Documentation for Variables Used in The Bell Curve Regression Analyses Files are saved as Macintosh text files with labels, tab indicating end of field, and CR indicating end of line. The data files are contained in self-extracting archives. Double-click on the icon to extract the files. ÒCYÓ: Calendar year. ÒSYÓ: Survey year. For numeric binary variables, 1=yes, 0=no. Means, standard deviations, and frequency distributions are provided for purposes of checking that the file has been accurately transferred. These statistics are computed without sample weights; hence standardized variables, which were created using sample weights, do not show a mean of 0 and an SD of 1. Note when replicating analyses: ¥ The analyses shown in Chapters 5Ð12 use sample weights applied to the entire white sample. The parallel analyses reported in Appendix 4 do not use sample weights, and are based on the white cross-sectional sample. ¥ The analyses shown in graphs in Chapter 15 and reported in Appendix 6 are all based on the entire white, black, and Latino samples, applying sample weights. ¥ The analyses reported in Appendix 6 using the HOME index or any of the child assessment indexes include TestYr and the appropriate age category (e.g., PPVTAge for the analysis of PPVT6Bot) as independent variables. If you have problems or questions, call Charles Murray at (301) 834-7425, fax the information to (202) 862-7177, or use e-mail address chasmurray@aol.com. NATION.TXT 12,686 cases, 50 variables This file includes variables scored for all NLSY subjects, one line per subject. Variables in the file are in the order listed below. The first 11 (through EdSamples) are the most commonly used variables. From then on, the variables are listed in alphabetical order. Core variables: ID ID number. AFQT80 AFQT score using the 1980 scoring system, excluding invalid scores (R6148, code 67). Not corrected for skew. Valid cases=11878 Mean=65.29 Std. Dev.=22.13 AFQT89 AFQT score using the 1989 scoring system. Not corrected for skew. See discussion of scoring procedure on p. 570 of The Bell Curve. Valid cases=11878 Mean=190.26 Std. Dev.=37.95 zAFQT89 Standard score for the AFQT 1989 scoring system corrected for skew. Assigns z-scores corresponding to the value of AFQT89 centile in a normal distribution, cutting off scores at 3 SDs from the mean. This is the variable used in all the logistic analyses reported in The Bell Curve. Valid cases=11878 Mean=-.26 Std. Dev.=1.01 zIQ*Yr89 Age-equated AFQT standard scores corrected for skew. The reference group is the cohort of subjects born in a given subjectÕs birth year. Valid cases=11878 Mean=-.27 Std. Dev.=1.01 Wgt90 Sample weight for SY90. Valid cases=12686 Mean=2645.39 Std. Dev.=2463.79 WASVAB The sample weight assigned in 1980 to all those who took the AFQT. Valid cases=11878 Mean=2818.25 Std. Dev.=2282.71 zAge Standard score for age of the respondent. Valid cases=12686 Mean=.05 Std. Dev.=.99 zSES The SES index expressed as a standardized score. Construction of the SES index is described in Appendix 2, pp. 573Ð575. Valid cases=11876 Mean=-.36 Std. Dev.=1.06 Race4 SubjectÕs race. Race for Black and Hisp is assigned according to the NLSYÕs cohort categorization (variable R2147). The remaining cohort identified by the NLSY as non-black and non-Latino on variable R2147 is split into Other (using variable R96, includes persons of self-identified American Indian, Asian, and Pacific descent), and White, consisting of the rest. Black 3174 Hisp 2002 Other 569 White 6941 Total 12686 EdSamples Identifies the two homogeneous-education samples used in Part II. ÒCollegeÓ indicates a person with a FinDegree of BA or BS and 16 years of education as of SY90. ÒHSÓ indicates FinDegree consisting of HS and 12 years of education as of SY90. College 1098 HS 3645 Total 4743 Variables in alphabetical order: 1MoNoWk89 Identifies people who worked at least one week in 1989 but less than 48. Excludes people who reported not working because of school in SY89 or SY90. 