Utilising DV2 (error scores) as the dependent variable

Utilising the error scores as the dependent variable in the same analytical approach as above revealed further insight into how participants rated the various foods at the differing levels of healthiness. Results are shown in Table 5.4 for those cases where they differed from DV1. Of the within subject factors, by far the largest effect was seen for the food categories (Food) indicating that the distance of participants’ ratings from the objective scores (i.e. error scores for each food), varied across the different food categories regardless of which label format they were shown.

In addition, utilising DV2 a larger effect size was observed in the two-way interaction Food*Healthiness than was observed for DV1. This demonstrates that the degree of healthiness of the foods influenced the distance of participants’ ratings from the objective score (Fig. 5.5), with the healthiness ratings for the low health variant in each category being further from the objective score than for the high health variant, although the extent of this varied across the different food categories (Table 5.3).

Overall, participants tended to underestimate the healthiness of the pizzas and yoghurts and overestimate the healthiness of the biscuits with their subjective healthiness ratings. However, further exploration of the three-way interaction Food*FOP*System demonstrated that when the FOP systems were present, the overestimation of the healthiness of the biscuits appeared to be slightly reduced which is a promising outcome (Fig. 5.6). The underestimation of the healthiness of the pizzas and yoghurts were also reduced.

Figure 5.5 Food*Healthiness interaction utilising DV2 (Mean error scores). F¹(3.7, 7542.3) = 106.54, p ≤ .001, ^ɳp2 = .391

Figure 5.6 Food*FOP*System interaction utilising DV2 (Mean error scores). F¹(6, 4047.3) = 16.20, p ≤.001, ^ɳp2 = .02

5.4 Discussion

The results of this study suggest that although the FOP systems tested can result in some small improvements to objective understanding across different foods, portion sizes and levels of healthiness, there was little difference observed over the provision of an FOP label containing basic numerical nutritional information alone or between the various systems under these conditions. Therefore the energy information as an FOP label may be sufficient to enable consumers to detect the healthier alternative within a food category should they wish to do so, or perhaps are forced to do so within an experimental environment. However, it should be noted that in this study participants were making their decisions of healthiness between the foods within one FOP labelling system. In reality, the presence of multiple FOP systems in the marketplace would make the task of comparing foods more difficult.

The results of this study are in line with those of previous research (Grunert

& Fernández-Celemin, 2010; Malam et al., 2009), which found that the vast majority of people can successfully identify healthier products using any of the prominent labelling formats. The novel aspect of our research is the direct systematic comparison of the FOP systems using the same food categories with foods at differing levels of health across different countries, the comparison of these to the provision of numerical nutrition information alone as the FOP label and comparison to an objective health score.

Despite only testing three food categories in this study, results suggest that people may rate different food categories differently. However, the tendency of the participants to underestimate the healthiness of pizza and yogurt and to overestimate the healthiness of biscuits in their subjective healthiness ratings may in fact be affected more by the portion size/reference amount (portion effect) in which the food is being presented, rather than any pre-conceived ideas about the relative healthiness of the different food categories. For example, in comparison to the pizza presented as either 200g or 100g, the biscuit was relatively small at 18g or 9g. The

numerical values for energy and the four nutrients on the label for biscuits would therefore have been small in magnitude and potentially perceived as being healthier than they objectively are. It is interesting to note that when the FOP systems were present there appeared to be some impact, albeit small, on this possible portion effect in the biscuits which is a promising outcome. Further research exploring the effect of reference amounts utilised in FOP nutrition labelling and their influence on evaluation of product healthfulness have reinforced the finding that pre-conceived ideas about the relative healthiness of the different food categories do not appear to have had a strong influence on healthiness ratings. However, consumers do appear to factor the reference amount, that is, the quantity of food for which the nutritional information is being presented, into their judgements of healthfulness. (Raats et al.

2015).

5.5 Conclusions

Under experimental conditions, results suggest that any structured and legible presentation of key nutrient and energy information on the FOP may be sufficient to enable consumers to detect a healthier alternative within a food category when provided with foods with distinctly different levels of healthiness.

However, in real life settings, personal factors and context must also be considered (Barker, Lawrence & Robinson, 2012) and often take precedence over health considerations in driving choice. Whilst FOP labels have the potential to facilitate healthier choices, in reality they can only do so when the motivation and intention to shop more healthily has been established.

When considering the implications of these results on future FOP labelling policy, one must bear in mind that although basic nutritional information alone might be sufficient to enable consumers to detect the healthier alternative from a limited choice set when specifically asked to do so under experimental conditions, in the real world it may not be. It is known that most consumers do not have the motivation or the time to process nutritional information when they are shopping (Gerrior, 2010;

van Herpen & van Trijp, 2011). The additional elements of Traffic Light colour, Guideline Daily Amounts or the presence of a health logo may have a greater impact in engaging consumers with the nutritional implications of their food. Overall the results from this study suggest that one may be more likely to identify the optimal FOP labelling scheme in real world studies than in studies which involve

forced exposure under experimental conditions. Future research should therefore focus on a given FOP scheme’s potential to engage consumers’ attention, motivate them towards healthier food choice and effect behavioural change within real world shopping environments.

CHAPTER 6

GENERAL DISCUSSION

6.1 Overview.

Food composition data, front-of-pack nutrition labelling and nutrition and health claims are fundamental to the development of appropriate policy, regulation and public health interventions (e.g. mandatory nutrition labelling) aimed at reducing the burden of diet-related chronic disease. The overarching aim of this thesis was to explore whether the communication of healthier food choice, through the provision of FOP nutrition labelling and nutrition and health claims, can be enhanced by the development of consumer derived frameworks of these domains, a greater understanding of the degree to which the different FOP labelling schemes impact on consumer health inferences and an improved approach to the sharing of food composition data between stakeholders.

The findings of the studies within this thesis are summarised and discussed below in terms of their contributions to theory, methodology and practice.

6.2 Improved approach to the sharing of food composition data