What ’ s cooking, if not meat? Effects of repeated home-use, recipe inspiration and meal context on perception of plant-based meat analogues

Plant-based meat analogues (PBMA) may help consumers in shifting towards more plant-based diets, but PBMA are not widely used yet, and little is known about their longer-term acceptance. This study investigated whether consumer acceptance of PBMA changed with repeated home-use, and whether providing recipe suggestions in the form of meal boxes could influence PBMA acceptance. To this end, Dutch regular meat eaters (n = 61) prepared, consumed and evaluated two meals (one from a meal box, one self-created) with PBMA (PB mince and PB chicken, counterbalanced across meal types) per week at home for four weeks. As a secondary objective, potential longer-term effects of repeated home-use of PBMA on meat (analogue) consumption habits and attitudes (e.g. motives for choosing PBMA, attitudes toward eating less meat) were assessed in a pre-vs post-intervention survey. Responses were compared with a control group of consumers not participating in the home-use study (n = 179). Results provided no evidence that PBMA liking changed with repeated home-use, nor that the provision of meal boxes increased liking of PBMA. Instead, PBMA liking was strongly influenced by the meal context, which may have overruled potential effects of repeated exposure. Findings from the pre-vs. post-intervention survey suggest that repeated exposure may stimulate longer-term consumption of PBMA, although more seems needed to bring about a structural shift toward a less animal-based consumption pattern. Future research should investigate whether more sophisticated recipes that provide a suitable meal context for PBMA and elevate consumers ’ meal experiences may improve PBMA acceptance and facilitate the transition toward more sustainable diets.


Introduction
Current Western diets with high levels of animal-derived foods, especially meat, are unsustainable, having negative impacts on climate and the environment, human health, animal welfare and global food security (Parlasca & Qaim, 2022;Springmann et al., 2018;Willett et al., 2019).To address these issues and achieve the UN Sustainable Development Goals on a better and more sustainable future for all by 2030 (UN, 2023), a shift towards diets higher in plant-based foods and lower in animal-derived foods is necessary (FAO/ WHO, 2019;Springmann et al., 2018;Willett et al., 2019).One way to facilitate the reduction of meat consumption is the use of plant-based meat analogues (PBMA), which are designed to mimic the appearance, taste, and texture of animal meat (Boukid, 2021).While the market of PBMA is expanding globally, PBMA are still not widely accepted by consumers (Boukid, 2021;Jahn et al., 2021).Recent research investigating motivational barriers that withhold consumers from switching to PBMA reported negative attitudes and beliefs about the sensory appeal of PBMA (e.g.meat attachment), unfamiliarity with PBMA and a tendency to avoid new foods, as well as social-cultural aspects, including the lack of knowledge on how to prepare PBMA (Eckl et al., 2021;Jahn et al., 2021;Onwezen et al., 2021b;Sijtsema et al., 2021;Vural et al., 2023).New and unfamiliar foods (such as PBMA) are often initially rejected (Michel et al., 2021;Pliner & Hobden, 1992), but perceptions of foods can change with repeated consumption.Here, we refer to liking as an affective response to foods, and desire-to-eat as (immediately) wanting or intending to consume a food (Berridge, 1996;Mela, 2000).
Several studies have shown that liking of foods increases with repeated exposure (Birch et al., 1982(Birch et al., , 1987;;Crandall, 1984;Hausner et al., 2012;Pliner, 1982;Porcherot & Issanchou, 1998;Zandstra et al., 2000).This so-called "mere exposure" effect has particularly been observed for unfamiliar products and flavours.Through repeated exposure, unfamiliarity or uncertainty with a novel product can be diminished, which may lead to improved acceptance over time.By contrast, other studies have shown that repeated exposure can result in decreased liking of foods caused by satiation with specific attributes of the consumed foods, especially for products that are initially highly liked (Porcherot & Issanchou, 1998;Stubenitsky et al., 1999;Weijzen et al., 2008;Zandstra et al., 2000).With repeated consumption, the product (or the idea of eating the product repeatedly) can become boring, which is reflected in a decrease in liking or desire to eat the food (Zandstra et al., 2004).For some foods (e.g.staple foods), liking does not change with repeated consumption (Schutz & Pilgrim, 1958;Siegel & Pilgrim, 1958).So far, little is known about how liking of PBMA develops over time.A study among non-users, light users (< once per week) and heavy users (≥ once per week) of PBMA in the Netherlands and the UK (Hoek et al., 2011) found that most heavy users started to consume PBMA more than 10 years ago, while the majority of light-/medium-users only started 5 years ago.This may suggest an increase in acceptance of PBMA over time.One study (Hoek et al., 2013) investigated changes in liking of meat substitutes (mycoprotein pieces and tofu) and meat (chicken), when consumed at home twice a week for 10 weeks.Overall, a decrease in liking was found for all products, but there were big individual differences: whereas some respondents experienced product boredom, others showed an increase in liking over repeated consumption.Having previous experience with meat substitutes was associated with overall higher liking scores.The same study also found that changes in liking over time were influenced by the meal context and variety.It was observed that consumers who gradually liked the meat substitutes less, used the products in more varied meal contexts (e.g., rice, potato, pasta or noodle dish, pizza, soup, salad), possibly to mitigate the experienced boredom (Hoek et al., 2013).While meal variety may be a driver for consumer acceptance of PBMA, many consumers lack the inspiration, knowledge, and skills to prepare a variety of tasty vegetarian meals with PBMA (Jahn et al., 2021;Sijtsema et al., 2021).In a qualitative study with focus group discussions among flexitarians in the Netherlands (Sijtsema et al., 2021), several opportunities were defined to support the consumption of plant-based meat substitutes, such as inspiration and recipes for easy-to-prepare tasty meals, information on product preparation and a convenient preparation similar to animal meat.Also, meal box solutions that provide all necessary ingredients for well-explained, convenient, and tasty new dishes were suggested by the study participants.These qualitative insights were taken along in the current study, which aimed to investigate the impact of repeated exposure in combination with recipe inspiration on consumer acceptance of PBMA in home-cooked meals.Specifically, we examined whether and how consumer acceptance of PBMA changes with repeated in-home cooking and consumption, and to what extent PBMA acceptance is influenced by the provision of recipe suggestions in the form of a meal box scheme.
To address our research questions, a four-week intervention study was executed during which consumers weekly prepared, consumed and evaluated two evening meals (one from a meal box scheme and one selfcreated) with a PBMA (one meal with PB chicken and one meal with PB minced meat) at home.We hypothesized that repeated exposure may improve acceptance of PBMA by overcoming initial unfamiliarity and taste concerns.Moreover, the provision of weekly varying recipe suggestions was hypothesized to have a positive effect on PBMA acceptance by ensuring meal variety and preventing boredom over time.PBMA liking was selected as the primary outcome measure to assess consumer acceptance of PBMA over the course of the four weeks.Perceptions of sensory PBMA quality (juiciness, tenderness) and evaluations of the full meal (liking, desire to eat, preparation difficulty) were additionally measured to further specify the factors underlying (changes in) PBMA acceptance.As a secondary objective, potential longer-term effects of repeated home-use of PBMA on meat (analogue) consumption habits and attitudes were investigated in a pre-vs.post-intervention survey, where participants' responses were compared with a control group of consumers not participating in the home-use study.
With this study, we contribute to the limited number of studies investigating consumer acceptance of PBMA in real-life settings.In particular, it may provide new insights into whether and how repeated exposure and recipe inspiration interact in influencing PBMA acceptance.

