THE EFFECT OF PACKAGE DISPENSER
AND LABEL ON SUNSCREEN APPLICATION
By
Alyssa Lee Harben

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Packaging — Master of Science
2017

ABSTRACT
THE EFFECT OF PACKAGE DISPENSER
AND LABEL ON SUNSCREEN APPLICATION
By
Alyssa Lee Harben
As skin cancer rates rise, sunscreen use is recommended as a way to decrease the
risk of developing skin cancer. Studies show that most consumers of sunscreen do
not apply the recommended amount of sunscreen to achieve full protection from UV
rays. This study investigates the effect of using two different package dispenser
types on human behavior in the context of sunscreen application in an effort to
quantify differences in application volume. Additionally, labeling interventions
designed to either encourage or discourage application of sunscreen were placed on
the bottle to test the effectiveness of using label messaging to influence application
amounts. Participants were asked to apply sunscreen from different package styles
to their arms and legs and complete a survey regarding the perceived risk of
developing skin cancer and frequency of sunscreen use and other demographic
factors. Results indicate that participants applied more sunscreen from the squeeze
bottles than the pump bottles (p<.0001), and that there was no evidence of a
discernable difference between different labels. Of the personal characteristics
included in the model (including age, gender, skin tone saturation, history of sun
burn, and study location) worry about developing skin cancer and frequency of
sunscreen use were also significant at α=.05. The overall findings of this study was
packaging design affects sunscreen application practices, but labeling warnings are
less influential.

Copyright by
ALYSSA LEE HARBEN
2017

ACKNOWLEDGEMENTS

“ To all the little girls watching… never doubt that you are valuable and powerful
and deserving of every chance and opportunity in the world.”
-

Hillary Rodham Clinton

I am so grateful to have the chance and opportunity to attend graduate
school at Michigan State University. I’d like to thank my advisor Dr. Laura Bix for
everything she has taught me about academia so far. I cannot thank you enough for
your advise, encouragement, and insight and ideas. Your passion for research and
learning is contagious. Dr. Twede and Dr. Huddleston, your insight and knowledge
made me a better researcher. Dr. de la Fuente, your research on modeling the
intersection of design, psychology, and packaging is what sparked my desire to go to
graduate school, and I am so grateful for your influence on my academic career.
I would like to thank my family: my parents and grandparents for their
financial and emotional support throughout the entire course of my education, but
especially my undergraduate and graduate school years, and my aunt for taking me
on my first tour of MSU and encouraging me to Go Green.
I’d also like to thank the rest of the HUB team for the support and laughter
throughout the past two years. Paula and Eric, thank you for taking the time to help
with my data collection.
Luke, thank you for all of the emotional support over the course of this
project. You keep me sane.

iv

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
Chapter 1 Introduction
Research Objectives

vi
vii
1
4

Chapter 2 Literature Review
Regulation of Sunscreen
Perception of Warning Labels
Fear Motivation and Risk Perception
Warning and Personal Characteristics in Packaging
The Children and Sunscreen Study

5
5
6
9
11
13

Chapter 3 Methods and Materials
Materials
Recruitment of Participants
Sample Preparation
Testing
Experimental Design
Statistical Analysis
Surface Area Estimation

15
16
16
18
22
26
26
27

Chapter 4 Results and Discussion
Summary Statistics
Statistical Model
Package Effect
Label Recognition
Worry Effect
Frequency of Sunscreen Use Effect

29
29
31
32
36
37
39

Chapter 5 Conclusions, Limitations and Ideas for Future Research
Ideas for Future Research
Limitations

42
42
43

APPENDICES
Appendix A Approved Forms
Appendix B Sunscreen Labels Used in Study
Appendix C Test Plan

45
46
53
58

REFERENCES

60

LIST OF TABLES

Table 1. Attributes that Impact Warning Perception ……………………………………….………7
Table 2. Age and Sex of Participants by Location …………………………………………….…….29
Table 3. Results from ANOVA of the Full Random Effects Model………………………………31
Table 4. Predicted Mean Application Amounts ……………………………………………………...33
Table 5. Predicted Means Level of Worry……………………………………………….……………...36
Table 6. Tukey’s HSD for Frequency of Sunscreen Use…………………………………………….39

vi

LIST OF FIGURES

Figure 1. Package Label Combinations……………………………………………….………….………...18
Figure 2. “Encouraging Label” describing effective dosing ……………………………………….19
Figure 3. “Discouraging Label” describing the presence of nanoparticles…………………20
Figure 4. Fill Levels of Sunscreen Bottles ……………………………………………….…………….…..21
Figure 5. Predicted Mean Amount of Sunscreen, Package and Label Pairwise ………….32
Comparisons
Figure 6. Anatomy of the SKS Pump 2585 Series……………….……………………………………..35
Figure 7. Amount of Sunscreen Applied by Level of Worry ………………………………………..38
Figure 8. Predicted Amount of Sunscreen Use by Frequency of Sunscreen Use…………...40
Figure 9. Incorrect Option for Survey Response #1………………………………..………..…...…...54
Figure 10. Incorrect Option for Survey Response #2…………………………………....…………...55
Figure 11. Incorrect Option for Survey Response #3………………………………..…..…………...56
Figure 12. Incorrect Option for Survey Response #4………………………………..…..…………...57

vii

Chapter 1 Introduction

Protecting the skin from the sun is not a modern notion. Ancient Egyptians,
Greeks, and Native Americans utilized a variety of plants and oils in attempts to
protect their skin from the sun (Aldahan, AS et. all). Modern sunscreen has advanced
in formulation and added a secondary function. The earliest and primary function of
sunscreen is to prevent sunburn, and the second is to prevent skin cancer and
premature signs of aging. The precursors to modern sunscreen, lotions designed to
prevent sunburn, were developed in the 1930’s, modern broad-spectrum sunscreen
was developed in the 1970’s (Aldahan, AS et. all ). Broad Spectrum refers to the
ability of the lotion to protect the skin from both ultraviolet A radiation (UVA) and
ultraviolet B radiation (UVB). While both UVA and UVB can cause skin cancer,
premature aging, and sunburn, UVB is the primary source of sunburn (FDA Sheds
Light on Sunscreen 2011). UVA has been found to have a stronger link to skin cancer
than UVB. (Kuritky, L. A and Beeker J., 2015). There are two subtypes of broadspectrum sunscreen: physical sunscreens with the ingredients zinc or titanium
oxide that work by physically blocking radiation from penetrating the skin, and
chemical sunscreens with UV filtering chemical compounds derived from carbon
that work by absorbing the harmful radiation (Mitchell, Heidi, 2014). The level of
protection the sunscreen product provides is measured in the unit of Sun Protection
Factor (SPF), and the relationship between the SPF value and the amount of
protection provided is exponential (Faurschou & Wulf, 2007).
Regular use of sunscreen is recommended to reduce the risk of sunburn and
the development of skin cancers. Routine use of sunscreen starting in childhood has
1

been found to reduce the risk of developing non-melanoma skin cancers by as much
as 78%, (Stern, Weinstein, & Baker, 1986) but, like most drug products, there are
also inherent risks to consumers which accompany the benefits.
There are multiple, documented risks associated with sunscreen usage. The
primary risk is that people do not apply adequate amounts of sunscreen, demanded
via application requirements, to achieve the appropriate SPF. SPF is calculated
based on the assumption that consumers will apply a layer of sunscreen that is
2mg/cm, but in reality consumers have been measured to apply approximately 0 .5
mg/cm2 (Srinivas, C et. all, 2006)(Isedeh, Osterwalder, & Lim, 2013).Improper
dosage of sunscreen can lead to a myriad of unwanted effects associated with a lack
of full protection. Without the application of a full dosage amount, the sunscreen is
rendered less effective at preventing skin cancer, premature signs of aging, and
sunburn (Stern, Weinstein, & Baker, 1986).The amount of sunscreen applied by the
average consumer has been indicated to be the equivalent of applying a sunscreen
with a quarter of the SPF reported on the label (Kuritky, L. A and Beeker J., 2015). In
addition to proper dosage, in order to maintain full sun protection, reapplication of
sunscreen is recommended every 2 hours, or after sweating or swimming (FDA
Sheds Light on Sunscreen, 2011).
Secondary risks linked to sunscreen use concern factors related to both
consumer health and the environment. It has been suggested that the UV filtering
ingredients contained within some sunscreens have the potential to, in high doses,
act as endocrine disruptors. Recently published research on the effect of UV filtering
ingredients on the development of Chironomus riparius, a common insect used for

2

chemical testing, found an increased risk of disruption of stress and sex hormones
during the embryotic phase compared to the larvae stage (Ozáez, Morcillo, &
Martínez-Guitarte, 2016)(Krause et al., 2012). While evidence is not conclusive that
endocrine disruptors are a proven risk to humans, there is evidence that compounds
found in sunscreen act as harmful endocrine disruptors in small mammals,
amphibians, insects and fish populations (Waring & Harris, 2005). The specific UV
filtering ingredients that are thought to be Endocrine Disruptors are:
benzophenone-3 (BP-3), 3-benzylidene camphor (3-BC), 3-(4- methyl-benzylidene)
camphor (4-MBC), 2-ethylhexyl 4-methoxy cinnamate (OMC), Homosalate (HMS), 2ethylhexyl 4-dimethylaminobenzoate (OD- PABA) and 4-aminobenzoic acid (PABA).
These ingredients are absorbed through the skin when sunscreen is applied, and
have been found in urine and breast milk samples. 8 Additionally, current research is
exploring whether or not vulnerable populations, such pediatric users, are at
particular risk of harm.

