All posts by Andreas Kaufmann

Final Thesis: QDAcity Quality Metrics

Abstract: The cloud-based QDAcity platform allows to conduct collaborative research projects applying Qualitative Data Analysis methods and research validation through crowdsourcing. When using Qualitative Data Analysis as a research method, expressive measurements of the quality and maturity of the results are essential to prove the validity of the research Findings. Common measures for validity of ratings are inter-coder agreement metrics, and for measuring the maturity of a qualitative research project, a prevalent approach is to calculate saturation. However, with a high number of raters, inter-coder agreement metrics become inconvenient to evaluate and the calculation of saturation requires a clean documentation of many project variables. The cloud environment of QDAcity can solve both of these problems, because it can efficiently store all ratings and project variables, and thus integrate both metrics more conveniently for the researcher. This thesis presents the implementation of the two inter-coder agreement metrics Krippendorff’s Alpha and Fleiss’ Kappa in QDAcity and a new approach with its implementation of theoretical saturation.

Keywords: Intercoder Agreement, Theoretical Saturation, Qualitative Data Analysis, QDA

PDFs: Master Thesis, Work Description

Reference: Matthias Schöpe. QDAcity Quality Metrics. Master Thesis, Friedrich-Alexander-Universität Erlangen-Nürnberg: 2017.

Final Thesis: Text Mining for Relationship Extraction

Abstract: Qualitative Data Analysis (QDA) methods are based on manual coding of texts. To extract a domain model from a text corpus using QDA, information has to be extracted and compiled into the domain model by hand. This is especially a problem for cases where large amounts of data have to be analyzed. For this purpose, We present a relationship extraction approach based on Natural Language Processing. It automates the extraction of relationships between codes that were provided by the coder. This speeds up the analysis process and helps to uncover relationships the human coder might have missed. Our method produces a graphical overview of relationships that were found to exist between codes. It is evaluated by comparison with previously generated models from existing Qualitative Data Analysis projects.

Keywords: Information Retrieval, Text Mining, Natural Language Processing, Qualitative Data Analysis, QDA

PDFs: Master Thesis, Work Description

Reference: Martin Hofmann. Text Mining for Relationship Extraction. Master Thesis, Friedrich-Alexander-Universität Erlangen-Nürnberg: 2017.

Final Thesis: A Metamodel for Code Systems

Abstract: Requirements elicitation is an important factor in software engineering. Mainly the information needed is elicited through interviews and other qualitative sources. The analysis that follows is often an ad-hoc process that relies on expertise of the analyst(s) and therefore is hardly replicable. Additionally, the process is not transparent as the resulting modeling elements cannot be mapped to the initial data. First attempts to solve this issues by adapting the clearly defined steps of Qualitative Data Analysis (QDA) suggest that the approach should be followed up. In order to further formalize the process this thesis suggests a metamodel which allows to derive structure and behavior models from the same coding process. The metamodel is derived by analyzing an existing metamodel and by comparing different existing coding systems and their resulting modeling artifacts. The metamodel is extended with a rule system and tested on an exemplary data set. For validation the resulting models are compared to models from an ad-hoc modeling process and evaluated by experts. Results show that utilizing QDA with a code system metamodel allows for an increase in transparency and makes it more easy to vary detail levels of the derived models.

Keywords: Domain Model, Domain Analysis, Requirements Engineering, Qualitative Data Analysis, QDA

PDFs: Master Thesis, Work Description

Reference: Sindy Salow. A Metamodel for Code Systems. Master Thesis, Friedrich-Alexander-Universität Erlangen-Nürnberg: 2016.

Final Thesis: Validating Requirements Specification Using Surveys

Abstract: The quality of the software requirements is vital to a project’s success, and the means of assessing the software requirements specifications quality are of high importance. This thesis presents a novel method of validation large requirements specifications using surveys. We developed a detailed method description, sketching a draft at the beginning of the research and constantly improving it after each method application on different industrial projects at Siemens AG. Moreover, the results were compared with the linguistic analysis of ambiguity indicators that present in the requirements and with the expert assessment on the requirements clearness. The result of method application is the metric, which shows the portion of non ambiguous requirements in the software requirements specification under study.

Keywords: Software Requirement Specification, Requirements Engineering, Requirements Validation

PDFs: Master Thesis, Work Description

Reference: Katsiaryna Krauchanka. Validating Requirements Specification Using Surveys. Master Thesis, Friedrich-Alexander-Universität Erlangen-Nürnberg: 2016.

Technical Report: Using Students as a Distributed Coding Team for Validation through Intercoder Agreement

Abstract: In qualitative research, results often emerge through an analysis process called coding. A common measure of validity of theories built through qualitative research is the agreement between different people coding the same materials. High intercoder agreement indicates that the findings are derived from the data as opposed to being relative results based on the original researcher’s bias. However, measuring such intercoder agreement incurs the high cost of having additional researchers perform seemingly redundant work. In this paper we present first results on a novel method of using students for validating theories. We find that intercoder agreement between a large number of students is almost as good as the intercoder agreement between two professionals working on the same materials.