Participants
Ethical approval for the involvement of human subjects in this study was granted by the Institutional Review Board of Wageningen University and Research (date of approval: January 22, 2022).Study participants were recruited from the Food Research participant panel of Wageningen University & Research, a database of over 2000 consumers living in or close to Wageningen and willing to participate in scientific research on food-related topics.
Sixty-three panelists were selected for participation in the home-use study.Inclusion criteria were: Dutch speaking; not following a vegetarian, pescatarian or vegan lifestyle; and being in the possession of a smartphone (which was needed to install the research app used for realtime data collection; for details, see Section 2.5 below).Panelists reporting an allergy for any of the ingredients used in the study were excluded.From the eligible panelists, 63 were selected based on their habitual meat and PBMA consumption behaviour (most frequent meat eaters; least frequent PBMA eaters), and randomly assigned to one of two subgroups, which were used for counterbalancing the experimental design.During the home-use period, all participants weekly prepared two meals (once a meal box meal and once a self-created meal) with two PBMA products (once with PB minced meat and once with PB chicken), but the combination of meal types and PBMA products was counterbalanced, such that subgroup A always prepared meal box meals with PB chicken and self-created meals with PB mince, whereas subgroup B prepared meal box meals with PB mince and self-created meals with PB chicken (see Section 2.2 below for details).Two participants were excluded (one drop-out, one due to substantial difficulty using the research app, which led to a high number of missing responses), leaving 61 participants (Subgroup A: N = 32, 20 female, mean age 50.8 (SD 16.5) years; Subgroup B: N = 29, 20 female, mean age 48.9 (SD 15.5) years) for analysis.Socio-demographics and other baseline characteristics (assessed in the pre-intervention survey) did not significantly differ between subgroups, except for meat attachment, to the extent that subgroup B was slightly more attached to meat than subgroup A (p = .05)(see Table 1).
One hundred and seventy-nine panelists served as a control group, which was selected based on online questionnaires.All panelists not participating in the home-use study received a link to a first questionnaire.Those who met the inclusion criteria (Dutch speaking, not following a vegetarian, pescatarian or vegan lifestyle) and completed this first questionnaire N = 317) received a link to a second questionnaire eight weeks later.All who completed both questionnaires were included in the analysis (N = 179, 143 female, mean age 49.8 (SD 15.5) years).Baseline characteristics of the control group did not differ from the intervention group, except for meat consumption frequency, which was significantly higher in the intervention group relative to the control group (p = .02)(see Table 1).
Home-use study participants provided written consent prior to the start of the home-use period; survey respondents provided consent at the first page of the online questionnaire before continuing to the questions.After the study, all were debriefed about the study aims via email.Participants in the home-use study received a monetary compensation of G. van Bergen et al.50 euros; ten gift cards of 20 euros were raffled among the survey respondents who completed both questionnaires.

