mm.»

2.92::
a

 

 

 

 

$11!
1 I}: .
. } "lr
. ‘l «, {.lt .’
. .1~ v eh}: :"..-...!I. v;
a .1 a V .i
1 v! $1.... I}. hi xlIerIII
t... 3.3:
7:"I‘ a.
RH! as . a
z , t .9...» .98.... ..
c2 . “34.4 L #9. .
.5.» may: no.5 .6; .51..» 2..
.u» \ 1..., “up n
.5! hr... In?!
.v 5.3.. 2!.

‘IV‘ ,3.
mm.“ H . :mu 1
.t : ‘civg‘ll uv‘
..

2r vvtpl‘
. .Ax..vi,!
3.111..

la 7 .5

421..
. ail:
I

.i 2v)

:3. . 5.

y: .32..
4!..3? . 333.:

v.

4.11:. A»!

3.99..
I a.“ v
Q}. .3!

. ; 59.th

5.3... r»...
3.1 3
.1... 33..

llttt' 1
Vt. ‘3 01. V. t. A
'10.}; :1

tea
eff! Ac. It

a IL. riff... a

. . r33 :0. ‘:

 

 

LIBRARY
Michigan State
University

o

 

 

 

This is to certify that the
thesis entitled

. MEASURING THE RELATIVE PROMINENCE OF
GRAPHIC SYMBOLS VS. TEXT FOR NUTRITION LABELS
USING EYE TRACKING

presented by
Chang Hoon Oh

has been accepted towards fulfillment
of the requirements for the

__-___M;§- __ degree in Packaging

,z/Wazz,
Major Professor's Signature
7/ 92? ./ /0

Date

 

MSU is an Affirmative Action/Equal Opportunity Employer

...._- -o-.---u---.----nn-.—-A

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE
JAN 2 1 2014

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5108 KIProlecc&PresICIRC/DateDue.indd

MEASURING THE RELATIVE PROMINENCE OF
GRAPHIC SYMBOLS VS. TEXT FOR NUTRITION LABELS
USING EYE TRACKING

By

Chang Hoon Oh

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Packaging

2010

ABSTRACT

MEASURING THE RELATIVE PROMINENCE OF
GRAPHIC SYMBOLS VS. TEXT FOR NUTRITION LABELS
USING EYE TRACKING

By
Chang Hoon Oh

As mortality caused by diseases with adult onsets increases, healthy lifestyles
become critical. As a result, people are paying more attention to diet and the selection of
food products. In this atmosphere, the nutrition information on food labels and functional
foods is rising, and the US Food and Drug Administration (FDA) is currently

considering revising nutrition labeling.

This study objectively examines the attentive behaviors of subjects when viewing
9 cereals with nutrition information presented in the traditional format and an altered
format, which includes the redundant cue of an iconic face for three nutrients that have
been linked to disease (sodium, sugar and fat) using eye tracking. Results indicate that the
iconic format enhanced the attentive behaviors of study participants for all the three
dependent variables analyzed (total time in zone, (P<0.0001), probability of noticing in
zone (P= 0.0051) and number of hits to the zone (P=0.0039)) compared to the traditional
Nutrition Facts Panel. Although further research is recommended, the fact that the
redundant cues increased the probability of viewing the nutritional information, and that

PCOple spent longer viewing this information holds promise for labeling policy.

ACKNOWLEDGEMENTS

It is truly hard to describe my deepest heartfelt thanks to Dr. Laura Bix for accepting me
and giving me the chance to work on this project. This study would not have been
possible unless her support. Due to her valuable advice and guideline during this research
project, I could complete this thesis successfully.

I would like to express my sincere appreciation to Dr. Joseph Kuszai and Dr.
Diana Twede for giving me a guidance and opinion for this project as committee
members.

I would like to thank to Dr. Wei Wang for helping me to organize the statistical
data and providing me a SAS code. I appreciate your continuous support for data analysis.

I would like to thank to Mr. Nate Deck and other engineers in Kellogg’s
Company (Battle Creek, MI, USA) to give me an opportunity to visit their headquarter
office and to provide valuable information about food label design for this project.

I would like to thank to Dr. Katie Anderson and Robert Wilson from Applied
Science Laboratories (Bedford, MA, USA) for training and helping to learn the Eye-
tracker instrument and to manage Gaze-tracker software for this study.

I am thankful to the Packaging Hub research group members, Raghav, Julia,
Carly, Wontae, Chris, Javier and all other people for their support and creative ideas.

I would like to show my gratitude to all school of packaging students for their
support especially to the PIC group members.

I would like to thank to Dongho Kang, Sungwook Hwang and Minjung Joo for
their encouragement and support during this study.

It is needless to say the extent of support and love from my family members. My
father HeungSei Oh, my mother HyunJung Joe, and elder sister DongEun Oh always
have encouraged me during my whole master’s years. I love you all.

Thanks to all my best friends, Yunhyung, Youngkuen, Jaeun, Sungjun Kim,
Seokwon, Kyunghun, Hyeyoung, Minji and Byungkook Kim, Gwantae, Yuichi, Mike
Snow, Ray and Garrett Deck for their support and encouragement. I also want to special
thank to Minjung Kyung for inspiring me at all times.

Chang Hoon Oh

East Lansing, MI, USA
July 23, 2010

iii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. vi
LIST OF FIGURES ......................................................................................................... viii
CHAPTER 1. Food and Health Eating Diet ........................................................................ 1
1.1. Health and Food in Everyday Life ...................................................................... 1

1.2. Bad Eating Habits and Risks ............................................................................... 1
1.2.1. Obesity and Health Risks ........... . .............................................................. 1

1.3. Risks from 3 Critical Nutrients (Fat, Sodium, and Sugar) .................................. 4

1.4. Public and Private Efi’orts for Healthy Eating .................................................... 4
Chapter 2; Ready to Eat Cereals for Healthy Diet ............................................................... 7
2.1. Cereals for Breakfast ............................................................................................ 7

2.2. Cereal as a Functional Food ................................................................................. 8
2.2.1 Various Positive Effects of Cereals .......................................................... 8

2.3. Choosing Better Cereals for Health .................................................................... 1 1
CHAPTER 3. Nutrition Fact Panels .................................................................................. 13
3.1. Nutrition Facts Panel ......................................................................................... 13
3.1.1. Importance of Nutritional Labeling ......................................................... 13

3.1.2. General FDA Regulations of Nutrition Facts .......................................... 15

3.2. Efforts for Nutrition Labeling Development .................................................... 17
3.2.1. Motivations of the Better Labels ............................................................. 17

3.2.2. Efforts for Nutrition Label Development ................................................ 19

3.3. Opinions for Label Development ...................................................................... 33
CHAPTER 4. Materials and Apparatuses ......................................................................... 34
4.1. Materials ............................................................................................................ 34

4.2. Apparatuses ....................................................................................................... 46
CHAPTER 5. Methods ...................................................................................................... 50
5.1. Working Hypothesis ......................................................................................... 50

5.2. Experiment Procedure ....................................................................................... 50
CHAPTER 6- Results, Analysis and Discussion ............................................................... 53
6.1. Arranging Data for Result Interpretation .......................................................... 53

6.2. Subject Demographics (Various Human Factors) ............................................ 55

Age ..................................................................................................................... 55
Gender............... ................................................................................................. 56

Visual Acuity ..................................................................................................... 57

BMI (Body Mass Index) .................................................................................... 57

The Opinion for Healthy Eating Question ......................................................... 58

6.3. Interpretations of Eye Tracking Data ................................................................ 59

iv

6.3.1. Section 1: The probability number of noticing in a zone (Binary response:

Noticed Yes or No) ............................................................................................ 60

6.3.2. Section 2: Numerical Response Variable- the Number of Hits in a Zone
........................................................................................................................... 68
6.3.3. Section 3: Numerical Response Variable- the Total Time Spent in a Zone
........................................................................................................................... 76
CHAPTER 7. Conclusions, Limitations and Future Study ............................................... 85
7.1. Effectiveness of the Graphic Nutrition Label and Consumer Behavior ........... 85
7.1.1. Effectiveness of the Facial Icon .............................................................. 86
7.1.2. Other Consumer Behaviors ..................................................................... 87
7.2. Limitations and Expectation for Future Study .................................................. 88
7.2.1. Limited Variation in the Subject Population ........................................... 88
7.2.2. Package Labels on a Two Dimensional surface ...................................... 88
APPENDIX 1. Research Consent Form ............................................................................ 90
APPENDIX 2. Recruitment Advertisement for Research Experiment ........................ 93
APPENDIX 3. Data Recording Sheet ............................................................................... 95
APPENDIX 4. Subject Filtering Data Based on Time Loss ............................................. 98
REFERENCE ................................................................................................... _ ................. 99

LIST OF TABLES

Table 1- “GDA Tables for Men and Women”; Institute of Grocery Distribution (2008)

[66] ..................................................................................................................................... 28
Table 2- “GDAs for children” Institute of Grocery Distribution (2008) [66] ................... 29
Table 3- Levels of Nutrition Contents (Fat, Sodium, and Sugar) ...................................... 32
Table 4- the standard quatity levels of 3 Critical Nutrition ............................................... 42
Table 5- Health Levels of Used Cereal Products ............................................................... 54

Table 6- The BMI Categories (Data Source from; http://www.nhlbisupport.com/bmi/) ..58

Table 7- Type III Test of Fixed Effects of the probability of noticing a given zone ......... 61
Table 8- Pairwise comparison of NIF with Zone for the probability noticing a zone ....... 62

Table 9- Pairwise comparisons of cereal health level with zone on the probability of
noticing a given zone ......................................................................................................... 65

Table 10- Correlation of BMI with each zone level and probability noticing in a zone ...68

Table 11- Type III Test of Fixed Effects of the Number of Hit in a Zone ........................ 69
Table 12- Pairwise comparison of NIF with Zone for number of hits in a zone ............... 70
Table 13- Pairwise comparison of cereal health level with Zone for number of hits ........ 72
Table 14- Correlation of BMI with each zone level for the number of hits in a zone ....... 73

Table 15- Pairwise comparisons of gender with zone for the number of hits in a given
zone .................................................................................................................................... 74

Table 16- Type III Test of Fixed Effects of the Total Time Spent in a Zone .................... 77

Table 17- Pairwise comparisons of nutrition formatting with zone for the total time spent
in a zone ............................................................................................................................. 78

Table 18- Pairwise comparison of a cereal’s health level with zone for the total time spent
in a given zone ................................................................................................................... 80

Table 19- Pairwise comparison of BMI with ZONE for the total time spent in a given
zone .................................................................................................................................... 82

vi

Table 20- Pairwise comparisons of Gender with Zone for the total time spent in a given
zone .................................................................................................................................... 83

vii

LIST OF FIGURES

Figure 1- Use of information on food and beverage packages; Data from “International

Food Information Council Foundation food and health survey” [39] ............................... 14
Figure 2- “A Food Labeling Guide” by US. Food and Drug Administration” [44] ......... 16
Figure 3- Health Claims ..................................................................................................... 19
Figure 4- “pick and tick labeling” in New Zealand [59] ................................................... 24
Figure 5- Green Keyhole labeling [60] .............................................................................. 25
Figure 6- “heart symbol” [62] ............................................................................................ 25
Figure 7- “Traffic Light Labeling” Food Standard Agency, UK. [69] ............................. 31
Figure 8- Near point Visual Acuity Card, Dow Corning Ophthalmics (Lefi; fi'ont, Right;

back) .................................................................................................................................. 35
Figure 9- Color Blind Test Book (Pseudo ISO-Chromatic Plate Booklet) ......................... 37
Figure 10- 9 Front Package Labels (PDP) of commercial cereal products ....................... 38
Figure 11- Normal nutrition facts labels of commercial cereal products .......................... 39
Figure 12- Normal Nutrition Facts Labels with PDP for the Control Data ....................... 40

Figure 13: a single example of the traditional NFP Label with PDP for the control data
(70% scales of the actual size) ........................................................................................... 41

Figure 14- Sample images of facial icons for the graphic nutrition facts panels (GNFP).43

Figure 15- Graphic Nutrition Facts Labels with PDP for the Test Data ........................... 44
Figure 16- a single example of icon inserted NFP (GNFP) Label

with PDP for the test data (70% scales of the actual size) ................................................. 45
Figure 17- Pan Tilt eye camera and eye-tracker setting for the experiment ...................... 46
Figure 18- the Layout of Eye tracker setting with a Chin Rest ......................................... 48
Figure 19- The bull’s eye target image for subjects’ eye calibrations ............................... 49
Figure 20- Frequency chart of subjects by age .................................................................. 56

viii

Figure 21- Frequency counts of subjects by gender .......................................................... 56
Figure 22- Subject frequency by visual acuity .................................................................. 57
Figure 23- Subject Demographics: BMI ............................................................................ 58

Figure 24- Frequency of Response to the statement, “Healthy Eating is Important for my
Life” ................................................................................................................................... 59

Figure 25- Bar chart for pairwise comparison of NFP with ZONE as Function of the
probability of noticing a zone ............................................................................................ 63

Figure 26- Bar chart for pairwise comparison of Health Level with Zone as function of
. the probability noticing in a zone ...................................................................................... 66

Figure 27- Bar chart for pairwise comparison of NIF with Zone as function of number of
hits in a zone ...................................................................................................................... 70

Figure 28- Bar chart for pairwise comparison of Health Level with ZONE as function of
the Number of hits ............................................................................................................. 72

Figure 29- Bar chart for pairwise comparison of Gender with Zone as function of number
of hit in a zone ................................................................................................................... 75

Figure 30- Bar chart for pairwise comparison of NIF with Zone as fimction of Time spent
in zone ................................................................................................................................ 78

Figure 31- Bar chart for pairwise comparison of Health Level with Zone as firnction of
total time spent in a given zone ......................................................................................... 81

Figure 32- Bar chart for pairwise comparison of Gender with Zone as function of total
time spent in a given zone ................................................................................................. 83

ix

CHAPTER 1. Food and Health Eating Diet

1.1. Health and Food in Everyday Life

There is a Western proverb, “An apple a day keeps the doctor away.” It means

that good eating helps a person to stay healthy.

Likewise, the Asian proverb “Rice is the best medicine” means the same. Both Proverbs
emphasize the important connection of food and health. If we eat meals with care, it helps

to maintain our health. We all understand that.
1.2. Bad Eating Habits and Risks

The National Health and Nutrition Examination Survey (N HANES) was
performed by the ERS (Economic Research Service), a part of the USDA, in order to
measure “the diet quality of the US.” The survey found that the dietary awareness of

Americans has changed.

During the period of 1989-91, about a half of adults in the US. rated their quality of diet
as very good or excellent. However, in 3 2005-2006 survey, only 33% scored their diets

as very good or excellent, while 18 to 27 % answered their diets are poor or fair. [1]
1.2.1. Obesity and Health Risks

According to the journal ‘Business Wire,’ more than 100 million Americans have
high cholesterol, and nearly 75 million have Diabetes or Pre-Diabetes. Moreover, almost

134 million are “either overweight or obese.” Since two-third of adults in America are

overweight or obese, the obesity problem represents a very significant issue in America.