3

Research Objectives
Since there are potential risks associated with both applying either too little
sunscreen or too much sun protection, we became interested in the factors that
affect application amount; specifically, the influence of label warnings and the
physical structure of the package. What is the effect of packaging type on the
amount of sunscreen applied? What is the effect of an alarmist label that alerts the
consumer to possible risks associated with nanoparticles compared to a label
encouraging generous application on sunscreen usage? What is the combined effect
of packaging and labeling? What subject characteristics influence application?
Namely, does personal history (e.g. skin tone, family history of skin cancer,
experience with severe sunburns) significantly affect the amount of sunscreen
applied? This research will investigate these questions.

4

Chapter 2 Literature Review
Regulation of Sunscreen
In the United States, the Food and Drug Administration (FDA) regulates
sunscreen as an over the counter drug (OTC). Sunscreen was first regulated as a
drug in 1978 (FDA “Sunscreen Drug Products for Over-The-Counter Human Use,”
1978). The Final Rule [21 CFR Part 352], “Sunscreen Drug Products for Over-TheCounter Human Use,” details the ingredients generally recognized as safe (GRAS),
labeling requirements specific to products with varied levels of UV protection, and
testing procedures for getting new products approved, but does not include product
packaging suggestions or discussion (FDA “Sunscreen Drug Products for Over-TheCounter Human Use,”1978). Labeling requirements include:
•

A drug facts panel

•

The sun protection factor (SPF) within a range of SPF 15 to SPF 50+

•

Water resistance claims with specific time limits

•

A warning alerting the consumer the product does not protect against
skin cancer or aging if the product has an SPF between 2-14.

Additionally, the regulation indicates that manufacturers may not use the phrases
“sunblock”, “sweat proof”, or “waterproof” since those terms have been deemed
misleading to the consumer (FDA Sheds Light on Sunscreen, 2011). In addition to
these labeling guidelines, there are guidance documents published in the Federal
Register with lists of approved ingredients and current best practices for
manufacturing (FDA “Sunscreen Drug Products for Over-The-Counter Human Use”,
1978) More recently (2011), the FDA expressed interest in better understanding the

5

relationship of packaging for these products and consumer behavior. Specifically,
how different dispenser types, primarily lotion dispensers, compare to both manual
and aerosol spray dispensers, and, ultimately, the influence they have on consumers’
usage of sunscreen products (Tan, 2011). Our review of the literature found a
dearth of information on this topic, in spite of the need expressed by the Agency.
There are a variety of sunscreen products available on the market, but the
FDA only allows those products in oil, cream, lotion, gel, butter, paste, ointment,
stick, or spray form without special approval. The FDA recognizes the following
delivery mechanisms for sunscreen as needing special approval before marketing:
powder, wipe, towellete, shampoo or body wash form (Tan, 2011). The number of
different application forms available makes the question of how varied package
designs influence consumer adherence to protection guidelines important.
Perception of Warning Labels
The warning process is described by Rogers, Lamson, and Rousseau as “four
components: notice, encode, comprehend, and comply,” (Rogers, Lamson, &
Rousseau, 2000). Other models, such as the Model of the Effects of Product Warning
Labels proposed by DeTurck, Rachlin, and Young, have expanded the four steps to
include a preliminary step of exposure (deTurck, Rachlin, & Young, 1994). The
Model of the Effects of Product Warning Labels is composed of the following steps:
exposure, awareness, comprehension, perception of risk, and compliance (deTurck,
Rachlin, & Young, 1994). Under these models, steps are serialized; that is, each is
prerequisite to the next. Researchers have expanded on the notion of the different
stages of the warning process to include the idea that warnings consist of three

6

parts: a signal word to alert, a description of the hazard, and instructions for
compliance (Rogers, Lamson, & Rousseau, 2000). The signal word, related to the
noticing or awareness step is intended to attract attention to the warning, which
will include a description, information about any risks and instructions for
mitigating the same. The ultimate intention is compliance, whereby the consumer
uses actionable steps to mitigate the risk.
Rogers et al. propose there are two, broad categories of input variables that
affect warning perception: personal variables, which are specific to the individual
viewers, and warning variables, that are unique to the specific warnings being
perceived. Some examples of personal and warning attributes are included below in
Table 1, but these lists are not exhaustive (Rogers, Lamson, & Rousseau, 2000).
Table 1. Attributes that Impact Warning Perception1

Personal variables include things like: age, familiarity, hazard perception,
gender, vision, cost of compliance, risk-taking style, and level of health literacy

1

Adapted from Rogers et al
7

(Rogers, Lamson, & Rousseau, 2000)(Davis et al., 2006). Warning variables are
attributes of how the warning is presented. Examples include: color, emphasis,
interactivity, placement, shape, size, signal word, explicitness, font size, layout,
length, text complexity, and tone (Rogers, Lamson, & Rousseau, 2000). Attributes of
both the warning and its target audience interact to produce different outcomes. For
example, if a person with vision impairment is the target of the warning, warning
variables such as font size are of paramount importance. Similarly, if a person with
low literacy skills is presented a warning at a reading level beyond their capabilities,
the effectiveness of the symbology utilized is vital.
One of the most important factors in warning perception is hazard
perception. Hazard perception is defined as a “subjective level of danger” (Rogers,
Lamson, & Rousseau, 2000), which can be described as the perceived likelihood of
risk or how dangerous a person perceives an object or action to be, as opposed to
how dangerous an object or action actually is. Wogalter, Brelsford, Desaulniers and
Laughery found that the willingness of a person to read the warning is positively
correlated with the perception of hazard despite level of familiarity (Wogalter,
Brelsford, Desaulniers, & Laughery, 1991). This is crucial to warning perception
because the literature shows that people tend to be less likely to notice warnings on
products they are comfortable with (Rogers, Lamson, & Rousseau, 2000).This
phenomenon could partially explain the number of people that noticed, and encoded
the warnings on the sunscreen bottles tested in this study, since familiarity with
sunscreen was one of the screening requirements for the subject pool. (This will be
discussed in more detail in Chapters 3 and 4.)

8

Fear Motivation and Risk Perception
Fear motivation describes behavior driven by an aroused emotional state
due to the threat of impending harm (Maddux & Rogers, 1983). Feather defines Fear
Motivation as “motivation to avoid a negative incentive or punishment,”(Feather,
1963).This theory assumes humans will act in their own self-interest, through
actions with the goal of protecting themselves from the impending harm (Maddux &
Rogers, 1983). Within the context of this study, Fear Motivation relates to acting to
avoid the negative consequences of unprotected UV exposure, such as a sunburn or
skin cancer. Whether or not they choose to act depends on their self-efficacy, or
belief about their own capability to perform the behavior presented to them as
necessary. Bandura argues that the stronger the self-efficacy, the more likely they
are to make a change (Bandura & Adams, 1977). Thus, self-efficacy is positively
linked to behavioral change, meaning as self-efficacy increases, the likelihood that a
person will follow through on changing a behavior also increases (Feather, 1963).
Similarly, whether or not one continues to act in a manner motivated by fear
depends on whether or not their fears are confirmed(Maddux & Rogers, 1983). In
the context of this study, someone who does not wear sunscreen regularly but does
not get sunburned, might not be motivated to apply sunscreen, even if avoiding
sunburn did not influence their risk of developing skin cancer.
Related to the Theory of Fear Motivation is the Theory of Risk Perception.
Risk Perception is the study of how humans behave based on how likely they
perceive the potential for harm. The fourth wave of risk perception research focused
on subjective attributes of individuals, such as having the intent to act cautiously or

