Please NOTE: this text was converted from WordPerfect to ASCII. Some information may have been lost in the conversion process. We suggest you obtain the WordPerfect version for any serious use. --DPLS Staff NELS:88 USER'S GUIDE FOR THE 1996 ELECTRONIC CODEBOOK SYSTEMS (ECBs) AND BASE YEAR THROUGH THIRD FOLLOW-UP PUBLIC-USE DATA FILES ON CD-ROM John R. Taylor Virginia H. Bartot TABLE OF CONTENTS I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 What's New in the 1996 Release?. . . . . . . . . . . . . . . . 1 1.2 How This Guide Is Organized. . . . . . . . . . . . . . . . . . 2 1.3 Overview of the NELS:88 ECB System . . . . . . . . . . . . . . 2 II. Structure of the Directories and Files on the 1996 CD . . . . . . . 3 2.1 The \DOCU Subdirectories . . . . . . . . . . . . . . . . . . . 5 2.2 The \NELS92 and \NELS94 Data File Subdirectories . . . . . . . 6 2.3 Merging Data Files: Merge keys and units of analysis in the ECB. . . . . . . . . . . . . . . . . . . . . . 9 2.4 Important Changes in the Structure of the First Follow-up Teacher Data on the 1996 Release.. . . . . . . . . . 11 III. The 1996 NELS:88 ECBs: Installation Procedures and System Features . . . . . . . . . . . . . . . . . . . . . . . . . 12 3.1 ECB System Files and Space Requirements. . . . . . . . . . . . 12 3.2 Installing the NELS:88 Electronic Codebooks. . . . . . . . . . 15 3.2.1 Installing the Windows ECBs 15 . 3.2.2 Copying the DOS ECB software to your PC . . . . . . . . 15 3.2.3 The installation menu: Configuring drives and paths to run the DOS ECB software . . . . . . . . . 17 3.3 Basic Features of the DOS ECB System . . . . . . . . . . . . . 19 3.3.1 The main ECB screen . . . . . . . . . . . . . . . . . . 19 3.3.2 The search feature. . . . . . . . . . . . . . . . . . . 21 3.3.3 The output menu . . . . . . . . . . . . . . . . . . . . 23 3.4 Variable Name and Reserved Code Conventions 23 . . 3.4.1 Variable name conventions . . . . . . . . . . . . . . . 23 3.4.2 BY-F2 reserved codes. . . . . . . . . . . . . . . . . . 24 3.4.3 F3 reserved codes . . . . . . . . . . . . . . . . . . . 24 3.4.4 Conventions for displaying reserved codes and continuous variables. . . . . . . . . . . . . . . . 25 3.5 Editing SAS and SPSS ECB-generated Program Code. . . . . . . . 25 3.6 EXTRACT.EXE--A handy subsetting program. . . . . . . . . . . . 26 IV. New Base Year through Second Follow-up Data on the 1996 NELS:88 CD Releases.. . . . . . . . . . . . . . . . . . . . . . . . 28 LIST OF TABLES Table 2.1 Directory/Subdirectory Structure of the NELS:88 Public-use CD. . . . . . 4 Table 2.1.1 Subdirectories in \DOCU Containing the ECB Guide, the NELS:88 Data File User's Manuals (DFUMs) and the F3 Reports. . . . . 5 Table 2.2.1 \NELS92: The 1996 NELS:88 BY-F2 Public-use Data File and ECB . . . . . 7 Table 2.2.2 \NELS94: 1996 NELS:88 BY-F3 Public-use Data Files and ECB . . . . . . . 8 Table 2.3.1 1996 NELS:88 Public-use CD: Unique Keys and Merge Procedures for BY-F3 ECB Modules . . . . . . . . . . . . . . . . . 10 Table 3.1.1 Windows ECB files Contained in the directory \ECBW . . . . . . . . . . . 12 Table 3.1.2 1996 NELS:88 BY-F2 Public-use ECB Files Contained in \ECBW\N2P . . . . . 13 Table 3.1.3 1996 NELS:88 BY-F3 Public-use ECB Files Contained in \ECBW\N4P . . . . . 13 LIST OF FIGURES Figure 3-1 DOS ECB Installation Menu . . . . . . . . . . . . . . . . 18 Figure 3-2 DOS ECB Search Menu . . . . . . . . . . . . . . . . . . . 22 Figure 3-3 DOS ECB Output Menu . . . . . . . . . . . . . . . . . . . 23 LIST OF APPENDICES Appendix A Example of SAS Code for Concatenating F1 Teacher Variables Appendix B Summary of NELS:88 Analytical Populations Appendix C Extracting Raw Data through the Extract Program I. Introduction This document is the user's guide for the 1996 NELS:88 public- use CD-ROM release. This release integrates data from the base year through third follow-up (BY-F3) surveys of the National Education Longitudinal Study of 1988 (NELS:88). The primary purpose of this guide is to instruct researchers in the use of the 1996 NELS:88 public-use CD and the Windows and DOS electronic codebook (ECB) systems included on the compact disc. A second purpose of this guide, and one that is of no less importance, is to inform users of the contents of the compact disc--the different data and documentation files contained within. The 1996 NELS:88 releases comprise three separate CDs: The first CD--the subject of this guide--contains public- use data files for the base year through second follow-up and base year through third follow-up NELS:88 samples. The second CD, which is not described in this guide, contains BY-F2 and BY-F3 restricted-use data files. This CD is available only to licensed users. The third CD contains files for the High School Effectiveness Study (HSES), an independent component of NELS:88. This CD is available only to licensed researchers. HSES was undertaken in the spring of 1990, in conjunction with the first follow-up of NELS:88, to extend the analytical utility of NELS:88 data. One of the primary purposes of HSES is to support hierarchical statistical analyses of the effects of high schools on the academic achievements of two longitudinal cohorts: 1988 eighth graders who were enrolled in tenth grade in the spring of 1990, and 1990 tenth graders who were still attending their tenth-grade schools two years later in the spring of 1992. The study is limited to investigations of school effects in urban and suburban schools in the 30 largest metropolitan statistical areas (MSAs). Specifications for the HSES CD-ROM are NOT included in this guide. 1.1 What's New in the 1996 Release? o Previously, datasets were released as separate files, requiring users conducting analyses of data from more than one file to merge files. Datasets at the same level (e.g., the student level) have been combined in the 1996 release to create "megafiles" with multiple records for each case. This new file structure simplifies the program code required to manipulate the NELS:88 datasets and makes file processing faster. o The 1996 release contains data for respondents to the 1994 third follow-up of NELS:88; third follow-up data have been integrated with data from the previous waves of NELS:88. o For composite variables derived from other variables delivered on the 1996 data files, the commented SAS code used to create the composite variables is provided in the ECB description window. Users who wish to re-create these derived variables may use the SAS code to do so; others may wish to use this code as the basis for altering NELS:88 composites to better fit their specific research questions. o The 1996 NELS:88 public-use CD includes--in addition to the third follow-up data--variables never before released on CD (see section IV). 1.2 How This Guide Is Organized This user's guide is divided into four sections: o Section I (this introduction) provides an overview of the ECB System for users who are new to this technology. o Section II furnishes comprehensive information on the directory and file structure of the 1996 CD. Tables in this section summarize the contents of the NELS:88 data files on the 1996 public-use CD release, as well as the organization of the data and the ECBs on the CD. o Section III explains how to install and operate the ECBs. o Section IV describes more fully what is new in the 1996 CD--in addition to the first release of the third follow- up of NELS:88, new base year through second follow-up data have been added to the previously-released files, and these data are not fully documented in user's manuals on the CD. 1.3 Overview of the NELS:88 ECB System While the ECB system is primarily an electronic version of a fully documented survey codebook, it is also more. The list below summarizes the major options that the ECB software provides to NELS:88 researchers: o users can electronically browse through a list of all the variables and composites contained on the NELS:88 data files; o using key words or variable names/labels, users can electronically search for variables that are relevant to their research questions; o the ECB provides an electronic display of the full question text of each variable in the database, along with notes and other pertinent information; o the ECB displays the SAS code that was used to create composite variables (if all of the variables that were used to construct the composite are also present on the data file); o the ECB includes electronic display of the distribution of