0 6107 1 2632 Total 8739 1MoOLF89 Identifies people who were out of the labor force for at least 4 weeks in 1989. Excludes people who reported not working because of school in SY89 or SY90. 0 6636 1 3005 Total 9641 1MoUnemp89 Identifies people who were unemployed for at least 4 weeks in 1989. Excludes people who reported not working because of school in SY89 or SY90. 0 8621 1 1358 Total 9979 30+Samp Identifies those whose age at the 1990 interview was ³30. 0 5850 1 4586 Total 10436 Adult14Bio Original (biological or adoptive) parents living with respondent at age 14. Mom 3181 Mom/Pop 8570 Neither 534 Pop 382 Total 12667 Adult14Sum Summary of parental configuration at age 14. Based only on formal marital status (live-in boyfriends or girlfriends donÕt count). Alternatives are Ò2OrigÓ (both original parents), ÒOrig/Step (either remarried original parent), ÒUnmarMomÓ (never married mother or a divorced, still unmarried mother), and ÒOtherÓ (includes arrangements in which child was living with original father, no stepmother, as well as all arrangements in which neither original parent was present). 2 Orig 8570 Orig/Step 1032 Other 725 UnmarMom 2340 Total 12667 AgeMarry Age at first marriage. Valid cases=6962 Mean=22.412 Std. Dev.= 3.489 AllvBA+ Binary variable distinguishing all those who had a bachelorÕs degree or higher as of SY90 from those who had less than a bachelorÕs degree. 0 8554 1 1745 Total 10299 BA AttainSamp Used to identify persons who were still enrolled as college undergraduates as of 1 May 1990. ÒEligibleÓ means the subject could be used for the analysis using AllvBA+ as the dependent variable. Eligible 12230 InSchl/NoD 456 Total 12686 BornWhere Country of birth. 0=US, 1=elsewhere. 0 11811 1 874 Total 12685 Div5Yrs Did the marriage end in divorce during the first five years? Limited to marriages that occurred prior to 1986. 0 4005 1 971 Total 4976 DOB Birthdate, expressed in years. Valid cases= 12686 Mean=60.84 Std. Dev.=2.26 Emp<>568 Subject was not in the military, or not working because unable to work or in school in either SY89 or SY90 (R30743, 34010). 1 9639 Total 9639 Emp<>Schl Subject did not respond Ònot working because in schoolÓ in either SY89 or SY90 (R30743, 34010). 1 10078 Total 10078 EverWed Had the subject ever been married as of SY90? 0 3345 1 6971 Total 10316 FinDegree Assignment of highest degree ever attained by SY90, discarding cases in which the reported years of education were markedly incommensurate with the reported degree. See discussion in Appendix 2, pp. 575Ð576. LTHS: Less Than High School; ProfDeg: PhD, law degree, medical degree; Assoc: AssociateÕs Degree. HS indicates a high school diploma achieved by the normal route, excluding all GEDs. Assoc 608 BA 1422 GED 1151 HS 5062 LTHS 1733 MA 243 ProfDeg 80 Total 10299 GEDvHSGrad Binary variable distinguishing GEDs from all who got a HS diploma through the normal route (including those who went on to further education). Persons with neither a GED or HS diploma are not coded. GED 1151 HSGrad 7415 Total 8566 Health89 ÒPreventÓ: subject prevented from working in CY89. ÒLimitÓ: subject limited in amount or type of work in CY89. ÒNoneÓ: no health limitations on work in CY89. Limit 445 None 9882 Prevent 278 Total 10605 HiIQOcc Classified 1 if subject is a civilian and his/her occupation is one of the high IQ occupations as defined in chapter 2, 0 if the person reports some other occupation. 0 8546 1 403 Total 8949 IntAge90 Age at time of 1990 interview, in whole years. 