9

perceive hazards (Kluckhohn, 1962). Risk perception is seen as a tie between
planned behavior and desired health outcomes, and has been shown to be an
important factor in both the planning and execution of sun protection behaviors
(Craciun, C et. all 2010). Self-efficacy mediates risk perception and planned
behavior. If something, such as sun exposure, is perceived as risky and the person
has high self-efficacy, meaning they believe they are capable of mitigating that risk,
planned behavior is more likely to be implemented (Feather, 1963).
There are many ways of manipulating risk perception. One of the major
approaches to manipulate perception of risk, and germane to this study, is via
dissemination of information regarding the possibility of risk and how that
information affects the behavioral outcome. How familiar a consumer is with the
nature of the product and what information is included on the product warning
labels affect the perception of risk (Leonard & Hill, 1993). In a study examining
consumer perception of risks associated with genetically engineered soy beans,
researchers found that if the benefits associated with a product or technology are
high, the product or technology was perceived as low risk. Meaning, perceived risk
can be manipulated by a third factor, simply changing the perceived benefits (Lynne
& Ping (2003). This idea has been extrapolated to sunscreen by examining the
change in perceived risks based only on the inclusion of warning labeling. One study
used labeling of sunscreens made with nanotechnology as a proxy variable for the
labeling of genetically modified foods (Siegrist & Keller, 2011).
Siegrist and Keller split participants into six groups and varied the labeling
stimuli each group was exposed to. The different stimuli were: a picture of a

10

sunscreen bottle without a nanotechnology label (control), a picture of a sunscreen
bottle with a nanotechnology label, and then four different combinations of the
picture of a sunscreen bottle with the nanotechnology label and either literature
describing: nanotechnology in general terms, the same general information
combined with information about the risks, only the risk literature, or literature on
the benefits of nanotechnology. Thus, there were six treatments: control, the
treatment picture without literature, and then four different treatment pictures and
literature combinations. While there was a statistically significant difference
between the control sunscreen label and the experimental labels alerting the
consumer to the presence of nanoparticles, no statistical difference was evident
between the perceived risks of the five bottles of sunscreen with the label indicating
nanotechnology, despite the different educational pamphlets included with the
different treatments. In other words, simply adding the label alerting consumers to
the presence of nanoparticles increased the perceived risk, whether or not
additional literature about risks or benefits was included (Siegrist & Keller, 2011).
Warning and Personal Characteristics in Packaging
There are many components of packaging that influence a person’s
perception of the product. The size, shape, color, labeling, and context all convey
different information about what the package contains, how to access the product,
what to do with the empty package after its contents are used, and many more
possible messages (Silayoi & Speece, 2007). An optimally designed package
immediately conveys information about opening, dispensing, usage, closing, storage,
and disposal (De La Fuente, Gustafson, Twomey, & Bix, 2015). The context

11

surrounding a task, for instance, time constraints facing the consumer and the
number of alternatives surrounding the product, affect the interaction that occur
and the behaviors that are exhibited (Silayoi & Speece, 2007). The definition of
context need not be narrowly defined as those encompassed in the physical
environment, but can also include things within the social context including who the
product will be used by or the opinion of others (De La Fuente, 2013).
One theory about the interactions between human processing systems and
package messaging systems is the Human Package Interaction Model (HPIM).
Proposed by de la Fuente (De La Fuente, 2013). The model purports people interact
with packaging using perception, cognition, and motor functions to process
information using five steps exposure, perception, encodation, comprehension, and
execution to accomplish a given task, such as dispensing the product (De La Fuente,
2013).
HPIM, when applied in the context of package dispensing mechanisms,
interacts with Affordance Theory to provide a theory that explains consumer
behavior and package use. According to Affordance theory, as applied to physical
objects by Don Norman, an affordance can be defined as, “the relationship between
a physical object and a person…An affordance is a relationship between the
properties of an object and the capabilities of the agent that determine just how the
object could possibly be used,”(Norman, D. A. 2013). In an earlier article, Norman
states, “ Affordances specify the range of possible activities, but affordances are of
little use if they are not visible to the users,” (Norman, 1999). A perceived
affordance differs from the traditional idea of an affordance by having the additional

12

requirement of allowing for the perception of a specific intended action, rather than
traditional idea of only the capability to preform an action (Norman, 2004). For
example, a handle affords gripping or a button affords pushing. Within the context of
this study, the pump bottle will afford metered dispensing of the product more
strongly than the bottle without a pump.
The different components of a package communicate different information
about the different actions the package affords the user to take, which is why the
messaging, both covert and overt, is important to consider during the design
process (Norman, 2004). Thus, the label on the bottle is not the only component of a
sunscreen package that communicates information to the consumer, each
component should be chosen carefully.
The Children and Sunscreen Study
In 2012, researchers at the University of Queensland, in the public health
department launched an investigation into the application of children’s sunscreen.
Researchers investigated the effects the package dispenser type and the age of the
child on the thickness of the application (Diaz, Neale, Kimlin, Jones, & Janda, 2012).
Families that participated were given three packaging alternatives filled with
sunscreen, and the directions to use one each week for a period of three weeks. The
package alternatives were a squeeze bottle, a pump bottle and a roll on; all
alternatives were different sizes and volumes. The child participants were
instructed to apply the sunscreen to themselves as their first daily application. The
children that participated in this study ranged in age from 7-12 years old.

13

The methods used in the Queensland study to calculate the amount of
sunscreen applied by the children involved weighing the packages and using a
formula to calculate the surface area of the entire body using height and weight. The
sunscreen was weighed before it was distributed to the participants, and again at
the end of the study. The differences in weight of the packages were used in
conjunction with the surface area of the body to calculate an average application
thickness. Since no researchers supervised or weighed the packages while the
participants were using them, and the packages were with the families for a period
of three weeks, the thickness of the individual applications unable to be known.
The study concluded that package type statistically significantly affected the
amount of sunscreen the children applied to themselves, regardless of age or other
demographic variables. The pump had the thickest average application, followed by
the squeeze bottle. The roll on had the thinnest average application. The children
still only applied about half of the recommended dosage; even with the pump bottle
(Diaz et al., 2012). The study presented herein differs from the Children and
Sunscreen Study by targeting an adult population and looking at the covariant of
labeling. The hypothesis that the package type will influence application thickness is
based on The Children and Sunscreen Study.

14

Chapter 3 Methods and Materials
In order to better understand the potential effects of packaging and labeling
on the amount of sunscreen applied by users, 96 research participants were
recruited to participate in the study presented herein during the summer of 2016.
Forty-eight participants between the ages of 18- 36 were recruited using the College
of Communication Arts and Science SONA system and tested in laboratories at
Michigan State University (MSU). A second population of subjects was recruited
California Polytechnic State University through in-class announcement provided in
the Graphic Communications course “Consumer Packaging”, the Industrial
Technology and Packaging courses “Packaging Fundamentals” and “Supply Chain
Management in Manufacturing and Services”, and the Business course “Information
Systems”.
The objective of this study was to investigate the effects of packaging and
labeling to better understand how consumers of sunscreen determine the amount of
sunscreen to use, and what factors influence that amount of sunscreen they selfapply. Specifically, a dosing pump system was compared to a non-dosing squeeze
bottle; these were crossed with two levels of label messaging (Figure 1), for a total
of four test treatments.

15

Materials
•

8 oz. Natural Low Density Polyethylene Boston Round Bottles, SKS Bottle &
Packaging International (Watervliet, New York)

•

Polypropylene Pumps, SKS Bottle and Packaging International

•

Polypropylene Flip Top Spout Caps, SKS Bottle and Packaging International

•

Fisher Science Education TM Portable Balance, Capacity: 300g, Readability:
0.001g

•

Rocky Mountain Sunscreen’s SPF 30 Kids Broad Spectrum Sunscreen

•

Adhesive Vinyl labels, FedEx (Memphis, Tennessee)

•

Measuring Tape

•

Baby Wipes

Recruitment of Participants
A total of 96 participants were recruited from the two locations of test,
Michigan State University and California Polytechnic State University SLO.
Participants were recruited through multiple methods in accordance with the
Internal Review Board (IRB) approvals from both Michigan State and California
Polytechnic State University of San Luis Obispo documented as #16-574. At
Michigan State University, the primary method of recruiting students was the
College of Communication Arts and Science SONA System. The study was posted to
the paid pool with an incentive of $10 and to the student pool with an incentive of
0.75 SONA credit. At California Polytechnic State University SLO, participants were
recruited by presenting the opportunity to participate in the study in the multiple
classes listed above and received $10 as an incentive to participate in the study.

16

In order to be eligible to participate in the study, subjects had to be between
the ages of 18 and 36 years old, be sunscreen users, and have no history of a skin
condition such as eczema. Participants between the ages of 18 and 36 belong to the
“Millennial” generation2. This generational cohort was chosen to assure that
consumers were familiar with the application of sunscreen, since they were all born
after sunscreen was first regulated by the FDA and were, thus, likely to have lifetime
sunscreen usage. Exclusion of participants with history of skin conditions was to
protect subjects with sensitive skin from using an unfamiliar product that could
potentially result in an adverse reaction. Participants were asked to wear clothing
that would facilitate the application of sunscreen to their entire arms and lower legs.
T-shirt or tank tops and shorts or knee length skirts were recommended, but it was
stressed that any clothing that could be rolled up to allow comfortable application of
sunscreen to the entire arm and lower leg was also allowed.
The IRB approved documents, including the consent form that details
eligibility criteria, the data sheet created for the study, and the demographic and
health history questionnaire the participants were asked to fill out as part of their
participation, are included in Appendix A.

