Documentation for student graduation rate dataset
Dataset format:
SASv8.2 for Windows, created on W/2000.
Dataset names:
Sdfxset4_nf New freshmen entering fall 1987 to spring of 1999
Sdfxset4_tr External transfers entering fall 1987 to spring of 1999
Flcoll Format for college code to text
Gleth Format for ethnic code to text
Ulcurnc Format for curriculum to program code
Ulcip Format for program code to text
Definitions and variables in sdfxset4_nf:
New freshmen are officially students starting within a year of graduating from high school and with less than 12 transfer credits. A new freshmen starting in the summer session before the first fall counts as starting in the fall for calculating graduation rates. NC State systems record the student with an entering status of new in both semesters.
Blindid randomized unique identifier
Ssubdate date of record, YYYYS, S=semester, 1=spring, 7=SS1, 8=SS2, 6=fall (yes I know they are out of order)
Sbyr birth year
Shsgryr high school grad year
Senrstat enroll status, 1=new freshmen, 2=external transfer
Ssex gender
Sbmth birth month
Scitiz citizenship, C=citizen, R=resident, N=non-citizen
Srescrhr credit hours this (first) semester, 12 and above=full-time
Ssatv SAT verbal
Ssatm SAT math
Sact ACT composite
Shspctl high school rank percentile in class (99 = top 1%)
Shsgpa high school weighted grade point average
Supgpa predicted grade point average
Scoll college code first semester
Sclass classification, should be freshman for all in this population
Scurr curriculum first semester
Sschch1 college choice 1 on admissions application (1994 and later)
Scurch1 curriculum choice 1 on admissions application
Sschch2 college choice 2 on admissions application
Scurch2 curriculum choice 2 on admissions application
Srace ethnic code
Hscode high school code
Rescode residence code
Ret2date retained date for 1-year retention (data=enrolled)
Ret2scoll retained college for 1-year retention
Ret2scur retained curriculum for 1-year retention
Ret3date retained date for 2-year retention check (data=enrolled)
Ret3scoll retained college for 2-year retention
Ret3scur retained curriculum for 2-year retention
Gcurr degree awarded in this curriculum
Gdegtype degree awarded (B=bachelors, blank=no degree)
Ggpa graduation grade point average
Gcollege degree awarded in this college
Yyyys degree awarded date
Fread reading score advanced placement
Fvocab vocabulary score advanced placement
Ftswe score advanced placement
Fpge predicted grade in English
Fpgm predicted grade in math
Fmac1 first math course
Fmag1 grade in first math course
Fenc1 first English course
Feng1 grade in first English course
Fchc1 first chemistry course
Fchg1 grade in first chemistry course
Fbic1 first biology course
Fbig1 grade in first biology course
Gpa_spring grade point average at end of spring semester 1st year (missing value implies student did not finish spring semester)
Gpa_fall grade point average at end of fall semester 1st year (missing value implies student did not finish fall semester)
Definitions and variables in sdfxset3_tr:
External transfers are officially students starting with more than 12 transfer credits and have a transcript from at least one institution of higher education. It is possible to be an external transfer from NC State via the LLE program. The starting classification is dependent on the number of credits transferred into the degree program. The freshman attributes are typically not used in making admissions decisions (test scores, HS performance, etc.)
Blindid randomized unique identifier
Ssubdate date of record, YYYYS, S=semester, 1=spring, 7=SS1, 8=SS2, 6=fall (yes I know they are out of order)
Sbyr birth year
Shsgryr high school grad year
Senrstat enroll status, 1=new freshmen, 2=external transfer
Ssex gender
Sbmth birth month
Scitiz citizenship, C=citizen, R=resident, N=non-citizen
Srescrhr credit hours this (first) semester, 12 and above=full-time
Strcrhr transfer credit hours, depending on program student is in, number of transfer credits accepted into program will be less than or equal to this number
Scoll college code
Sclass classification, should be freshman for all in this population
Scurr curriculum
Srace ethnic code
Rescode residence code
Gcurr degree awarded in curriculum
Gdegtype degree awarded (B=bachelors, blank=no degree)
Ggpa graduation grade point average
Gcollege degree awarded in this college
Yyyys degree awarded date
Rules and notes:
The rule for calculating graduation rate is based on entering semester and number of years to graduation. Part-time students in the first semester are not in the calculation population. Students entering in fall of 1991 and graduating by the end of summer sessions in 1995 have graduated in 4 years. The standard calculation used by NCAA and federal rules is 150% of normal degree period, so the comparison norm for undergraduates is the 6-year rate. If a student has multiple degree records, the first one in time is the one that counts.
