Quality of Conclusions sections: a qualitative inquiry

worked on by: Ailis Oßwald


If you regularly read scientific software engineering articles the thought, that a conclusion section seems unsatisfying, has probably crossed your mind, more often than not. This rather subjective thought inspired Prechelt to introduce me to the topic of my master thesis, which shall start to find the problems in conclusion sections.

To find such problems I use Grounded Theory Methodology (GTM) to find codes and, more importantly, hints to identifiy shortcomings of conclusions. With such qualitative research methods I try to explain what kind if bad conclusions exist and guess what the authors purpose for this conclusion was. At the end of my thesis we hope to have codes, quality problems and ways to identify them grounded in data. These findings should then be used to quantify a large amount of conclusion sections with the purpose of making the subjective impression of "bad conclusions" more objective.

A first very preliminary codebook has been developed and introduced to a dozen students learning GTM. These students got a slimmed down version of my codebook, modified by Prechelt. This codebook version contains only codes and very short descriptions, no unfinished guesses and thoughts about what those codes may mean, to give the students freedom of thought and let them learn GTM. By now the students have modified the codebook and are moving from theier Open Coding Phase into the Content Analysis, which will generate some more coded data. I will use this data for further analysis like the students and can hopefully refine some of my ideas regarding quality problems.