2

Millennial Generation was defined as born between 1980 and 1998. This cohort
was chosen because it the oldest members were born after sunscreen was first
regulated by the FDA and adopted for wider use, and the youngest members are of
the age of majority (18 years old).
17

Sample Preparation
Four complete sets of the test packages were assembled. Each set of four
included a bottle with a pump (hereafter interchangeable with “Dosing Package”)
and a label advising the application of 9 teaspoons of sunscreen to the entire body
(hereafter referred to as “Encouraging Label”)(Figure 2), a bottle with a pump and a
label indicating the product contained nanoparticles (hereafter referred to as
“Discouraging Label”)(Figure 3), a bottle with a flip top (hereafter interchangeable
with “Non-Dosing Package”) advising the application of 9 teaspoons of sunscreen,
and a bottle with a flip top indicating it contained nanoparticles.
Figure 1. Package Label Combinations

18

Figure 2. “Encouraging Label” describing effective dosing

19

Figure 3. “Discouraging Label” describing the presence of nanoparticles

20

While each block of test packages consisted of four unique
bottle/label/dispenser combinations, (two levels of physical structure crossed with
two levels of label treatment), a total of sixteen test packages were prepared to
rotate throughout the testing in case of any damage to the sunscreen bottles or
labels. All bottles were filled to the top of the display panel (i.e. at the shoulder of the
bottle, see Figure 4) with the Rocky Mountain Sunscreen between uses. Physical
structure of the bottles was chosen as the fill level to account for undetectable
inconsistencies with label application and the level of fill was chosen to account for
any displacement from adding the pump. The full bottle of sunscreen weighed
approximately 298 g. The labels were designed to be the exact height of the display
panel of the bottle and wrapped fully around the circumference of the bottle.
Figure 4. Fill Levels of Sunscreen Bottles

As part of staging between participants, filled bottles of sunscreen were
weighed and weight was recorded on the data sheet with the corresponding
participant number. Staging also included laying out the labels for the order of

21

sunscreen application to each limb (see Appendix C for the detailed list of the orders
sunscreen were applied), and refilling the bottles if their weight dropped below
280g.
Additional preparation for the study included preparing stimuli for the label
recall task in the survey by laminating actual size print outs of four additional labels
in an attempt to verify that subjects actually viewed each label. Laminated labels
indicated as A-F and displayed with laminated print outs of the labels used in the
study where the participants sat to fill in the demographic survey. Lastly, in order to
help the participants remember the order of body parts to apply the sunscreen to,
place cards were for the sunscreen to clarify the application area (left arm, right
arm, right leg, left leg). The place cards were rotated as part of the staging for each
participant to reflect the order in which they were to apply the sunscreen.
Testing
When participants arrived at the testing facility, the researcher provided
them with a consent form and verbally indicated the screening criteria, time
commitment, incentive and that they could withdraw at anytime or refuse any
testing that made them uncomfortable without penalty. The sunscreen was set up,
with the labels facing the participant, in the predetermined application order (based
on the predetermined counterbalanced scheme- see Appendix C) presented to them
in the order from left to right, on a table. The bottles of sunscreen were set upon
place cards with the body part the sunscreen was intended for. The table also had
baby wipes, access to a trashcan, and a chair or stool.

22

After the participants read and signed the consent form, the researcher asked
their handedness and walked them through a short exercise in order to determine
laterality. The term laterality, often interchangeable with “limb preference” is ,”used
to express the preferential use of one limb in voluntary motor acts,” (Sadeghi,
Allard, Prince, & Labelle, 2000). An example of laterality would be the leg that tends
to be on top while sitting cross-legged or the thumb that is on top when hands are
clasped. Laterality was determined by asking the participants to clasp their hands
together and report which thumb was on top and then to cross their arms and
report which arm was on top. This was measured to potentially use as a variable in
the model if order of application, and thus which limbs were used to apply the
sunscreen, influenced application amount . Handedness was reported as right, left,
or ambidextrous. Laterality was reported as “right-right”, “right-left”, “left-right”, or
“left-left.” For example, if the right thumb was on top when the hands are clasped,
and the left arm is on top when the arms are crossed, laterality would be “right-left”.
This information was recorded on the data collection sheet in Appendix A.
The final information collected before the application of sunscreen was
measurements of the circumference of the upper arm (right above the start of the
bicep), circumference of the wrist, the length of the arm (from shoulder to wrist),
the circumference of the calf (directly below the knee), the circumference of the
ankle, and the length of the lower leg (from the bottom of the knee to the ankle).
These measurements were recorded from a tape measure with accuracy to the
nearest 1/8 of an inch. This information was collected to estimate the surface area
where the participants would be applying sunscreen.

23

The participants were directed to, “Apply the sunscreen to their arms and
lower legs as if they would be spending the entire day outside on a sunny day”. Due
to the similarity of the label designs, the researcher also instructed the participants,
”While these packages all look the same at first glance, they are different. Please
look at each package before applying the sunscreen.” The test plan was developed
with the intention of analyzing the differences within each participant rather than
across each participant. Because of this, each participant was provided each
package-label combination to apply. Since there are four packages with different
dispenser-label combinations, each package would be applied a single time, and
each limb would receive one application of sunscreen. The sunscreen bottles were
refilled between participants once the weight of the package was below 280g.
Each possible order in which each participant could apply the four packages
of sunscreen was assigned randomly to a number between 1 and 24. Likewise, the
order of which limb the sunscreen could be applied to was also assigned to a
number between 1 and 24. The participant number within each block of 24
participants corresponded with the application orders. The test plan is included in
Appendix C to better illustrate how the order of application rotated throughout each
block of 24 participants. In order to clearly communicate the instructions to the
participants, the sunscreen packages were placed on labels with the body part
receiving the sunscreen laid out in the order of application starting from right to left.
To further clarify which package of sunscreen to use next, the researcher removed
the bottles as the participants finished each round of application. Baby wipes were

24

provided to the participants to clean their hands between each application of
sunscreen.
After applying the sunscreen, participants were asked to complete a twelvequestion survey. The survey is included in Appendix A as part of the IRB-approved
documents. In an effort to screen subjects that had failed to access the label
information, the survey started by asking the participants to identify which
sunscreens they applied out of six possible sunscreen label possibilities. These six
included the two labels that subjects had, indeed seen (figures 2 and 3), and four
that they had not (Appendix B). The survey also contained questions pertaining to
their personal history of sun burns and skin cancer, frequency regarding sunscreen
usage, their perception of risk for developing skin cancer, and demographic
questions related to age, gender, skin tone, and parental status. Skin tone was
determined by asking the participants to self identify the option that best matched
their skin tone from Pantone SkinTone Guide (Pantone LLC Carlstadt, NJ), and
record the corresponding value.
The sunscreen packages were weighed after the application of sunscreen,
and the final weight was recorded. Notes and observations were also recorded if
anything out of the ordinary happened in that particular trial. For example, one
participant mistakenly applied sunscreen to the same leg twice, and that error was
noted. The total amount of sunscreen applied was calculated by subtracting the final
weight of the packaging from the initial weight of the package, as recorded at the
prior to the participant’s arrival. Any data that was determined to have an

25

application amount of less than zero grams of sunscreen was removed from analysis
due to human error in the recording of the data.
Experimental Design
The experiment was designed as a Counterbalanced Randomized Complete
Block. Participants were randomly assigned to the order in which to apply the
sunscreens to their limbs based on when they arrived to participate in the study.
Since there were 24 possible combinations for both the order of using the different
package-label combinations and the order of the limbs sunscreen was applied to,
blocks of 24 participants were recruited. Each order of package-label was married
to the order of limbs sunscreen was applied to. Post-hoc analysis did not reveal a
run order effect on the amount of sunscreen applied and the response was analyzed
in units of thickness of application in order to eliminate any application differences
due to limb size differences.
Statistical Analysis
The collected data was analyzed using R (R Core Team 2016). Data was fit to
a Random Effects Model (Also referred to as a Mixed Linear Model). The response
variable of amount of sunscreen applied was analyzed as a thickness in the unit of
mg/cm2, and the categorical predictor variables were selected based on correlations
with the response and variability within the collected data. Some variables were
reverse coded in order to correspond correctly with the other variables included in
the model. The Binomial Distribution was used to calculate the probability of 55%
correctly identifying the labels by random chance in order to include the labels as a
predictor variable.

