COR 690 MAJOR DECISIONS MADE IN THE CLEANING AND EDITING OF THE MARRIAGE, FERTILITY AND WORK HISTORY SECTIONS (EVER-MARRIED FEMALES ONLY) OF THE QUESTIONNAIRE OF THE 1975 RESPONDENT STUDY AND THE 1977 SIBLING STUDY I. Marriage Section. A. Due to the minor number of annulments and the lengthy duration of marriages ending in annulments, annulments were recoded as divorces. There were no annulments reported in 1977, so these were not a problem. B. It is possible for spouse's earnings to be non-zero in the 1975 survey (qs.103-105) despite the fact that a spouse's current or last job was not recorded. This inconsistency results for those spouses who are currently keeping house or are a student but worked last year. No editing was done to resolve this inconsistency. C. If the month in questions 64a, d, e, or 75a, d, e was reported as,"Don't know", for the 1977 Sibling Study, it was allocated to June for the construction of Century Months. If the allocation to June resulted in an inconsistency, such as the next marriage occurring before the termination of the prior marriage, the prior event was moved to 3 months before the date reported for the subsequent event for the construction of Century Months. If the months for either separation and divorce, or divorce and remarriage, or death of former spouse and remarriage, were both reported as "Don't know," which resulted in both events being allocated to June, no changes were made. There is one respondent who reported living with her spouse for 5 years after her divorce. No changes were made, and therefore, there will be one case where the date of divorce is prior to the separation date. II. Fertility Histories. In the fertility histories, all adopted, foster or step children that we knew of were omitted from q's. 117 & 118 in the 1975 survey. This editing was done for males and females, although q's 125m & n (asked of ever-married females only) provided an additional opportunity for the respondent to indicate that a given child was not her biological child. Q's 125-128g of the 1975 survey were changed to correspond to the new number of children. The decision to eliminate non-biological was necessitated by the fact that the interviewer attempted to exclude all adopted, step or foster children from q's 117 & 118 of the 1975 survey and only included them when the respondent was adamant about their inclusion. Thus, this editing made all responses consistent with the interviewers' policy. III. Children Section The child roster was expanded in the 1977 interview to count adopted, step, and foster children. For women, XKIDSNO should be an accurate count of natural or biological children due to the addition of questions 80 and 81. Men did not complete the child roster in 1977 which causes males to be coded inappropriate in more variables in 1977 than in 1975. Five people had more than 10 children in 1977 and this is noted in the text of the codebook when it affects variables. IV. Work Histories of Ever-married Women. A. There were two major decisions made while editing this section of the 1975 data. These decisions were guided by one overriding principle: There should be no inaps in q's 125a or m for any column lower than the highest column checked and the highest column checked should be consistent with the number of children (q's 117 & 118). Editing in this section of the 1977 Sibling Study was designed to be consistent with decisions made in the 1975 study. Therefore, there should be no inaps in Q83a for any column lower than the highest column checked, and the highest column checked in Q83 should be consistent with the number of children (Q79). 1. Twins If a woman had n children (where n is less than or equal to 4) and she had a set of twins, the highest column checked in q 125 was n-l and the nth column had an inap in q 125a & m for the 1975 survey. Thus, the highest column checked in q 125 was not consistent with the number of children in q 117 & 118. In editing, the highest column checked in q 125 was changed to n and a "false interval" was created. For the interval between the births of the twins, a 2 was inserted as the response to 125a and 125m & n were given the same values as 125m & n for the first twin despite the fact that a woman could not be employed between the births of the first and second twins; this is creating a "false interval". When n>4 and the interval between the births of twins was not covered in the table, no changes were made. When n>4 and the interval between the births of the twins was covered in the table, appropriate changes (according to the above guidelines) were made. The same procedure was followed for the coding of the 1977 Sibling Study. Q83a was coded as 2 (not employed) for the interval between the twins. Questions 83b, c and d were coded as inaps. This procedure insures that the highest column checked in Q83 is consistent with the number of children (Q79). 2. Premarital first births. The interval marriage to first birth is not a relevant interval for ever-married women with premarital first births. If the respondent didn't inform the interviewer that her first birth occurred before marriage, the respondent was asked about this interval. If the respondent informed the interviewer that her first birth was before marriage, the interviewer skipped column 1 of Q125 of the 1975 survey and moved on to the next column (when appropriate). This decision of the interviewers resulted in inaps in 125a and 125m for a column lower than the highest column checked for some of the women with premarital first births. Two options for cleaning these data were available: 1) inap col 1 of Q 125 for all those with premarital first births; or 2) for those with a birth before marriage and inaps in col 1 of Q 125, put a 2 in 125a col 1 and 9's in col 1 125m and 125n. Given the aforementioned "overriding principle", the latter alternative was chosen. The months of work experience in the interval marriage to first birth will be 0 for women with premarital births. A similar strategy was followed in the coding of 1977 Sibling Study Data. If the respondent informed the interviewer that her first birth occurred before marriage, some inaps may have been recorded for Q83a(l). This violates the guideline established earlier that there should be no inaps in Q83a for any column lower than the highest column checked. Therefore, any inaps in Q83a(l) for women with premarital births were changed to "2". If the respondent did not inform the interviewer that her first birth occurred before marriage and responded "yes" to Q83a(l), no editing changes were made. The months of work experience in the interval marriage to first birth will be 0 for all women with premarital births (see section III). B. Months worked during an interval. If a woman said that she was employed in a particular interval but the calculated number of months employed during that interval was < 0, she was assigned a 0 on months worked during that interval and her response to 125a in the 1975 survey and Q83a in the 1977 Sibling Study for that interval was not changed. C. Months Worked During an Interval In constructing number of months worked during the intervals next to last and last birth, and end of last period to time of interview for the five persons with more than ten children, the birth dates of the appropriate children were taken from the interview schedule and these variables were constructed by hand for the 1977 Sibling Study data (see Notes on XWWM03L and XWIVMOAL in the codebook).