Keywords: Qualitative Data Analysis, Theory Triangulation, Intercoder Agreement, Distributed Coding, Collective Coding

Reference:  Andreas Kaufmann, Ann Barcomb and Dirk Riehle, “Using Students as a  Distributed Coding Team for Validation through Intercoder Agreement,” Friedrich-Alexander-Universität Erlangen-Nürnberg, Dept. of Computer Science, Technical Reports, CS-2016-01, April 2016.

The report is available as a PDF file.

Final Thesis: The Use of Domain Knowledge for Structuring and Describing Code Systems

Abstract: Domain knowledge is a widely used but seldom defined term in research. Since every research project is located in at least one dedicated domain, its influence to the research work seems to be natural but is not investigated, especially not for qualitative research. An exploratory study was performed interviewing eight qualitative researchers located in different domains: economic education, computer science and business sciences. The interviews were qualitatively analysed using Grounded Theory Methodology (GTM). Of particular interest were possibly existing schematic procedures the researchers apply during their qualitative research process and especially during their coding process and if existing how they are rooted in the domain. With the set-up and reuse of a domain specific code schema in one domain the existence of schemata and its integration to QDA could be confirmed. Furthermore, a suggestion was given on how to generate artefacts for Requirements Engineering (RE) in the form of User Stories based on a code schema. Additionally, a code system meta model has been created to clarify the influence of domain knowledge from an abstract perspective.

Keywords: Qualitative Data Analysis, Grounded Theory Methodology, Requirements Engineering, Domain Knowledge

PDFs: Master Thesis, Work Description

Reference: Rebecca Reuter. The Use of Domain Knowledge for Structuring and Describing Code Systems. Master Thesis, Friedrich-Alexander-Universität Erlangen-Nürnberg: 2016.

Final Thesis: Integrating Multiple Views In A Code System

Abstract: Mature requirements engineering is one of the key success factors of professional software development. This thesis proposes an approach that adapts core principles of qualitative data analysis, which is well established in social science, and transforms them into a holistic method for requirements analysis. Our method addresses the process steps between the analysis of the target domain and the generation of a requirements specification and different types of domain models. In particular, it aims at inherently providing a high degree of traceability by the institution of an additional artifact, the so-called code system. We state this code system to introduce significant advantages to the quality of requirements analysis with respect to systematics, documentation, maintainability and efficiency. The proposed approach features the direct deduction of different types of domain models out of the gathered domain knowledge and a substantial support for the development of a requirements specification.

Keywords: Domain Analysis, Domain Modeling, Qualitative Data Analysis, Requirements Engineering, Requirements Specification

PDFs: Diplomarbeit, Work Description

Reference: Florian Schmitt. Integrating Multiple Views In A Code System. Diplomarbeit, Friedrich-Alexander-Universität Erlangen-Nürnberg: 2016.

DATEV Appreciation of Students for AMOS 2015 Projects

DATEV eG sponsored two AMOS projects this last summer. They went the extra mile to show their appreciation towards our students for their hard work in these projects: Each student received a document about their participation expressing our industry partner’s satisfaction about it.

Continue reading DATEV Appreciation of Students for AMOS 2015 Projects

Final Thesis: Improving Domain Modeling And Requirements Analysis Using Grounded Theory

Abstract: One of the key factors of the success of professional software development is mature requirements engineering. This thesis focuses on the elicitation and analysis of requirements and addresses the process steps between the elicitation of requirements and the gathering of the system requirements speci cation and domain model. We consider the state-of-the-art of these aspects of requirements engineering as suboptimal and propose an approach that implements the elicitation and analysis of requirements as an adaption of the Grounded Theory approach, which is methodically sound at the eld of social studies. Our approach will implicate the institution of an additional artifact within the process, the  so-called code system. Furthermore this approach enables the direct mapping from the  gathered information about the target domain, represented in the code system, into a  domain model.

Keywords: Domain Analysis, Domain Modeling, Qualitative Data Analysis, Requirements Engineering

PDFs: Studienarbeit, Work Description

Reference: Florian Schmitt. Improving Domain Modeling And Requirements Analysis Using Grounded Theory. Studienarbeit, Friedrich-Alexander-Universität Erlangen-Nürnberg: 2015.

Final Thesis: Developing a Domain Analysis Procedure based on Grounded Theory Method

Abstract: Domain analysis is the process of analyzing and modelling the domain in which a future software system is supposed to operate. It is an essential step in requirements engineering (RE) and therefore critical for the success of software development projects. However, common methods for deriving a domain model from natural language descriptions do not address the difficulties of abstracting a complex domain sufficiently and depend on the analyst’s experience and expertise. Grounded theory method (GTM) offers a techniques for breaking up and abstracting qualitative data by developing and relating concepts. Its use can therefore improve the procedure of extracting the important entities of a domain and their relationships, while ensuring traceability between the data and the derived domain model. This thesis shows how GTM has to be adapted for its successful utilization in RE. For this purpose, we applied GTM to a domain analysis example and derived a systematic procedure for domain analysis.

Keywords: Domain Analysis, Domain Model, Requirements Engineering, Qualitative Data Analysis, Grounded Theory Method

PDFs: Master Thesis, Work Description

Reference: Katharina Kunz. Developing a Domain Analysis Procedure based on Grounded Theory Method. Master Thesis, Friedrich-Alexander-Universität Erlangen-Nürnberg: 2015.