Interventions to change purchasing behaviour in supermarkets: a systematic review and intervention content analysis

Figures & data

Figure 1. Objectives of study and key steps taken in coding and mapping process. BCTs = Behaviour change techniques. BCTTv1 = Behaviour change techniques taxonomy version one. BCW IF = Behaviour change wheel intervention functions. TIPPME = Typology of interventions in proximal physical microenvironments.

Figure 2. PRISMA Flow Diagram for Included Studies. From: Moher et al. (2009).

Table 1. Assessment of Risk of Bias in Studies; Rating for Each Question in GATE Checklist.

	GATE Checklist Question Number
	Population			Method of allocation to intervention										Outcomes						Analyses						Summary
Authors (Year)	1.1	1.2	1.3	2.1	2.2	2.3	2.4	2.5	2.6	2.7	2.8	2.9	2.10	3.1	3.2	3.3	3.4	3.5	3.6	4.1	4.2	4.3	4.4	4.5	4.6	5.1	5.2	Overall
Achabal et al. (1987)	+	++	++	++	++	NR	+	++	++	++	++	++	++	++	++	+	++	++	+	NA	NA	++	NR	++	+	++	++	++
Ball et al. (2015)	++	++	++	++	+	++	+	++	+	++	++	++	++	++	++	++	++	++	++	++	++	++	−	+	−	++	++	++
Bevelander et al. (2011)	++	++	+	+	++	NR	+	++	++	++	++	++	++	++	++	++	++	++	NA	++	NA	++	++	++	NR	++	++	++
Chevalier (1975)	+	++	++	+	+	NR	+	+	+	++	++	++	++	++	++	+	++	NA	−	NR	NA	−	++	+	+	+	−	−
Cox (1964)	+	++	++	+	++	NR	+	++	+	++	++	++	++	+	++	+	++	NA	−	NR	NA	+	+	++	+	+	+	+
Cox (1970)	+	++	++	+	++	NR	+	++	++	++	++	++	++	+	++	+	++	NA	−	NR	NA	+	+	++	−	+	+	+
Curhan (1974)	NR	++	++	+	+	NR	+	++	++	++	++	++	++	+	++	+	++	NA	−	NR	NA	−	++	−	−	−	−	−
De Wijk et al. (2016)	+	++	+	+	++	NR	+	++	++	+	++	++	++	++	++	++	++	NA	−	−	NA	+	++	+	+	+	−	−
Dhar & Hoch (1996) Study 1	++	++	++	+	++	+	NR	++	++	++	++	++	++	++	++	++	++	++	+	++	NA	++	++	++	+	++	++	++
Dhar & Hoch (1996) Study 2	++	++	++	+	++	+	NR	++	++	++	++	++	++	++	++	++	++	NA	NA	+	NA	++	++	++	+	++	++	++
Foster et al. (2014)	++	++	++	++	++	+	+	++	+	++	++	++	++	++	++	++	++	NA	−	++	NA	++	NR	++	+	++	++	++
Gamburzew et al. (2016)	++	++	++	−	++	−	+	++	++	++	++	++	++	++	++	++	++	NA	−	+	NA	+	++	++	+	+	++	+
Geliebter et al. (2013)	++	++	+	++	++	NR	+	++	++	++	−	++	++	++	++	+	++	++	+	++	+	++	++	++	NR	++	+	+
Heilman et al. (2002)	+	+	+	−	+	NR	+	++	++	++	+	++	++	++	+	+	++	NA	−	NR	NA	+	++	+	+	+	+	+
Iyengar & Lepper (2000) Study 1	+	++	+	+	++	NR	+	+	+	++	+	++	++	++	++	+	++	NA	−	NR	NA	+	NR	+	+	−	−	−
Jeffery et al. (1982)	++	++	++	+	++	NR	+	++	+	++	++	++	++	+	++	++	++	NA	−	NR	NA	+	+	+	NR	+	+	+
Kiesel & Villas-Boas (2013)	NR	NR	NR	−	−	−	NR	+	+	+	++	++	++	++	++	−	++	+	+	NR	NA	NR	++	−	NR	+	−	−
Levy et al. (1985)	++	++	++	−	++	−	+	+	++	+	++	++	++	++	++	++	++	NA	−	−	NA	++	++	+	−	−	−	−
McCabe et al. (2015)	−	NR	+	++	++	++	++	++	++	++	++	++	+	++	++	++	++	NA	−	++	NA	+	++	++	+	++	+	+
Meuldijk et al. (1996)	NR	+	+	−	++	−	−	++	+	++	NR	++	++	++	+	+	++	NA	−	NA	NA	−	++	++	−	−	−	−
Milliman et al. (1982)	+	++	++	++	++	NR	+	++	++	++	++	++	++	++	++	++	++	NA	−	NR	NA	+	++	++	+	++	++	++
Milliron et al. (2012)	+	++	+	++	++	+	+	++	++	++	++	++	++	++	++	++	++	NA	−	++	NA	++	++	++	+	++	+	+
Muller (1984)	+	++	++	+	++	NR	+	++	+	++	++	++	++	++	++	++	++	NA	−	NR	NA	−	NR	++	+	+	−	−
Ni Mhurchu et al. (2010)	++	++	++	++	++	++	+	++	++	++	++	++	++	++	++	++	++	++	++	++	++	++	++	++	++	++	++	++
Ogawa et al. (2011)	+	++	++	−	++	NR	NR	++	++	++	++	++	++	++	++	++	++	NA	−	+	NA	−	++	++	++	+	+	+
Papies et al. (2014)	+	+	+	NR	++	NR	+	++	++	++	++	++	++	++	++	++	++	NA	−	++	NA	+	++	++	+	+	+	+
Patterson et al. (1992)	+	++	++	−	++	−	NR	+	++	++	++	++	++	++	++	+	++	NA	−	−	NA	++	++	+	+	−	−	−
Payne et al. (2015)	NR	++	++	+	++	NR	+	++	+	++	++	++	++	++	++	++	++	NA	−	NR	NA	NR	++	++	+	+	+	+
Payne et al. (2016)	+	++	++	+	++	NR	NR	++	++	+	++	++	++	++	++	++	++	NA	−	+	NA	−	++	+	+	+	+	+
Russo et al. (1986) Study 1	++	++	++	−	−	NR	NR	++	+	−	+	++	++	++	++	++	++	NA	−	NR	NA	−	++	−	NR	−	−	−
Russo et al. (1986) Study 2	NR	++	++	−	+	NR	NR	++	+	++	++	++	++	++	++	++	++	++	+	NR	NA	−	NR	+	NR	−	−	−
Saarela et al. (2013)	+	++	+	−	++	−	−	++	−	++	++	++	+	++	++	+	++	NA	−	NA	NA	+	NR	−	−	−	−	−
Salmon et al. (2015)	+	++	+	+	++	+	+	++	++	++	++	++	++	++	++	++	++	NA	−	++	NA	++	++	++	+	++	+	+
Schwartz et al. (2014)	+	++	+	+	++	NR	+	++	++	++	++	++	++	++	++	+	++	NA	−	++	+	++	++	++	+	++	+	+
Sigurdsson et al. (2011) Study 2	+	+	++	++	++	+	+	++	+	++	++	++	++	++	++	+	++	NA	−	NR	NA	+	++	+	−	−	+	−
Sigurdsson et al. (2014)	+	++	++	−	++	NR	+	+	+	+	++	++	++	++	++	++	++	+	+	NR	NA	−	−	+	−	−	−	−
Smith & Curnow (1966)	+	++	−	NR	−	NR	NR	+	+	++	++	++	++	−	++	+	+	NA	−	NR	NA	−	NR	−	+	−	−	−
Surkan et al. (2016)	++	++	++	−	+	NR	NR	+	++	++	++	++	++	++	++	++	++	NA	−	++	NA	−	++	+	−	−	−	−
Tal & Wansink (2015) Study 1	NR	NR	−	+	++	NR	+	++	+	++	−	++	++	++	+	−	−	NA	−	NR	NA	−	NR	−	+	−	−	−
Teisl & Levy (1997)	+	++	++	−	++	NR	+	+	+	+	++	++	++	++	++	++	++	NA	−	NR	NA	++	++	−	−	−	−	−
Vlaeminck et al. (2014)	+	+	+	+	++	NR	+	++	++	++	++	++	++	++	++	++	++	NA	−	++	NA	+	++	++	+	++	+	+
Wansink et al. (1998) Study 1	NR	++	++	+	+	NR	+	++	++	+	++	++	++	++	++	+	++	NA	−	NR	NA	++	++	−	−	+	+	+
Wansink et al. (1998) Study 2	+	−	+	−	++	NR	+	+	++	++	++	++	++	+	+	+	++	NA	−	NR	NA	+	++	++	+	+	−	−
Waterlander et al. (2013)	++	+	++	++	++	++	+	++	++	++	−	++	++	++	++	++	++	++	++	++	+	++	++	+	+	++	++	++
Wilkinson et al. (1982)	+	++	++	−	+	+	+	++	+	++	++	++	++	+	++	+	++	NA	−	NA	NA	+	NR	++	+	+	−	−
Winett et al. (1991)	NR	+	+	+	++	NR	+	+	++	+	++	++	+	++	+	+	++	++	+	NR	+	−	++	+	+	−	−	−
Winett et al. (1997)	+	+	−	+	++	NR	+	+	++	++	+	++	+	++	+	++	++	++	++	++	+	−	++	+	+	−	−	−
Winkler et al. (2016)	++	+	++	+	++	NR	NR	+	+	+	++	++	++	++	++	++	++	++	+	+	NA	+	++	+	++	+	++	+
Woodside & Waddle (1975)	NR	++	++	+	+	NR	+	+	+	++	++	++	++	+	++	+	++	NA	−	NR	NA	+	NR	++	+	+	−	−

