Terminology

A major obstacle in developing dietary guidelines has been the confusion associated with the terminology used^{( 64 )}. Consumers have difficulty in understanding the terms PS and SS⁽ Reference Anderson, Barton and Craigie ^{48 ,} Reference Brown, Timotijevic and Barnett ^{53 )} as these tend to be used interchangeably, even within the same scheme⁽ Reference Britten, Haven and Davis ^{50 , 52 )}. Table 3 provides an overview of various cited definitions of PS and SS. These terms are sometimes believed to have the same meaning: the amount of food eaten at a single eating occasion⁽ Reference Britten, Haven and Davis ^{50 )} or one sitting⁽ Reference Anderson, Barton and Craigie ^{48 )}. In fact, in a US study, few consumers were aware that their PS could represent more than one SS⁽ Reference Britten, Haven and Davis ^{50 )}. PS can be the amount of food offered in a restaurant, or on a label or plate^{( 68 )}. Phrases associated with the term PS have been reported as ‘daily allowance’, ‘restriction’, ‘enough for one person’ and the weight of food in grams^{( 73 )}.

Table 3.

Definitions of portion size (PS) and serving size (SS) as cited in the literature

* The definitions of PS and SS used in this review.

The majority of the definitions listed in Table 3 can be loosely translated as PS being the amount of food consumed at a single eating occasion, and SS being the amount that is recommended. However, evidence from the most recent European-wide consumer research is not in accord with these definitions^{( 71 )}. It is imperative that there are universally agreed definitions of both PS and SS established and communicated effectively to the consumer. As mentioned earlier, this review utilises definitions of PS and SS that were recently cited in the UK^{( 52 )} (i.e. PS is the amount of food intended to be consumed by an individual in a single eating occasion and SS is the quantity recommended to be consumed in a single eating occasion), in order to provide a basis for comparison with future research.

Units of measurement

Various units of measurement are used in SS guidance to convey what constitutes an appropriate SS including weight (e.g. grams or ounces), household measures (e.g. cup), units (e.g. one piece of fruit), a fraction or slice (e.g. slice of bread) or proportions of a plate. Statements such as ‘a balanced diet’⁽ Reference Brown, Timotijevic and Barnett ^{53 )}, advising to consume ‘more’ of something⁽ Reference Williams ^{74 )} or terms such as small/medium/large may be too ambiguous and subjective⁽ Reference Young and Nestle ^{75 )}. Consumers interpreted SS to be the amount suitable to fulfil daily nutrient requirements rather than that appropriate to have a satiating effect. The authors suggested communicating SS in specific gram weights⁽ Reference Ueland, Cardello and Merrill ^{70 )}. However, consumers reportedly find it difficult to adapt from the imperial system to the metric system, particularly those over 45 years, while others show no knowledge of either approach^{( 73 )}. Furthermore, weights can be difficult to understand, especially for cereals, unless the SS can be related to a common object or household measure^{( 73 )}. Indeed, consumers generally understand amounts expressed in terms of household measures and units better than specific weights, e.g. the cup as a measure for fruit and vegetables⁽ Reference Britten, Haven and Davis ^{50 ,} Reference Brown, Timotijevic and Barnett ^{53 ,} Reference Pollard, Daly and Binns ^{62 )}. In addition, household measures used in conjunction with specific food examples, i.e. one cup of oatmeal (one serving of grains), reportedly helps to ‘visualise’ amounts⁽ Reference Britten, Haven and Davis ^{50 )}. However, care must be taken to ensure consumers fully understand the differences between household measures. For example, a group of UK consumers speculated that a tablespoon was: ‘three dessert spoons’, ‘probably one ounce’, or ‘is it a dessert spoon or a soup spoon?’^{( 73 )}.

Given that few consumers actually weigh their food^{( 73 )}, SS guidance using household measures with food examples may be most appropriate, with the caveat that consumers need to understand the type of the household measure in question.