raw counts and percentages for each variable in the database; o the ECB permits users to select or "tag" variables of interest. Users can subsequently:  print a hardcopy codebook that displays the distributions of the tagged variables;  generate SAS-PC, SPSS-PC+ or SPSS-for-Windows program code for the tagged variables (that in turn can be used with a user's own SAS or SPSS statistical software);  generate a "tag" file which will save the set of tags for import in a future application. The NELS:88 ECBs run on IBM-compatible PCs equipped with compact disc (CD-ROM) readers. Windows and DOS versions of the ECBs are included on the CD-ROM. Note that the ECB does not directly access the NELS:88 data on the CD. Instead, the ECB system generates program code that can later be used by the researcher with his/her own copies of SAS or SPSS. To use the ECB's SAS-PC, SPSS-PC+ or SPSS for Windows program code option, users must have access to, preferably, the PC versions of these statistical software packages; however, users may modify the ECB- generated code for use with versions of SAS or SPSS running on other platforms, such as a mainframe or Unix box. Statistical software is not provided with the NELS:88 ECBs. II. Structure of the Directories and Files on the 1996 CD The NELS:88 CD contains a tremendous amount of information. Table 2.1 outlines the directory structure of the CD and describes the contents of each subdirectory. Table 2.1 Directory/Subdirectory Structure of the NELS:88 Public-use CD \ This "root directory" contains README.TXT, an ASCII file with basic information about installing the CD and its contents. (The WordPerfect 5.1 file for this document, ECBGUIDE.WP5, appears in the \DOCU subdirectory--see below). \ECBW This subdirectory contains all of the files (including program, resource and miscellaneous files) required to run the two ECB systems residing on this CD. Section III of this guide explains which of the ECB files must be copied from the CD to your PC to operate the system. This subdirectory also contains HOW2USE.ECB, a brief explanation of the DOS ECB. \NELS92 This subdirectory contains the base year through second follow-up (BY-F2) student-level data file (as opposed to ECB files) for the core NELS:88 sample. See table 2.2.1 below for file name, size and contents. \NELS94 This subdirectory contains base year through third follow-up data files for the third follow-up respondents to NELS:88. (The third follow-up sample consists of a subsample of second follow-up sample members.) See table 2.2.2 below for file names, sizes and contents. \DOCU This subdirectory contains ECBGUIDE.WP5, a comprehensive manual for the ECBs and the NELS:88 data files on this CD, as well as EXTRACT.TXT, a brief explanation of the extract program. It also contains the NELS:88 base year through second follow-up data file user's manuals (DFUMs) that fully document the design and implementation of the NELS:88 survey components. The 1996 third follow-up publications, the Methodology Report and Descriptive Summary Report, also appear in this subdirectory. Table 2.1.1 below contains additional information on the files contained by \DOCU. \MISC -- Contains the text files, including file frequencies, question text and variable descriptions, that were used to create the resource files for the ECB systems. These large files provide technical documentation on every variable in the ECBs. On the public-use CD these files are named "N2P.TXT" (BY-F2) and "N4P.TXT" (BY-F3). 2.1 The \DOCU Subdirectories The directory \DOCU and its subdirectories contain WordPerfect 5.1 files of the NELS:88 BY-F2 data file user's manuals and ASCII text and WordPerfect 6.0/6.1 files of third follow-up reports. These documents contain information about the purpose of the study, the data collection instruments, the sample design and the data collection and processing procedures. A data file user's manual has been created for each component survey conducted within each wave of NELS:88. Since some of the DFUMs are very large, each has been divided into multiple files, the largest of which is just over 300 KB. While this makes it easier for users to view the individual files on their PCs, it also creates a large number of files. In order to clarify which files correspond to which manuals, the \DOCU subdirectory has been further divided to create a subdirectory for each DFUM. Table 2.1.1 lists the full subdirectory structure of \DOCU: Table 2.1.1 Subdirectories in \DOCU Containing the ECB Guide, the NELS:88 Data File User's Manuals (DFUMs) and the F3 Reports Subdirectory in \DOCU Data File User's Manual \DOCU\ECBGUIDE.WP5 This manual describes the ECB and the contents of the 1996 NELS:88 public-use CD-ROM release \DOCU\BYF1STDM\ Student: Base Year & First Follow-up Combined; \DOCU\F2STDFUM\ Second Follow-up \DOCU\F1DODFUM\ Dropout: First Follow-up; \DOCU\F2DODFUM\ Second Follow-up \DOCU\BYSCDFUM\ School: Base Year; \DOCU\F1SCDFUM\ First Follow-up; \DOCU\F2SCDFUM\ Second Follow-up \DOCU\BYPADFUM\ Parent: Base Year; \DOCU\F2PADFUM\ Second Follow-up \DOCU\BYTEDFUM\ Teacher: Base Year; \DOCU\F1TEDFUM\ First Follow-up; \DOCU\F2TEDFUM\ Second Follow-up \DOCU\F3DSRRPT\ Third Follow-up: Descriptive Summary Report \DOCU\F3METRPT\ Third Follow-up: Methodology Report All second follow-up manuals are in WordPerfect 5.1/Courier 10 font. The base year and first follow-up manuals are in WordPerfect 5.1/Courier 12 font. Users who wish to copy and/or print document files from the CD may use table 2.1.1 above to select the \DOCU subdirectory that contains the files for the manual or report they wish to read or print. It is important to note that these WordPerfect documents have been stripped of all figures and tables that were not produced in WordPerfect. As such, the data file user's manuals are not exact replicates of the bound hardcopy versions that users can, and should, obtain directly from NCES. You may obtain one or more of the bound manuals, complete with all tables, by contacting Peggy Quinn at NCES (202-219-1743). Each of the data file user's manuals fully explains the background, study design and interrelationships among survey components. Appropriate use of the thousands of NELS:88 variables requires a thorough knowledge of sample selection, data collection and data processing procedures. The NELS:88 Second Follow-up Student Data File User's Manual is the most comprehensive resource and is recommended as the best starting point for becoming acquainted with the NELS:88 methodology and products. Note that this manual has been expanded since its initial release on the 12/93 NELS:88 ECB/CD. A Guide To Using NELS:88 Data is an extremely useful document on the special characteristics of the longitudinal NELS:88 database. This guide is unique in that it contains hints and practical advice on how to appropriately use the database. The guide can be obtained by contacting Peggy Quinn at NCES (202-219-1743). 2.2 The \NELS92 and \NELS94 Data File Subdirectories In the 1996 NELS:88 CD-ROM release, datasets with the same unit of analysis (i.e., the student) that were previously released as separate data files have been combined into single "megafiles" with multiple records for each case. The new ECBs reflect this structure and operate faster as a result. There are two data subdirectories on each CD: o \NELS92 contains all BY-F2 data for the core NELS:88 sample. On the public-use CD, the corresponding DOS ECB is ECBN2P (located in directory \ECBW\ECBN2P). o \NELS94 contains the BY-F3 student and F3 postsecondary institution files for the sample of F3 respondents only. On the public-use CD, the corresponding DOS ECB is ECBN4P (located in directory \ECBW\N4P). A single, integrated ECB contains all of the variables on all of the data files in any one data subdirectory. Tables 2.2.1 and 2.2.2 below outline the contents of the 1996 NELS:88 public-use CD data file subdirectories and ECBs. These tables describe for each subdirectory: o the names and sizes of the files contained within the subdirectory; o the level of the data (for example, whether the unit of analysis is the student or postsecondary institution) on each file; o the NELS:88 survey waves and the number of cases that are represented on each file; and o the contents and structure of each file, including new variables that have been added to previously- released data. Table 2.2.1 \NELS92: The 1996 NELS:88 BY-F2 Public-use Data File and ECB The BY-F2 