Note: (++) = low risk of bias, (+) = moderate risk of bias, (−) = high risk of bias, NR = not reported, NA = not applicable.

Table 2. Characteristics of Studies Included in Narrative Review and Intervention Content Analysis.

Authors (Year)	Design	Unit of Allocation	Country	Setting & Store Demographics	Participants	Theoretical Basis	Intervention and Control Details	Duration and Fidelity	Outcome	Key Findings
Achabal et al. (1987)	N-RCT	Store (B)	Dallas, San Francisco & Washington DC, USA	372 (ctrl = 124, int 1 = 124, int 2 = 124) Safeway Stores across greater metropolitan area; stores matched as triplets across 3 areas, based on size	Whole store intervention. N = 26,784 individual weekly sales observations	No explicit theoretical basis. Examined effect of ‘provision of information’	POP signage by target F&V items; signage presented either produce selection information or purchase selection information plus nutritional information	4 wks pre-test, 4 wks int (signs in place), 4 wks post-test. F/U not conducted. Fidelity NR beyond initial implementation.	Sales data for 6 F&V items (broccoli, cabbage, carrots, cauliflower, kiwi fruit, tomatoes)	Considered ineffective. No effect seen of POP signage on purchases of 6 types of F&V. F/U planned but not conducted, as no intervention effect
Ball et al. (2015)	RCT	Individual (B)	Melbourne, Victoria, Australia	Coles supermarket chain. Recruitment targeted at two stores; one each from an advantaged and disadvantaged area	N = 642, F = 100%, m age = 43.7 yrs, ethnicity NR, low SES = 44%, high SES = 55%	Theoretical basis: social ecological and social cognitive theories. Examined effect of ‘nutrition education & skill-building’ & ‘fiscal interventions’	Skill-building behaviour change int for healthy eating, including budgeting worksheets, self-monitoring exercises, recipes and web-based forum (n = 160), 20% price-reduction on target F&V items, water & diet/low kcal drinks (n = 161), combined skill-building and 20% price-reduction (n = 160), ctrl, no int (n = 161)	3 mth retro B/L data, 3 mth int, F/U at 6 mth post-int. Fidelity NR for education element. Discounts were applied to loyalty card, but no control over card use beyond int period.	Sales data for target F&V & beverage items	Considered partially effective (education alone ineffective). Price reduction group purchased more F&V compared with B/L (+15% vegetables & +35% fruit). Combined price-reduction & skill-building group purchased more fruit. Skill-building and ctrl groups, no increases in F&V. No changes in bottled water or low kcal beverages across any group. Effects not maintained at 6 mth F/U
Bevelander et al. (2011)	RCT	Individual (B)	The Netherlands	8 stores, chain NR	N = 89, F = 100%, (grade 5 students n = 50, m age = 10.48 yrs; grade 6 students n = 39, m age = 11.3 yrs), SES & ethnicity NR, underweight = 3.4%, normal weight = 91%, overweight or obese = 5.6%	Theoretical basis: social modelling (shoppers are influenced by their peers). Examined effect of ‘social modelling’	Participants (low kcal n = 31, middle kcal n = 30, high kcal n = 28) instructed to shop with a peer to purchase 5 items for lunch for a school trip. Peers were confederates, previously instructed to purchase 5 low, middle or high kcal items	Each peer pair only shopped once, no F/U. Study lasted 6 mth. Fidelity: 14 shopping couples excluded as confederate did not follow instructions	Kcal content of purchases & whether participant’s choices matched confederate’s items	Considered effective. Food match: compared with confederate, 11.2% of participants purchased identical items, 20.2% purchased same number of items, 76.4% purchased more items, 3.4% did not match purchased items at all & 12.4% purchased 1 similar item. Snacks most matched item. Food purchase: high kcal condition purchased higher kcal products compared to those in low kcal condition. No difference between high & medium kcal or medium & low kcal conditions
Cox (1964)	RCT	Store (W)	Austin, Texas, USA	6 stores from same chain, chain NR	Whole store intervention. N for customers or transactions NR	No explicit theoretical basis. Examined effect of ‘changes in shelf space’	Varied shelf space: 4 products, each with 6 conditions, ranging from low (2, 4 or 6 spaces) to high (12, 14 or 21) numbers of shelf spaces.	Summer of 1962 (6 wks total). Display changes made once per week. No F/U. Fidelity NR.	Sales (measured via 4 audits per week) of 4 products: baking soda, tang, hominy & powdered coffee cream.	Considered partially effective. Sales of hominy affected by shelf-space; greater sales related to greater number of product-facings. No effects seen for powdered coffee cream, tang or baking soda
Cox (1970)	RCT	Store (W)	Houston, Texas, USA	6 large stores in Weingarten chain; 4 from regional chain & 2 from local chain	Whole store intervention. N for customers or transactions NR	No explicit theoretical basis. Examined effect of ‘changes in shelf space’	Varied proportion of shelf space (1/3, 1/2 or 2/3 of shelf) between 2 brand options (high or low consumer acceptance)	3 wks total; each display in place for 1 week. No F/U. Fidelity NR.	Sales (measured via daily shelf inventory) of 2 brands of salt & 2 brands of powdered coffee cream	Considered partially effective. Increased shelf space increased sales of Coffeemate (high consumer acceptance) but not Creamora (low consumer acceptance). No effect seen for branded salt items.
Dhar & Hoch (1996) Study 1	RCT	Store (B)	Chicago, USA	All 86 (ctrl = 43, int = 43) stores in Dominick’s Finer Foods chain; 20% of Chicago’s metropolitan market	Whole store intervention. N for customers or transactions NR	Theoretical basis: price theory, push promotions, pull promotions, price perception (reference prices), attitudes towards coupons, conditioning, novelty, opportunity costs of time. Examined effect of different in-store presentations of price promotion	Compared presentation of price discounts. In-store coupons (hung on shelf directly in-front of promoted product, to be taken to till-point for redemption) compared with ‘Bonus Buys’ (shelf tag displaying price discount). Price discounts varied: 13% – 30% off regular price	Study conducted over 6 mth. Target items promoted for 1 week. 2 mth F/U. Fidelity: each store audited once per week of promotion to ensure compliance (any issues with non-compliance are NR)	Sales data for target product against rest of product category (RTE cereals & soft drinks); described as popular package size of leading national brand.	Considered effective. Increased purchases seen in RTE cereals & soft drinks for both bonus buy promotions and in-store coupons. Bonus buy: RTE cereals = 199% increase, soft drinks = 218% increase. In-store coupons: RTE cereals = 172% increase, soft drinks = 368% increase. F/U – no evidence of post-promotion sales increase.
