Errata for NLSY97 Round 6 Release

Errata for NLSY97 Round 6 Release

Significant changes made to the October 2004 Round 6 event history release 

1) Respondent race variable
The code block for the round 6 respondent race variable SYMBOL!KEY!RACE (S15284.00) is incorrect. The correct code block and frequencies for this variable are listed below.

4562 1 White
2136 2 Black or African American
61 3 American Indian, Eskimo, or Aleut
153 4 Asian or Pacific Islander
178 5 Something else? (SPECIFY)
662 6 R REFUSES TO CLASSIFY RACE EXCEPT AS HISPANIC/LATINO
75 7 MIXED
15   Refusal
54   Don't Know

The above error is in the documentation only; no changes to the data are necessary. This error will be corrected in the next release of the NLSY97 data.

2) Created weeks worked per year variables
Corrections were made to the values of the CV_WKSWK_YR series (CV_WKSWK_YR_ALL, CV_WKSWK_YR_ET, CV_WKSWK_YR_SE) for 269 respondents; all rounds were affected by these changes.

Three programming changes affected 69 respondents in round 1. First, a small number of previously unaccounted for within-job gaps were included in the program. This variable was also amended for respondents who reported an invalid month and a valid year to reflect the weeks worked in subsequent years. Finally, the respondent who was assigned an invalid skip in YEMP-200 was reassigned a 0 for all years rather than an invalid skip. Additionally, this update standardized this variable years for respondents who skipped out of the employment section due to age constraints. A 0 was assigned rather than a valid skip for these respondents in some years; this update imposed the same rule on all years.

In addition to reflecting the corrections made during round 1, the round 2 version of this variable was incorrectly calculated by the archivist for those who reported a job beginning before the date of the previous interview (e.g., back-reporters). There were 74 respondents affected by the round 2 changes.

Beginning in round 3, back-reported jobs were included as part of the program rather than hand editing each case. This program used the prior year's weeks arrays to generate the correct number of weeks worked. However, due to an incorrect input file, back-reported jobs included data from only those weeks per year reported in the previous round rather than the weeks per year previously reported across all rounds. For example, the new count reflected only the weeks worked in round 3 and round 2, but not round 1; this error continued through round 6. This correction resulted in the values of 126 respondents in round 3 begin updated to include all rounds of data. In round 4, 250 respondents were affected by this correction; in round 5 the number of respondents was 243; and in round 6, there were 235 respondents affected.

3) Created income and poverty variables
In the round 6 program that creates the CV_INCOME_GROSS_YR variable and the CV_HH_POV_RATIO variable, the CV_UI_YR.01 variables was mistakenly added into the total rather than that CV_AMT_UI_YR.01 variable. This change affected the CV_INCOME_GROSS_YR variable for 96 respondents and the CV_HH_POV_RATIO for 93 respondents. In addition, a change in the topcoded value resulted for additional respondents.

4) Created residence variable
The round 6 created residence variable CV_TTL_RESIDENCES_EDT was updated to limit the data the 858 respondents who answered YCHR-1070. For all other respondents the data should have been a valid skip (-4).


 Significant changes made between the round 5 event history release and the round 6 main release:

1) Created highest grade completed variables - Round 2
Corrections were made to the values of the round 2 variables CV_HGC_EVER (793 cases) and CV_HGC_YEAR (794 cases). Most corrections reflect updates in cases where the respondent was read an incorrect grade for the highest grade completed during the round 1 interview, but the respondent did not corrected this information. Previously, CV_HGC_EVER was incorrectly coded using the loop information, which relies on a correct report from the round 1 interview, but now reflects the respondent's report of the highest grade completed since the round 1 interview. Additional corrections were made to standardize conditions used in creating these variables from round 3 to the present. The updated version of the round 2 program can be found in the following file: cvhgcR2_n.sas.

2) Created hours worked variables - Round 5
There were corrections to 67 respondents in the round 5 CV_HOURS_WK_ADULT_ET variable and 10 respondents in the round 5 CV_HOURS_WK_TEEN variable. These corrections were due to small coding errors in the program. The PUBIDs for the CV_HOURS_WK_ADULT_ET variable are the following: 335, 382, 426, 486, 539, 769, 962, 1072, 1085, 1226, 1274, 1343, 1534, 1576, 1790, 1854, 1860, 2011, 2076, 2163, 2403, 2568, 2579, 2679, 2744, 2829, 2869, 3132, 3312, 3325, 3337, 3548, 3650, 4131, 4179, 4206, 4276, 4489, 4561, 4563, 5094, 5132, 5146, 5222, 5658, 5721, 5998, 6214, 6459, 7061, 7147, 7191, 7192, 7227, 7393, 7478, 7614, 7865, 8027, 8240, 8312, 8481, 8535, 8565, 8572, 8582, and 8637. The PUBIDs for the CV_HOURS_WK_TEEN variable are the following: 2135, 2510, 2820, 3475, 4444, 5600, 5640, 6964, 10930, and 15755.