Design
The home-use study followed a mixed within-between subjects design.For a period of four weeks, participants weekly prepared and consumed two evening meals (one from a meal box scheme and one of their own choice) with a PBMA (once PB minced meat, once PB chicken pieces).The duration of the home-use period (four weeks) and the number of PBMA exposures (eight in total) were based on previous research and recommendations by Zandstra et al. (2004).All participants were exposed to the same experimental conditions, but the combination of meal types and PBMA products was counterbalanced, such that half of participants prepared the meal box meals with PB chicken and the self-created meals with PB mince (subgroup A), and the other half prepared the meal box meals with PB mince and the self-created meals with PB chicken (subgroup B).The meal box scheme consisted of four different meal box meals, which were the same for all participants (except for the PBMA in the meals).The order of the four meal box meals was counterbalanced.A description of the meal box meal ingredients is provided in Table 2; Fig. 2 shows example plates of the meal box meals as prepared by participants.

Products
Every week, the meal box contained a different recipe for a basic, easy-to-prepare, warm meal for two persons with all belonging ingredients (see Table 2), with instructions for storage and preparation.All   meal box ingredients were provided at no cost.For self-created meals, only the PBMA product was provided for free (also with storage and preparation instructions); other meal ingredients had to be purchased by the participants themselves.
PBMAs were delivered frozen to the research location in the week prior to the home-use period, where they were stored in the freezer at − 20 • C. On Fridays, the batch for the coming week was taken from the freezer and stored in the fridge at 4-7 • C. Other meal box ingredients were purchased from an online grocer.Non-perishable ingredients were delivered to the research location in the week prior to the home-use period, where they were stored at room temperature.Perishable ingredients were delivered to the research location on pick-up days and stored in the fridge at 4-7 • C until the pick-up moment.
Every week, participants received a cooler bag with two PBMA products (PB mince and PB chicken) plus the ingredients for the meal box meal of that week.The PBMA products were provided blind (carton product sleeves including brand information were removed); other ingredients were provided in their original packaging.To ensure participants would use the PBMAs in the intended meal type (which differed across subjects), a yellow sticker was placed on the PBMA product that had to be used separately from the other products (i.e., in the selfcreated meal).

Procedure
The study was executed in January-April 2022.One week prior to the start of the home-use period (week 0), participants installed the research app on their smartphone, and completed an online questionnaire assessing their baseline meat/PBMA consumption habits and attitudes, as well as general personal traits.A link to this questionnaire was simultaneously sent to the control group.During the home-use period (weeks 1-4), participants weekly came to the campus of Wageningen University & Research on Mondays to pick up the ingredients for the study meals for that week, plus instructions for storage and preparation.They prepared the two meals on two weekdays (Mon-Fri) of their choice, taking the expiry dates of perishable ingredients (e.g.pre-cut vegetables) into consideration.Over the course of the four weeks, ecological momentary assessments were obtained via the research app (for details, see section 2.5).At the end of the home-use period (week 4), participants completed an online questionnaire assessing how they experienced the home-use period.Four weeks later (week 8), they filled out a final questionnaire assessing their consumption habits and attitudes in the past four weeks (i.e. the weeks following the home-use period); a link to this questionnaire was simultaneously sent to the control sample.A schematic overview of the procedure is presented in Fig. 1.

Ecological momentary assessments
During the home-use period, a smartphone application called Real-Life Exp on the participants' end and LifeData on the researchers' end (LifeDatacorp, Marion, IN, USA) was used to obtain ecological momentary assessments, which enabled studying participants' real-time behaviour in their natural environments to maximize ecological validity (Shiffman et al., 2008).Each weekday at 19h45, participants received a notification on their smartphone asking whether they had prepared and consumed a study meal that evening.If they answered 'no', the session ended immediately; if they answered 'not yet', the same notification was sent 90 min later.Upon a 'yes'-response, they indicated which meal type they had prepared (meal box or self-created meal), which triggered meal-specific follow-up questions.For meal box meals, participants were asked whether they had deviated from the provided recipe (yes/no) and if so, which ingredients they had added and/or left out (open question).For self-created meals, they provided a short written description of the meal and indicated where they got their inspiration for the meal from (multiple choice question: from a meal I usually prepare with meat; from the ingredients I had at home; from a recipe; from the offer in the supermarket; from friends/family; other).In addition, participants were asked to take a photo of their plate and upload it in the app.Responses and photos were used to monitor compliance with the study procedure and to gain qualitative insight in how the PBMAs and the study meals were prepared (see Section 3.1.1below).
Subsequently, participants evaluated the PBMA and the meal they had prepared that day on visual analogue scales (VAS) presented horizontally on the smartphone screen (Table 3).Over the course of the four weeks, a total of eight scores per VAS measure were collected per participant (two responses per week), which were used for analysis.
PBMA liking was selected as the primary outcome measure to reflect consumer acceptance of PBMA.PBMA juiciness and tenderness were measured because PBMA acceptance has been shown to crucially depend on sensory quality (Hoek et al., 2011;Weinrich, 2019), and these attributes have been defined as highly valued attributes of meat (Grunert et al., 2004) as well as the most important textural attributes for PBMA liking (Godschalk-Broers et al., 2022).Liking of the full meal was included because of the previously shown context-sensitivity of PBMA liking (Elzerman et al., 2011).Desire-to-eat was included as a potential indicator of "concept boredom" (Zandstra et al., 2004), that is, boredom with the idea of having to eat the same foods repeatedly.Lastly, meal preparation difficulty was measured to assess whether preparing meals with PBMA would become easier over time (especially with the provision of recipes and preparation instructions).This measure was included as preparation difficulty has been identified as a barrier for using PBMA (Jahn, et al., 2021;Sijtsema et al., 2021).

