International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
Can a Repeated Opt-Out Reminder remove hypothetical bias in discrete choice experiments? An application to consumer valuation of novel food products
Mohammed Hussen Alemu, Søren Bøye Olsen

Last modified: 28 March 2017


Recent studies such as Ladenburg and Olsen (2014) and Varela et al. (2014) have suggested that use of a so-called repeated opt-out reminder (ROOR) might mitigate hypothetical bias in stated Discrete Choice Experiments (DCE), but evidence so far has only been circumstantial. These studies tested the impact of the ROOR in purely hypothetical setups without assessing whether hypothetical bias was actually present in the first place, so drawing any conclusions concerning reductions in hypothetical bias in these studies would be inappropriate. To fill this void, we provide the first comprehensive empirical test of whether a ROOR can actually mitigate hypothetical bias in stated DCE. The data originates from a field experiment concerning consumer preferences for a novel insect-based food product, namely buns where some of the wheat flour is replaced by cricket powder (CP). Three split sample treatments are used; 1) a control in terms of a fairly standard hypothetical DCE (labelled HYPO), 2) a hypothetical DCE framed with a ROOR (labelled ROOR), and a financially incentivized DCE (labelled REAL). Based on focus group interviews as well as a desire to ensure that all product combinations could be actually delivered to participants, three relevant attributes were identified for the experimental design. These are the amount of CP, whether some portion of the wheat flour is fortified with added micronutrients or not, and a price for a 160g bag of buns. Using priors informed by two rounds of pilot studies, a D-efficient fractional factorial design was used to generate the 12 choice sets that participants were subjected to. And in the ROOR treatment subjects were provided with the following opt-out reminder before each choice set:


"If the prices of "Bag 1" and "Bag 2" are higher than what you think your household will pay, you should choose "Bag 3".

'Insert figure 1 about here, see supplementary file'

As the aim of this paper is to investigate preferences and WTP for buns with and without CP, which is more in line with a marketing study, a forced choice setup is used where the opt-out alternative is defined as a standard bag of scones without CP or fortified flour. This alternative is constant across all choice tasks. Participants received an amount of Kenyan Shillings (KShs) 30 as a show-up fee and the participants in the REAL treatment received an additional KShs 90 which they could later spend fully or partly on paying for one of the bun products according to their choices in the experiment. We followed the procedure used in Alfnes et al. (2006) and Gracia et al. (2011). At the end of the experiment, each individual participant was asked to draw a random number from 1 to 12 (the total number of choice sets) to determine which choice set would be realized, and participants then had to pay the listed price to receive the product they had chosen in this specific choice set. Implementing a stratified random sampling procedure, a total of 334 consumers participated in the field experiments which were conducted in Kenya. While the HYPO and ROOR treatments are based on 109 participants each, 116 participants took part in the REAL treatment.

We analyse the data in willingness-to-pay (WTP) space as WTP is the measure of interest for this study. The results show that a higher share of participants chose the Standard buns more frequently in ROOR than in the HYPO, which elucidates that the ROOR may indeed have enabled participants to make the required trade-offs between attributes that may reflect their true choices. In line with our a priori expectations, participants chose the Standard buns more frequently in the REAL treatment than in the HYPO or in the ROOR. We find that the marginal WTP estimates in the HYPO treatment are up to 85% higher than the ones in REAL treatment, confirming the usual presence of hypothetical bias. More interestingly, comparing this to a hypothetical setting where the ROOR is introduced, we find that the ROOR effectively eliminates hypothetical bias for one attribute and significantly reduces it for the rest of the attributes. Our results further suggest that these reductions of hypothetical bias are brought about by a decrease in the tendency to ignore the price attribute. Overall, we thus find strong support for the effectiveness of the ROOR.


Alfnes, F., A. G. Guttormsen, G. Steine, and K. Kolstad. 2006. Consumers' Willingness to Pay for the Color of Salmon: A Choice Experiment with Real Economic Incentives. American Journal of Agricultural Economics 88 (4):1050-1061.

Gracia, A., M. L. Loureiro, and R. M. Nayga. 2011. Are Valuations from Nonhypothetical Choice Experiments Different from Those of Experimental Auctions? American Journal of Agricultural Economics 93 (5):1358-1373.

Ladenburg, J., and S. B. Olsen. 2014. Augmenting short Cheap Talk scripts with a repeated Opt-Out Reminder in Choice Experiment surveys. Resource and Energy Economics 37:39-63.

Varela, E., P.-A. Mahieu, M. Giergiczny, P. Riera, and M. Soliño. 2014. Testing the single opt-out reminder in choice experiments: An application to fuel break management in Spain. Journal of Forest Economics 20 (3):212-222.


Conference registration is required in order to view papers.