What nutrition label to use in a catering environment? A discrete choice experiment

Abstract

Worksite and university canteens are increasingly used for daily main meal consumption. Following the use of front-of-pack nutrition labels on pre-packed foods, the provision of easily accessible nutrition information on foods prepared and consumed out of home is a highly topical policy issue with the potential to help consumers making better informed and more healthy food choices when eating out. Information that is presented in a format preferred by the target group is more likely to be used. A sample of 1725 university canteen users participated in a web-based choice experiment designed to identify and understand individual preferences for alternative nutrition labels on canteen meals. Participants valued the presence of nutrition labels on canteen meals and showed a preference for more detailed formats. Ability and motivation to process information as well as socio-demographics explained differences in label preferences. Observed decreasing marginal utility from combinations of two simple label formats as well as from combinations of two detailed formats, signalled information insufficiency versus information overload, respectively. In order to satisfy most canteen users’ information needs, a nutrition label that contains basic Guideline Daily Amount (GDA)-type of numerical information in combination with familiar visual aids like stars and colour codes is proposed to be used in university canteens.

Introduction

Consumers’ increased demand for information about the health characteristics of foods has motivated food manufacturers and retailers to provide nutrition information on their food labels (Verbeke, 2005). Recently, different front-of-pack (FOP) nutrition labels have been introduced voluntarily as a complementary scheme to the European Union (EU) regulated back-of-pack (BOP) nutrition table (EC, 1990), and nutrition and health claims on pre-packed foods (EC, 2007). Examples are the Guideline Daily Amount (GDA) and Traffic Light label (TL). These FOP nutrition labels differ from the traditional nutrition tables in the amount and presentation of the nutrition information and therefore, would require less time and effort to be processed. From a public health and food policy point of view, providing consumers with summarised nutrition information at the point of purchase may help them to easily identify and (hopefully) choose the healthier foods. Marketers, for their part, have an interest in nutrition labelling to develop strategies to better differentiate their food products in the market and to build and maintain nutrition- and health-related competitive advantages. Both from a public and private perspective it is important to identify and understand consumers’ preferences for alternative nutrition information formats (Grunert and Wills, 2007).

Due to rising overweight and obesity rates (WHO, 2011) and the increased reliance on foods away from home (FAFH) (Guthrie et al., 2002, Orfanos et al., 2007), there is ongoing debate on whether to adopt mandatory nutrition labelling in the FAFH sector. FAFH have been blamed for hindering the potential beneficial effects associated with label use due to the increased substitution of at-home consumption of pre-packed foods, which the EU regulation supports, with away-from-home consumption (Drichoutis et al., 2006). Frequent out-of-home consumption has been associated with higher energy intakes, and a higher prevalence of overweight and obesity (Ayala et al., 2008, Orfanos et al., 2007, Vandevijvere et al., 2009). Additionally, most consumers seem to underestimate the nutrient content of FAFH (Burton et al., 2006). Given the possible mismatch between the perceived and actual nutritional value of FAFH, the inclusion of nutrition information on the menu could benefit consumers by effectively transforming the nutrient content, a typical credence attribute, into a search attribute (Caswell and Mojduszka, 1996), and herewith reduce uncertainty and information asymmetry for FAFH choices. Moreover, consumers demand high quality foods, apparently with as much information as possible, but at the same time they often experience time constraints and a lack of motivation and skills to process the information (Verbeke, 2005). This suggests that FOP nutrition labelling is a potentially interesting policy tool also for FAFH. Other cues or indicators of quality that are frequently used in situations of time pressure, are the price and brand name of the food product (Gracia et al., 2009). In the context of out-of-home eating, brand names might not be important, but the price is (Hwang and Lorenzen, 2008) and will, therefore, be included in the choice experiment to allow a monetary valuation of the food information attributes.

