Effectiveness of measures assessing response to price information
| DOI | https://doi.org/10.1108/JPBM-06-2015-0908 |
| Pages | 676-686 |
| Date | 21 November 2016 |
| Published date | 21 November 2016 |
| Author | Pierre Desmet |
| Subject Matter | Marketing,Product management,Brand management/equity |
Effectiveness of measures assessing response
to price information
Pierre Desmet
DRM UMR 7088, PSL Université Paris-Dauphine, Paris, France and Department of Marketing, ESSEC Business School, Cergy, France
Abstract
Purpose – Questionnaire measures of consumers’ willingness to pay (WTP) and price sensitivity are biased, yet these declarative methods
can aid managerial decision-making. Additional choices involve which question formats to use (open-ended or discrete choice) and how many
questions (unique versus multiple). This paper aims to inform such choices for online data collection with an empirical evaluation of the size
of the bias induced by four methods (price acceptability, price judgements, multiple discrete choices and single discrete choices) in a realistic
choice context.
Design/methodology/approach – An experimental framework collects online data about a staple product whose price should be well known. Price
sensitivity, WTP and their confidence intervals are derived from a logistic binary model of acceptability, then ranked to evaluate the size of the bias
of each method, relative to an indirect benchmark.
Findings – Online data collections with self-administrated questionnaires lower respondents’ involvement and create substantial bias; hypothetical
methods overestimate WTP and underestimate price sensitivity, especially with methods using unique questions (both discrete choice and price
acceptability). Multiple questions (price judgements and repeated random discrete choices) increase attention to price information and reduce the
bias. The round price effect also is notable in data collected by open-ended methods.
Practical implications – To measure declarative WTP and price sensitivity with online data collections, researchers should use a random discrete
choices method. Price acceptability questions and split tests are not recommended. Price judgements provide reliable information about consumer
reactions to prices, but the strong round price bias is problematic.
Originality/value – This study adds to marketing and economic literature by comparing actual measurement methods used by firms, rather than
hypothetical versions, and offers strong external validity.
Keywords PSM, Pricing, Willingness to pay, Discrete choice, Price sensitivity, Acceptable price, Gabor & Granger, Open-ended question,
Price elasticity, Psychological price
Paper type Research paper
Assessing consumers’ reactions to prices is critical to the
development of a price policy, so firms collect information to
assess the maximum price that potential buyers would be
prepared to pay and thus to derive a buy-response curve
(Monroe, 1971). Although several methods exist for
determining this price and response curve, including reliable
behavioural approaches (Jedidi and Jagpal, 2009), consumer
goods firms frequently use declared data, collected with
surveys (Lipovetsky, 2006;Steiner and Hendus, 2012;
Voelckner, 2006;Wedel and Leeflang, 1998), because these
data offer important practical advantages (e.g. ease of
implementation and cost and time savings; Hofstetter et al.,
2013). Yet hypothetical survey methods do not require an
obligation to buy, as incentive methods do (Breidert et al.,
2006), so they create biases, generally resulting in
overestimations of people’s willingness to pay (WTP)
(Bateman et al., 2001;Cummings et al., 1995;Lusk and
Hudson, 2004;Voelckner, 2006;Wertenbroch and Skiera,
2002). According to one comparative study though (Miller
et al., 2011), the WTP hypothetical bias for consumer goods
actually is low, and another pricing indicator, price sensitivity,
is less biased in hypothetical compared with incentive settings.
Overall, hypothetical methods efficiently supply a demand
curve that is relatively close to reality, leading to minimally
biased normative recommendations (Miller et al., 2011).
Hypothetical methods demand two additional choices,
pertaining to question formats that are open-ended
(self-stated WTP) or discrete choices (contingent valuation
methods) and the number of questions to use (unique or
multiple questions). Empirical research generally indicates
that the WTP hypothetical bias increases for discrete choices
compared with open-ended questions (Balistreri et al., 2001),
though it also depends on the respondent’s ability to evaluate
the offer; furthermore, this difference is insignificant for
private goods (Halvorsen and Soelenminde, 1998). With
multiple questions, yea-saying effects can induce a higher
WTP and a lower price elasticity, but weariness effects have
opposite results (Bateman et al., 2001). These differences
between methods might arise because the preferences that
The current issue and full text archive of this journal is available on
Emerald Insight at: www.emeraldinsight.com/1061-0421.htm
Journal of Product & Brand Management
25/7 (2016) 676–686
© Emerald Group Publishing Limited [ISSN 1061-0421]
[DOI 10.1108/JPBM-06-2015-0908]
The author thanks Renaud Dédéyan for his suggestions on a previous
version of this article, Toluna Quick Surveys® for online data collection
and IRI for supplying retailer panel data.
676
To continue reading
Request your trial