Table 3 Likelihood of selecting a product with a health-related claim

Ares (2010) [27]

Part-worth utilities: multinomial logit regression.

Cluster analysis. Cluster 1 more diet and health concerned than cluster 2.

Yes

Yogurt.

RDR (fibre - cancer), (antioxidants - heart disease + cancer).

C1: 1.28 (1.06, 1.56)

Y

C2: 1.38 (1.11, 1.71)

Aschemann-Witzel [28](2010)

Proportion of products that carry claims & chosen.

OR calculated from the number of choices of a product with a claim and the number of expected choices of products with a claim, if the claim was chosen at random.

No

Yogurt, muesli, pasta.

HCs: NOF (calcium + vitamin D - bones/teeth), (folic acid - mental function), (fibre - bowel function), RDR (calcium + vitamin D - osteoporosis), (folic acid - dementia), (fibre - cancer).

1.21 (0.98, 1.43)

Y

Aschemann-Witzel (2013) [29]

Percentage products with claim chosen (number of choices) .

As Aschemann-Witzel (2010).

No

Yogurt, breakfast cereal, pasta.

NCs: Ncon (calcium, vitamin D), NOF (calcium, vitamin D - osteoporosis, Ncon (folic acid), NOF (folic acid - brain/mental functions).

1.10 (0.87, 1.32)

Y

Barreiro-Hurle (2010) [30]

Coefficient: random parameter logit.

Same population tested both products.

No

Pork Frankfurter sausage.

Ncon (fat)

1.67 (1.48, 1.87)

Y

RDR (CVD)

1.97 (1.74, 2.24)

Ncon (fat) & RDR (CVD)

0.58 (0.49, 0.69)

Yogurt.

RDR (CVD)

1.25 (1.21, 1.28)

Casini (2014) [33]

Parameter estimates: conditional logit model.

Average of NOF and RDR used for meta-analysis as same products and population.

No

Olive oil.

NOF (polyphenols - oxidative stress)

1.44, (1.29, 1.60)*

Y + N

RDR (polyphenols - cholesterol)

1.23 (1.09, 1.39)*

Average

1.33 (1.19, 1.49)

HRIC (polyphenol)

0.88 (0.77, 1.01)

Contini (2015) [35]

Parameter estimates: latent class model.

Average used as same product, population, and claim-sub type

No

Olive oil.

RDR (polyphenols - blood lipids)

1.41 (1.27, 1.57)*

Y

RDR (olive oil - CHD)

1.66 (1.54, 1.80)*

RDR (olive oil - blood lipids)

1.70 (1.52, 1.89)*

Average

1.58 (1.44, 1.75)

De Marchi (2016) [45]

Parameter estimates: random parameter logit with an error component.

Results adjusted for time preferences. Interaction terms not included.

No

Yogurt.

RDR (saturated fat & cholesterol - heart disease)

1.25 (1.13, 1.38)

Y

de-Magistris (2016) [36]

Parameter estimates: Random Parameters Logit model.

.

No.

Cheese.

Ncom (fat)

2.23 (0.01, 570.66)

Ncom (sodium)

0.56 (39.38, 0.01)

Fernández-Polanco (2013) [37]

Coefficients: conditional logit model, (+WTP).

Was not included in meta-analyses as SE not reported.

No

Fish (seabream).

Ncon (omega-3)

1.63

Y

Gracia (2009) [38]

Coefficient: Parameters Logit model.

Interaction terms not included.

No

Breakfast cookies.

Ncon

1.46 (1.21, 1.75)

Y

Krystallis (2012) [42]

Coefficient (+WTP).

Averaged by product as same population and same claim sub-types.

No

Crisps.

Ncon (calcium)

2.31 (1.07, 5.00)*

Y + N

Ncon (vitamins)

1.86 (0.92, 3.75)*

Ncon (omega-3 fatty acids)

0.77 (0.39, 1.49)*

Ncon (fibres)

1.54 (0.75, 3.18)*

Average

1.50 (0.73, 3.07)

Croissant.

Ncon (calcium)

1.31 (0.85, 2.00)*

Ncon (vitamins)

1.44 (0.99, 2.09)*

Ncon (omega-3 fatty acids)

0.74 (0.51, 1.07)*

Ncon (fibres)

0.83 (0.56, 1.22)*

Average

1.04 (0.70, 1.53)

Loose (2013) [44]

Aggregated multinomial logit model/part worth utility estimates

No

Oysters.

Logo (Heart tick)

1.11 (1.08, 1.13)

Y

Maubach (2014) [46]

Hazard ratio: binominial logit regression.

OR for health claim scaled by OR for ‘no claim’. Results for NC not included.

Interactions not included

Yes

Breakfast cereal.

RDR (wholegrain - cholesterol)

1.17 (1.13, 1.22)

Y

McLean (2012) [47]

Utility: multinomial logit regression (no FoP label model used)

600 participants 300 with hypertension,

300 without.

Recruited from same database

Averaged by product as same population and same claim sub type.

Yes

Low sodium Baked beans.

Ncom (sodium)

0.66 (0.53, 0.83)*

Y + N

Ncon (sodium)

1.55 (1.22, 1.97)*

Average

1.11 (0.88, 1.45)

High sodium baked beans.

Ncom (sodium)

1.21 (0.94, 1.55)

Mohebalian (2012) [48]

Odds ratio

Cluster analysis

No

Fruit juice.

NOF (antioxidants - immune system).

C1: 1.98 (1.51, 2.59)

Y

C2: 1.63 (1.18, 2.24)

C3: 2.50 (2.2.4, 2.79)

C4: 1.72 (1.44, 2.06)

Mohebalian (2013) [49]

Coefficient

Different populations for each product. Results adjusted for demographic characteristics.

Yes

Elderberry jelly.

NOF (antioxidants - immune system).

1.96 (1.52, 2.52)

Y

Elderberry juice.

NOF (antioxidants - immune system).

1.71 (1.31, 2.25)

Moon (2011) [50]

Rating: willingness to try 5 point scale

Different populations for each claim.

No

Soy foods.

RDR (protein - heart disease) FDA approved.

1.48 (1.32, 1.66)

Y

RDR (protein - heart disease) Not FDA approved.

1.52 (1.35, 1.71)

Van Wezemael (2014) [54]

Parameter estimates: error component model (best fit) (+multinomial logit model, +WTP)

Average by claim sub-type, same product, results by country.

No

Lean beef steak.

Ncon (iron), Ncon (fat), Ncon (protein).

NL: 3.42 (2.50, 4.69)

Y

BE: 2.98 (2.17, 4.10)

FR: 3.61 (2.61, 5.00)

UK: 3.06 (2.18, 4.30)

NOF (iron - cognitive function), NOF (saturated fat - cholesterol), NOF (protein - muscle).

NL: 5.68 (4.06, 7.95)

BE: 4.08 (3.00, 5.56)

FR: 4.60 (3.38, 6.26)

UK: 3.06 (2.37, 3.97)