Dhar & Hoch (1996) Study 2	RCT	Store (B)	Chicago, USA	All 86 (ctrl = 43, int = 43) stores in Dominick’s Finer Foods chain; 20% of Chicago’s metropolitan market	Whole store intervention. N for customers or transactions NR	Theoretical basis: price theory, push promotions, pull promotions, price perception (reference prices), attitudes towards coupons, conditioning, novelty, opportunity costs of time. Examined effect of different in-store presentations of price promotion	Varied level of price discount offered using in-store coupons: 20%, 30%, 35%, 40% & 50%	5 wks, each coupon offered for 1 week. No F/U. Fidelity: each store audited once per week of promotion to ensure compliance (any issues with non-compliance are NR)	Sales data for leading national brands in the RTE cereal category	Considered effective. Coupons increased sales of promoted brands of RTE cereals and decreased sales of rest of category. Effect increased as discount increased, e.g., 20% discount = +172% promoted brand & −2.43% rest of category; 50% (3 brands) = +731% to +959% promoted brand & −3.20% to −5.55%
Foster et al. (2014)	RCT	Store (B)	Philadelphia & Wilmington, USA	8 stores (ctrl = 4, int = 4); 2 chains, 4 stores each. Urban location, low income, high minority ethnic communities, chain NR	Whole store intervention. N for customers or transactions NR	Theoretical basis: 4 Ps framework (price, promotion, product, placement). Examined effect of all 4 Ps, focused on placement and promotion	Range of product placements to give preference to healthier foods: (1) multiple facings (> shelf space), (2) prime location, e.g., eye level, (3) signage, (4) secondary placement (e.g., aisle end), (5) taste-testing & (6) cross promotion	3 mth B/L, 6 mth int. No F/U. Fidelity: visual planograms provided to store staff to promote consistency across stores. Assessed with weekly unannounced visits by research team. Average compliance across 4 stores varied from 64% to 85%. Average compliance across 5 product categories varied from 42% (beverages) to 92% (milk)	Sales data for 5 categories: milk, frozen meals, RTE cereal, in-aisle beverages, checkout cooler beverages	Considered partially effective. Possible intervention effect for skimmed & 1% milk (sales stable in int stores, but decreased in ctrl stores). Some healthier ready meal purchases and bottled water purchases increased. No other int effects identified
Gamburzew et al. (2016)	RCT	Store (B)	Marseille, France	4 (ctrl = 2, int = 2) stores from DIA retail chain in 2 disadvantaged neighbourhoods	Whole store intervention. N = 6,625 loyalty card customers included, who made 100% of purchases in same store during B/L and intervention	No explicit theoretical basis (but considers multiple influences to explain attitude-behaviour gap). Examined effect of ‘shelf labelling & social marketing strategy’	Combined strategies to target products (inexpensive, good nutritional food): (1) shelf labels, (2) posters and leaflets (3) taste testing booth and leaflets focused on canned fish, pulses and eggs	6 mth B/L (same time period in previous year), 6 mth int. No F/U. Fidelity: implementation of intervention performed by research team	Sales data for 6 categories: F&V, starches, meat/fish/eggs, mixed dishes & sandwiches, dairy products & other. Calculated contribution of promoted items to total spending on food and food category	Considered partially effective. No effect seen for overall sales of all targeted products. Sales of F&V & starches categories were higher in int stores compared to ctrl
Geliebter et al. (2013)	RCT	Individual (B)	Manhattan, USA	2 stores in D’Agostino chain	N = 47, F n = 33, m age = 37.5 yrs, Caucasian = 56%, African American = 19%, Asian = 9%, Hispanic = 13%, Other = 2%, no response = 2%, BMI m = 30.2 kg, median family income similar across stores	No explicit theoretical basis. Examined effect of ‘environmental economic intervention’	50% price discount on target items: ctrl (n = 19), int (n = 28)	4 wks B/L, 8 wks int, 4 wks F/U. Fidelity: price discounts enabled via store loyalty card. 10 participants excluded due to non-compliance or incomplete purchasing data	Sales data for low-energy density F&V, bottled water & diet sodas	Considered partially effective. Weekly purchasing of target F&V greater in discount group (US$5.50) compared to B/L & ctrl group (US$1.69). No effect seen for water & diet sodas. F&V effect not maintained at F/U
Heilman et al. (2002)	N-RCT	Individual (B)	St Louis, Missouri, USA	2 stores, middle class suburb, chain NR	N = 192, gender & ethnicity NR, Income: <US$35k = 28%, US$35-70k = 37%, >US$70k = 35%, employment status: 13% = retired, 20% = stay-at-home, 26% = full-time, 33% = part-time, 4% = self-employed, 4% = unemployed	Theoretical basis: psychological income effect & elevated mood from unanticipated gains, priming effect, memory effect. Examined effect of ‘surprise coupons’	In-store discount coupons (US$1.00 off) for items not on shopping list (unplanned purchases): ctrl n = 87, int n = 105	Study period NR; one-off shop for each shopper once. No F/U. Fidelity: Research team intercepted all participants and provided intervention group with coupons	Sales (as checked via receipts) of target unplanned items (spaghetti sauce, cereal & 2 non-food items) compared to sales on planned items (checked to shopping lists)	Considered effective. Customers who received coupon made average 47% more unplanned purchases that ctrl customers (11.37 unplanned purchases, worth $27.26 vs. 7.76 worth $19.58). Total spend was greater amongst those who received coupon ($83.91) compared to those who didn't ($75.22)
Jeffery et al. (1982)	RCT	Store (B)	Minneapolis, St. Paul, & suburban areas, Minnesota, USA	8 stores (ctrl = 4, int = 4) from same chain; matched on basis of shoppers’ education level. Located in inner-city, middle-class & affluent areas	Whole store intervention. N for customers or transactions NR	No explicit theoretical basis. Examined effect of ‘nutrition education’	Nutrition education in-store. Wave 1: 6 posters, 12 signs identifying low-fat options. Wave 2: 1 poster & 6 tear-off recipe cards. Wave 3: 1 poster & 12 tear-off cards with low-fat recommendations	6 mth B/L, 6 mth int: Wave 1 = 12 wks, Wave 2 = 6 wks, Wave 3 = 6 wks. No F/U. Fidelity: materials in each store checked weekly for damage & replaced as necessary	Sales of 25 targeted dairy & frozen dessert items, measured via weekly shipments to stores from wholesalers	Considered ineffective. No effect seen on ratio of low-fat to higher-fat targeted dairy & frozen dessert items
McCabe et al. (2015)	RCT	Individual (B)	Mid-west state, USA	1 store, chain NR	N = 114, m age = 47.53 yrs, m education = 13.48 yrs, m BMI = 30.17, Hispanic n = 11, non-Hispanic n = 76, unreported n = 11, gender & SES NR	Theoretical basis: terror management health model & prototype willingness model. Examined effect of ‘subtle thoughts of mortality’ & ‘socially informed images (prototypes)’	Factorial design. (1) Mortality salience (MS) induced; participants completed a true-false scale about either fear of pain or of death. (2) Prototype priming: participants led to visualise a healthy eater or a typical person. Conditions: MS x healthy n = 29, MS x typical n = 29, pain x healthy n = 31, pain x typical n = 25	Study period NR. One-off shop. No F/U. Fidelity: Research team intercepted all participants to administer intervention materials	Whole basket nutritional score (measured via receipts)	Considered effective. Interaction between MS & prototype; shoppers exposed to MS & healthy prototype purchased more nutritious food than customers in any other condition.
