AN EVALUATION OF ATTENTIONAL SCANPATHS ACROSS DRUG LABELS USING
CHANGE DETECTION AND EYE TRACKING
By
Cory Jay Wilson

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Packaging – Master of Science
2013

ABSTRACT
AN EVALUATION OF ATTENTIONAL SCANPATHS ACROSS DRUG LABELS USING
CHANGE DETECTION AND EYE TRACKING
By
Cory Jay Wilson
Medication errors can occur when drug labeling fails to communicate critical
information necessary for safe and effective use. Tracking the eye’s visual scanpath
across drug labels affords investigation into attentive behavior to identify attention
assets and deficits. Eye-tracking and change detection methodologies were combined
in this study to examine attentive behavior in the context of drug labeling during visual
search. Twenty-six participants viewed images on a computer screen of six mock drug
labels that were designed based on 6 commercially available pharmaceutical products.
Labels were sectioned into a 3 x 3 grid (9 square zones, top-left, bottom-right,
center, etc.) during analysis. Post hoc pairwise comparisons indicated the center zone
(zone 5, center of label) garnered significantly more visual hits (P<0.0001) and time in
zone (P<0.0001) than any other zone. The center zone was also the fastest zone
fixated on by participants (P<0.0001). Scanpath data showed agreement and also
indicated frequent fixation on the center zone throughout the visual search task. The
bottom-right zone (zone 9, bottom-right corner of label) was consistently grouped
among the lowest in terms of attention. Information design implications are discussed.

ACKNOWLEDGEMENTS
I would like to thank everyone who contributed to the development of this
research. Thank you, Dr. Laura Bix, for providing your mentorship, patience and
guidance throughout the last few years. I would also like to thank the graduate students
and staff in the MSU Packaging community who provide everyday support.
Special thanks to Dr. and Mrs. Sundar for your willingness and help coding
programs to organize the enormous amount of data generated. Thanks to the
statisticians who took the time to help me manage and understand this data and to
Forrest Pharmaceuticals and Pfizer for providing samples for stimulus development.
Finally, thank you to my committee members, Dr. Susan Selke and Mary K. Smith for
your questions and ours conversations.

iii

TABLE OF CONTENTS
LIST OF TABLES…….…….…………………………………………………………..vi
LIST OF FIGURES…….…….………………………………………………………..viii
CHAPTER 1 – LITERATURE REVIEW……………………………………………..1
Introduction………………………………………………………………1
Visual Attention………………………………………………………….1
Pre-attentive parallel processing..………………………………2
Attentive serial processing.………………………………………2
Selective Visual Attention……..………………..………………………3
The Spotlight Model….….……………………………………….3
The Zoom-lens Model.….………………………………………..4
Inattentional Blindness….……………………………………….5
Fixation.…………………………………………………………...6
Saccadic movements….………………………………………...6
Visual Search……………………………………………………………7
Bottom-up processing.…………………………………………..7
Top-down processing.………………………………………...…8
Center of Gravity.………………………………………………...9
Studies of Visual Attention and Drug Labeling……………………..10
Conspicuousness and Comprehension.….…………………..10
Legibility…….….………………………………………………...11
Health Literacy.….………………………………………………12
Text and Pictorial Formats..……………………………………13
Content Accentuation.……….………………………………….14
Study Approach………………..………………………………………15
CHAPTER 2 – MATERIALS AND METHODS………………………..…………..16
CHAPTER 3 – ANALYTICAL METHODS AND RESULTS.…….……….……...26
Demographic Statistics…………………………….………………….26
Prescription Medication Use………………………………….26
Perceptual Aptitude……………………………………………27
Gaze Tracking Results………………………………………………..28
Time in Zone……………………………………………………29
Visual Hits………………………………………………………32
Time to Zone Fixation…….……………………………………34
Order of Gaze (Discrete Zone Fixation Order)……………………..35
Scanpath Length……………………………………………….36
Discrete Zone Fixation Order…………………………………37
Scanpath: 1st Discrete (Zone) Fixation……..….……………38
Scanpath: 2nd Discrete (Zone) Fixation……….………….…39
Scanpath: 3rd and 4th Discrete (Zone) Fixation.…..…….…40
Scanpath: 5th Discrete (Zone) Fixation…..……….…………41
iv

Scanpath: 6th Discrete (Zone) Fixation…..…………………42
Scanpath: 7th Discrete (Zone) Fixation…..…………………43
Scanpath: 8th Discrete (Zone) Fixation…..…………………44
Scanpath: 9th Discrete (Zone) Fixation…..…………………45
Scanpath: 10th Discrete (Zone) Fixation…....………………46
CHAPTER 4 – DISCUSSION.………………………………………..………………48
APPENDICES...……...………………………………………………………………..54
Appendix A – Example Content of IRB Submitted Documents..…55
Appendix B – Mock Drug Label Stimulus Images……..……..…….62
BIBLIOGRAPHY…....…………………………………………………………………92

v

LIST OF TABLES
Table 1– Shortened version of the Realm-R word list.………………………………..23
Table 2 – Number of participants by gender and by age…………….……………….26
Table 3 – Participant Product Familiarity: Participants by gender that were familiar
(I.E. brand name recognition) with the genuine brands upon which the mock
designs were based.…………………………………………………………………………27
Table 4– Participants who were at risk for health literacy……………..………..……28
Table 5 – Average percentage of time spent fixating in zones (sqrt transformed
data; different letters indicate statistical significance at α=0.05).……...…………...32
Table 6 - Average percentage of hits by zone (log transformed data; different
letters indicate statistical significance at α=0.05).………...………………………...…34
Table 7 – Number of trials that generated a scanpath (by length which
corresponds to the number of unique zone fixations recorded); Note: Four trials
of the total 1,404 did not generate data due to user error or noncompliance........37
Table 8 (to 17) – First Zone Fixation: Zones with observed values greater than the
expected values (of a random distribution model) and that have a chi-square
contribution greater than the chi-square critical value (or chi-square contribution
plus compliment if the contribution alone is not greater than the critical value)
have a significant magnitude of effect that contributes to a non-random
effect.………………………………………………………………………………….……….38
Table 9 - Second Zone Fixation: Significant magnitude of effect for individual
zones.…………….………………...…………………..……………………………………...39
Table 10 - Third Zone Fixation: Significant magnitude of effect for individual
zones.…………….………………...…………………..……………………………………...40
Table 11 – Fourth Zone Fixation: Significant magnitude of effect for individual
zones.…………….………………...…………………..……………………………………...41
Table 12 – Fifth Zone Fixation: Significant magnitude of effect for individual
zones.…………….………………...…………………..……………………………………...42
Table 13 - Sixth Zone Fixation: Significant magnitude of effect for individual
zones.…………….………………...…………………..……………………………………...43

vi

Table 14- Seventh Zone Fixation: Significant magnitude of effect for individual
zones.…………….………………...…………………..……………………………………...44
Table 15 - Eighth Zone Fixation: Significant magnitude of effect for individual
zones.…………….………………...…………………..……………………………………...45
Table 16 - Ninth Zone Fixation: Significant magnitude of effect for individual
zones.…………….………………...…………………..……………………………………...46
Table 17 - Tenth Zone Fixation: Significant magnitude of effect for individual
zones.…………….………………...…………………..……………………………………...47

vii

LIST OF FIGURES
Figure 1 – Spotlight Model Representation……………………………………………….4
Figure 2 –Broad and selective attention: How many squares are pictured above?
A broad distribution of attention allows you to easily see there are 4 squares that
appear to be similar. Selective attention facilitates detail extraction; if you focus
on the bottom-right square you can see it is missing the top of its black border.
For interpretation of the references to color in this and all other figures, the
reader is referred to the electronic version of this thesis.……………………………..5
Figure 3 – Fixation and saccade components demonstrating visual attentive
behavior……………...………………………………………………………………………….7
Figure 4 – Bottom-up processing example: The prominent features of the large
red letter B directs and draws your attention more readily than any of the
individual X’s…….……………………………………………………………………………..8
Figure 5 – Example of Top-down processing: Specific nutrition information draws
contextual attention. A hypertensive person concerned about sugar consumption
would implicitly decide to direct their attention towards carbohydrates
information when making a purchasing decision.…..…………………………………..9
Figure 6 – Center of gravity example: The eye tends to drop lower to center its
vision as we look from left to right at each individual square, even though both
squares sit on the same plane...……………………………….………………………….10
Figure 7 – Flicker technique sequence using change detection..…………………..18
Figure 8 – Demonstration Labels: Two examples of demonstration label images
and the change in the labels that takes place during the flicker technique. Visual
layout example – text specifics are not pertinent..…………………………………….19
Figure 9 – Dow Corning Ophthalmics Near Point Visual Acuity Card. Visual layout
example – text specifics are not pertinent....................................……………………21
Figure 10 – Pseudo isochromatic plates (Richmond Products)..……………………22
Figure 11 – Square format mock drug label image example: sectioned into a 3 x 3
grid, each zone contains approximately 40,000 pixels. Visual layout example –
text specifics are not pertinent.…………………………………….……………………..24
Figure 12 - Zone spatial orientation and correspondence to square shaped
stimulus images. Zone 1 (200 x 200 pixels) corresponds to the top left corner area
of the image. Visual layout example – text specifics are not pertinent...……...…..28
viii