User experience questionnaire
At the end of the home-use period, an online questionnaire was sent out using EyeQuestion® (version 5.11.2,EyeQuestion Software, the Netherlands) to gain insight in how participants experienced the homeuse period, and to assess potential differences between subgroups.Participants indicated to what extent they found participating in the study enjoyable (7-pt Likert scale: 1 = totally disagree; 7 = totally agree), and how they evaluated the study meals relative to their usual evening meals

Pre-vs. post-intervention survey
Measures from the pre-vs.post-intervention survey were obtained via EyeQuestion® (version 5.11.2,EyeQuestion Software, the Netherlands).Data were collected to address our secondary research question, and for further characterization of our sample.Repeated questions assessed consumption habits (meat consumption frequency: "On average, how many days per week do you eat a warm meal with meat?" 0-7 days; PBMA consumption frequency: "How often do you usually eat a warm meal with a PBMA?" never; less than once per month; less than once per week; 1-2 times per week; > 2 times per week) motives for choosing PBMA (adapted from the single-item food choice questionnaire by Onwezen et al., 2019); meat attachment (Meat Attachment Questionnaire, Graça et al., 2015) and attitudes toward reduced meat consumption ("To me, eating less meat is … important/healthy/good for the environment"; 7 = pt Likert scale, 1 = totally disagree, 7 = totally agree).
Responses were used to assess baseline differences between (sub)groups (see Table 1) as well as intervention-related changes in consumption habits and attitudes.Non-repeated questions were used for further characterization of our study samples (see Table 1), and assessed prior experience with the PBMA products used in the study (y/n) as well as personal traits: cooking skills (not at all skilled; somewhat skilled; fairly skilled; highly skilled); food neophobia (Pliner & Hobden, 1992), variety seeking tendency (Van Trijp & Steenkamp, 1992) and a subset of food choice motives: health, convenience (Steptoe et al., 1995) and animal welfare (Lindeman & Vaananen, 2000)).

Analysis
Analyses were performed in R version 4.0.5 (R core team, 2021) and the rstatix package (Kassambara, 2021).During the home-use study, one participant (from subgroup A) accidently indicated to have prepared a meal box meal twice in the same week, by which follow-up questions for the self-created meal were incorrect.Responses to these incorrect questions were treated as missing data.Another participant (also from subgroup A) completed the home-use period but did not fill out the post-intervention survey; this participant was included in the analysis of home study measures, but excluded from the survey analysis (leaving participants in subgroup A and 62 participants in the intervention group).

Ecological momentary assessments
Horizontal positions on the visual analogue scales were converted to scores ranging from 0 to 100.For presentation purposes, meal preparation difficulty scores were inversed such that higher scores represented easier preparation.To address our primary research question, effects of repeated exposure and the provision of a meal box scheme on PBMA liking were assessed by means of 4 (Week; within-subjects) x (Meal Type: meal box vs. self-created; within-subjects) x 2 (Subgroup: A vs. B; between-subjects) analyses of variance.Because of the mixed design, liking of PBMA products (PB chicken vs. PB mince) could not be

Survey responses
To assess intervention-related changes in consumption habits and attitudes, pre-vs post-intervention responses were compared between the intervention group and the control group by means of 2 (week 0 vs. week 8; within-subjects) x 2 (intervention vs. control group; betweensubjects) ANOVAs.Within groups, changes over time (week 0 vs. week 8) were assessed with paired samples t-tests; differences between groups (at baseline and in week 8) were assessed with independent ttests.Intervention-related changes in PBMA consumption frequency (which was a categorical question) were assessed with chi-squared analyses.