25 285 26 1382 27 1405 28 1420 29 1358 30 1385 31 1197 32 1212 33 792 Total 10436 Jail Those ever interviewed in jail, excluding those who were not interviewed in jail but are not part of the 1990 sample. Note: the other measure of crime in chapter 11, those in the top decile of self-reported crime as of the 1979 interview, was based on the white male distribution and computed only for white males (because of known large ethnic differences in self-report behavior) and is not included in this data file. 0 10052 1 430 Total 10482 KInc89 Total net family income in CY89 expressed in 1990 dollars, using CPI. Valid cases=8688 Mean=33866.27 Std. Dev.=26994.14 KWage89 Total salary and wages for the respondent in CY1989 expressed in 1990 dollars. Valid cases=10200 Mean=16320.27 Std. Dev.=16308.31 LTHSvAll Binary variable distinguishing those without either a HS diploma or GED from all others with a GED, HS diploma, or more advanced degree. AllOther 8566 LTHS 1733 Total 10299 LTHSvGED Binary variable distinguishing those without either a HS diploma or GED from all others with a GED. Those who got their HS diploma through the normal route are not coded. GED 1151 LTHS 1733 Total 2884 LTHSvHSGrad Binary variable distinguishing those without a HS diploma from all who got a HS diploma through the normal route (including those who went on to further education). Persons with a GED are not coded. HSGrad 7415 LTHS 1733 Total 9148 MarDate1 Date of first marriage, in years. Valid cases=6962 Mean=83.208 Std. Dev.=3.861 MCV Index The Middle Class Values Index, scored 1 and 0 as described in chapter 12. Single people and men in school or unable to work in 1989 or 1990 who otherwise would have scored 1 are treated as missing. 0 5420 1 2939 Total 8359 Occ90Type Type of occupation in 1990, using the 1970 census occupational categories. Clerical 1698 Craft 1061 Farm 96 Labor 530 Mgr 989 NA 1459 Operatv 929 Prof/Tech 1417 Sales 415 Service 1444 Transpt 356 Total 10394 Pov89 SubjectÕs family was below the poverty line in calendar year 1989 0 7316 1 1372 Total 8688 RelAtt Attendance at religious services. ÒWkly+Ó indicates more than once per week. 1/mo 1175 1/wk 2728 2-3/mo 1537 Infreq 3501 None 2543 Wkly+ 1185 Total 12669 Sample The original samples for the NLSY. Military 1280 Supplement 5295 XSection 6111 Total 12686 Sex Gender. Female 6283 Male 6403 Total 12686 UnempSamp Denotes civilians who were in the labor force throughout 1989 and did not respond Ònot working because in schoolÓ in either SY89 or SY90 (R30743, 34010). 1 5959 Total 5959 Wed90 Marital status as of SY90. Modification relies on the Fertility file characterization, but uses the cross-sectional information to identify separated and widowed persons. Married 5367 NevMar 3345 Sep/Div 1523 Widow 45 Total 10280 WedBy30 Was the respondent married prior to the age of 30? 1 Yes 3268 2 No 1261 Total 4529 Work89 Weeks worked in CY89. Valid cases=10436 Mean=38.53 Std. Dev.=19.49 YrRndWrkr Work89=52, with KWage89 greater than 0. 1 5217 Total 5217 zEd90 Standard score for the years of education as of SY90. Valid cases=10358 Mean=-.16 Std. Dev.=1.02 WOMEN.TXT 6,283 cases, 28 variables Variables scored for all women in the NLSY (one case per subject). Core variables: ID ID number. zAFQT89 Standard score for the AFQT, 1989 scoring system, corrected for skew. See description in NATION.TXT documentation. Valid cases=5927 Mean=-.268 Std. Dev.=.958 zIQ*Yr89 Age-equated AFQT standard scores, corrected for skew. The reference group is the cohort of subjects born in a given subjectÕs birth year. Valid cases= 5927 Mean=-.278 Std. Dev.=.953 Wgt90 Sample weight for SY90. Valid cases=6283 Mean=2624.257 Std. Dev.=2369.636 WASVAB The sample weight assigned in 1980 to all those who took the AFQT. Valid cases=5927 Mean=2778.969 Std. Dev.=2260.305 zAge Standard score for age of the respondent. Valid cases=6283 Mean=.055 Std. Dev.=.981 zSES The SES index expressed as a standardized score. Construction of the SES index is described in Appendix 2. Valid cases=5925 Mean=-.368 Std. Dev.