26

Surface Area Estimation
The surface area of the limbs the sunscreen was applied to was estimated by
taking measurements and then treating arms and legs mathematically as truncated
cones.
Equation 1. Surface Area of a Truncated Cone
=
Where: R= Greater Radius

[

+

+ 2 +

r= less radius

]
s= slant height

The measurements used in the arm surface area estimations are: the
circumference of the upper arm (around the bicep), circumference of the wrist, and
the length from shoulder to wrist. The measurements used in the leg surface area
estimations are: the circumference of the calf directly below the knee, the
circumference of the ankle, and the length of the shin (from the bottom on the knee
to the ankle). The circumferences were used to solve for the radii to use in the
surface area of a truncated cone equation, and the length of the arm or shin was
used as the slant height. This estimation of surface area was used to standardize the
amount of sunscreen applied by the participants to the units of mg/cm2 commonly
used in sunscreen research.
The data was fit to a Random Effects Model (also referred to as a linear mixed
model) using R (Vienna, Austria). In the model illustrated by Equation 1, y is a
vector of the number of participants (94) multiplied by the number of observations
per participant (4). X is made up of a matrix of the predictor variables; the number
of participants (94) multiplied by the number of predictor variables (10). β is the
fixed effects vector made up of the coefficients for the predictor variables, and Z is

27

the matrix for the random effects that make this model different from a traditional
linear model. The error term, to explain the parts of Y that are not explained by the
rest of the model is ε, the vector of the residuals (Jones, G. 2011).
Equation 2. Random Effects Model
=

+

+

~

0, !"# )

!"# = .6644062

'~

0, !"( )

!"( = .5885008

The response variable was the thickness of the sunscreen application,
determined by dividing the weight of the applied sunscreen by the estimated
surface area of the body part the sunscreen was applied to, and the factors included
in the analysis were: package type, label messaging, whether or not they recognized
the labels during the survey, frequency of worry about developing skin cancer, skin
tone saturation, whether they know someone with skin cancer, frequency of
applying sunscreen when spending more than two hours outside in the sun, gender,
location of participation in the study, and age.
In order to select which of the factors generated by the survey data to include
in the statistical model, Pearson Correlations were run with all of the survey
response variables. To avoid redundancy and artificially inflating the number of
significant variables, if variables were highly correlated only one was selected to be
in the model. While interactions between the factors were tested while developing
the statistical model, no interaction terms were included in the final model because
none of the tested interactions were significant.

28

Chapter 4 Results and Discussion
Summary Statistics
Data collected from 94 participants were used for analysis after 2
participants were excluded from the final data set due to recording error; namely,
the weights recorded after usage were higher than those prior to use. Of those
included in the analysis, 47 participants were recruited from MSU and 47
participants from Cal Poly. The mean age for the total sample was 23 years old with
a median age of 21. The mean and median ages for the MSU portion of the sample
were 25 and 24 years old respectively, and the mean and median ages for the Cal
Poly sample were calculated as 20. The overall sample was 41% male and 59%
female, with the MSU portion of the sample being 29% male and 71% female and
the Cal Poly portion of the sample being 52% male and 48% female. Two
participants reported having children. All data was collected between July 6th
August 26th 2016.
Table 2. Age and Sex of Participants by Location
Total Sample

MSU Sample

Cal Poly Sample

Sample Size

94 (96)*

47 (48)*

47 (48)*

Mean Age (years)

23

25

20

Median Age (years)

21

24

20

Male

41%

29%

52%

Female

59%

71%

48%

*94 participants were included in the statistical analysis since two were excluded
from the study due to error in data recording.

29

Of the sample of 94 participants, 52 (55%) correctly identified the labels of
the sunscreen they applied and 42 (45%) did not. Sixteen participants (17%)
reported having experienced severe sunburns, 37 (39%) reported knowing
someone with skin cancer, and the most prevalent skin tone saturations, matched
from a Pantone SkinTone Guide (Pantone LLC Carlstadt, NJ) on a scale of 1(fair) to
15 (deep), each with 12 respondents, were saturation levels 2, 3, and 8.
When comparing the actual amount applied by participants in the study to
the amount recommended in the FDA guidelines for testing2, only 8 of the 94
participants (9%) applied sunscreen in the thickness denoted by the FDA (≥2
mg/cm2) for all four of their applications. In other words, less than 10% of study
participants consistently applied enough sunscreen in each application to truly have
30 SPF as the protection level from UV rays. None of the predicted means (see figure
5 below) were above the recommended level for sunscreen thickness.

30

Statistical Model
After fitting the random effects model to the data using the maximum
likelihood method, a type three analysis of variance (ANOVA) was run to determine
if there were any significant factors. The results of this ANOVA are reported below
in Table 3. The significant factors: package type, self-reported worry about
developing skin cancer, and self-reported frequency of sunscreen use when outside
for more than two hours on a sunny day, are reported in bold as determined by an
evaluation threshold of α=0.05.
Table 3. Results from ANOVA of the Full Random Effects Model
Factor

Num.
DF
1

Den. DF

F-value

P-value

280

23.5016

<. 0001

Label Type

1

280

0.0645

0.7997

Gender

1

68

0.0488

0.8259

Know Someone with Skin Cancer

1

68

0.0528

0.8189

Worry about developing Skin Cancer

3

68

3.2871

0.0258

Recognize Labels

1

68

1.5688

0.2147

Self-reported frequency of Sunscreen

4

68

2.9601

0.0258

Skin Tone Saturation

13

68

1.4044

0.1803

Age

1

68

0.6098

0.4376

Location

1

68

0.1942

0.6608

Package Type (Pump or Squeeze)

Use When Outside for 2 hours

The significant factors from the model will be discussed individually in further
detail. The predicted means were calculated using the Least Square Means function
in R.

31

Package Effect
The first significant factor that will be discussed is the effect of the different
physical structure on the amount of sunscreen that participants applied. The model
indicates that regardless of the label, participants applied 30% less sunscreen when
using the pump as compared to the squeeze bottle (p<.0001). This is further
illustrated in figure 5. Pairwise comparisons of treatments are shown in different
colors at α=0.05. This is an interesting finding because one of the hypotheses was
that a dosing dispensing system (i.e. a bottle outfitted with a pump dispenser) could
encourage consumers to use more of the product since it would be dispensed in a
relatively consistent amount each time, at similar fill levels.
Figure 5. Predicted Mean Amount of Sunscreen, Package and Label Pairwise
Comparisons
FDA Reccommended Thickness

2

Amount of sunscreen in mg/cm^2

1.8

B

A

1.6
1.4
1.2

1.086398

1.070375

Pump Bottle with
Encouraging Label

Pump Bottle with
Discouraging Label

1.376245

1.392268

Squeeze Bottle with
Encouraging Label

Squeeze Bottle with
Discouraging Label

1

0.8
0.6
0.4
0.2
0

The type of package was found to significantly affect the thickness of
sunscreen application, whereby participants applied more sunscreen from the
squeeze bottles when compared to the pump bottles. A ranking of the four

32

sunscreen package and label combinations shows participants applied the most
sunscreen from the squeeze bottle with the discouraging label, followed by the
squeeze bottle with the encouraging label, the pump bottle with the encouraging
label, and the least from the pump bottle with the discouraging label, but the
difference between the amounts applied from the same package with different
labels did not suggest significant differences.
Table 4. Predicted Mean Application Amounts
Package Label Combination
Pump Bottle with Encouraging Label

Predicted Mean Application Amount (mg/cm2)
1.086398A*

Pump Bottle with Discouraging Label

1.070375A

Squeeze Bottle with Encouraging Label

1.376245B

Squeeze Bottle with Discouraging Label

1.392268B

*Means followed by the same letter are not significantly different from each other using Tukey’s
HSD at a significance level of 0.05.

Affordance theory can be leveraged to offer insight into this finding about the
effect of the physical dispensing structure on the amount that participants applied.
It is possible that consumers interpreted the amount the pump metered out each
time it was pressed down as the correct amount of sunscreen to apply, despite one
single pump not being adequate to provide full protection to the arms or legs of
most adults. A 95% confidence interval for the amount of sunscreen in one pump
from the bottle is 491mg to 523mg; this suggests it would take approximately 3
pumps to adequately cover an arm or lower leg with the average surface areas
represented in this sample, 808.41cm2 and 691.04cm2 respectively.
The pump could be considered a misleading or false perceived affordance,
and thus ended up under-dosing when compared to the squeeze bottle in this
experiment. While the pump affording dispensing the product does not change

33

based on the amount in each pump, the perceived correct amount could be affected,
and that signal would lead to consumers stopping the dispensing process before
complying with application recommendations. The potential ramification of the
under-dosing pump was 91% of the participants were noncompliant with FDA
sunscreen application guidelines. Without further replication, the data implies a
squeeze bottle should be the preferred package for sunscreen to ensure greater
usage.
By contrast, results from Davis et al indicate that the use of a pump bottle
results in significantly greater amounts of product application as compared with a
squeeze tube and roll on dispenser. 30 We postulate two reasons the findings
presented here run counter to those reported by Davis et al. Firstly, the study
presented here utilized adults applying sunscreen to themselves while Davis et al.
investigated how packaging influences the sunscreen application behaviors of
children. Secondly, the average volume dispensed by the pump used on the pump
package in the Children and Sunscreen Study was not disclosed in the paper, so it is
unknown the exact difference in volume between the pump the children used and
the pump used in this study.
Figure 6, a specification graphic created by SKS Bottle & Packaging Inc, below
shows a cross section of the pump used in this study to dispense the sunscreen.
When the pump is primed, meaning the accumulator fills with product on each
stroke of the pump rather than air, the pump dispenses .5mL per stroke. For the
product used in this study, .5mL was equal to the roughly .5 grams dispensed per
pump. If the volume of the accumulator is greater, more product is drawn up the dip

