Here is a description of files in this directory grab-nels2.lsp This file extracts 15 variables from nels88_2.dat and puts them in a clean format, with spaces between variables. --------------------------------------------------------------------- multi2.dat This is the output of running grab-nels2.lsp. The variables are in the following order: "School_ID" "Stud_ID" "Sex" "Race" "BYS79A" "G8CTRL" "SES" "BYPARED" " BYTXMSTD" "BYSC47D" "BYSCENRL" "G8URBAN" "G8REGON" "G8MINOR" "BYRATIO" I have split multi2.dat into separate student and school files: (MISSING CODES ARE IN PARENTHESES) multi2-stud.dat "SCH_ID"() "STUD_ID"() "SEX"() "RACE"(8) (24,599 X 8) "BYS79A"(96 98) "SES"(99.998) "BYPARED"(8) "BYTXMSTD"(999.998) multi2-sch.dat "SCHOOL_ID"() "G8CTRL"() "BYSC47D"(8) "BYSCENRL"() (24,599 X 8) "G8URBAN"() "G8REGON"(8) "G8MINOR"(8) "BYRATIO"() multi2-school.dat Same as above, except redundant cases have been (1052 X 8) deleted. ----------------------------------------------------------------------- reg.lsp A useful file for loading multi2.dat into XLISP-STAT and defining variables. ------------------------------------------------------------------------------- The nels88_2.dat master data file has been split into two files, one for school level variables, the other for student level variables. 1) nels88-school.dat (1052 X 37) 2) nels88-stud.dat (24,599 X 52) Note: the two files yield 89 (instead of 85) variables since sex and race have been added, student ID is separated from school ID in the student file, and school ID appears in both files. (thus, we add 4 to the original 85 to get 89.) ----------------------------------------------------------------------------= new-grab.lsp A file that employs dialogs to allow the user to select variabes from nels88-school.dat and nels88-stud.dat. To use this file: > (compile-file "new-grab.lsp") > (load "new-grab") > (school-data) ; access function school-data > (stud-data) ; and stud-data, no arguments Note, the file nels.lsp contains the stud and school matrices used in the Monte Carlo simulations, and their corresponding labels.