State Augmentation

Appendix H
State Augmentations to HSLS:09

Contract Addition: State Augmentations

NCES has received funds from NSF to produce state-level datasets for 10 states in HSLS – California, Florida, Georgia, Michigan, North Carolina, Ohio, Pennsylvania, Tennessee, Texas, and Washington. This effort is called the State Supplement to HSLS. The original HSLS design was developed to produce efficient national estimates, for which schools were selected from lists sorted by state to ensure a geographical representation. But with the State Supplement, power analyses determined that 40 responding schools from each of the 10 states would be sufficient to meet the goals of a state representative dataset that could be used to conduct state representative analyses. Samples for the national design within California and Texas met this minimum size criterion. A Keyfitz sampling procedure was implemented for each of the remaining augmented states with the goal of maximizing the overlap with the national design sampled schools, and minimizing the overlap with the schools selected for the HSLS:09 field test and 2009 PISA.

In addition to sampling, recruiting, and collecting data from the additional schools for the State Supplement, data will be collected from State administrative records for each participating student, contingent upon agreements with each State, as part of the SDRA (State Data Record Augmentation). State officials and data personnel will be contacted early to gain their cooperation and to determine the requirements for collecting the State records, a process that is already underway. To comply with the privacy and data security regulations of individual States, NCES will enter into a binding agreement with each State that identifies the specific data items requested and documents the security procedures that will be in place to protect the data. These agreements will also document the procedures to be followed in transmitting the data and in returning the data to the State (or procedures for destroying the data after the merge).

States which participate in the SDRA will be asked to provide data on seven variables: (1) Course titles; (2) Course grades; (3) Entry/exit codes; (4) Retention – yes/no; (5) Test scores; (6) Attendance records and (7) a Data dictionary to learn how States define the data elements they have. Despite our efforts for simplicity, this is a complex process which will be highly variable by state. States will be asked to provide as much of the requested information as they have available for all of the 9^th graders in selected schools. This means that the State knows the schools, not the students, which are participating in HSLS, thereby maintaining student anonymity. To facilitate the linkage between the student and the State data, each State will be asked to identify one or more variables in common at the state and school levels, such as a state-level ID number, to be included on the student list collected from the school.

The security agreement with each State will bind NCES to an obligation to apply disclosure procedures consistent with ESRA to protect the identities of all students and schools in any public releases of the data. Further, each security agreement will stipulate that NCES will not include any direct student identifiers in restricted access (use) research files.

Parent consent forms provided to schools in each State that agrees to participate in SDRA will be customized to inform parents of the specific information that will be provided by the state on behalf of their teenager. Student data collection in the schools is scheduled to end in mid-December 2009. In early January 2010, NCES will share agreed-upon information about participating schools with the States to commence the state-level administrative records collection. States will be asked to provide the requested data by the end of February 2010. After quality control checks have been performed on the data, NCES’s contractor, RTI International, will merge the HSLS:09 and available State data. After delivery of a master HSLS data file, State data will be returned or destroyed (per agreement with the particular state).

With the SDRA, there is a possibility that administrative records held at the State level in data warehouses or other such entities will be merged with HSLS data. The process of preparing the State administrative records file will require at least one senior and one junior staff person in each State to conduct the data file preparation process and upload the file to NCES. NCES’s contractor, RTI International, will merge the State data file with HSLS:09 data. The estimated burden presented in Table 8 of the clearance package’s Supporting Statement represents a rough estimate of the time this process will take and the burden imposed on each state. The expected response rate is likely to be less than 100%, however the numbers in this table assume the best case scenario for data collection and collaboration with States.

0