=================================================== ** ** NEPS STARTING COHORT 3 - RELEASE NOTES a.k.a CHANGE LOG ** changes and updates for release NEPS SC3 7.0.1 ** (doi:10.5157/NEPS:SC3:7.0.1) ** =================================================== * Known Issues * General: - an updated data manual is not yet published; please use the appropriate data manual from version 1.0.0 instead CohortProfile: - variable 'Current type of school (reconstructed)' [t723080_g1] for wave 6 currently does not correctly distinguish between branches in schools with several educational programs; this leads to the school type of students in such schools being erroneously sclassified as 'School with several educational programs: Unclear' [value 8]; this will be fixed in an upcoming release =================================================== * Changes introduced to NEPS:SC3 by version 7.0.1 * =================================================== General: - due to withdrawn consent by target persons' parents between survey waves, several observations usually have to be omitted from releases; for release 7.0.0 it turned out that, for some of these observations, only consent for parent interview data, not target person data, had been withdrawn; the erroneously dropped observations have been re-integrated, and weights calculated accordingly Weights: - joint weights for students and parents have not been available in recent releases; they have been calculated by now and have been integrated into this release xTargetCompetencies: - uncorrected WLE score for mathematical competence in Grade 9 [mag9_sc1u] has been incorrectly linked; it should not been used for mean-level comparisons with preceding grades (longitudinal correlational analyses or cross-sectional analyses are not affected); correctly linked WLE scores will be included in a future SUF release for Starting Cohort 3 =================================================== * Changes introduced to NEPS:SC3 by version 7.0.0 * =================================================== General: - data from wave 7 have been incorporated into this release xTargetCompetencies: - items from the competency assessment for domain "maths" in wave 5 had not been delivered by the responsible data curators in due time for release 5.0.0 through 6.0.1; they have been integrated into the 7.0.0 release =================================================== * Changes introduced to NEPS:SC3 by version 6.0.1 * =================================================== Weights: - survey design weights (w_t_cal) erroneously had been calibrated to the population totals of sex, federal state and school type in school year 2010/2011 for version 6.0.0; this only should have happened for the weight of participation in wave 1 (w_t1_cal); this eventually led to all cross-sectional and longitudinal weights being based on the uncalibrated weight (w_t) in version 6.0.0; this has been fixed xTargetCompetencies: - linkage of WLE estimators for domains "maths" (mag*_sc1u), "reading" (reg*_sc1u), and "ict" (icg*_sc1u) had been errouneous in all previous releases; this has been fixed =================================================== * Changes introduced to NEPS:SC3 by version 6.0.0 * =================================================== General: - data from wave 6 have been incorporated into this release CohortProfile: - a mistake in the variable tx80230 "panel frame" led to erroneously coded values in the variable ID_t and the generated variable t723080_g1 "current type of school (constructed)"; this has been fixed =================================================== * Changes introduced to NEPS:SC3 by version 5.0.0 * =================================================== General: - data from wave 5 have been incorporated into this release pParent: - the generated variables ISCED, CASMIN and YEARS OF EDUCATION for the surveyed parent and its partner (p731802_g* and p731952_g*) contained wrong values for wave 4 in Version 4.0.0; this has been fixed =================================================== * Changes introduced to NEPS:SC3 by version 4.0.0 * =================================================== General: - data from wave 4 have been incorporated into this release CohortProfile: - in version 3.1.0 and 3.0.0, dummy variables indicating availibility of context data from headmasters [tx80524], class courses [tx80525], german courses [tx80526] and maths courses [tx80527] erroneously had been set to 0 for wave 1 data for all students sampled in special educational needs schools or the migrant oversampling sub-sample; this has been fixed pTarget: - in version 3.1.0, coded variables containing the target persons', parents' or grandparents' "country of birth" [t400000_g*,t400070_g*,t400090_g*,t400220_g*,t400240_g*,t400260_g*,t400280_g*] suffered a coding error in the pTarget dataset version 3.1.0, leading to wrong values for all variables in wave 1, this also affected the calculated migrational background variables [t400500_g*]. This has been fixed. In version 3.1.0, the