATUS OF THIS MARRIAGE Number: 11 Type: Num Width: 1 Decimals: 0 Location: 30 Status of This Marriage This variable indicates the status of this marriage in the wave in which marriage history was most recently collected for this individual (V16). 16407 1 Marriage intact 2324 3 Marriage ended in widowhood 8564 4 Marriage ended in divorce or annulment 1214 5 Spouses separated 31 7 Other: more recent marriage also reported, but no evidence of divorce or widowhood. Individual is probably a bigamist. 135 8 NA; DK 14545 9 Inap.: never married (V17=0) Missing Data: 8,9 _________________________________________________________________________________ V12 MONTH WIDOWED OR DIVORCED Number: 12 Type: Num Width: 2 Decimals: 0 Location: 31-32 Month in which This Marriage Ended in Widowhood or Divorce 733 1 January 605 2 February 733 3 March 616 4 April 662 5 May 772 6 June 640 7 July 670 8 August 616 9 September 647 10 October 611 11 November 663 12 December 21 21 Winter 33 22 Spring 52 23 Summer 28 24 Fall; autumn 2921 98 NA or DK month; NA or DK status of marriage (V11=8) 32197 99 Inap.: marriage has not ended (V11=1,5,7); never married (V17=0) Missing Data: 98,99 _________________________________________________________________________________ V13 YEAR WIDOWED OR DIVORCED Number: 13 Type: Num Width: 4 Decimals: 0 Location: 33-36 Year in which This Marriage Ended in Widowhood or Divorce 9767 1901-1999 Year marriage ended 1256 9998 NA or DK year; NA or DK status of marriage (V11=8) 32197 9999 Inap.: marriage has not ended (V11=1,5,7); never married (V17=0) Missing Data: 9998,9999 _________________________________________________________________________________ V14 MONTH SEPARATED Number: 14 Type: Num Width: 2 Decimals: 0 Location: 37-38 Month of Separation for This Marriage The values for this variable indicate the month or season in which spouses stopped living together. See Section IV: Idiosyncracies, Data Cleaning and Variable Detail for more information. 728 1 January 507 2 February 513 3 March 510 4 April 485 5 May 673 6 June 544 7 July 489 8 August 552 9 September 471 10 October 450 11 November 500 12 December 22 21 Winter 23 22 Spring 38 23 Summer 13 24 Fall; autumn 3426 98 NA or DK month; NA or DK status of marriage (V11=8) 33276 99 Inap.: married or widowed (V11=1,3); never married (V17=0) Missing Data: 98,99 _________________________________________________________________________________ V15 YEAR SEPARATED Number: 15 Type: Num Width: 4 Decimals: 0 Location: 39-42 Year of Separation for This Marriage The values for this variable indicate the year in which spouses stopped living together. See Section IV: Idiosyncracies, Data Cleaning and Variable Detail for more information. 8111 1901-1999 Year separated 1833 9998 NA or DK year; NA or DK status of marriage (V11=8) 33276 9999 Inap.: married or widowed (V11=1,3); never married (V17=0) Missing Data: 9998,9999 _________________________________________________________________________________ V16 YEAR MOST RECENTLY REPORTED MARRIAGE Number: 16 Type: Num Width: 4 Decimals: 0 Location: 43-46 Year of the Most Recent Marriage Report for This Individual Values for this variable indicate the interviewing year in which marriage data were most recently collected for this individual. 737 1985 Most recently collected in 1985 746 1986 Most recently collected in 1986 698 1987 Most recently collected in 1987 708 1988 Most recently collected in 1988 612 1989 Most recently collected in 1989 1116 1990 Most recently collected in 1990 1240 1991 Most recently collected in 1991 2145 1992 Most recently collected in 1992 1960 1993 Most recently collected in 1993 2169 1994 Most recently collected in 1994 5838 1995 Most recently collected in 1995 6101 1996 Most recently collected in 1996 1749 1997 Most recently collected in 1997 17401 1999 Most recently collected in 1999 _________________________________________________________________________________ V17 NUMBER OF MARRIAGES OF THIS INDIVIDUAL Number: 17 Type: Num Width: 2 Decimals: 0 Location: 47-48 Number of Marriages of This Individual The values for this variable indicate the total number of marriages reported for this individual through the wave in which his or her marriage history was most recently collected (V16). 14545 0 None 16905 1 One 8628 2 Two 2182 3 Three 393 4 Four 114 5 Five 24 6 Six 31 7 Seven 8 8 Eight 390 98 NA; DK Missing Data: 98 _________________________________________________________________________________ V18 LAST KNOWN MARITAL STATUS Number: 18 Type: Num Width: 1 Decimals: 0 Location: 49 Most Recently Reported Marital Status of This Individual Values for this variable indicate the final status of the current or last marriage reported for the individual through the wave in which his or her marriage history was most recently collected (V16). If an individual has multiple marriages, values here do not necessarily refer to the status of this marriage. See V11 for a record-specific status. 20914 1 Married 14544 2 Never married 2118 3 Widowed 4024 4 Divorced, annulment 1492 5 Separated 128 8 NA; DK Missing Data: 8 _________________________________________________________________________________ V19 NUMBER OF MARRIAGE RECORDS Number: 19 Type: Num Width: 1 Decimals: 0 Location: 50 Number of Records for This Individual The values for this variable indicate the total number of records on the file for this individual. For example, a value of 2 in this variable indicates that he or she has two records, that is, two marriages. A value of 1 indicates an individual with one record on the file. The single record for an individual is used to record either one marriage, no marriages, or entirely missing data about marital history. Therefore, if this variable contains a value of 1, use values in V17, Number of Marriages, to determine which of these three possibilities is indicated. 31834 1 One record for this individual: one marriage, no marriages, or an unknown number of marriages 9415 2 Two records for this individual 1587 3 Three records for this individual 268 4 Four records for this individual 75 5 Five records for this individual 12 6 Six records for this individual 21 7 Seven records for this individual 8 8 Eight records for this individual _________________________________________________________________________________ V20 RELEASE NUMBER Number: 20 Type: Num Width: 1 Decimals: 0 Location: 51 Release Number 43220 1 Release number