34

tube with each stroke of the pump, and more product is dispensed. There are many
different pumps available to be used for products, and thus without the disclosure of
the delivery volume per stroke, it is unknown if the package used in the Children
and Sunscreen study dispensed more or less per pump than the package used in this
study.
Figure 6. Anatomy of the SKS Pump 2585 Series3

3

Pumps (2585 Series) reproduced with permission from SKS Bottle & Packaging Inc

35

At similar fill levels, a pump with a greater delivery volume per stroke would
implicitly communicate through affordances a greater quantity for the correct
dosage, and thus be more efficient at promoting compliance with sunscreen
application guidelines. While the fill level could influence how much is dispensed
per stroke by also drawing air up the dip tube, that impact is primarily for lower
levels of fill, when it is difficult to dispense the last bit of the product from most
packages.
Label Recognition
While the model did not show label messaging to be statistically significant, it
is important to note that there is evidence to indicate some participants did interact
with the labels. One of the questions in the survey (included in Appendix A) asked
participants to identify the two labels of the sunscreens applied during their
participation in the study out of six possible options. This question was included as a
mechanism to determine whether or not it would be appropriate to conduct
statistical analysis examining differences in application amount due to the two
different labels. Out of the 94 participants included in the final analysis, 55% of the
participants were able to correctly recall both of the sunscreen labels. Fifty-five%
correct identification was compared to the probability of randomly selecting the
correct responses (33.6%) using the Binomial Probability Distribution with
p<0.0001 (p=0.000004578957).
A contrast was applied to the model to determine if recognition had a
significant effect on the label’s effect on sunscreen application, and that finding was
also inconclusive. Despite the lack of significance that label message effected the

36

amount of sunscreen used in an application, the percentage that correctly identified
the labels is different from random chance, which is evidence that the participants
did interact with the labels to at least a limited extent. One potential reason for this
is that not all of the participants interacted with the labels in the same way. While
55% of the participants were able to correctly identify the labels, 45% were not able
to correctly identify the labels despite being specifically asked to look at the entire
package prior to applying the lotion. Perhaps a different warning message could
have performed differently because of the tendency for consumers to perceive a risk
when a warning was placed on the label, perhaps it would have been better for the
negative warning to be about the damage UV rays can cause, emphasizing the
benefits of proper sunscreen usage to mitigate that damage, rather than an
ingredient warning
Worry Effect
The second significant factor in the model was the level of worry about
developing skin cancer scaled from 1 (Rarely/Never) to 4 (All of the Time), see
Appendix A to reference survey question 9 which was used to measure worry. While
worry about developing skin cancer was a significant factor within the statistical
model, multiple comparisons using Tukey’s Honestly Significant Difference (HSD)
did not find any significant differences in the amount of sunscreen that was applied
based on the levels of worry at the .05 significance level. The numerical value of the
HSD is dependent on the number of replications of each treatment. Thus, the small
proportion of the sample that experience increased levels of worry limits inferences
about Worry’s effect on sunscreen application.

37

Table 5. Predicted Means Level of Worry

1

1.4082173

1.1572449

1.659190

Frequency of
Participants at
self-reported
level of worry
49

2

1.7767686

1.4784189

2.075118

35

3

1.2127755

0.6693801

1.756171

9

4

0.5275249

-0.9509662

2.006016

1

Level of
Worry

Predicted Mean
(mg/cm2)

Lower Limit to
95% Confidence
Interval

Upper Limit to
95% Confidence
Interval

Table 5 presents the predicted means based on the level of worry, but as the
confidence intervals are all different sizes due to different number of people
responding to each level of worry, they all overlap and none are significantly
different from each other. Since the number of participants who selected the lower
levels of worry is so much greater than those who selected the higher amount of
worry, the power is reduced and the confidence intervals are wider for the higher
level of worry and it is not possible to determine a statistically significant difference.
Specific testing of this effect is recommended for further research.

Amount of Sunscreen Applied in
mg/cm^2

Figure 7. Amount of Sunscreen Applied by Level of Worry

2.5
2
1.5
1
0.5
0
-0.5
-1
-1.5

1.4082173

1.7767686

1.2127755

0.5275249

1

2

3

4

Occasionall
y

Often

Rarely/Neve
r

Level of Worry about Developing Skin Cancer

38

All of the Time

The discrepancy in sample size between levels of worry about developing
skin cancer and the predicted amount of sunscreen that would be applied is
illustrated in figure 7. The worry level 4 has a wide confidence interval that extends
both higher than all of the other levels and below zero into a negative amount of
sunscreen applied.
Frequency of Sunscreen Use Effect
The final significant factor influencing the amount of sunscreen participants
applied was their self-reported frequency of sunscreen use when planning on
spending 2 or more hours outside in the sun. This factor was scaled from 1/Always
to 5/Never, with 2, 3/Sometimes, and 4 as the intermediate options, as collected by
question 5 in the survey in Appendix A. The distribution of the responses was more
even for frequency of sunscreen use than it was for worry, Pairwise comparisons of
the application thickness by varied levels of use were conducted using Tukey’s HSD
(see Figure 8). The predicted mean application amounts and the statistically
significant differences are reported below in table 6.
Table 6. Tukey’s HSD for Frequency of Sunscreen Use
Frequency of Sunscreen
Use

1 (Always)

1.754555*

A

Frequency of
Participants at selfreported frequency of
Sunscreen Use
14

2

1.015278

B

16

AB

36

B

21

AB

9

3 (Sometimes)
4
5 (Never)

Predicted Mean Application
Thickness (mg/cm2)

1.266836
1.017533
1.102405

*Means followed by the same letter are not significantly different from each other at the .05
significance level using Tukey’s HSD.

39

Participants tended to center their answers on the scale; as such, no evidence
of statistical significance resulted when comparisons of application were made by
use levels 2, 3, and 4. By contrast, people who reported always using sunscreen
when spending 2 hours or more outside on a sunny day had signifanctly higher
predicted amounts of sunscreen applied. The people who selected the intermediate
levels of the scale applied significantly less. Figure 8 illustrates the different
amounts of sunscreen applied, with 95% confidence intervals, by participants with
different levels of frequency of sunscreen usage when spending 2 hours of more
outside on a sunny day. None of the participants who responded that they have
experienced severe sunburn selected “never” using sunscreen when going outside
for more than 2 hours on a sunny day, but the number of people in the sample who
experienced severe burns is limited, and thus it was not also included in the model.

Amount of Sunscreen Applied in
mg/cm^2

Figure 8. Predicted amount of Sunscreen Use by Frequency of Sunscreen Use
2.5
2

A

A
B

1.754555

1.5

1.266836
1.017533

1.015278

1.102405

1
0.5
0

Always

1

Sometimes

2

3

Never

4

5

Level of Sunscreen Usage Frequency
One finding from this study is that the frequency of use of sunscreen when
spending 2 hours or more outside on a sunny day is a predictor for how much
sunscreen one will apply. Familiarity with the product can come from repeated use,

40

and a self-report of “Always” using sunscreen when spending two hours or more
outside on a sunny day proposes repeated use. Additionally, it would follow that
consumers who are more likely to burn when out in the sun for 2 hours or more are
more likely to be repeat users of sunscreen products. This familiarity with a
product used to prevent sunburns might also be related to the consumer being more
cognizant of the amount they personally use to prevent burning.

