Field experimental evaluation of how experience with meat substitutes affects meat substitution and substitute adoption rates
How to cite
Fesenfeld, Lukas; Maier, Maiken; Michel, Fabienne; Bernardic, Ursa; Bättig, Fabian. 2024. "Field experimental evaluation of how experience with meat substitutes affects meat substitution and substitute adoption rates." AEA RCT Registry. AEARCTR-0013528. https://doi.org/10.1257/rct.13528-1.1
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
An increasing body of literature identifies consumption habits, particularly of affluent population segments, as an important barrier and key lever of climate change mitigation efforts (Creutzig et al., 2016; IPCC, 2022; Mundaca et al., 2019; Poore & Nemecek, 2018). One area of such currently unsustainable consumption habits are food consumption patterns threatening global food security due to high diet-related greenhouse gas emissions, land and water use, biodiversity loss, as well as negative health consequences (Clark et al., 2020; Crippa et al., 2021; IPCC, 2022; Scarborough et al., 2023; Willett et al., 2019). Thus, there is a growing consensus in the literature that a demand-side shift toward more plant-based diets is a key lever to sustainably change the food system and dietary patterns (Jetzke et al., 2020; Rosenzweig et al., 2020; Willett et al., 2019). This is due to animal foods having a higher inefficiency in water, energy, soil, and land usage compared to plant-based foods, for example, due to animal feed production (Poore & Nemecek, 2018). Further, excessive meat consumption increases the risk for different health problems, such as obesity, cardiovascular diseases, and cancer (Jetzke et al., 2020). Yet, despite the well-understood negative consequences of meat consumption, average global meat consumption levels are still increasing due to an overall increasing demand, especially in developing and emerging countries, while demand is stagnating mostly on rather high levels in industrialized countries (Jetzke et al., 2020; OECD & FAO, 2021). Thus, there is an urgent need for strategies to encourage demand-side shifts toward more plant-based diets.
With the so-called Planetary Health Diet, the EAT-Lancet commission proposes a sustainable and healthy diet that includes a maximum of 15 kilograms of meat per person per year (Willett et al., 2019). However, changing consumer behavior in the realm of food is challenging. Individuals typically do not base their consumption decisions on rational arguments, nor are they perfectly informed. Instead, individuals often rely on simple cues or heuristics (Kahneman et al., 1982; Montibeller & von Winterfeldt, 2015), such as perceived naturalness, feelings of disgust, or trust in the food industry (Siegrist & Hartmann, 2020). Additionally, previously formed habits play a significant role in food decision-making (Rees et al., 2018; Saba & Di Natale, 1998), particularly regarding meat consumption (de Boer & Aiking, 2017). Those deeply rooted food consumption habits result, amongst others, from culture, traditions, learned behavioral patterns, values, and the social environment, and are therefore not easily changed (Constantino et al., 2022; Hartmann & Siegrist, 2020; Kukowski et al., 2023).
One important enabling factor that is increasingly being discussed to accelerate a shift toward more plant-based diets are meat substitute products. In the last years, the market for plant-based meat substitutes, such as plant-based burger patties, sausages, cold cuts, etc., was growing exponentially, although it is still relatively small compared to the global meat market (Polaris Market Research, 2020). Due to economies of scale and technological learning, especially plant-based meat substitute products become increasingly available and more affordable for consumers, as well as more like meat regarding their taste, texture, and appearance (He et al., 2020; Smetana et al., 2021). Although different studies still find that meat substitute products are on average perceived to be inferior to meat regarding taste, texture, and price (Michel et al., 2021; Taylor et al., 2022), others find that meat substitutes are perceived positively, especially regarding animal welfare, health, and the environmental impact (Fesenfeld et al., 2024; Michel et al., 2021; Taylor et al., 2022). However, we lack causal evidence about how experience with such substitute affects future consumption of meat substitute (i.e., substitute adoption rates) and whether more substitute consumptions leads to significant reductions in meat consumption (i.e. meat substitution effects) (Cuffey et al., 2022; Neuhofer & Lusk, 2022; Zhao et al., 2022).
Several, mainly observational and/or survey experimental studies, highlight the essential role of familiarity and repeated (positive) experiences with meat substitutes in influencing individuals’ attitudes, intentions, and subsequently potentially behavioral patterns (Fesenfeld et al., 2023; Graça et al., 2019; Hartmann & Siegrist, 2017; Hoek et al., 2013; Van Bergen et al., 2024). While some of the observational and survey-based studies do not find any positive effects of substitute experience on meat substitution nor substitute adoption (e.g., Cuffey et al., 2022; Van Bergen et al., 2024), other studies suggest that growing exposure to meat substitutes might alter consumers’ product perceptions, substitute adoption and meat substitution – at least for some products (Fesenfeld et al., 2023, 2024; Zhao et al., 2022). Overall, existing studies are thus inconclusive regarding the effects of meat substitute experience on adoption rates and substitution effects. In particular, we currently lack field-experimental studies, in which randomly selected treatment group from a larger population representative sample gains repeated experiences with meat substitute and subsequently their attitudes as well as real-world purchasing behaviors are compared to a control group without the exposure treatment. In this study, we close this research gap by conducting a population-representative survey- and field experiment in Switzerland.