Consumer perceptions v. recommendations

Consumer understanding of appropriate SS can show little correspondence with actual recommendations⁽ Reference Burger, Kern and Coleman ^{16 ,} Reference Anderson, Barton and Craigie ^{48 ,} Reference Schwartz and Byrd-Bredbenner ^{69 ,} Reference Hogbin and Hess ^{76 –} Reference Shah, Adams-Huet and Elston ^{79 )}. With respect to specific foods, consumers estimated SS of pasta, breakfast cereals, meat and rice to be larger than recommended⁽ Reference Flynn, O'Brien and Faulkner ^{49 ,} Reference Britten, Haven and Davis ^{50 ,} Reference Schwartz and Byrd-Bredbenner ^{69 , 73 )}, while SS of banana were estimated with ease⁽ Reference Brogden and Almiron-Roig ^{77 )}. In contrast, salad items⁽ Reference Schwartz and Byrd-Bredbenner ^{69 )} and mashed potato⁽ Reference Flynn, O'Brien and Faulkner ^{49 )} were estimated to be smaller than the recommended SS. A UK study found that none of the GOV or HCP SS guides under consideration correlated with a group of normal and overweight men's perceptions⁽ Reference Brogden and Almiron-Roig ^{51 )}. However, appropriate PS can vary considerably depending on age and sex⁽ Reference Rangan, Schindeler and Hector ^{80 )}, therefore it is imperative that recommended SS consider the variable needs of the whole population.

Recommended SS often bears little resemblance to consumers’ habitual eating patterns. For instance, in Ireland, one slice of bread equates to one serving; however, people are most likely to consume two slices in a sandwich or two slices of toast at one eating occasion. Consequently, efforts are being made to improve advice so that it resonates better with consumers⁽ Reference Flynn, O'Brien and Ross ^{81 )}, in a manner similar to Australian guidelines⁽ Reference Smith, Kellett and Schmerlaib ^{82 )}. In the US and the UK, self-reported PS of ethnic minority groups were often multiple times the recommended SS⁽ Reference Gans, Risica and Kirtania ^{83 ,} Reference Sharma, Cade and Landman ^{84 )}. These larger habitual PS can easily promote a distorted view of recommendations, causing over-estimated SS of various foods⁽ Reference Shah, Adams-Huet and Elston ^{79 )}. Industry are also urged to use realistic SS^{( 85 )} since manipulation of SS to make the nutritional content appear more favourable may be misleading for many consumers^{( 86 )}.

Foods that have been identified as being the most difficult for consumers to select appropriate PS include: starchy foods (rice, pasta, breakfast cereals and potatoes), meat, fats, cheese, alcohol and foods sold loose or in multi-serve packs^{( 63 )}. Composite dishes also add to the confusion. Although consumers understood that composite dishes such as pizza could account for servings from more than one food group, they were unsure why a range of servings were recommended⁽ Reference Britten, Haven and Davis ^{50 )}. While the ‘5-a-day’ campaign for fruit and vegetables is widely promoted, consumers found it difficult to understand what constitutes a SS of fruit or vegetables⁽ Reference Pollard, Daly and Binns ^{62 , 73 ,} Reference Ashfield-Watt ^{87 )}. When more guidance was available to increase consumer awareness of fruit and vegetable SS their consumption of this food group increased considerably⁽ Reference Ayala ^{88 )}.

Overall, it is apparent that recommendations may need to be reconsidered to be more reflective of consumers’ perceptions of SS and their habitual PS, in order to be more meaningful. Particular consideration and perhaps additional education may be needed for some foods.

