The work of our institute follows the “Standards zur Qualitätssicherung in der Markt- und Sozialforschung” (standards of quality assurance in social and market research) and already sticks to the rules of the just recently published standard DIN ISO 20252 "Markt- und Sozialforschungsdienstleistungen" (social and market research services).

We attach special importance to the following issues:

Basic Quality Management

In our institute, we established a quality management system. We use this system to analyze the research process of each project in order to detect existing weaknesses and to avoid them in future.
On a regular basis the quality system is subject to internal examination to prove its full functionality. Causes of possible problems or inconsistencies are analyzed and immediately eliminated to prevent further damage.
The research process is under continuous monitoring and control to prove that the performance standards agreed upon are being carried through. The most important features of the research process include drawing of a sample, development of a questionnaire, data survey and processing, analysis and reporting.
All confidential or sensitive material, documents, products and data bases provided by the client are kept strictly confidential and are only accessible to the persons involved in the project.

Selection and Training of Interviewers.

The interviewers are selected and trained for their cooperation with us in a multilevel screening process. An initial aptitude test is given to candidates at the application stage. Via a first telephone conversation our experienced staff gets a general idea of the communication skills of the applicants. Only applicants who meet the basic requirements are interviewed personally and get further training.
After a first training session of several hours when the applicants are acquainted with the basic aims, methods and principles of social and market research and the requirements of data protection, an introduction into computer technology and questioning techniques follows.
Subsequently, the applicants are faced with several test interviews under supervision of our staff, who give them help and advice throughout the time. Only those applicants, who prove their aptitude in direct contact with respondents and achieve a high level of sample coverage in interviews are finally accepted in our team.

Before starting a project all interviewers again receive intensive training and advice on the specifications of the new project.
On average 10 to 15 percent of the interviewers are monitored, mainly by direct control. With those who are not able to meet the required quality standards, despite additional training, collaboration is not continued.

Validation and Protection of Collected Data.

The collected data is continuously checked for completeness, plausibility and consistency. If necessary, missing data is resurveyed and wrong data is excluded from the stored portfolio.
Reviewing data also comprises the regular check for completeness and compliance with the sample parameters, the check for plausibility of the answers and a comparison of the answers with standard data or among different field interviewers. Some of these checks are carried out manually, but most are computer-based.
By the way we store our data we ensure that it is neither getting lost nor being changed and that confidentiality is maintained. All project records stored electronically are secured by at least one data copy not stored on the institute’s premises.
All steps of data processing are kept continuously and several backups are stored (RAID technology, various backup systems). All data is maintained in original. Thus every step of data processing and cleaning can be reconstructed in detail and all changes made to the original data are recorded.
Additionally, all data cleaning defaults as well as the weighing methods and programs used are documented and stored.
Working with data files we ensure that there are no double entries, that we always use the newest version and that we are always able to compare them with the original data. All files of data processing are unmistakably labelled and thus can be easily allocated to the working steps concerned.
As far as statistic interpretation is concerned we present a weighted and non weighted basis separately.
By scrutinizing all data processing and interpretation steps again retrospectively we guarantee that data do not get lost or distorted.