Here, we build on dual process models of human decision making (Chaiken & Trope, 1999; Kahneman, 2011; van der Linden, 2014). According to the theory, it is crucial to differentiate between affective (System I) and rational (System II) decision making. System I, the affective (experiential) decision making system makes rather unconscious, intuitive, automatic decisions with low cognitive effort based on learned behavior and experience. In contrast, System II, the analytical (deliberative) decision making system makes conscious and reflected decisions with higher cognitive effort based on logical thinking, knowledge, and the available information. In decision-making, both systems typically work together to help individuals navigate various cognitive tasks. Food consumption preferences and resulting food consumption choices are typically dominated by System I, which builds on past experiences (Godfray et al., 2018; Smith & DeCoster, 2000). Thus, becoming familiar with and tasting meat substitute products can be a crucial first step towards changing habitual consumption patterns. This is confirmed by several studies showing that familiarity with meat substitutes - i.e. building experience with such products by consuming them repeatedly - is necessary to establish favorable attitudes towards consuming them (Fesenfeld et al., 2023; Hellwig et al., 2022; Onwezen et al., 2022; Siddiqui et al., 2022; Szenderák et al., 2022).
Thus, in the present study we contribute to the literature by causally analyzing the impact of repeated meat substitute experiences with high-quality plant-based meat substitutes on the intention to reduce meat consumption, to increase meat substitute consumption, to support food policy measures, as well as on the revealed purchasing patterns over time. To experimentally evaluate this, we conduct a survey and a field experiment with a representative sample of Swiss residents (obtained from the Federal Statistical Office) in cooperation with the Swiss retailers Coop and Migros and the Swiss plant-based meat substitute producers Planted and Green Mountain. The study was approved by the ETH Ethics Commission. In the first part of the study, we conduct a first survey wave to get to know individuals’ attitudes and intentions on meat and meat substitute consumption, as well as other important control variables influencing food consumption behavior. Then we randomly assign individuals to either the control group getting no substitute products or the treatment group that receives a package with meat substitute products from the Swiss brand Planted and/or vouchers to get meat substitute products from the Swiss brand Green Mountain to cook at home (enough for about 2 dishes) after the first survey wave has been completed. Following the treatment phase, we conduct a second survey wave to ask individuals about their attitudes and intentions on meat and meat substitute consumption post treatment. In the second field experimental part of the study, we then causally evaluate the impact of the random meat substitute experience treatment on individuals’ food purchasing behavior over time by obtaining food purchasing data via Migros and Coop (based on previous informed consent of the survey participants). To our knowledge, this is the first study to collect longitudinal, experimental data from a representative sample on the meat substitution effect and substitute adoption effect of gaining experience with meat substitute products. Furthermore, the study allows causal inference about the effect of gaining experience with meat substitutes on the size of the intention-behavior gap. Several scholars have emphasized the need for such experimental studies (Fesenfeld et al., 2023; Kwasny et al., 2022; Nguyen et al., 2022; Onwezen et al., 2022; Taufik et al., 2019). Furthermore, this study aims to explore the mediating factors of the plant-based meat substitute experience treatment, examining their relationship with stated and revealed consumption adaptations, as well as changes in policy perceptions resulting from experience with meat substitute products.
Based on a review of the existing literature, we formulated the a number of testable hypotheses on the effects of the experience treatment. These hypotheses are outlined in the section "Experimental Details".
Intervention
Independent variable: Experimentally varied treatment
The study consists of two survey waves with an interim meat substitute experience treatment period and a field experimental part in which we will receive retailer purchasing data from Coop and Migros for the period of 24-36 months prior to the experience treatment until about 12-24 months after the experience treatment.
In the following part, the choice of an experience treatment in between two survey waves followed by a field experiment will be justified as an appropriate method to answer the research question. This study's quantitative experience treatment part is a randomized controlled experiment with a control group that does not get treated and a group that receives meat substitute products to try at home. Participants will be randomly assigned to either the control or the treatment group. Due to randomization, the only systematic difference between the control and the treatment group should be the treatment, which allows estimating a causal effect of the treatment on the dependent variables, if the randomization works as intended and the sample is generally large enough (Stock and Watson 2020).
Due to the two survey waves, the first prior to the treatment and the second after the treatment, we can analyze the effect of the experience treatment on changes in attitudes and intentions towards meat and meat substitute consumption, as well as policy support. Moreover, the first survey wave is used to randomly assign participants to the control and the treatment group. In addition, a conjoint experiment is added to the second survey wave to randomly vary different policy attributes to assess the effects of these attributes on the participants’ food policy preferences (Hainmueller et al., 2014) and to determine the policy packages that are most likely also supported in real-world voting scenarios (Hainmueller et al., 2015). Further, the survey respondents are asked to share their Migros and/or Coop customer account numbers in the survey for the field experimental part of the study, such that we can later anonymously use the real-world food consumption purchasing data that is stored within their account to match them with the survey data and monitor potential changes in purchasing due to being assigned to the treatment.
