Using mutual information and information gain ratios on entrepreneurial research: an empirical case for Greek I.T. start-ups

Abstract

The majority of Entrepreneurial quantitative research focuses on Correlation Coefficients. However, new statistical analysis based on Entropy, such as Mutual Information and Information Gain Ratios cast a new light on understanding the relationships among variables and offer a view of non-linear relationships.The study examines key entrepreneurial variables using Mutual Information and Information Gain Ratios and compares findings using the same dataset which examined I.T. Greek Start-Ups. Use of Mutual Information and Information Gain ratios reveals much more relationships between the variables examined, in comparison to Pearson Correlation. Furthermore, the study compares results from Pearson Correlation and Mutual Information and Information Gain ratios to drawn new conclusions on the perceptions of Greek I.T. start-up founders. The findings indicate that use of Mutual Information reveals a set of factors that contribute to entrepreneurial perception of success which differs significantly from the conclusions based on Correlation Coefficient Analysis. More specifically factors such as Operation Years and Previous Start-Ups play a far more crucial role than B2B and Sales. The study offers an original contribution to entrepreneurial science, introducing the use of entropy-based mathematical ratios, such as Mutual Information and Information Gain in Entrepreneurial Research. The study highlights that use and results derived of such ratios enable researchers to identify more information regarding (non-linear) relationships between variables, compared to Correlation Coefficient methods.