Impact

The impact of SS aids and training has mainly been evaluated in the context of dietary assessment. Relatively few studies have assessed their direct impact on consumer SS estimation and selection⁽ Reference Anderson, Barton and Craigie ^{48 )} (n 17) and these are summarised in Table 4. Although the majority of these studies have reported a reduced error in SS estimation or a more favourable nutrient intake (positive impact), these effects were only measured in the short term and usually in the laboratory setting⁽ Reference Ayala ^{88 –} Reference Ollberding, Wolf and Contento ^{97 )}. A follow-up study conducted 3 months after initial training about SS showed that the immediate positive impact had not been maintained⁽ Reference Colapinto and Malaviarachchi ^{96 )}. Some studies showed that SS guidance had no impact⁽ Reference Ueland, Cardello and Merrill ^{70 ,} Reference Ashfield-Watt, Welch and Day ^{98 –} Reference Vermeer, Steenhuis and Leeuwis ^{101 )} and others resulted in under- or overestimation of SS (negative impact)⁽ Reference Tavelli, Beerman and Shultz ^{102 ,} Reference Riley, Beasley and Sowell ^{103 )}. These equivocal results cannot be explained by the different methods of communicating SS information, e.g. computer based and food labelling v. SS aids, because each of these methods could have either positive, negative or no impact depending on the study. Group training using food models and other aids appears to be the most consistently effective method, albeit in the short term only. The majority of these studies were carried out in the US. Moreover, there have been no long-term or intervention studies on the impact of SS guidance on weight management^{( 68 )}.

Table 4.

Impact of serving size guidance

SS, serving size; min, minutes; F, female; M, male; F&V, fruit and vegetables.

* Reduced the error in SS estimation, i.e. closer to recommended or reference size, or resulted in a more favourable nutrient intake.

† Under or over-estimation of SS.

‡ No change in SS estimation.

Awareness of SS guidance does not necessarily result in its implementation⁽ Reference Nayga and Capps ^{104 )}. For instance, the current 5-a-day campaign for fruit and vegetables is well known by the public, but there is little evidence of its effectiveness⁽ Reference Williams ^{74 )}. A UK study demonstrated that guidance on SS estimation of fruit and vegetables using household measures and food photographs as aids had no effect on PS selection⁽ Reference Ashfield-Watt, Welch and Day ^{98 )}. There is some evidence to suggest that portion control tools may be effective for aiding weight-loss in the clinical setting⁽ Reference Hannum, Carson and Evans ^{105 –} Reference Pedersen, Kang and Kline ^{107 )}, but further work is needed to evaluate the impact of SS guidance in the general population⁽ Reference Anderson, Barton and Craigie ^{48 )}.

In the US, dietary guidelines are revised every 5 years, but there is no mandate to evaluate their impact⁽ Reference Guthrie and Smallwood ^{108 )}, nor is this compulsory in the UK or Ireland. More support is clearly needed to aid the implementation and evaluation of guidance⁽ Reference Trubswasser and Branca ^{109 )}. Even when information on SS is available, consumers still tend to inaccurately assess their PS^{( 68 )}. A positive impact is more likely if consumers are aware that the change is important for their health⁽ Reference Wood, Robling and Prout ^{110 )}.

Acceptability

In general, consumers tend to be interested and recognise that SS guidance may be helpful, but do not always consider it relevant to them personally⁽ Reference Vermeer, Steenhuis and Seidell ^{26 ,} Reference Anderson, Freeman and Stead ^{111 )}, and the idea that ‘one size does not fit all’ has been evident since the 1980s⁽ Reference Hunter, Sampson and Stampfer ^{112 )}.

It would appear that consumers will only consult SS guidance when they feel intervention is required, e.g. when aiming for weight loss⁽ Reference Brown, Timotijevic and Barnett ^{53 , 73 ,} Reference Wood, Robling and Prout ^{110 ,} Reference Anderson, Freeman and Stead ^{111 )}. Furthermore, SS guidance is reportedly more accepted by consumers for main meals and foods such as rice and pasta, but is generally not considered for snack foods, drinks or staples like bread and milk^{( 73 )}. Acceptability also appears to be dependent on sex, socio-economic status, level of interest in the diet and the perceived credibility of the source⁽ Reference Driskell, Meckna and Scales ^{25 ,} Reference Vermeer, Steenhuis and Seidell ^{26 ,} Reference Anderson, Barton and Craigie ^{48 , 66 , 73 ,} Reference Anderson, Freeman and Stead ^{111 , 113 )}.