Figure 13 – Failed stimulus detection counts for each product. Certain change
task stimuli in Products 3 and 6 were likely too difficult or subtle for people to
detect.......................................................................................…………………………..30
Figure 14 – Mean trial time to change detection by product (with 95% confidence
intervals). Products 3 and 6 garnered significantly longer mean trial time due to
higher rates of failed stimuli detection (different letters indicate statistical
significance at α=0.05).…….……………………………………..…………………………31
Figure 15 – Total number of visual hits across all subjects (back transformation of
log-transformed data) categorized by product. Products 3 and 6 demonstrated
significantly higher zone hit rates than the other 4 products, due to higher rates
of failed stimuli detection (different letters indicate statistical significance at
α=0.05).………………………………………………………………………………….……...33
Figure 16 – Time to Zone Fixation: Mean elapsed time (log transformed data; with
95% confidence intervals and significance; different letters indicate statistical
significance at α=0.05) until a particular zone is first fixated in.…….……………...35
Figure 17 - Top: Raw scanpath output. Bottom: Numbers in brackets represent
each sequential category (first discrete zone fixation, second… third.., etc.) in a 4sequence scanpath; values to the right of the brackets indicate the actual
discrete fixation zones obtained from the raw scanpath.….…………………………36
Figure 18 – Fixation regions and sequential search behavior with zones
demonstrating the likelihood of attention (zones with high frequency of discrete
zone fixation and individual magnitudes of effect that significantly contributes to
a non-random distribution).…………………………………………….…………………..51
Figure 19 - Theoretical centralized visual search mechanism: The effects of
inattentional blindness and resource allocation on search behavior may cause a
viewer to frequently reorient their visual attention to the center (center of gravity
of the elements in the image) of the image.…………………………..…………………52
Figure 20 – Altered Mock Drug Labels 1 and 1.1: Left – Product 1 label, Right –
Product 1 Zone 1 (top left) altered. Visual layout example – text specifics are not
pertinent...…………………………………………………………….…………………….....62
Figure 21 – Altered Mock Drug Labels 1.2 and 1.3: Left – Product 1 Zone 2 (top
center), Right – Product 1 Zone 3 (top right). Visual layout example – text
specifics are not pertinent...……………………….………………...………………….....63
Figure 22 – Altered Mock Drug Labels 1.4 and 1.5: Left – Product 1 Zone 4 (center
left), Right – Product 1 Zone 5 (center). Visual layout example – text specifics are
not pertinent.……………………………………………...…………………………………..64
Figure 23 – Altered Mock Drug Labels 1.6 and 1.7: Left – Product 1 Zone 6 (center
right), Right – Product 1 Zone 7 (bottom left). Visual layout example – text
specifics are not pertinent.……………………………………………..…………………..65
ix

Figure 24 – Altered Mock Drug Labels 1.8 and 1.9: Left – Product 1 Zone 8 (bottom
center), Right – Product 1 Zone 9 (bottom right). Visual layout example – text
specifics are not pertinent.……………………………………………..…………………..66
Figure 25 – Altered Mock Drug Labels 2 and 2.1: Left – Product 2 label, Right –
Product 2 Zone 1 (top left) altered. Visual layout example – text specifics are not
pertinent.………………………………………………………………………………..……..67
Figure 26 – Altered Mock Drug Labels 2.2 and 2.3: Left – Product 2 Zone 2 (top
center), Right – Product 2 Zone 3 (top right). Visual layout example – text
specifics are not pertinent.…………………………………………………………..……..68
Figure 27 – Altered Mock Drug Labels 2.4 and 2.5: Left – Product 2 Zone 4 (center
left), Right – Product 2 Zone 5 (center). Visual layout example – text specifics are
not pertinent.…………………………………………………………..…………………..….69
Figure 28 – Altered Mock Drug Labels 2.6 and 2.7: Left – Product 2 Zone 6 (center
right), Right – Product 2 Zone 7 (bottom left). Visual layout example – text
specifics are not pertinent.…………………………………………………………….…..70
Figure 29 – Altered Mock Drug Labels 2.8 and 2.9: Left – Product 2 Zone 8 (bottom
center), Right – Product 2 Zone 9 (bottom right). Visual layout example – text
specifics are not pertinent.……………………………………………………………..….71
Figure 30 – Altered Mock Drug Labels 3 and 3.1: Left – Product 3 label, Right –
Product 3 Zone 1 (top left) altered. Visual layout example – text specifics are not
pertinent.…………………………………………………………….………………………...72
Figure 31 – Altered Mock Drug Labels 3.2 and 3.3: Left – Product 3 Zone 2 (top
center), Right – Product 3 Zone 3 (top right). Visual layout example – text
specifics are not pertinent.……………………………………………………………..…..73
Figure 32 – Altered Mock Drug Labels 3.4 and 3.5: Left – Product 3 Zone 4 (center
left), Right – Product 3 Zone 5 (center). Visual layout example – text specifics are
not pertinent.…………………………………………………………….……………………74
Figure 33 – Altered Mock Drug Labels 3.6 and 3.7: Left – Product 3 Zone 6 (center
right), Right – Product 3 Zone 7 (bottom left). Visual layout example – text
specifics are not pertinent.…………………………………………………………….…...75
Figure 34 – Altered Mock Drug Labels 3.8 and 3.9: Left – Product 3 Zone 8 (bottom
center), Right – Product 3 Zone 9 (bottom right). Visual layout example – text
specifics are not pertinent.…………………………………………………………….…...76
Figure 35 – Altered Mock Drug Labels 4 and 4.1: Left – Product 4 label, Right –
Product 4 Zone 1 (top left) altered. Visual layout example – text specifics are not
pertinent.………………………………………………………………………………….……77

x

Figure 36 – Altered Mock Drug Labels 4.2 and 4.3: Left – Product 4 Zone 2 (top
center), Right – Product 4 Zone 3 (top right). Visual layout example – text
specifics are not pertinent.…………………………………………………………………78
Figure 37 – Altered Mock Drug Labels 4.4 and 4.5: Left – Product 4 Zone 4 (center
left), Right – Product 4 Zone 5 (center). Visual layout example – text specifics are
not pertinent.………………………………………………..………………………………...79
Figure 38 – Altered Mock Drug Labels 4.6 and 4.7: Left – Product 4 Zone 6 (center
right), Right – Product 4 Zone 7 (bottom left). Visual layout example – text
specifics are not pertinent.……………………..………..…………………………………80
Figure 39 – Altered Mock Drug Labels 4.8 and 4.9: Left – Product 4 Zone 8 (bottom
center), Right – Product 4 Zone 9 (bottom right). Visual layout example – text
specifics are not pertinent.……………………..………..…………………………………81
Figure 40 – Altered Mock Drug Labels 5 and 5.1: Left – Product 5 label, Right –
Product 5 Zone 1 (top left) altered. Visual layout example – text specifics are not
pertinent.……………………..………..………………………………………………………82
Figure 41 – Altered Mock Drug Labels 5.2 and 5.3: Left – Product 5 Zone 2 (top
center), Right – Product 5 Zone 3 (top right). Visual layout example – text
specifics are not pertinent.……………………..………..…………………………….......83
Figure 42 – Altered Mock Drug Labels 5.4 and 5.5: Left – Product 5 Zone 4 (center
left), Right – Product 5 Zone 5 (center). Visual layout example – text specifics are
not pertinent.……………………..………..………………………….………………….......84
Figure 43 – Altered Mock Drug Labels 5.6 and 5.7: Left – Product 5 Zone 6 (center
right), Right – Product 5 Zone 7 (bottom left). Visual layout example – text
specifics are not pertinent.……………………..………..…………………………….......85
Figure 44 – Altered Mock Drug Labels 5.8 and 5.9: Left – Product 5 Zone 8 (bottom
center), Right – Product 5 Zone 9 (bottom right). Visual layout example – text
specifics are not pertinent.……………………..………..…………………………….......86
Figure 45 – Altered Mock Drug Labels 6 and 6.1: Left – Product 6 label, Right –
Product 6 Zone 1 (top left) altered. Visual layout example – text specifics are not
pertinent.……………………..………..……………………………....................................87
Figure 46 – Altered Mock Drug Labels 6.2 and 6.3: Left – Product 6 Zone 2 (top
center), Right – Product 6 Zone 3 (top right). Visual layout example – text
specifics are not pertinent.……………………..……..…………....................................88

xi

Figure 47 – Altered Mock Drug Labels 6.4 and 6.5: Left – Product 6 Zone 4 (center
left), Right – Product 6 Zone 5 (center). Visual layout example – text specifics are
not pertinent.……………………..……..………………….………....................................89
Figure 48 – Altered Mock Drug Labels 6.6 and 6.7: Left – Product 6 Zone 6 (center
right), Right – Product 6 Zone 7 (bottom left). Visual layout example – text
specifics are not pertinent.……………………..……..………………….……….............90
Figure 49 – Altered Mock Drug Labels 6.8 and 6.9: Left – Product 6 Zone 8 (bottom
center), Right – Product 6 Zone 9 (bottom right). Visual layout example – text
specifics are not pertinent.……………………..……..………………….……….............91

xii

CHAPTER 1 – LITERATURE REVIEW

Introduction
In a report that gained national attention in 1999, the Institute of Medication
(IOM) estimated that at least 7,000 deaths occur each year in hospitals alone due to
medication errors [1]. In the time since, the prevalence of medication errors and issues
of noncompliance have been well documented [2] and federal prevention programs
have been charged with reducing error rates.
Medication error is defined as “any preventable event that may cause or lead to
inappropriate medication use or patient harm while the medication is in control of the
health care professional, patient, or consumer” [2]. Labeling is one way to convey
information critical to the safe and effective use of a drug. The importance of this
strategy cannot be underestimated due to the fact that it stays with the product,
providing timely information at the point of use [3], in an accessible and affordable way
[4]. Drug labels are intended to convey critical information to health care providers,
pharmacists and consumers; thus the noticeability of this information is essential for
safe and effective use [1, 5]. The noticeability of information is dependent on its capacity
to garner an individual’s visual attention [6].

