COR #336-Attachment III Location of 1957 Raw Data Status of Variables and Recommendations January, 1983 (slightly revised 9/99) The following is a description of the variables on the Master Tapes from the Questionnaires of April, 1957 and October, 1957. For a key to the questionnaire items see the questionnaires reproduced in Appendix A. THIS IS THE DETAILED DESCRIPTION OF THE CREATION OF THE RECODED OR "Q" VARIABLES. NO ATTEMPT HAS BEEN MADE TO UPDATE THIS DOCUMENT WITH CURRENT FILE NAMES OR COLUMN NUMBERS. IT IS PROVIDED FOR THE RECORD IN CASE A USER WANTS TO COMPARE THE RESULTS OF EARLIER WORK WITH CURRENT WORK. THIS DOCUMENT MAY EXPLAIN SOME CHANGES IN THE CODING OVER THE LONG HISTORY OF THE WLS PROJECT. I. High School Rank HSR on SWL21 Little's original 1/6 sample is coded correctly but the other 5/6 were rounded to the nearest number ending in a "2" or "7". This appears to have been done to create 20 categories for analysis which were convenient to tabulate on a card sorter. Unfortunately, no documentation existed on this recoding procedure and the adjustment was discovered for the first time in Spring, 1982. 00 is a legal code and has also been used for missing data. In this recode - code 000-099 and use 999 for missing data (-- or blank). This coding appears on all tapes through SWL21-Change Notice F. HSR = 100 - (Rank in class/# of students in class X 100) COPY HSR from SWL21 for IDSAMP=1 - Little's original 1/6 sample; N=4905 (This N does not agree with the 5,675 described in early publications by Little because a number of the original cards were damaged and some cases were excluded because of missing data on IQ.) Using the original HSR written in ink on the April, 1957 questionnaires for cols.7-8 - VERIFY the 1/6 sample codes and CODE AND VERIFY all others. (The original score was crossed out in pencil and the adjusted score written next to - or sometimes on-top-of the original score.) UPDATEVAR = HSRANKQ UPDATEVAR = HSRSCORQ (3-digit transform) NOTE: The existing variables HSR and HSRNRM will remain on the master. The codebook will indicate these earlier variables were used for all analyses prior to this version of the master, explain the discrepancies and recommend the new version of the variable. II. Henmon-Nelson Test of Mental Ability IQ on SWL21 The variable has the same problems as those described for HSR (I). The problem with the adjusted coding for the 5/6 individuals not in Little's original sample was discovered in COR #332 during an additional search for the Henmon-Nelson scores for Freshman and Juniors for respondents whose siblings were in the sibling sample (N=2438). The Henmon-Nelson Score written on the questionnaires are junior Henmon-Nelson Scores. Missing data for the IQ variables IQ,IQNLNR and IQSCOR were filled in with any score that could be found. (This included Henmon-Nelson Scores from other years, conversions of other scores reported in the State Testing Service booklets and conversions of test scores requested from the schools by Professor Sewell in 7/62 and 10/66.) The variables coded in COR #332 for Freshman and Juniors do not fill in missing data. COPY IQJCR from DE5344 (DAYFILE.DAT) as per COR #332 for N=2438 and flag these entries as "coded" and "verified" for these 2438 individuals on IQJCR. If the respondent does not have a code for IQJCR and is in the original 1/6 sample (IDSAMP=1) then COPY IQ from SWL21. Using the original HN score written in ink on the April, 1957 questionnaires for cols.9-10 - VERIFY the Henmon-Nelson score for those in the 1/6 sample with IQ copied from SWL21 and CODE and VERIFY the score for all others. (The group of students with IQJCR and those in the 1/6 sample are not mutually exclusive - IQJCR has priority and will not be substituted by IQ for the 1/6 sample.) DELETEVAR = IQNLNR NEWVARS = IQFRS, IQFCR, IQFSCOR, IQJRS, IQJCR, IQJSCOR (from DAYFILE.DAT for N = 2438) UPDATEVAR = IQHNQ (Best Measure of Henmon-Nelson Centile Rank. Contains IQJCR if available ELSE recoded value from the original 1957 questionnaire.) UPDATEVAR = IQHNSCRQ (3-digit transform of IQHNQ - Best Measure of IQ Score on the Master Tape.) NOTE: Retain the existing IQ and IQSCOR on the master and provide detailed footnotes similar to those for HSR with recommendations to use the new variables. Delete non-linear transformation of the original IQ and provide details necessary to recreate the variable if desired in Attachment VII. III. Student's Name DETAILS OMITTED FROM THE PUBLIC VERSION IV. Age in April, 1957 from the April, 1957 Questionnaire AGERSP on SWL21 and Birthdate from 1975 Questionnaire - CMBRDX and BRDXDA on SWL21. VARIABLES OK - DO NOT CODE V. Sex of Respondent SEXRSP on SWL21 Coded originally from April, 1957 Questionnaire and corrected whenever errors were discovered during subsequent follow-ups. VARIABLE OK - DO NOT CODE VI. School ID DETAILS OMITTED FROM THE PUBLIC VERSION VII. 1957 Address of the Student and Parent DETAILS OMITTED FROM THE PUBLIC VERSION VIII. Father's Name DETAILS OMITTED FROM THE PUBLIC VERSION 1a. Plans Next Year PLNS58 on SWL21. Original coding was not consistent with multiple coding and the treatment of the "like" option. Respondents checking "no plans" and giving school information in Q2a were coded "school" in Q1a. Multiple responses had the first checked as the variable coded. COPY PLNS58 from SWL21 and using April,1957 questionnaires provide for multiple responses and VERIFY all codes. UPDATEVAR = PLNS58Q 1b. Like to Do What is Planned" SWL1A - col.14 coded for 1/6 sample only. COPY variable from SWL1A - col.14; using the April,1957 questionnaire VERIFY the 1/6 sample coding and CODE AND VERIFY all others. ADD NEWVAR = PLNSLKQ 1c. Circumstances Preventing Doing What You Like SWL1A - col.15 coded for 1/6 sample only. SKIP - DO NOT CODE 1d. Certainty of Plans is located on SWL18A - col.15 and on SWL1A - col.16 (coded on both tapes for the 1/6 sample only.) COPY variable from SWL1A (col.16); using the April, 1957 questionnaire VERIFY the 1/6 sample coding and CODE AND VERIFY all others. ADD NEWVAR = PLCERTQ 2a. Kind of School PLNSED on SWL21 (condensed); on SWL18A - cols. 16, 17 and 342 (condensed); on SWL1A - col. 17 raw data appears to be coded correctly but was not consistent with assignment of "multiple" codes, responses to "like" and Q1 "no definite plans". COPY variable from SWL1A - col. 17 and VERIFY with April, 1957 questionnaire. UPDATEVAR = PLNSEDQ 2b. Attend School in Wisconsin on SWL1A - col. 18 COPY variable from SWL1A - col. 18 and VERIFY with April, 1957 questionnaire ADD NEWVAR = EDWISCQ 2c. Attend School Full or Part-Time on SWL1A - col. 19 for the 1/6 sample only. COPY variable from SWL1A - col. 19 and using the April, 1957 questionnaire VERIFY the 1/6 and CODE AND VERIFY all