There is evidence that consumers were not inclined to implement SS guidance even when they were aware of it^{( 66 )}. Indeed, only half of UK consumers reported that they would use serving demarcations on food labels^{( 73 )}. Negative connotations were associated with measuring SS as it was considered time consuming, impeded the enjoyment of meal time, and could be frowned upon by others⁽ Reference Mian and Brauer ^{114 )}. Consumers are generally amenable to tools and household measures as guides to SS^{( 73 )} and the provision of computer-based SS information is generally well accepted by younger groups, e.g. students⁽ Reference Ayala ^{88 ,} Reference Carlton, Kicklighter and Jonnalagadda ^{115 ,} Reference Cousineau, Franko and Ciccazzo ^{116 )}. Although proportional pricing strategies were considered to be unattractive for large households⁽ Reference Vermeer, Steenhuis and Seidell ^{26 )}, consumers were reportedly most receptive to pricing strategies, SS labelling and the provision of a larger range of SS⁽ Reference Vermeer, Steenhuis and Seidell ^{26 ,} Reference Vermeer, Steenhuis and Seidell ^{99 )}. However, despite being viewed as acceptable, pricing strategies had no effect on PS selection⁽ Reference Vermeer, Alting and Steenhuis ^{117 )}.

The available data suggest that SS guidance should be simple, non-prescriptive, consistent, practical and flexible in order to be accepted by consumers^{( 73 ,} Reference Anderson, Freeman and Stead ^{111 )}. Ultimately, the uptake of any recommendation about SS is unlikely unless they are both realistic and achievable⁽ Reference Watts, Hager and Toner ^{118 )}.

Usability

As previously discussed, consumers have difficulty comprehending incoherent and contradictory advice⁽ Reference Britten, Haven and Davis ^{50 )} available from various sources, each with different perspectives⁽ Reference Goldberg and Sliwa ^{119 )}. For example in the US, the inconsistencies between SS from the Pyramid^{( 120 )} (half a cup of cooked pasta) and the Nutrition Facts label on food packaging (1 cup cooked pasta)^{( 121 )}, add to consumer confusion⁽ Reference Britten, Haven and Davis ^{50 )}. Labelling SS for products from the grains group can vary considerably^{( 121 )}. US market-place PS were found to be larger than those recommended⁽ Reference Young and Nestle ^{8 ,} Reference Young and Nestle ^{47 )}, e.g. cookies were seven times bigger, while cooked pasta was almost five times the recommended SS⁽ Reference Young and Nestle ^{8 )}. Moreover, reference SS for food labels were derived from food consumption surveys conducted between 1977–78 and 1987–88^{( 122 )}. In Australia, SS for snacks varied greatly (18–100 g), while the SS for drinks frequently did not correspond to the size of the container⁽ Reference Walker, Woods and Rickard ^{123 )}. Furthermore, SS of some own brand foods such as ready meals and pizza were significantly smaller than their market brand equivalents⁽ Reference Cleanthous, Mackintosh and Anderson ^{124 )}. In Europe, there are no regulations or clarification of SS on labels⁽ Reference Anderson, Barton and Craigie ^{48 )}. A comparison of the recommended UK medium PS^{( 59 )} with food label SS demonstrated several inconsistencies, e.g. the average SS for a medium slice of steak pie is 120 g but this ranges between 138 and 300 g on food labels⁽ Reference Anderson, Barton and Craigie ^{48 )} making it difficult to compare products. On the other hand, UK consumers thought SS on front-of-pack labels helped them make quick comparisons between the nutritional composition of foods^{( 86 )}.

Recommended SS of foods within the same food group should also be nutritionally comparable to enable consumers to exchange foods. For instance, a SS of rice should be comparable in terms of energy content with an SS of pasta. In the present Irish dietary guidelines, there is a wide range in energy content between SS of bread, cereals and potato ranging from 314 to 1046 kJ (75–250 kcal), although these were deemed to be equivalent⁽ Reference Flynn, O'Brien and Faulkner ^{49 )}. In theory, this limits the consumer's ability to effectively use SS in the intended way, although efforts have been made to rectify this issue^{( 125 )}.