Visual Attention
Visual attention is a process through which awareness is obtained by cognitively
organizing and then interpreting visual elements, or stimulus information, within the
1

visual field. The process of visual attention operates by means of parallel and serial
processing of information. During parallel processing, visual attention is broadly
distributed wherein multiple stimuli can be spatially processed in terms of contextual
organization of information. Visual attention can also narrow selectively during serial
processing where details of specific visual elements are the focus. Parallel and serial
processing of information distinguish two functionally independent hierarchical stages of
visual attention [7].
Pre-attentive parallel processing

The first stage is described as a pre-attentive state where parallel processing
occurs across the entire visual field. Processing during this pre-attentive state functions
simultaneously across various spatial locations, divorced of strategic control and in
unlimited capacity [8]. Spatial orientation of rudimentary information is elucidated such
that prominent characteristics of stimuli in the visual field can garner selective attention.
Attentive serial processing
The second stage of visual attention is a focused, attentive state where serial
processing of a selected spatial location occurs. Processing during this stage operates
with limited capacity and within a limited spatial area [8]. During the first pre-attentive
stage, multiple visual elements have the potential to garner attention, while attentive
serial processing represents the focused extraction of information once an element has
gained selective attention.

2

Selective Visual Attention
Selective visual attention operates by appropriating available visual field
processing resources from a broad field to a more concentrated spatial locus. There are
at least two prevailing models that describe the general behavioral mechanisms of
selective visual attention [9].
The Spotlight Model
The spotlight model describes selective visual attention as comprised of a focus,
fringe, and margin elements [10]. The focus is a geometrically centered and limited area
of high resolution from which visual attention is precisely directed and detailed
information can be extracted. Low resolution gradations that extend from the outer
boundary of the focus area form the fringe area and predispose visual attention to
rudimentary information. The margin frames the limits of the fringe area (See Figure 1).

3

Figure 1 – Spotlight Model Representation
The Zoom-lens Model
The zoom-lens model incorporates all three focus, fringe and margin elements of
the spotlight model with a size-change mechanism [11]. Under this model, the range of
selective visual attention can be enhanced by changing the size of the focus area.
Reducing the focus area allows for efficient processing of smaller visual elements while
increasing the focus area allows larger visual elements to be comprehensively
encompassed. The zoom lens, in essence, is a “sliding scale” between pre-attentive
processing, where many things are attended but not focally processed, and selective
visual attention, where limited items are selectively attended but heavily processed.

4

The processing capacity of the human visual system is limited by the distribution
of finite resources. An increase in the size of the focus area will slow or reduce
processing efficiency because our limited amount of processing resources will be
distributed over a larger area. Thus, selective visual attention is characterized by focus
area sizes that allow for a dense distribution of processing resources and facilitates fine
detail extraction (See Figure 2).

Figure 2 –Broad and selective attention: How many squares are pictured above?
A broad distribution of attention allows you to easily see there are 4 squares that
appear to be similar. Selective attention facilitates detail extraction; if you focus
on the bottom-right square you can see it is missing the top of its black border.
For interpretation of the references to color in this and all other figures, the
reader is referred to the electronic version of this thesis.

Inattentional Blindness
A consequence of visual attention that is directly selective is that in order to
maintain avid focus, peripheral fringe information must be ignored. Inattentional
5

blindness occurs when elements that fall into range of our visual field fail to be noticed
[12]. It is not possible to pay attention to everything we see because minimal processing
resources are available to the fringe area of the visual field.
Fixation
To overcome our visual attention distribution limitations without losing detail
resolution, the eye moves by shifting the locus of the focus area from one location to
another. Fixation is a metric that can be used to measure visual attention by comparing
relatively static eye behavior to dynamic eye movement. Fixation occurs when the locus
of the focus area remains at a relatively stationary spatial location for a given amount of
time. Specifics of fixation duration and location dispersion are arbitrarily defined and
lack general consensus because our eyes are constantly in motion [13-16]. Very rapid
involuntarily micro movements (micro-saccades, ocular drifts, and micro-tremors) of
ocular motor behavior occur as part of our natural physiology [17]. The generally
accepted practice is to define fixation with relation to saccadic movements.
Saccadic movements
Saccades are the linear bridges of eye movement between fixations. While
fixations are relatively stationary, saccades are characterized by their displacement
length and directional velocity over a very short period of time. Information extraction is
not a principal function of saccadic movement although low level rudimentary
processing may occur. Eye movements are generally distinguished as either saccades
or fixation components that help illustrate visual search dynamics and, ultimately,
demonstrate visual attention behavioral patterns (See Figure 3).
6

Figure 3 – Fixation and saccade components demonstrating visual attentive
behavior

Visual Search
Mapping visual search can give researchers insights into behavioral models of
visual attention. Search behavior can be primarily guided by the noticeability of
elements in our visual field or by goal oriented control of the viewer [18-20]. There are
two processes that describe these effects.
Bottom-up processing
Bottom-up processing refers to involuntary, stimulus-driven attentive behavior.
Prominent attributes, such as a sudden motion, can capture attention whether we want
to attend to it or not. Visual attention is spatially directed by stimuli attribute cues within
the visual field.
7

Figure 4 – Bottom-up processing example: The prominent features of the large
red letter B directs and draws your attention more readily than any of the
individual X’s.

Top-down processing
Visual search can also be goal-driven. Top-down processing directs search
behavior contextually under the viewer’s explicit control. Goal-driven search behavior
involves cognitive processing aspects of conflict resolution, working memory and
inhibition to parse information and cultivate attentive decisions while pursuing a related
objective.

8

Figure 5 – Example of Top-down processing: Specific nutrition information draws
contextual attention. A hypertensive person concerned about sugar consumption
would implicitly decide to direct their attention towards carbohydrates
information when making a purchasing decision.

Center of Gravity
Another important element of visual search is the tendency to cognitively balance
the sum total of stimulus elements in our visual field and orient (or gravitate) towards the

9

center of their collective mass [21, 22]. This center of gravity tendency allows the visual
system to frame information and begin to build spatial relationships.

Figure 6 – Center of gravity example: The eye tends to drop lower to center its
vision as we look from left to right at each individual square, even though both
squares sit on the same plane.

Studies of Visual Attention and Drug Labeling
Studies of drug labeling related to visual attentive behavior investigate the
relative prominence and effective communication of label information. Researchers
employ methodological approaches that attempt to characterize the conspicuousness
and comprehension of drug label information.
Conspicuousness and Comprehension
As previously mentioned, fixations are generally employed as measures that can
describe characteristics of visual attentive behavior. The faster a stimulus is visually
fixated on suggests greater conspicuousness or attention capturing properties [23]. The
number of fixations a particular stimulus receives and the duration of each fixation are
also descriptive of attention. However, it is an important to understand that attention

10

does not necessarily signify comprehension, but comprehension issues can affect
attentive behavior.
Studies investigating comprehension distinguish attention as a precursor to
comprehension and develop structured tasks amenable to analysis of specific goals and
objectives. In one study participants were asked to perform several tasks that required
them to visually fixate on a stimulus, extract information from it and make a response
based on the assimilated stimulus information. Results suggested a link between
fixation duration and stimulus processing time [24]. Prolonged fixation duration can
indicate difficulty in extracting information. In another study, it was also found that high
rates of fixation to a particular area can be indicative of difficulty extracting meaningful
information [25]. As such, researchers must consider comprehension issues as a
dynamic that can influence attentive behavior. Several factors that can influence visual
attentive behavior by affecting the conspicuousness and/or comprehension of
information are commonly researched in drug label studies, including health literacy,
legibility, content format, and content accentuation.
Legibility
Drug labels contain a myriad of information such as warnings, instructions for
safe and effective use, active ingredients, expiration date, etc., that all have to compete
for limited label space to be displayed. Several information design approaches exist to
address this issue, including; using varying font sizes, relocating specific information to
secondary labels, and the use of foldout or extendable labels and tags. Legibility

11

concerns arise when drug label content is constricted and lacks adequate print surface
area conducive to discernable presentation and readability of information.
The prevalent focus of most drug label legibility studies is examining the
characteristics of textual print. Research findings indicate font width, height and white
spacing between lettering as key factors that impact drug label legibility. Studies have
found proportionally smaller print sizes have negative legibility impacts on both young
and elderly populations [26-28]. In one study, participants interacted with 12 labels of
varying print sizes and white space amounts on two label formats
(standard/extendable). Both young and elderly populations expressed preference for
larger print types. Data suggested white space between lettering was more impactful on
readability than print size for the younger population (likely because a wider range of
print sizes were legible for that group); whereas for the elderly population print size was
significantly more important in terms of legibility [29]. Although FDA (Food and Drug
Administration) regulation for drug labeling specifies limits on minimum print font sizes,
this is problematic because many font types of the same specified size vary
proportionally.
Health Literacy
Low health literacy has increasingly become recognized as a patient safety
issue. In one study [30] participants were presented with label instructions “Take two
tablets by mouth twice daily”. Participants were tested to see if they could accurately
read, recite and demonstrate the correct amount of medication according to the
instructions. Only 34 percent of those at risk for poor health literacy could demonstrate