Milliman et al. (1982)	RCT	Store (W)	South-west city, USA	1 store; chain NR, customers described as middle-class, Anglo-Saxon	Whole store intervention. N for customers or transactions NR	No explicit theoretical basis. Examined effect of ‘in-store music’	Varied tempo of in-store music: no music, slow tempo, fast tempo	9 wks; each condition rotated daily, randomly across study period. No F/U. Fidelity: music maintained at constant volume, decibel metre used to ensure precision	Sales data across whole store	Considered effective. Highest sales seen in slow-tempo condition, which was significantly different from fast-tempo condition, but not from no-music condition. Fast-tempo and no-music conditions not significantly different
Milliron et al. (2012)	RCT	Individual (B)	Surprise, Phoenix, Arizona, USA	1 store of an Arizona-based chain; socio-economically diverse area	N = 153, F = 81%, m age = 41 yrs, white = 78%, average annual income < US$60,000 = 39%	No explicit theoretical basis. Examined effect of ‘environmental manipulations’ & ‘nutrition education’	Nutrition education. Nutritional shelf signs only (ctrl n = 83) or shelf signs plus 10-min face-to-face healthful nutrition education, delivered by nutrition educator at front of store (int n = 70). Educators in-store 6 h / day	4 mth study period, one-off shop. No F/U. Fidelity NR	Whole basket nutritional analysis: purchases of total fat, sat fat & trans-fat (grams / 1,000 kcal) F&V (servings / 1,000 kcal)	Considered partially effective. No difference seen between ctrl & int groups for total, sat, or trans-fats, or for servings of total veg. Int group purchased more fruit & dark green/bright yellow veg. than ctrl
Ni Mhurchu et al. (2010)	RCT	Individual (B)	Wellington, Wanganui & New Plymouth, New Zealand	8 stores in Foodstuffs (Wellington) Cooperative Society Ltd chain	N = 1,104, F = 86%, m age = 44 yrs, Ethnicity: Maori = 23%, Pacific = 9%, New Zealand European or other = 68%, average annual income < NZ$60,000 = 52%, >NZ$60,000 = 42%	No explicit theoretical basis. Examined effect of ‘nutrition education’ and ‘economic incentives’	12.5% price discount on eligible healthy foods (n = 275), tailored nutritional education, weekly mailings (n = 274), combined price discount & education (n = 277), ctrl, no int (n = 278)	12 wks B/L, 24 wks int, 24 wks F/U. Fidelity NR	Primary: change in % sat fat in food purchases at 6 mth. Secondary: change in % sat fat, total fat, protein, carbs, energy density, sodium & sugar in food purchases at 6 & 12 mth	Considered partially effective (discounts) & ineffective (education). No effect seen for any nutrients (sat fat, total fat, protein, carbs, energy density, sugars & sodium) from discounts or education. Those with discounts purchased more healthier discounted foods than those without discounts (+0.79 kg/wk, +11% from baseline). This effect was maintained at 24 wks follow-up (after discount removed). No difference seen for products in those who received education
Ogawa et al. (2011)	RCT	Store (B)	Gosen City & Tagami Town, Niigata Prefecture, Japan	2 stores (ctrl = 1, int = 1) from same chain	Whole store intervention. Average daily store visitors: ctrl n = 1,098, int n = 1,684	No explicit theoretical basis. Examined effect of ‘POP health information’	POP health & nutrition information: A4 posters for vegetables & A5 posters for fruits. Posters changed every 2 wks.	30 days pre-int, 60 days int (split equally into early & late int; reason NR). No F/U. Fidelity NR	Sales increases of F&V	Considered partially effective. Total veg. purchases in int store = +9.1% from B/L, compared to −1.2% in ctrl store. Increases in sales in 4 out of 6 categories of veg. in int store, but in only 2 veg. categories in ctrl store. No increases in fruit purchases in either store
Papies et al. (2014)	RCT	Individual (B)	The Netherlands	1 store, chain NR	N = 99, F = 94, m age = 54.18, m BMI = 24.73, lower education n = 58, higher education n = 38	Theoretical basis: hedonic food cues, priming, attentional resources. Examined effect of ‘subtle health primes in the environment’	In-store low kcal recipe flyers handed out: ctrl (n = 50) ‘new recipe’ flyer compared to health prime flyer (int n = 49), e.g., ‘good for your figure’	Study period = 5 days, 4 h each day. One-off shop. No F/U. Fidelity: Research team intercepted all participants to provide int or ctrl flyers	Purchases of unhealthy & high kcal snacks (cakes, cookies, sweets, chocolates & crisps)	Considered effective. No effect of health prime seen in normal-weight shoppers; amongst overweight shoppers, those who received health prime flyer purchased fewer unhealthy snacks than those who received ctrl flyer
Payne et al. (2015)	RCT	Store (B)	El Paso, Texas & Southern New Mexico, USA	4 stores (ctrl = 1, int = 3) from Pay & Save Inc. chain in areas with similar demographics, approx. 95% Hispanic, 53% F.	Whole store intervention. Collected N = 971,706 individual person transactions across study stores	Theoretical basis: social norms. Examined effect of ‘shopper marketing nutrition intervention’ & ‘social norms messaging’	Social norms messages placed in grocery carts: ‘in this store, most people choose at least X produce items’, graphics of most popular F&V, a smiley face & list of top 10 F&V purchased in that store	Stores 1 & 2 (pilot): 57 days B/L, 14 days int, Store 3: 78 days B/L & 28 days int, Store 4: 106 days B/L & 28 days int. No F/U. Fidelity NR	Total purchases of F&V, and proportion of F&V spending to total spend	Considered effective. F&V purchases in 3 int stores increased from B/L during int period: pilot store vs. control store +16% F&V spending. Store 3 = +7.5% F&V spending, Store 4 = +12.6% F&V spending. No impact on overall customer spending
Payne et al. (2016)	RCT	Store (B)	USA	4 stores (ctrl = 2, int = 2) from Pay ‘N’ Save; stores with different demographics & poverty levels	Whole store intervention. N for customers or transactions NR	Theoretical basis: social norms. Examined effect of ‘shopper marketing nutrition intervention’	10 large green arrows on shop floor; placed to direct shoppers’ attention to F&V section, e.g., ‘Follow green arrow for healthy [heart, weight]’	Pilot store: 14 days B/L & 14 days int, Store 3: 123 days B/L & 19 days int, Store 4: 96 days B/L & 25 days int. No F/U. Fidelity NR	Proportion of F&V spending to total spend	Considered effective. F&V purchases in 3 int stores increased from B/L during int period: store 1 = +8.0%, store 2 = +15.0%, store 3 = +4.6%. No impact on overall customer spending
