National Data Management Unit

The work of the National Data Management Unit follows the data collection phase (field operations). After the test and survey instruments are processed and their results recorded, the national data management unit provides the study's scientific directors with data records prepared from those results. These data records constitute the basis for further scientific analysis and reports on study results. 

National data management may be broken down into the following specific areas: 

Incoming packages containing test booklets and questionnaires are first subjected to a receiving inspection. As a second step, the contents of the test packages are inspected for completeness by a database-based collection system, which uses a barcode-reading system to compare printed and returned instruments. Apart from thorough documentation of the returned materials, other actions might include, for example, contacting the schools or outside test administrators to locate missing test instruments. At the end of the logistics phase the instruments are sorted by instrument type for inclusion in the next processing stages.

Data collection begins with the creation of what are referred to as codebooks. These contain the definitions necessary for translating pupils’ answers into interpretable numeric codes. Depending on the instrument type or scope, manual or scanner-assisted data collection methods may be used.

The data from smaller batches of questionnaires and from such survey documents as pupil lists and test session records are thus entered manually for the most part. Input masks are set up for this purpose, and these also enable transcription of answers to open-ended questions (e.g., occupational information on questionnaires).

Scanner-assisted data collection is accomplished using high-performance scanners with a daily capacity of approximately 80,000 sheets. Once the test booklets and questionnaires have been processed, they are scanned and stored as image files. At the same time, selected details from the scanned pages are provided as material to be used for coding open-ended answers.

Markings in multiple-choice answer fields are recorded automatically with the use of pre-defined data entry masks and converted to data or numbers that can be analyzed. Invalid answers are identified during the data collection process and manually corrected as necessary (verification of data).

The data matrix derived from the acquired data is subjected during the data cleaning process to formal and content validity checks. A formal validity check involves a comparison of the collected data with the printed, returned and scanned test booklets and questionnaires. For a content validity check, the values of the individual variables (i.e., the answers actually given for items or to questions) are compared with the range of possible answers previously defined in the codebook. All stages in the data cleaning process are controlled through programs. At the end of the data cleaning process, edited SPSS or SAS system data records can be provided to the clients.

The images created during data collection are transferred to an electronic archiving system and likewise made available to the project's scientific directors. Those materials that are not electronically recorded are transferred to external storage and preserved in strict compliance with the provisions of data protection law until the project is completed.

Jens Gomolka
IEA Data Processing and Research Center
Mexikoring 37
22297 Hamburg
Germany

Phone: +49 40 48 500 602
Fax: +49 40 48 500 501
Email: mailiea-dpc.de