Study meal preparations
Fig. 2 shows example plates of the meal box recipes prepared with PB chicken (subgroup A; top) and PB mince (subgroup B; bottom).Overall, compliance with the meal box recipes was high: participants reported to have followed the recipe 91% of the time.In two cases, one ingredient was left out (in both cases the cheese sauce in the potato meal).In 15 cases, one or more ingredients were added, mostly condiments (e.g., hot sauce, sweet soy sauce, tomato paste, tabasco) and spices (chili powder, cumin, paprika), but some also mentioned to have used extra vegetables that they would have otherwise thrown away (half a bell pepper, tomatoes).Compliance did not differ between subgroups (p = .50).
As for self-created meals, the meal descriptions and uploaded photos revealed considerable variation in how PBMAs were used.For comparison with the meal box meals, the first and second author classified all meal descriptions into five meal context categories, four of which matched the four meal box meal contexts (Mexican, Asian, Italian and Dutch); meal descriptions that did not fit any of these contexts were classified as 'other/unknown'.Intercoder agreement was 96%, disagreements were resolved through discussion.Both PB mince and PB chicken were used in all meal contexts, but with different frequencies: PB mince was used most often in Italian (32%) and least often in Asian (12%) meal contexts, whereas PB chicken was used most often in Asian (35%) and least often in Mexican (4%) meal contexts (Table 3).
Individual creativity over the course of the four weeks varied across participants.18% used the PBMA in a different meal context each week, against 7% who always used the PBMA in the same meal context.Most participants (54%) prepared the PBMA in three different meal contexts; 21% used the PBMA in two different meal contexts.Example plates from a creative (four different meal contexts) and a less creative cook (only Italian meal contexts) are provided in Fig. 3. Individual creativity did not differ across subgroups (p = .25).
Inspiration for the self-created meals predominantly came from dishes usually prepared with meat (43%) and ingredients participants had at home (34%), suggesting that they mostly relied on their habitual cooking routines.A recipe was used in 13% of the cases.Less frequently mentioned sources of inspiration included the offer in the supermarket (5%), friends/family (1%) and other (3%).Sources of inspiration did not differ between subgroups (p = .70).

PBMA/meal evaluations over time
No evidence was found that PBMA liking changed over the course of the home-use period, (p = .32;Fig. 4, top left), nor that PBMA liking differed across meal types (p = .73).However, the analysis showed a significant Meal Type × Subgroup interaction (F (1,57) = 11.2, p = .001,η p 2 = 0.16).Post-hoc comparisons showed that subgroup A liked PBMA better in self-created meals than in meal box meals (difference 17.2, t (30) = 5.91, p < .001,d z = 1.06), whereas subgroup B tended to like PBMA better in meal box meals than in self-created meals (difference 10.3, t (28) = 1.77, p = .089,d z = 0.33).Given that combinations of meal types and PBMA products were counterbalanced across subgroups, we interpret this result as a main effect of product, that is, both subgroups liked PB mince better than PB chicken.
As for desire-to-eat (Fig. 4, center right), no evidence was found that participants' desire to eat the study meals changed over time (p = .14),but a main effect of Meal Type (F (1,57) = 7.13, p = .010,η p 2 = 0.11) indicated that participants desired to eat their self-created meals more than the meal box meals (difference 5.4).A main effect of Subgroup (F (1,57) = 8.15, p < .001,η p 2 = 0.13) furthermore indicated that Subgroup A had overall more desire to eat the study meals than Subgroup B (difference 8.0).

Liking across meal contexts
Primary analyses provided no evidence that PBMA liking changed with repeated home-use, but note that PBMAs were always prepared and consumed in context, that is, as part of a meal.A correlation analysis between PBMA liking and meal liking scores (averaged over meal types and weeks) showed that PBMA liking and meal liking were strongly related (r = 0.86, p < .001), in line with previous research demonstrating the context-sensitivity of PBMA perception (Elzerman et al., 2011).For further interpretation of this result, we explored whether liking of (meals with) PBMA varied across meal contexts instead of over time.To this end, exploratory analyses were performed separately for meal box meals (with four meal contexts which were identical for all participants) and self-created meals (with five meal context categories which differed between participants; see Table 4) (Fig. 5).
For self-created meals (Fig. 5, bottom panels), meal liking did not significantly differ across the five meal context categories (p's ≥ 0.12).Overall, self-created meals with PB mince (in subgroup A) were liked better than self-created meals with PB chicken (in subgroup B) (F (1,232) = 18.4,p < .001,η p 2 = 0.07).This preference was reduced if PBMA were prepared in Asian meal contexts, but the Meal Context × Subgroup interaction was not significant (p = .40).For PBMA liking, a significant Meal Context × Subgroup interaction F (4,232) = 3.04, p = .018,η p 2 = 0.05) indicated that the preference for PB mince over PB chicken was context-specific: PB mince was preferred in Mexican, Italian and Dutch meals (p's < 0.001), but not in Asian or other/unknown meal contexts (p's ≥ 0.32).Put differently: the overall preference for PB mince over PB chicken was reduced if consumed in identical (meal box) meal contexts, as well as in (self-created) meal contexts less typical for PB mince (yet most typical for PB chicken, i.e.Asian meals; see Table 4 above).
Taken together, these exploratory findings suggest that liking of PBMA strongly depends on the context in which it is consumed, which raises the possibility that the contextual variation over time overruled potential effects of repeated exposure.

User experiences
Participation in the home-use study was overall experienced as enjoyable (M = 5.4; 77% agree, 8% disagree).Participants found the study meals about equally healthy and satiating, but somewhat less tasty, relative to their usual evening meals.30% of participants indicated that preparing PBMA was totally comparable with preparing meat; others found the preparation methods partly (49%) or not at all (21%) comparable.The most frequently mentioned differences were the substantially shorter preparation time of PBMA, and the increased amount of oil needed to prevent PBMA from burning.When asked whether participants would use one or both of the PBMA products again in the future, 77% of participants indicated they would.The willingness to use PB mince again was higher (62%) than the willingness to use PB chicken again (36%).