12

the correct amount of pills to take daily was four. The results of the study demonstrated
an association between low health literacy and misinterpretation of instructions on
prescription medication labels. Other studies evaluating people at risk for low health
literacy and their ability to interpret instructions or warnings have found similar results
[31-33]. The Institute of Medicine estimates that 90 million adults in the United States
may have trouble comprehending medication labels [34].
Text and Pictorial Formats
Drug labels communicate information in the form of text and nonverbal graphic
symbols or pictorials. Studies of visual attention related to text and pictorial formats
have also shown that illustrations may increase visual attention, especially when the two
formats communicate redundant information [35-37]. In one study, 234 patients who
had visited an emergency room were given printed instructions after receiving treatment
for lacerations. Patients were randomly given printed instructions with half receiving text
instructions only and the other half given text instructions with the addition of illustrated
information. Participants were interviewed by phone three days later and asked if they
read the instructions, followed by a series of questions about the instructions. Patients
who received printed instructions with illustrations were significantly more likely to report
they paid attention to the instructions [38]. They also answered all the interview
questions correctly at a significantly higher rate than those who only received text
instructions.
In a related marketing study [39], participants were instructed to view print
advertisements on a computer screen that contained both text and pictures while their

13

eye movements were tracked. Viewers were told to learn as much as they could about
the products advertised under the guise of a purchase decision. Results showed
participants fixated for longer durations on the images but had a greater number of
fixations in the text part of the ad. More fixations do not necessarily mean more
attention; higher fixation counts are typically observed when textual information is
present because the eye fixates more when reading each word. It is important to
distinguish text and pictorial information because visual attentive behavior varies
between these presentation formats.
Content Accentuation
Products with similar looking and sounding drug names are common due to the
abundance of products on the market that are constituted from the same drug families.
For example, drug names Ephredrine and Epinephrine look similar while Benadryl and
Benazepril sound similar. Drug name confusion is a known cause of incidents of
medication error [40, 41]. Accentuation of drug names is a key part of strategic
reduction of errors associated with drug name confusion.
The use of partial uppercase lettering known as “tall man” (I.E. ePHEDrine and
EPINEPHrine) and the use of color are commonly researched. In one study [42], the
use of highlight color versus tall man lettering against standard lettering controls was
investigated across three experiments. In the first two experiments, participants were
tasked with recognition of differences between word pairs that appeared on a computer
screen. Data from experiment one and two demonstrated for recognition tasks tall man
lettering does make look-alike drug names more distinguishable. The third experiment

14

consisted of memorizing a drug name list and recognition of names on the list when
similar distractors were presented. Results from experiment three indicated tall man
lettering and color do not make memory tasks less confusing but do increase attention.
Tall man lettering has gained wide acceptance as a method to differentiate drug names
that appear similar.

Study Approach
Visual search methodologies frequently characterize attention by presenting
multiple visual stimuli and requesting that a viewer identify a particular stimulus as
quickly as possible. Drug labels contain a large amount of information where
comprehension issues can influence attentive behavior. To characterize attentive
behaviors, we employed a change detection methodology combined with an eye
tracking methodology.
Tracking a viewer’s visual search behavior allows researchers to quantify visual
attentive behavior and identify qualitative elements that influence it. Framing visual
search tasks specifically with drug labels while tracking eye movement will afford
investigation into visual attention as a contextual behavioral mechanism to identify
attention deficits and assets that influence attentive behavior when viewing drug labels.

15

CHAPTER 2 – MATERIALS AND METHODS

The objective of this study is to determine the attentional scan paths of
participants as they view novel designs, based on commercially available prescription
drugs, in order to characterize attentive behaviors. Subjects were recruited with email
fliers and by word of mouth to investigate this objective (See Appendix A for IRB
approved flier). Subjects were excluded if they: were not at least 18 years of age, were
legally blind, or had a known history of seizure.
Researchers provided participants with a verbal explanation detailing all test
procedures. All participants were provided with an IRB approved consent form (IRB 11980 - see Appendix A) which they were asked to review and sign. Consenting study
participants were assigned a subject number and asked to wear any corrective eye
wear they normally require for reading. After informed consent was obtained and
subject number was assigned, demographic information was collected. All data
collected from participants were identified by their subject number and no reference was
made to subject number on the consent form. Collected data (See Appendix A for data
collection form) included: gender, age, profession (work setting and history- if
healthcare professional), education level, self-declared ethnicity, requirements for
eyewear, prescription drug use, and familiarity with label brands the mock drugs were
based on.
Subjects were then calibrated to the pan-tilt optics of our ASL (Applied Science
Laboratories; Boston, MA) 504 eye tracker. This system consists of a small camera that
sits in front of a computer monitor. The camera is equipped with a near infrared beam
16

that is directed into the eye of the subject that, upon calibration, enables the research
team to detect where the subject is looking as they view images that appear on a
computer screen.
Subjects were seated in a chair with no wheels at a desk with a large computer
monitor (Hyundai 24.2 inch Model W242D). The desk was also equipped with an
adjustable chin rest covered in foam to allow subjects to comfortably stay in position.
Subjects were asked to rest their chin on the chin rest. Minor adjustments to the optics
system and the chin rest were made, when necessary, so that they were comfortable for
each subject. Researchers asked that the subject hold as still as possible while they
calibrated the eye tracker using a nine-point calibration with a fixed grid. Having the
subject remain relatively still (through the use of a chin rest) enabled the eye tracker to
precisely track the position of the subject's eye in space and minimized parallax error.
After calibration, subjects viewed a series of “change detection” demo trials on
the computer screen. Change detection tests were conducted with EPrime Software
(Psychology Software Tools, Inc.). During change detection testing, a test image was
shown on a computer screen, followed by a grey screen, then an altered image (the
same as the first but slightly altered in a single locale) ending with another grey screen.
This sequence: test image, grey screen, altered image, grey screen is shown in a
continuous, iterating loop, giving the location of the change a “flickering” appearance.
As such, change detection testing is sometimes referred to as a “flicker task.” (See
Figure 7).

17

Original Image
240 milliseconds

Gray Screen
80 milliseconds

Altered Image
240 milliseconds

Gray Screen
80 milliseconds

Figure 7 – Flicker technique sequence using change detection.

The EPrime software was set such that a test label image appeared on the
computer screen for a period of 240 milliseconds, followed by a gray screen for 80
milliseconds, followed by the altered label image for a period of 240 milliseconds with a
second gray screen at 80 milliseconds [6]. This sequence "loops," giving a "flickering"
appearance in the area of change (See Figure 7) until the subject indicates that they
have detected the change by pressing the space bar, or times out for the trial (at a
period of 1 minute).
Subject testing commenced with two change detection “demo pairs” (See Figure
8). Researchers instructed, "This is a demonstration of the change detection software
that will be used throughout testing. You will see two images flash over one another with
a single change between the two. We are trying to see how long it takes people to
notice changes in these images. Please hit the space bar on the computer as soon as
you notice a change in the label." Timings from each of the two demo pairs were not
analyzed in the final results, but used as an acclimation period as subjects became
acquainted with the test protocol.
18

Test Label (Pair 1)

Test Label (Pair 2)

Altered Label (Pair 1)

Altered Label (Pair 2)

Figure 8 – Demonstration Labels: Two examples of demonstration label images
and the change in the labels that takes place during the flicker technique. Visual
layout example – text specifics are not pertinent.

After hitting the space bar, subjects were asked to click on the portion of the label
that they saw change. The coordinates of the click relative to the label was recorded by
the software enabling researchers to evaluate a correct (within a 50 x 50 pixel area) or
incorrect response with regard to change location. The next trial started after the label
had been clicked, regardless if the response was correct or incorrect. Correct and
incorrect responses with regard to change location were coded (Yes/No) during
analysis. In the event that subjects did not locate the change by pressing the space bar,
the trial timed out after 1 minute, and the subsequent trial began.

19

After the demonstration trials, subjects viewed two sets of 27 flicker pairs (a total
of 54 test trials) with a break between the two sets in order to allow for a period of visual
rest. On-screen instructions “press any key to begin / continue” were presented and
mirrored verbal instructions. After 27 flickers pairs were completed, an informational
screen appeared instructing the participant to take a break to provide visual rest. During
the break period, the subject’s visual acuity, color perception, and health literacy (a
shortened version of the Rapid Estimate of Adult Literacy in Medicine REALM- R) were
measured.
Each subject’s visual acuity was tested using a Dow Corning Ophthalmics Near
Point Visual Acuity Card (See Figure 9). The card was placed approximately 16" from
the subject’s eyes and they were instructed to read (aloud) the lowest line on the card
that they could. In the event that the subject missed any of the letters within the line,
they were asked to read the line above. This continued until they correctly identify all
letters within a given line. Their visual acuity was recorded (20/20, 20/30, etc.) based on
the results of this test.

20

Figure 9 – Dow Corning Ophthalmics Near Point Visual Acuity Card. Visual layout
example – text specifics are not pertinent.

To examine the color perception of the subject, a set of 15 pseudo-isochromatic
plates manufactured by Richmond products was shown to them during the break period.
Subjects were instructed to indicate, aloud, what number, if any, is present on each
card. Responses were recorded and tabulated. Each subject was classified as either at
risk for color blindness or able to perceive color. (See Figure 10)

21

Figure 10 – Pseudo isochromatic plates (Richmond Products)

The Realm-R [43] is a shortened version of the REALM test (Rapid Estimate of
Adult Literacy in Medicine). The original REALM test comprises a list of 66 words; the
REALM-R consists of 11 words (See Table 1). The words were shown printed on a
sheet of white paperboard using 20 pt. sans-serif font. Subjects were instructed to
pronounce (aloud) each word on the list, and say "blank" if there was a word that they
could not pronounce. They were scored based on their ability to correctly pronounce
each word. However, the words fat, flu and pill were not scored because they served as
an acclimation period. Subjects with a score of 6 or less were considered to be at risk
for poor health literacy.