Foods labelled as containing multiple servings can be problematic. An entire packet of a food product is often eaten oblivious to the fact that it contained several servings^{( 68 )} and unless the whole pack contained just a single SS consumers were confused about how to interpret the nutrition information^{( 73 )}. Details of the number of servings in packaged amorphous foods, demarcations of individual SS on packets of foods such as rice and butter, and individually packaged SS can be useful^{( 73 )}. Despite this less than a quarter of consumers use food labels to aid their estimation of PS^{( 63 )}. With regard to the elderly, some may not be able to remember SS recommendations that would limit their ability to follow them⁽ Reference Ervin and Smiciklas-Wright ^{126 )}.

Food photographs⁽ Reference Ovaskainen, Paturi and Reinivuo ^{127 )}, household measures and other practical tools can be useful in PS estimation although these have mainly been investigated with respect to their use in dietary assessment⁽ Reference Anderson, Barton and Craigie ^{48 )}. Photographs and food models can help the consumers to visualise their typical PS but they may not be useful for composite dishes, e.g. sauce covering meat or rice in a meal, making it difficult to interpret PS of the individual food items on the plate⁽ Reference Wrieden and Momen ^{128 )}.

Another issue is that SS guidance does not always consider the types and amounts of foods typically eaten by ethnic minorities⁽ Reference Gans, Risica and Kirtania ^{83 ,} Reference Sharma, Cade and Landman ^{84 )}. ‘The plate model’ was considered useful, but concerns have been expressed about the large proportion of vegetables recommended because of the customary addition of fats and oils to vegetables by some ethnic groups. For example, a typical PS of vegetable soup for a British–African–Carribean group (which was more than double that of the Caucasian population in the UK) could contain up to 25 g of fat⁽ Reference Sharma, Cade and Landman ^{84 )}. A Canadian study conveyed that the ‘hand jive’ method (e.g. using the palm of the hand to estimate a serving of protein) was too vague and difficult to remember for immigrants from South Asia, who found themselves thinking of SS in terms of household measures⁽ Reference Anderson, Freeman and Stead ^{111 )}. Consequently, such tools may need special consideration of the traditions and language of ethnic minority groups⁽ Reference Mian and Brauer ^{114 )}. Further research is needed to assess the validity and relevance of SS tools and aids such as food photographs and household measures for ethnic minority groups⁽ Reference Cypel, Guenther and Petot ^{129 )}.

Potential barriers

The potential barriers to the uptake of SS guidance have been widely documented in the literature. There are various environmental factors that can act as potential barriers to the implementation of SS guidance both within and outside of the home, such as package size^{( 63 ,} Reference Wansink and Park ^{130 )}, eating food in units⁽ Reference Geier, Rozin and Doros ^{131 )}, poor nutrition knowledge^{( 63 )}, value for money⁽ Reference Vermeer, Steenhuis and Seidell ^{26 , 73 ,} Reference Vermeer, Steenhuis and Leeuwis ^{101 ,} Reference Watts, Hager and Toner ^{118 )}, irregular eating patterns^{( 63 )}, time constraints⁽ Reference Anderson, Barton and Craigie ^{48 ,} Reference Britten, Haven and Davis ^{50 ,} Reference Brown, Timotijevic and Barnett ^{53 ,} Reference Vartanian, Herman and Wansink ^{132 )}, taste⁽ Reference Vartanian, Herman and Wansink ^{132 )}, social interactions/distractions⁽ Reference Wansink ^{133 ,} Reference Koh and Pliner ^{134 )}, food cue exposure⁽ Reference Tetley, Brunstrom and Griffiths ^{135 )} (especially in unrestrained eaters)⁽ Reference Ferriday and Brunstrom ^{136 )}, language barriers⁽ Reference Mian and Brauer ^{114 )} and literacy and numeracy skills⁽ Reference Huizinga, Carlisle and Cavanaugh ^{137 ,} Reference Rothman, Housam and Weiss ^{138 )}. Larger plates have been linked to larger PS⁽ Reference Wansink ^{133 )}, although manipulation of plate size does not affect food or energy intake⁽ Reference Rolls, Roe and Halverson ^{139 ,} Reference Shah, Schroeder and Winn ^{140 )}. One of the main reasons for failing to adhere to such guidance is consumer habit and experience⁽ Reference Anderson, Barton and Craigie ^{48 , 66 , 73 ,} Reference Brunstrom, Shakeshaft and Alexander ^{141 )}. The tradition of ‘cleaning one's plate’^{( 66 , 73 ,} Reference Wansink, Painter and North ^{142 )} occurs at over 91% of meals⁽ Reference Fay, Ferriday and Hinton ^{143 )} with just over half of consumers admitting to this when eating out^{( 73 )}. Consumers have been found to ignore SS guidance when eating outside of the home^{( 66 )}, albeit they considered the home to be the most challenging setting in which to control their PS^{( 63 )}.