fix can be manually implemented by using wave 1 and 2 data can be used from SUF version 2.0.0, for instance by merging variables using the following Stata syntax: * -------------------------BEGIN Stata------------------------------- // adjust this with the file path to your version 2.0.0 pTarget file local oldfile Z:/SUF/On-site/SC3/SC3_O_2-0-0/Stata14/SC3_pTarget_O_2-0-0.dta // adjust this with the file path to your version 3.0.0 pTarget file local newfile Z:/SUF/On-site/SC3/SC3_O_3-0-0/Stata14/SC3_pTarget_O_3-0-0.dta // variable list to be merged local updatevarlist t400500_g* t400000_g* t400070_g* t400090_g* t400220_g* t400240_g* t400260_g* t400280_g* // open 3.0.0 dataset use `"`newfile'"' , clear // fill in variables from 2.0.0 dataset merge 1:1 ID_t wave using `"`oldfile'"' , keepusing(`updatevarlist') update replace nogenerate assert(master match match_conflict) // done exit 0 * ---------------------------END Stata-------------------------------- - the concept of reflecting migrational background in NEPS SUFs has been improved in order to also represent migrants in 3.75th generation; thus, the older variables on migrational background [t400500_g1,t400500_g2,t400500_g3] in the pTarget dataset have been renamed using the "v1" suffix [t400500_g1v1,t400500_g2v1,t400500_g3v1], and the new ones have been introduced - data from the questionnaires on spelling (waves 1 and 3) has been removed from xTargetCompetencies and integrated into pTarget upon request by the responsible item developers pParent: - the concept of reflecting migrational background in NEPS SUFs has been improved in order to also represent migrants in 3.75th generation; thus, the older variables on migrational background [p400500_g1,p400500_g2,p400500_g3] in the pParent dataset have been renamed using the "v1" suffix [p400500_g1v1,p400500_g2v1,p400500_g3v1], and the new ones have been introduced ParentMethods: - variable "Willingness: panel participation" [px80400] in dataset ParentMethods erroneously did not contain value labels for values 0 ["Unavailable"] and 1 ["Available"] in version 3.1.0; this has been fixed xTargetCompetencies: - maths items that have been developed for graede 5 and are repeatedly measured in grade 7 erroneously had been misnamed since versino 3.0.0; variables mag5q301_c_sc3g7, mag5d051_c_sc3g7, mag5d052_c_sc3g7, mag5r251_c_sc3g7, mag5v321_c_sc3g7 and mag5r191_c_sc3g7 should (in accordance with NEPS' naming convetions for test items) have been named mag5q301_sc3g7_c, mag5d051_sc3g7_c, mag5d052_sc3g7_c, mag5r251_sc3g7_c, mag5v321_sc3g7_c and mag5r191_sc3g7_c; this has been fixed - some items from the domain procedural metacognition that had been surveyed in wave 3 were missing in versions 3.0.0 and 3.1.0; this has been fixed - data from the questionnaires on spelling (waves 1 and 3) has been removed from xTargetCompetencies and integrated into pTarget upon request by the responsible item developers =================================================== * Changes introduced to NEPS:SC3 by version 3.1.0 * =================================================== pTarget: - variables t31135a and t31035a suffered from a coding error in wave 1 & wave 2 in version 3.0.0; this has been fixed In version 3.0.0, the fix can be manually implemented using the following Stata syntax: * -------------------------BEGIN Stata------------------------------- // adjust this with the file path to your version 3.0.0 pTarget file use "Z:\SUF\On-site\SC3\SC3_O_3-0-0\Stata14\SC3_pTarget_O_3-0-0.dta" , clear local recodevars t31135a t31035a foreach var of local recodevars { recode `var' (4=1)(1=2)(2=3)(3=4) if wave ==1, copyrest recode `var' (1=4)(2=1)(3=2)(4=3) if wave ==2, copyrest } save, replace exit 0 * ---------------------------END Stata-------------------------------- - in version 3.0.0, coded variables containing the target persons', parents' or grandparents' "country of birth" [t400000_g*,t400070_g*,t400090_g*,t400220_g*,t400240_g*,t400260_g*,t400280_g*] suffer a coding error in the pTarget dataset version 3.0.0, leading to "Germany" erroneously being coded as "Afghanistan"; this also affects the calculated migrational background variables [t400500_g*]. This has been fixed In version 3.0.0, the fix can be manually implemented by using wave 1 and 2 data can be used from SUF version 2.0.0, for instance by merging variables using the following Stata syntax: * -------------------------BEGIN Stata------------------------------- // adjust this with the file path to your version 2.0.0 pTarget file local oldfile Z:/SUF/On-site/SC3/SC3_O_2-0-0/Stata14/SC3_pTarget_O_2-0-0.dta // adjust this with the file path to your version 3.0.0 pTarget file local newfile Z:/SUF/On-site/SC3/SC3_O_3-0-0/Stata14/SC3_pTarget_O_3-0-0.dta // variable list to be merged local updatevarlist t400500_g* t400000_g* t400070_g* t400090_g* t400220_g* t400240_g* t400260_g* t400280_g* // open 3.0.0 dataset use `"`newfile'"' , clear // fill in variables from 2.0.0 dataset merge 1:1 ID_t wave using `"`oldfile'"' , keepusing(`updatevarlist') update replace nogenerate assert(master match match_conflict) // done exit 0 * ---------------------------END Stata-------------------------------- - Variable [t27111d_O] has been completely removed CohortProfile: - an indicator containing the field of determined special educational needs from the list of students has been added [tx80505_D] is a dummy indicating special educational needs [tx80505_R] describes the need more specific (only available in RemoteNEPS or via on-site access) - two indicators containing the region [tx80109_g1] and the German federal state [tx80109_g2R] of the sampled institution have been added; the latter is only available in RemoteNEPS or via on-site access xTargetCompetencies: - in the xTargetCompetencies data set file, items from the orthography testing in wave 3 were missing; this has been fixed, including updated values of wave 1 orthography test items spSiblings: - in the spSiblings data set, system missing values in variables "Sibling’s date of birth - month" [p73221m] and "Sibling’s date of birth - year" [p73221y] are incorrectly coded in version 3.0.0, leading to implausible birth dates; this has been fixed. pParent: - the variables [p727001] and [p727002] (formerly anonymized because they contain federal states) are now available with full information in RemoteNEPS or via on-site access - variable [t751016] now has been correctly renamed to [p751016] - variable [askl] now has been correctly renamed to [p723400] =================================================== * Changes introduced to NEPS:SC3 by version 3.0.0 * =================================================== General: - starting with this release, all NEPS Scientific Use Files will ship with an additional, unicode-enabled Stata data set version; this version is only readable in Stata version 14 or younger, and is placed in the subdirectory "Stata14" - translation for all meta data (variable and value labels, question texts, etc) have been revised and completed - meta data for all variables have been revised and updated where appropriate - additional wave 3 has been incorporated into the data, including observations from a sample refreshment in wave 3 - regional information for German federal states have been added (for waves 1 through 3, restrospectively) to the download SUF, more fine-grained information to the RemoteNEPS and onsite variant - information from a mid-year status update mailing regarding the school status of target persons in the individual field have been incorporated into CohortProfile pParent: - forwarded general and vocational educational information for parent and partner in time to construct more reliable ISCED-97 and CASMIN scores - ISCED-97: added code "4A" for parent and partner when reporting both "Abitur" etc. and "Vocational Training" (not university in this context) - minor changes to achieve more precise ISCED-97 and CASMIN values für parents and partner pTarget: - variables "idealistic educational aspiration" [t31035a] and "realistic educational aspiration" [t31135a] suffered an encoding error for wave 1 data in version 2.0.0; this has been fixed; a temporary workaround can be achieved by using version 2.0.0 data sets and correctly recode those values using the following Stata syntax snippet: . recode t31035a t31135a (4=1) (1=2) (2=3) (3=4) if wave==1 // - variables for persons who did not respond to a part of the questionnaire in special needs schools erroneously contained the value "missing by design" (-54) instead of "not participated" (-56) in the pTarget data set in version 2.0.0; this has been fixed - when generating variable "Global self-esteem" [t66003a_g1] in the pTarget data set, variable "Global self-esteem: competence" [t66003d] erroneously had been ignored in version 2.0.0; t66003a_g1 could temporarily be re-generated from version 2.0.0 using the following Stata syntax: * -------------------------BEGIN Stata------------------------------- nepsmiss t66003a t66003b t66003c t66003d t66003e t66003f t66003g t66003h t66003i t66003j tempvar t66003b_r t66003e_r t66003f_r t66003h_r t66003i_r rowmissings recode t66003b (1=5) (2=4) (3=3) (4=2) (5=1), generate(`t66003b_r') recode t66003e (1=5) (2=4) (3=3) (4=2) (5=1), generate(`t66003e_r') recode t66003f (1=5) (2=4) (3=3) (4=2) (5=1), generate(`t66003f_r') recode t66003h (1=5) (2=4) (3=3) (4=2) (5=1), generate(`t66003h_r') recode t66003i (1=5) (2=4) (3=3) (4=2) (5=1), generate(`t66003i_r') egen `rowmissings'=rowmiss(t66003a `t66003b_r' t66003c t66003d /// `t66003e_r' `t66003f_r' t66003g `t66003h_r' `t66003i_r' t66003j) egen `target_variable'=rowtotal(t66003a `t66003b_r' t66003c t66003d /// `t66003e_r' `t66003f_r' t66003g `t66003h_r' `t66003i_r' t66003j) if `rowmissings'==0 & wave==3 replace `target_variable'=-54 if wave!