22

Table 1– Shortened version of the Realm-R word list
Realm-R (shortened version) word list
fat

fatigue

flu

directed

pill

colitis

allergic

constipation

jaundice

osteoporosis

anemia

As mentioned, during testing subjects were asked to perform two series of
"flicker tests" that consist of 27 trials each, for a total of 54 flicker trials per subject with a
brief visual break between the two sets. Subjects were recalibrated to the eye tracking
equipment before beginning the second set of trials. Time to detect visual changes in a
series of mock prescription drug labels was recorded as the dependent variable of
interest for the change detection trials. Six mock drug brand labels were developed so
that each a brand had nine images for a total of 54 tests per subject (See appendix B).
Visual changes were distributed equally across all mock brands in order to examine the
effect of location on a change’s noticeability. To identify varied locations, grid lines were
created such that nine zones were created. Each label image was broken down into a 3
x 3 grid, or 9 zones (not visible to the participants during testing, but used during the
analysis, see Figure 11). Each square zone within the grid measured 200 x 200 pixels;
as such, the total image/package was 600 x 600 pixels. Changes were created so that

23

they were approximately equivalent in each zone; i.e. the same number of pixels
disappeared within each of the nine zones of the images.
All 54 pairs of stimulus images (six brands x nine zone locations) were created
using Adobe Photoshop CS5. Square images were scaled to 600 pixels x 600 pixels so
that the images appeared square at a screen resolution of 1024x768. The six mock
brand labels consisted of white backgrounds and a black border. Each mock brand was
designed such that all graphical and text elements distributed across the label were
roughly equivalent (7,632 pixels/zone (+/-5%) in each of the nine zones of the label
when it is apportioned into a 3 x 3 grid.

Figure 11 – Square format mock drug label image example: sectioned into a 3 x 3
grid, each zone contains approximately 40,000 pixels. Visual layout example –
text specifics are not pertinent.
24

Each change was composed of 1,789 pixels (+/-4%) that represented graphic
elements and/or text and “flickered” disappearing within a single zone. Order of
presentation of the 54 trials was randomized across subjects.
Change detection provided two response variables for potential analysis (time to
detect a change- variable data and binary data; ability to detect within the allotted timedetect yes/no). Multiple response variables were captured with the eye tracker for
potential analysis. These included: the time spent in each zone, number of visual hits in
each zone, time until each zone was first fixated on, and the order of gaze, i.e. which
zone did the eye fixate on first, second, third etc. Order of gaze was analyzed via a 10
sequence scanpath because beyond the 10th sequence the data lost resolution due to
the number of subjects tested and incremental complexity of each sequence.
Eye (gaze) tracking was conducted concurrently with change detection testing,
and fixation was defined as four or more coordinate readings (66 milliseconds or
greater) occurring within a 20 x 20 pixel range. These settings were based on the
sensitivity of our equipment and previous research [44]. Visual hits, or when the eye’s
gaze enters a zone, were defined as at least one reading within a given zone.
Response variables characterized the attentional scan paths of subjects as they viewed
prescription drug labels in order to establish “viewing patterns” so that placement of
pertinent information can be identified.

25

CHAPTER 3 – ANALYTICAL METHODS AND RESULTS

Demographic Statistics
A total of 35 participants were tested for this research study. Four participants
could not be calibrated and tracked due to the refractive properties of their eyewear or
the size of their pupils. Five additional participants were eliminated from the data
analysis because post-hoc evaluation of tracking indicated that for at least 10% of their
total trial time their eye was not tracked. As such, demographic information and
generated data for the remaining 26 participants were analyzed.
Of the 26 participants, 13 were male and 13 female. Twenty-two participants
reported their ethnicity as Caucasian, two reported African American and two identified
as Asian. Two participants reported their native language was not English. Table 2
provides frequencies by age and gender.

Table 2 – Number of participants by gender and by age
18-24
Male
9
Female 6
Total
15

25-34
1
3
4

35-49
3
3
6

50+
0
1
1

Total
13
13
26

Prescription Medication Use
Participants were asked about their use of prescription drugs as part of the data
collected form (See appendix A). Fifteen participants reported that they did not
currently take prescription medication (57.7%), nine reported taking 1-2 prescriptions
26

daily (34.6%) and two reported taking 3 -4 prescriptions on a daily basis (7.7%).
Participants were further polled about their familiarity with the products upon which the
stimulus materials were based. Table 3 provides frequency reports of the participants,
by gender, who indicated familiarity with these brands, which have been listed as A-F.

Table 3 – Participant Product Familiarity: Participants by gender that were familiar
(I.E. brand name recognition) with the genuine brands upon which the mock
designs were based.
Participants
Male
Female
Total

Product A
0
0
0

Product B
0
0
0

Product C
5
5
10

Product D
1
0
1

Product E
7
5
12

Product F
10
7
17

Perceptual Aptitude
Participants were tested for visual acuity, color perception and health literacy.
Eleven participants indicated they wore either contact lenses or glasses (42.3%) when
reading; these participants were instructed to wear correction during the testing.
Thirteen participants tested at 20/20 vision (50%), 12 at 20/30 (46.2%) and one at 20/40
(3.8%). None of the 26 participants tested as at risk for color blindness. Three
participants of the 26 (11.5%) tested at risk for poor health literacy based on the
REALM-R results. Table 4 indicates characteristics of participants (labeled A, B and C)
who were at risk for health literacy.

27

Table 4 - Participants who were at risk for health literacy
Participants Gender

Age

A
B
C

18-24
35-49
18-24

Male
Male
Female

Ethnicity

Native
English
Speaker
Caucasian Yes
Asian
No
Asian
No

Visual Acuity

Product
Familiarity

20/30
20/20
20/30

None
None
None

Gaze Tracking Results

Each of the 26 participants viewed a series of 54 images (See appendix B) for a
total of 1,404 change detection trials, as described in the Methods chapter. Zones 1
through 9 refer to specific equilateral sections (600 x 600 pixels) of the stimulus images
(See Figure 12). Statistical analysis was carried out using The SAS System (SAS
Institute Inc., Cary, NC, USA). The model included product (6 levels), gender (2 levels),
zone (9 levels) and their interactions as fixed independent variables, while subject was
considered a random factor in the analysis of variance.

Figure 12 - Zone spatial orientation and correspondence to square shaped
stimulus images. Zone 1 (200 x 200 pixels) corresponds to the top left corner area
of the image. Visual layout example – text specifics are not pertinent.
28

Visual hits to a given zone (an ordinal response), time spent in zone (a
continuous response), and time to each first discrete zone fixation (a continuous
response) were evaluated as dependent variables to characterize attentional behavior
over the entire duration of the change detection trials. Analysis of residuals indicated
that response variables time in zone, time to zone fixation, and zone hits were positively
skewed. In order to fulfill model assumptions, time in zone data were square roottransformed for analysis, while zone hits and time to zone fixation were log-transformed.
Time in Zone
There was no evidence to suggest that gender impacted time in zone (p = 0.
9727). Analysis indicated that the interaction term, product x zone was significant (p
<.0001). The significant effects interaction term can be partially informed by the
variance (See Figure 13) in the frequency of failed stimulus detection for products 3 and
6. Despite the fact that changes were purposefully created with the intent of being equal
across all products (See methods, page 24), changes in products 3 and product 6 (See
appendix B) were likely too difficult or subtle for people to detect, leading to significantly
higher time in overall search (See Figure 14) for these items.

29

Failed Stimulus Detection by Product
Count of Failed Detection

30
25
20
15
10
5
0
1

2

3
4
Product

5

6

Figure 13 – Failed stimulus detection counts for each product. Certain change
task stimuli in Products 3 and 6 were likely too difficult or subtle for people to
detect.

30

Mean Time to Change Detection (ms) by Product
95% CI for the Mean

Time to Detect (ms)

18000
15000
12000
9000
6000
3000
0
1

2

3
4
Product

5

6

Figure 14 – Mean trial time to change detection by product (with 95% confidence
intervals). Products 3 and 6 garnered significantly longer mean trial time due to
higher rates of failed stimuli detection (different letters indicate statistical
significance at α=0.05).

To account for product effects, time spent fixating in each of the nine zones was
characterized as the average percentage of each trial’s total time; in essence
standardizing the response variable. Post hoc zone comparisons using Tukey HSD
suggested three “thresholds” of viewing interest. The average percentage of time spent
in zone 5 was significantly greater than any of the other 8 individual zones (p < 0.001).
This zone, located directly in the middle of the package, accounted for approximately
one-fifth (20%) of the time spent by subjects. Pairwise comparisons of the percentage of
time spent in zones 1-4 and zone 7 did not yield evidence of significant difference
(α=0.05), and as a group, three-fifths of the total time was spent by participants on

31

these five zones. A third threshold group was comprised of zones 6, 9 and 8 (the three
zones that comprise the lower right corner of the stimulus images (See Table 5)). Study
participants spent a significantly lower percentage of their viewing time in these areas
and in aggregate only account for one-fifth of total fixation time).

