Predicting new venture survival: A Twitter-based machine learning approach to measuring online legitimacy
Journal
Journal of Business Venturing Insights
ISSN
2352-6734
Type
journal article
Date Issued
2019-01-07
Research Team
IWI6, Crowdsourcing, CCC
Abstract
Research indicates that interactions on social media can reveal remarkably valid predictions about future events. In this study, we show that online legitimacy as a measure of social appreciation based on Twitter content can be used to accurately predict new venture survival. Specifically, we analyze more than 187,000 tweets from 253 new ventures' Twitter accounts using context-specific machine learning approaches. Our findings suggest that we can correctly discriminate failed ventures from surviving ventures in up to 76% of cases. With this study, we contribute to the ongoing discussion on the importance of building legitimacy online and provide an account of how to use machine learning methodologies in entrepreneurship research.
Language
English
HSG Classification
contribution to scientific community
HSG Profile Area
SoM - Business Innovation
Refereed
Yes
Publisher
Elsevier
Volume
11
Number
June
Start page
1
End page
8
Pages
8
Subject(s)
Eprints ID
256299
File(s)![Thumbnail Image]()
Loading...
open.access
Name
AntretterEtAl_JBVI_2019.pdf
Size
353.53 KB
Format
Adobe PDF
Checksum (MD5)
f9e5246a5b5eebd4b14a31114c755d84