![[Table of Contents]](/icons/up.gif){kind=icon}
The National Education Longitudinal Study of 1988 (NELS:88) is a survey that began with a nationally representative sample of 1988 8th-graders and surveyed them every two years. The most recent follow-up survey occurred in 1994. Respondents’ teachers and schools were also surveyed in 1988, 1990, and 1992, while parents were surveyed in 1988 and 1992. In contrast to previous longitudinal studies, NELS:88 began with 8th-graders in order to collect data regarding the transition from elementary to secondary education. The first follow-up in 1990 provided the data necessary to understand the transition. Dropouts were administered a special survey to understand the dropout process more thoroughly. For the purpose of providing a comparison group to 1980 sophomores surveyed in one High School and Beyond Institutional Study (HS&B), the NELS:88 sample was also "freshened" with new participants who were 10th-graders in 1990.
In spring of 1992, when most of the NELS:88 samples were 12th-graders, the second follow-up took place. This survey focused on the transition from high school to both the labor force and postsecondary education. The sample was also "freshened" in order to create a representative sample of 1992 seniors for the purpose of conducting trend analyses with the 1972 and 1982 senior classes (NLS-72 and HS&B). Students identified as dropouts in the first follow-up were also resurveyed in 1992. In spring of 1994, the third follow-up was administered. It was this follow-up that provided information about postsecondary enrollment experiences used in this report. Sample members were also questioned about their labor force participation and family formation. For more information about the NELS:88 survey, consult the NELS:88/94 Methodology Report.[7]
Because this analysis was concerned with "resilient" students at risk--high school graduates who, in the 8th grade, were at risk of dropping out but who did not drop out--the NELS 1992 high school graduate cohort was used as the base sample (i.e., F3QWT92G weight variable was used). In addition, only students considered at moderate to high risk of dropping out of high school were included in the analysis (see description of risk variables under the BYRISK2 entry in appendix A). Moderate to high risk was defined as having two or more of six possible risk factors. Among this sample, students also had to have values for the three outcomes analyzed (for description of outcome variables, see "Statistical Procedures" below).
There were three outcome variables analyzed in the study. Two describe at-risk students’ postsecondary education enrollment status within two years after high school graduation: (1) attended a 4-year college versus all others (including a non-4-year college or no college at all); and (2) attended any postsecondary education versus those who never attended. The third outcome was a measure of postsecondary persistence and classified students with (and without) indicators of strong postsecondary persistence among those who enrolled. Because all three outcomes are dichotomous variables (0=No and 1=Yes), a logistic regression model was used to perform multivariate analyses. In order to take into account the complex nature of the NELS survey design, SUDAAN software application was used. SUDAAN uses a Taylor series approximation technique to estimate standard errors for the logistic regression estimates, a method that takes into account the stratified sample design of the survey. For a detailed description of SUDAAN, please refer to SUDAAN Users Manual. The syntax of the SUDAAN program is illustrated by the following example, which produced the output for step 4 in Table 7:
PROC LOGISTIC DATA="c:\\logcoll" FILETYPE=sas DESIGN=WR; NEST strata psu; WEIGHT f3qwt92g; SUBGROUP f2ptalk f1parexp f1attend f1excur f1frstud f2frcoll aidinfo1 aidtalk1 outreach f2helpex f2helpap SINGLE HELDBACK sibdrop mthqual8; LEVELS 4 6 4 4 4 4 4 6 2 3 3 2 2 2 9; MODEL coll4yr = f2ptalk f1parexp f1attend f1excur f1frstud f2frcoll aidinfo1 aidtalk1 outreach f2helpex f2helpap single heldback sibdrop byses byp40 bygrd68 preach mthqual8; SETENV COLWIDTH=11 DECWIDTH=2 COLSPCE=2 LINESIZE=78; PRINT BETA SEBETA T_BETA P_BETA; PRINT / RISK=all; PRINT / TESTS=default; TITLE "Attended a 4-year college or not".
In this program, the first command is procedure statement that indicates a logistic regression procedure. "LOGCOLL" is the SAS system file that stores the data for the study. Design statement is "with replacement." "STRATA" and "PSU" are stratum and psu variables derived from the student ID in NELS:88. The subcommands of "SUBGROUP" and "LEVELS" indicate the levels of all categorical variables included in the model. For variables with 10 percent or more missing cases, a "missing" category was included in the analysis. The subcommand of "MODEL" specifies the logistic regression model with the dichotomous outcome variable of "COLL4YR" (enrolled in a 4-year college) and all independent variables that we were interested in examining.