=1.070 Race4 SubjectÕs race. See description in NATION.TXT documentation. Black 1561 Hisp 1002 Other 292 White 3428 Total 6283 EdSamples Identifies the two homogeneous-education samples used in Part II. See description in NATION.TXT documentation. College 577 HS 1831 Total 2408 Variables in alphabetical order: 30+Samp Identifies those whose age at the 1990 interview was ³30. 0 2906 1 2418 Total 5324 ADCBin Did the woman ever receive AFDC? Coded ÒyesÓ for all women reporting at least one year of AFDC recipiency. Coded ÒnoÓ for women without children and with no reported AFDC recipiency; and for women with children who had no missing AFDC data from SY79ÐSY90, a child born after 1977, and no reported AFDC recipiency. 1 Yes 1368 2 No 3562 Total 4930 ADCEarly Had the woman received AFDC by the end of the CY following birth of the child? 0 2130 1 673 Total 2803 Adult14Bio Original parents (biological or adoptive) living with respondent at age 14. Mom 1609 Mom/Pop 4229 Neither 273 Pop 163 Total 6274 Adult14Sum Summary of parental configuration at age 14. See description in NATION.TXT documentation. 2 Orig 4229 Orig/Step 514 Other 352 UnmarMom 1179 Total 6274 Age1stBirth MotherÕs age at first birth, in years. Valid cases=3686 Mean=21.736 Std. Dev.=3.973 BStatus Legitimacy status of the first-born child. Illegit 1288 Legit 2343 Total 3631 ChronBin Contrasts chronic welfare recipients with women with children who had no welfare experience. Chronic welfare recipiency (=1) is defined as five or more years of welfare recipiency from CY78ÐCY89. The comparison group (=0) is women whose first child was born after 1977 but prior to 1986, with complete AFDC data but 0 years of welfare. 0 1211 1 572 Total 1783 DOB Birthdate, expressed in years. Valid cases=6283 Mean=60.820 Std. Dev.=2.243 DOBKid Date of birth of the first child. Valid cases=6283 Mean=82.505 Std. Dev.=4.178 IllegMom Mother has given birth to a child out of wedlock (0 scores include women with no children). 0 3981 1 1288 Total 5269 IntAge90 Age at time of 1990 interview, in whole years. 25 122 26 693 27 696 28 688 29 707 30 712 31 623 32 661 33 422 Total 5324 MarDate1 Date of first marriage, in years. Valid cases=3822 Mean=82.498 Std. Dev.=3.968 NKids Number of children as of SY90. Valid cases=5324 Mean=1.443 Std. Dev.=1.295 Pov1st3Yrs First child lived in poverty throughout its first three years. 0 1142 1 283 Total 1425 PreBthPov Was the woman in poverty in the CY before the birth of the first child? 0 1970 1 570 Total 2540 Sample The original samples for the NLSY. Military 456 Supplement 2719 XSection 3108 Total 6283 Wed90 Marital status as of SY90. Modification relies on the Fertility file characterization, but uses the cross-sectional information to identify separated and widowed persons. Married 2880 NevMar 1432 Sep/Div 893 Widow 36 Total 5241 zEd90 Standard score for the years of education as of SY90. Valid cases=5276 Mean=-.127 Std. Dev.=.995 CHILD1.TXT 8,513 cases, 26 variables Variables scored for all NLSY children, representing one case per child for whom data were available through SY90. Core Variables: KidID ChildÕs ID number MomID MotherÕs ID number. zAFQT89 Standard score for the AFQT, 1989 scoring system, corrected for skew. See description in NATION.TXT documentation. Valid cases=8041 Mean=-.525 Std. Dev.=.939 zIQ*Yr89 Age-equated AFQT standard scores, corrected for skew. The reference group is the cohort of subjects born in a given subjectÕs birth year. Valid cases=8041 Mean=-.556 Std. Dev.=.920 KidWgt90 ChildÕs sample weight for SY90. Valid cases=5803 Mean=3559.027 Std. Dev.=2396.266 zAge Standard score for age of the mother. Valid cases=8513 Mean=.184 Std. Dev.=.963 zSES The SES index expressed as a standardized score (refers to the SES background of the mother). Construction of the SES index is described in Appendix 2. Valid cases=8037 Mean=-.626 Std. Dev.