Table 5 – Average percentage of time spent fixating in zones (sqrt transformed
data; different letters indicate statistical significance at α=0.05).
Zones
5
7
2
3
4
1
8
9
6
Total

Time in Zone (%)
21.14
12.66
11.78
11.55
11.13
11.13
7.98
7.17
5.47
100

ANOVA (Tukey)
A
B
B
B
B
B
C
D
E
p<0.001

Zone Representation
Greatest
Attentional Time:
Intermediate
Attentional Time:
Minimal
Attentional Time:

Visual Hits
Consonant with our previous results, when the response variable was visual hits
to a given zone, there was no evidence of a significant effect of gender (p=0.6664), but
a significant interaction term product x zone (p<0.0001) was indicated. However
dropping product from our model did not increase the error associated with our means
or impact our ability to observe significant effects within the response variable. When all
products were considered in aggregate, post hoc zone comparisons using Tukey HSD
indicated zone 5 (the center zone, see Table 6) garnered a greater percentage of hits

32

than any other zone, while zones 9 and 6 received the least amount of hits (p < .0001),
with zones, 1-4 and 7 and 8 intermediate; following a similar pattern to that evidenced in
the standardized time in zone data.

Total Number of Hits by Product
3000

Sum of Hits

2800
2600
2400
2200
2000
1

2

3
4
Product

5

6

Figure 15 – Total number of visual hits across all subjects (back transformation of
log-transformed data) categorized by product. Products 3 and 6 demonstrated
significantly higher zone hit rates than the other 4 products, due to higher rates
of failed detection (different letters indicate statistical significance at α=0.05).

33

Table 6 - Average percentage of hits by zone (log transformed data; different
letters indicate statistical significance at α=0.05).
Zones Hits per Zone (%)
5
4
2
7
1
8
3
9
6
Total

26.91
12.34
11.75
10.87
9.55
9.50
8.02
5.72
5.34
100

ANOVA
(Tukey)
A
B
B
BC
C
CD
D
E
E
p<0.001

Zone Representation
Highest
Attentional Hits:
Intermediate
Attentional Hits:
Lowest
Attentional Hits:

Time to Zone Fixation
When the response variable was time to zone fixation (or how long it took until a
given zone received its first fixation) only zone 5 (p <0.0001) was shown to be
significant. Pairwise comparisons yielded six groupings of statistically significant unique
zones based on the time to first fixation (See Figure 16). Zone 5 was fixated the fastest,
while zones 1, 2 and 4 (the upper left hand corner) took significantly longer than zone
five, but less than the next group comprised of zone 3, which induced statistically faster
fixation than zone 7. There was no evidence of difference when zones 8 and 6 were
compared but they were fixated significantly faster than zone 9 which was the last zone
visually fixated. This suggests the possibility of a specific directionality or ordered
search behavior within the context of the stimulus images.

34

Time to Zone Fixation
95% CI for the Mean

4000
3500
Time (ms)

3000
2500
2000
1500
1000
500
0
5

4

2

1

3
Zone

7

8

6

9

Figure 16 – Time to Zone Fixation: Mean elapsed time (log transformed
data; with 95% confidence intervals and significance; different letters
indicate statistical significance at α=0.05) until a particular zone is first
fixated in.

Order of Gaze (Discrete Zone Fixation Order)

Directed by the previous results, order of discrete zone fixations (scanpath) was
analyzed in a 10-sequence transition matrix. The first element in the scanpath indicates
the specific zone where the first fixation occurs when a participant is viewing an image.
The nth element indicates the nth discrete zone that is fixated (See Figure 17). Thus the
scanpath represents sequential search order by zone.

35

Consider, for instance, the example scan path provided below:
Raw scanpath sequence:

6, 6, 6, 6, 4, 4, 4, 5, 5, 7, 7, 7, 7, 7, 7

Registered scanpath sequence: [1] 6, [2] 4, [3] 5, [4] 7

Figure 17 - Top: Raw scanpath output. Bottom: Numbers in brackets represent
each sequential category (first discrete zone fixation, second… third.., etc.) in a 4sequence scanpath; values to the right of the brackets indicate the actual
discrete fixation zones obtained from the raw scanpath.

Scanpath Length
Length of scanpath sequences varied by trial due to variation within and between
participants in visual scanning behavior, combined with nature of the change detection
task, where the trial ends when the participant identifies the stimulus. Table 7 identifies
the number of trials that generated a scanpath (by scanpath length: unique zone
fixations). Four trials of the total 1,404 did not generate data due to user error or
noncompliance (I.E. inadvertently pressing the spacebar in quick succession at the end
of a trial thereby skipping the subsequent trial). Trials where the subject identified an
incorrect detection area for the change stimulus were included as long as the trial
generated a scanpath preceding detection.

36

Table 7 – Number of trials that generated a scanpath (by length which
corresponds to the number of unique zone fixations recorded); Note: Four trials
of the total 1,404 did not generate data due to user error or noncompliance.
Scanpath Length
(unique zone fixations)
-1
2
3
4
5
6
7
8
9
10

Number of Trials
1404
1400
1382
1343
1258
1173
1107
1026
950
875
795

Discrete Zone Fixation Order

The dependent variable is the order in which discrete zone fixations occur. ChiSquare Goodness of Fit Test was used to determine if the distribution of discrete zone
fixations fit the assumptions of a random distribution model. Chi-Square Contribution
and Chi-Square Contribution Compliment were calculated to determine which specific
zones were significant relative to the contribution of a particular outcome. Each of the
nine zones can be described in terms of one of three outcomes. Discrete zone fixations
for each zone can (1) occur at a frequency that is statistically significantly greater than a
random distribution, (2) occur at a frequency that is not statistically significantly different
from a random distribution or, (3) occur at a frequency that is statistically significantly
less than a random distribution. Each of the nine zones of interest were analyzed and

37

described at each sequential category (first discrete zone fixation, second… third.., etc.)
across a 10-sequence scanpath.
Scanpath: 1st Discrete (Zone) Fixation
There is evidence that the distribution of the first zone fixation across the nine
2

zones of interest is not random c (8, N = 1400) = 1549.09, p < .001. The first region
participants fixated in was more likely in zones 5, 4, and 2, respectively, than the other
six zones (See Table 8).

Table 8 (to 17) – First Zone Fixation: Zones with observed values greater than the
expected values (of a random distribution model) and that have a chi-square
contribution greater than the chi-square critical value (or chi-square contribution
plus compliment if the contribution alone is not greater than the critical value)
have a significant magnitude of effect that contributes to a non-random effect.

Most Likely:

Zone
(Proportion =
0.1111)
5
4
2
1
6
8
3
7
9

1st Discrete (Zone) Fixation
Random:

Observed
(Expected = 156)
558
267
212
143
66
46
45
42
21

Chi-Square
Contribution
1041.18
79.84
20.48
1.01
51.56
77.16
78.57
82.90
116.39
38

Least Likely:

Chi-Square
Critical (df=
1) (P=0.05)

3.841

Chi-Square
complement
+contribution
---1.35
------

Scanpath: 2nd Discrete (Zone) Fixation
Among the nine zones of interest there is evidence that the distribution of second
2

discrete (zone) fixation is not random c (8, N = 1382) = 539.767, p < .001. Fixation was
more likely to occur in zones 5, 4, 1, and 2, respectively, than the other five zones (See
Table 9).

Table 9 - Second Zone Fixation: Significant magnitude of effect for individual
zones.

Most Likely:

Zone
(Proportion =
0.1111)
5
4
1
2
3
7
6
8
9

2nd Discrete (Zone) Fixation
Random:

Observed
(Expected = 154)
292
264
239
231
117
79
73
61
26

Chi-Square
Contribution
124.820
79.437
47.545
39.058
8.702
36.199
42.260
55.788
105.958

39

Least Likely:

Chi-Square
Critical (df=
1) (P=0.05)

3.841

Chi-Square
complement
+contribution
----------

Scanpath: 3rd and 4th Discrete (Zone) Fixations
2

Distributions of the third discrete (zone) fixation c (8, N = 1343) = 353.229, and
2

fourth discrete (zone) fixation c (8, N = 1258) = 210.202, p < .001 as indicated, were
shown to be non-random. Fixation was more likely to occur in zones 5, 2, 1, and 3
across the third and fourth fixation distributions (See Table 10 and Table 11
respectively) than in the other 5 zones.

Table 10 - Third Zone Fixation: Significant magnitude of effect for individual
zones.

Most Likely:

Zone
(Proportion =
0.1111)
2
5
1
3
4
7
8
6
9

3rd Discrete (Zone) Fixation
Random:

Observed
(Expected = 149)
254
246
206
200
151
105
75
70
36

Chi-Square
Contribution
73.571
62.765
21.604
17.279
0.021
13.105
36.918
42.059
85.907

40

Least Likely:

Chi-Square
Critical (df=
1) (P=0.05)

3.841

Chi-Square
complement
+contribution
----0.034
-----

Table 11 – Fourth Zone Fixation: Significant magnitude of effect for individual
zones.

Most Likely:

Zone
(Proportion =
0.1111)
5
2
1
3
4
7
8
6
9

4th Discrete (Zone) Fixation
Random:

Observed
(Expected = 140)
232
202
187
164
144
93
92
88
56

Chi-Square
Contribution
60.846
27.698
15.953
4.198
0.128
15.655
16.331
19.180
50.213

Least Likely:

Chi-Square
Critical (df=
1) (P=0.05)

3.841

Chi-Square
complement
+contribution
----0.203
-----

Scanpath: 5th Discrete (Zone) Fixation
The distribution of the fifth discrete (zone) fixation across the nine zones of
2

interest also demonstrated non-random effects c (8, N = 1173) = 159.223, p < .001.
Fixation was more likely to occur in zones 5, 3, and 2 than the other six zones (See
Table 12).