Previous consumer research on nutrition labelling has mainly focused on the use and understanding of on-pack nutrition information and its effect on consumer decision-making (reviewed by Campos et al., 2011, Cowburn and Stockley, 2005, Drichoutis et al., 2006, Grunert and Wills, 2007). It was concluded that most consumers are able to find and use simple numerical information for making simple comparisons between products for consumption at home, but their ability to interpret nutrition labels decreases as the complexity of the tasks increases. Therefore, it was suggested to add interpretational aids like verbal descriptors (high, medium, low) and recommended reference values to assist consumers in making more informed food choices. Since the introduction of the traditional BOP nutrition tables, several stakeholders (food manufacturers and retailers, governmental and non-governmental organisations) have been focussing on these interpretational aids resulting in a wide range of additional FOP nutrition information labels. Different classifications of FOP nutrition labels have been suggested in literature (Bonsmann et al., 2010a, Feunekes et al., 2008, Grunert and Wills, 2007) based on their level of simplicity (simple versus complex), comprehensiveness (basic versus detailed) and/or coerciveness (non-directive, semi-directive, directive). An overview of FOP labels can be found in EHN (2007). To date the question remains in which format FOP nutrition information should preferably be made available to consumers (Möser et al., 2010).

The presence of multiple nutrition labels on food packages is likely to further confuse consumers (van Trijp, 2009). Nutrition labels have been found to be more valued by consumers when presented individually than when appearing together (Barreiro-Hurlé et al., 2009, Barreiro-Hurlé et al., 2010). Consumers are clearly facing increasing nutrition information on pack which may eventually yield information overload. Although consumers say they want more information, being faced with too much information, and limited time and motivation to process information may simply cause them to opt out of the nutrition information search process in order to protect themselves from information overload and resulting uncertainty (Caswell and Padberg, 1992).

A few studies have been examining differences in consumer preferences for FOP labelling. Feunekes et al. (2008) found that simple symbols were more appropriate in a shopping environment compared to the more detailed labels due to the lower processing time needed, while differences in perceived consumer friendliness and usage intention between both formats were minor. Best performing FOP labelling formats according to the study by UK Food Standards Agency (FSA) are Multiple Traffic Lights (MTLs) and Colour-coded Guideline Daily Amounts (CGDAs) compared to Simple Traffic Lights and Monochrome Guideline Daily Amounts (FSA, 2005). MTL performed best in the individual product evaluation, while CGDA performed better when comparing two products. A recent study by the British Market Research Bureau for FSA concluded that the strongest FOP labels are those which combine text (high, medium, low), Traffic Light colours and % GDA information (Malam et al., 2009). Regarding labelling characteristics (such as display size, position of the label on FOP, colour scheme), research by Bialkova and van Trijp (2010) found that the presence rather than absence of a nutrition label printed on a consistent location on the package, with a doubled display size and with mono-rather than polychromatic colouring, were key success factors for attracting consumers’ attention. A choice experiment assessing consumers’ preferences regarding nutrition labels revealed that the easy-to-use format may benefit more shoppers than the detailed format (Berning et al., 2007). These findings are important for the outlook of nutrition labels in the FAFH sector. To date no other study has been covering the FAFH market with the exception of the study by Drichoutis et al. (2009), which concluded that consumers are willing to pay more for FAFH with nutrition information and that they value the EU nutrition table and TL label more than the US nutrition facts panel.

The current study extends the literature on nutrition information preferences by an evaluation of FOP nutrition labels in a specific out-of-home context, namely university canteens. Presenting canteen customers with nutrition information in a preferable format may increase its use and eventual health impact. Observed heterogeneity in the preference of nutrition information in grocery stores is also expected to exist in a canteen environment, which highlights the importance for canteens to identify label format preferences of their customers. Specifically, the present study addresses two research questions: (1) To what extent do consumers prefer simple nutrition labels to more detailed labels on meals served in a catering environment? and (2) To what extent do specific determinants of label use and liking influence the nutrition label preferences?

These research questions were analysed using a stated choice modelling approach since consumers’ preferences for various nutrition labelling formats were assumed to be the result of trade-offs between different attributes due to preferences for ease of use, for being fully informed and for not being pushed into particular food choices (Grunert and Wills, 2007). Also the possibility to evaluate the potential use of existing and new label formats or combinations of label attributes that are not yet available in the market, made discrete choice experiments of interest.

The paper is structured as follows. The second section describes the choice experiment setting in detail including the attribute selection, choice design and modelling approach followed in the analysis. The third section presents the selection between alternative choice models and the results of the best-performing model. The final section contains a discussion of the results and concluding policy implications of the model outcome.