A Meta-analysis of Innovation and Organizational Size

TLDR

Prior research on the relationship between organizational size and innovation has produced contradictory results, prompting Damanpour’s 1992 proposal for a meta‑analysis to clarify the issue. This study replicates and extends Damanpour’s meta‑analysis, updating the evidence from 1970–2001 and examining how different operationalizations of size affect the size‑innovation relationship. The authors conducted a meta‑analysis of 87 correlations drawn from 53 empirical studies published in leading business‑administration journals. The analysis confirmed a significant positive correlation between size and innovation and showed that prior contradictions stem from divergent size‑measurement methods, with a detailed breakdown of size definitions helping to explain the variability.

Abstract

Findings regarding the direction and intensity of the relation between size and innovation in the literature are contradictory. In the journal Organization Studies in 1992, Damanpour proposed a meta-analytical study in an attempt to clarify the diversity of existing conclusions. The present article is a replica and an extension of that study using the same methodology. Our aim is to (1) bring the pool of accumulated knowledge up to date, examining the time span 1970–2001, and (2) review in greater depth the effects of alternative ways of measuring organizational size. The sample used was made up of 87 correlations drawn from 53 empirical studies published in the most important journals on business administration. The analysis enabled us to confirm the existence of a significant and positive correlation between size and innovation. It also provided evidence showing that the contradictory results obtained in previous studies are due to divergences in the methods used to operationalize one, or more, of the variables to be analysed. The main contribution made by our work stems from the fact that the empirical analysis performs a more thorough breakdown of the definitions of the size variable used in the literature. This may well be a first step toward justifying the differences in the results of the primary studies that analyse the relation under examination.