The prevalence of overweight and obesity in the United States has increased dramatically 1. One factor that many believe to be an important contributor to this increase is the number of meals and snacks eaten away from home. Over the past several decades the proportion of total food expenditures spent on food away from home has increased from 34% in 1974 to about half in 2004 2. Foods available at restaurants and other away from home eating locations tend to be higher in calories and fat 3456 and often larger in portion size 78 compared to foods eaten from home, which may contribute to energy intake in excess of energy expenditure.

In recognition of the potential role of meals eaten away from home on excess energy intake, a number of strategies to promote more healthful food choices when eating out have been proposed 910111213. One recommended approach is to increase the availability of nutrition information for foods eaten and prepared away from home 910111213. Moreover, it has been suggested that fast food chain restaurants be required to provide calorie information on their menu boards or on product packaging 10. In theory, the provision of calorie information at the point-of-purchase for restaurant products may help consumers limit excess calorie intake.

Aside from providing point-of-purchase nutrition information, it has also been suggested that the food industry reduce its use of value size pricing as a marketing technique 12. Value size pricing involves structuring product prices such that the per unit cost (e.g., price per ounce) decreases as portion size increases. It has been speculated that this product pricing structure leads to higher energy intake when eating out because value conscious consumers may be prone to purchase larger-sized food items 12.

Although a number of studies have been conducted to evaluate the effect of point-of-purchase nutrition promotions on foods purchased away from home 141516171819202122232425262728293031323334353637, the relevance of most of these studies to calorie labeling in a restaurant context is limited for a number of reasons. First, many of the studies evaluated point-of-purchase promotional activities that focused solely on identifying and promoting more healthful food choices (e.g., low-fat foods) 161920212223272830313233 versus providing nutrient content information, such as calorie information, for the full range of available foods. Many of the studies were conducted in a workplace or university setting 1415161718192031323334353637 instead of a restaurant. Among those studies that evaluated comprehensive calorie labeling (e.g., providing calorie information for most food items) 14151718242634353637, a number evaluated self-reported behavioral intentions 142425262937 rather than actual food purchases.

In consideration of the limitations of previous studies, we conducted an experimental trial to examine the effects on food purchases of providing calorie information at the point-of-purchase for food items on a fast food restaurant menu. The influence of value size pricing on fast food meal choices and consumption was also examined.

The demographic characteristics of participants are shown in Table 1. More females (59.4%) than males (40.6%) participated. Approximately three-fourths of participants were white, with blacks comprising the second largest racial/ethnic group (10.9%). The demographic characteristics of participants in each experimental group were similar.

When asked to rate the importance of price, taste, nutrition and convenience when purchasing food from a fast food restaurant and when buying groceries, taste was the most highly rated factor for both; 97.6% and 98.5% reported taste as very important or somewhat important when buying fast food and groceries, respectively (Table 2). In contrast, nutrition was the least likely to be rated as very important or somewhat important, with 58.2% and 83.5% of participants rating nutrition as very important or somewhat important when buying fast food and groceries respectively.

The average energy and nutrient composition of meals ordered and consumed by those in each experimental group (Table 3) were similar. For example, there were no significant differences (p = 0.25) in the average number of calories consumed by those in the calorie, price, calorie plus price, and control menu conditions were 805, 813, 761, and 739 respectively. Selection and consumption of major food categories (e.g., sugar-sweetened soft drinks, diet soft drinks, French fries, salads, etc.) and portion sizes were also examined, with no significant differences found (data not shown).

The average energy content of meals selected and consumed were similar across experimental conditions among those in each age, education level, and body weight strata. However, significant differences in energy intake across experimental condition were observed among males (p = 0.01), those who reported that nutrition was important when buying food from a fast food restaurant (p < 0.01), and those who reported price was not important when buying food from a fast food restaurant (p = 0.01). Average energy intake was higher among males in the calorie, price, and calorie plus price experimental conditions compared to those who selected their meal from the control menu (Figure 7). Among those who reported that nutrition was important when buying fast food, average energy intake was significantly lower among those who received the control and calorie plus price menus relative to those in the other two experimental conditions (Figure 8). Among those who reported that price was not important when buying food from a fast food restaurant, average energy intake was lowest among those in the control condition (598 kcal) and highest among those in the calorie plus price condition (948 kcal) (Figure 9).

