Access code
Access code list and representativeness
The access code list is created as follows:
- The university informs ISTAT of the number of graduates to be surveyed.
- ISTAT generates the corresponding number of anonymous access codes and sends these to the university.
- The access codes (Zugangscodes) are distributed randomly by the university among the graduates.
- The names and addresses of the graduates are deleted.
This results in a list structured as follows:
- The university sends the list, set out like this, to the Institute for Applied Statistics.
- The information is thus supplied without any connection to names or addresses.
- When the participant logs in with the access code, the information on the field of study (Studienfach) and degree type (Abschlussart) is retrieved from this list and automatically carried over into the survey.
- There is an option to correct the information during the survey.
Demographic basic values for verifying representativeness
Every university can add to this access code list with the following attributes:
- Final grade (Abschlussnote)
- Gender (Geschlecht)
- Year of birth (Geburtsjahr)
- Information as to whether the participant is a German citizen (Staatsbürgerschaft: nur Deutsch), has both German and a different citizenship (Staatsbürgerschaft: Deutsch + andere) or just a different citizenship (Staatsbürgerschaft: nur andere)
This results in an expanded access code list for verifying representativeness, which can contain the following information at most:
- The university sends the list, set out like this, to the Institute for Applied Statistics.
- The information is thus supplied without any connection to names or addresses.
- This information forms an exhaustive picture of individual parameters of the population set, enabling scientific analysis.
Quality assurance according to representativeness check
An essential quality criterion for the meaningfulness of a study is whether or not the group of participants is comparative with the population set (all graduates). If this comparability is present, the results are deemed “representative”.
Various errors can mean that this goal is not achieved.
This happens, for example, when human error creeps into the process of sending out the invitations: a study programme is forgotten or the mistakes are made in preparing the letters (a sorting error in Excel, for instance) and they fail to reach their destinations. Errors may also occur in writing to students living abroad (wrong postage, wrong address).
Apart from these sources of error, errors may pop up that were unknown before.
This makes it essential that changes in participation are monitored as closely as possible so that anomalies can be countered quickly. Ultimately, there should be usable results at the end of the project so that the time invested by scientists, university staff and the participating graduates is not in vain.
For this reason, the information described above undergoes constant examination using a statistical testing process (chi-squared test). A few of the test results are shown below as an example.