PBMA/meat consumption habits
Prior to the home-use period (week 0), survey respondents indicated how often they usually ate warm meals with PBMA/meat; four weeks after the end of the home-use period (week 8), they indicated how often they had eaten warm meals with meat/PBMA in the past four weeks, i.e., in the weeks following the home-use intervention.
As for PBMA consumption (Fig. 6, left panel), most respondents indicated to usually eat PBMA less than once per month in week 0; no difference was found between the intervention and the control group (p = .28).In week 8, reported PBMA consumption frequency significantly differed across groups (X 2 (3) = 12.9, p < .005).The percentage of respondents who did not eat PBMA in the four weeks following the intervention period was smaller in the intervention group relative to the control group.A comparison of (within-subjects) changes in reported PBMA consumption frequency (week 0 vs. week 8) showed that the number of changers vs. no-changers significantly differed between the intervention group and the control group (X 2 (2) = 8.2, p = .016).Reported PBMA consumption frequency did not change for the majority of respondents in the control group (65%), but it did for the majority of participants in the intervention group (55%): 38% ate PBMAs more often (most of whom ate PBMA less than once per month in week 0), against 17% who ate PBMAs less often (most of whom ate PBMA 1-2 times per week in week 0).This finding suggests that repeated home-use of PBMA induced an increase in PBMA consumption in less experienced users.
As for meat consumption (Fig. 6, right panel), reported meat consumption frequency in week 0 was significantly higher in the intervention group (M = 4.7 days/week) than in the control group (M = 4.1 days/week), t (236) = 2.41, p = .017,d s = 0.40 (which is not surprising given that we intended to include frequent meat eaters in the home-use study).In the intervention group, reported meat consumption frequency was significantly reduced in week 8 relative to week 0 (Δ = -0.35,t (59) = 2.82, p < .01,d z = 0.37).However, a similar reduction was found in the control group (Δ = -0.37,t (177) = 5.74, p < .001,d z = 0.43), and the ANOVA provided no evidence for a Week × Group interaction effect (p = .86).

Attitudes toward meat (analogue) consumption
Findings regarding intervention-related changes in consumer attitudes toward meat (analogue) consumption were largely inconclusive.Some positive attitude changes were observed in the intervention group: after the intervention, participants were slightly less attached to meat (Δ − 0.14), considered eating less meat to be more important (Δ +0.30) and more strongly agreed that PBMA products were fairtrade (Δ +0.38), but only the latter significantly differed from the control group.Results are provided in Appendix A (Table A1).

