Breaking the Boundaries of Knowledge Space: Analyzing the Knowledge Spanning on the Q&A Website through Word Embeddings

TL;DR

This study investigates how the distance and hierarchy of knowledge influence the appeal of questions on a large online platform, revealing a nonlinear relationship with an optimal span.

Contribution

It introduces a novel analysis of knowledge spanning in question asking, integrating word embeddings and hierarchy theories to understand question appeal.

Findings

Knowledge spanning has a parabolic effect on question appeal.

The positive impact of knowledge spanning is contingent on knowledge hierarchy.

Optimal knowledge span exists beyond which question appeal declines.

Abstract

The challenge of raising a creative question exists in recombining different categories of knowledge. However, the impact of recombination remains controversial. Drawing on the theories of knowledge recombination and category spanning, we propose that both the distance of knowledge spanning and the hierarchy of knowledge shape the appeal of questions. Using word embedding models and the data collected from a large online knowledge market (N = 463,545), we find that the impact of knowledge spanning on the appeal of questions is parabolic: the appeal of questions increases up to a threshold, after which point the positive effect reverses. However, the nonlinear influence of knowledge spanning is contingent upon the hierarchy of knowledge. The theoretical and practical implications of these findings for future research on knowledge recombination are discussed. We fill the research gap by conceptualizing question asking as knowledge spanning and highlighting the theoretical underpinnings of the knowledge hierarchy.