Distortion in retrospective measures of word of mouth

Robert East

Kingston Business School and Ehrenberg-Bass Institute

Mark D. Uncles

University of New South Wales

Jenni Romaniuk

Ehrenberg-Bass Institute

Chris Hand

Kingston Business School

Introduction

Survey responses may be subject to bias, and this is particularly so when respondents are asked to report on what they have done some time ago. Infrequent behaviour such as word of mouth (WOM) cannot easily be observed as it happens and, for this reason, WOM is often measured via surveys, by asking about the volume (number of instances in a time interval) of positive and negative word of mouth (PWOM, NWOM) that respondents have given or received. Such evidence is required for comparing the effects of PWOM and NWOM, and for relating WOM to brand and company performance.