Errata for NLSY97 Round 7 Release

Errata for NLSY97 Round 7 Release

Significant changes made between the round 6 event history release and the round 7 main release

1) Created weeks worked per job variables - Round 5 and Round 6
A review of the employment variables revealed that the weeks worked variables for respondents who worked at a job but took a break from it for a round or more such that it was not reported during the intervening round were missing the data from past rounds. This problem was confined to the round 5 and round 6 data and has been corrected for the round 7 main release. Variables that have been updated for 10 or more cases are listed below:

Round 5 Cases Round 5 QNAME
11 CV_WKSWK_JOB_YR.01.97
23 CV_WKSWK_JOB_YR.01.98
35 CV_WKSWK_JOB_YR.01.99
11 CV_WKSWK_JOB_YR.02.98
19 CV_WKSWK_JOB_YR.02.99


Round 6 Cases Round 6 QNAME
63 CV_WKSWK_JOB_YR.01.00
11 CV_WKSWK_JOB_YR.01.01
17 CV_WKSWK_JOB_YR.01.97
34 CV_WKSWK_JOB_YR.01.98
68 CV_WKSWK_JOB_YR.01.99
34 CV_WKSWK_JOB_YR.02.00
11 CV_WKSWK_JOB_YR.02.97
12 CV_WKSWK_JOB_YR.02.98
43 CV_WKSWK_JOB_YR.02.99
13 CV_WKSWK_JOB_YR.03.00
13 CV_WKSWK_JOB_YR.03.98
12 CV_WKSWK_JOB_YR.03.99

2) Created jobs worked variables - Round 6
Due to a coding error in matching jobs across rounds, the number of all jobs per year (CV_TTL_JOB_YR_ALL.01/.02), the number of employer-type jobs per year (CV_TTL_JOB_YR_ET.01/.02), and the number of self-employed jobs per year (CV_TTL_JOB_YR_SE.01/.02) in 2001 and 2002 were incorrectly specified in the round 6 release. In addition, a small number of updates were made to the variables measuring the number of all adult jobs (CV_TTL_JOB_ADULT_ALL) and the number of adult employee-type jobs (CV_TTL_JOB_ADULT_ET). All of these variables have been updated in the round 7 main release.

3) Created hours worked variables - Round 6
Due to a small coding error, the CV_HOURS_WK_TEEN variable was updated for 328 respondents. This was corrected as of the round 7 main data release.

4) Created number of schools variables - Round 4 to Round 6
A small programming error caused the following variables to be incorrect in the round 6 event history release: CV_SCH_ATTEND_EVER (affects 15 in round 4, 38 in round 5, 57 in round 6) and CV_SCH_ATTEND_YR (affects 15 in round 4, 37 in round 5, 56 in round 6). These values were corrected for the round 7 main release.

5) Created migration variable - Round 6
A review of the migration variables revealed that the CV_MIGRATION.01 variable was miscoded for 26 respondents. This variable has been updated in the round 7 main release.

Significant changes made between the Round 6 Event History Release and the Round 7 Event history Release

1) Schooling Event History Arrays - all rounds
Users who compare the round 6 NLSY97 Schooling Event History Arrays with round 7 data for the same time periods will see that a number of respondents have their schooling status changed from enrolled to not enrolled. After the round 6 schooling event histories were created, a detailed check was done to ensure an accurate set of created variables were being released. During the checking process a number of problems were uncovered that were fixed. Fixing these problems caused the schooling status changes.

The first issue found was that a number of respondents stated they were taking college courses but these individuals were not enrolled for a college degree. In addition some people were found to be taking GED courses at a college. Prior to the round 7 release all of these people were incorrectly marked as being enrolled in college, when in fact they were only sitting in a college classroom.

To fix this issue, code was added to all rounds. The code first captures the type of diploma the person is pursuing. If that degree is not a true college degree (i.e. not an associates, bachelors, or masters) then the respondent as of the round 7 release is marked as not enrolled.

The second issue was that a number of respondents were found taking high school courses after they reported having their high school graduation. Some of these classes were enrichment classes, while others appear to be misreported or inconsistent dates. Prior to round 7 these graduates were marked as being enrolled in high school, no matter what their degree status. To fix these issues the schooling event histories now consider someone enrolled in high school only until they graduate After high school graduation, all courses and dates that the respondent provides for high school classes are not included in the event histories.

The third issue found was that a number people were taking both college courses and attending secondary school at the same. Prior to the round 7 release the program gave priority to college and it marked the respondent in at least grade 13. As of the round 7 release, the event history program now considers secondary school a higher priority than college. In the data there are a number of respondents who took a few college courses during their high school years. Because college was given higher priority in the round 6 release and earlier these people, who were in 9th to 12th grade suddenly jumped to the 13th grade. If the college program lasted just one semester the respondent not only jumped ahead but also jumped backward in grades resulting in both extra promotions and extra grade regressions. By ensuring high school takes a higher priority these sharp transitions in the event histories are avoided.

Significant changes made to the July 2005 round 7 Mainfile

1) Partners roster variable
Due to a documentation error, the year variables for start and stop date of cohabitation were missing for the 2002 and 2003 PARTNERS roster variables. These variables were included in the next release.

2) Created MSA variable - Round 6 and Round 7
The respondents residing abroad were miscoded as a valid skip (-4) in the round 6 and round 7 CV_MSA variables; the correct code is 5. This change will affect 27 respondents in round 6 and 23 respondents in round 7. These variables will be updated as of the round 7 event history release

3) Other parents roster variable
A number of the year 2002 variables on the OTHERPARENTS roster were inadvertently left off of the public data release. These variables were included in the next release.