ECB (N2P) variables from 1 data file in subdirectory \NELS92. Components are listed in the order in which they appear in the ECB. STMEG.PUB: BY-F2 Student-level File Size: 241.0 MB N= 27,394 Unit of analysis: Core student Contents: Universe variables BY, F1 and F2 student components (including new student supplement variables) F2 early graduate supplement variables (appended to the F2 student questionnaire variables) F1 and F2 dropout components BY, F1 and F2 school variables at the student level BY and F2 parent components BY, F1 and F2 teacher components Table 2.2.2 \NELS94: 1996 NELS:88 BY-F3 Public-use Data Files and ECB The BY-F3 ECB (N4P) displays variables from 3 data files in subdirectory \NELS94. Files and components are listed in the order in which they appear in the ECB. 1. STMEG3.PUB: BY-F3 Student-level File Size: 163.6 MB N= 14,915 (F3 respondents only) Unit of analysis: Core student Contents: Universe variables BY, F1 and F2 student components (including new student supplement variables) F2 early graduate supplement variables (appended to the F2 student questionnaire variables) F2 student-level transcript variables F1 and F2 dropout components BY, F1 and F2 school variables at the student level BY and F2 parent components BY, F1 and F2 teacher components F3 questionnaire data and derived variables 2. PSE1994.DAT: F3 Postsecondary Education Attendance File Size: 2.8 MB N= 11,560 Unit of analysis: Student and postsecondary institution pair Contents: Data on periods of attendance at postsecondary institutions, including enrollment dates and major field studies. 3. INST1994.PUB: F3 Postsecondary Institution File Size: 38.8 KB N= 2,771 Unit of analysis: Postsecondary institution Contents: For postsecondary institutions applied to or attended by F3 respondents: sector, tuition/fee deciles, enrollment. 2.3 Merging Data Files: Merge keys and units of analysis in the ECB As table 2.2.2 above indicates, the 1996 BY-F3 ECB includes variables from files with different units of analysis; for example, the BY-F3 ECB includes modules at the level of the respondent and the postsecondary institution. Data from different files may be merged on a common variable, known as a merge key. Merging data from different files requires that you be familiar with and select during your ECB session the appropriate merge keys (ID variables) from each module. Note that if you select any variables from the student module, the ECB-generated program code will automatically include the unique key for the student file--the student ID code--identified as "ID" in the program code. (The term "unique key" refers to an ID variable, or set of ID variables, that uniquely identifies cases on a file; unique keys may or may not be merge keys, depending on the files to be merged.) However, even in the student module you must select the merge key that is common to the files you plan to combine. When you select variables from other modules, in which the student ID code is named "STU_ID", no unique key or merge key is automatically included in the ECB-generated program code; you must tag any required keys yourself and edit the ECB-generated program code accordingly. Table 2.3.1 below displays the unique and merge keys for each BY-F3 ECB module. (Note that the same key, such as STU_ID, may have different names on different files, requiring editing of the ECB-generated program code.) The table lists for each data file the file name, the unit of analysis, the unique key and the names and merge keys for files with which the file can be merged. Table 2.3.1 1996 NELS:88 Public-use CD: Unique Keys and Merge Procedures for the BY-F3 ECB Modules 1. File = STMEG3.PUB Unit of Analysis = F3 respondents Unique Key = STU_ID (included automatically in ECB- generated program code as "ID" when student variables are selected) Merge with the postsecondary education institution and student file (PSE1994.DAT) on STU_ID (PSE1994.DAT)/ID (STMEG3.PUB). You must select STU_ID from the PSE1994.DAT module of the ECB. Merge with the postsecondary education institution file (INST1994.PUB) on INCODE (INST1994.PUB)/INCOD1A1 or INCOD1A2 (STMEG3.PUB). You must select INCODE from the institution module of the ECB and INCOD1A1 or INCOD1A2 from the respondent module. 2. File = PSE1994.DAT Unit of Analysis = Student-postsecondary education institution pairs Unique Key = STU_ID + INCODE + INSTNUM Merge with the postsecondary education institution file (INST1994.PUB) on INCODE. You must select INCODE from both the PSE1994.DAT and INST1994.PUB modules of the ECB. See above for information on merging PSE1994.DAT with STMEG3.PUB. 3. File = INST1994.PUB Unit of Analysis = Postsecondary education institutions Unique Key = INCODE See above for merging INST1994.PUB with PSE1994.DAT and STMEG3.PUB. 2.4 Important Changes in the Structure of the First Follow-up Teacher Data on the 1996 Release. In previous releases of the first follow-up teacher data, the data were at the level of the student-teacher pair, with a maximum of two records (teacher reports) per student appearing on the files. In the 1996 release, the data are presented at the student level only. To create one student level record from multiple teacher records per student, teacher reports were combined for each student. Data from the first teacher for any given student appear in the 1996 release under the variable names used in the original release of the data, TCH_IDA through F1T4_11G; these variable names correspond to the F1 teacher questionnaire. Data from the second teacher report, if any, have been written to TCH_IDB through F1T8_11G. Variables for the second report essentially have been renamed, from TCH_IDA--F1T4_11G to TCH_IDB--F1T8_11G, and have been appended to the first report. It was not possible to segregate teacher reports by subject area between the two parts of the student-level record. Consequently, data from mathematics teachers, for example, can appear in either variable grouping TCH_IDA--F1T4_11G or TCH_IDB--F1T8_11G for different students. Research investigations calling upon the use of all teacher data in a particular subject area will require some renaming of variables prior to the analysis of the data, since data provided by math teachers, for example, may appear in TCH_IDA--F1T4_11G for some students and in TCH_IDB--F1T8_11G for others. An example of the program code required to rename variables and create a file that has up to two teacher records per student (depending on the number of teachers--0, 1 or 2--responding for the student) appears in Appendix A. Consult this appendix before attempting to use the F1 teacher data. III. The 1996 NELS:88 ECBs: Installation Procedures and System Features This section briefly describes the NELS:88 ECB software files and provides step-by-step instructions for installing and operating the Windows and DOS versions of the NELS:88 ECB systems on your PC. 3.1 ECB System Files and Space Requirements The ECBs are designed to run on an IBM (or compatible) PC. In general, the greater the resources of your PC system (RAM, hard drive space, etc.) the easier and faster it will be to operate the ECBs. However, even '286' systems may be used (although operation will be slower than with more powerful configurations) with the DOS ECBs. The Windows ECBs are compatible with Windows 3.1 and Windows 95. Each ECB system comprises program and resource files. The resource files, with filename extension .CDC , contain the question text and frequencies displayed in the ECB and are the largest ECB files by far. Table 3.1.1 below lists the Windows program files found in the directory \ECBW. These files are automatically installed on a user-designated hard drive or network drive during the installation of the Windows ECBs. (A number of setup files are also contained in this directory on the CD.) Tables 3.1.2 and 3.1.3 display the DOS ECB program files and the ECB resource files shared by the DOS and Windows versions of the ECBs. Note how the file names in the two ECB systems identify each of the systems: o public-use BY-F2 ECB file names include "N2P"; o public-use BY-F3 ECB file names include "N4P"; Table 3.1.1 Windows ECB files contained in the directory \ECBW Filename Description Size ECBW.EXE Windows ECB program file 407.7 KB ECBW.HLP Windows ECB help file 158.5 KB TBPRO1W.DLL - TBPRO6W.DLL, CTL3.DLL Dynamic link library files for Windows ECB program 830 KB (7 files) Total size 1.4 MB Table 3.1.2 1996 NELS:88 BY-F2 