Salmon et al. (2015)	RCT	Store (W) x Individual (B)	The Netherlands	1 store, chain NR	N = 127, F = 71.8%, m age = 50.4 yrs, education level: lower = 10.3%, average = 31.6%, higher = 58.1%, ethnicity NR	Theoretical basis: social proof heuristic, decision-making in low self-control contexts, ego-depletion. Examined effect of ‘social proof cues’ in conditions of ‘low self-control’	Factorial design: (1) Social proof heuristic (or none): banners next to low-fat cheese ‘most sold in this supermarket’ (banners present or absent), (2) Ego-depletion (or none): induced using a speech control task. Conditions: no heuristic x no depletion n = 41; no heuristic x depletion n = 26; heuristic present x no depletion n = 32; heuristic present x depletion n = 28	One-off shop per person. Study conducted over 4 days. No F/U. Fidelity: Research team intercepted all participants and administered intervention components	Purchases of low-fat cheese	Considered effective. Marginal interaction between ego-depletion & social proof heuristic: customers who were ego-depleted & saw social proof bought more low-fat cheese (instead of regular cheese) than customers in other 3 conditions, (but very wide CIs as only small number of customers actually bought low-fat cheese)
Schwartz et al. (2014)	RCT	Individual (B)	South Africa	Stores from Pick n Pay chain	N = 4,073 households using HealthyFood cashback bonus program (25% cashback on target items applied to store-specific credit card balance each month), demographic information NR	Theoretical basis: loss aversion. Examined effect of ‘commitment device with the potential of a loss’	Cashback program used to incentivise increased purchases of healthy food items compared to personalised B/L. Pre-commitment int (n = 632): participants agreed in advance to increase household spending on target healthy items by 5% each mth, over 6 mth; for each mth this increase was not met, cashback bonus was reversed (i.e., they forfeit bonus). No commitment int (n = 2,112): participants asked to consider the commitment hypothetically (but cashback not actually reversed). Ctrl (n = 1,329): no exposure	6 mth B/L, 6 mth int. No F/U. Fidelity: all participants completed initial online survey; int components embedded within survey. Ongoing intervention emails delivered monthly; details about whether emails were read NR	Changes in percentage of health items as proportion of whole shopping basket	Considered effective. Committed households purchased higher percentage of healthy items during 6 mth commitment period, compared to own baseline and to ctrl. Average healthy items purchased = +3.52%, average unhealthy items purchased = −1.30%, average neutral items purchased = −2.22%. No changes seen in non-commitment & ctrl households.
Vlaeminck et al. (2014)	RCT	Individual (B)	Belgium	1 store from Delhaize Group chain in suburban location; customers representative of urban, rural & country areas	N = 150, F = 54% (age, education & income reported by condition)	No explicit theoretical basis. Examined effect of making ‘information of environmental impact more accessible’	Varied labels by food items (apples, tomatoes & protein items) to present details of environmental impact. Standard label (ctrl n = 50), treatment least label, information not immediately coherent (n = 50), treatment most label, information presented clearly (n = 50)	Study period 9 days. No B/L or F/U as test area in store was created for study. Fidelity: implementation performed by research team	Eco-friendliness of target items purchases. Market share between target items	Considered partially effective. Average eco-friendliness / kg of products bought was higher in treatment-most condition compared to both treatment-least and ctrl. Treatment-least & control not different. For target products, customers more likely to purchase most eco-friendly option in 2 of 3 product categories in the treatment-most condition
Wansink et al. (1998) Study 1	RCT	Store (B)	USA	86 stores (ctrl = 43, int = 43), chain NR	Whole store intervention. N for customers or transactions NR	Theoretical basis: anchoring. Examined effect of ‘quantity anchors’	Varied presentation of price promotions by unit quantity: either a single unit discount (e.g., US$0.75) or multi-unit discount (e.g., 2 for US$1.50). Discounts varied across products between 9% & 44%, but were matched across units for products	All discounts in place for 1 week. B/L previous 6 mth. No F/U. Fidelity NR	Percentage increase in sales of target items: candy, breakfast cereal, cookies, frozen dinners, soft drinks, canned soup & canned tuna	Considered partially effective. Overall effect across all items: price discount increased sales by 125% in single-unit presentation and by 165% in multi-unit presentation, which was significant difference. Sales higher in multi-unit presentation for frozen dinners with 12% discount & canned tuna. No difference seen between different presentations for all other categories
Waterlander et al. (2013)	RCT	Individual (B)	The Netherlands	4 stores, 2 from PLUS chain & 2 from C1000 chain; across 4 provinces in rural areas	N = 199, F = 96%, m age = 51.7 yrs, education: primary = 3%, secondary = 40.5%, higher secondary = 7.7%, intermediate vocational = 48.8%, ethnicity: Dutch = 96.5%, Other = 3.5%, Income: low = 32.5%, medium = 11.9%, high = 29%, not answered = 26.6%	No explicit theoretical basis for overall intervention. Theoretical basis for education intervention: MI and TPB. Examined effect of ‘nutrition education’ and ‘pricing strategies’	50% price discount on F&V (n = 55), nutritional education, recipe books & telephone counselling, based on motivational interviewing (n = 49), combined price discount & education (n = 50), ctrl, no int (n = 45)	6 mth int, plus F/U at 3mth post-int. Fidelity: all discount coupons received by participants in line with protocol (redeem rate varied from 15% to 45%); all education participants received 2 planned recipe books, 54.5% received all 4 counselling calls, 16.2% 3 calls, 9.1% 2 calls, 6.1% 1 calls, 14.1% 0 calls	Increases in F&V purchases, in grams	Considered effective (ineffective for education alone). Purchases of F&V higher in discount group & combined discount & education group compared to ctrl & education-only group. Effect not maintained at 3 mth F/U
Winkler et al. (2016)	N-RCT	Store (B)	Bornholm & Odsherred, Denmark	28 stores (ctrl 1, Bornholm n = 12, ctrl 2 Odsherred n = 12, int Bornholm n = 4) from Coop chain	Whole store intervention. N for customers or transactions NR	Theoretical basis: social marketing principles. Examined effect of ‘product placement at POP’	Product changes at checkout; sugar confectionery replaced with fresh fruit, fruit bars, carrot sticks & dried fruit	B/L 4 wks, int 4 wks, F/U 7 wks. Fidelity: variation in implementation between stores	Purchases of target items	Considered ineffective. No difference seen between ctrl & int stores on sales of target products (sugar confectionary did not decrease & healthy snacks did not increase)