3) Created weeks worked variables - Round 5
Due to an incorrectly specified input variable, the round 5 weeks worked at a self-employed job (CV_WKSWK_YR_SE.02 for 2186 respondents) and at an employer job (CV_WKSWK_YR_ET.02 for 82 respondents) were overestimated for respondents who held both job types during these weeks. Respondents who only held one type of job were not affected. These variables were updated as of the round 6 main release.

4) Created fertility variables - All rounds
A review of the created fertility data revealed some between-round inconsistencies in coding when a respondent's children lived outside the household. The variables affected were the following: CV_BIO_CHILD_HH, CV_BIO_CHILD_NR, CV_CHILD_BIRTH_DATE~M, CV_CHILD_BIRTH_DATE~Y, CV_CHILD_BIRTH_MONTH, CV_CHILD_STATUS, CV_CHILD_DEATH_DATE~M, CV_CHILD_DEATH_DATE~Y, and CV_CHILD_DEATH_MONTH. Updates were made to the round 6 releases for the following respondents: R1 - 2310; R2 - 2310, 5848, 7138, 8228, 8625; R3 - 11, 441, 1580, 1854, 2310, 3299, 3806, 4119, 4120, 4468, 6653, 7138, 8037, 8066, 8246, 8346, 8701, 8748, 8781; R4 - 441, 769, 2310, 3299, 7138, 7675, 8037, 8052, 8346, 8729; R5 - 227, 441, 2310, 3299, 5325, 6491, 7138, 8052, 8346, 8729.

5) Created hourly wage and created hourly compensation variables - All rounds
Corrections reflect (1) a standardization of the wage programs across all rounds, (2) the addition of wages from previous rounds when the jobs lasts less than 13 weeks in total and spans two rounds, and (3) small coding errors.

Observations affected by changes to the hourly wage variable (CV_HRLY_PAY.xx):
R2: job 1 =59, job 2=41, job 3=26, job 4=17, job 5=4, job 6=4, job 7=1;
R3: job 1 =50, job 2=70, job 3=46, job 4=24, job 5=6, job 6=3, job 7=2;
R4: job 1 =102, job 2=143, job 3=95, job 4=49, job 5=17, job 6=6, job 7=3, job 8=1;
R5: job 1 =44, job 2=18, job 3=3, job 4=1, job 5=1.

Observations affected by changes to the hourly compensation variable (CV_HRLY_COMPENSATION.xx):
R2: job 1=177, job 2=83, job 3=36, job 4=22, job 5=8, job 6=4, job 7=1;
R3: job 1 =226, job 2=112, job 3=50, job 4=24, job 5=5, job 6=2, job 7=3;
R4: job 1 =280, job 2=204, job 3=106, job 4=50, job 5=14, job 6=6, job 7=1, job8=2;
R5: job 1 =1, job 2=5, job 3=1, job 4=4, job 6=1, job 8=1.


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

1) Reference Number changes - Event History Arrays for the year 2000
Due to data processing constraints, the reference numbers (refum) for the event history arrays referring to year 2000 data were changed between the round 5 release and the round 6 release. The file R6EH_RUNM.xls contains a crosswalk between the round 5 and the round 6 rnumbers. Click here to download R6EH_RUNM.xls in a new window.

2) Created net worth variables - Round 4, Round 5, and Round 6
A small coding error in the CV_HH_NET_WORTH_Y program used in previous rounds was discovered. The previously released data included both the amount of a loan from friends, etc... (YAST-3812) and the amount of the loan not paid back (YAST-3882) rather than including only the amount still owed. Incorrect data in round 4 (pubid=539, 2273), in round 5 (pubid=1838, 2601, 3897, 3957, 4072, 4140, 8640) and in round 6 (pubid=2704, 3897, 4191, 4891, 5495, 6424) were updated in the October 2004 release.

3) County City Data Book variables - Round 6 and beyond
County level data from the County City Data Book will no longer be included on the NLSY97 geocode release. Users wishing to use these data will find the 2000 edition of the County City Data Book is available from the U.S. Census Bureau at http://www.census.gov/statab/www/ccdb.html.

4) Created enrollment variable - Round 6
During preparation for the round 6 event history release, it was discovered that the created enrollment variable (CV_ENROLLSTAT) contained a coding error that incorrectly listed the status of 47 respondents as not enrolled. This error was corrected in the round 6 event history release. The PUBIDs of the respondents affected by the CV_ENROLLSTAT changes are the following: 753, 855, 1008, 1054, 1410, 1516, 1709, 2083, 3009, 3181, 3301, 3555, 3784, 3865, 3968, 4067, 4116, 4346, 4459, 4942, 5464, 5798, 5860, 5973, 6002, 6065, 6237, 6487, 6518, 6579, 6687, 6777, 6826, 7050, 7227, 7259, 7582, 8194, 8221, 8501, 8724, 8727, 8833, 8922, 8944, 8947, and 9021.

5) Sampling weight variables - Round 6

Following the round 6 release, it was discovered that the documentation was reversed for the following round 6 variables: SAMPLING_WEIGHT_CC and SAMPLING_PANEL_WEIGHT. This documentation is correct on the round 6 event history release.