Discussion
This real-life intervention study investigated whether consumer acceptance of PBMA changes with repeated in-home preparation and consumption, and assessed potential positive effects of providing recipe inspiration in the form of a meal box scheme.As a secondary objective, potential long-term effects of repeated home-use of PBMA on consumption habits and attitudes were assessed in a survey prior to and   after the intervention period.
Findings from the home-use study provided no evidence that liking of PBMA changed with repeated home-use, neither positively nor negatively.The lack of evidence for positive changes in liking over time was unexpected, as such mere exposure effects have particularly been observed for unfamiliar foods or flavours (Birch et al., 1982(Birch et al., , 1987;;Pliner, 1982;Zandstra et al., 2004).Our sample was relatively unfamiliar with PBMA: prior to the home-use period, the majority of participants consumed PBMA less than once a month, and about half had no prior experience with the products provided.A possible explanation may be the relatively low number of exposures (i.e., 8 exposures to PBMA, and 4 exposures to individual products).Although changes in liking of repeatedly consumed foods have previously been found with as little as five to eight exposures (Kahneman & Snell, 1992;Meiselman et al., 2000;Pliner, 1982;Spill et al., 2019), most studies reporting (either positive or negative) repeated exposure effects used 10 or more exposures.It should also be noted that PBMA were not rejected at first exposure (possibly because our sample was relatively food neophilic; see Table 1), by which room for improvement over time was limited.
Although a positive change in PMBA liking over time could not be demonstrated, the products were generally liked.Moreover, we found no evidence that PBMA liking decreased over time, which is promising and a prerequisite for continued PBMA consumption (also evidenced by most participants' willingness to use the PBMA products again).This finding contrasts with a previous real-life intervention study (Hoek et al., 2013), where repeated home-use of meat substitutes (mycoprotein and tofu) was found to lead to a decrease in liking over time.The difference may again relate to the number of exposures (participants in that study used the same product twice per week for 10 weeks), but also to the (lack of) product variation.By providing participants with two different PBMA products per week, we simulated a more realistic use situation as consumers naturally vary between products; see (Van Trijp et al., 1996).The variation that came with the alternate use of PB mince and PB chicken, which differ in taste, shape, and texture, may have prevented participants from becoming bored with PBMA.That variety can alleviate boredom effects has also been shown in Zandstra et al. (Zandstra et al., 2000), where subjects who were allowed to choose between three different flavours of a meat sauce every week, got less bored over time relative to subjects who received the same flavored sauce every week.Lastly, our exploratory findings revealed that PBMA liking was strongly context-sensitive.The contextual variation over the course of the home-use period may have overruled potential effects of repeated exposure.Given the realistic product variety and the contextual creativity displayed by most participants in their preparations of self-created meals, this raises the question whether boredom effects regarding PBMA are likely to be found in more realistic home-cooking situations.
Our hypothesis that providing consumers with inspiration (in the form of meal boxes with recipes, ingredients and preparation instructions) would have a positive effect on PBMA acceptance was not confirmed.Instead, we found that participants had less desire to eat the meals from the meal box scheme when compared with their self-created meals.Note that our sample mainly consisted of fairly to highly skilled cooks, who displayed considerable creativity in preparing self-created meals.This suggests that the meal boxes (which were deliberately kept basic to remove any barriers regarding preparation difficulty) may have been too simple to satisfy their expectations.Indeed, at the end of the home-use period, various participants spontaneously mentioned that the meal box meals (although easier to prepare than self-created meals) were uninspiring and rather flat in taste, and criticized the use of prepackaged sauces and spice mixes.Alternatively, participants' reduced desire to eat meal box meals relative to self-created meals could be due to the idea of having to cook and eat fixed recipes with fixed ingredients.With self-created meals, participants could more directly respond to their wants: they could decide themselves which meal to prepare and choose the ingredients they wanted at that specific moment.This autonomy was lacking for meal box meals, where the recipe and ingredients were imposed (with the aim of keeping experimental control over meal context effects).This lack of autonomy may have led to a reduced anticipated reward, reflecting a form of 'concept boredom' (see also (Zandstra et al., 2004)).For future studies investigating effects of repeated exposure on PBMA acceptance in home environments, it is recommended to strike a better balance between experimental control and autonomy.Researchers could have participants use the same PBMA product(s) repeatedly, but provide a more flexible meal box scheme, for instance by offering a list of product-specific recipes to choose from, varying in complexity to serve participants with diverging cooking skills.Lastly, given that branding and price can also influence food reward (Motoki & Suzuki, 2020), it should be noted that the meal box ingredients were predominantly private label products (provided in their original packaging) and were provided free of charge, both of which may have influenced participants' quality perceptions.One way to avoid this would be to have participants purchase additional meal ingredients themselves.
Findings from the exploratory analyses underline the importance of the meal context in PBMA perceptions (Elzerman et al., 2011;Hoek et al., 2013).Not only did PBMA liking increase with increased meal liking, product-specific preferences were also found to be context-sensitive.The overall preference for PB mince over PB chicken largely disappeared in identical (meal box) meal contexts, as well as in (self-created) meal contexts less typical for PB mince (yet more typical for PB chicken).These findings are in line with (Elzerman et al., 2011), who found similar product-specific contextual preferences forPBMA (mycoprotein pieces and mycoprotein mince), and provide further suggestive evidence that other meal components can mask the sensory and hedonic qualities of PMBA (Elzerman et al., 2011).Taken together, our exploratory findings suggest that suitable meal contexts are crucial for consumer acceptance of PBMA, and contextual variation may positively contribute to PBMA liking, although further confirmatory research is needed.It should also be kept in mind that contextual effects on product-specific preferences in the current study were based on between-subjects comparisons, by which we cannot rule out that differences were (partly) due to individual variation.Although no differences between subgroups were found in cooking skills, consumption habits or personal traits, note that subgroup B (who prepared self-created meals with PB chicken and meal box meals with PB mince) was (slightly) more attached to meat, which may have contributed to the liking differences between PBMA products in self-created meals.
Survey findings provided limited evidence for intervention-related changes in meat (analogue) consumption habits and attitudes.A significant increase in self-reported PBMA consumption frequency was observed in the intervention group, especially among participants with a low use frequency at baseline (less than once per month).This is a promising result, as it suggests that repeated home-use may facilitate the adoption of PBMA in consumers' habitual (animal-based) cooking routines.A significant decrease in self-reported animal meat consumption was also found, but a similar effect in the control group prevented us from attributing this to the intervention.This finding possibly reflects a more general trend towards eating less meat observed in Dutch consumers (Natuur & Milieu, 2020;Petit et al., 2021).Another possibility is that repeatedly asking consumers about their meat consumption increased their self-awareness on the topic, which may have induced a social desirability bias.The lack of evidence for further intervention effects may in part be due to our choice for a parallel rather than cross-over design.Note, however, that two recent intervention studies (also using parallel designs) did find long-term positive effects of repeated home use of PBMA on meat consumption and attitudes up to three months after the intervention period (Bianchi et al., 2022;Trewern et al., 2022).Both these studies differ from the current study in that the interventions consisted of additional components besides the provision of free PBMA products and recipe inspiration, such as online cook-along sessions, information about the health and environmental benefits of eating less meat, an online community to exchange ideas and inspiration, and challenges to implement new knowledge and skills.All of these additional factors may have contributed to the positive effects we failed to observe here.We tentatively conclude that provision of PBMA with (little inspiring) recipe suggestions alone may not be enough to bring about a structural change toward a more sustainable consumption pattern or a more positive attitude towards PBMA.Additional research is needed to specify which (combination of) intervention strategies is (are) most successful.
Lastly, this study makes a case for research in real-life settings when investigating consumer acceptance of alternative proteins.Most research in this domain is directed at pre-implementation phases, i.e., investigating attitudes towards and intentions to eat alternative proteins through surveys and focus group discussions, or assessing perception in single tasting sessions (Weinrich, 2019).Recent reviews mention that even intervention studies aimed at reducing meat intake often measure attitudes and intentions rather than actual consumption behaviour, and highlight the need for studies assessing actual consumption (Harguess et al., 2020;Kwasny et al., 2022;Mathur et al., 2021).Real-life interventions are important because food perception and eating behaviour are known to be sensitive to the environment in which a food is consumed (Boutrolle et al., 2007;de Graaf et al., 2005;Petit & Sieffermann, 2007).Also, the social context can influence food perception, food choice and food consumption behaviour (Boothby et al., 2014;Herman et al., 2003;Robinson et al., 2013), and a lack of social support has been argued to be an important barrier for reducing meat consumption (Graça et al., 2019;Sijtsema et al., 2021).It is therefore crucial to measure the effectiveness of an intervention there were the change needs to be made.So far, only few real-life intervention studies have been conducted in the field of alternative protein consumption and reducing meat intake (Onwezen et al., 2021a;Taufik et al., 2019).With the current study, we have contributed to filling this gap by gaining insight in how consumers use, perceive, and appreciate PBMA over time in their own homes.Note, however, that intervention-related changes in consumption patterns were assessed with self-report questions.For future studies, a more reliable measure that is less susceptible to socially desirable responses (such as a food diary, Bianchi et al., 2022, or actual consumption behaviour) is advised.