As part of the post-meal interview, participants in the calorie, price, and calorie plus price menu conditions were asked questions to assess whether they noticed the menu manipulations. Those who received a menu with calories listed were asked, "Did you notice that calorie counts were listed on your menu this evening?". About half (54%) of those in the calorie condition and 59% of those in the calorie plus price menu conditions reported noticing the calorie information (Table 4). Those with a higher level of education, whites, and those 15–25 years of age were more likely to report noticing the calorie information (data not shown). Participants who received the price or calorie plus price menu were asked, "Did you notice on the menu from which you ordered tonight that prices were set such that for a given unit of food (like a chicken nugget or ounce of soda), you paid the same price for that unit no matter what size you ordered?" Due to the complexity of the question, an illustration of this pricing structure was provided and explained as part of the question. Less than one-fifth reported noticing this pricing structure (Table 4). The demographic characteristics of those who reported noticing the price modification were similar to those who did not (data not shown).

To evaluate whether calorie information may have influenced food choices among those who reported noticing this information relative to those who did not, a linear regression analysis was conducted. This analysis was restricted to participants in the calories and calories plus price experimental conditions. Covariates included in the analysis were factors found to differ between those who reported noticing the calorie information and those who did not notice this information, specifically age, race, education level, and site. Results from the multivariate analysis indicated that average energy intake was comparable between those who reported noticing the calorie information and those who did not (690 kcal versus 671 kcal; p = 0.65). Results were similarly null in a comparable analysis conducted to compare energy intake of those who noticed and did not notice the pricing structure modification (p= 0.90).

Results of the present study showed that providing calorie information at the point-of-purchase on a fast food restaurant menu had little effect on food selection and consumption among a sample of adolescents and adults who eat regularly at fast food restaurants. These results contribute to a limited literature on point-of-purchase calorie labeling. To date, seven studies have examined the influence of providing calorie composition information at the point-of-purchase for most food items available in a cafeteria 1415171837 or restaurant 2426 setting. Among these studies, one found no evidence of an effect of calorie labeling on food choices 18. In contrast, six of the seven studies found some evidence in support of the hypothesis that calorie information may positively influence food choices 1415241726, however, results from most of these studies were weak or inconsistent. For example, Conklin et al. found that only 18% of college freshman living on a campus where point-of-purchase nutrition information was available in the dining commons agreed that the available information affected their choice of food 37.

A host of factors may explain the weak and inconsistent results in the literature. Firstly, the calorie labeling formats utilized varied across studies. For example, in one study 5 cm by 5 cm cards with calories in red ink were placed as close as possible to food items in a hospital cafeteria 17. In contrast, in another cafeteria study calorie information for all menu items was presented on two large posters at the cafeteria entrances, with leaflets distributed to patrons to encourage use of this information 15. Also, in three studies calorie information was provided along with other nutrient composition information such as saturated fat and fiber 142437. In the present study, calorie information alone was provided on a restaurant menu in a column between the food item name and price. To draw attention to it, the column was highlighted in bright yellow. Nonetheless, only slightly more than one-half of those who ordered from a menu with calories listed reported noticing this information. The calorie content of meals selected by those who noticed the information compared to those who did not were similar, suggesting our null results are not solely due to the failure of some to notice the calorie information.

The designs of studies conducted to date have varied greatly, with all having limitations. Most notably, four studies measured behavioral intentions 14242637 rather than actual food choices. Consequently, social desirably bias in reporting is a significant concern in these studies. Other major weaknesses include use of quasi-experimental designs 151718 where factors other than the experimental conditions being tested may have differed across test periods due to lack of randomization. The present study is the first to measure actual food choices within an experimental design where participants were randomly assigned to experimental condition. However, it has methodological weaknesses. Participants were exposed to the calorie information on only one occasion. This is a critical shortcoming if repeated exposure to calorie information is required before awareness or behavior change may be expected.

