On the use of balanced item parceling to counter acquiescence bias in structural equation models
(A2020-62936)

AUTHORS

Bert Weijters, Ghent University; Hans Baumgartner, Penn State Smeal College of Business

KEYWORDS

SEM; parceling; Acquiescence bias

ABSTRACT

Net Acquiescence Response Style or NARS refers to respondents’ relative preference for positive vs. negative response options in agreement rating scales and has been found to bias responses. The use of reversed items offers an important remedy against this bias. While researchers often randomly combine individual items from self-report scales into parcels that are subsequently used as indicators in a Confirmatory Factor Analysis (CFA) or Structural Equation Model (SEM), we propose that creating balanced parcels, that is, parcels that each combine a regularly worded item with a reversed item (i.e., an item that is formulated in a direction opposite to that of the construct it measures) has the important advantage that it corrects for Net Acquiescence Response Style at the parcel level and consequently makes it unnecessary to include a method factor in the measurement model. We demonstrate in two empirical studies that the balanced parceling approach leads to models that fit well and yield unbiased parameter estimates.