Big data and its epistemology

Research output: Contribution to journalArticle

55 Citations (Scopus)

Abstract

The article considers whether Big Data, in the form of data-driven science, will enable the discovery, or appraisal, of universal scientific theories, instrumentalist tools, or inductive inferences. It points out, initially, that such aspirations are similar to the now-discredited inductivist approach to science. On the positive side, Big Data may permit larger sample sizes, cheaper and more extensive testing of theories, and the continuous assessment of theories. On the negative side, data-driven science encourages passive data collection, as opposed to experimentation and testing, and hornswoggling ("unsound statistical fiddling"). The roles of theory and data in inductive algorithms, statistical modeling, and scientific discoveries are analyzed, and it is argued that theory is needed at every turn. Data-driven science is a chimera.

Original languageEnglish (US)
Pages (from-to)651-661
Number of pages11
JournalJournal of the Association for Information Science and Technology
Volume66
Issue number4
DOIs
StatePublished - Apr 1 2015

Fingerprint

epistemology
Testing
science
Big data
Epistemology

Keywords

  • knowledge

ASJC Scopus subject areas

  • Information Systems and Management
  • Library and Information Sciences
  • Computer Networks and Communications
  • Information Systems

Cite this

Big data and its epistemology. / Fricke, Martin H.

In: Journal of the Association for Information Science and Technology, Vol. 66, No. 4, 01.04.2015, p. 651-661.

Research output: Contribution to journalArticle

@article{406f11d597b14154997b6d50c31dcca9,
title = "Big data and its epistemology",
abstract = "The article considers whether Big Data, in the form of data-driven science, will enable the discovery, or appraisal, of universal scientific theories, instrumentalist tools, or inductive inferences. It points out, initially, that such aspirations are similar to the now-discredited inductivist approach to science. On the positive side, Big Data may permit larger sample sizes, cheaper and more extensive testing of theories, and the continuous assessment of theories. On the negative side, data-driven science encourages passive data collection, as opposed to experimentation and testing, and hornswoggling ({"}unsound statistical fiddling{"}). The roles of theory and data in inductive algorithms, statistical modeling, and scientific discoveries are analyzed, and it is argued that theory is needed at every turn. Data-driven science is a chimera.",
keywords = "knowledge",
author = "Fricke, {Martin H}",
year = "2015",
month = "4",
day = "1",
doi = "10.1002/asi.23212",
language = "English (US)",
volume = "66",
pages = "651--661",
journal = "Journal of the Association for Information Science and Technology",
issn = "2330-1635",
publisher = "John Wiley and Sons Ltd",
number = "4",

}

TY - JOUR

T1 - Big data and its epistemology

AU - Fricke, Martin H

PY - 2015/4/1

Y1 - 2015/4/1

N2 - The article considers whether Big Data, in the form of data-driven science, will enable the discovery, or appraisal, of universal scientific theories, instrumentalist tools, or inductive inferences. It points out, initially, that such aspirations are similar to the now-discredited inductivist approach to science. On the positive side, Big Data may permit larger sample sizes, cheaper and more extensive testing of theories, and the continuous assessment of theories. On the negative side, data-driven science encourages passive data collection, as opposed to experimentation and testing, and hornswoggling ("unsound statistical fiddling"). The roles of theory and data in inductive algorithms, statistical modeling, and scientific discoveries are analyzed, and it is argued that theory is needed at every turn. Data-driven science is a chimera.

AB - The article considers whether Big Data, in the form of data-driven science, will enable the discovery, or appraisal, of universal scientific theories, instrumentalist tools, or inductive inferences. It points out, initially, that such aspirations are similar to the now-discredited inductivist approach to science. On the positive side, Big Data may permit larger sample sizes, cheaper and more extensive testing of theories, and the continuous assessment of theories. On the negative side, data-driven science encourages passive data collection, as opposed to experimentation and testing, and hornswoggling ("unsound statistical fiddling"). The roles of theory and data in inductive algorithms, statistical modeling, and scientific discoveries are analyzed, and it is argued that theory is needed at every turn. Data-driven science is a chimera.

KW - knowledge

UR - http://www.scopus.com/inward/record.url?scp=84944159210&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84944159210&partnerID=8YFLogxK

U2 - 10.1002/asi.23212

DO - 10.1002/asi.23212

M3 - Article

AN - SCOPUS:84944159210

VL - 66

SP - 651

EP - 661

JO - Journal of the Association for Information Science and Technology

JF - Journal of the Association for Information Science and Technology

SN - 2330-1635

IS - 4

ER -