others. ADD NEWVAR = EDTIMEQ 3. Major Field of Study - SWL1A - cols. 20-21 no key to codes. This variable is to be coded using the codes used for the 1975 and 1977 follow-up studies in Appendix N of the SWL22 codebook. Use codes 000-299 for vocational school or apprenticeships and codes 300 and above for college or university major fields. CODE AND VERIFY directly from the April, 1957 questionnaires using the codes from SWL21-Appendix N. ADD NEWVAR = EDMJ57Q 4a and 4b. Job Plans - Status of Search and Job Title SWL1A - cols. 22-23 no key to codes or marginals available. These variables have been coded for 1/6 sample only. 4a. Job Plans - Status of Search CODE AND VERIFY directly from the April, 1957 questionnaire. ADD NEWVAR = JBST57Q 4b. Job Title SKIP - DO NOT CODE 4c. Name of Firm of 1957 Job - This variable has never been coded. SKIP - DO NOT CODE 4d. Weekly Pay for 1957 Job - SWL1A - col. 24 coded for the 1/6 sample only. SKIP - DO NOT CODE 5a and 5b. Discussions with Teachers and Teachers' Influence SWL21 - TCHCNT and TCHINF and preceding tapes have collapsed the missing data code into the "not at all" category. SWL1A combined both variables into one. COPY TCHCNT and TCHINF from SWL21 - ASSIGN a missing data code in appropriate cases and VERIFY. UPDATEVAR = TCHCNTQ UPDATEVAR = TCHINFQ 6a and 6b. Discussions with Parents and Parents' Influence SWL21 - PARCNT and PARINF and preceding tapes have collapsed the missing data code into the "not at all" category. SWL1A combined both variables into one. COPY PARCNT and PARINF from SWL21 - ASSIGN a missing data code in appropriate cases and VERIFY. UPDATEVAR = PARCNTQ UPDATEVAR = PARINFQ 7a. Education of Father. EDFA57 on SWL21 (condensed); SWL1A - col. 27 SWL21 and earlier tapes combined "did not attend high school", "not ascertained", and "do not know" as one variable and also combined "attended trade or business school" with "graduated from high school" as a single variable. COPY variable from SWL1A - col. 27 and VERIFY with April,1957 questionnaire UPDATEVAR = EDFA57Q NOTE: Use the coding currently used on the SWL21 tape with the addition of "13" for "trade or business school", and the codes 77 and 99 for do not know and N.A. 7b. Education of Mother EDMO57 on SWL21 (condensed); SWL1A - col. 28 SWL21 and earlier tapes combined "did not attend high school", "not ascertained", and "do not know" as one variable and also combined "attended trade or business school" with "graduated from high school" as a single variable. COPY variable from SWL1A - col. 28 and VERIFY with April,1957 questionnaire UPDATEVAR = EDMO57Q NOTE: Use the coding currently used on the SWL21 tape with the addition of "13" for "trade or business school", and the codes 77 and 99 for do not know and N.A. 8a. Education of Older Brother SWL1A - col. 29 coded for the 1/6 sample only. SKIP - DO NOT CODE 8b. Education of Older Sister SWL1A - col. 30 coded for the 1/6 sample only. SKIP - DO NOT CODE 9a. Occupation of Father in 1957 OCF57 on SWL21 (condensed); SWL1A - col. 31 (Other variables for Father's Occupation in 1957 were coded from Wisconsin Tax Department sources.) SKIP - DO NOT CODE 9b. Occupational Aspiration of Student OCCL57, OCST57, and OCPR57 on SWL21 VARIABLE OK - DO NOT CODE 9c. Mother's Occupation in 1957 SWL1A - col. 33 coded for the 1/6 sample only. (Other variables for Mother's Occupation in 1957 were coded from Wisconsin