[2]

Obesity and diabetes are increasing in the U.S., and more than 300,000 American
adults die of obesity-related factors every year. Research has indicated overweight and
obesity classifications are critically associated with “diabetes, high blood pressure, high
cholesterol, asthma, arthritis, and poor health status.” Obesity and adult diseases
(Diabetes, strokes, and cardiovascular diseases) are very significant because they are

major reasons of adult mortality in United States. [3]

Other studies also indicate weight gain and obesity are correlated with an
increased danger of diabetes or other adult diseases. Also, adult people who were
overweight in their youth have more adult disease such as diabetes, strokes, and

cardiovascular diseases. [3]

As such, eating habits developed in youth are very critical to the prevention of
adult diseases because they affect the health status of adults, too. Eating habits need to be
controlled from childhood. Recently, the Obama Administration emphasized the
“importance of health and wellness” and first lady, Michelle Obama and the President’s
special project team announced a recommendation report focused on decreasing the
obesity rate for children in the US. Obviously, the obesity problem is an important

concern in the US. [2]

Between 1977 and 1996, junk food consumption by children increased nearly

300%. Those changes can partially explain “the rise in childhood obesity observed in the

past few years. [4] This is important because as previously suggested, bad eating habits

such as these lead obesity and mortality in adulthood.

Specifically, obesity is now the most prevalent nutritional disease of children and
adolescents in the United States. Although obesity-associated morbidities occur more
frequently in adults, significant consequences of obesity as well as the antecedents of

adult disease occur in obese children and adolescents.[5]

Of course, the obesity problem sometimes results from genetic tendencies,
however, it can also be acquired through poor eating habits and the intake of too many
calories. In the United States, obesity began to garner attention on the national policy
stage after the Surgeon General warned that “obesity had become a new, national

epidemic.” [6]

This was, and continues to be, an alarming trend due to the fact that obesity is
frequently an antecedent to serious diseases. “Increases in obesity and diabetes among
US adults continue in both sexes, all ages, all races, all educational levels, and all

smoking levels. Obesity is strongly associated with several major health risk factors. [3]

The American Heart Association (AHA) has focused on obesity as “a major risk
factor” and has issued appeals for more studies. The AHA report indicates that a 5%-10%
weight loss can decrease many health risks. Weight reduction reduces risks associated
with high blood pressure, elevated blood cholesterol and diabetes. AHA recommended

fat restriction as a good choice for weight loss. [7]

“Obesity greatly increases risks for many serious and morbid conditions,
including diabetes mellitus, hypertension, dyslipidemia, coronary artery disease, and
some cancers. Obesity is clearly associated with increased risk for mortality, but there has

been controversy regarding optimal weight with respect to mortality risk.” [8]
1.3. Risks from 3 Critical Nutrients (Fat, Sodium, and Sugar)

As a result of these realities, how to choose and eat for healthy lives is a concern.
According to one FDA survey, more than half of the public (52%) are concerned about
potentially harmful health effects from the foods they eat. The survey suggested that

many are anxious about their intake of fat, sugars, sodium and etc. [9]

As such, the US. Department of Health and Human Services (HHS) strongly
recommends limits on the intake of fat, salt, and sugars, which have been linked to
cardiovascular disease, hypertension and diabetes. Nutrient information, in particular
these nutrients, which are dictated to appear on the Nutrition Facts Panel (N FP) as a
result of the Nutritional Labeling Education Act (N LEA), should be of paramount

importance during the purchasing process.
1.4. Public and Private Efforts for Healthy Eating

A good diet is important for good health. A healthy and varied diet can help to
maintain a healthy body weight, enhance general wellbeing and reduce the risk of a

number of diseases including heart disease, stroke, cancer, diabetes and osteoporosis. [10]

Absence of health awareness leads to bad eating habits; bad eating habits are

associated with obesity, which, in turn, can lead to mortality that is the result of adult
4

diseases. For all of these reasons, public policies and guidelines for healthy eating are
emerging.

Healthy Eating Guidelines are intended to promote overall health while reducing
the risk of developing nutrition-related diseases, like cancer and heart disease. They are
directed at all healthy individuals over the age of 14. There is nothing difficult about
healthy eating. It is simply a common-sense approach to food that is easy to live with,

once you get used to it. [11]

The US. Department of Health and Human Services (HHS) recommends food
buying plans and has made other guidelines regarding pre-purchase behaviors. For
instance, it is strongly recommend that consumers use the nutrition labels mandated on
most food products. Specifically, it is recommended that consumers, “Know the limits on
fats, salt, and sugars. Read the Nutrition Facts label on foods. Look for foods low in
saturated fats and trans fats. Choose and prepare foods and beverages with little salt

(sodium) and/or added sugars (caloric sweeteners)” [12]

The National Consumer Council (U.K.) made a healthy eating index table for

consumers to formulate their healthy eating plan.

They were given marks out of 10 based on criteria including salt and fat levels in
own-brand foods, nutritional information on packaging, the amount of shelf space given
to fruit as opposed to sweets and biscuits, the presence of sweets at check-outs, and the

availability of advice on healthy eating. [10]

Also, Nutrition Foundation (U .K.) indicates healthy eating tips to enhance

people’s healthy life. They recommend the following, specific guidelines:
5

0 Eat small quantities of these foods
- Choose low fat or reduced sugar foods where possible

- Use spreads and oils sparingly — opt for fats and oils with a high proportion of
monounsaturates and polyunsaturates

- Try to limit consumption of sugar-containing foods and drinks between meals
0 Try not to add fat to foods when cooking
Additionally, this same organization proposes to enable better use of the

information contained on food product labels. [10]

Keeping in step with varying governments’ efforts and consumers’ needs, many
food companies try to advocate healthy diets through ad campaigns and emphasize that

their products are healthy foods in commercials.

Chapter 2; Ready to Eat Cereals for Healthy Diet

2. 1. Cereals for Breakfast

Ready to eat cereal (RTEC) is one of the most popular foods for breakfast around
the world. According to one food industry newsletter, most people eat cereals for

breakfast, and, as a result, it is the most commonly consumed breakfast item. [13]

Public institutions have emphasized the importance of breakfast as part of a
healthy diet, so the importance of the choice of cereal should not be underestimated. The
UK’s British Nutrition Foundation (BNF) recommends that people have breakfast
everyday and has suggested that cereal is an item that can be added to breakfast as a good
source of nutrients. [10] The UK is not the only country that has officially recognized the
importance of breakfast. The US. Departments of Agriculture (USDA) and Health and
Human Services (HHS) have indicated that people should eat breakfast as part of their
healthy guidelines and that breakfast is an important part of a healthy diet. Additionally,

they recommend breakfast cereals because many cereals offer whole grains. [14]

As a result, it is not surprising that the consumption of cereal products is
increasing. In China, demand and consumption of cereal is growing and this trend is

expected to continue. [15]

According to the New York Times, the cereal consumption rate of young adults
and teenagers is increasing in the US. and college students have been documented to eat
cereals for lunch or dinner, as well as breakfast. In the Times survey, college students

answered that they eat cereal because it is easy and cheap. Also, “NPD Food-world” in
7

Chicago reports that young people consume more cereals than before because they have

regularly consumed cereals since they were toddlers. [16]
2.2. Cereal as a Functional Food

2.2.1. Various Positive Effects of Cereals

Cereal is one of the important sources of energy for the day. Since ready to eat
cereals have the potential for fluent nutrition, individuals who eat cereal products
regularly have been reported to be healthier than those who consmne them less frequently.

[17]

Cereals represent simple, fast and tasty meals (or snacks) to consumers and they

frequently have healthful nutrition contents. [18]
1. Physical Health Benefits Associated with Cereal Consumption

As part of “the public health nutrition program of UC Berkeley”, Gladys Block,
PhD, said that cereal eaters usually eat fewer amounts of fat and cholesterol. Meanwhile

they eat more fiber per day. Cereals eaters are likely to make smart choices at other

meals.” [19]

Some research suggests that, the benefits of eating cereals regularly extend

beyond nutritional gains, into reducing the chance of diseases.

Cereal usually offers fiber, calcium, iron, folic acid, vitamin C and zinc, as well as
a range of micronutrients such as vitamin E, some of the B vitamins, magnesium and zinc.

Those contents are critical health maintenance. [20-21]

8

Research has indicated that people who ate cereals consume more fiber, calcium,

folic acid, vitamin C and zinc and less fat and cholesterol. [20]

Vitamins, minerals, fiber and oil in cereal play an important role in lowering
“cardiovascular risk factors” and, in 2007, a health study about the relationship between
“Breakfast Cereals and Risk of Heart Failure” suggested that “a higher intake of whole

grain breakfast cereals is associated with a lower risk of Heart Failure.”[22-23]

Fiber, found in many cereals in the form of whole grains, has a preventative
effect on chronic diseases such as “coronary heart disease, diabetes and colorectal cancer.”
[21] The American Diabetes Association recommends that people eat “20-35 grams of
fiber” everyday. Eating this amount of fiber each day has been indicated to help to

decrease cholesterol levels and reduce the risk of colon cancer. [18]

Many cereals, particularly the bran cereals, which are packed with fiber, reduce
the risk of “constipation, diverticulosis, heart disease, and possibly cancer.” [24] Cereals
which have “oat” contents have been indicated to lower cholesterol levels because the
soluble fibers in oats have “beta glucan.” The beta glucan content may help to reduce

cholesterol levels. [25]

One Italian study announced that ready to eat cereals which contain “Inulin”
potentially lowers lipid levels in the blood, reducing the risk of diseases related to blood

vessels. [26]

Whole grain cereals have also been linked to the prevention of type 2 diabetes in
women. Consumption of cereals containing whole grain may reduce “inflammatory

markers among diabetes patients.” [27]
2. Maintaining Body Weight

Body weight is very important to maintain one’s health because overweight (or
obesity) is positively correlated with many diseases. Millions of American people are

dieting to reduce their weight because of health concerns. [28]

Research has suggested that cereal is also helpful to dieters. According to studies
which focused on the dietary effects of cereal products, ‘whole-grain’ breakfast cereal is
helpful to managing one’s weight. [29] Research suggests that people who regularly eat
cereals manage their weight more successfully than people who don’t. [30] Furthermore,
according to a study conducted at Wayne State University, people who consumed “cold-
cereal with low fat milk after dinner,” ultimately consumed fewer calories than people

who did not eat cold cereals.[31]

Recent research of the American Diabetic Association supports the relationship
between cereals and health. The research found that consuming oat cereals helps weight

loss in both genders with obesity. [32]
3. Better Mental Activities

Studies performed in children have also suggested that breakfast may help with
mental performance (memory and attention). In one study, groups of children that had no

breakfast or glucose drink for breakfast exhibited declines in attention and memory;
10

however, the memory and attention declines were decreased among the children who had
cereals. This research indicates “a typical breakfast of cereal rich in complex
carbohydrates can help maintain mental performance over the morning.” [33] A larger
study also suggested a relationship between eating breakfast cereal and health. Adults
who ate cereals every day for breakfast maintained “better mental and physical health,
compared to those who consumed it less frequently.” In the research, perceived stress,
fatigue, anxiety, depression levels were remarkably low among the subjects who ate

cereals daily. [34]
2. 3. Choosing Better Cereals for Health

It is helpful that consumers choose the right cereals based on their health status.
Cereals may contain inappropriate levels of fat, sugar or sodium while other cereal
products are rich sources of fiber, vitamins and minerals like iron, magnesium and
selenium. For this reason, choosing right kind of cereal provides nutritional advantages

and helps to reduce many chronic diseases with adult onsets.[35]

Nutrition experts have recommended purchasing “flaky or puffy cereals” rather
than those that are the “dense, nugget” type because the former typically have far fewer
calories. Also, it has been recommended that consumers should look for “whole grain”

cereal for health diet. [24]

“Whole grains” are rich in fiber content and have been shown to prevent many

diseases.

11

There are two kinds of fibers; insoluble fiber and soluble fiber. The former is rich
in ‘Bran’ and the latter is high in ‘Oat.’ The insoluble fiber helps to prevent “constipation.
Diverticulosis, and heart diseases.” Moreover, animal studies found that insoluble fiber
may reduce colon cancer. The main effect soluble fiber is that it helps “lower cholesterol

and lower risk of heart disease.”[24]

Because of the potential positives of consuming the right kind of cereals, it is
important that consumers understand the nutrient content of what they are choosing. This
make the information contained within the nutrition facts panel (N FP) especially

“calories, fiber grams, and sugar grams” particularly important. [35]

12

CHAPTER 3. Nutrition Fact Panels

3.1. Nutrition Facts Panel

3.1.1. Importance of Nutritional Labeling

A primary vehicle for supplying point of purchase nutrition information to

consumers is “Nutrition Labeling.” [36]

In fact, research conducted in the United Kingdom (UK) suggests that the major
source of nutrition information that is used by consumers comes from labels. Studies
published in the United Kingdom indicate that 30 % of consumers always read the food

labels. [37]

The importance of nutritional labeling to consumers has also been studied in the
US, where product labeling is required to conform to FDA and/or USDA requirements

for the vast majority of food items sold. [3 8]

In a study conducted in 2009 by the International Food Information Council
Foundation, consumers were asked about “the type [of] information” they find on the
package of food and beverage when they decide to purchase food or beverage products.
According to the survey, the “Nutrition Facts panel (NF P)” was the most common label
element used by consumers to get nutrition information. The other factors that consumers
focused on were the “Expiration Date”, “Brand name”, “Ingredients”, “Size of Product”,

“Nutrition Statements”, “Statement about Health Benefits.” [3 8]

13

 

What information do you look for on the food or
beverage package when deciding to purchase or eat
a food or beverage?

Select all apply.

2009(n=1064)

Nutrition Facts Panel
Expiration Date
Brand Name
Ingrdients
Size of Product
Statements about Nutrition
Statements about Health Benefits
Country of Origin Labeling
Organic ~ *

Allergen Labeling

Health Symbol or Health Icon E

None of the Above

    

 

 

 

 

 

l

0% 20% 40% 60% 80%

 

 

 

Figure 1- Use of information on food and beverage packages; Data from “International

Food Information Council Foundation food and health survey” [3 9]

According to research conducted by the International Food Information Council
Foundation in 2006, health conscious consumers use the NFP information more

frequently. [40]

It has been suggested that the importance of the role of the “Nutrition Facts” label
is increasing because reading (and using) the information contained on these labels can
help to prevent chronic disease and obesity; an issue of growing concern around much of
the globe. Americans are showing an increasing interest in the link between food and

health, nutrition, diet and the prevention of chronic disease. [41]

14

According to a recent report of the US Food and Drug Administration (FDA), the
percentage of people who read nutrition labels before buying food products reflects this
increased interest. In 2002, 44% of people indicated that they read nutrition labels before

buying a product, where the number had increased to “54% in 2008.” [41]

F urtherrnore, research at the University of Nebraska, which gauged consumer
attitudes, indicated that the presence of a nutrition facts label in dining halls impacted the
food choice on campus and outside. Afier posting the nutrition contents in the dining
halls, food consumption was reduced. The nutrition facts labeling is very important

because it has been linked to consumption behaviors as well as purchase decisions. [42]
3.1.2. General FDA Regulations of Nutrition Facts

In the US, the Nutrition Labeling and Education Act of 1990 (N LEA) has
required food manufacturers to provide nutrition information in a standardized format in
a truthful and complete manner. Regulated products were required to have this labeling
in place by the May 1994 deadline imposed in the associated regulations. The goal of this
requirement was to reduce the negative effects of untruthful and exaggerated claims as
well as improve the accessibility of nutrition information at the point of sale so as to

increase the extent to which consumers could process and use it in their choices. [43]

FDA Regulations:

Details regarding the standardized content and formatting of information can be

found in 21CFR 101 .9(d).