Public-use ECB Files Contained in \ECBW\N2R Filename Description Size DOS Program Files: ECBN2P.EXE The ECB program file--a DOS executable file 76.3 KB EXTRN2P.EXE The DOS executable file for the extract program 30.8 KB ECB Resource Files (used by the Windows and DOS ECBs): N2P01.CDC Descriptions and distributions 321 KB N2P02.CDC Descriptions and distributions 2.4 MB Other ECB Files ECBSPEC.N2P Specifies number of ECB modules and number of records per module 105 bytes N2P.ICO Windows icon file 766 bytes ECB.HLP The DOS ECB HELP File 6.7 KB Total size BY-F2 public-use ECB files 2.8 MB Table 3.1.3 1996 NELS:88 BY-F3 Public-use ECB Files Contained in \ECBW\N4P Filename Description Size DOS Program Files: ECBN4P.EXE The ECB program file--a DOS executable file 76.3 KB EXTRN4P.EXE The DOS executable file for the extract program 30.8 KB ECB Resource Files (used by the Windows and DOS ECBs): N4P01.CDC Descriptions and distributions 341 KB N4P02.CDC Descriptions and distributions 2.7 MB Other ECB Files: ECBSPEC.N4P Specifies number of ECB modules and number of records per module 172 bytes N4P.ICO Windows icon file 766 bytes ECB.HLP The DOS ECB HELP File 6.7 KB Total size of BY-F3 public-use ECB files 3.1 MB Tables 3.1.2 and 3.1.3 are important for DOS users who are uncertain whether their PC has sufficient disk space to accommodate both the program and the resource files for the DOS ECB. The tables indicate that, for the two ECB systems, the space required to copy all of the files for either DOS ECB system ranges from 2.8 MB to 3.1 MB. The space required for the program files for each DOS ECB (ECBN2P.EXE and EXTRN2P.EXE or ECBN4P.EXE and EXTRN4P.EXE), the only files that must be copied to a hard drive or network drive, is much less at only 107.1 KB. See section 3.3 for procedures for installing the DOS ECB software. The Windows setup program installs all of the files listed in table 3.1.1, plus the ECB data files (*.CDC), the ECBSPEC file and the icon file (listed in tables 3.1.2. and 3.1.3) for the selected ECB. The setup program on the CD guides the user through the software installation. Additional space requirements. In addition to the space needed for installing the ECB program and data files, you will need additional disk space for: 1. storing output files generated by the ECB system such as list files of selected variables, customized codebook files or the SAS-PC or SPSS-PC+/Windows program code files; 2. creating statistical program system files; and, 3. performing additional SAS-PC and SPSS-PC+/Windows functions such as recodes of variables, statistical analyses, merges, sorts and other basic data manipulation and analysis functions. Requirements for storing output files. Program code files may be as small as 5K if only a few variables are selected, or as large as 60-80K if 200 or more variables are selected. The size of codebook files also depends upon the number of variables, as does the size of variable list files. Requirements for creating system files. You can estimate the size of the final system file produced from the SAS or SPSS program code by using the following formula: NVAR * NREC * SYS = file size in bytes where: NVAR is the number of selected variables (see discussion of the Tag Window) NREC is the number of observations or data records on the file(s) from which you selected variables. The N, or number of cases, for each file on the 1996 release is reported in the tables in section II above. SYS is 2.2 for SPSS and 4.5 for SAS files In most cases this formula will overestimate the needed space but it serves as a useful guide to the size of the final SAS-PC and SPSS-PC+ system files. What is not taken into account is the space needed by your statistical program to merge files and/or perform analyses. These functions will require space beyond that derived from the application of this formula. 3.2 Installing the NELS:88 Electronic Codebooks The following two sections provide step-by-step procedures for installing the Windows and DOS NELS:88 ECBs. 3.2.1 Installing the Windows ECBs Installation of one Windows ECB requires 4.1 to 4.4 MB of space on any writeable drive. The installation of the second ECB requires 2.7 to 3.0 MB of additional space. To install one or both Windows ECBs: 1. Read the contents of README.TXT (in the root directory of the CD). 2. Turn on your PC, allowing it to boot up. 3. Place the NELS:88 CD in your CD-ROM reader and turn it on. 4. Determine whether there is enough free disk space to install the ECB. 5. From the Windows program manager, under File, Run, run the ECB setup program (SETUP.EXE) contained in the CD subdirectory \ECBW. 6. The setup program will guide you through installation of one or both ECBs, prompting you with questions concerning the ECB that you wish to install, the CD drive, and the installation drive and subdirectory. 7. The setup program will create an Electronic Codebook program group and program item for the selected ECB(s). Run an ECB by double clicking on its icon. 3.2.2 Copying the DOS ECB Software to Your PC The DOS program files contain the software for running the DOS ECB system and the extract program. Only the program files must be copied from the CD to a hard drive or network drive to use the system (installation to a floppy drive is not recommended because of the slowness of such drives). The program files are easily identified by the .EXE extension in their names. The ECB resource files (*.CDC) may remain on the CD or be copied to a hard drive or network drive. However, response is very slow when data files are run from the CD-ROM, particularly when the search feature is used. In deciding whether to install the ECB data files in addition to the program file, you will need to assess whether you have sufficient disk space; the topic of the next section is ECB space requirements. Once you have selected the ECB system that you wish to install, follow the directions below, which use the public-use BY-F2 ECB, ECBN2P, as an example: 1. Place the NELS:88 CD-ROM in your CD reader. (If you have an external reader, be sure to turn it on). 2. Build a subdirectory (on your hard drive or network drive) to hold the ECB software (program file only or program, resource and miscellaneous files). For example, to create subdirectory \N2P on your hard drive ("C:"), at the DOS prompt (">") type: C:\>MD N2P 3. Determine how much free disk space you have on your drive. The ECB will operate faster if all files related to the ECB system you wish to use are copied to your PC. (Consult tables 3.1.2 and 3.1.3 above to determine how much disk space you need to copy all files for any single ECB system). If you do not have enough space to copy all of the files, you can still use the ECB by copying only the program files from the CD. 4. Depending on how much space you have, copy files from directory \ECBW\N2P on the CD to the subdirectory you built in Step #2 above. As noted above, it is optimal to copy all \ECBW\N2P files to the hard drive or network subdirectory: >COPY E:\ECBW\N2P\*.* C:\N2P Or you can copy just the program (executable) files and the ECBSPEC file. For example: >COPY E:\ECBW\N2P\ECBN2P.EXE C:\N2P >COPY E:\ECBW\N2P\EXTRN2P.EXE C:\N2P 5. If you plan to store SAS-PC or SPSS-PC/Windows program code in a subdirectory that is different from the one built in #2, create it now. For example: C:\>MD NELSPGM NOTE: If you do create a subdirectory for program code, you can input this information to the ECB program via the INSTALLATION MENU (see Step #10 below). 6. Change subdirectories so that the default subdirectory is the subdirectory built in Step #2: C:\>CD N2P 7. Execute the ECB software by typing: C:\N2P>ECBN2P Note that the last three characters of this command identify the ECB system. The command for executing the BY-F3 ECB is ECBN4P. 8. The software will usually clear the screen and tell you that it cannot find "needed files." Press any key to engage the INSTALLATION MENU. (Skip the key press if you are placed directly into the INSTALLATION MENU.) 9. Once at the INSTALLATION MENU, modify all of the drive and subdirectory (path) settings to suit your equipment and setup. (Press function key 1 [F1] to get HELP.) See section 3.4 below. NOTE: Although the ECB provides context sensitive help at any point in the software, you may wish to print out \ECBW\N2P\ECB.HLP to learn more about using the ECB software. You can always activate HELP by pressing F1. 