41

Table 12 – Fifth Zone Fixation: Significant magnitude of effect for individual
zones.
5th Discrete (Zone) Fixation
Random:

Most Likely:

Zone
(Proportion =
0.1111)
5
3
2
1
7
4
8
6
9

Observed
(Expected = 130)
224
177
163
129
128
116
95
78
63

Chi-Square
Contribution
67.315
16.709
8.188
0.0136
0.0418
1.576
9.579
21.014
34.786

Least Likely:

Chi-Square
Critical (df=
1) (P=0.05)

3.841

Chi-Square
complement
+contribution
-----2.509
----

Scanpath: 6th Discrete (Zone) Fixation
The distribution of the sixth discrete (zone) fixation demonstrated evidence of
2

non-random effects c (8, N = 1107) = 99.252, p < .001. Fixation was more likely in
zones 5 and 4 than the other seven zones (See Table 13).

42

Table 13 - Sixth Zone Fixation: Significant magnitude of effect for individual
zones.
6th Discrete (Zone) Fixation
Random:

Most Likely:

Zone
(Proportion =
0.1111)
5
4
2
8
3
1
7
6
9

Observed
(Expected = 123)
197
145
137
131
126
122
110
76
63

Chi-Square
Contribution
44.520
3.935
1.594
0.520
0.073
0.00
1.374
17.959
29.268

Least Likely:

Chi-Square
Critical (df=
1) (P=0.05)

3.841

Chi-Square
complement
+contribution
--2.522
0.823
0.119
-----

Scanpath: 7th Discrete (Zone) Fixation
Non-random effects were also evident across the distribution of the seventh
2

discrete (zone) fixation c (8, N = 1026) = 76.456, p < .001. Fixation was more likely in
zones 5, 7, and 4 than the other six zones (See Table 14).

43

Table 14 - Seventh Zone Fixation: Significant magnitude of effect for individual
zones.

Most Likely:

Zone
(Proportion =
0.1111)
5
7
4
2
8
1
3
9
6

7th Discrete (Zone) Fixation
Random:

Observed
(Expected = 114)
170
141
132
118
116
107
106
77
59

Chi-Square
Contribution
27.509
6.394
2.842
0.140
0.035
0.430
0.561
12.009
26.535

Least Likely:

Chi-Square
Critical (df=
1) (P=0.05)

3.841

Chi-Square
complement
+contribution
--4.565
0.232
0.058
-----

Scanpath: 8th Discrete (Zone) Fixation
Among the nine zones of interest the distribution of the eighth discrete (zone)
2

fixation indicated non-random effects c (8, N = 950) = 78.236, p < .001. Fixation was
more likely in zones 5, 7, and 8 than the other six zones (See Table 15).

44

Table 15 - Eighth Zone Fixation: Significant magnitude of effect for individual
zones.

Most Likely:

Zone
(Proportion =
0.1111)
5
7
8
3
2
4
1
9
6

8th Discrete (Zone) Fixation
Random:

Observed
(Expected = 106)

Chi-Square
Contribution

159
141
133
102
96
92
86
82
59

27.060
11.901
7.136
0.120
0.865
1.741
3.623
5.26
20.53

Least Likely:

Chi-Square
Critical (df=
1) (P=0.05)

3.841

Chi-Square
complement
+contribution
-----3.093
6.500
---

Scanpath: 9th Discrete (Zone) Fixation
Distribution of ninth discrete (zone) fixation by zone within the scanpath
2

demonstrated non-random effects c (8, N = 875) = 93.266, p < .001. First fixation was
more likely to occur in zones 5, 7, 4, and 8 than in the other 5 zones (See Table 16).

45

Table 16 - Ninth Zone Fixation: Significant magnitude of effect for individual
zones.

Most Likely:

Zone
(Proportion =
0.1111)
5
7
4
8
2
9
3
1
6

9th Discrete (Zone) Fixation
Random:

Observed
(Expected = 97)
148
139
120
114
81
76
73
71
53

Chi-Square
Contribution
26.520
17.952
5.337
2.895
2.707
4.633
6.035
7.072
20.115

Least Likely:

Chi-Square
Critical (df=
1) (P=0.05)

3.841

Chi-Square
complement
+contribution
---4.604
4.646
-----

Scanpath: 10th Discrete (Zone) Fixation
Similar to the eighth fixation distribution, the distribution of the tenth discrete
(zone) fixation distribution indicated non-random effects among the nine zones of
2

interest c (8, N = 795) = 61.o, p < .001. Fixation was more likely in zones 5, 7, and 8
than the other six zones (See Table 17).

46

Table 17 - Tenth Zone Fixation: Significant magnitude of effect for individual
zones.

Most Likely:

Zone
(Proportion =
0.1111)
7
5
8
4
9
3
6
2
1

10th Discrete (Zone) Fixation
Random:

Observed
(Expected = 88)
130
115
114
92
81
75
69
61
58

Chi-Square
Contribution
19.654
8.050
7.458
0.152
0.609
2.012
4.231
8.458
10.416

47

Least Likely:

Chi-Square
Critical (df=
1) (P=0.05)

3.841

Chi-Square
complement
+contribution
---0.253
-3.456
----

CHAPTER 4 – DISCUSSION

Discussion
Data for our three response variables characterizes each of the nine zones in
terms of attentional behavior. Zone 5 (the center zone) was statistically significantly
more effective at garnering attention (α= 0.05) than any other zone as evidence by: (1)
the highest average percentage of zone fixation time, (2) the greatest number of visual
hits, and (3) the fastest average zone fixation time.
A viewer’s inclination to gravitate towards the center of the elements in their
visual field (known as center of gravity) is well recognized [21, 45-47]. When viewing an
image a viewer’s gaze initially gravitates towards the center of the elements in the
image in order to orient their attention to the whole of the image. Two mechanisms,
parallel processing (wide distribution of attention) and serial processing (narrow or
focused distribution of attention), facilitate the extraction of visual information. Parallel
processing allows the viewer to grasp the gist of the image as a whole but not the
specific elements that comprise it. This is because the human visual system has limited
available cognitive processing resources and it is not possible to focus on every detail at
once.
Serial processing allows detailed information to be extracted, when attention is
selectively focused. Research indicates selective visual attention is thought to function
like a spotlight with a zoom-lens mechanism that can vary its focus over time [10, 11]. If
visual search is carried out primarily by a spotlight with a zoom-lens mechanism we
would expect that the time it takes to shift the spotlight of attention between elements in
48

the visual field would be influenced by the distance between them. However, Kwak et
al.[48] found that shifts in visual attention are independent of time; attention can be
immediately directed to new elements in the visual field, regardless of distance.
Other researchers have refuted the notion of spotlight or zoom-lens model
functionality. Egly and Homa [49] demonstrated that when visual attention is directed to
a particular location, stimuli outside the range of the spotlight area were no less
detectable than stimuli inside the spotlight. However, Theeuwes [8] showed that when
attention is focused visually, abrupt stimuli outside the spotlight do not readily draw
attention whereas inside the spotlight they do. This discrepancy infers that visual
attention is comprised of multiple mechanisms that operate on different functional levels
of attentional distribution.
Analysis of data from our ten-sequence-scanpath suggests a centralized search
mechanism for visual search. In our experiment, attentional preference was
characterized in terms of zones that represent specific subdivisions of the product
image. Zones that showed attentional preference at a given scanpath succession (first
zone participants viewed, second zone, etc.) were identified as regions (See Figure 18)
that were characterized by evidence of statistical difference using varied dependent
variables. All regions were composed of two or more zones. Scanpath data showed
agreement with response variable data promoting the attentional prominence of zone
five (the center zone). Zone 5 demonstrated attentional preference in every one of the
first ten regions sequentially fixated on by participants (See Figure 18). This data
suggests the center of the label frequently and quickly garnered visual attention.

49

If we consider the effects of inattentional blindness and resource allocation on
search behavior it may be necessary for a viewer to frequently reorient broad visual
attention to the center of gravity of the image (our center zone, zone 5). Humphreys &
Bruce [50] offer evidence that serial processing is a self-terminating process. During
serial processing in order to visually focus on any specific element and extract detailed
information, other elements in our visual field must be ignored [12]. Inattentional
blindness towards other elements may lead to a brief discontinuous state of inattention
(in the absence of a cue to attend another element within spotlight proximity) between
the currently attended element and the next element that is selected to be attended.
This condition may characterize the transition between serial and parallel processing
during visual search. Jonides [9] indicates effective visual search must include a
mechanism that frequently and swiftly reallocates resources and Remington[51]
suggests this is accomplished by shifting gaze or fast processing concentration on a
new location in the visual field. Our scanpath data suggests the center zone is an
integral allocation hub during visual search. Frequently reorienting to the center of
gravity of the image with a broad distribution of attention reduces cognitive processing
load and can explain how shifts of visual attention are independent of time regardless of
their distance in the visual field. The viewer reacclimatizes to the whole of the image
and any visual element has the potential to garner selective attention. This continuous
loop of parallel and serial processing during visual search would also explain how
elements outside the spotlight effect are not readily attended until serial processing
terminates (Figure 19).

50

Scanpath Regions of Sequential Search Behavior
1st

2nd

3rd

4th

5th

6th

7th

8th

9th

10th

Zone Representation

Most Likely Zone of Fixation (Attention)

Figure 18 – Fixation regions and sequential search behavior with zones
demonstrating the likelihood of attention (zones with high frequency of discrete
zone fixation and individual magnitudes of effect that significantly contributes to
a non-random distribution).

51

Figure 19 - Theoretical centralized visual search mechanism: The effects of
inattentional blindness and resource allocation on search behavior may cause a
viewer to frequently reorient their visual attention to the center (center of gravity
of the elements in the image) of the image.

