Eric A. Hanushek, John F. Kain, Jacob M. Markman, and Steven G. Rivkin, 
"Does peer ability affect student achievement?", Journal of Applied 
Econometrics, Vol. 18, No. 5, 2003, pp. 527-544.

Because of federal and state confidentiality laws and regulations, the data
used in this paper are not currently available to other researchers.
Nonetheless, interested researchers should contact the authors, because
there are on-going efforts to open access to the data.

The UTD Texas Schools Project Microdata

The matched panel data sets used in this paper combine information from 
several data files produced by the Texas Education Agency. The test scores 
come from the Texas Assessment of Academic Skills file, the information on 
family background is contained in a Demographic file, and teacher 
information comes from the Staffing and Classroom Assignment files.  The 
student data sets are merged using unique student identifiers, and the 
teacher information is merged with the student data by year, grade and 
school. Note that it is not possible to separate students by classroom. 
Therefore peers are defined as students in the same school, grade, and year.
The data track the universe of three successive cohorts of Texas public 
elementary students as they progress through school, beginning with 
students who attended third grade in 1992. The student data contain a 
limited number of student, family, and program characteristics including 
race, ethnicity, gender and eligibility for a free or reduced price lunch 
(students eligible for any type of subsidy are lumped together) and 
compensatory education services.

Beginning in 1993, the Texas Assessment of Academic Skills (TAAS) was
administered each spring to eligible students enrolled in grades three 
through eight. The criteria referenced tests evaluate student mastery of 
grade-specific subject matter; we use the mathematics test results in this 
paper. Each math test contains approximately 50 questions. Because the 
number of questions and average percent right varies across time and 
grades, we transform test results for each year and grade into standardized 
scores with a mean of zero and variance equal to one. In order to avoid 
complications associated with classification as limited English proficient 
(LEP) or disabled, all LEP and special education students are dropped from 
the analysis but not from the calculation of the peer variables. The peer
variables are calculated over schoolmates in the same grade and year. Peer
average achievement in grade g equals the average test scores in grade g-2
for schoolmates in grade g, while peer percent eligible for a subsidized
lunch equals the percent of grade g peers so eligible.