Tax Department sources.) SKIP - DO NOT CODE 10. Perception of College Costs Away from Home SWL1A - col. 34 coded for the 1/6 sample only. COPY variable from SWL1A (col.34); using the April, 1957 questionnaire VERIFY the 1/6 sample and CODE AND VERIFY all others. ADD NEWVAR = EDCOSTQ 11. Parental Support for College PARSUP on SWL21 (condensed); on SWL18A - col. 25 (condensed); and on SWL1A -col.35 (condensed). "Cannot afford" and "must work" are condensed on all tapes. This was a reasonable decision - a number of cases on SWL21 - PARSUP=1 were dumped and compared with the April,1957 questionnaires. Correct the codebook in Change Notice G. 12. Family Wealth Among Community SESP57 on SWL21 VARIABLE OK - DO NOT CODE 13a and 13b. Ever Plan to Attend College and Would You Consider If You Had Money? PLHIED on SWL21; SWL18A - col.41 and SWL1A - col.37 (13a combined with 13b). The combined variable on SWL18A and SWL21 cannot be split into the original variable without recoding. The variable as combined on SWL1A can be split into two distinct variables. COPY variable from SWL1A (col.37) and SPLIT into two distinct variables; using the April,1957 questionnaire VERIFY the codes. ADD NEWVAR = PLCOLLQ ADD NEWVAR = PLCLMNQ 13c. Would Borrow for College if Payback After College? SWL18A - col.42 (Missing data collapsed); SWL1A - col.38 coded for the 1/6 sample only. NOTE: During the testing of COR _#336 it was discovered that the coding of the Sewell 1/6 sample for this variable in 1962 was not done correctly. The code of "2" appears to be assigned to all individuals. Therefore, on all tapes throughout the history of the project - the coding of this variable is correct for only the original 1/6 sample and the rest is garbage. COPY variable from SWL18A (col.42) using April,1957 questionnaire CODE a unique missing data code and VERIFY all others. ADD NEWVAR = EDBOROQ 13d. Parents' Financial Contribution to College in 1958 PRSP58 on SWL21 VARIABLE OK - DO NOT CODE 14a. Applied for Admission to School or College? SWL1A - col.40 (If R checked more than one entry, the entry checked having the highest code was the one assigned. There is an exception when R checked two of the "refused" options - the lower of the three possibilities is assigned.) COPY the variable from SWL1A (col.40) and VERIFY. ADD NEWVAR = APPLIEDQ 14b and 14c. Names of Schools Receiving Applications SKIP - DO NOT CODE 14d. Number of College Applications SWL1A - col. 41 coded for the 1/6 sample only. COPY variable from SWL1A (col.41) and using the April,1957 questionnaire VERIFY the 1/6 sample and CODE AND VERIFY all others. ADD NEWVAR = NCOLAPLQ 15. National Merit Scholarship SWL1A - col.42 COPY variable from SWL1A (col.42) and VERIFY. ADD NEWVAR = NAMERITQ 16a. College Entrance Board Examinations SWL1A - col.43 COPY variable from SWL1A (col.43) and VERIFY. ADD NEWVAR = CEBEXAMQ 16b. Scholastic Aptitute Test Never coded. SKIP - DO NOT CODE 16c. Subject matter of College Entrance Tests Never coded. SKIP - DO NOT CODE 17a. Scholarship Applications SWL18A - col.43 (recoded); SWL1A - col.44 SKIP - DO NOT CODE 17b, 17c and 17d. Places to Which R Applied for Scholarships. Never coded. SKIP - DO NOT CODE 18. Scholarship Pays SWL1A - col.45 coded for the 1/6 sample only. SKIP - DO NOT CODE 19. Earn Toward School Expenses SWL1A - col.46 COPY variable from SWL1A (col.46) and VERIFY. ADD NEWVAR = EDEARNQ 20. Marriage Influenced Plans MARPLN on SWL21; SWL18A - col.56; and SWL1A - col.47 on all tapes - coded for the 1/6 sample only. (Some coding errors were discovered and corrected in 1962. See the description in Appendices in G and H. It is not clear whether this recoding was done properly.) Missing data cannot be retrieved without recoding the original questionnaires. COPY MARPLN from SWL21 and using April,1957 questionnaires ASSIGN a unique missing data code and VERIFY all others.