3.2.3 The installation menu: Configuring drives and paths to run the DOS ECB software The first time you attempt to execute a DOS ECB, the system will tell you that it cannot find needed files. At this message, complete installation of the ECB software by pressing any key except [ESCAPE] to call up the INSTALLATION MENU (the [ESCAPE] key terminates the ECB). _________________________________________________________________ Figure 3-1 Installation Menu _________________________________________________________________ Default drive & path C: [Drive and path to which ECB software was copied/ installed] SPSS/SAS/Extract files drive and path code C: [Drive and path output files, e.g., SAS or SPSS program code file, codebook file, variable list file] Compact disc drive E: [Drive letter for compact disc reader] Warning BEEP for errors is ON [Toggle warning beep on or off] Toggle context sensitive HELP for errors OFF [Toggle HELP resulting from errors on or off] Help [Information on installation options] _________________________________________________________________ The example above includes a brief explanation of each menu option; in the actual menu, these explanations appear at the bottom of the menu as you cursor through the options. You can move around the menu with left/right and up/down arrow keys or by entering the first letter of each menu option; for example, entering the letter "C" will take you to the compact disc drive line. Press to exit the menu. The first three menu items are the most critical installation options. They specify: 1) the location of the ECB program file; 2) the drive and directory to which program code, codebook and variable lists are to be written; and 3) the drive assigned to the CD reader (the location of the raw data files). The system will not operate without correct specifications for these three menu options. If the default drives and path names displayed in the installation menu are not the drives and paths to which you copied the ECB program file or to which you would like your output files written, you must change the settings for first three installation options. In the present example, C: is specified as the default drive for the location of the ECB software files; however, based on the earlier example given, these files were actually copied to C:\N2P. As such, you need to modify the default drive and path to reflect the exact location of the ECB software. You would change the current default C: to C:\N2P." Similarly, in the previous example you created a special subdirectory to which your output files are to be written. You must change the default output drive and path code from "C:" to "C:\NELSPGM." You must specify the full path name of the drive and directory to which the ECB software has been copied or to which you would like your output files written. Finally, in the above example, "E:" is specified as the drive assigned to your compact disc reader This is correct, according to the earlier example, and no change is necessary. Additional Installation Features. On the Installation Menu you may also choose whether a WARNING BEEP is sounded for user errors. When ON, a beep is sounded when errant keystrokes and other errors are detected by the system. This warning beep may be toggled OFF by moving the cursor to the warning beep menu option (or by entering "W") and then exiting the option; this action changes the setting from ON to OFF (or vice versa). The system will also provide HELP screens that are related to any errors you make; this feature is context-sensitive in that the HELP instructions are pertinent to your location in the ECB system and the current function when the error occurs. You may toggle this feature OFF in the same manner as the warning beep. Note that when OFF, HELP is provided only when invoked by the F1 key. You can access the installation menu (to inspect or modify the settings) at any time during an ECB session by entering I or ALT-N or from the main ECB menu . 3.3 Basic Features of the DOS ECB System This section reviews the main features of the DOS ECBs. The Windows ECBs incorporate similar or identical features. Complete documentation of these features is provided by the online help function of the Windows ECB system; it is strongly recommended that the user review this on-line documentation before using the Windows ECB. 3.3.1 The main ECB screen Once the ECB software has been installed, it presents the main ECB screen. From top-to-bottom, this screen displays the project title, the banner bar, the header bar, a list of variables and the keystroke bar. Project Title identifies the database described by the ECB. Since there are two ECBs and databases on the CD, the title is a very useful reminder of which ECB you are operating. The banner bar lists the major command groups of ECB functions. The highlighted letters are 'hot' keys that activate the commands. In the DOS ECB, the entire set of commands is displayed by function key 10 [F10] or [M]. From this display, the user may activate any of the commands. The header bar labels the columns of variable names and variable labels. In addition, the header bar identifies the module (data file) and wave and component or section of a highlighted variable. For example, "STMEG.PUB F2 STUDENT PUB" indicates that the variable is on the student-level BY-F2 public- use (privileged-use) file and is a second follow-up student component variable. The variable list displays each variable's name and a brief descriptive label. Variables are color-coded in this list: green=categorical, white=alphanumeric, and magenta=continuous. The NELS:88 BY-F2 variables are ordered within modules by survey component (student, transcript, dropout, school, parent, teacher) and within component by wave (base year, first follow-up, second follow-up). Third follow-up student-level variables are appended to the end of the student-level module, following the F2 teacher variables. The "List" function displays all variables that have been tagged in the ECB session (see section 3.3.2 for procedures for tagging variables). List may be activated from the main menu, or it may be toggled by L and ALT-L. The keystroke bar at the bottom of the screen lists the valid keystrokes (other than hot keys) for scrolling through the list of variables. In addition, at the far right, a count of the number of currently selected variables (or "tags") is displayed. From the main menu (invoked from the main screen by pressing M or F10) or the banner bar, users may pursue the four major applications of the ECB: 1. The ECB displays information about each variable including: o the wording of the questionnaire item corresponding to the variable or an explanation for a composite variable; o the SAS code used to construct composite variables on the file (if all of the variables used to create the composite are present on the data file); o for each categorical variable (and some continuous variables with relatively few values), frequency/percentage distributions are displayed (only unweighted distributions are available; for BY-F2 variables, valid percentages are displayed); o for each continuous variable, the valid range is presented as a label for all cases (instead of each value listed separately, which would involve an excessively long display); o footnotes and comments are included in variable descriptions to indicate variables that have been recoded or suppressed in the public-use version of the data to protect confidentiality, notify users of potential bias from high item nonresponse or further elucidate the meaning of a constructed variable (including pointing users to related constructs in the same or different NELS:88 modules). 2. The search capabilities of the ECB allow you to find and tag variables required for extract and analyses. By coupling the first and second features (full information on a variable and search capabilities), you can investigate the breadth and depth of the dataset. See section 3.3.2 below. 3. A subset of variables for extraction and analysis may be selected or "tagged. To tag a variable, highlight the variable and press T . To untag the variable, press T again. 4. Finally, program code (SPSS PC+/Windows or SAS-PC) or codebook text may be output for the tagged variables. See sections 3.3.3 and 3.5 below. 