Improving usability of nutrition information through graphical visualization

Pratt, Nathan

Permalink

https://hdl.handle.net/2142/73079 Copy

Description

Title

Improving usability of nutrition information through graphical visualization

Author(s)

Pratt, Nathan

Issue Date

2015-01-21

Director of Research (if dissertation) or Advisor (if thesis)

Nakamura, Manabu T.

Doctoral Committee Chair(s)

Donovan, Sharon M.

Committee Member(s)

• Nakamura, Manabu T.
• Ellison, Brenna D.
• Lee, Soo-Yeun

Department of Study

Nutritional Sciences

Discipline

Nutritional Sciences

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Nutrition
• nutrition labeling
• consumer behavior
• data visualization
• weight management

Abstract

Effective use of nutrition information is an important method of preventing or managing diet-related diseases. However, it is difficult to effectively communicate information to those in need when decision making occurs. Methods of simplifying nutrition information exist to help users interpret nutrient content of foods; those which present specific information for individual nutrients and interpret their content with color-coding are the most preferred by users and the most effective with regard to improving user understanding of nutrition information. However, this approach does not lead to changes in consumer behavior when implemented in realistic settings. The goal of this research was to develop a method of presenting nutrient-specific information for multiple nutrients and color-based reference information in a format that is effective at influencing behavior in realistic settings. A secondary goal was to design the method to guide users toward healthful food choices and balanced meal formation rather than emphasize which foods to avoid. We hypothesized that presenting nutrient content of foods graphically relative to a target recommendation would improve the ability of users to process that information under time constraint to improve its usability for decision making. As a first step, a graphical method of presenting nutrition information two-dimensionally was developed within the scope of weight management which plots shows fiber per calorie on the x axis and protein per calorie on the y axis, with a target in the center of the plot representing recommendations for these nutrients. Validity of nutrient selection, criteria, and presentation of the method in achieving its objectives are tested. The method was shown to be able to differentiate beneficial foods to include in the diet from foods whose intake should be limited, visualizing many dietary recommendations such as decreasing intake of discretionary fats and added sugars, choosing lean proteins, low-fat dairy, fruits, vegetables, and whole-grains, and choosing whole fruits over fruit juice. In addition, plotting nutrient content per calorie allows for visualization of how foods combine to form a meal to help users understand nutrient content in a balanced meal context. To determine if presenting nutrition information graphically instead of numerically allows users to more effectively process that information under time constraint, a cued-recall experiment was conducted. University students (n=63) were presented fiber and protein content of foods either numerically or with the graphical method for 15 seconds per food and asked to recall nutrient content after completion of the test via survey. Graphical presentation of information improved recall by up to 43% compared to recall of those shown numeric information, suggesting the graphical method was effective at communicating the desired information in a time constrained situation. To determine if this advantage can lead to changes in consumer food choices in a realistic setting, a cafeteria field experiment was conducted to determine the impact of signposting nutrition information graphically at the point of purchase on diner choices. Nutrient content of diner meals (n=362) was compared between periods of nutrition signposting using the graphical method, nutrition facts panels (NFP), or no nutrition label. Surveys were collected to determine predictors of nutrients purchased and consumer understanding of the tool. Graphical signposting improved nutrient content of purchased meals in the intended direction while NFP had no effect compared to baseline for measured nutrients. Calories ordered from total meals, entrées, and sides were significantly less during graphical signposting than the no label period (16%, 7% and 52% decrease, respectively) and NFP period (20%, 14% and 57% decrease, respectively). Graphical signposting remained a predictor of calories purchased in regression modeling and was well understood by diners. These results indicate that consumers can make healthy food purchase decisions based on nutrition information when it is presented two-dimensionally to facilitate time constrained food comparisons. Finally, the method was developed into a software tool usable for applied settings, such as dietary counseling, and formative evaluation of the tool was conducted in a potential target market of health educators. The goals of this stage of work were to 1) determine the value of the software tool to the target end users and 2) identify how the prototype can be improved to create a workable application that is able to be used for larger scale studies. Local practicing health educators (n=8) were recruited to test the software features. Formalized quantitative and qualitative feedback was collected via a survey, which included questions targeting usability (user interface, data entry, profile tools, navigation) and functionality (usefulness of current features, improvements to be made, helpful features not currently present on the software). The primary features of the software, such as plotting nutrient content graphically, color coding of the graph, and visualizing meal formation, were highly rated by users. Data entry/retrieval and a graph to monitor weight were the most consistently disliked and difficult to use features. Qualitative feedback provided specific direction in how to improve the software. This work is one of the first to demonstrate an effect of simplified nutrition information for multiple nutrients on consumer behavior in realistic, time constrained settings. The findings support the method of graphically presenting quantitative information of two nutrients and interpreting that nutrient content using color-coding and a target recommendation focused on balanced meal formation.

Graduation Semester

2014-12

Permalink

http://hdl.handle.net/2142/73079

Copyright and License Information

Copyright 2014 by Nathan Pratt

Owning Collections