15

The regulation indicates that there are no size requirements for the “Box” which
contains the nutrition facts information. However, the letter size of the "Nutrition Facts"
heading must be larger than all other fonts that appear on the nutrition label. Also, a
minimum type size of 6 and 8 points are required for the various pieces of required

information, and there are minimum spacing requirements between lines of text. [44]

Overall Rules:

The Nutrition Facts label is boxed into what is referred to as the “Nutrition Facts
Panel,” which is located to the right of a package’s principal display panel (PDP). The
‘NFP’ is placed with lists of “Ingredients” and “Name and Address (name and address of

distributor, manufacturer and packer). (See Figure 2)

\

Nutrition Facts

 

 

PDP

 

 

 

- 4— lnforamtion Panel

Ingredients:

 

Manufactured
By:

 

 

 

Figure 2- “A Food Labeling Guide” by US. Food and Drug Administration” [44]

16

The box and information are presented in all black or one color type printed on a

white or neutral background.

Tmetace and Size:

The letters in Nutrition Facts Panel use six point or larger and it is suggested that
information appears in “Helvetica Black or Helvetica Regular” fonts. Key nutrients and
their % Daily Value must be presented in a minimum of 8 points Helvetica Black (but
"%" is set in Helvetica Regular).

1. “Nutrition Facts” is set in either Franklin Gothic Heavy or Helvetica Black to fit the
width of the label flush left and flush right (a justified format).

2. “Serving Size and Servings per container” are set in 8 point Helvetica Regular with
1 point of leading.

3. The “table labels (for example, ‘Amount per Serving’)” are set in 6 point Helvetica
Black.

4. “Absolute measures” of nutrient content (for example, "1g") and nutrient subgroups
are set in 8 point Helvetica Regular with 4 points of leading.

5. “Vitamins and minerals” are set in 8 point Helvetica regular, with 4 points of
leading, separated by 10 point bullets.

6. All type that appears under vitamins and minerals is set in 6 point Helvetica Regular
with 1 point of leading.

Box:

All labels are enclosed by 1/2 point box rule within 3 points of text measure. [44]

3.2. Efforts for Nutrition Labeling Development

As mentioned previously, it has been suggested that “consumers are becoming
health conscious and most agree that eating healthily is better way to manage illness than

using medications” [45]

3.2.1. Motivations of the Better Labels

17

Despite the growing importance of nutrition facts labeling, research suggests that
consumers are struggling with interpreting the information that they provide. Nutrition
facts have to be read easily, however, this labeling has been called “confusing,”

“misleading,” and even “inaccurate.” [46]

According to a survey of the Wall Street Journal, even “well-educated”
individuals who have “the good reading and interpreting skills” have difficulties

understanding the label data. [47]

Moreover, a study published in the “Journal of Preventive Medicine,” suggested
that comprehension errors of labels occurred in populations of not only “lower literacy
and mathematic skills” but also “better-educated groups.” In other words, the members of
well-educated group failed to understand the label information as well as those with
diminished educational backgrounds. In the study, the authors report that only 37 % of
test subjects calculated “the total grams of carbohydrates in a 20 ounce bottle of soda”
correctly. In the case of a whole bagel, 40 % of participants couldn’t calculate “how

many grams of carbohydrates they would consume if they ate only half a bagel.” [47]

Obviously, many consumers are confirsed by nutrition facts information. As a
result, efforts, both public and private, have been started to make labels that are better

understood.

The majority of efforts seem to take one of two approaches (1) to create labels
that effectively warn and are easily comprehended by consumers (2) to promote a food

product highlighting the healthful content of that product. Studies of the former type are

18

mostly carried out by public institutions or consumer groups, while those of the latter

focus are mostly developed by food companies [48]

Globally, there has been a movement to change the formatting of nutritional
information. The use of “Health Claims” and varied schemes for Front of Package (FOP)
Labeling are good examples of this movement. New formats rely heavily on the use of
color cues and icons, combined with text, to convey nutrition information or imply claims

of health.

3.2.2. Efforts for Nutrition Label Development

I. Health Claims

Need at Health Claims

Health Claims are an example of a nutrition-profiling tool; Nutrition profiling is a
method for “categorizing the healthfulness of foods” based on their nutrition contents.
[49-50] When food manufacturers attempt to communicate the “health related properties
of a food”, this can be regarded as a “health claims.”[5 1] See below for example images

of health claims.

 

Figure 3- Health Claims
19

Functional Food and Health Claims;

The term “firnctional food” means that the food provides various benefits ‘beyond
basic nutrition.’ People can get the various nutrients from functional foods which help
their nutrition diet and it may lead not only deliveries of the good nutrition as the whole

food itself but lowering the rate of various disease occurrences. [52]

Because of the demand for “functional foods” and their promise of healthful
benefits, the number of food product carrying “health and nutrition claims” is growing.
[49] Various health claims are seen on the labeling of many food products throughout the
world and each health claim is one of the important tools for the marketing of functional
foods.[45]Therefore, companies are using health claims (scientific evidence) to raise the

marketing benefit. [49]

Clearly, inserting a health claims can improve “the quality of dietary choices and

knowledge of diet-disease relationships.” [53]

Regulations to Protect consumers

Although printed health claims must be based on “scientific evidence” [49],
sometimes health claims are based on dubious scientific information. Put simply,
untruthful statements are used as health claims at times. Early in the history of label-
located health claims, the value of the claims made by food companies was very low.
Many, for instance, put ambiguous claims on their package labels to appeal to consumers’

desires for healthful products, creating the potential to mislead consumers. [48]

20

It has been reported that most consumers do not precisely comprehend ‘labeling
information;” however, information, in the form of a health claim, has been reported to
change “consumers’ perceptions toward specific food products.” [54] Therefore, the
potential for misleading consumers exists. As a result, regulations regarding health claims

have been prepared as preventive measures by public institutions.

Prior to 1984, in United States, “health claims for food products” were banned by
FDA. In 1990, under “the Nutrition Labeling and Education Act (N LEA)”, FDA issued
regulations authorizing ‘Health Claims’ for dietary supplements after reviewing the

scientific evidence submitted in ‘health claim petitions’.[5 5]

After that, the first trial of a health claim (located on the package) was made by
Kellogg Company. Kellogg’s suggested the health benefits of certain cereal products as

agents of “high fiber content and lower risk of cancer.” [43]

By 1990, food labeling was an important policy issue. The US Food and Drug
Administration (FDA) developed “a new regulatory regime for health claims” as well as
regulations under the directive of the “Nutrition Labeling and Education Act of 1990.”
Although the regulation of food labels that are subject to NLEA requirements has been
relatively stable, arguments regarding the appropriate use of ‘health claims’ are still

ongoing. [53]

During this same period of time, the FDA further reformed nutrition labeling rules
by creating regulations that governed the use of “Free Claims.” They suggested ‘free’
claims used on food products that are “inherently free of a nutrient” are misleading unless

accompanied by a statement that all foods of that type are inherently free of that nutrient.
21

(Federal Register 1993) For example, a head of broccoli could not be labeled as “fat free,”

but instead had to be labeled as “broccoli, a fat free food.”[56]

The US. has not been the only country involved in discussions of health claims.
In 2003, the joint agency Food Standards of Australia and New Zealand (FSANZ)
attempted to standardize health claims to prevent them fi'om being misleading. Currently,
“the food standard cord (Standard 1.1A.2)” allows the “nutrition content claims” because
they are claims for maintenance. However, “disease reduction claims, (with the exception
of those directed at the consumption of folate (folic acid), a substance known to reduce

the risk of neural tube defects)” are prohibited. [45]

F SANZ implemented “the Claims Framework in the Policy Guideline” and

classified three types of claims. See below 3 levels of claims.

0 Nutrition content claims

; Statements regarding the amount of a nutrient, energy or a biologically active

substance in the food

0 General level health claims

; The presence of a nutrient or substance in a food and its effect on a health

function

0 High-level health claims

; Claims that make reference to a serious disease or biomarker

22

Health claims on the labeling of food products are becoming more prevalent in
many countries and there is some evidence that the claims of health give a positive effect

on healthy choices when consumer wants to purchase premium food. [45]

Concerns about Health Claims

According to a recent survey of “the global regulatory environment for health
claims performed by World Health Organization (WHO)”, a majority of countries do not
regulate these claims. Among those that report regulation of said claims, requirements

vary significantly. [53]

Moreover, consumers, particularly those unfamiliar with nutritional information
and those that are illiterate have great difficulty. To prevent such problems, education
regarding health nutrition and disease and efforts to make international regulations are

necessary. [57]

Although it is touted that ‘Health Claim’ labels have the potential to
affirmatively impact food choices, research is still needed.

The claims mostly include healthy nutrition contents or health support
declarations. Sometimes they can lead to misunderstandings or actually misinform
consumers, especially if the claims are not created by public institutes. Claims like “good
for health”, “healthy” or “better choice” are not clear expressions to the consumers. For
that reason, FDA recommends that consumers continue to check the Nutrition Facts
Panel and read and rely on the percent of ‘Daily Value’ content to avoid confusion and

misinfonnation. [5 8]

23

Examples; Claims for Public Health

I The ”Pick and Tick" symbol
The Heart Foundation of New Zealand created “pick and tick” symbols. In general, the

“pick and tick” symbol is approved on food packages which contain foods lower in “total
fats”, “saturated fat”, “added sugar” and “sodium.” The mission of this symbol is “To
improve the nutrition health through a food information program which encourages a

healthier food supply.”

 

Figure 4- “pick and tick labeling” in New Zealand [59]

I The “Green Keyhole” symbol
The National Food Administration in Sweden introduced the symbol “Green keyhole”
during a nutrition campaign in 1989. The symbol is made to give health information on a
package label and the symbol indicates “low-fat and high-fibre foods” The aim of the
“Green Keyhole” campaign is to help consumers make “low-fat, high-fibre food
choices.”[60]

24

 

the

bee.

the

but

 

 

 

 

Figure 5- Green Keyhole labeling [60]

I The “Heart” symbol

In 2000, the Finnish Heart Association and Finnish Diabetes Associations introduced
the ‘heart symbol,’ a symbol which indicates that the food is a better choice for health
because the food has less fat and salt. [61] The “protect heart symbol,” of Slovenia, gives
the information not only regarding “the amounts of fat, cholesterol, sugars and sodium

but [also] the amount of fibre and energy.” [62]

 

Figure 6- “heart symbol” [62]

II. Front of Pack [FOP] Labeling

Front at Pack (FOP) label

25

Front of package (FOP) labeling is a relatively new approach to labeling which
places a summary of key information from the more complete nutritional information (e.g.
the Nutrition Facts Panel in the US) on the front panel. “Nuval” and “Nutrition

Highlights” in cereal products are the examples of FOP labeling. [5 8]

It has been suggested that many consumers think that FOP labeling is helpful
because they do not need to use the “Nutrition Facts” section. However, concerns exist
that new FOP systems may result At present, the FDA is gathering information for
consideration in order to standardize an FOP labeling system that people can fully trust

and adopt. [58]

T mes ot FOP Labeling

Several types of FOP systems are currently commercial. [63]:

I “Fact based systems provide FOP information that brings to the front panel
information that is provided in the NFP (nutrition facts panel)”

I “Better-for-you systems use symbols to indicate how a product ranks against a
defined set of nutritional criteria.”

I “Graded better-for-you systems also use symbols, but with indicators to convey
good, better, and best nutritional quality.”

I “Numerical rating systems use numbers to rank the overall nutritional quality of
a food or beverage.”

l “Color-coded systems use colors to provide at-a-glance information about the

levels of individual nutrients in a food or beverage”

26

Many of the FOP systems that have been introduced around the world in the recent
past incorporate ‘graphic symbols’ in an attempt to garner consumer attention and aid in

their understanding of the nutritional aspects of a product.

Suggested FOP labeling;

1. Guideline Daily Amounts (GDAs) labeling

In 1996, Guideline Daily Amounts (GDAs) were developed as “Daily Guideline
intakes (DGI)” by the UK Ministry of Agriculture Fisheries and Food (MAFF), now were
the Food Standard Agency (FSA). In 1998, a set of GDAs for food labels were suggested
to communicate the Govemment’s nutrient intake recommendation. GDAs were created
in order to help consumers’ comprehension of nutrition information on a package [64].

GDAs have been used for about a decade by many food companies in the UK.[48]
GDAs offer consumers the information regarding a product’s “energy, fat, saturated fat,

carbohydrate, total sugars, fibre, sodium, salt and protein”.[65]

Recentrly, the GDAs labeling is spread throughout EU countries. In 2006, the
Confederation of food and Drink industries (CIAA) introduced an EU set of GDAs which
is based on Eurodiet recommendations and, in 2010, the Members of the European
Parliament (MEPs) supported GDAs are the best way to get nutrition information when
consumers purchase food products. In July 2008, according to a research performed by

APCO Worldwide Insight, many food companies in EU voluntarily inserted GDAs across

a broad range of foods. According to the research conducted based on food companies in

27

France, Italy, Spain, the United Kingdom and the Netherland, nearly half of all

respondents (44%) are voluntarily labeling their products with GDA information [64].

See Table 1 and 2 for GDAs for adults and children.

Table l- “GDA Tables for Men and Women”; Institute of Grocery Distribution (2008) [66]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NUTRITION INFORMATION GUIDELINE DAILY
AMOUNTS
Typical Values 11(3):); 2:31;? Women Men
Energy — kj 480kj 1680kcal - -
-kcal 1 15kcal 405kcal 2000 2500
Protein 9.5g 33.3g 45g 55g
Carbohydrate 8.6g 30.1g 230g 300g
Of which sugars 2.0g 7.0g 90g 120g
Fat 4.6g 16.1 g 70g 95g
Of which saturates 3.0g 10.0g 20g 30g
Fibre 1.5g 5.3g 24g 24g
Sodium* 0.3g 1.1g 2.4g 2.4g
*Equivalent as salt 0.8g 2.8g 6g 6

 

28

 

Table 2- “GDAs for children” Institute of Grocery Distribution (2008) [66]

 

 

 

 

 

 

 

 

 

 

 

 

 

GUIDELINE DAILY AMOUNT
5- 10 years

Energy (Calories) 1800
Fat (g) 70
Saturated Fat (g) 20

Carbohydrate (g) 220
Total Sugars (g) 85
NMES (g) 50
Protein (g) 24
Dietary Fibre (AOAC) 15
Dietary Fibre (N SP) 11
Sodium 1.4

Salt (g) 4

 

 

 

 

GDAs are based on “the estimated daily intake of an average individual eating a
diet” from ‘Dietary Reference Values’ report performed by Committee on Medical

Aspects of Food Policy [67]

2. Traflic Light System Labeling in U.K.

An FOP labeling system that is garnering a lot of discussion and attention is the

“Traffic light labeling” system, used in the UK.