Notes about the data:
- A student can start more than once, and a student can have more than 1 degree. Here are some examples:
§ ID 001000069 has two entry records, summer2=19928 and fall=19926. (For the purposes of this analysis, the earlier can be ignored / deleted.)
§ ID 001000042 has three degree records all in the same semester.
§ ID 001000092 has two degree records in different semesters.
- Missing data is represented by blanks or the SAS standard dot (.) for numeric fields.
- Some data has been coded as unknown with a value of 9s. An SAT or ACT of 9s should be treated as missing data.
- Some students come in with no test scores or no high school record. These include home school-ers, foreign students, and several other special cases.
- In the “old days” before about 1996, the high school grade point average was unweighted, meaning the raw score was recorded and not weighted to reflect the maximum grade point possible at that high school. For instance, some high schools will award a grade above 4.0 for advanced placement work, while other high schools have no advanced placement available and will award no grades above 4.0. In the “old days” there was no data mechanism to identify the maximum possible GPA and weight each student’s score appropriately.
- Typically the fall semester full-time population is used to calculate the graduation rate. The federal rules allow for inclusion of the spring entrants in the same cohort, however they are not given time credit for being a semester behind. I would like to know if they help or hurt the graduation rate. The semester schedule at NC State may penalize a spring entrant in some disciplines where there are course pairs that must be taken in fall-spring order.
- The 1-year and 2-year retention rates are measured in a mechanical check of student’s enrollment in the system at the year-later time point.
- Program codes are called CIP for Classification of Instructional Programs. This is a federal NCES taxonomy based on 6 digit codes where the first two digits are general disciplinary groups, 4 digits more specific, and 6 digits are very specific. For example all engineering are in the 14xxxx group, and mathematics and statistics are both in the 27xxxx group. The 4 and 6-digit codes map to our curriculum codes, and the 2 digit codes map somewhat to college codes. A general rule is that a student is considered to have graduated in the same discipline if graduating with the same 2 digit discipline as they started in,. A CIP code of all zeroes means undesignated, or no program, and the student cannot graduate with a non-program code.
Formats included with datasets:
College code to college text
Curriculum code to 6-digit discipline code
Discipline code to text
Race code to text
Questions
- What are the attributes that predict graduation? Ideally admissions attributes (test scores, admissions choice, residence, high school, etc) can be separated from behaviors as an enrolled student (course grades, first year GPAs, discipline same or different, etc.)
- What are the attributes that predict retention for the first or second year?
- Admissions choice has been a big topic lately. The “choice” phrase is shorthand notation for whether the student can matriculate into the program they want to or not. Typically, students have to meet a GPA minimum to get into the more competitive programs such as Design, Engineering, Pre-Med, or Management. Some students meet the minimum and enter directly as new freshmen, others may not meet minimums and instead enter the First Year College (FYC starting in 1995). This may be seen as a choice for one of those colleges, but enrollment in FYC. Later the student has the grades to transfer into the original choice and hopefully graduates in the originally desired program, or does not have the grades and finds another ‘happy’ choice, or stays unhappy and drops out, or otherwise does not graduate. See IDs 001000014 (no choice for FYC) and 001000122 (FYC was second choice) for examples of students enrolling in FYC who may have preferred to be in another program. Does the choice difference predict a behavior that we can anticipate?
- Does Fall or Spring entry make a difference in graduation or retention?
Documentation for student graduation rate dataset.doc
1/27/2003 Page 4 of 4