* UPDATEVAR = MARPLNQ *We should have copied for the 1/6 sample only but did not discover the 1/6 only coding situation until the 1982 recoding job was in progress. 21a. Number of Semesters of Each Academic Subject SWL1A - cols.48-57 COPY variables from SWL1A and VERIFY with April,1957 questionnaires. ADD NEWVAR = ALGEBRAQ ADD NEWVAR = GEOMTRYQ ADD NEWVAR = TRIGQ ADD NEWVAR = BIOLOGYQ ADD NEWVAR = CHMSTRYQ ADD NEWVAR = PHYSICSQ ADD NEWVAR = ENGLISHQ ADD NEWVAR = HISTORYQ ADD NEWVAR = SOCSTDYQ ADD NEWVAR = FORNLNGQ 21b. College Preparatory Course HSPROG on SWL21 (condensed); SWL18A - col.45 and 338 (condensed); SWL1A - col.58. (The variables on SWL21 and SWL18A represent whether the student had taken a program which would meet the entrance requirements for the University of Wisconsin in 1957 and was not a record of the student's response to question 21b. See SWL21-Appendices G and H for problems with this variable in the past. Missing data cannot be retrieved without recoding original questionnaires.) COPY SWL1A - col.58 and using april,1957 questionnaires ASSIGN a unique missing data code and VERIFY all others. ADD NEWVAR = COLPREPQ 22. H.S. Practice in College Type Experiences SWL1A - col.59 coded for the 1/6 sample only and coded only if R checked only one option. SKIP - DO NOT CODE 23. Necessity of College Education for Future Occupation NCCLED on SWL21; SWL18A - col.40; and SWL1A - col.60 coded for the 1/6 sample only. (See SWL21-Appendix H for notes on coding errors. It appears that the variable has never been coded for more than the 1/6 sample on all tapes.) COPY NCCLED from SWL21 and using April,1957 questionnaires VERIFY all codes. (It may be necessary to code all not in original 1/6 sample.) UPDATEVAR = NCCLEDQ 24. Teacher's Encouragement for College TCHENC on SWL21 (condensed); SWL18A - col.51 (condensed); and SWL1A - col.61. COPY TCHENC from SWL21 and using April,1957 questionnaires ASSIGN a unique code for missing data and VERIFY all codes. UPDATEVAR = TCHENCQ 25. Parent's Encouragement for College PARENC on SWL21 (condensed); SWL18A - col.54 (condensed); SWL1A - col.62. COPY PARENC from SWL21 and using April,1957 questionnaires ASSIGN a unique code for missing data and VERIFY all codes. UPDATEVAR = PARENCQ 26. Friend's Plans FRNDPL on SWL21 (condensed); SWL18A - col.55 (condensed); SWL1A - col.63. (The original coding was not consistent when coding multiple responses.) COPY FRNDPL from SWL21 and using April,1957 questionnaires ASSIGN a unique code for missing data, provide codes for multiple responses, and VERIFY all codes. UPDATEVAR = FRNDPLQ 27. Student's Interest in High School Studies SWL18A - col.44 (condensed); SWL1A - col.64. COPY col.44 from SWL18A and using April,1957 questionnaires ASSIGN a unique code for missing data and VERIFY all codes. ADD NEWVAR = HSNTRSTQ 28 and 29. Influence of Military Service on Plans MILPLN on SWL21; SWL18A - col.57; and