41

Chapter 5 Conclusions, Limitations and Ideas for Future Research
Overall, this study reinforced that users of sunscreen do not apply the
amount that is tested to determine SPF level, and thus do not receive the full
benefits associated with sunscreen use. A key finding of this study is the magnitude
of the effect the packaging can have on sunscreen application, despite it being a
relatively ignored factor in the literature surrounding sunscreen use and skin cancer
prevention as well as the FDA guidelines for sunscreen manufacturers. While the
information provided on a label is important for consumers, the packaging of
sunscreen is also an important contributor to providing consumers the tools to
protect their skin from UV rays and thus should also be considered by regulators
and manufacturers of sunscreen products. The findings of this study indicate that
even though packaging influences the amount of sunscreen people use, not enough
research has been done to compare the different product package systems for their
influence on compliance with sunscreen use guidelines, and the work is not
complete. These findings may be useful in understanding what factors are
associated with increased sunscreen use and how packaging can better encourage
healthy skin protection practices, but they are only the starting point.
Ideas for Future Research
This study attempted to fill in the knowledge gap of how packaging
influences the amount of sunscreen consumers apply, but it did not answer all of the
questions relevant to this line of inquiry. Some relevant directly related research
questions I would recommend pursuing are: influences of packaging on the
application of nontraditional sunscreens such as spray bottles, aerosol sprays,

42

powders, or sticks, how consumers learn to apply sunscreen and if their application
process changes over time, self application verses application to others (e.g.
children), the effects of pump mechanics as a potential means to influence dosing,
and differences in sunscreen application based on marketing message (e.g. specially
formulated for the face as opposed to traditional sunscreen).
Additionally, previous research by de la Fuente has found a difference in grip
strength between children and adults (Javier & Fuente, 2004). It is possible that this
difference in grip strength affects the ability to dispense product from a squeeze
bottle. Thus, further research that compares the differences between how adults
and children apply sunscreen to themselves from different package types based on
the target market would better inform the packaging decision making process.
Limitations
The primary limitations of this study were the distribution of ages within the
sample and the method of estimating surface area. The range of ages sampled at the
two study locations were different from each other, with MSU’s participants ranging
from 20 to 35 years old and Cal Poly’s participants ranging from 19 to 23 years old.
In Sociology, a cohort effect is influenced by both age and the time period (Keyes,
Utz, Robinson, & Li, 2010), thus it is possible that there were factors that influenced
the sunscreen application behaviors of the different sample populations that were
not controlled for, despite only sampling from the millennial generation.
A second limitation was the method used to estimate surface area of
application. A rudimentary method was selected for ease of data collection at
multiple locations, but because the method of measuring with a tape measure is

43

subject to human error, arms and legs are not perfect truncated cones, and
variability across limbs, even on the same body, the surface area estimates are not
perfect. Thus, we are limited in the amount of inferences that can be made.

44

APPENDICES

45

APPENDIX A

Approved Forms

46

Research Participant Information and Consent Form
Title: Effect of Packaging and Labeling Interventions on Sunscreen Application
Principal Investigator:
Dr. Laura Bix, School of Packaging, Michigan State University 517-355-4556
Secondary investigator:
Alyssa Harben, Grad. Student, Michigan State University, 530-210-3864
To participate in this research you must:
• be between 18- 36 years old
• have no known history of skin condition (e.g. eczema, latex allergy, etc.)
• Be a user of sunscreen
Purpose of the research:
You are being asked to participate in a research study, which investigates the link
between packaging and labeling and the application of sunscreen. This experiment
will take no more than 30 minutes.
What you will do:
First, we will ask you to fill out some basic demographic information (age, gender,
etc.) and we will take measurements of your arm and lower leg so that we can
calculate the surface area later.
Then, you will be asked to apply sunscreen to each of your arms and each of your
lower legs. We will ask you to roll up any t-shirt sleeves so they are not in the way of
application. We will ask you to use each of the four packages once, for a total of four
(4) different applications. You will be given hand wipes to clean your hands in
between trials.
After you have applied the four different sunscreen package-label options, you will
be asked to fill out a secondary survey. This survey will include information about
your health history, risk perception, and risk awareness. We will then analyze the
amount of sunscreen applied by weighing the package.
Benefit
Although there is no direct benefit to you for participating in this research, it is our
hope that the data gathered could be used to understand the interface between
people and sunscreen packaging in order to create designs that will facilitate proper
application.
Risk
You will be asked to apply sunscreen in SPF 30. The sunscreen being used is Rocky
Mountain Sunscreen Kids Broad Spectrum SPF 30. It is free of PABA, gluten, peanut
oil and anything else that regularly irritates sensitive skin. That said, as with any
lotion or cream, it may cause irritation in small segments of the population. To

47

minimize possible staining or other damage to clothing, participants are asked to
wear shorts and either a tank top or t-shirt to participate in the study.
Privacy & confidentiality
All information about subjects will be tied to a subject number and you will not be
identifiable by name (even to the research team). Information collected during this
study will be stored in a password-protected computer in spaces controlled by the
Principle Investigator in the School of Packaging . Research records will be accessible
only to authorized researchers and members of MSU HRPP (Human Research
Protection Program) at MSU. Occasionally, publications ask for raw data sets
associated with published work. In the event that these are requested, they would be
furnished to the journal (de-identified). Records will be kept for a minimum of three
years after the closing date of the project.
Your rights to participate
Participation is voluntary. Refusal to participate will involve no penalty or loss of
benefits to which you are otherwise entitled. You may discontinue participation at
any time without penalty or loss of benefits to which you are otherwise entitled. You
may change your mind at any time and withdraw. You may choose not to answer
specific questions or to stop participating at any time.

Costs and Compensation
There is no cost for being in this study. In exchange for your participation, you may
choose $10 cash OR SONA credit if you are enrolled in a participating class in the
College of CAS. Participants who consent to take part in this study, for SONA credit, will
be awarded SONA credits through http://msucas.sona-systems.com. In the SONA
system, 1 hour of research participation is worth 1 SONA credit and this credit is prorated in 15-minute increments. It is up to individual course instructors to determine
how many points this converts to in their classes (this should be specified in the
syllabus for each course).
The duration of this study is approximately 30 minutes. Hence, participants who
complete this study will receive .5 SONA credits.
Participation in this study is voluntary. You may withdraw at any time without
penalty. This means that no SONA credits will be deducted from your account, nor will
withdrawal have any effect on your relationship with any of your instructors.
Participants who withdraw partway through the study will be awarded credit based
on the portion of the study they complete (1/2 completion = .25 SONA credit). Students
who view the materials but do not participate in any part of the research will receive 0
SONA credit.

48

The right to get help if injured
If you are injured as a result of your participation in this research project, Michigan
State University will assist you in obtaining emergency care, if necessary, for your
research related injuries. If you have insurance for medical care, your insurance
carrier will be billed in the ordinary manner. As with any medical insurance, any
costs that are not covered or in excess of what are paid by your insurance, including
deductibles, will be your responsibility. The University’s policy is not to provide
financial compensation for lost wages, disability, pain or discomfort, unless required
by law to do so. This does not mean that you are giving up any legal rights you may
have. You may contact Dr. Laura Bix, MSU, 517-355-4556, ext. 153 or Alyssa Harben
530-210-3864 with any questions or to report an injury.

Contact Information
If you have concerns or questions about this study, such as scientific issues, how to
do any part of it, or to report an injury, please contact the researcher, Dr. Laura Bix,
Laura Bix 517-355-4556; 153 Packaging Building East Lansing MI 48824
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@ora.msu.edu or regular mail at Olds Hall, 408 West
Circle Drive #207, MSU, East Lansing, MI 48824.
Documentation of Informed Consent
Your signature below means that you voluntarily agree to participate in this research
study.
________________________________________ _________
Signature
You will be given a copy of this form to keep.

49

Date

Demographic & Health History Questionnaire: The Effect of
Package Dispenser and Label on Sunscreen Application
Subject #______________________

Date_______________

Location of Study

Cal Poly

Handedness
Ambidextrous

Right

Laterality

Right-Right

Arm Measurements:
___________

MSU
Left

Right- Left

Left-Right

Left-Left

Wrist: ______________Upper Arm: ____________Length:

Leg Measurements: Ankle: ______________Below Knee: ____________Length: ___________

Trial

Treatment
(AA, AB,
BA, BB)

Application Initial Weight
Area
(1,2,3,4)

1
2
3
4

50

Final Weight

Difference

If you are uncomfortable answering any of the following questions, feel free to
leave the response field blank.
1. Please record which two sunscreens you applied from the options of the board:

______________

______________

2. Do you have children?
YES

NO

3. Do you have any personal experience with severe (second or third degree)
sunburns?
YES
NO
4. Please list if you or anyone you know has a history of skin cancer? If not you
please state their relation.