29

The Food Standard Agency (FSA) in the U.K. created “Traffic Light Labeling,” a
system that employs both colors and text to convey information regarding key nutrients.
As with most FOP systems, the intention of traffic light labeling is to induce a healthy

diet by providing nutritional information to consumers in a format that is easily read and

understood. [48]

Traffic light labels are basically an extension of GDAs (Guideline Daily Amount)
labeling. While “GDAs” usually indicate the amount of nutrition using a percentage

format, “Traffic Light” labeling uses Colors. [68]

30

 

 

7. 7gF per Aservlng

 

 

09 per serving

 

 

.. SUGAR
'7 42. 29 per serving

. SALT
2. 0g per serving
L

LOW LOW :
FAT SAT FAT SUGAR: SALT‘

sArunATES:)
2

 

 

 

 

 

 

7.79 2.09 42.29 2.09
per serving per serving per serving per serving

 

 

 

Figure 7- “Traffic Light Labeling” Food Standard Agency, U.K. [69]

FSA standards recommend that the Traffic Light system include the amounts of 3
nutrients (health fat, sodium and sugar) per 100g serving. These nutrients were selected
for inclusion in the FOP format because of their links to detrimental health effects. Color
serves as a redundant cue for the information; red, amber and green light levels,

correspond with high, moderate and low amounts of the nutrients, respectively. [48]

Cutoffs for each of the nutrient levels (Fat, Sodium and sugar) are posted on the

31

“U. K. Food Standard Agency” webpage; See Table 3.

Table 3- Levels of Nutrition Contents (Fat, Sodium, and Sugar)

 

Density Level FAT (I100g) SODIUM (/100g) SUGAR (/100g)

 

High (Red) More than 20g More than 1.5g More than 12.5 g

 

more than 3g & more than 3g & more than 5g &

Medium (Amber)
less than 20g less than 1.5g less than 12.5g

 

Low (Green) less than 3g less than 0.3 g less than 5g

 

 

 

 

 

 

Research investigating the “Traffic Light” labeling performed by FSA U.K.
suggests that consumers showed strong reaction to the number of “reds” presented on a
product. Simply put, consumers have a tendency to look for products that contained
“fewer reds” as opposed to “more green.” Since people regard the “red” color as
cautionary, warning of danger, the effect of the “red” color was noted to be bigger than
“green” or “amber”. Also, the report states that, “when looking at a single product, the
number of ‘reds’ or ‘greens’ often influenced their decision. In some cases, the presence
of any red provoked further investigation. In other cases seeing more than two reds may

lead to a rejection. [63]

As such, the use of the, “Traffic Light” labeling system has potential to

significantly impact behavior. However, eating a varied and balanced diet is the goal;

32

some “foods which are considered HF SS [high fat, sugar and salt] also contribute a wide
variety of beneficial nutrients to their diet and can be consumed as part of a healthy
balanced diet.” Currently, the “traffic light” system is being voluntarily inserted in these
categories of food: sandwiches, wraps, filled baguettes and similar foods, ready meals,
burgers, sausages, pies, pasties, quiches, breaded, coated or formed meat products and

alternatives, pizzas and breakfast cereals. [63]
3. 3. Opinions for Label Development

As mentioned above, there are many efforts related to label development, using

various formats.

Consumers usually do not use nutrition labels because of “lack of interest” and
“lack of time.” According to research completed by McIlveen and Semple (2002), people
who didn’t read food labels commented that they didn’t read labels because they can’t
understand the information on the labels. Lack of interest can lead to consumers that

don’t read nutrition labels and they cannot improve their diet. [70]

Literacy level affects the ability to read, interpret and understand nutrition labels.
Also, graphic presentations have been indicated to be helpful to comprehension rates in

consumers with low literacy. [57]

33

cl;

C0

Ht

012.

in

4.1

 

 

CHAPTER 4. Materials and Apparatuses

In this chapter, the first section provides a description of the materials used
during the course of study, and the second includes a description of the eye tracking

apparatus and the settings employed.

A convenience sample of 24 subjects was recruited via word of mouth and an
advertisement (see APPENDIX 2) that was announced during a required undergraduate

class at the school of Packaging and through the employee pool at MSU.

In order to participate, subjects had to be 18 years of age, could not wear hard
contact lenses, or be legally blind. In exchange for participation, subjects were provided a

$10 Best BuyTM gift card.

Upon arrival, after informed consent was obtained, subjects were assigned a
number; recorded data is tied to this number, not subject name. Subjects began by filling
out survey questions which included: demographic information, information about
healthy eating and consumption patterns (frequency, preferences, and etc.), and self-
reports of personal information (e. g. age, gender, height and weight and etc). See

APPENDIX 3 for more detailed information.

4.1. Materials

Visual Acuity Test Card (Dow Corning Ophthalmics Product)

34

Following the short survey previously described, each subjects’ visual acuity
was tested a using a Dow Corning Ophthalmics Inc. (Norfolk, Virginia 23508) Near

Point Visual Acuity Card (see Figure 8).

The dimensions of the card are ‘7"(w) x 4"(h)’ and the size of printed letters

moderately decreases from the top to the bottom of the test card.

Subjects were tested in good light and instructed to hold the card 16” fiom their
eyes and read the lowest line that they could decipher. In accordance with Dow’s
instructions, the visual acuity recorded for the subject coincided with the lowest line on
the card where they were able to correctly report all letters; the potential range of

recorded acuities was 20/20 to 20/ 120.

 

Figure 8- Near point Visual Acuity Card, Dow Corning Ophthalrrrics (Left; front, Right;

back)

35

n!
a? i

1313

 

Visual acuity was measured with corrected eyesight. Subjects wore corrective

eyewear that would normally be required for reading tasks during the visual acuity tests.

Color Blind Test Book (Pseudo Isa-Chromatic Plate Booklet)

Immediately following the test of visual acuity, each subject’s ability to see color
was testing using the Pseudo ISO-chromatic Plates Booklet manufactured by Richmond

Products (Albuquerque, New Mexico).

This blue covered booklet has 14 plates of numbers which screen for the ability
to see red and green colors (see Figure 9). Each subject was asked to read the number
printed on each page of the book. Subjects were allowed to view each stimulus for
approximately 2 seconds before responding to a given test plate. In accordance with
manufacturer’s instructions, when the number of correct responses was 2:10 (out of a

possible 14 test plates), the response was recorded as “Normal” color vision.

36

d.

OD
WE

NF

 

Figure 9- Color Blind Test Book (Pseudo Iso-Chromatic Plate Booklet)
Labels for Eye Tracking Stimulus (NFP labels with PDPs)

Nine, commercially-available cereals were selected across a range of nutritional
densities (See figure 10). The corresponding Nutrition Facts Panels (N F Ps) were
modified, resulting in two levels of presentation for the eye tracking stimulus: (1) text
only (the commercially-available format) and (2) facial icons inserted. As such, there
were a total of 18 unique files to be viewed by subjects (9 cereals x 2 presentations of

NFP).

37

 

RMSEN

BEER

I

'--' mt mar. (P0515 1313 7A;

 

 

 

 

Figure 10- 9 Front Package Labels (PDP) of commercial cereal products

1. Text-only (Commercially Available) NFPs

38

Figure 11 shows the nine NFPs that corresponded to the cereals selected for use
in the study. 9 PDPs and NFPs were scanned at a resolution of 300 dpi. The PDP was

saved as a JPEG picture format.

Each cereal’s corresponding NFP was scanned at the same resolution with the
same scanner and sized to 2500 pixels by 3600 pixels, and, as with the PDP image, saved
in a JPEG picture Format. The Scanned images were then combined (see Figure 12 and
13) using Adobe Illustrator CSZ® (San Jose, CA), and saved as a single JPEG image in
which the PDP was on the left hand side of the computer screen with the NFP to its
immediate right. In the Adobe Illustrator CS2® Software, a border was created to flame
and unify the size of all slide images. Therefore the size of all PDP +NFP, combined

images were 3184 pixels (X) by 3601 pixels (Y), including into the box flame.

 

ll

it

‘ It iii] {I

 

 

 

 

 

   

 

 

 

Figure 11- Normal nutrition facts labels of commercial cereal products

39

 

 

Past ,. ..
. ‘ l, - m .1
gym (ll/(1 /‘

I bra/10'4”" I

Cheerios

 

 

Figure 12- Normal Nutrition Facts Labels with PDP for the Control Data

40

 

 

Figure 13: a single example of the traditional NFP Label with PDP for the control data

(70% scales of the actual size)

41

2. Icons inserted NFPs : GNFP
Scanned images of the NFPs were modified using Adobe Illustrator CS2® (San Jose,
CA), to create the second level of presentation (iconic) for the NFPs. Iconic or graphics

were created for three nutrients, Based on the critical role that they have been identified

to have in lifestyle diseases of importance in the US (heart disease, hypertension and

diabetes). See Chapter 1 for more information about health risks caused from Fat,

Sodium and Sugar.

Table 4- the standard quatity levels of 3 Critical Nutrition

 

 

 

 

 

 

 

 

 

Density SODIUM SUGAR FACIAL
Lever FAT (/ 100g) (/100g) (/100g) ICON
. More than
ngh More than 20g More than 1.5g 12.5g FROWN
more than 3g more than 3g more than 5 g
Medium & & & NORMAL
less than 20g less than 1.5g less than 12.5g
Low less than 3g less than 0.3g less than 5g SMILE

 

(http://www.eatwell.gov.uk/healthydiet/fss/)

42

*Data in Table 4 founded in U.K. Food Standard Agency website.

 

Using Illustrator CSZ® software, the researcher designed facial icons intended to
represent three levels of nutrient density. (3 levels: Smile, Normal, and Frown faces) See

the examples of facial icons below.

 

Sodium Sugar

 

 

Figure 14- Sample images of facial icons for the graphic nutrition facts panels (GNFP)

NF Ps were modified by adding these facial icons set based on the nutrient levels
(fat, sodium and sugar) present in the specific products, just below the calorie report in

the traditional NFP.

See figure 15 for a presentation of all 9 cereals with modified NFPs and Figure

16 for a single example. (70% scale of the actual size)

43

 

 

A‘ "
. ,. ‘ , airsrrrs
.- 1.];)/)/(‘(,.{(M”{‘r Past,
v.3. ”Um/rt!
1:” ,1 Rarsrtr
*,,._ . eaau

u‘m irilif. m {ill-l TM“!

nary/16m

Cheerios '1

 

 

Figure 15- Graphic Nutrition Facts Labels with PDP for the Test Data

 

 

Figure 16- a single example of icon inserted NFP (GNF P) Label

with PDP for the test data (70% scales of the actual size)

45

4. 2. Apparatuses

The ASL Pan Tilt Optics Eye Tracker (Model 504)

In order to test for an effect of iconic format on the noticeability of information
related to the select nutrients, an Applied Science Laboratories (Boston, MA), Pan Tilt

Optics eye tracker (ASL Eye-Tracker model 504) was used. (See Figure 17)

 

Figure 17- Pan Tilt eye camera and eye-tracker setting for the experiment

46

This eye-tracker is designed to accurately measure a person’s pupil diameter and
point of gaze on a stationary (room fixed) scene space. The followings are the

explanations of apparatus features and operation.

Major Component and Operations: The pan tilt eye camera’s infrared LED ring
a (round the eye camera lens) provides eye illumination. The beam from LED illuminates
the pupil and creates a small bright spot of light (known as corneal reflection). The Pan
tilt provides 25° of tilt capability and 100° of pan angle. This pan tilt eye camera optic

module can be controlled by remote controller.

Eye Calibration and Data Recording: The reflections of corneal and pupil data
were saved when subjects watched the 9 points on the bull’s eye target screen. After the
eye calibration, the gaze points of subject during a slideshow consisting of the 18 images
of cereal boxes could be recorded in Eye Response® Gaze Tracker® software (Boston,

MA).
Chin Rest

To rrrinimize unconscious movement and to get more accurate data from Pan Tilt
eye-tracker, the researcher asked subjects to rest their heads on a chin rest, fixing the eye
relative to the scene. The height of the chin rest can be adjusted and research did not

begin until a subject was comfortable. (See Figure 18)

47

 

Figure 18- the Layout of Eye tracker setting with a Chin Rest

48

Bull’s Eye Target

The Bull’s Eye Target image was set as a desktop background picture and it was

used to calibrate each subject’s eye before eye tracking. See Figure 19.

 

Figure 19- The bull’s eye target image for subjects’ eye calibrations

49

CHAPTER 5. Methods

5. 1. Working Hypothesis

0 Nutrition information presented in varied formats (iconic and text) is more notice
able than text-only formats of the same.
0 Attentive behaviors are influenced by a confluence of factor related to the consu

mer. (Age, Gender BMI and etc.)

5. 2. Experiment Procedure

Each subject who participated was tested using the procedure below. The
experiment took less than an hour per subject. The sequence of experiment provided

below.

Consent Form:
A consent form for the experiment was provided to each participant and all the
participants had enough time to read the consent form. A copy of the consent form can

see in the APPENDD( 1.

Additional Explanation
If any subject needed a more specific explanation about the research experiment
or the instrumentation, the researcher provided a verbal explanation and answered all

questions.

Visual Acuity Test

50

Before Eye Tracking began, a visual acuity test was performed to qualify their

eyes’ ability to resolve images (i.e.:20/20, 20/30, 20/40 and etc).

The researcher held the visual acuity test card on a wall in front of the subjects.
The distance between the card and a subject was controlled as 16” per the test method
instructions. Subjects were asked to read from the upper to the lower letters until they
were unable to decipher the line of text. The subjects’ visual acuity coincided with the

lowest line where they could decipher all letters correctly.

Color Blindness Test
To verify a subjects’ ability to see color, a test of red-green color blindness was
performed. Subjects were asked to read the numbers on contained within Pseudo-

Isochromatic Plates manufactured by Richmond Products.

Collection Subject lnforrnation
A data recording sheet was given to each subject and demographic information
was collected. This included: gender, the highest level of education completed, age,

height, etc. The data recording sheet is offered in Appendix 3.

Eye Tracking
Subjects were seated in front of Pan Tilt eye camera and asked to place their
head on the chin rest. After subjects were determined to be comfortable, the researcher

calibrated the subjects’ eye using a file that contained nine frxed points (see Figure 19).

51

After the eye calibration, eye tracking began. Subjects were shown the 18 image
slide show, with each image appearing a single time for a period of 10 seconds each.
Slides were separated by a grey screen, which appeared for a period of three seconds, as
well as at the beginning and ending of the slide show. Slide order was randomized to

rrrinirnize any effect of fatigue or run order.

Using Gaze Tracker Software, look zones were constructed for: 1. the nutrition
information, 2. the net contents, 3. the brand name, 4. the graphic image that appeared on
the PDP 5. Health claims. It is important to note that the health claims were not included
in the analysis, as only two products contained this zone. Zones were drawn so that they
accurately followed the image of interest. Data from Gaze Tracker was exported into an
Excel Spreadsheet to create a flat‘file that was analyzed in SAS statistical software (Cary,

NC).

