Statistical noise or valuable information

Mehr zum Buch

Contemporary marketing research often relies on empirical studies and quantitative data analysis. Researchers frequently encounter the challenge of a small number of subjects that deviate significantly from the rest of the data. This raises the question of whether these outliers are errors needing correction or phenomena deserving further exploration. Clemens Pirker’s dissertation addresses this issue, highlighting how leading journals typically handle outliers by either removing them from the dataset or adjusting the data statistically. While these methods enhance model fit and support hypothesis verification, they often overlook potentially valuable insights contained within the deviating data, risking lost opportunities for scientific advancement. The author proposes alternative approaches to managing outliers, grounded in a philosophical discussion of science. He concludes with a comprehensive empirical study demonstrating how these atypical observations can yield new insights beneficial for both scientific progress and management decisions. This dissertation provides valuable suggestions for those interested in empirical research and encourages a thoughtful methodological dialogue rooted in the philosophy of science. It is essential reading for emerging academics in the field.