Conclusion
With the present study, we have gained valuable insights in how regular meat eaters use, perceive and appreciate PBMA in home-cooked meals.We investigated (a) to what extent PBMA liking is influenced by repeated home-use and by inspiration in the form of a meal box scheme; and (b) whether repeated PBMA consumption can facilitate a shift toward a more sustainable consumption pattern.No evidence was found that liking of PBMA changed with repeated consumption, and positive effects of a meal box scheme could not be demonstrated (likely due to the simplicity of the meal box recipes).Exploratory findings highlighted the importance of a suitable meal context for consumer perceptions of PBMA.Although findings suggest that repeated exposure can stimulate longer-term use of PBMA, more seems needed to bring about a structural reduction in meat consumption.Future research should investigate whether more appropriate recipe suggestions that are tailored to consumers' cooking skills and (meat-centered) meal expectations can improve perceptions of PBMA and facilitate the transition towards more sustainable diets.

Table A1
Mean (SD) attitudes (week 0 and week 8) and attitude changes (Δ) in the intervention group and the control group.Significant changes in bold, non-significant changes in grey.Superscript letters indicate significant differences between groups.

Fig. 1 .
Fig. 1.Experimental design and study procedure.Grey boxes indicate the measures obtained during the study.

Fig. 2 .
Fig. 2. Example plates of the four meal box meals prepared with PB chicken (top) and PB mince (bottom).

G
.van Bergen et al.

Fig. 3 .
Fig. 3. Example plates of self-created meals prepared by a creative (top) and a less creative cook (bottom).

G
.van Bergen et al.

Fig. 4 .
Fig. 4. Mean evaluations of PBMA (left) and full meals (right) per week.Solid lines indicate meal box meals, dashed lines represent self-created meals.Red lines indicate PB chicken, blue lines represent PB mince.Error bars show ± SEM.

Fig. 5 .
Fig. 5. Mean PBMA and meal liking scores per meal context in meal box meals (top) and self-created meals (bottom).Red bars indicate PB chicken, blue bars represent PB mince.Different superscript letters indicate significant differences between meal contexts; Asterisks indicate significant differences between PBMA products within meal contexts.Error bars show ± SEM.

Fig. 6 .
Fig. 6.Reported consumption frequency of PBMA (left) and meat (right) with warm meals prior to and after the home-use period in the intervention group (yellow trian-gles) and the control sample (brown circles).

Table 1
Sample characteristics of intervention group (full sample and subgroups) and control group.Different superscript letters indicate significant differences between groups.

Table 2
Meal box meal ingredients.
Additional questions (not reported here) concerned the use of the research app to learn lessons for future home-use studies.

Table 3
Ecological momentary assessments.directly in the same analysis; these were derived from the Meal Type × Subgroup interactions.Greenhouse-Geisser corrections were applied to factors violating the sphericity assumption (assessed with Mauchly's test of sphericity).F-statistics, p-values and effect sizes (partial η 2 ) are reported.Significant effects (p < .05)were further interpreted with either paired (within-subjects effects) or independent (between-subjects effects) t-tests; test statistics with corresponding pvalues (Bonferroni-corrected in case of multiple comparisons) and effect sizes (Cohen's d) are reported.The same analyses were performed for the other VAS scores. compared

Table 4
Frequencies (%) of self-created meal contexts of PB chicken (subgroup B) and PB mince (subgroup A).