Note: Unless otherwise stated, all purchases were measured using store sales data (e.g., electronic scanner data).

Abbreviations: B = between; B/L = baseline; BMI = body mass index; carbs = carbohydrates; CIs = confidence intervals; ctrl = control; F = female; F/U = follow-up period; F&V = fruit and vegetables; int = intervention; kcal = calories; kg = kilograms m = mean; MI = motivational interviewing; min = minutes; mth = month; N = number in total sample; n = number in sub-group; N/A = not applicable; NR = not reported; N-RCT = non-randomised controlled trial; POP = point-of-purchase; RCT = randomised controlled trial; retro = retrospective; RTE = ready-to-eat; sat = saturated; SES = socioeconomic status; TPB = theory of planned behaviour; USA = United States of America; W = within; wks = weeks; yrs = years.

Table 3. Coding of Studies According to Behaviour Change Taxonomy v1 (BCTv1) by BCT groups.

Authors (Year)*	1.Goals & Planning	2.Feedback & Monitoring	3.Social Support	4.Shaping Knowledge	5.Natural Consequences	6.Comparison of Behaviour	7.Associations	8.Repetition & Substitution	9.Comparison of Outcomes	10.Reward & Threat	11.Regulation	12.Antecedents	13.Identity	14.Scheduled Consequences	15.Self-belief	16.Covert Learning
Effective
Bevelander et al. (2011)						6.1X
Dhar & Hoch (1996) Study 1							7.1XX			10.1XX 10.2XX		12.5XX
Dhar & Hoch (1996) Study 2							7.1XX			10.1XX 10.2XX		12.5XX
Heilman et al. (2002)							7.1X			10.1XX 10.2XX
McCabe et al. (2015)					5.2X	6.2X
Milliman et al. (1982)												12.1XX
Papies et al. (2014)				4.1X	5.1XX		7.1XX
Payne et al. (2015)						6.2XX 6.3X	7.1XX
Payne et al. (2016)						6.3X	7.1XX
Salmon et al. (2015)						6.2XX			9.1XX
Schwartz et al. (2014)	1.9X	2.2XX
Waterlander et al. (2013), Discounts										10.1XX 10.2XX
Waterlander et al. (2013), Discounts & Education	1.2XX 1.4X		3.1XX	4.1XX					9.1XX	10.1XX 10.2XX
Partially effective
Ball et al. (2015), Discounts										10.1XX 10.2XX
Ball et al. (2015), Discounts & Education	1.1XX 1.2XX 1.4X		3.1XX	4.1XX	5.1X				9.1XX	10.1XX 10.2XX
Cox (1964)												12.1XX
Cox (1970)												12.1XX
Foster et al. (2014)							7.1XX	8.1X		10.1XX 10.2XX		12.1XX
Gamburzew et al. (2016)				4.1X			7.1XX	8.1X				12.1XX
Geliebter et al. (2013)										10.1XX 10.2XX
Ni Mhurchu et al. (2010), Discounts										10.1XX 10.2XX
Milliron et al. (2012)				4.1XX	5.1X		7.1XX		9.1XX
Ogawa et al. (2011)				4.1X	5.1XX		7.1XX
Vlaeminck et al. (2014)					5.3XX		7.1XX
Wansink et al. (1998) Study 1										10.1XX 10.2XX
Ineffective
Achabal et al. (1987)				4.1XX			7.1X
Ball et al. (2015), Education	1.1XX 1.2XX 1.4X		3.1XX	4.1XX	5.1X				9.1XX
Jeffery et al. (1982)				4.1X			7.1XX		9.1XX
Ni Mhurchu et al. (2010), Education				4.1X
Ni Mhurchu et al. (2010), Discounts & Education				4.1X						10.1XX 10.2XX
Waterlander et al. (2013), Education	1.2XX 1.4X		3.1XX	4.1XX					9.1XX
Winkler et al. (2016)												12.1XX

Note: *Studies grouped by intervention effectiveness. 1.1 = Goal setting (behaviour); 1.2 = Problem solving; 1.4 = Action planning; 1.9 = Commitment; 2.2 = Feedback on behaviour; 3.1 = Social support (unspecified); 4.1 = Instruction on how to perform the behaviour; 5.1 = Information about health consequences; 5.2 = Salience of consequences; 5.3 = Information about social and environmental consequences; 6.1 = Demonstration of the behaviour; 6.2 = Social comparison; 6.3 = Information about others’ approval; 7.1 = Prompts/cues; 8.1 = Behavioural practice/rehearsal; 9.1 = Credible source; 10.1 = Material incentive (behaviour); 10.2 = Material reward (behaviour); 12.1 = Restructuring the physical environment; 12.5 = Adding objects to the environment. (XX) = clearly present. (X) = most likely present.

Figure 3. Frequency of behaviour change techniques identified across interventions, by intervention effectiveness.

Table 4. Coding of Studies According to Intervention Types within the Typology of Interventions in Proximal Physical Microenvironments (TIPPME).

Authors (Year)*	Availability	Position	Functionality	Presentation	Size	Information
Effective
Dhar & Hoch (1996) Study 1	XX
Dhar & Hoch (1996) Study 2	XX
Heilman et al. (2002)						XX
Milliman et al. (1982)				XX
Papies et al. (2014)						XX
Payne et al. (2015)						XX
Payne et al. (2016)						XX
Salmon et al. (2015)						XX
Partially effective
Cox (1964)	XX
Cox (1970)	XX
Foster et al. (2014)	XX	XX		X		XX
Gamburzew et al. (2016)				X		XX
Milliron et al. (2012)						XX
Ogawa et al. (2011)						XX
Vlaeminck et al. (2014)						XX
Wansink et al. (1998) Study 1						XX
Ineffective
Achabal et al. (1987)						XX
Jeffery et al. (1982)						XX
Winkler et al. (2016)		XX

Note: *Studies grouped by intervention effectiveness. (XX) = clearly present. (X) = most likely present.

