Reversed polarity items in tourism scales: best practice or dimensional pitfall?

Abstract

ABSTRACT Disciplines such as marketing and education have begun to question the benefits of incorporating reverse polarity items (e.g. a mixture of negatively and positively worded items) into multi-item scales due to such items’ degradation of scale dimensionality. The tourism literature, however, has yet to critique this practice due to the commonly held belief that reverse polarity items reduce acquiescence bias. With a limited critique of its practice within the tourism literature, the purpose of this ‘Methods and Practice’ paper is to provide a literature review of the topic and to conduct psychometric analyses on four tourism scales including reversed polarity items. EFA and CFA results from 703 responses to the Psychological, Social, and Political Empowerment Scales of the Resident Empowerment through Tourism Scale and 300 responses to the Perceived Stress Scale reveal that the inclusion of reversed polarity items had significant negative impacts on unidimensionality, model fit, factor loadings, and AVE in each instance. Differences were also found in the strength of regression coefficients and variance explained between reverse and non-reverse polarity scales when regressed on theoretically relevant dependent variables. Implications for future scale development are discussed highlighting the need to simultaneously reduce acquiescence bias and ensure scales demonstrate construct validity.