3.3.2 The search feature. Searches may operate on one, two or three information fields--variable name, label, or description. Variable name is always included in searches. Figure 3-2 displays the DOS ECB search menu. _________________________________________________________________ Figure 3-2 DOS ECB Search Menu _________________________________________________________________ Search for text Find next [CTRL - L] [Forward search will begin] Search string is [Modify search string] Labels search is ON [Toggle search of labels ON and OFF] Description search is OFF [Toggle search of descriptions ON and OFF] Tag all matches [Untags if currently tagged] Import one file of tags [Loads a set of tags from one import file] Combine two files of tags [Combine with AND/OR logic] Advance to next tag [Moves down to next tagged variable] _________________________________________________________________ You should also be aware that, in the DOS ECB, once you begin a search-and-tag session, you cannot escape. Depending on your equipment, a search and tag session may take several minutes or more. Before initiating a search and tag session, carefully consider the nature of the search and the string that is the basis of the search. Searches that are similar to Dbase/Paradox queries may be used to identify sets of variables. The procedure for query searches is: 1. Clear all tags. 2. Enter the desired search string. 3. Tag all matches. You may delete some of the tagged variables, if desired, after viewing their descriptions and frequencies. 4. Output the selected tags through the output menu (see section 3.3.3 below). 5. After repeating the above steps, combine multiple output files of tags with AND/OR logic using Search, Combine Tags. 3.3.3 The output menu Figure 3-3 displays the DOS ECB Output Menu. From this menu you can specify the kind of program code--SAS or SPSS (PC or Windows)--desired. Codebook text files and import files of tagged variables may also be output from this menu. See section 3.5 for instructions on editing program code generated by the ECB system. _________________________________________________________________ Figure 3-3 Output Menu _________________________________________________________________ Code: Select Desired Output A -- SAS-PC program code [Creates file of SAS-PC extract program code] S -- SPSS-PC program code [Creates file of SPSS-PC extract program code] W -- SPSS for Windows include file [Creates include (*.inc) for SPSS for Windows] C -- Codebook text file [Creates text file codebook] T -- Save file of TAGS for import [Creates file for Import] (see Search Menu) _________________________________________________________________ 3.4 Variable Name and Reserved Code Conventions 3.4.1 Variable name conventions. Base year through second follow-up variable names typically begin with the survey wave for which the variable was created: "BY" = Base Year = 1988; "F1" = First Follow-up = 1990; "F2" = Second Follow-up = 1992. Certain school-level composites begin with the modal grade of the student sample--for example, "G12" = Grade 12. Statistical weights include the string "WT" in the variable name. The case selection flags that must be used in tandem with statistical weights ordinarily have the same variable name stem. For example, the selection flag for use with F2QWT is F2QFLG. Note that third follow-up variable names do not follow the same conventions as the prior rounds; third follow-up variables names do not necessarily begin with "F3". Third follow-up respondent-level variables are appended to the end of the BY-F2 student-level variables, making them easy to identify. 3.4.2 BY-F2 reserved codes. A set of standard reserved codes are used in the NELS:88 BY-F2 datasets to help users differentiate among reasons for missing items. These reserved codes and their meanings are: 6 = MULTIPLE RESPONSE 7 = REFUSED 8 = MISSING 9 = LEGITIMATE SKIP When a variable is more than one column in length, the right-hand column contains one of the above missing codes and the remainder of the columns are filled with nines (such as "96," "996," etc.). Blanks may also indicate data are missing because of unit nonresponse. (Standard reserved codes are assigned only to items that are missing for respondents.) In the student-level megafiles on the 1996 CD, blanks (usually labelled "NONR/NOT IN SAMPLE THIS WAVE") for base year through second follow-up items indicate either that an eligible sample member did not complete the survey instrument on which the item appeared (was a nonrespondent) or that the sample member was not eligible to complete the instrument in that particular round of the survey (was not in the sample in that wave). 3.4.3 F3 reserved codes. The third follow-up survey uses negative values as reserved codes for missing data. This convention has become the new standard at the Department of Education. These codes are: -2 CURRENTLY ATTENDING (This code is used when an ending date, such as the ending date of school attendance, is asked.) -3 NOT ASKED IN SAQ (hardcopy self-administered version of questionnaire) -4 UNCODABLE VERBATIM -5 NOT APPLICABLE -6 MISSING -7 REFUSED -8 DON'T KNOW -9 LEGITIMATE SKIP -10 INSTITUTION NOT IN 1993/1994 IPEDS FILE (This code is used when the appropriate value for a postsecondary institution cannot be determined since it is not present in the 1993/1994 IPEDS file.) -11 MILITARY TRAINING (Military training programs are not considered postsecondary education, but were listed by some respondents--academic programs, such as West Point, are not assigned this missing code.) -12 FOREIGN INSTITUTION (This code is applied to foreign institutions that do not have programs affiliated with U.S. postsecondary education institutions.) Codes "0" and "-1" are reserved for use in the DAS (Data Analysis System) created for the third follow-up and are not found in the F3 ECB data files. 3.4.4 Conventions for displaying reserved codes and continuous variables. Within the frequency/percentage window for each variable, reserved codes may be mixed with valid data codes. All reserved codes are enclosed by braces, {}. For continuous variables, valid data is typically assigned a code of {CONT} and, for third follow-up variables, labeled with {minimum-maximum,mean/standard deviation}. If the first (top) value label for a frequency/percentage distribution is enclosed by braces, the ECB software assumes that the variable is continuous, not categorical. 3.5 Editing SPSS and SAS ECB-generated Program Code Though the ECB system will generate code with appropriate data steps and frequency commands, editing of the code is required to: 1. Merge files. The ECB-generated code does not include data steps that merge files when variables are selected from different files in the ECB. You must add your own merge steps, and you should be certain to select the appropriate merge keys--refer to tables 2.3.1 and 2.3.2 above--during the ECB session. 2. Select a subset of cases. You must add statements to the program code to select a sample representing a population of analytic interest. Otherwise, all cases in the dataset will be included in the analysis. For example, to select from the BY-F2 student-level file eligible female members of the tenth grade cohort who completed an F1 student questionnaire, you would add to the SAS program code the statements: IF F2SEX=2; (F2SEX is more complete than F1SEX) IF G10COHRT=1; IF F1QFLG=1; More information on how to define meaningful NELS:88 BY-F2 analytic samples is provided in Appendix B. The BY-F3 student-level file includes data for only F3 respondents. The F3 data include variables that identify BY-F3 analytic samples, such as 1988 eighth graders in 1994 (the BY-F3 longitudinal panel.) Refer to variable descriptions in the F3 portion of the BY-F3 ECB for additional information on identifying F3 analytic samples. 3. Save temporary work files in SAS as a permanent SAS dataset. You should also review the file names used within the code. While the file names embedded in the generated code should reflect the file names, drives and paths specified in the installation menu, it is always prudent to check the file names used throughout the generated code before actually running it. In addition, before running a SAS job, you should review the PROC FORMAT command. The ECB-generated code creates unique format names by adding an 'F' to the end of the variable name. If a variable name contains eight characters, quite common for NELS:88 variables, 'F' will be substituted for the last character of the original variable name. If variables with names that are unique only because of the eighth character are selected from the ECB, the user must edit the PROC FORMAT command to make unique format names for each variable. 