All the survey participants in both waves are asked about their food purchasing and consumption habits, attitudes, perceived social norms, food purchasing intentions, and their policy support for different measures aiming to reduce meat consumption and increase meat substitute consumption.
Based on a review of the existing literature, we formulated the following testable hypotheses on the effects of the experience treatment.
Main effects
H1a: The experimental experience with plant-based meat substitutes positively affects intended meat substitute consumption (i.e., positive intended stated adoption effect).
H1b: The experimental experience with plant-based meat substitutes positively affects intended reduction of current meat consumption (i.e., positive intended stated substitution effect).
H2a: The experimental experience with plant-based meat substitutes will lead to an increase of observed meat substitute consumption (i.e. positive revealed adoption effect).
H2b: The experimental experience with plant-based meat substitutes will lead to a reduction of observed meat consumption (i.e. positive revealed substitution effect).
H3: The experimental experience with plant-based meat substitutes will affect observed (revealed) meat and meat substitute consumption less than the intended (stated) preferences for meat and meat substitute consumption (i.e. intention-behavior gap effect).
H4: The experimental experience with plant-based meat substitutes will increasingly reduce the observed (revealed) meat consumption and increase the observed (revealed) meat substitute consumption over time.
H5: The experimental experience with plant-based meat substitutes will increase respondents' stated support for stricter political regulation of the agri-food sector and policies to alter food consumption and production (i.e., policies fostering meat reduction, more plant-based food, more plant-based meat substitutes).
Mediators
H6: The experimental experience with plant-based meat substitutes will affect the intended (stated) and observed (revealed) meat and meat substitute consumption via
a) changes in respondents’ affective reactions towards the substitute product
b) changes in respondents’ evaluations of the substitute product, especially through,
i) changes in respondents’ product quality perceptions.
ii) changes in respondents’ product sustainability perception.
iii) changes in respondents’ product knowledge.
iv) changes in respondents’ perceived social norms.
Sponsors and Partners:
The study is publically funded by ETH Zurich (Speed2Zero project) and the Swiss Network for International Studies (Project on the political economy of meat system transformation, C2107). The study sample is based on a official population representative sampling frame provided by the Swiss Federal Statistical Office. The study has been approved by the ETH Ethics commission.
This survey and field experiment is developed and conducted in cooperation with the Migros Genossenschaftsbund and the Coop Genossenschaft, the two largest food retailers in Switzerland. Additionally, Planted Foods AG, a Swiss plant-based meat substitute producer as well as the retailer Coop provide the meat substitute products for the experience treatment (Planted products are delivered home to selected participants via the online shop of the Planted Foods AG and Green Mountain products are distributed via product vouchers that selected respondents can use in stores of the Coop Genossenschaft). The sponsoring companies do not have any influence on the study design or outcomes of this pure scientific study. The authors declare no competing interests. The funding of the larger project that this study is a part of is independent and neither sponsored by Migros, Coop, nor Planted. As part of the cooperation, Migros, Coop, and Planted will use the results of our independent scientific analyses as a source to evaluate and better understand potential substitute adoption rates and meat substitution effects induced by meat substitute product experiences.
Experimental design
Survey Panel:
Participants are only eligible for the treatment group(s) if they provide their membership number (Cumulus card number for Migros or Supercard number for Coop) for one of the two retailer loyalty programs. Via the first survey wave, we identify meat-eating respondents with at least one completed membership number and that provided their confirmed consent to analyze their purchasing data. For these respondents we will have access to observed purchasing data from Coop and/or Migros. After the sampling of the first survey wave will be finished, we randomly assign around 1000-2000 respondents to the treatment(s) (i.e., Planted and/or Green Mountain product exposure, see details below) and control (i.e. no product exposure) groups of equal size. Balance checks will be conducted to test that randomization worked.
Conjoint Experiment:
We ask respondents to evaluate profiles that combine multiple randomly assigned attributes. We used a conjoint design of fully randomized paired profiles in which each respondent was shown profiles of two different hypothetical policy packages displayed side by side. Hence, each policy measure constituted an attribute in the package to which it belonged, and the attribute values were randomly assigned such that the two policy packages in each pair differed in one or more attribute values. This paired-profiles design was chosen because research suggests it performs well at reducing social desirability bias and replicating real-world behavior (Hainmueller et al., 2015). To help participants understand how the relevant policy measures function, before showing the pairs of policy packages, we showed a page with brief descriptions of each policy measure (see section Secondary Outcomes above).
Design details
Cited by
No citation data available for this registration yet.