Survey
After eye-tracking, subjects were asked to respond to a survey regarding their cereal

consumption behaviors and the importance of healthy eating. (See APPENDIX 3)

Compensation

All subjects were given a 10$ Best-buy Gift card. (See APPENDIX 2)

52

CHAPTER 6- Results, Analysis and Discussion

6. 1. Arranging Data for Result Interpretation

As discussed in the Methods section of this document, the researcher collected
both data that characterized the user (e. g. visual, acuity demographics, consumption
preferences, self reported height and weight etc. See Data Recording Sheet in
APPENDIX 3) and ‘eye tracking data.’ This enabled the researcher to get a basic
understanding of the subject’s physical conditions, cereal purchasing habits, opinions

about healthy eating, calculate each subject’s BMI, etc.

The eye tracking data was collected as a video file which was later processed
using Gaze Tracker® software. For Analysis, five "look zones (LZs)” of interest were

drawn on each of the 18 files created to serve as visual stimulus for the subjects.

The 5 look zones were: the nutrition information, the product’s Net Weight, the
Brand name, the Large Graphic Image on the PDP and Health Claim on PDP. However,
only four zones were analyzed. The look zone 5, “Health Claims,” was removed from the
analysis because not all products contained such claims. The data measured in the other

four look zones were used for the purpose of analysis.

As explained in the Methods Chapter, 18 slide image files consisting of 9 brands
of cereal at two levels were presented to subjects in randomized order of appearance. One
treatment level was represented by the nine brands in their commercial presentation
(identified as NFP configurations) and the other level was the same nine brands that had

their NFPs modified to include iconic faces.
53

Henceforth, treatment one will be referred to as the NFP configuration, while
those modified (with the addition of iconic faces) will be referred to as ‘Graphic Nutrition

Facts Panels (GNFP).’

Nine cereal products were purposefully chosen for the research so that they were
nearly evenly distributed into three levels of health (i.e. healthy= 2 or more smiles;

moderate = 2 or more straight faces and unhealthy= 2 or more frowns). (See Table 5)

Table 5- Health Levels of Used Cereal Products

 

 

 

 

 

 

 

 

 

 

 

 

Fat Sodium Sugar Overall
Cereal Brand 1(6):; 1(6):; 1(6):; Healthfulness=Code
Shredded Wheat Smile Smile Smile Healthy = 3
Berry Burst Cheerios Smile Frown Frown Unhealthy = 1
Cocoa Pebbles Moderate Frown Frown Unhealthy = 1
Honey Nut Cheerios Moderate Frown Frown Unhealthy = 1
“1:52;;th Smile Smile Frown Healthy = 3
Kashi Honey Puffs Moderate Smile Frown Moderate = 2
Grape Nuts Smile Moderate Moderate Moderate = 2
Raisin Bran Smile Moderate Frown Moderate = 2
Select Apple Caramel Moderate Moderate Frown Moderate = 2
Pecan Crunch

 

 

 

 

 

54

 

Data collected to serve as response variables included: the proportion of people
that noticed a zone (i.e. the probability of hitting a zone), the number of hits to a zone and
time in zone. To compare the interaction between eye tracking data and consumer
information, a flat file was created and analyzed using SAS Statistical Software (Cary,

NC).
6. 2. Subject Demographics (Various Human Factors)

A total of 31 people 18 years and older participated in the eye tracking experiment.
Inevitably, some data is lost during the course of eye tracking (e. g. eye blinks, head
movements, etc). Collected data was examined to determine the total loss of time per
subject; subjects that were not tracked for more than 20% of the total viewing time were
not analyzed. After the filtering, data from 24 subjects was included in the analysis; i.e.

seven subjects. were removed. (See the APPENDIX 4)

Subjects were asked to respond to a series of survey questions so that their

consumption behaviors and preferences could be characterized.
Age

A convenience sample was drawn primarily from the student population at MSU
through word-of-mouth advertising. Subjects ranged in age from 20 to 63 years

(ave.28.7i11.54). Most of the subjects (20 of 24 participants) were in their twenties.

See Figure 20 for frequencies of subject by age.

55

 

Subject Demographics: Age

 

 

 

 

 

 

 

 

 

 

Number of Subjects
8

 

 

 

 

205 305 403 50s 60s

 

 

 

Figure 20- Frequency chart of subjects by age
Gender

Among the 24 participants included in the data analysis, 11 were male (code=1)

and 13 were female (code=2). (See Figure 21 for frequency by gender)

 

Subject Demographics: Gender

Number of Subjects

 

1= Male 2= Female

 

 

 

Figure 21- Frequency counts of subjects by gender
56

Visual Acuity

Visual acuity, based on a subject’s corrected eyesight, was measured and recorded

(visual acuity code;- 20/20 = 1, 20/30 = 2, 20/40 = 3 and 20/50 = 4)

About half (13 members out of 24) of tested participants were measured to have

visual acuity of 20/30. (See Figure 22 for subject frequency by visual acuity)

 

Subject Demographics: Visual Acuity

 

 

 

 

 

 

 

 

Number of Subjects
0‘

 

 

l=20/20 2=20/ 3 0 3=20/40 4=20/50

 

 

 

Figure 22- Subject fi'equency by visual acuity
BMI (Body Mass Index)

BMI is a measure of body fat based on height and weight that applies to adult men
and women. BMIs were calculated using the self-reported height and weight of subjects
and the calculator available from the National Heart, Lung and Blood Institute website
(Source of BMI; www.mhlbisupport.com/bmi/). This calculation provided 4 levels of
BMI, which gives an indication of the degree of obesity of an individual. (BMI Code;

Underweight = 1, Normal weight = 2, Overweight = 3 and Obesity = 4) (See Table 6)
, 57

Table 6- The BMI Categories (Data Source from; http://www.nhlbisupport.com/bmi/)

 

 

 

 

 

 

 

 

Level of Obesity BMI Value Code
Underweight <18.5 1
Normal weight 18.5—24.9 2
Overweight 25—29.9 3
Obesity BMI of 30 or greater 4

 

The majority of subjects were indicated to be normal weight, but 10 out of 24

subjects (41.7%) were overweight or obesity. (See Figure 23 for subject frequency by

BMI status)

 

Subject Demographics: BMI

 

an:
A

 

 

.—
N

 

 

—
O

 

 

0°

 

 

 

 

Number of Subjects

 

 

 

 

 

1= Under weight 2= Normal weight 3= Overweight

 

4= Obese

 

 

Figure 23- Subject Demographics: BMI

The Opinion for Healthy Eating Question

58

 

Subjects answered a series of survey questions. Using a 5-point Likert scale,
subjects indicated their level of agreement with the statement “Healthy Eating is
Important for my Life.” The 5 answers were coded as ‘Strongly Disagree = 1, Disagree =

2, neither disagree nor agree = 3, Agree = 4, Strongly Agree = 5.’

All but two subjects indicated that healthy eating was important to some degree,
with 13 agreeing and 9 strongly agreeing with the statement. (See figure 24 for frequency

reports)

 

Subject Demographics: Healthy Eating is Important

Number of Subjects

 

l=Strongly 2=Disagree 3=Neigther 4=Agree =Strongly
Disagree . Agree nor Agree
Disagree

 

 

 

Figure 24- Frequency of Response to the statement, “Healthy Eating is Important for my

Life”

6. 3. Interpretations of Eye Tracking Data

59

Data were analyzed for all 24 subjects using binary (noticed, yes/no), ordinal

(number of hits) and variable (time in zone) outcomes.

Statistical analysis and more specific explanations of each dependant variable are

continued below.

6.3.1. Section 1: The probability number of noticing in a zone (Binary

response: Noticed Yes or No)

The success or failure of a subject to notice one of the zones of interest was
modeled as ‘a binary response variable (noticed/did not notice)’, with zone being a four

level categorical value.

As mentioned above, the 4 zones were: the Nutrition Information, the product’s
net weight, the brand name, and the large graphic image on the PDP. The zone “Health
Claims” was removed from the analysis because not all products contained such claims.
Usable data was collected fiom a total of 24 subjects after filtering subjects with data

losses of greater than 20%.

The GLIMMIX (General Linear Mixed Model) procedure of SAS (SAS Institute
Inc., Cary, NC) was used for statistical analysis. The probability of noticing a zone was
modeled as a function of: subject, brand (9 levels), the health level of a cereal (3 levels),

nutrition information formatting (2 levels) and zone (4 levels).

Nutrition Information Formatting (N IF : NFP or GNFP) and health level of a

cereal were modeled as fixed effects. BMI, age, gender, their visual acuity, the healthful
60

level of the person’s favorite cereal, and the importance that subjects placed on eating
healthy were considered for possible inclusion on the model, but excluded if they did not

improve the model fit (P>0.10).

The final model included the fixed effects: Nutrition Information Formatting
(NFP v GNF P), zone and health level and their interactions and the random effects BMI
and age. (See Table 7) In NIF, traditional Nutrition Facts Panel (NF P) was coded as 1,
and icon inserted graphic Nutrition Facts Panel (GNFP) coded as 2.

Table 7- Type III Test of Fixed Effects of the probability of noticing a given zone

 

 

 

 

 

 

 

 

 

 

Efi’ect Num DF , Den DF F Value Pr>F
r . I . v' \

Nutrition Info Formatting (NIF) 1 .. 1488 0.01 0.9150
Zone 3 1488". 19.96 <0.0001
NIF x Zone 3 1488 2.90 0.0337
Reported Health Level 2 1488 1.27 0.2806
Reported Health Level x NIF 2 1488 0.65 0.5199
Reported Health Level x Zone 6 1488 3.40 0.0024
Reported Health Level x NIF x 6 1488 0.11 0.9956

Zone
BMI x Zone 4 1488 3.17 0.0132
Age 1 1488 4.48 0.0345

 

 

 

 

 

 

 

*Bolded effects are significant at a=0.05.

1. NIF x Zone

61

Based on the significant interaction term that was indicated between Nutrition
information and zone when the dependent variable was the probability of noticing a given
zone,_pairwise comparisons were examined. As mentioned above, code numbering NFPl
was the traditional nutrition facts panel (N FP: control) and NFP2 represented the icon

inserted GNFP; graphic nutrition facts panel (test).

The JLSD method was used for the pairwise comparisons in this section.
I r, L r r

. r ’ ’
r ‘1

Table 8- Pairwise comparison of NIF with Zone for the probability noticing a zone

 

 

 

Zone 1 Zone 4
. . Zone 2 Zone 3 (large
(““mtm“ (net wei ht) (brand) hie on
facts panel) g gralfDP)
NFPI 0.8049 i 0.1062 :1: 0.9429 :1: 0.9675 :1:
(=traditional a, A b, A c, A c, A
NFP) 0.03 0.02 0.17 0.01
(=icon 0.9041 :1: 0.0964 :I: 0.933 It 0.9404 :1:
inserted 0 028’ B 0 02'”A 0 02M 0 02“
GNFP) . . . .

 

 

 

 

 

 

"' Values in the same row with different lowercase letters provide evidence of statistical
significance at 0r=0.05.

** Values in the same column with different capital letters provide evidence of statistical
significance at a=0.05.

62

 

Probability Noticing in a Zone

 

0.0
«gage Mi staffed?

v v a N“ " "
e, C
s 5 S $7 5 5? s
Figure 25- Bar chart for pairwise comparison of NFP with ZONE as Function of the

probability of noticing a zone

Although the main effect of NIF was not indicated to be significant (0.9150), this
is to be expected, as it compares all zones across the two levels of treatment (the images
that had the altered and unaltered nutritional information). However, when the interaction
term was explored, data suggested that participants were significantly more likely to
notice the nutritional information when presented in GNFP format (90.4%) than the
traditional NFP format (80.5%) (P= 0.0051). This suggests that icon presentation of
nutrients in the Nutrition Facts Panel (NF P) has the potential to attract the attention of

consumers to this section.

63

There was no evidence of difference when the formats were compared across
other zones (column comparisons). This was to be expected, as all other zones of the two
formats were identical. Data analysis provided no evidence of statistical difference on the
probability of noticing the brand name (94.2% and 93.3%) (P=0.6666) as compared with

the graphics presented on the PDP (96.8% and 94.0%) (P=0.2370).

However, in the case of the traditional nutrition facts panel (NFPl), subjects were
significantly less likely to view the nutrition facts information (8.0%) compared with the

brand name (9.4% at P <0.0001) or the PDP graphics (9.7% at P <0.0001).

Not surprisingly, for both treatments, the net contents (Zone 2) was the element

least likely to be noticed (10.6% and 9.6%) (P=0.7372).

2. Health level of Cereal Products x Zone
Table 9 indicates the pairwise comparison for the probability of noticing a zone. LSD

method was used for the pairwise comparisons in this section.

64

Table 9- Pairwise comparisons of cereal health level with zone on the probability of
noticing a given zone

 

Zone 1 Zone 4
(nutrition facts Zone 2 Zone 3 (large
anel) (net weight) (brand) graphic on
p PDP)

 

Health Level] 0.8537 a 0.08847 a 0.9605 a 0.9750 e
(III-healthful) 0.03“A 0.02",A 0.028A 0.01“

 

“3:33.13? 03756 i 0.07121 :1: 0.95183: 0,9053 :1:
Healthy) y 0.03“A 0.02":A 0.02%A 0.02393

 

Health Level=3 0.8542 :1: 0.1609 :1: 0.87923: 0.9647 :1:
(Healthful) 0.04“A 0.04",A 0.04”"B 0.028AB

 

 

 

 

 

 

 

* Values in the same row with different lowercase letters show evidence of statistical
significance at a=0.05.

** Values in the same column with different capital letters provide evidence of statistical
significance at a=0.05.

65

 

 

 

 

 

 

 

 

 

 

 

 

1.2
O
E] 1.0 - 13.2: \\
g a ‘5' i
-—- 0-8 . \ E
fa." h E
a a... N E
o 352:3? \ E
o _ \ E
H 0'2 2213:} \ E
°‘ § 2 [I]
’°ss§§ 583s sass
«9 A? A? «\9 A? A? A? A? A? A? «9 A?

’84
’82
’8;
’8;
’81
’81

i??? .8”; e e e e e.»
gees @éssrssfie
Figure 26- Bar chart for pairwise comparison of Health Level with Zone as function of

the probability noticing in a zone

Data depicted in Table 9 and Figure 26 did not provide evidence of an effect of
health level on the probability of noticing the nutrition information (see column one,
Table 9). In other words, cereals that were categorized as “healthful,” moderately
healthful,” or “unhealtlrful” (based on the key nutrients identified in Table 9) did not
appear to elicit differing attentive responses when the probability of noticing the nutrition

information was the dependent variable of interest.

There was evidence of a difference in the probability of noticing the brand name

and large graphics on the PDP across levels of healthfulness. Participants were

66

significantly less likely to notice the brand name when it was level 3 (a “healthy cereal”
88%) as compared to cereals characterized as “unhealthy” 96% (P=0.0204) or
“moderately healthy” (95%; P=0.0292) There was no evidence of difference in the
probability of noticing a brand name when the unhealthy (96%) and moderately healthy

cereals (95% ) were compared (P=0.6900).

Pairwise analysis of zone 4, the graphic information presented on the package
PDP, also provided evidence of significant differences in the probability of noticing the
graphic, based on level of healthfulness. Moderately healthy cereals had graphics that
were significantly less likely to be viewed (90.5%) when compared to “unhealthy”
cereals (97.5%; P= 0.0256). However, when the probability of viewing the graphic
elements on moderately healthy cereals (90.5%) was compared with healthy cereals
(96.4%), no evidence of a difference existed (P= 0.0733). Similarly, data provided no
evidence of a difference when the probability of viewing the graphic element of a

“healthy” cereal (96.5%) and an “unhealthy” cereal (97.5%) were compared.(P=0.6513)

3. BMI x Zone

BMI was also examined for correlation to the probability of viewing a given zone.