Table 5. Coding of Studies according to the Intervention Functions of the Behaviour Change Wheel (BCW).

Authors (Year)*	Education	Persuasion	Incentivisation	Coercion	Training	Restriction	Environmental Restructuring	Modelling	Enablement
Effective
Bevelander et al. (2011)								XX
Dhar & Hoch (1996) Study 1			XX				XX
Dhar & Hoch (1996) Study 2			XX				XX
Heilman et al. (2002)			XX
McCabe et al. (2015)		XX		XX
Milliman et al. (1982)							XX
Papies et al. (2014)	XX
Payne et al. (2015)		XX					XX
Payne et al. (2016)		XX					XX
Salmon et al. (2015)		XX					XX
Schwartz et al. (2014)				XX
Waterlander et al. (2013), Discounts			XX
Waterlander et al. (2013), Discounts & Education	XX	XX	XX
Partially effective
Ball et al. (2015), Discounts			XX
Ball et al. (2015), Discounts & Education	XX		XX		XX
Cox (1964)							XX
Cox (1970)							XX
Foster et al. (2014)			XX				XX
Gamburzew et al. (2016)	XX						XX
Geliebter et al. (2013)			XX
Milliron et al. (2012)	XX				XX	XX
Ni Mhurchu et al. (2010), Discounts			XX
Ogawa et al. (2011)	XX					XX
Vlaeminck et al. (2014)	XX						XX
Wansink et al. (1998) Study 1			XX
Ineffective
Achabal et al. (1987)					XX		XX
Ball et al. (2015), Education	XX				XX
Jeffery et al. (1982)	XX						XX
Ni Mhurchu et al. (2010), Education	XX
Ni Mhurchu et al. (2010), Discounts & Education	XX		XX
Waterlander et al. (2013), Education	XX	XX
Winkler et al. (2016)							XX

Note: *Studies grouped by intervention effectiveness. (XX) = clearly present. (X) = most likely present.

Table 6. Theoretical Coherence between Coded Behaviour Change Techniques (BCTs) and Coded Intervention Functions (IFs) of the Behaviour Change Wheel (BCW).