The adoption of SS guidance is difficult in a culture where larger PS have become ‘the norm’⁽ Reference Vermeer, Steenhuis and Seidell ^{26 ,} Reference Anderson, Barton and Craigie ^{48 ,} Reference Vermeer, Steenhuis and Leeuwis ^{101 )} and recommended SS are perceived as being too small⁽ Reference Vermeer, Steenhuis and Seidell ^{26 )}. Consumers are generally oblivious to the fact that these consumption norms coupled with the underestimation of energy content in large PS can often result in consumption beyond their needs⁽ Reference Wansink and van Ittersum ^{144 )}. Another obstacle to adopting SS guidance is the increasing concern about avoiding food wastage⁽ Reference Vermeer, Steenhuis and Seidell ^{26 , 63 , 66 ,} Reference Vermeer, Steenhuis and Leeuwis ^{101 )}. Lack of consideration for PS has also been attributed to hunger and satiety cues⁽ Reference Vermeer, Steenhuis and Seidell ^{26 , 73 ,} Reference Vermeer, Steenhuis and Leeuwis ^{101 ,} Reference Mian and Brauer ^{114 ,} Reference Vartanian, Herman and Wansink ^{132 ,} Reference Fay, Ferriday and Hinton ^{143 ,} Reference Brunstrom and Shakeshaft ^{145 –} Reference Shimizu, Payne and Wansink ^{151 )}. Approximately a fifth of consumers eat what is on their plate until they feel full^{( 73 )}. Hunger has been typically shown to cause an increase in PS^{( 73 ,} Reference Brunstrom, Rogers and Pothos ^{152 )}. However, PS estimates of a range of foods and beverages were reported to be significantly smaller under hungry compared with full conditions⁽ Reference Brogden and Almiron-Roig ^{51 )}. PS has also been shown to correlate with food liking and familiarity, and larger PS were estimated for foods expected to be less satiating⁽ Reference Brogden and Almiron-Roig ^{77 )}.

Consumers are sceptical of labelled SS as they feel the recommended SS may be manipulated by manufacturers to mislead them^{( 86 )}. The ‘health-halo’ effect associated with low-fat foods may also lead consumers to disregard SS information⁽ Reference Matthiessen, Fagt and Biltoft-Jensen ^{10 ,} Reference Wansink and Chandon ^{153 )}. The so-called healthier options may not be as satisfying and may lead to an increased consumption^{( 73 )}. In addition, the food industry is reluctant to reduce SS as this would increase packaging costs and consumers could perceive it as a strategy to increase profits⁽ Reference Vermeer, Steenhuis and Seidell ^{99 )}. Marketing techniques such as product naming, reduced cost, labelling and presentation can make the larger portions more appealing⁽ Reference Vermeer, Steenhuis and Seidell ^{99 )}.

Health professionals have also identified barriers to providing SS advice, i.e. determining the level of details required in SS guidance communications, conflicting information in the media (especially for carbohydrates), and the absence of national quantitative guidance⁽ Reference Anderson, Barton and Craigie ^{48 )}.