Just one outlier can often mess things up for theory, practice, and hypothesis tests. An outlier is an observation far from the normal and disproportionately influential on the statistical and behavioral properties of a system. The primary goal of this NSF project was to create and disseminate a theoretical framework for explaining and predicting the emergence of outliers in entrepreneurship. The framework proposes that four meta-constructs--Endowments, Expectations, Engagement, and Environments--are the primary drivers of outliers and highly right-skewed power law distributions of all input and outcome variables in the domain, as found by Crawford et al. (2015) in Journal of Business Venturing. The conceptual framework--with power law distributions underlying each construct and larger circles represent the size of companies--is shown in Image 1.

These underlying power laws are important. Within these distributions, there is a critical threshold (the vertical dotted line), where data change from linear to nonlinear. This threshold and other aspects of the distribution are displayed in Image 2. Outliers exist beyond the threshold. Beyond that threshold, one interaction can lead to disproportionately large cascades of unexpected outcomes. As well, beyond that threshold, the probability of capturing resources in the environment is higher compared to those below the threshold--over time, this can lead to infinite variation in outcomes. Beyond that threshold, outliers have an increased likelihood of successfully pushing back on selection forces and changing the normal rules of the system. Finally, beyond that threshold, each observation accounts for a much higher percentage of the total (see elongated tail of the distribution in Images 2 & 3); here, outliers are like stars with gravity, they have the potential to pull in additional resources.

The general purpose of the project is to have the framework accepted into both the scholarly community and mainstream culture. The three primary means of having the framework accepted throughout the domain include (1) thought leadership by creating and organizing Professional Development Workshops (PDW), (2) publishing in journals, conferences, and mainstream outlets (3) presenting to both internal and external stakeholders.

For the domain, I position this framework as a "Power Law Perspective," which accounts for all observations in a distribution, including the myriad low performers, the few elite outcomes that are orders of magnitude larger, and all points in between--this inclusiveness is a unique aspect that makes a significant contribution to research. For the mainstream media, I position the framework as "Rock Star Theory," with the primary focus on breaking through the threshold to become an outlier.

The NSF grant facilitated the dissemination of this framework all over the world. In Summer 2018, I organized and chaired a PDW, "Modeling the Emergence of Outliers in Entrepreneurship: Theory, Mechanisms, and Methods, at the largest conference in the domain, the Academy of Management Meeting. The panelist presentations, hands-on method workshop, newly created website (see Image 4), and publicly available agent-based model provide resources for other scholars to test their own hypotheses about outlier emergence.

After the PDW, thought leadership pivoted to publishing in a special issue about knowledge accumulation in the domain at Entrepreneurship Theory and Practice. Our  team of eight co-authors and eight data analysts attempted to conduct exact replications on 19 seminal studies about new venture creation, with the intent to identify how the handling of skewed distributions and outliers influenced published results. We failed a lot, successfully replicating only a third of the studies. However, our narrative about the extensive drama we experienced--and the ten lessons we convey to the domain--was worthy of acceptance.  

The framework has been disseminated via other means, including a book chapter, two peer-reviewed journal articles, four conference proceedings, 10 conference papers, and one article distributed to the masses in Research Outreach. The latter work was distributed to more than 60,000 people; an interview on the Let's Talk Sales podcast has been listened to 10,000+ times; and an animated sketchbook on YouTube has 8,000 views. I presented the framework at a dozen conferences with both academic and executive audiences. The Rock Star framework has been presented to more than 2,000 students in the US, Italy, and China (see Image 5)--both in-person and online. The total views of this framework from publications, podcasts, Academy emails, and presentations are more than 100,000.

This research demonstrates that outliers are unique in that they think, act, and grow in a different way. In many fields, outliers are typically dismissed as "freak" observations that skew a system's behaviour or performance, and the role they play in explaining the phenomena is oftentimes downplayed or entirely ignored. In research, outliers are vital in providing scholars with a more thorough understanding of the entire scope of entrepreneurial processes. In social systems, outliers are vital to move our society forward. Indeed, we need more freaks. Be the one.

Last Modified: 10/31/2021
Modified by: G Christopher Crawford