See Table 10.

67

Table 10- Correlation of BMI with each zone level and probability noticing in a zone

 

 

 

Zone 4
Z0 1
(nutritilbir facts Zone 2 Zone 3 (large
anel) (net weight) (brand) graphic on
P PDP)
BM] -0. 1873 :1: -0.3782 :l: -0.6715 :1: -O.3819. :1:
0.1742 0.2237 0.2345 0.2399
If ? t; 0.2824 0.0912 0.0042 0.1 l 16

 

 

 

 

 

 

 

Table 10 suggests that there is less negative correlation between BMI and noticing

in zone] than those in zone 2, 3, and 4.

6.3.2. Section 2: Numerical Response Variable- the Number of Hits in a

Zone

As with the probability of noticing, the number of hits in a zone was also modeled.
Nutrition Information Formatting (N FP vs. GNFP), Zone, and Health Level of a Cereal
(see Table 11) were modeled as fixed effects. BMI, age, Gender, Favorite Cereal, Healthy
Eating Importance question and Visual Acuity were considered for incorporation into the

model, but later removed as their inclusion did not improve model fit (p-value >0.10).

The GLIMMD( (General Linear Mixed Model) procedure of SAS (SAS Institute

Inc., Cary, NC) was used. The final model included 3 fixed effects and their interactions,

68

as well as BMI, Age, Gender and Healthy Eating Importance. Pairwise comparisons were

conducted using an LSD procedure (See Table 11).

Table 11- Type 111 Test of Fixed Effects of the Number of Hit in a Zone

 

 

 

 

 

 

 

 

 

 

 

Eflect 1%: Den DF F Value Pr>F

Nutrition Info Formatting (NIF) 1 1485 4.64 0.0314
Zone 3 1485 16.58 <0.0001
NIF x Zone 3 1485 7.61 <0.0001

Reported Health Level 2 1485 3.77 0.0234
Reported Health Level x NIF 2 1485 0.34 0.7127
Reported Health Level x Zone 6 1485 15.30 <0.0001
Reported Health Level x NIF x Zone 6 1485 1.49 0.1775
BMI x Zone . 4 1485 19.91 <0.0001

Age 1 1485 3.20 0.0739

Gender x Zone 4 1485 12.80 <0.0001

 

 

 

 

 

 

 

*Bolded effects are significant at u=0.05.

1. NIF x Zone
Pairwise comparisons of NIF formatting (N F P vs GNFP) with Zone were examined

for possible effects on the number of hits in a zone.

69

Table 12- Pairwise comparison of NIF with Zone for number of hits in a zone

 

 

 

Zone 1 Zone 4

(nutrition facts Zone 2 Zone 3 (large
anel) (net weight) (brand) graphic on

P PDP)
(fixing 4.5162 :1: 0.03270 a 3.6430 a 4.5034. 3:

NFP) 0.21”"A 0.02M 0.188A 0.21M

NFP2

(=icon 5.1122 :1: 0.2333 i 3.49273: 3.8852 3:

i333! 0.23":B 0.03",B 0.18““ 0.19“

 

 

 

 

 

 

 

* Values in the same row with different lowercase letters provide evidence of statistical
significance at 0t=0.05.

** Values in the same column with different capital letters provide evidence of statistical
significance at 01=0.05.

Number of Hit in a Zone

 

$~23 save:

g

X 38 «8° «8 of «8° «8°
.8 s 48‘ s . 48’s” 48
Figure 27- Bar chart for pairwise comparison of NIF with Zone as function of number of
hits in a zone

70

Comparisons of the nutritional information suggest that the iconic format

garnered more hits than did the traditional (text only) NFP (5.11 vs. 4.52; P = 0.039).

When comparing the zones when nutritional information was presented in a
traditional (text only) format (N FPl), the nutritional information (4.50) and the graphic
on the PDP (4.50), data did not suggest that the zone elicited a different number of hits
(P= 0.9492). However, the number of hits elicited by these zones was significantly

different than the brand name (3.64; P <0.0001) or the net contents (0.03; P<0.0001).

When this same comparison was done for the stimulus material that contained the
nutritional information in the GNFP format, all zones were statistically significantly
different from one another (0r=0.05). The icon inserted nutritional information (GNFP)
had a higher number of hits than all other zones (5.11), and the net contents had the least

number of hits (0.23).

2. Health level x Zone
Table 13 indicates the pairwise comparisons by a cereal’s health level when the
dependent variable was the number of hits in a given zone. For the pairwise comparison,

the LSD method was used. (See Table 13 and Figure 28)

71

Table 13- Pairwise comparison of cereal health level with Zone for number of hits

 

 

 

 

 

 

 

 

 

Zone 1 Zone 4
. . Zone 2 Zone 3 (large
(““mm” (net weight) (brand) bio on
facts panel) wig)”
Health Level] 4.86934: 0.30643: 4.43452 4.71633:
(Unhealthful) 0.28““ 0.05",A 0.27”"A 0.28““
Health Level=2
4.9900: 0.219421: 3.895221: 3.29383:
“$53? 0.26”"A 0.03",A 0.21“,A 0.17‘LIB
Health Lever=3 4.56564: 0.31344: 2.62763: 4.71113: '
(Healthful) 0.32“ 0.06““ 0.218B 0.33M

 

* Values in the same row with different lowercase letters provide evidence of statistical
differences at a=0.05.

** Values in the same column with different capital letters provide evidence of statistical
differences at o.=0.05.

Number of Hit in a Zone

 

 

 

 

 

 

 

 

 

 

 

e
s - _-
_ 8‘ a
3. § E Z
28 E 2% E
352323; \ E / 5555233
2 - =‘s§s§s \ 2.5 % 2325;:
55%;? \ E % 35553;?
255;: § 5 / sgzgzgs
8 E /
1 ‘ 2;: \ = / :;:;:;:
is: N E / 3:52:25
0 98% En . E ,
«7 e 3‘ «'7 "' e? «'7 "' 3? 8'
fff’ 8888 seies
M88 88W88‘8888

iii

Figure 28- Bar chart for pairwise comparison of Health Level with ZONE as function of

iiii

diff

the Number of hits

72

 

 

From Table 13 and Figure 28, data provided no evidence of a statistical difference
of a cereal’s healthfulness on the number of hits when viewing the nutritional information

at a=0.05.

There was an effect of the healthfulness of a cereal on the number of hits to zones
three and four (brand name and graphic information). “Healthy” cereals registered
significantly fewer hits to the brand name (2.63) than did the “Moderately Healthy” (3.90)
(P<0.0001) or those that were unhealthy (4.43) (P<0.0001). There was no evidence of
difference when hits to the brand name were compared for those that were moderately

healthy (3.90) and those that were “unhealthy” (4.43) (P=0.0732).

The number of hits to the graphic information also indicated evidence of
difference when the healthfulness of the cereals was compared. The graphics of
“moderately healthy” (3.30) cereals registered significantly fewer hits than those that

were “healthy” (4.71) (P <0.0001) or “unhealthy” (4.72) (P <0.0001).
3. BMI x Zone
Data presented in Table 14, suggests a positive correlation of BMI with Zone 1.

Table 14- Correlation of BMI with each zone level for the number of hits in a zone

 

 

 

Zone 1 Zone 4
(nutrition facts (11:23; 2h t) (€233) (large
panel) 3 graphic on PDP)
BMI 0.2114. -0.4916 -0.4657 i -0.0538 :1:
300372 :tO. 1340 0.0402 0.0397
Pr ’> t < .0001 0.0003 0.2466 0.1764

 

 

 

 

 

 

73

 

4. Genderx Zone

Table 15 indicates the pairwise comparisons of gender with zone for the number of

hits in a given zone. The LSD method was used for the pairwise comparisons. (See Table

15 and Figure 29)

Table 15- Pairwise comparisons of gender with zone for the number of hits in a given

 

 

 

zone
Zone 4
Zone 1
(nutrition facts Zone 2 Zone 3 (large
panel) (net weight) (brand) gmggi’c) on
Gender 1 4.2574. :1: 0.4439 :1: 3.3827 :1: 3.9285 :1:
(Male) 0.23‘LA 0.05",A 0.19“,A 0.218A
Gender 2 5.4230 :1: 0.1718 :1: 3.7565 :1: 4.4537 :1:
(Female) 0.275"B 0.03",B 0.203A 0.22“LA

 

 

 

 

 

* Values in the same row with different lowercase letters provide evidence of statistical
significance at a=0.05.

** Values in the same column with different capital letters provide evidence of statistical
significance at a=0.05.

74

 

Number of Hit in a Zone

 

Figure 29- Bar chart for pairwise comparison of Gender with Zone as function of number

of hit in a zone

From Table 15 and Figure 29, females made significantly more hits (5.42) to
nutritional information than their male counterparts (4.26) (P<0.0001). However, men
registered a higher number of hits to the net contents (0.44) than did females (0.17)
(P<0.0001). No gender differences were evident when the number of hits to the brand
name were compared (P= 0.1141). There was a marginal difference in the number of hits
between at zone 4 (graphic images on the PDP) females (4.45) registered compared with

males (3.93) (P=0.0508).

75

When within gender comparisons were made between zones of interest, there was
evidence of a difference in the number of hits of all zones of interest when viewed by
females (a=0.05) (see Table 15). Females had the greatest number of hits on the nutrition
facts zone (5.42), followed by the PDP graphics (4.45), then the brand name (3.76), and,

finally, the net contents (0.17).

Males followed a similar pattern, registering the most number of hits to the
nutritional information (4.26), followed by the PDP graphics (3.93), brand name (3.39)
and, finally, the net contents (0.44). However, for males, data analysis provided no
evidence of a statistical difference when their viewing behaviors of nutritional

information (4.26) and PDP graphics (3.93) were compared (P=0.0939).

6.3.3. Section 3: Numerical Response Variable- the Total Time Spent in a

Zone

Nutrition Information Formatting (N FP vs. GNFP), Zone, and Health Level of
Cereals were modeled as fixed effects. BMI, Age, Gender, Favorite Cereal, Healthy
Eating Importance and Visual Acuity were added to the model, but later removed because

they did not improve model fit (p-value > 0.10).

For statistical analysis The Mixed procedure of SAS (SAS Institute Inc., Cary,
NC) was used. The final model included 3 fixed effects and their interactions, as well as
BMI and Gender. See Table 16. Pairwise comparisons were conducted using the LSD

method.

76

Table 16- Type 111 Test of Fixed Effects of the Total Time Spent in a Zone

 

 

 

 

 

 

 

 

 

 

Efi'ect Num DF gel? F Value Pr>F

Nutrition Info Formatting (NIF) 1 1485 3.97 0.0465

Zone 3 1485 59.05 <0.0001

NIF x Zone 3 1485 17.27 <0.0001

Reported Health Level 2 1485 2.11 0.1217

Reported Health Level x NIF 2 1485 0.22 0.8005

Reported Health Level x Zone 6 1485 8.77 <0.0001

Reported Health Level x NIF x 6 1485 0.39 0.8848
Zone

Gender x Zone 4 1485 2.89 0.0213

BMI x Zone 4 1485 13.72 <0.0001

 

 

 

 

 

 

 

*Bolded effects are significant at a=0.05.

1. NIF x Zone
Pairwise comparisons of nutrition information formatting (N FP vs. GNFP) with

zone were examined for possible effects on the time spent in each zone.

77

Table 17- Pairwise comparisons of nutrition formatting with zone for the total time spent

 

 

 

 

in a zone
Zone 4
(nutfi‘fifirlfacts Zone 2 Zone 3 (large
anel) (net weight) (brand) graphic on
p PDP)
(41:55.1... 1.3004 4 0.04533 :1: 1.0205 3: 1.1344 :1:
NFP) 004“ 0.04‘m 0.04“A 0.04M
NFP2
(=icon 1.7204 i 0.04977 :1: 0.9490:|: 1.0172 :I:
inserted 0.0433 0.04",A 0.04“A 0.04%B
GNFP)

 

 

 

 

 

 

* Values in the same row with different lowercase letters provided evidence of statistical
significance at a=0.05.

** Values in the same column with different capital letters provided evidence of
statistical significance at 0t=0.05.

Time Spent in a Zone (sec.)

 

Figure 30- Bar chart for pairwise comparison of NIF with Zone as function of Time spent

in zone
78

When the results (time in zone) are compared across the two formatting
treatments (NF P vs. GNFP), the nutritional information (P <0.0001) and PDP graphics
(P=0.0478) indicated evidence of statistical differences in viewing time. For the
nutritional information, the GNFP format was viewed significantly longer by subjects
(1.72) than the NFP (traditional) (1.30) (P<0.0001). However, study participants spent
more time viewing the PDP graphics of the traditionally formatted packaging (1.13) than
they did on the package that contained the GNF label (1.02) (P=0.0478). For the net
contents, there was no significant differences between the GNFP formatted label (0.05)
and NFP (0.05) at P=0.9402. Comparison of the brand names across the two formats did
not appear to elicit a significant difference in viewing time; the time was 1.02 seconds for
the brand names that appeared in the traditional formats, and 0.95 seconds for the brand
names that appeared with nutritional information that had been modified to include icons

(P= 0.2271).

When zones were compared within nutritional format, both formats exhibited a
similar pattern. When time is zone was considered to be the measure of attentive behavior,
no statistical difference (a=0.05) was evident when brand name and PDP graphics were
compared for both formats (traditional NFP and icon inserted GNF P). In the traditional
NFP a comparison of the brand name (mean =1 .02) and graphic information (mean=l.13)
provided evidence of marginal significance (P=0.0548).

Like the labels that appeared in traditional formats (NF P), a comparison of the
brand name (mean =0.95) and graphic information (mean=1.01) for those that appeared
with the GNFP formats provided no evidence of statistical significance P=0.2500. Both

formats elicited the greatest amount of time viewing on the nutrition information (1.30
79

for traditional formats and 1.72 for icon inserted GNFPs) and the least amount of time

was spent (for both) on the net contents (0.05 at traditional NFP and 0.05 at GNFP).

2. Health level x Zone
Table 18 indicates the pairwise comparisons for the total time spent in a given zone.

The LSD method was used for the pairwise comparisons. (See Table 18 and Figure 31)

Table 18- Pairwise comparison of a cereal’s health level with zone for the total time spent
in a given zone

 

 

 

Z0 1 Zone 4

“.8 . Zone 2 Zone 3 (large
(nutrition (net wei ht) (brand) hie on

facts panel) g ngP)

Health Level=l 1.4583 :1: 0.05201 :1: 1,1175 4 1.1217 i

(Unhealthful) 0.05“,A 0.051% 0.05c,A 0.05c'A
113;: 1:1th 15709 i 003375 is 1.0332i 0.8454 :1:

Healthy)y 0.04“ 0.04",A 0.04“,A 0.04“~B

 

Health Level=3 1.50204 0.056892: 0.80371: 1.2604:
(Healthful) 0.06M 0.06M O.06°'B 0.06M

 

 

 

 

 

 

 

* Values in the same row with different lowercase letters provide evidence of statistical
significance at a=0.05.