3.6 EXTRACT.EXE--A handy subsetting program The sections above describe how to subset variables through the ECB and how to subset cases using selection statements you add to the ECB-generated SAS or SPSS program code. You may also subset both variables and cases with the separate extraction program that is stored under \ECBW on the CD. The EXTRACT program, which runs from the DOS prompt, works with the ASCII codebook files produced by the ECBs to subset data files, especially useful for non-SAS/SPSS users. (Note: The extract program [\ECBW\N2P\EXTRN2P.EXE or \ECBW\N4P\EXTRN4P.EXE] is not installed by the Windows ECB setup program and must be copied to a hard drive or network drive.) Before using the program, you must create through the ECB a codebook text file containing the variables to be extracted and any variables on which cases are to be selected. By entering the correct extract command at the DOS prompt, a description of how to use the program will appear on your screen. Please read \DOCU\EXTRACT.TXT (which also appears as Appendix C of this guide) for additional information. It is important to note here that you MUST use the correct extract command; the correct syntax varies with the ECB system you are using. The command is always of the form: EXTRN??.EXE where ?? = 2P for the public-use BY-F2 ECB, or 4P for the public-use BY-F3 ECB. Note also that the default name assigned to the ASCII codebook text file generated by the ECB system varies by ECB. For example, for the public-use BY-F2 ECB the name assigned by the system to the codebook text file is "N2PCB.TXT." If only one data file (e.g., STMEG.PUB, the public-use student file containing BY-F2 data) is being used, you may also set conditions to select cases. (You must make sure that the attribute[s] on which you are subsetting cases [e.g., gender] is [are] one of the variables in the codebook text file [e.g., SEX].) IV. New Base Year through Second Follow-up Data on the 1996 NELS:88 CD Release. This section of the guide describes the new student-level data that appears on the 1996 NELS:88 public-use CD. Expanded universe variables. A total of six universe variables appear on the 1996 NELS:88 student-level files. These universe variables document how each sample member entered the sample and show each sample member's enrollment/eligibility status in each round. The universe variables that were delivered on previous releases of the NELS:88 second follow-up student data have been expanded to include all cases on the 1996 BY-F2 student-level files, including BY sample members who were subsampled out in the first follow-up. Two new universe variables have been added to the BY-F3 student files on the 1996 CDs to document F3 status and BY-F3 status. Centiles. Centile variables (with categories of one percent each) have been constructed for selected F2 test score composites and F2SES1 (the first of three F2 versions of socioeconomic status) and appear on the BY-F2 and BY-F3 student-level files. In addition to allowing more fine-grained analyses than the corresponding quartile variables, the centile variables permit analysts to collapse centiles in order to create other categories (e.g., deciles, tertiles) that are appropriate to their research questions. The new centile variables appear on all 1996 student- level files. At-risk variable. BYRISK, a new variable, is a count of the number of factors related to dropping out of high school exhibited by each sample member. BYRISK appears on all 1996 student-level files. Early graduate supplement data. Data from the early graduate supplement to the core F2 student questionnaire are provided on all 1996 student-level files. Early graduate supplement data are provided for respondents who: - completed the main portion of the F2 student questionnaire; - answered "Already graduated" to Q. 6A in the main student questionnaire ("What grade are you in?"); - answered at least one item in the early graduate supplement (Q. 114 - Q. 127B of the F2 student questionnaire). The NELS:88 supplement paralleled the High School and Beyond (HS&B) early graduate supplement and collected information about when the student graduated, why he or she chose to graduate early and who helped in making the decision and the student's activities since early graduation (continuing his/her education, working, participating in a training program, actively serving in the military, etc.) If the student attended a two- or four-year college or vocational school, additional information was sought about when, where and how often the student attended the school. If the student worked, information about the type and length of employment was requested. The NELS:88 early graduate supplement differs from the HS&B supplement in one respect: NELS:88 included students who graduated by alternative means, such as the GED, whereas HS&B did not. Thus, dropouts who eventually earned a GED are included in the NELS:88 early graduate sample. Early graduates who earned a GED can be separated from those who earned a high school diploma to compare NELS:88 and HS&B early graduates, using responses to NELS:88 second follow-up student questionnaire item F2S6B. Note that some sample members with early graduate data appear not to have graduated from high school or received a GED before the standard graduation date for their cohort. Other sources in the NELS:88 files (such as "F2EGDFLG", the early graduate flag, and "F2TRSTYP", a cross-component estimator of dropout status) may be used to investigate early graduation status. Appendix A Example of SAS Code for Concatenating F1 Teacher Variables /*Example of SAS code for concatenating the two sets of F1 teacher variables (TCH_IDA--F1T4_11G AND TCH_IDB--F1T8_11G) to create one set of variables containing all F1 teacher data*/ LIBNAME N2P 'C:\N2P'; PROC FORMAT; VALUE F1T2_3F 1='ACADEMIC' 2='ADVANCED OR HONORS' 3='GENERAL' 4='VOC-TECH-BUSINESS' 5='OTHER' 6='{MULTIPLE RESPONSE}' 8='{MISSING}' /*Create temporary SAS dataset containing both sets of teacher variables. Note that, in addition to the ID and subject variables and the F1 teacher questionnaire flag (F1TEQFLG), you may need to select the appropriate sample flag and corresponding statistical weight for your analysis.*/ DATA TEACH1;INFILE 'E:\NELS92\STMEG.PUB' LRECL=1024 PAD; INPUT ID 1-7 / / / / / / / TCH_IDA $ 107-114 F1SUB_A $ 115-115 F1CLSIDA 116-117 F1T2_3 150-150 TCH_IDB $ 699-706 F1SUB_B $ 707-707 F1CLSIDB 708-709 F1T6_3 742-742 / ; /*Select only cases with F1 teacher data*/ F1TEQFLG=1; RUN; /*Create temporary SAS dataset containing only second set of teacher variables*/ DATA TEACH2;SET TEACH1 (KEEP=TCH_IDB F1SUB_B F1CLSIDB F1T6_3); /*Rename second set of variables to match names of first set*/ DATA TEACH2A; SET TEACH2; RENAME TCH_IDB=TCH_IDA F1SUB_B=F1SUB_A F1CLSIDB=F1CLSIDA F1T6_3=F1T2_3; RUN; /*Create a dataset that concatenates the data in the first and second portions of the teacher record. Note that the two source datasets must contain the same variables with matching attributes. The concatenated dataset, TEACH, will contain twice as many records as STMEG.PUB and will be at the level of student- teacher report pairs. Some records will be blank, except for STU_ID (ID), because only one teacher report was received for some sample members. Selecting only records in a particular subject area, such as math, will result in a student-level file, since only one report per student per subject is possible.*/ DATA TEACH; SET TEACH1(KEEP=TCH_IDA F1SUB_A F1CLSIDA F1T2_3) TEACH2A (KEEP=TCH_IDA F1SUB_A F1CLSIDA F1T2_3); /*To select records in only one subject area--math, for example-- add the code: F1SUB_A='M' The resulting file contains only one teacher record per student. */ LABEL TCH_IDA='TEACHER ID NUMBER' F1SUB_A='CLASS SUBJECT' F1CLSIDA='CLASS ID NUMBER' F1T2_3='TRACK CLASS IS CONSIDERED TO BE' ; FORMAT F1T2_3 F1T2_3F. ; RUN; PROC FREQ DATA=TEACH; TABLES F1CLSIDA F1T2_3 F1SUB_A; PROC UNIVARIATE;VAR F1SCH_ID ; RUN; Appendix B Summary of NELS:88 Analytical Populations There are eight distinct analytic samples in the NELS:88 BY - F2 student databases. The following table summarizes these samples and the population that is represented by each one. This table is an essential tool for selecting the sample on which to test your research questions -- the sample used depends upon the nature of your investigation (cross-sectional vs. longitudinal, for example). The table lists the sample selection flag(s) and the associated statistical weight to use with each sample. You must tag these flags and weights in the ECB session in order to properly define your sample of interest. ____________________________________________________________________________________________________ Summary of NELS:88 populations, samples, level of analyses,sample identification flags and weights _____________________________________________________________________________________________________ Population of Interest 1988 8th Graders in 1990 (Panel): The population of 1988 eighth graders two years later (as of 1990). 