Do more heads imply better performance? An empirical study of team thought leaders' impact on scientific team performance

Technical Summary: Do More Heads Imply Better Performance? An Empirical Study of Team Thought Leaders' Impact on Scientific Team Performance

Problem Statement
While scientific innovation is increasingly a team endeavor, the specific impact of the number of "thought leaders" within a team on scientific performance remains understudied. Existing literature often conflates managerial leadership with the conceptual generation of ideas or relies on author order as a proxy for leadership, potentially overlooking middle authors who contribute to idea conception. Furthermore, it is unclear whether increasing the number of individuals responsible for conceptualizing research ("more heads") linearly improves team performance or if diminishing returns and conflicts arise. This study addresses two primary research questions: (1) What is the relationship between the number of thought leaders and team performance? and (2) How can a causal relationship between these variables be investigated?

Methodology
The study utilizes a dataset of 141,550 research articles published in Public Library of Science (PLOS) journals between 2007 and 2015.

• Definition of Thought Leaders: Unlike studies using author order, this research defines thought leaders specifically as authors who contributed to the task of "conceived and designed the experiments," based on PLOS contribution statements.
• Data Processing: Author names were parsed and matched to contribution statements with 99.19% accuracy. The dataset was linked with the PubMed Knowledge Graph (PKG) for citation data and author disambiguation, and with OpenAlex for affiliation data. Gender was inferred using NamSor.
• Variables:
  • Dependent Variables: Team impact (measured by 5-year cumulative citation counts) and team disruptiveness (measured by the Disruption Index, CD).
  • Independent Variable: The ratio of thought leaders (number of thought leaders divided by total team size).
  • Control Variables: Team size, institutional diversity, international collaboration, number of references, gender diversity, publication year, field, and thought leader characteristics (average academic age and prior productivity).
• Analytical Approach:
  1. Multivariable Regression: Used to quantify the correlational relationship, including squared terms to test for non-linear (curvilinear) effects.
  2. Coarsened Exact Matching (CEM): Employed to infer causality by creating balanced treatment and control groups based on covariates (team size, international status, references, etc.), thereby reducing confounding effects and simulating a quasi-experimental design.

Key Results

• Inverted U-Shaped Relationship with Impact: The study found an inverted U-shaped relationship between the ratio of thought leaders and team impact. Team impact increases with the number of thought leaders up to a point, peaking when the ratio of thought leaders is approximately 0.70 (70%). Beyond this threshold, adding more thought leaders negatively affects team impact.
• Negative Relationship with Disruptiveness: Teams with a higher number of thought leaders tend to produce less disruptive ideas. The relationship between the ratio of thought leaders and the disruption index is significantly negative; as the proportion of thought leaders increases, the disruptiveness of the output decreases.
• Causal Findings (CEM): The causal analysis confirmed that increasing the number of thought leaders generally increases team impact (up to a peak) but decreases team disruptiveness. Specifically, teams with a high ratio of thought leaders (e.g., >0.8) showed a significant reduction in disruptiveness compared to teams with low ratios.
• Control Variable Effects:
  • Team Size: Larger teams correlate with higher impact but lower disruptiveness.
  • International Collaboration: Improves team impact but lowers disruptiveness.
  • Gender Diversity: Positively associated with team impact but negatively (though non-significantly) associated with disruptiveness.

Significance and Contributions
The paper claims three primary contributions:

1. Conceptual Extension: It extends the concept of "thought leadership" from organizational management to the domain of scientific innovation, distinguishing it from managerial or supervisory roles by focusing on the specific task of idea conception.
2. Empirical Evidence: It provides new evidence regarding the non-linear impact of thought leader quantity on two distinct performance metrics: impact (citations) and disruptiveness (novelty). It challenges the assumption that "more heads" always yield better performance, demonstrating that an optimal ratio exists for impact and that high numbers of thought leaders may hinder breakthrough innovation.
3. Methodological Innovation: It pioneers the application of Coarsened Exact Matching (CEM) in the science of science to investigate the causal effects of team composition, moving beyond simple correlation to address potential selection biases.

Implications
The authors suggest that for policymakers and team managers, funding for high-impact, incremental innovation should be directed toward teams with a larger number of thought leaders (up to the optimal ratio), whereas funding for disruptive, breakthrough innovation should be allocated to teams with fewer thought leaders. The study concludes that "more heads do not always imply better performance," highlighting the need for a balanced composition of conceptual contributors within scientific teams.