=1.041 Race4 MotherÕs race. See description in NATION.TXT documentation. Black 2462 Hisp 1575 Other 430 White 4046 Total 8513 EdSamples Identifies the two homogeneous-education samples used in Part II. See description in NATION.TXT documentation. College 404 HS 2789 Total 3193 Variables in alphabetical order: 30+ Sample Mother was 30 or older as of SY1990. 1 4066 Total 4066 ADCEarly Mother received AFDC by the end of the calendar year following birth of the child. 0 4671 1 1881 Total 6552 ADCYr0 Mother received AFDC in the calendar year the child was born 0 5750 1 1373 Total 7123 ADCYr1 Mother received AFDC in the calendar year following the birth of the child. 0 5300 1 1461 Total 6761 AgeBirth Age of mother at birth, in years. Valid cases=8511 Mean=23.092 Std. Dev.=3.956 AgeMarry Age at which the mother married, in years. Valid cases=6335 Mean=20.692 Std. Dev.=3.277 BStatus Legitimacy status of the child. Illegit 2355 Legit 5191 Total 7546 DOBKid ChildÕs date of birth, in years. Valid cases=8511 Mean=83.616 Std. Dev.=4.077 EvFosOrRel For children with complete residence data, was he/she ever listed as living in foster care or as living with nonparental relatives? 0 8088 1 361 Total 8449 LBW Low birth weight, defined as ²88 ozs. Children under 88 ozs. whose weight was appropriate for gestational age were coded as missing. LBW 521 Not LBW 7306 Total 7827 Pov1st3Yrs Child lived in poverty throughout its first three years 0 2228 1 854 Total 3082 PreBthPov Mother in poverty in the calendar year prior to birth. 0 4503 1 1795 Total 6298 PreNat1stTri Mother received prenatal care in the first trimester. 0 1666 1 5989 Total 7655 Reside90 Custodial adults residing with the child in SY90. ÒDeadÓ refers to deceased children. 2 Pars 4507 Adoptive 57 Dead 101 Foster 19 Mom 2594 Pop 204 Relatives 138 Total 7620 Sample The original samples of the NLSY Military 415 Supplement 4245 XSection 3853 Total 8513 SmokeBin Mother smoked during pregnancy. No 5347 Yes 2529 Total 7876 zEd90 MotherÕs years of education as of SY90, expressed as a standard score based on the entire NLSY population. Valid cases=7622 Mean=-.466 Std. Dev.=.926 CHILD2.TXT 17,040 cases, 40 variables Each case represents one child for one test year. A given child may therefore be represented in up to three cases. TY=test year. Percentiles on the developmental and behavioral indicators all represent within-gender percentiles. Core variables: KidID ChildÕs ID number MomID MotherÕs ID number. TestYr Test year. 86 4971 88 6266 90 5803 Total 17040 zAFQT89 Standard score for the AFQT, 1989 scoring system, corrected for skew. See description in NATION.TXT documentation. Valid cases=16294 Mean=-.60 Std. Dev.=.93 zIQ*Yr89 Age-equated AFQT standard scores, corrected for skew. The reference group is the cohort of subjects born in a given subjectÕs birth year. Valid cases=16294 Mean=-.63 Std. Dev.=.91 KidWgt ChildÕs sample weight for the TY in question. Valid cases=17040 Mean=3008.82 Std. Dev.=2270.10 zAge Standard score for age of the mother. Valid cases=17040 Mean=.18 Std. Dev.=.96 zSES The SES index expressed as a standardized score (refers to the SES background of the mother). Construction of the SES index is described in Appendix 2. Valid cases=16289 Mean=-.69 Std. Dev.=1.03 Race4 MotherÕs race. See description in NATION.TXT documentation. Black 5493 Hisp 3399 Other 785 White 7363 Total 17040 EdSamples Identifies the two homogeneous-education samples used in Part II. See description in NATION.TXT documentation. College 700 HS 6283 Total 6983 Variables in alphabetical order: AgeBirth Age of mother at birth, in years. Valid cases=17040 Mean=22.40 Std. Dev.=3.66 AgeCA Age of child in months at child assessment. Valid cases=17040 Mean=68.09 Std. Dev.=44.49 AgeMA Age of child in months at mother assessment. Valid cases=17040 Mean=68.09 Std. Dev.=44.49 BehProbBot The worst decile in the BehProb index, based on BehProbZ for whites, ages 4-11. 