A novel method for detecting careless respondents in survey data: floodlight detection of careless respondents


DOĞAN V.

JOURNAL OF MARKETING ANALYTICS, vol.6, no.3, pp.95-104, 2018 (ESCI) identifier identifier

  • Publication Type: Article / Article
  • Volume: 6 Issue: 3
  • Publication Date: 2018
  • Doi Number: 10.1057/s41270-018-0035-9
  • Journal Name: JOURNAL OF MARKETING ANALYTICS
  • Journal Indexes: Emerging Sources Citation Index (ESCI), Scopus
  • Page Numbers: pp.95-104
  • Eskisehir Osmangazi University Affiliated: Yes

Abstract

The current paper proposes a novel method for detecting careless respondents, namely, floodlight detection of careless respondents. This novel method consists of two steps: (1) creating a nonsense regression model and (2) testing a moderator role of response time on the nonsense regression model. An illustration of the floodlight detection of careless respondents method was performed with online survey data collected from 341 Turkish participants. The floodlight detection of careless respondents method is the first systematic approach to calculate a cut-off value for response time, which distinguishes careless respondents from careful respondents. According to the results of floodlight detection of careless respondents, the percentage of careless respondents was 59.8%, which is little higher than the percentage of careless respondents calculated through Instructional Manipulation Check (40.7%) and bogus item (44.2%). The floodlight detection of careless respondents method is described herein, and implications are provided for future research.