Portion sizes in restaurant meals come in for particular criticism because calorie information is not mandatory and therefore consumers must estimate the calorie content of the meals they choose. A thorough understanding of consumer food choices, and in particular calorie estimation, is critical for both regulators and food manufacturers if they are to address the complex set of issues involved and promote beneficial behaviour changes while maintaining corporate profitability.
This working paper, to be published in the February 2007 issue of the Journal of Marketing Research, is one of a series of three investigations into the factors that influence food choices, conducted by Pierre Chandon, Assistant Professor of Marketing at INSEAD and Brian Wansink, the John S. Dyson chair of marketing and nutritional science, department of applied economics and management, Cornell University.
This paper examines how consumers estimate the calorie-content of their meals and reveals a new finding that has important implications for policy makers, food companies and health professionals.
The authors developed a perceptual model of meal size estimation, and used it to test the assumption that people of higher body mass index (BMI) are more likely to underestimate calories than those of normal BMI. They then used the model to examine the reasons behind the failure of education campaigns, such as the FDA's "Count Calories" initiative, to promote weight loss, and identified an alternative method for helping consumers estimate better the calorie content of their meals.
Presenting survey participants with a typical fast food meal, the authors found that the calorie content of such meals was usually underestimated, but that, contrary to expectations, underestimation was related to the size of the meal rather than the BMI of the subject. They determined that, among both overweight and normal weight subjects, smaller meals were estimated more accurately than larger meals, and that underestimation increased with the size of the meal.
A second study found that higher BMI subjects tended to choose larger meals, which supported the assumption that the underestimation of calories associated with higher BMI subjects was a consequence of meal choice rather than BMI. This study also found that calorie estimation could be improved in subjects who were told to use a piecemeal procedure (estimating calories for each item separately) to calculate total calories.
In a field study, the authors again found that calorie estimation was more accurate for smaller meals than for larger meals, and that higher BMI individuals chose larger meals. Nutrition involvement - whether participants considered the health impact of what they ate - was tested and again the researchers found that calorie estimations were driven by meal size rather than nutrition involvement.
These findings led the researchers to test dieticians' assumptions about meal size and estimation biases, because these could have important implications for the clinical treatment of obesity. They found that, while dieticians underestimated the calorie content of meals, their estimations were more accurate than those of normal consumers.
They also found that dieticians wrongly assumed that high BMI individuals were less accurate at estimating calorie content than lower BMI individuals. Using the piecemeal decomposition method of estimating calories improved the accuracy of consumers' estimations, and influenced their consumption decisions in favour of smaller meals.
As larger portion sizes are readily available, and are actively marketed, the calorie underestimation bias can explain why average obesity rates are increasing over time. While the food industry is not exonerated by this study, it demonstrates that BMI cannot be linked to an individual's inability to estimate calories.
Efforts to promote counting calories as a way to lose weight may therefore be misguided, and authorities should look for alternative methods that have a better chance of success.