When 'significant' is not significant

Abstract

Big data is here for some and coming for many. It promises access to new knowledge along with some challenges, but let's not forget the important lessons of the past to ensure that we are advancing knowledge and making the right decisions from the data we have. In this paper, we submit that marketing's emphasis on statistical significance is misplaced, especially in the new world of big data. We include case examples to demonstrate how statistical significance is easy to find, but not necessarily important. We will also discuss the alternative route for generating robust knowledge. Specifically, we espouse the tradition pioneered by Andrew Ehrenberg of Many Sets of Data (MSoD) and descriptive models as the way to advance marketing science, and as a solid foundation for data interpretation in market research studies. We offer insights for market research practitioners and marketers alike, to ensure they are getting the best from their data for robust marketing decision-making.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 28 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.