** Values in the same column with different capital letters provide evidence of statistical
significance at a=0.05.

80

1.8

 

 

 

 

 

 

 

 

 

 

 

 

1.6 - =3

:3; 1.4 - $ E
g 1.2 - E g}
o .: 5: xx E /
N _ :33; ‘Q E , _,, /
43 1'0 Egigig \ E / /
an. §§ E 3% Z
8. 0 6 _ 232: § 5 % zazsza /
m - \ E % /
o \ E % %
.§ 0-4 ‘ 2333;? § '=" % 53323: %
[— \ E / //
w- 3% Efléfi Z
00 ~ ,, .\\ E g 35;? 2,
.7. .. a a g a a g; 3

$00? 060$..sv
3%?” $335” $5.55
5555 4455 5555

Figure 31- Bar chart for pairwise comparison of Health Level with Zone as function of

 

total time spent in a given zone

When the two formats of labeling were compared, there was no evidence of an
impact of the cereal’s level of healthfulness on the time spent in the when the nutritional
information was compared at 01=0.05. This was also the case for the net contents.
However, data analysis did indicate a difference in the attentive behaviors of subjects (as
indicated by time in zone) when they viewed the brand name and graphic information.
“Unhealthy” cereals (1.12) and “moderately healthy” cereals (1.03) registered the greatest
amount of time in the brand name as compared “healthy” (0.80) (P<0.0001; P=0.0019,

respectively).

81

Comparisons of cereal healthfulness and the attentive behaviors (as measured by
time in zone) were also evident when subjects viewed the PDP graphics. The level of
attention garnered by “healthy” (1.26) and “unhealthy” (1.12) was not significantly
different (P=0.0750). However, the level of attention gamered by the “moderately healthy”
products (0.85) was significantly less than the others; when compared to “healthy”

products (mean 1.26) P<0.0001 and “unhealthy” products (mean 1.12) P <0.0001).

3. BMI x Zone

Table 19 suggests that the mean estimate value is positive only at the correlation
of BMI with Zone 1. It also indicates less negative correlation between BMI and hitting

in Zone] than those in Zone 2, 3, and 4.

Table 19- Pairwise comparison of BMI with ZONE for the total time spent in a given

 

 

 

zone
Zone 1 Zone 4
(nutrition facts Zone 2h in”: (large
panel) (”mag ') ( "n ) graphic on PDP)
BMI 0.0512 -0.0054 -0.1409 -0.2101
:l:0.03486 :l:0.03486 i0.03486 i0.03486
Pr>t < .1419 0.8768 < .0001 < .0001

 

 

 

 

 

 

 

4. Gender x Zone

Table 20 presents the pairwise comparisons of gender with zone when the total
time spent in a zone was the resultant variable. The LSD method was used for the

pairwise comparisons. (See Table 20 and Figure 32 below)

82

Table 20- Pairwise comparisons of Gender with Zone for the total time spent in a given

 

 

 

 

zone
Zone 4
Z.°."° 1 Zone 2 Zone 3 (large
(nutrrtron facts . .
panel) (net werght) (brand) gulps; on
Gender 1 1.5461. :1: 0.05675 :1: 1.0689 :1: 1.1069 :1:
(Male) 0.04“ 0.04"'A 0.04“A 0.04”"
Gender 2 1.4747 i 0.03835 :1: 0. 90073: 1.0447 :1:
(Female) 0.038,A 0.04M 0.04“,B 0.04‘LA

 

 

 

 

 

* Values in the same row with different lowercase letters provide evidence of statistical

significance at 01=0.05.

** Values in the same column with different capital letters provide evidence of statistical
significance at a=0.05.

Time Spent in a Zone (sec.)

 

Figure 32- Bar chart for pairwise comparison of Gender with Zone as firnction of total

83

time spent in a given zone

 

When comparing the attentive behaviors of the genders in each of the zones, only
one zone, the brand name, elicited a statistically different response in terms of the time
spent. Women spent significantly less time (0.90) viewing the brand name than men

(1.07) (P=0.0032).

When making comparisons across the zones within gender, men spent
significantly more time on the nutrition information than any other zone (mean 1.55) and
significantly less time viewing the net contents (0.06) (0t=0.05). Moderate to the brand
and net contents, were the PDP graphics and brand name, which were not significantly

different from one another: brand (1.07) and PDP graphics (1.11) at P=0.5294.

Women followed a similar pattern to men, viewing the nutritional information
(1.47) significantly longer than any other zone (a=0.05). Unlike men, however, evidence
collected from women suggested significant differences in the time that they spent
vieWing the brand name and PDP graphics, such that they spent significantly more time
on the graphics (1.04) than the brand name (0.90) (P=0.0100). As with their male
counterparts, females also spent significantly less time viewing the net contents (0.04)

compared to all other zones 01=0.05.

84

CHAPTER 7. Conclusions, Limitations and Future Study

This study investigated the attentive behaviors of 24 subjects (aged 20—63) while
viewing 18 flattened graphic images (PDP + nutritional information) of cereal packages
using a pan-tilt, bright-pupil eye tracker. Of the 18 images, 9 appeared in the
commercially available format (traditional NFP) and 9 were modified to include a
redundant cue (facial icons) for three nutrients: fat, sodium and sugar (GNF P). Cereals
were purposefirlly selected to be roughly equally distributed across three levels of

“healthfulness” (“healthy,” “unhealthy,” and “moderately healthy”)

Analysis of eye-tracking data concentrated on four Look zones (zone1= NFP,
zone2=Net Weight, zone3=Brand, and zone4=Large Image on the PDP). Each subject
saw each of the 18 images for a time period of ten seconds and the order of presentation

was randomized to minimize effects of run order.

Study results can be used not only to investigate the effectiveness of facial icons
to convey nutritional information, but also can be used to develop insights into the
attentive behaviors of consumers based on their characteristics (gender, BMI, etc.) and a

given cereal’s level of “healthfulness.”

7.1. Effectiveness of the Graphic Nutrition Label and Consumer

Behavior

About half of the subjects were overweight (BMI: 25 — 29.9) or obese (BMI:
greater than 30). Yet, when asked about the importance of eating and health, no one

suggested this to be unimportant.
85

7.1.1. Effectiveness of the Facial Icon

The research suggested that the icon inserted into the nutrition facts label (GNFP)
is more noticeable than the traditional nutrition facts label in cereal products. In all 3
sections of data analysis (the probability number of noticing in a zone, the number of hits
in a zone and the total time spent in a zone), the nutritional information with the icon
inserted registered significantly higher values than the traditional NFP for the nutritional
information. Specifically, when the dependent variable was the probability of noticing in
the nutrition zone, the mean value of the iconic formats was 90%, as compared with 80%
for the traditional NFP (P =0.0051). When the dependent variable was the number of hits
to a zone, the mean value nutritional information that appeared in the altered format was
5.1122 hits as compared with the NFP (mean 4.5162) (P= 0.0039). When the total spent
time in a zone was considered the dependent variable, mean values for the nutritional
information that appeared in altered formats was 1.7204 seconds as compared with the

traditional NFP (mean 1.3004 seconds) at P<0.0001.

This is somewhat expected, as a review of the literature indicates that facial
stimuli capture attention when people are engaged in another task [71], thereby making
face stimuli immune to the phenomena of inattentional blindness [72] or attentional blink

[7374].

Additionally, facial expressions of emotion have been indicated to be readily
evaluated in the near absence of attention [75-77]. Finally, it has been suggested that the
inclusion of pictorial icons can increase the ease with which information is encoded and

remembered [78]. People extract the basic meaning from pictures extremely rapidly [79]

86

and form relatively long lasting memories of them [80], even when people or not actively
attempting to form memories of them [81]. As such, further research into the use of

facial icons relating to their ability to convey nutritional information is suggested.

7.1 .2. Other Consumer Behaviors

Some interesting results that have the potential to impact future research
occurred regarding the differences in genders. When comparing the attentive behaviors
of the genders using the number of hits to the zone as the dependent variable, although
both genders “hit” the nutrition information more than any other zone, woman

(mean=5.4230) hit significantly more times than men (mean=4.2574) at P<0.0001.

Another interesting finding concerned a cereal’s “healthfulness” and the attentive
behaviors of subjects when the dependent variable was the probability of noticing a given
zone. Participants were significantly less likely to notice brand names for “healthful”
cereals than those that were “unhealthy” or “moderately healthy.” Specifically, the mean
value (at brand name) of healthful cereals was 88%; this was less than the registered
mean for unhealthy cereals 96% (P =0.0204). Also, the mean value of the healthful cereal
(88%) was significantly less than that of the moderate healthy cereal 95% at P= 0.0292.
Further exploration into this finding is warranted as it could be the result of a confluence

of factors (color, graphics, marketing, etc.)

Comparisons of the probability of noticing the graphics found on the PDP across
varied levels of “healthfirlness” also showed significance. The graphics of the two
extremes of “healthfulness” (i.e. “healthy” and “unhealthy”) were more likely to be

viewed than their “moderately healthy” counterparts. Specifically, the mean value of
87

“unhealthy” cereal (98%) was greater than “moderately healthy” (91%) at P=0.0256.
Also, the mean value of “healthy” cereal (96%) was greater than “moderately healthy”
(91%) at P=0.0433. One possible reason for this is that the graphics of the extremes are
tuned to provide congruence with the level of health (i.e. they are more extreme), where
the moderate cereals are not “tuned” to the message at hand, eliciting less attention to the

graphics.
7. 2. Limitations and Expectation for Future Study

7.2.1. Limited Variation in the Subject Population

The data was collected from 24 subjects using a convenience sample of a
relatively uniform age. (See Chapter 6) Small sample size of a very limited subject pool
means that the results presented here cannot be generalized. However, data does provide
pilot information regarding the attentive behaviors of people viewing cereals. The ASL
Pan Tilt eye-tracker is set in a laboratory of school of Packaging. Undoubtedly, the

laboratory nature of this setting influences the behaviors of people as they view items.
7.2.2. Package Labels on a Two Dimensional surface

In commercially-available, FDA regulated products, the PDP and NFP are placed

on different surfaces. Namely, cereal packages are usually 3 dimensional boxes.

However, in this study, 2 dimensional formats (flat files) were used (See Chapter
2). As such, subjects were automatically exposed to the nutritional information, where

in more ecologically valid settings; they would have to rotate the package for exposure to

88

ensue. As such, it is recommended that the experiment be repeated using a more

ecologically valid setting (e. g. homes) with Mobile eye tracking equipment.

89

APPENDIX 1. Research Consent Form

Michigan State University
School of Packaging

103 Packaging Bldg. East Lansing, Ml48824

INSTRUCTIONS ANDfiRESEARCH CONSENT FORM -
Employing eye tracking to measure attentive behaviors regarding labels

You are being asked to participate in a research study. The purpose of this research
study is to understand how people attend to various components of food labels
while they are making purchase decisions.

As part of this research, we will record your gender, height, weight educational
background, age and ask you to fill out a brief survey regarding your consumption of
cereal and dietary behaviors. This information will be tied to a subject number; you
will not be identified by name and your confidentiality will be maintained to the
maximum extent of the law. Information retrieved during this entire study will be
stored in a password protected computer in locked laboratories/offices in the School
of Packaging for a minimum of 3 years. The room will be accessible only to
authorized researchers of Dr. Laura Bix’s research team and the University’ s
Institutional Review Board (IRB). This study on the will take no longer than 1 hour.

We will also test your visual acuity (20/20, 20/30, 20/40, etc.) and color blindness.
These tests will be conducted by asking you to view a series of cards and asking you
to decipher images to the best of your ability.

E ye-tracking

After these tests have been run, we will begin the eye tracking portion of the study.
The eye tracker is a very sensitive eye movement monitor, which can tell us exactly
where your eyes are looking while you are viewing a package label. A beam of light
that cannot be detected by the human eye will be shown into your eye. The
instrument tracks the movement of your eye by tracking the movement of the beam.

90

Setup and Calibration:

You will be seated in front of our computer. In front of the computer, there sits a
small camera system that has the light beam. We will ask you to stay still, with your
chin resting on a chin rest as we adjust the camera, chair and chinrest so that it
shines the light into your eye. To calibrate the instrument, you will be asked to move
nothing but your eyes to a series of dots on the computer screen. Once calibrated,
testing will begin. During testing you will be asked to sit as still as possible with your
chin on the chinrest, and move nothing but your eyes.

Experiment Procedure:

A series of 18 cereal labels will appear on the screen, one at a time for ten seconds
each. The researcher will prompt you to look at the labels as you would before
deciding whether or not you would want to choose to purchase this cereal to eat for
breakfast.

If you have any questions at any time please ask.

If you choose to discontinue your participation you will still receive the $10 Best Buy
card.

The results of the study will be treated in strict confidence and that you will remain
anonymous. Your privacy will be protected to the maximum extent allowable by law.
Within these restrictions, results of the study will be made available to you at your
request.

Risks associated with participating include minor discomfort during the time when
you are asked to rest your head on the chin rest and remain still. Additionally, we
will be asking personal information that you may not feel comfortable disclosing
such as your height and weight. Please feel free to omit questions that you are not
comfortable answering, or choose to discontinue research at any point during the
study without penalty. You will not be penalized for discontinuing the research or
omitting answers i.e. you will still be eligible for the gift card.

Your participation in the study does not guarantee any beneficial results to you. The
study however does carry a potential benefit to society. Using the data generated in
this study, future research could focus on redesigning food labels.

91

As mentioned previously, your participation is voluntary: You may choose not to
participate at all, or you may refuse to participate in certain procedures or answer
certain questions without consequence.

If you have any concerns or questions about this research study, such as scientific
issues, how to do any part of it, or to report an injury, please contact the researcher,
Dr. Laura Bix, 153 Packaging East Lansing, MI 48824 at 517-355-4556 or

bixlaura@_msu.edu.

If you have questions or concerns about your role and rights as a research
participant, would like to obtain information or offer input, or would like to register a
complaint about this study, you may contact, anonymously if you wish, the Michigan
State University's Human Research Protection Program at 517-355-2180, Fax 517-
432-4503, or e-mail irb@msu.edu or regular mail at 207 Olds Hall, MSU, East Lansing,
MI 48824.

I voluntarily agree to participate in the Eye Tracking study of cereal labels.

 

Date:

 

You will be provided with a copy of this consent form.

92

APPENDIX 2. Recruitment Advertisement for

Research Experiment

Cereal Labeling Research

Participants wanted for research concerning cereal labeling. In
exchange for your participation, you will receive a $10 Best Buy
gifl certificate. The study will take no longer than 1 hour. To
participate in this study:

0 You must be able to provide your own transportation to
the School of Packaging at Michigan State University
You must be over 18 years of age

You must not be legally blind

You must not wear hard contact lenses

You must be willing to provide a contact phone number or email so that researchers can
contact you to remind you of your appointment

 

You are being asked to participate in a study of cereal labels being conducted by graduate student
Chang-hoon Oh.

As part of his Master’s Program, Chang-boon is investigating cereal labels. To do this, he is using
an ASL eye tracker. If you choose to participate, you will be hooked up to the eye tracker which
will track your eye movements as you look at 18 different labels.