1988 8th Grade Cross- Section: The population of all students enrolled in the eighth grade in 1988. Sample and Sample N (Participants) Base year retained sample members who completed both a base year and first follow-up questionnaire. N=17,424. Note: undercoverage bias; 5% of potential base year sample excluded. Base year selected sample members who participated in the base year. N=24,599. Note: undercoverage bias; 5% of potential base year sample excluded. Level of Analysis Cross-wave, longitudinal level of analysis. Cross-sectional level of analysis. Sample ID Flag Use the first follow-up student data file and select for the 8th to 10th grade panel, F1PANFLG=1. Use the base year student data file and select for BYQFLG=1. Weight Use the panel weight (F1PNLWT). Use the base year questionnaire weight (BYQWT). Summary of NELS:88 populations, samples, level of analyses, sample identification flags and weights, cont'd. Population of Interest 1990 10th Grade Cross- Section: The population of all students enrolled in the tenth grade in 1990. 1990 (First Follow-Up) Cross-Section: The population of all first follow-up eligible persons who were either in the eighth grade during the 1987- 1988 school year or were in the tenth grade during the 1989-1990 school year. Sample and Sample N (Participants) Representative sample of students enrolled in tenth grade in the spring term of 1990. Includes freshened students and excludes dropouts and out-of- sequence sample members. N=17,544 on the original F1 release. On the F2 re-release, BYI's who completed a F1 questionnaire appear producing a new N=17,754. 1990 sample members; combines 1988- eligible eighth- grade cohort and 1990 tenth-grade freshened sample. N=19,264. On the F2 re-release, BYI's who completed a F1 questionnaire appear producing a new N=19,394. Level of Analysis Cross-sectional analysis; Trend analyses with HS&B 1980 sophomores and F1 1990 sophomores. Sample ID Flag Use the first follow-up student file and select for F1SEQFLG=0. Use the first follow-up student file and select for F1STAT=0. Weight Use the first follow-up questionnaire weight (F1QWT). Use the first follow-up questionnaire weight (F1QWT). _______________________________________________________________________________________________________ Summary of NELS:88 populations, samples, level of analyses, sample identification flags and weights, cont'd. _______________________________________________________________________________________________________ Population of Interest 1988 8th Graders in 1992 (Panel): The population of 1988 eighth graders four years later (as of 1992). 1990 10th Graders in 1992 (Panel): The population of 1990 sophomores two years later (as of 1992). Sample and Sample N (Participants) Base year retained sample members who completed a questionnaire in all three waves of NELS:88--base year, first follow-up, and second follow-up. N=16,489. Note: undercoverage bias; 5% of potential base year sample excluded. Representative sample of students enrolled in tenth grade in the spring of 1990 (see definition of tenth grade cross- section above) who completed a questionnaire in both the first follow-up and second follow-up. N=16,749. Level of Analysis Cross wave, longitudinal level of analysis. Cross wave, longitudinal level of analysis; longitudinal trend analyses with HS&B 1980 sophomore cohort and F1 1990 sophomore cohort. Sample ID Flag Select for F2PNLFLG=1. Select for F2F1PNFL=2. Weight Use the second follow-up panel weight (F2PNLWT). Use the second follow-up panel weight (F2F1PNWT). _____________________________________________________________________________________________________ Summary of NELS:88 populations, samples, level of analyses, sample identification flags and weights, cont'd ____________________________________________________________________________________________________ Population of Interest 1992 12th Grade Cross- Section: The population of all students enrolled in twelfth grade in 1992. 1992 (Second Follow- Up) Cross-Section: The population of all second follow-up eligible persons who were in eighth grade during the spring of 1988 or in tenth grade during the spring of 1990 or in twelfth grade during the spring of 1992. Sample and Sample N (Participants) Representative sample of students enrolled in twelfth grade in the spring of 1992. N=16,114. Includes freshened students and excludes dropouts and out-of- sequence sample members. All second follow-up 1992 sample members; combines 1988- eligible eighth-grade cohort and 1990-eligible tenth grade cohort and 1992 twelfth-grade freshened sample. N=19,220. Level of Analysis Cross sectional analysis; Trend analyses with NLS 1972 seniors, HS&B 1980 seniors and NELS:88 1992 seniors. Sample ID Flag Select for F2SEQFLG=0. Select for F2STAT=00. Weight Use the second follow-up questionnaire weight (F2QWT). Use the second follow-up questionnaire weight (F2QWT). Note: For information on how to use school, parent and transcript data--and respective weights--with student data, consult the separate data file user's manuals for each of these components. Appendix C Extracting Raw Data through the Extract Program EXTRN[2P/4P].EXE is a program that allows users to subset variables and cases from data files on the compact disc. It is stored in ECBW\N2P or ECBW\N4P on the CD. EXTRACT operates on the codebook text file (N[2P/4P]CB.TXT, by default). It begins with the variables that have been selected for the codebook file during an ECB session and extracts these from the data files on the CD, writing them out to a separate ASCII data file. Optional filtering of cases that meet specified conditions is possible when variables are extracted from a single data file (the syntax for filter logic is explained below). At the MS-DOS prompt on your PC, enter the ECB-specific extract command and the program describes its operation. (The remainder of this appendix uses the BY-F2 public-use ECB, ECBN2P, as an example.) Use. The EXTRACT command takes the form: EXTRN2P C=%1 F=%2 [E=%3] [L=%4] [B="%5"] where the required parameters are: %1 = CD Drive Specification (ex. - "L:") %2 = Codebook Text File Name (ex. - "C:\NEL\N2PCB.TXT") and where the [Optional Parameters] are: %3 = Extract Data File Name [default = "XTRACT01.POS"] %4 = Record Length for %3 [default = 1024] %5 = Filter Logic ["%5" MUST be surrounded by double quotes AND there can be NO embedded blanks in the statement] [] = OR () = AND = Equals > Greater than < Less than >= Greater than or Equals <= Less than or Equals <> Not Equals Example of filter logic: "(F2SMPFLG=0)(F2SES1Q=4)" This statement instructs the computer to select all cases in which the respondent is a member of the eighth grade cohort (value of F2SMPFLG="0") and is in the highest socioeconomic status quartile (value of F2SES1Q= "4"). Please note that filter logic (the %5 specification) may NOT be used for extracts from more than one data file on %1 (the CD). If and only if there is NO filter logic (i.e., %5 is not used), then variables may be extracted from multiple modules. Operation. Login to the DOS directory that houses the %2 file (on your local drive, see example below) and enter as follows: >EXTRN2P C=%1 F=%2 [E=%3] [L=%4] [B="%5"] Example: L:\ECB\EXTRN2P C=L: F=C:\NEL\N2PCB.TXT E=XTRACT.POS L=1024 B="(F2SMPFLG=0)(F2SES1Q=4)" Given the instructions in the example above, the computer will: - read the codebook file that is specified in %2; - reference files stored on the CD drive specified in %1; - build an ASCII extract file named %3 on the user's hard drive (after applying the filter logic involving Grade Cohort and Socioeconomic Status as specified in %5) with logical record length = 1024; - create a new file with the name %2.NEW to describe the extract file.