Authors (Year)*	Theoretical basis for intervention	Coherent behaviour change techniques & intervention functions	Non-coherent behaviour change techniques	Non-coherent intervention functions	Outcomes with observed intervention effect	Outcomes with no observed intervention effect
Effective
Bevelander et al. (2011)	Social modelling	6.1 = Modelling	–	–	Kcal purchased highest when confederate purchased high kcal items	–
Dhar & Hoch (1996) Study 1	Price theory, push promotions, pull promotions, price perception, attitudes towards coupons, conditioning, novelty, opportunity costs of time	7.1 & 12.5 = Environmental Restructuring, 10.1 & 10.2 = Incentivisation	–	–	RTE cereal and soft drink purchases increased when price discounted	–
Dhar & Hoch (1996) Study 2	Price theory, push promotions, pull promotions, price perception, attitudes towards coupons, conditioning, novelty, opportunity costs of time	7.1 & 12.5 = Environmental Restructuring, 10.1 & 10.2 = Incentivisation	–	–	Target RTE cereal purchases increased when price promoted; rest of category purchases decreased	–
Heilman et al. (2002)	Psychological income effect, elevated mood from unanticipated gains, priming effect, memory effect	10.1 & 10.2 = Incentivisation	7.1 = Environmental Restructuring	–	Total basket spend, unplanned purchases (any category), unplanned treat purchases, unplanned purchases of products cognitively-related to coupon product, and unplanned purchases of products in proximity to coupon-products were all greater amongst those who received surprise discount coupon for either cereal, spaghetti sauce, laundry detergent, or paper towels	–
McCabe et al. (2015)	Terror management health model, prototype willingness model	5.2 & 6.2 = Persuasion	–	No BCTs could be mapped to Coercion	MS x prototype interaction. Healthier rating of whole basket purchases for those exposed to MS and healthy eater prototype	–
Milliman et al. (1982)	No explicit theoretical basis	12.1 = Environmental Restructuring	–	–	Higher gross sales during slow tempo music compared to fast tempo music	–
Papies et al. (2014)	Hedonic food cues, priming, attentional resources	5.1 = Education	4.1 = Training, 7.1 = Environmental Restructuring	–	Prime x Weight interaction. Purchases of unhealthy snacks were fewer amongst overweight shoppers exposed to health prime	–
Payne et al. (2015)	Social norms	6.2 & 6.3 = Persuasion, 7.1 = Environmental Restructuring	–	–	Spending on produce (F&V) items increased in presence of social norms messages in grocery carts	–
Payne et al. (2016)	Social norms	6.3 = Persuasion, 7.1 = Environmental Restructuring	–	–	Spending on produce (F&V) items increased in presence of floor arrows with health messages directing to produce areas	–
Salmon et al. (2015)	Social proof heuristic, decision-making in low self-control context, ego-depletion	6.2 & 9.1 = Persuasion	–	No BCTs could be mapped to Environmental Restructuring	Ego-depletion x social proof heuristic interaction. Purchases of low-fat cheese greater in those who experienced ego-depletion and exposure to social proof	–
Schwartz et al. (2014)	Loss aversion	1.9 & 2.2 = Coercion	–	–	Purchases of healthy items increased, and neutral and unhealthy items decreased, in those who made a pre-commitment to increase purchases of healthy items by 5% (compared to their own B/L) or lose a cashback bonus	–
Waterlander et al. (2013), Discounts	No explicit theoretical basis	10.1 & 10.2 = Incentivisation	–	–	Purchases of F&V greater by weight with price discounts compared to ctrl	–
Waterlander et al. (2013), Discounts & Education	No explicit theoretical basis for Discounts element. Motivational interviewing and theory of planned behaviour used in education element	9.1 = Persuasion, 10.1 & 10.2 = Incentivisation	1.2, 1.4 & 3.1 = Enablement, 4.1 = Training	–	Purchases of F&V greater by weight with nutrition education & price discounts compared to ctrl	–
Partially effective
Ball et al. (2015), Discounts	Social ecological and social cognitive theories	10.1 & 10.2 = Incentivisation	–	–	Purchases of total vegetables, frozen vegetables, total fruit, and fresh fruit greater in those who received price discounts	No differences in purchases of fresh, dried, or canned vegetables. No differences in purchases of frozen, dried, or canned fruit. No differences in bottled water, sugar-sweetened beverages, or diet beverages
Ball et al. (2015), Discounts & Education	Social ecological and social cognitive theories	4.1 = Training, 5.1 = Education, 10.1 & 10.2 = Incentivisation	1.1, 1.2, 1.4, 3.1 = Enablement, 9.1 = Persuasion	–	Purchases of total fruit and fresh fruit greater in those who received price discounts and nutrition education	No differences in purchases of total, frozen, fresh, dried, or canned vegetables. No differences in purchases of frozen, dried, or canned fruit. No differences in bottled water, sugar-sweetened beverages, or diet beverages
Cox (1964)	No explicit theoretical basis	12.1 = Environmental Restructuring	–	–	Purchases of hominy greater when shelf space increased	No difference in purchases of powdered coffee cream, tang, or baking soda when shelf space increased
Cox (1970)	No explicit theoretical basis	12.1 = Environmental Restructuring	–	–	Purchases of Coffeemate (high customer acceptance brand) greater when shelf space increased	No difference in purchases for Creamora (low customer acceptance brand) or for high acceptance or low acceptance salt brands
Foster et al. (2014)	4 Ps framework (price, promotion, product, placement)	7.1 & 12.1 = Environmental Restructuring, 10.1 & 10.2 = Incentivisation	8.1 = Training	–	Purchases of 1% milk, skim milk, meat ready meals (chicken nuggets & turkey dinner), and bottled water were greater in stores with changes made to product facings, placement, & promotional activity	No difference in purchases of whole milk, 2% milk, cheerios cereal, honeycomb cereal, meat ready meal (steak), diet Pepsi, regular Pepsi, regular beverages, or low calorie beverages when changes made to product facings, placement, & promotional activity
Gamburzew et al. (2016)	No explicit theoretical basis	7.1 & 12.1 = Environmental Restructuring	4.1 & 8.1 = Training	No BCTs could be mapped to Education	Purchases of 2 categories (F&V and Starches) greater in intervention stores	No difference in purchases of 4 categories (Meat/fish/eggs, Mixed dishes & sandwiches, Dairy products, Other)
Geliebter et al. (2013)	No explicit theoretical basis	10.1 & 10.2 = Incentivisation	–	–	Purchases of F&V higher when price discounted	No differences in purchases for bottled water or diet sodas when price discounted
Milliron et al. (2012)	No explicit theoretical basis	4.1 = Training, 5.1 = Education,	7.1 = Environmental Restructuring., 9.1 = Persuasion	No BCTs could be mapped to Restriction (BCW book indicates there are no BCTs in BCTTv1 that are linked to this IF)	Purchases (per 1000 kcal) of fruit and dark-green/yellow vegetables higher in those who received nutrition education	No differences in purchases (per 1000 kcal) of total fat, saturated fat, trans-fat, or total vegetables in those who received nutrition education
Ni Mhurchu et al. (2010), Discounts	No explicit theoretical basis	10.1 & 10.2 = Incentivisation	–	–	Purchases of all foods, all healthier products, healthier cereals, healthier F&V, healthier meat & alternatives, and healthier milk & milk products in those who received price discounts compared to those who did not	No differences in purchases of all less-healthy products and healthier fats & oils in those who received price discounts. No differences (as % energy) in sat fat, total fat, protein, carbohydrate, energy density, sugars, or sodium in those who received price discounts
Ogawa et al. (2011)	No explicit theoretical basis	5.1 = Education	4.1 = Training, 7.1 = Environmental Restructuring	No BCTs could be mapped to Restriction (BCW book indicates there are no BCTs in BCTTv1 that are linked to this IF)	Purchases of all vegetables, green leafy vegetables, fruit & flower vegetables, F&V juice, and other vegetables increased when nutritional POP material present	No differences in purchases of leafy vegetables, root vegetables, and fruits when nutritional POP material present
Vlaeminck et al. (2014)	No explicit theoretical basis	5.3 = Education, 7.1 = Environmental Restructuring	–	–	Purchases of most eco-friendly tomatoes and most eco-friendly protein options greater when environmental label present	No difference in purchases of apples when environmental label present
Wansink et al. (1998) Study 1	Anchoring	10.1 & 10.2 = Incentivisation	–	–	Purchases of cookies, 2 types of frozen dinners, frozen entrees, 1 type of canned soup, soft drinks, and canned tuna greater when price promotion expressed in multiple units (instead of single units)	No differences in purchases of candy, breakfast cereal, or 1 type of canned soup when price promotion expressed in multiple units
Ineffective
Achabal et al. (1987)	No explicit theoretical basis	4.1 = Training. 7.1 = Environmental Restructuring	–	–	–	Purchases of broccoli, cabbage, carrots, cauliflower, kiwi fruit, or tomatoes did not increase in presence of nutrition POP information
Ball et al. (2015), Education	Social ecological and social cognitive theories	4.1 = Training, 5.1 = Education,	1.1, 1.2, 1.4 & 3.1 = Enablement, 9.1 = Persuasion	–	–	No differences in purchases across 10 categories of F&V, in bottled water, in sugar-sweetened beverages, or in diet beverages.
Jeffery et al. (1982)	No explicit theoretical basis	7.1 = Environmental Restructuring	4.1 = Training, 9.1 = Persuasion	No BCTs could be mapped to Education	–	No difference in sales for 25 items within dairy and frozen dessert items
Ni Mhurchu et al. (2010), Education	No explicit theoretical basis	–	4.1 = Training	No BCTs could be mapped to Education	–	No differences in purchases for all foods, all less-healthy products, all healthier products, healthier cereals, healthier fats & oils, healthier F&V, healthier meat & alternatives, or healthier milk & milk products in those who received tailored nutrition education compared to those who did not. No differences (as % energy) in sat fat, total fat, protein, carbohydrate, energy density, sugars, or sodium in those who received tailored nutrition education
Ni Mhurchu et al. (2010), Discounts & Education	No explicit theoretical basis	10.1 & 10.2 = Incentivisation	4.1 = Training	No BCTs could be mapped to Education	–	No differences in purchases for sat fat. Inferential statistics not reported for other outcomes for this group; apparently no effect, but not clear if lack of observed differences is meaningful
Waterlander et al. (2013), Education	Motivational interviewing and theory of planned behaviour	9.1 = Persuasion	1.2, 1.4 & 3.1 = Enablement, 4.1 = Training	–	–	No difference by weight in purchases of F&V between ctrl and nutrition education groups
Winkler et al. (2016)	Social marketing principles	12.1 = Environmental Restructuring	–	–	–	No differences in purchases of sugar confectionary, fresh fruit, fruit bars, carrot snack packs, or dried fruit when snack options at checkout displays were altered

Note: *Studies grouped by intervention effectiveness. 1.1 = Goal setting (behaviour); 1.2 = Problem solving; 1.4 = Action planning; 1.9 = Commitment; 2.2 = Feedback on behaviour; 3.1 = Social support (unspecified); 4.1 = Instruction on how to perform the behaviour; 5.1 = Information about health consequences; 5.2 = Salience of consequences; 5.3 = Information about social and environmental consequences; 6.1 = Demonstration of the behaviour; 6.2 = Social comparison; 6.3 = Information about others’ approval; 7.1 = Prompts/cues; 8.1 = Behavioural practice/rehearsal; 9.1 = Credible source; 10.1 = Material incentive (behaviour); 10.2 = Material reward (behaviour); 12.1 = Restructuring the physical environment; 12.5 = Adding objects to the environment; B/L = baseline; ctrl = control; F&V = fruit and vegetables; kcal = calories; MS = mortality salience; POP = point-of-purchase; RTE = ready-to-eat; sat = saturated.

Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & The PRISMA Group, X. (2009). Preferred Reporting items for systematic Reviews and Meta-Analyses: The PRISMA statement. PLoS Medicine, 6(7), e1000097. https://doi.org/https://doi.org/10.1371/journal.pmed1000097

 PubMed Web of Science ®Google Scholar

Data availability

The raw data files for quality assessments and for coding according to the behavioural science frameworks are available as supplemental materials.