Curtailing data biases in business research: Introducing a hybrid approach

Muhammad Aftab Alam, Omar Khalid Bhatti

Research output: Contribution to journalArticleResearchpeer-review

Abstract

This study elucidates the falsity of business research in relying on either respondents or informants alone for data collection, and argues that with the biased data, business research cannot provide unbiased solutions. We compare 400 reports (200 respondents and 200 informants) on the workplace deviance and assess the goodness of both the techniques. Analysis of variance and posthoc (descriptive discriminant analysis) indicate significant disparities between the two approaches across all items. In the informant’s role, people tend to overreport, whereas in the respondent’s part they underreport an undesirable behavior. Further, we find that conventional techniques for assessing the construct’s validity and common-method bias neither assures realistic measurement nor eliminate the response bias. Drawing on the theory of psychological projection, we propose a hybrid approach that curtails some of the main biases in data and measurement. Qualitative confirmation through informal interviews with managers in the investigated firms validates the proposed method.
LanguageEnglish
Pages1-15
Number of pages15
JournalJournal of Management and Organization
Early online date2018
DOIs
Publication statusE-pub ahead of print - 2018

Fingerprint

Hybrid approach
Business research
Psychological
Response bias
Discriminant analysis
Data collection
Managers
Workplace deviance
Analysis of variance

Keywords

  • bias
  • data
  • informant
  • method
  • research
  • respondent

Cite this

@article{1f24cfdfe824486e846d59e4a9307de0,
title = "Curtailing data biases in business research: Introducing a hybrid approach",
abstract = "This study elucidates the falsity of business research in relying on either respondents or informants alone for data collection, and argues that with the biased data, business research cannot provide unbiased solutions. We compare 400 reports (200 respondents and 200 informants) on the workplace deviance and assess the goodness of both the techniques. Analysis of variance and posthoc (descriptive discriminant analysis) indicate significant disparities between the two approaches across all items. In the informant’s role, people tend to overreport, whereas in the respondent’s part they underreport an undesirable behavior. Further, we find that conventional techniques for assessing the construct’s validity and common-method bias neither assures realistic measurement nor eliminate the response bias. Drawing on the theory of psychological projection, we propose a hybrid approach that curtails some of the main biases in data and measurement. Qualitative confirmation through informal interviews with managers in the investigated firms validates the proposed method.",
keywords = "bias, data, informant, method, research, respondent",
author = "Alam, {Muhammad Aftab} and Bhatti, {Omar Khalid}",
year = "2018",
doi = "10.1017/jmo.2018.7",
language = "English",
pages = "1--15",
journal = "Journal of Management and Organization",
issn = "1833-3672",
publisher = "Australian and New Zealand Academy of Management",

}

Curtailing data biases in business research : Introducing a hybrid approach. / Alam, Muhammad Aftab; Bhatti, Omar Khalid.

In: Journal of Management and Organization, 2018, p. 1-15.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Curtailing data biases in business research

T2 - Journal of Management and Organization

AU - Alam, Muhammad Aftab

AU - Bhatti, Omar Khalid

PY - 2018

Y1 - 2018

N2 - This study elucidates the falsity of business research in relying on either respondents or informants alone for data collection, and argues that with the biased data, business research cannot provide unbiased solutions. We compare 400 reports (200 respondents and 200 informants) on the workplace deviance and assess the goodness of both the techniques. Analysis of variance and posthoc (descriptive discriminant analysis) indicate significant disparities between the two approaches across all items. In the informant’s role, people tend to overreport, whereas in the respondent’s part they underreport an undesirable behavior. Further, we find that conventional techniques for assessing the construct’s validity and common-method bias neither assures realistic measurement nor eliminate the response bias. Drawing on the theory of psychological projection, we propose a hybrid approach that curtails some of the main biases in data and measurement. Qualitative confirmation through informal interviews with managers in the investigated firms validates the proposed method.

AB - This study elucidates the falsity of business research in relying on either respondents or informants alone for data collection, and argues that with the biased data, business research cannot provide unbiased solutions. We compare 400 reports (200 respondents and 200 informants) on the workplace deviance and assess the goodness of both the techniques. Analysis of variance and posthoc (descriptive discriminant analysis) indicate significant disparities between the two approaches across all items. In the informant’s role, people tend to overreport, whereas in the respondent’s part they underreport an undesirable behavior. Further, we find that conventional techniques for assessing the construct’s validity and common-method bias neither assures realistic measurement nor eliminate the response bias. Drawing on the theory of psychological projection, we propose a hybrid approach that curtails some of the main biases in data and measurement. Qualitative confirmation through informal interviews with managers in the investigated firms validates the proposed method.

KW - bias

KW - data

KW - informant

KW - method

KW - research

KW - respondent

UR - http://www.scopus.com/inward/record.url?scp=85042524661&partnerID=8YFLogxK

U2 - 10.1017/jmo.2018.7

DO - 10.1017/jmo.2018.7

M3 - Article

SP - 1

EP - 15

JO - Journal of Management and Organization

JF - Journal of Management and Organization

SN - 1833-3672

ER -