0 7866 1 1003 Total 8869 BehProbZ Same-sex std. score on the behavioral problems index. Uses national norms, mean=100, SD=15. This was used to create the Òbottom decileÓ coding for the behavioral problems index reported in chapter 10, based on the worst decile for the white sample. Note: The racial discrepancies in this variable, based on the motherÕs report of behavioral problems, were large, with black mothers reporting substantially lower levels of behavioral problems than other mothers. This is inconsistent with other data in the NLSY, and consistent with literature on comparable behavioral self-reports suggesting that such differences represent inter-ethnic differences in interpretation of the items rather than inter-ethnic differences in behavior. We therefore elected to define the bottom decile on the basis of the white sample in chapter 10. In chapter 14, we applied the scoring based on the white decile to the other races as well, and inserted a caveat about interpretation. Valid cases=9855 Mean=108.078 Std. Dev.=14.376 BPAgeCat Behavioral problems index age categories, in years. 4-7 5956 8-11 3558 Total 9514 BStatus Legitimacy status of child. Illegit 5416 Legit 10961 Total 16377 DevProbIndex Was the child in the worst decile on any of the four developmental indexes? (DiffBot, FriendBot, MoSocBot, BehProbBot) 1 Yes 1747 2 No 13178 Total 14925 DiffBot In the worst decile in the Difficulty index, for infants ages 12Ð23 mos. 0 1314 1 185 Total 1499 Difficult% Same-sex percentile score on the Difficulty index, normed for age. Valid cases=1499 Mean=53.221 Std. Dev.=29.502 DOBKid ChildÕs date of birth, in years. Valid cases=17040 Mean=82.95 Std. Dev.=3.73 Friend% Same-sex percentile score on the Friendliness index, normed for age. Valid cases=1510 Mean=46.698 Std. Dev.=29.297 FriendBot In the worst decile in the Friendly index, for infants ages 12Ð23 mos. 0 1304 1 206 Total 1510 Home% Percentile score on the HOME scale. Normed for NLSY children (no national norms available). Valid cases=16086 Mean=45.90 Std. Dev.=29.42 HomeAgeCat HOME index age categories, in years. 0-2 4777 3-5 4860 6+ 7403 Total 17040 HomeBot In the worst decile on the HOME scale. 0 13293 1 2793 Total 16086 LBW Low birth weight, defined as ²88 ozs. Children under 88 ozs. whose weight was appropriate for gestational age were coded as missing. LBW 1051 Not LBW 14981 Total 16032 MoSoc% Same-sex percentile score on motor & social development index, normed for age. Valid cases=5873 Mean=51.075 Std. Dev.=28.535 MoSocBot In the worst decile on the Motor & Social Development index. 0 5371 1 502 Total 5873 MSAgeCat Motor & social development age categories, in months. 0-23 3153 24-47 3299 Total 6452 Pov1st3Yrs Child lived in poverty throughout his/her first three years 0 5065 1 2097 Total 7162 . PPVT PPVT standard score, based on national norms, expressed in IQ points. Children with scores lower than 40 (4 SDs below the mean) were assigned a score of 40. Valid cases=6069 Mean=86.039 Std. Dev.=20.895 PPVT6 PPVT scores for children ages 6 and older at the time of the test, and for whom the mother has an AFQT score. Valid cases=2651 Mean=87.863 Std. Dev.=17.324 PPTV6Bot In the worst decile on the PPVT, based on national norms, limited to children ages 6 and older at the time they took the test. 1 Yes 902 2 No 1861 Total 2763 PPVTAge Age in months at time of PPVT assessment. Valid cases=15289 Mean=73.61 Std. Dev.=44.02 PreBthPov Mother in poverty in the calendar year prior to birth. 0 8803 1 3893 Total 12696 PreTYADC ChildÕs mother was receiving welfare in the calendar year prior to TY. 0 13265 1 3721 Total 16986 PreTYPov Child was below the poverty line in the calendar year prior to TY. 0 9807 1 5001 Total 14808 Sample The original NLSY samples. Military 55 Supplement 8495 XSection 8490 Total 17040 TempAgeCat Temperament age categories, for Difficulty & Friendly scales, in months. 0-11 1533 12-23 1620 Total 3153 Page