Table 1.

Study level characteristics

	Eligibility	Recruitment	Attrition and sample size*	Publication status	Effective Public Health Practice Project Quality Assessment Tool for quantitative studies†
Randomised controlled trial‡
Bacon and Krpan (2018), UK41	Individuals had to be resident in the UK, have English as their first language, and not follow a diet precluding the choice of meat	Individuals were recruited through the Prolific Academic research platform	T1: n=564; attrition unknown§	Peer reviewed publication	Medium
Kongsbak et al (2016), Denmark42	Individuals had to be male university students aged between 18 and 29 years old	Individuals were recruited through advertisement on social media and on Aalborg University campus	T1: n=65, attrition 0%	Peer reviewed publication	Medium
Kunst and Hohle (2016), study 2b, Norway43	Individuals had to be Americans from the USA	Individuals were recruited through Amazon Mechanical Turk	T1: n=101, attrition unknown§	Peer reviewed publication	Low
Kunst and Hohle (2016), study 5, Norway43	NA	Individuals were recruited through Amazon Mechanical Turk	T1: n=190, attrition unknown§	Peer reviewed publication	Low
Kunst and Palacios Haugestad (2018), American sample, Norway44	Individuals had to be Americans from the USA who consumed meat and were 18 years or older	Individuals were recruited through Amazon Mechanical Turk	T1: n=178, attrition unknown§	Peer reviewed publication	Low
Kunst and Palacios Haugestad (2018), Ecuadorian sample, Norway44	Individuals had to be Ecuadorians who consumed meat and were 18 years or older	Individuals were recruited through snowball sampling on social networks	T1: n=183, attrition unknown§	Peer reviewed publication	Low
McClain et al (2013), USA45	Individuals had to be 18–23 years old, have a meal plan with the residence dining hall, and eat at the dining hall at least three days per week	A convenience sample was recruited by approaching students who used one of the four participating cafeterias	T1: n=525 (individual responses), attrition NA	Peer reviewed publication	Strong
Sorensen et al (2005), USA39	Small businesses had to be manufacturing industries with 50–150 employees, with at least 25% of workers being first generation or second generation immigrants or people of colour, a turnover rate during the past year of less than 20%, and the capacity to decide to participate	Eligible small businesses were actively approached and asked to participate	T1: n=1740 (individual responses), attrition NA (8% of worksites withdrew)	Peer reviewed publication	Medium
Vermeer et al (2010), The Netherlands46	Individuals had to be 18 years or older	Individuals visiting a Dutch fast food outlet were approached and asked to participate in the study after they made their purchase	T1: n=137, attrition 9%	Peer reviewed publication	Medium
Crossover randomised controlled trial‡
Reinders et al (2017), The Netherlands47	Individual meals had to be of the relevant menu items (eg, exclusion of vegetarian meals, child menus, and special offerings), coming from parties with fewer than 12 orders, and from customers who completed questionnaires	All eligible individual orders placed during the study period in participating restaurants were recorded; restaurants were actively approached for recruitment	T1: n=1006, attrition NA	Peer reviewed publication	Strong
Rolls et al (2010), USA48	Individuals had to be 20–45 years old, have a BMI between 18 and 40 kg/m2, regularly eat three meals per day, and like and be willing to eat all three foods served in the test meals; individuals were excluded if they were dieting to gain or lose weight, had food allergies or restrictions, were taking medications known to affect appetite, or were smokers, athletes in training, pregnant or breastfeeding, had symptoms of depression, or had disordered attitudes towards food	Individuals were recruited through advertising in local newspapers and university mailing lists	T1: n=48, attrition 0%	Peer reviewed publication	Medium
Factorial randomised controlled trial‡
Campbell-Arvai et al (2014), USA49	Undergraduate students living on campus	Individuals were actively approached and invited to take part in the experiment upon entering the dining facilities on campus	T1: n=319, attrition 0%	Peer reviewed publication	Medium
Multiple treatment reversal‡¶
Stewart et al (2016), study 1, UK38	University dining halls with appropriate booking systems	The staff of eligible university dining halls were actively approached; all bookings placed during the study were recorded	T1: n=5280 (individual orders), attrition NA	Unpublished	Medium
Stewart et al (2016), study 2, UK38	University dining halls with appropriate booking systems	The staff of eligible university dining halls were actively approached; all bookings placed during the study were recorded	T1: n=782 (individual orders), attrition NA	Unpublished	Medium
Stewart et al (2016), study 3, UK38	University dining halls with appropriate booking systems	The staff of eligible university dining halls were actively approached; all bookings placed during the study were recorded	T1: n=61 (individual orders), attrition NA	Unpublished	Medium
Pre-post design‡
Clark (2017), UK50	Individuals had to be aged between 21–50 years, have a BMI of 18–28 kg/m2, be healthy men or women (premenopausal), have good spoken and written English, consume four to five portions of red or processed meat per week, not smoke, not have a chronic disease, not be pregnant or breast feeding, not use chronic medication (excluding over the counter medication and oral contraceptives), not have participated in other research 3 months before screening, and not have clinically significant findings at screening	Individuals were recruited through advertising in newspapers, on social media pages, and in different online and offline facilities of the University of Nottingham	T1: n=26, attrition 39·5%; T2: n=22, attrition 48·8%	Unpublished	Medium
Flynn et al (2013), USA51	Individuals had to have access to transport to attend study activities, be willing to try new recipes, and be contactable by telephone	Individuals were recruited through advertisement in and referral from emergency food pantries	T1: n=63, attrition 26%	Peer reviewed publication	Strong
Holloway et al (2012), UK52	Individuals had to consume meat at least four to five times weekly, be 18–30 years old, not take regular meals in halls of residence or not live with parents or partners, be free of chronic disease, and have a BMI of 22–27 kg/m2	Individuals were recruited through a brief advertising presentation to around 350 students in Nottingham University	T1: n=19, attrition 27%	Unpublished	Medium

NA=not available. BMI=body-mass index.

^*T1 and T2 respectively refer to the shortest and longest available post-intervention follow-up. This information refers to data underlying the analyses of meat demand.

^†The Effective Public Health Practice Project Quality Assessment tool for Quantitative Studies rating is based on study design, selection bias, confounders, blinding, data collection method, withdrawal, and dropouts. Studies with more than two weak ratings in the aforementioned dimensions were assigned a low overall rating, studies with one weak rating were assigned a medium overall rating, and studies with no weak ratings were assigned a strong overall rating.

^‡The study design refers to the design underlying the main comparison reported in this review.

^§These studies used a one-off survey with an experimental component and might only have sourced data from participants who started and completed the survey. For these studies, we consider attrition to be unknown.

^¶Multiple treatment reversal designs refer to experimental studies in which intervention periods and control periods are sequentially alternated over an extended time period.