=3 label variable `target_variable' "Global self-esteem" replace `target_variable'=-55 if missing(`target_variable') * ---------------------------END Stata-------------------------------- this issue has been fixed; spParentSchool: - for the sake of consistency between NEPS Starting Cohorts, the data set "spSchool" has been renamed to "spParentSchool" spParentGap: - for the sake of consistency between NEPS Starting Cohorts, the data set "spGap" has been renamed to "spParentGap" =================================================== * Changes introduced to NEPS:SC3 by version 2.0.0 * =================================================== General: - SPSS data sets now ship with the same "VARIABLE ATTRIBUTES" as Stata data sets' "characteristics" - metadata for all datasets has been revised and updated where appropriate - variables now ship with a characteristic 'NEPS_instname' attached in Stata datasets, reporting the variable name used in the survey - wave 2 data has been fully integrated into the data - a new dataset "Weights" has been added, reflecting panel weights for the cohort; documentation is available online - a new dataset "spSibling" has been added, reflecting the target person's siblings reported in the parent's interview - a new dataset "pTargetMicrom" has been added for onsite access, reflecting spatial data from "microm Micromarketing-Systeme und Consult GmbH" - a new dataset "pInstitutionMicrom" has been added for onsite access, reflecting spatial data from "microm Micromarketing-Systeme und Consult GmbH" - several bugfixes and enhancements have been integrated into this new release, influencing various variables; only the most important ones are listed in this change log pParent: - as wave 2 data makes this a panel dataset, the filename has changed from "xParent" to "pParent" - the interview process in parent's interviews does not guarantee unique ids for parents; thus, the identifier in this dataset is no longer "ID_p", but the target person's "ID_t" - three variables with information about the target person's migrational status have been calculated [p400500_g1, p400500_g2, p400500_g3]; a working paper on the generation process and theoretical background is forthcoming - values 5, 6 and 7 have been recoded to 7, 8 and 9 in 'Highest education qualification (ISCED)' [p731802_g1] - values 5, 6 and 7 have been recoded to 7, 8 and 9 in 'Partner: Highest education qualification (ISCED)' [p731852_g1] - value 96 has been recoded to -20 in 'Partner: (Highest) vocational education certificate' [p731863] in accordance with official NEPS missing codes - EGP generation syntax was adjusted due to errors in the derivation syntax (particularly classes IVc and V) [p731904_g8, p731954_g8] - German EGP value labels have been corrected [p731904_g8, p731954_g8] - CASMIN [p731802_g2 & p731852_g2]: Class assignment slightly modified - ISCED [p731802_g1 & p731852_g1]: Civil servants of the medium grade are now identifiable - 'SDQ-Scale: Prosocial behaviour' [p67801a_g1] has been corrected to only contain a sum score if all included items are non-missing CohortProfile: - older weighting variables ('Standardized design weight' [weight_design_std] and 'Design weight' [weight_design]) are now deprecated and have been removed - the interview process in parent's interviews does not guarantee unique ids for parents; as "ID_p" therefore has been replaced by "ID_t" and is no longer needed to link datasets, it has been removed from CohortProfile - variable 'Test: survey day (month)' [test] has been renamed to [testm] - variable 'Test: survey day (year)' [test] has been renamed to [testy] pEducator: - as wave 2 data makes this a panel dataset, the filename has changed from "xEducator" to "pEducator" pInstitution: - as wave 2 data makes this a panel dataset, the filename has changed from "xInstitution" to "pInstitution" pCourseClass: - as wave 2 data makes this a panel dataset, the filename has changed from "xCourseClass" to "pCourseClass" pCourseGerman: - as wave 2 data makes this a panel dataset, the filename has changed from "xCourseGerman" to "pCourseGerman" pCourseMath: - as wave 2 data makes this a panel dataset, the filename has changed from "xCourseMath" to "pCourseMath" pTarget: - as wave 2 data makes this a panel dataset, the filename has changed from "xTarget" to "pTarget" - three variables with information about the target person's migrational status have been calculated [t400500_g1, t400500_g2, t400500_g3]; a working paper on the generation process and theoretical background is forthcoming - 'SDQ-Scale: prosocial behaviour' [t67801a_g1] has been corrected to only contain a sum score if all included items are non-missing - the scale of variable 'Helps other voluntarily' [t67801i] has been erroneously reversed in generation of 'SDQ-Scale: prosocial behaviour' [t67801a_g1]; this has been fixed spGap: - the interview process in parent's interviews does not guarantee unique ids for parents; thus, the identifier in this dataset is no longer "ID_p", but the target person's "ID_t" spSchool: - the interview process in parent's interviews does not guarantee unique ids for parents; thus, the identifier in this dataset is no longer "ID_p", but the target person's "ID_t"