Bio: 

Irene Scopelliti is Professor of Marketing and Behavioural Science at Bayes Business School (formerly Cass), head of the Marketing group, and Co-Director of the Behavioural Research Lab. Her research interests are in the domain of consumer psychology, judgment, and decision making.

She holds a PhD in Management (major in Marketing) from Bocconi University, a MSc in Economics and Management from the University of Calabria (Italy), and a BA in Music. Before joining Bayes, Irene was a Post Doctoral Research Fellow at Carnegie Mellon University.

Her research has been published in leading journals including Management Science, Psychological Science, Journal of Consumer Psychology, Journal of Management, International Journal of Research in Marketing, Organizational Behavior and Human Decision Processes, Journal of Product Innovation Management, and has been featured by major news organizations including Forbes, Time Magazine, BBC News, the New York Times, the Independent, and the Atlantic. She is currently on the editorial boards of the Journal of Consumer Research and International Journal of Research in Marketing. 

At Bayes Irene teaches analytical methods for marketing in the MSc program, and experimental design and analysis in the PhD program.

Abstract:

The advent of big data has revolutionized how scientists, policy makers, and managers use empirical evidence to make decisions. We identify a critical mistake in how decision-makers interpret such data: the big data fallacy—the tendency to erroneously infer causation from correlation when sample sizes are large. This constitutes a fundamental error because sample size cannot overcome the endogeneity problems inherent in observational data, regardless of data volume. Twelve preregistered experiments (total N = 6,408) provide evidence that decision-makers are more likely to interpret correlational relationships as causal when observational data have large versus small sample sizes, even when presented with conflicting experimental evidence showing no (or the opposite) causal effect. The fallacy persists across decision-domains, affects also those with reasonable scientific reasoning skills and experienced executives, and resists conventional debiasing interventions. We show that decision-makers engage in attribute substitution, using readily available information about data quantity to judge the more complex question of whether data permit causal inference. Because large observational datasets can provide precisely wrong answers with high confidence, this bias has profound implications for evidence-based management and organizational decision-making in the era of big data.