SWL1A - col.65 coded for the 1/6 sample only. (See SWL21-Appendix H for notes on coding errors. It appears that the variable has never been coded for more than the 1/6 sample on all tapes.) COPY MILPLN from SWL21 and using April,1957 questionnaires VERIFY all codes or CODE AND VERIFY. UPDATEVAR = MILPLNQ 30. Value of College as Perceived by Respondent VALUCL on SWL21; SWL18A - cols.69-70 and 71-72); SWL1A - cols.66-68. The variable on SWL18A and SWL21 is Question #30 factor-weighted and was created by J. Michael Armer for his dissertation. See SWL21-Appendix EE for detailed information. It appears that SWL1A has Question #30 coded for the l/6 sample only and we do not know what the key to the codes is. VARIABLE OK - DO NOT CODE Sample Identification IDSAMP on SWL21; SWL18A - col.38; SWL1A - col.69 ADD NEWVAR = IDSAMPQ Teacher's Evaluations written on top of first page of April,1957 Questionnaire for students with special abilities TCHEVL on SWL21; SWL18A - col.48; SWL1A - col.70. VARIABLE OK - DO NOT CODE OCTOBER,1957 QUESTIONNAIRE TO 1/6 SAMPLE PARENTS October,1957 1/6 Sample Respondent CREATE a variable to indicate whether the parent was a 1/6 sample respondent or a 1/6 sample non-respondent to the October, 1957 Questionnaire. ADD NEWVAR = IDSOC57Q 1a. What is son or daughter doing in Fall, 1957? CODE AND VERIFY from October,1957 questionnaires, providing for multiple responses. ADD NEWVAR = ACTV57Q 1b. Full-time or Part-time school CODE AND VERIFY from October,1957 questionnaires ADD NEWVAR = EDFL57Q 2. Name of College Attended in Fall,1957 SKIP - DO NOT CODE 3. Parent's Aspiration for Children's Future Education CODE AND VERIFY from October,1957 questionnaire ADD NEWVAR = EDPAS57Q 4. If parents' wanted child to have further education and child is not enrolled in school, what prevented it? CODE AND VERIFY each option as a dummy variable from October,1957 questionnaire ADD NEWVAR = EDPV157Q ADD NEWVAR = EDPV257Q ADD NEWVAR = EDPV357Q ADD NEWVAR = EDPV457Q ADD NEWVAR = EDPV557Q ADD NEWVAR = EDPV657Q ADD NEWVAR = EDPV757Q DO NOT CREATE A VARIABLE FOR OTHER. 5. Parent Not Wanting Further Education for Child and Reasons CODE AND VERIFY each option as a dummy variable from October,1957 questionnaire ADD NEWVAR = EDNO157Q ADD NEWVAR = EDNO257Q ADD NEWVAR = EDNO357Q ADD NEWVAR = EDNO457Q ADD NEWVAR = EDNO557Q ADD NEWVAR = EDNO657Q ADD NEWVAR = EDNO757Q ADD NEWVAR = EDNO857Q DO NOT CREATE A VARIABLE FOR OTHER. 6a. Did Parent Advise College? CODE AND VERIFY from October,1957 questionnaires ADD NEWVAR = PARADVQ 6b1 - 6b3. Parental Report of Courses Taken in High School CODE AND VERIFY each option as a dummy variable from October,1957 questionnaires ADD NEWVAR = PRALGEQ ADD NEWVAR = PRGEOMQ ADD NEWVAR = PRFRLNGQ 7. Parental Estimate of Family Wealth in Community CODE AND VERIFY from October,1957 questionnaires ADD NEWVAR = FMWELTHQ 8. Parental Opinion of Value of College CODE AND VERIFY each option as a dummy variable from October, 1957 questionnaire ADD NEWVAR = PRVLC1Q ADD NEWVAR = PRVLC2Q ADD NEWVAR = PRVLC3Q ADD NEWVAR = PRVLC4Q ADD NEWVAR = PRVLC5Q ADD NEWVAR = PRVLC6Q ADD NEWVAR = PRVLC7Q ADD NEWVAR = PRVLC8Q ADD NEWVAR = PRVLC9Q ADD NEWVAR = PRVLC10Q ADD NEWVAR = PRVLC11Q ADD NEWVAR = PRVLC12Q ADD NEWVAR = PRVLC13Q DO NOT CREATE A VARIABLE FOR "OTHER" 9. Who responded to this questionnaire? CODE AND VERIFY from October,1957 questionnaire