5. How frequently do you apply sunscreen (a product with greater than 15 SPF)?

Less than
once a
month

1 to 3 times
per month

1 time per
week

Between 2-6
times a
week

Everyday

6. When you go outside for more than 1 hour on a warm, sunny day, how often do
you wear sunscreen?
1
Always

2

3
Sometimes

51

4

5
Never

7. How likely do you think it is that you will develop skin cancer in the future?
1
Very low

2

3
Moderate
Chance

4

5
Very High

7. Compared to the average person your age, would you say that you
are:
1
More likely to
get skin
cancer

2
About as likely to
get skin cancer

3
Less likely to
get skin cancer

9. How often do you worry about getting skin cancer?
1
Rarely/Never

2
Occasionally

3
Often

4
All of the time

10. Gender:_________________________

11. Age:____________________________
For Question 12 please use the back of your hand.
12. Skin Tone using the provided Pantone Booklet:____________________________

52

APPENDIX B

Sunscreen Labels Used in Study

53

The labels included in this section were the other options for the question testing label recall. These labels were designed to
look similar but have different messaging that is commonly used on sunscreen products to test whether or not the participants
could recall the messaging from the sunscreen they applied.
Figure 9. Incorrect Option for Survey Response #1

54

Figure 10. Incorrect Option for Survey Response #2

55

Figure 11. Incorrect Option for Survey Response #3

56

Figure 12. Incorrect Option for Survey Response #4

57

APPENDIX C

Test Plan

58

“The four products that you are applying are all different, please
inspect them carefully prior to applying them as you would if were to
spend the entire day outside on a very clear, sunny day.”
Key: P= pump bottle S=Squeeze Bottle IA=Label in increase amount DA=Label to
decrease amount
LA=Left Arm LL= Left Leg RA=right arm RL= Right Leg
Subject # Within Block

Package Order

Application Area Order

1

PIA, PDA, SIA, SDA

LA, RA, RL, LL

2

PDA, SIA, PIA, SDA

RA, LA, RL, LL

3

SDA, SIA, PIA, PDA

LA, RA, LL, RL

4

SIA, PDA, SDA, PIA

RA, LL, LA RL,

5

PIA, SDA, SIA, PDA

LL, RA, RL, LA

6

PDA, PIA, SIA, SDA

RA, LL, RL, LA

7

PIA, SIA, PDA, SDA

RL, RA, LL, LA

8

PDA, SIA, SDA, PIA

RA, RL, LA, LL

9

SIA, PIA, PDA, SDA

LL, LA, RL, RA

10

PDA, PIA, SDA, SIA

RA, LA, LL, RL

11

SDA, PIA, PDA, SIA

LL, LA, RA, RL

12

SIA, SDA, PDA, PIA

LA, LL, RA, RL

13

PDA, SDA, PIA, SIA

LL, RA, LA, RL

14

SDA, SIA, PDA, PIA

RL, LA, LL, RA

15

PDA, SDA, SIA, PDA

LL, RL, RA, LA

16

PIA, PDA, SDA, SIA

RL, LL, LA, RA

17

SDA, PDA, PIA, SIA

LA, RL, LL, RA

18

SIA, PIA, SDA, PDA

RL, LL, RA, LA

19

SDA, PIA, SIA, PDA

LA, LL, RL, RA

20

PIA, SDA, PDA, SIA

RL, LA, RA, LL

21

SIA, SDA, PIA, PDA

RA, RL, LL, LA

22

PIA, SIA, SDA, PDA

LA, RL, RA, LL

23

SDA, PDA, SIA, PIA

LL, RL, LA, RA

24

SIA, PDA, PIA, SDA

RL, RA, LA, LL

59

REFERENCES

60

REFERENCES

Bandura, A., & Adams, N. E. (1977). Analysis of self-efficacy theory of behavioural
change. Cognitive Therapy and Research, 1(4), 24.
Davis, T. C., Wolf, M. S., Ill, P. F. B., Middlebrooks, M., Durazo-arvizu, R., Bocchini, A.,
… Parker, R. M. (2006). Low Literacy Impairs Comprehension of Prescription
Drug Warning Labels. Journal of General Internal Medicine, 847–851.
http://doi.org/10.1111/j.
De La Fuente, J. (2013). USABILITY OF TABS IN SEMI-RIGID PACKAGING By Claudio
Javier de la Fuente Submitted to Michigan State University in partial fulfillment
of the requirements for the degree of Packaging - Doctor of Philosophy.
De La Fuente, J., Gustafson, S., Twomey, C., & Bix, L. (2015). An affordance-based
methodology for package design. Packaging Technology and Science, 28(2),
157–171. http://doi.org/10.1002/pts.2087
De La Fuente, J. (2004). The use of a universal design methodology for developing
child-resistant drug packaging. http://doi.org/10.3102/00346543067001043
deTurck, M. A., Rachlin, R. A., & Young, M. J. (1994). Effects of a Role Model and Fear
in Warning Label on Perceptions of Safety and Safety Behavior. Advances in
Consumer Research, 21(1), 208–212. Retrieved from
http://search.ebscohost.com/login.aspx?direct=true&db=bth&AN=83373816&
site=ehost-live
Diaz, A., Neale, R. E., Kimlin, M. G., Jones, L., & Janda, M. (2012). The Children and
Sunscreen Study. Archives of Dermatology, 148(5), 606–612.
http://doi.org/10.1001/archdermatol.2011.2586
Faurschou, A., & Wulf, H. C. (2007). The relation between sun protection factor and
amount of suncreen applied in vivo. British Journal of Dermatology, 156(4),
716–719. http://doi.org/10.1111/j.1365-2133.2006.07684.x
Feather, N. T. (1963). Mowrer’S Revised Two-Factor Theory and the MotiveExpectancy-Value Model. Psychological Review, 70(6), 500–15.
http://doi.org/10.1037/h0043736
Isedeh, P., Osterwalder, U., & Lim, H. W. (2013). Teaspoon rule revisited: Proper
amount of sunscreen application. Photodermatology Photoimmunology and
Photomedicine, 29(1), 55–56. http://doi.org/10.1111/phpp.12017

61

Keyes, K. M., Utz, R. L., Robinson, W., & Li, G. (2010). What is a cohort effect?
Comparison of three statistical methods for modeling cohort effects in obesity
prevalence in the United States, 1971-2006. Social Science and Medicine, 70(7),
1100–1108. http://doi.org/10.1016/j.socscimed.2009.12.018
Kluckhohn, C. (1962). Toward a General Theory of Action, 984–988.
Krause, M., Klit, A., Blomberg Jensen, M., Søeborg, T., Frederiksen, H., Schlumpf, M., …
Drzewiecki, K. T. (2012). Sunscreens: Are they beneficial for health? An
overview of endocrine disrupting properties of UV-filters. International Journal
of Andrology, 35(3), 424–436. http://doi.org/10.1111/j.13652605.2012.01280.x
Leonard, S. D., & Hill, G. W. (1993). Effect of labels on risk perceptions of consumer
products.
Lynne, J., & Ping, Y. (2003). Consumer perception of risk associated with eating
genetically engineered soybeans is less in the ...
Maddux, J. E., & Rogers, R. W. (1983). Protection motivation and self-efficacy: A
revised theory of fear appeals and attitude change. Journal of Experimental
Social Psychology, 19(5), 469–479. http://doi.org/10.1016/00221031(83)90023-9
Norman, D. A. (2004). Affordances and Design. Retrieved 2-7-2015, 19(3), 2007.
http://doi.org/10.1.1.115.4864
Ozáez, I., Morcillo, G., & Martínez-Guitarte, J. L. (2016). Ultraviolet filters
differentially impact the expression of key endocrine and stress genes in
embryos and larvae of Chironomus riparius. Science of the Total Environment,
557–558, 240–247. http://doi.org/10.1016/j.scitotenv.2016.03.078
Rogers, W. a, Lamson, N., & Rousseau, G. K. (2000). Warning research: an integrative
perspective. Human Factors, 42(1), 102–139.
http://doi.org/10.1518/001872000779656624
Sadeghi, H., Allard, P., Prince, F., & Labelle, H. (2000). Symmetry and limb dominance
in able-bodied gait: A review. Gait and Posture, 12(1), 34–45.
http://doi.org/10.1016/S0966-6362(00)00070-9
Siegrist, M., & Keller, C. (2011). Labeling of Nanotechnology Consumer Products Can
Influence Risk and Benefit Perceptions. Risk Analysis: An International Journal,
31(11), 1762–1769. Retrieved from 10.1111/j.1539-6924.2011.01720.x
Silayoi, P., & Speece, M. (2007). The importance of packaging attributes: a conjoint
analysis approach. European Journal of Marketing, 41(11/12), 1495–1517.
62

http://doi.org/10.1108/03090560710821279
Stern, R. S., Weinstein, M. C., & Baker, S. G. (1986). Risk reduction for nonmelanoma
skin cancer with childhood sunscreen use. Archives of Dermatology, 122(5),
537–45. http://doi.org/10.1001/archderm.1986.01660170067022
Sunscreen Drug Products for Over-The-Counter Human Use. (1978). Food and Drug
Administration.
Tan, R. (2011). Sunscreen Products for Over-the-Counter Human Use; Request for
Data and Information Regarding Dosage Forms, 76(117), 1–4.
Waring, R. H., & Harris, R. M. (2005). Endocrine disrupters: A human risk? Molecular
and Cellular Endocrinology, 244(1–2), 2–9.
http://doi.org/10.1016/j.mce.2005.02.007
Wogalter, M. S., Brelsford, J. W., Desaulniers, D. R., & Laughery, K. R. (1991).
Consumer product warnings: The role of hazard perception. Journal of Safety
Research, 22(2), 71–82. http://doi.org/10.1016/0022-4375(91)90015-N

63