You will also be asked to answer a brief questionnaire about your eating habits and cereal
purchasing behaviors and be asked basic demographic information. Instrument set up and the test
itself should take no longer than 45 minutes.

Prior to testing, your visual acuity (20/20, 20/30, 20/40, etc.) and a test for color blindness will be
administered. These tests involve viewing a series of cards.

93

If at any time you are uncomfortable with the testing or wish to discontinue the data collection
process, you may discontinue participation without penalty.

If you are interested in pursuing this opportunity, please contact Chang-hoon Oh at
ohchang@msu.edu to make an appointment.

If you have questions or comments regarding this study, please contact Dr. Laura Bix, Associate
Professor of Packaging at Michigan State University at 517-355—4556 or bixlaura@msu.edu.

94

APPENDIX 3. Data Recording Sheet

Data Recording Sheet

Cereal Labeling Study

Subject #

 

Male Female Other

 

Number of times per day you eat cereal

Number of different cereals purchased per week

 

Favorite Cereal

Most commonly consumed Cereal

 

Highest level of education completed

8th Grade High School Undergraduate Graduate Doctorate

Age

 

Height

95

Weight (lbs)

Visual Acuity

Color Blindness Score

 

 

 

Eye tracking

Do you have any of the following conditions? (circle all that apply)

Diabetes Hypertension High Cholesterol

Does anyone in your immediate family have any of the following? (circle all
that apply)

Diabetes Hypertension High Cholesterol

 

How often do you?

 

Never Rarely Sometimes Frequently Always

 

Read nutritional
labels

 

Diet to lose weight

 

 

Diet to gain weight

 

 

 

 

 

 

 

Healthy eating is important for you

 

 

 

1 2 3 4 5
Strongly Disagree Neither Agree Agree Strongly
Disagree nor Agree

Disagree

 

 

 

 

 

96

 

 

 

Healthy eating is important for your family

 

 

 

1 2 3 5
Strongly Disagree Neither Agree Agree Strongly
Disagree nor Agree

Disagree

 

 

 

 

 

 

How do you consider your physical health at present?

 

3

5

 

 

Poor

 

Fair

 

Neutral

 

Good

 

Excellent

 

Please indicate which of the following cereals you have eaten, of the ones that you
have eaten, please indicate how many times per week you have eaten them?

97

 

 

APPENDIX 4. Subject Filtering Data Based on Time Loss

Su Time Loss A 10sec.
0.103
0.247
0.108
7.386

0.312
0.685
6.384
0.141
0.515
5.004
0.657
0.531
0.321
0.51
1.615
1.082
0.166
4.621
0.6685
0.1 1 1
1.641
0.177
1.488
6.622
0.212
0.413
0.706
5.899
0.434
31 0.455
* Filtered data (average time loss more than 23ec./103ec.) marked as gray color.

(OQNQOI-th—ll

 

* 1 subject abandoned during the experiment. (Subject # 5)

98

10.

11.

12.

13.

14.

REFERENCE

Smith, T. Diet Quality and Food Consumption: Flexible Consumer Behavior
Survey 07 CBS). 2010 FEB.2010 [cited 2010 JULY.15]; Available from:
http://www.ers.uscEgov/briefmgjdietqualitv/flexible.htm.

Cohen, C. New Healthcare Marketing Vehicle Launches to help Americans Living
with Cholesterol, Diabetes and Obesity. 2008 [cited 2010.

Mokdad, A., et al., Prevalence of obesity, diabetes, and obesity-related health risk
factors, 2001. Jama, 2003. 289(1): p. 76.

St-Onge, M., K. Keller, and S. Heymsfield, Changes in childhood food
consumption patterns: a cause for concern in light of increasing boay weights.
American Journal of Clinical Nutrition, 2003. 78(6): p. 1068.

Dietz, W., Health consequences of obesity in youth: childhood predictors of adult
disease. Pediatrics, 1998. 101(3): p. 518.

KERSH, R., The politics of obesity: a current assessment and look ahead.
Milbank Quarterly, 2009. 87(1): p. 295-316.

Eckel, R. and R. Krauss, American Heart Association call to action: obesity as a
major risk factor for coronary heart disease. Circulation, 1998. 97(21): p. 2099.

Solomon, C. and J. Manson, Obesity and mortality: a review of the epidemiologic
data. American Journal of Clinical Nutrition, 1997. 66(4): p. 10448.

Heimbach, J ., Yesterday, today, and tomorrow: consumer perceptions of food
safety. Unpublished paper, Washington, DC, Bureau of Foods, Food and Drug
Administration (June 1981).

E., H.T., A whole diet approach to healthy eating. Nutrition Bulletin, 2004. 29(1):
p. 44-49.

Loading, C. and P. Fluids, HEALTHY EATING GUIDELINES.

MosocCo, D., Dietary Guidelines for Americans, 2005. Home Healthcare Nurse,
2006. 24(9): p. 552.

Harmening, J ., General Mills Sees More US Breakfast Cereal Consumption
Growth. 2010.

Peterkin, B., Dietary guidelines for Americans. Journal of the American Dietetic
Association, 1990. 90(12): p. 1725.

99

15.

16.

17.

18.

19.
20.

21.

22.

23.

24.

25.

26.

27.

28.

Yajima, Y., Y. Hayashi, and N. Alexandratos, China's consumption of cereals and
the capacity of the rest of the world to increase exports. Food Policy, 1997. 22(3):
p. 253-267.

Foderaro, L.W., These Days, the College Bowl Is Filled With Milk and Cereal, in
The New York Times. 2004.

Schwartz, M., et al., Examining the nutritional quality of breakfast cereals
marketed to children. Journal of the American Dietetic Association, 2008. 108(4):
p. 702-705.

Go with the grain. Health Module, 2000. 53(6).
Robertson, S., The perfect ‘thin ’ food, in Prevention. 2004.

Barton, B., et al., The relationship of breakfast and cereal consumption to nutrient
intake and body mass index: the National Heart, Lung, and Blood Institute
Growth and Health Study. Journal of the American Dietetic Association, 2005.
105(9): p. 1383-1389.

McKevith, B., Nutritional aspects of cereals. Nutrition Bulletin, 2004. 29(2): p.
1 1 1-142.

Anderson, J., Dietary fiber prevents carbohydrate-induced hypertriglyceridemia.
Current Atherosclerosis Reports, 2000. 2(6): p. 536-541.

Dj ousse, L. and J. Gaziano, Breakfast cereals and risk of heart failure in the
physicians' health study 1. Archives of internal medicine, 2007. 167(19): p. 2080.

Liebman, B. and J. Hurley, Cereals made simple. Nutrition action health letter.,
1999.

Karmally, W., et al., Cholesterol-lowering benefits of oat-containing cereal in
Hispanic Americans. Journal of the American Dietetic Association, 2005. 105(6):
p. 967-970.

Brighenti, F., et al., Effect of consumption of a ready-to-eat breakfast cereal
containing inulin on the intestinal milieu and blood lipids in healthy male
volunteers. European journal of clinical nutrition, 1999. 53(9): p. 726.

Qi, L., et al., Whole-grain, bran, and cereal fiber intakes and markers of systemic
inflammation in diabetic women. Diabetes care, 2006. 29(2): p. 207.

Melanson, K., et al., Consumption of whole-grain cereals during weight loss:
eflects on dietary quality, dietary fiber, magnesium, vitamin B-6, and obesity.
Journal of the American Dietetic Association, 2006. 106(9): p. 1380-1388.

100

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

Mattes, R., Ready-to-eat cereal used as a meal replacement promotes weight loss
in humans. Journal of the American College of Nutrition, 2002. 21(6): p. 570.

Wyatt, H., et al., Long-term weight loss and breakfast in subjects in the National
Weight Control Registry. Obesity, 2002. 10(2): p. 78-82.

Cereal Eflects, in Saturday evening post. 2005.

Maki, K., et al., Whole-Grain Ready-to-Eat Oat Cereal, as Part of a Dietary
Program for Weight Loss, Reduces Low-Density Lipoprotein Cholesterol in
Adults with Overweight and Obesity More than a Dietary Program Including
Low-Fiber Control Foods. Journal of the American Dietetic Association, 2010.
110(2): p. 205-214.

Wesnes, K., et al., Breakfast reduces declines in attention and memory over the
morning in schoolchildren. Appetite, 2003. 41(3): p. 329-331.

Smith, A., Breakfast cereal consumption and subjective reports of health.
International journal of food sciences and nutrition, 1999. 50(6): p. 445-449.

cereal think outside the bowl, in Mom's Best Naturals Blogs. 2007.

Grunert, K. and J. Wills, A review of European research on consumer response to
nutrition information on food labels. Journal of Public Health, 2007. 15(5): p.
385-399.

Krebs, J ., What's on the Label? Science, 2004. 306(5699): p. 1101.

Levy, A., et al., The impact of a nutrition information program on food purchases.
Journal of Public Policy & Marketing, 1985. 4: p. 1-13.

2009 Food & Health Survey: Consumer Attitudes toward Food, Nutrition and
Health. 2009, International Food Information Council Foundation.

Food label consumer research project: qualitative research findings. 2008,
International Food Information Council Foundation.

More Americans checking nutrition labels. 2010.

Driskell, J ., M. Schake, and H. Detter, Using Nutrition Labeling as a Potential
T 001 for Changing Eating Habits of University Dining Hall Patrons. Journal of
the American Dietetic Association, 2008. 108(12): p. 2071-2076.

Ippolito, P. and A. Mathios, New food labeling regulations and the flow of

nutrition information to consumers. Journal of Public Policy & Marketing, 1993.
12(2): p. 188-205.

101

44.

45.

46.
47.

48.

49.

50.

51.

52.

53.

54.
55.

56.

57.

58.

59.

VII. Nutrition Labeling; Questions LI through L153; Guidance for Industry; A
Food Labeling Guide, U.S.F.a.D. Administration, Editor. 2009.

Williams, P., Communicating health benefits-do we need health claims?
Australian Journal of Dairy Technology, 2005. 60(2): p. 192.

Olson, J ., Boost nutrition label literacy in Naturalfoods merchandiser. 2009.
Making the most of Nutrition Facts Labels. 2006.

Judith, A.B., Nutrient profiling : consumer friend or foe? Australian Journal of
Dairy Technology, 2009. 64(1): p. 6.

Ghosh, D., Functional food and health claims: regulations in Australia and New
Zealand. Australian Journal of Dairy Technology, 2009. 64(1): p. 152-154.

Drewnowski, A., Concept of a nutritious food: toward a nutrient density score.
American Journal of Clinical Nutrition, 2005. 82(4): p. 721.

Grunert, K., et al., Perception of health claims among Nordic consumers. Journal
of Consumer Policy, 2009. 32(3): p. 269-287.

C., T.L., Nutrients, foods and diets : challenging fimctional food development.
Australian Journal of Dairy Technology. 64(1): p. 3.

Williams, P., Consumer understanding and use of health claims for foods.
Nutrition reviews, 2005. 63(7): p. 256-264.

Colby, S., et al., Nutrition Marketing on Food Labels. 2010.

Sharfstein, J .M., Oversight of Dietary Supplements, U.S.F.a.D. Administration,
Editor. 2010.

Burke, 8., S. Milberg, and W. Moe, Displaying common but previously neglected
health claims on product labels: understanding competitive advantages,
deception, and education. Journal of Public Policy & Marketing, 1997. 16(2): p.
242-255.

Viswanathan, M., M. Hastak, and R. Gau, Understanding and facilitating the
usage of nutritional labels by low-literate consumers. Journal of Public Policy &
Marketing, 2009. 28(2): p. 135-145.

Sagall, R.J., Front of Package Labeling: Helpful or Harmful? Pediatrics for
Parents, 2010. 26(1&2): p. 1.

Stamier, R., Implication of the intersalt Stuay. hYPERTENSION, 1991 . 17(1): p.
5.

102

60.

61.

62.

63.

64.
65.
66.
67.
68.

69.

70.

71.

72.

73.

74.

75.

Larsson, I. and L. Lissner, The 'Green Keyhole 'nutritional campaign in Sweden:
do women with more knowledge have better dietary practices? European journal
of clinical nutrition, 1996. 50(5): p. 323.

Kinnunen, T., The Heart symbol: a new food labelling system in Finland.
Nutrition Bulletin, 2000. 25(4): p. 335-339.

Review of front of pack nutrition schemes. Heartnetconnection, 2007.

Schor, D., et al., Nutrition Facts You Can 't Miss: The Evolution of Front-of-Pack

Labeling: Providing Consumers With Tools to Help Select Foods and Beverages
to Encourage More Healthfid Diets. Nutrition Today, 2010. 45(1): p. 22.

GDA labelling. 2010 [cited 2010.

Guideline Daily Amounts (GDAs). 2006 [cited 2010.
Guideline Daily Amounts (GDAs). 2008 [cited 2010.
COMA report 1. 1991.

Fine, G., Give shoppers a clearer signal. A GDA-based food labelling scheme
used by some retailers and manufacturers, while helpful, lacks the directness of
the traflic light system. MARKETING WEEK, 2007: p. 29.

Traffic light labelling, Food Standards Agency.

Mannell, A., et al., French consumers' use of nutrition labels. Nutrition and Food
Science, 2006. 36(3): p. 159-168.

Langton, S., et al., Attention Capture by Faces. Cognition, 2008. 107(1): p. 330-
342.

Mack, A., et al., What we see: Inattention and the capture of attention by meaning.
Consciousness and Cognition, 2002. 11(4): p. 488-506.

Awh, E., et al., Evidence against a central bottleneck during the attentional blink:
Multiple channels for configural and faeatural processing. Cognitive Psychology,
2004. 48(1): p. 95-126.

Landau, A. and S. Bentin, Attentional and perceptual factors ajfecting the
attentional blink for faces and objects. Journal of Experimental Psychology:
Human Perception and Performance, 2008. 34(4): p. 818-830.

Bishop, 8., R. Jenkins, and A. Lawrence, Neural processing of fearfid faces:
Eflects of anxiety are gated by perceptual capacity limitations. Cerebral Cortex,
2007. 17(7): p. 1595-1603.

103

76.

77.

78.

79.

80.

81.

de Gelder, B., et al., Non-conscious recognition of aflect in the absence of striate
cortex. Neuroreport, 1999. 10(18): p. 3759-3763.

Whalen, P., et al., Masked Presentations of emotional facial expressions modulate
amygdala activity withou explicit knowledge. The Journal of Neuroscience, 1998.
18(1): p. 411-418.

Paivio, A., Mental comparisons involving abstract attributes. Memory and
Cognition, 1978. 6(3): p. 199-208.

Greene, M. and A. Oliva, The briefest of glances: The time course of natural
scene understanding. Psychological Science, 2009. 20(4): p. 464-472.

Shepard, R., Recognition, memory for words, sentences and pictures. Journal of
Verbal Learning and Verbal Behavior, 1967. 6(1): p. 156-163.

Becker, M. and I. Rasmussen, Guidance of attention to objects and locations by
long-term memory of natural scenes. Journal of Experimental Psychology:
Learning, memory and cognition, 2008. 34(6): p. 1325-1338.

104

ll

will

Hllllllllllllllll

3063

93

 

will”
U“
511i.
M1
“"2
N“
A“1
Wu“
H“