A final issue is that the weak and inconsistent results across studies may reflect heterogeneity in response to calorie labeling, with certain population subgroups more apt to utilize calorie information when it is provided. For example, a number of studies have found that females are more likely than males to use nutrition information on packaged food products 404142434445. Consequently, it is perhaps not surprising that Milich et al found calorie labeling to effect cafeteria food choices in a sample of female hospital employees 17, whereas Mayer et al. found no significant effect of calorie labeling on cafeteria food choices in a study involving male and females employees of a Fortune 500 company 18. The results of the present study are consistent with the notion that certain population subgroups may be more likely to use calorie information when it is provided. In the present study males appeared to use the calorie information to choose a higher calorie meal. This finding could be an artifact of multiple comparisons, as a significant number of subgroup analyses were conducted. Conversely, this result could reflect a desire among males for an energy dense meal. To our knowledge, only one other study 24 has reported findings suggesting an unintended consequence of calorie labeling. Yamamoto et al. conducted a study in which adolescents were asked to order meals from three different restaurant menus that did not contain nutrition information, and then reorder their meals after being shown a version of the menus that included calorie and fat content information for menu items. Approximately 17% of meal orders were changed in response to the calorie and fat information. Among the meals that were modified, 20.4% were modified in a way that resulted in a higher calorie content meal.

In the present study the elimination of value size pricing was found to have little influence on food selection or consumption. This finding is somewhat surprising given that a number of studies have documented that price changes may influence food choices 32464748. The price shifts we evaluated tended to be smaller in magnitude compared to those evaluated in previous studies, which could explain why our results conflict with previous findings. It is also possible that the null results are due to the study design which provided only one exposure to the price modification. When queried regarding whether they had noticed the modified pricing structure, less than one-fifth responded affirmatively. Since most fast food restaurant chains utilize a value size price structure, it is possible that study participants generally assumed the larger sized items were the better value without considering the prices listed on the menu. In consideration of this potential methodological issue, future studies designed to evaluate value size pricing should ensure repeated exposure to price modification.

The present study has a number of strengths. The study measured actual food choices rather than behavioral intent. Consequently, social desirability bias in reporting is likely less of a concern and internal validity is probably better than studies that only measured behavioral intent. A randomized design was employed ensuring that potential confounding factors, such as age and sex were equally distributed across experimental conditions. Another strength of the present study is that participants were a community sample of adolescents and adults who ate regularly at fast food restaurants. Consequently, the external validity of results may be stronger than many previous studies.

Limitations of the present study include that participants were exposed to the experimental condition only once. As mentioned earlier, this is problematic as it is possible that repeated exposure to calorie information and standardized pricing may be required before behavior is impacted. Although participants were blinded to the true intent of the study, and the ordering and dining procedure was set-up to be as naturalistic as possible, subject reactivity remains a concern. Of particular concern is the possibility that the participation incentive undermined price sensitivity.

It is important to note that most of the study limitations just described could have been avoided if the study had been conducted in fast food restaurants where menu boards and prices could be manipulated for prolonged periods of time. Unfortunately we were not able to find any fast food restaurants willing to collaborate with us, and thus we were not able to implement a more rigorously designed study.

In conclusion, results from this study indicate that providing calorie information for food items on fast food restaurant menus may have little effect on the food choices made by adolescents and adults who regularly eat at these establishments. It is possible that skills for using point-of-purchase nutrition information must be built before the information provided may be effectively used. For example, Hawthorne at el. found that knowledge of the basic use of the nutrition facts label on packaged food products was low among a sample of adolescents 49. After a brief training on use of the label, understanding of the label was significantly improved. More likely though is the need to increase concern about nutrition when eating at fast food restaurants, as factors such as taste and convenience appear to be far more important consideration for most consumers.

Although the design of the present study is more methodologically rigorous than most previous research, it has significant shortcomings. Consequently additional research is warranted to more rigorously evaluate calorie labeling and value size pricing in the context of fast food restaurants. Collaborating with fast food restaurant establishments will be a critical but challenging need.

The authors declare that they have no competing interests.

LH designed the study and developed the study procedures with input from SF, MS, RJ, MO, and SR. SR coordinated data collection and performed the statistical analysis under the direction of MO and LH. All of the authors were involved in drafting the manuscript. Also, all of the authors read and approved the final manuscript.