Abstract:

The classical way to store and analyze empirical research data in social science and education research, is a program application. Excel and SPSS are typical examples. Data is stored in generic formats and analyzed with specific program scripts. For example, binary or text-oriented memory formats are usual. This procedure is suitable if only small quantities as well as little branched measurement data have to be analyzed and treated by one or few researchers. However, this system is problematic whenever:
(1) many people work simultaneously on data sets (uploading, editing and analyzing),
(2) data have widely branched dependencies,
(3) complex questions, complex query and analysis methods are required, and
(4) if Excel and SPSS are overwhelmed with the storage and display of large data sets.

This requires a change in the classical analysis path, where relational database management systems (DBMS) can offer solutions.

A research project on teaching quality of university instructors (“SoKonBe”) is used to show the applicability of the database management systems. In this contribution, we present a new way for a better, easier, faster and more valid storage and analysis of research data. In addition we present a self-test to check the need for the use of databases in different research areas.

In the research project we actually store 30 million records in approximately 300 variables. This is required because of the branched dependencies we previously addressed. The total sample contains data from 44 lectures by 43 instructors (male = 55.8%; female = 44.2%). The number of participating students per lecturer varies between N = 10 and N = 124 students. The total sample of the student participants contains 1714 individual details. In the pretest of the project an average of 70 minutes of the lecture of every instructor were filmed with three cameras. Every student filled in multiple questionnaires about the lecture, the instructor and about attractiveness of the lecturer. After this up to five student research assistants analyzed the videos based on 72 categories. With a DBMS it was possible to store, manage and aggregate the data without the need of a dedicated data center.

Considering the steadily growing amount and heterogeneity of data (questionnaires, video, sensor data, interviews…), DBMS offer innovative, efficient and future-proof data analysis opportunities in the whole domain of psychological and educational research. Minimal data redundancy due to separate tables and relations, improved data consistency and data accessibility, multiple users as well as program-data independence are possible.

Keywords:

Database management system, research, big data, psychology, education.