Zones demonstrating the least amount of attention were also identified. Zone 9
was consistently grouped in the lowest attention capturing zones. Zones 8, 9, and 6
combined only represented twenty percent of the total time fixating in all zones by
52

participants (for comparison zone 5 individually accounted for twenty percent, See
Table 54). Zones 9 and 6 also garnered the lowest amount of visual hits and zone 9
took the longest amount of time to fixate on. Scanpath tracking results for the first 10
regions participants fixated on (where did they look first, second, third, etc.) showed
zones 1 through 5 exclusively demonstrated attentional preference within the first six
fixation regions. Zone 7 does not demonstrates attention preference until the seventh
region and onward. Zone 8 did not show attentional preference until the eighth region
while zones 9 and 6 did not show attentional preference in any of the first ten regions
participants fixated on. This pattern suggests zones 9 and 6 are likely the last zones to
be viewed by participants (See Figure 18).
Zone 9 is of particular note because many packages place warnings, critical
information, or overt labeling in the bottom right corner (zone 9) of the label such as: Do
not take with nitrates, refrigerate after opening, shake well, do not consume with
alcohol, avoid prolonged exposure to the sun, may cause drowsiness, chew before
swallowing, etc. This research indicates there may be a positional effect related to the
noticeability or attentional priority of information within the context of a drug label
(square display format). The attentional prominence of zone 5 (the center zone)
suggests the placement of critical information may be innately more effective in the
center zone and least effective in zone 9 (the bottom right corner).

53

APPENDICES

54

APPENDIX A – Example Content of IRB Documents
Recruitment Flyer
Re: Opportunity to participate in a research study
This email is to inform you of the opportunity to participate in a research study at the
School of Packaging. You are under no obligation to participate.
To participate you must:
• Be at least 18 years of age
• Have no history of seizure
• Not be legally blind
• Have transportation to the School of Packaging, where the study will take place
You are being asked to participate in a research study regarding the labeling of
prescription drug products that is being conducted by Cory Wilson for his Master’s
thesis. He is using an eye tracking system which consists of a small camera in front of a
computer screen and change detection software, or “Flicker” program to conduct this
research. The eye tracking system will be calibrated with your eye. You will be asked to
look at a computer screen and hit the space bar when you detect a change in the labels.
(During the flicker task a photo will alternate with a second photo that has one small
change; these two photos will be separated by a brief, blank display and will continue
“flickering” until you detect the change in the two). The time to detect the change will be
recorded as a way to quantify how prominent the change is to the scene and the eye
tracker will record your gaze and attentive behavior.
Your color blindness, visual acuity, and health literacy will also be tested. These tests
involve viewing a series of cards. You will also be asked to fill out a brief survey which
includes information about your ethnicity, gender, age, educational background, usage
of and familiarity with prescription medications.
The test should take no longer than 1 hour. In exchange for your participation, you will
receive your choice of a $10 Starbucks gift card OR 2 points extra credit in PKG 480,
whichever you choose.
If at any time you are uncomfortable with the testing or wish to discontinue the data
collection process, you may discontinue participation without penalty. You will still
receive the $10 Starbucks gift card or 2 points extra credit in PKG 480.
If you are interested in pursuing this opportunity, please contact Cory Wilson at
wilso279@msu.edu to make an appointment.

55

If you have questions or comments regarding this study, please contact Dr. Laura Bix,
Assistant Professor or Packaging at Michigan State University 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 202 Olds Hall, MSU, East Lansing, MI 48824.

56

Consent Form

Michigan State University
School of Packaging
INSTRUCTIONS AND RESEARCH CONSENT FORM – Applying eye tracking and
change detection to test the noticeability of components of prescription drug labels

You are being asked to participate in a research study. Your participation in this
study is voluntary. To participate in the study you must:

•
•
•

Be at least 18 years of age
Have NO HISTORY OF SEIZURE
Not be legally blind

In exchange for your participation in this study, you will receive your choice of a $10
Starbucks gift card OR 2 points extra credit in PKG 480.
As part of this research, we will record your gender, ethnicity, educational
background, age, eye wear requirements, and information about your professional
history. We will also test your visual acuity (20/20, 20/30, 20/40, etc.), health literacy
(word identification), and color perception. These tests will be conducted by asking
you to view a series of cards and asking you to decipher images or read aloud
common words related with healthcare to the best of your ability. You will also be
asked to fill out a brief survey regarding your usage of and familiarity with
prescription medications. We are interested in the things that people look at prior to
medication usage, and whether or not labeling design can be manipulated in ways
that draw your attention to important directions or information.
All 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 a
locked laboratory in the School of Packaging at Michigan State University for a
MINIMUM of 3 years. The room will be accessible only to authorized researchers of
Dr.Laura Bix’s research team. This study will take no more than one hour, and poses
little risk to your health and well-being.

57

For eye tracking and change detection testing, Cory Wilson from the School of
Packaging, will ask you to view a series of labels on a computer screen, and will time
how long it takes you to find changes in the images. You will be seated at a desk.
The desk is equipped with an adjustable chin rest made from foam for support and
comfort while testing. The pan-tilt optics ASL 504 eye tracker, which consists of a
small camera that sits in front of a computer screen, will be calibrated with your eye.
You will have 1 minute to view each label in succession, the eye tracking device will
record what your eye sees. This is a test of attentive behavior.
Two images will alternate on the computer screen in rapid fashion, with a blank
screen between the two. One of the images is slightly different than the other, and
you will be asked to hit the space bar as soon as you detect the difference in the two
images. If you correctly identified the change, the program will advance to another
pair of images; if not correctly identified, the previous images will play again until you
are able to detect the change. This is not a test of your speed, but of a test of the
ability of the change to draw your attention.
Important: You are free to discontinue your participation in the study at any time
without penalty. You may discontinue participation at any time and still be eligible for
a $10 Starbucks gift card OR 2 points extra credit in PKG 480.
There is no direct benefit to you in exchange for participating in this study. The hope
is that through studies like this, we will gain an understanding of the design features
that garner attention, so that this information can be used to make important
information, such as directions and warnings, prominent.
There is a possible risk of seizure that is associated with viewing flashing images. If
you are injured as a result of your participation in this research project, researchers from
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 Laura Bix at 517-355-4556 with any questions or to report an injury.
You are aware that the results of the study will be treated in strict confidence and
that you will remain anonymous. Raw results from your trials will be available to the
Institutional Review Board (IRB) at MSU and the research team that is conducting
this research. Your confidentiality will be protected to the maximum extent allowable

58

by law. Within these restrictions, results of the study will be made available at your
request.
If you have any concerns or questions about this research study, such as scientific
issues, how to do any part of it, or if you believe you have been harmed because of
the research, please contact the researcher 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 517432-4503, or e-mail irb@msu.edu or regular mail at 202 Olds Hall, MSU, East
Lansing, MI 48824.
I voluntarily agree to participate in the study of prescription drug labels.
Name: ____________________

Date:

____________________

You will be provided with a copy of your signed consent form.

I choose (circle 1)

$10 Starbucks gift card

2 points extra credit in PKG 480

If you choose the $10 Starbucks gift card, please indicate receipt of the money with
signature and date below.
Name: _____________________

Date: ______________________

59

Data Collection Form
Demographic Information
1. Gender
❑ Male ❑ Female
2. Profession
❑ Health Field Professional ______________________________________
❑ Student
❑ Other _______________________________________________________
3. Years of Experience in healthcare related field _________________________
4. Age
❑18～24 ❑25～34 ❑35～49 ❑ 50+
5. Eye wear
❑None ❑Glasses ❑Contact lenses ❑Other
6. Highest level of Education Achieved
❑High school ❑Associates ❑Bachelors ❑Graduate ❑Other
7. Ethnicity
❑White ❑Black ❑Hispanic ❑Asian ❑Other __________________
8. Native Language ______________________________________________

Prescription Drug Information
Please answer the following questions:
9. Please list any prescription medications that you are currently taking:
________________________________
________________________________
________________________________
________________________________
________________________________
60

________________________________

10. Circle the names of the drugs that are familiar to you and indicate if you have taken
them before, or (if you are a healthcare professional) prepared the drug for a patient.
❑Bystolic ________________________________________________
❑Campral _________________________________________________
❑Lexapro _________________________________________________
❑Namenda _________________________________________________
❑Nasonex _________________________________________________
❑Viagra
_____________________

61

APPENDIX B – Mock Drug Label Stimulus Images

Figure 20 – Altered Mock Drug Labels 1 and 1.1: Left – Product 1 label, Right – Product 1 Zone 1 (top left) altered.
Visual layout example – text specifics are not pertinent.

62

Figure 21 – Altered Mock Drug Labels 1.2 and 1.3: Left – Product 1 Zone 2 (top center), Right – Product 1 Zone 3
(top right). Visual layout example – text specifics are not pertinent.

63

Figure 22 – Altered Mock Drug Labels 1.4 and 1.5: Left – Product 1 Zone 4 (center left), Right – Product 1 Zone 5
(center). Visual layout example – text specifics are not pertinent.

64

Figure 23 – Altered Mock Drug Labels 1.6 and 1.7: Left – Product 1 Zone 6 (center right), Right – Product 1 Zone 7
(bottom left). Visual layout example – text specifics are not pertinent.

65

Figure 24 – Altered Mock Drug Labels 1.8 and 1.9: Left – Product 1 Zone 8 (bottom center), Right – Product 1
Zone 9 (bottom right). Visual layout example – text specifics are not pertinent.

66

Figure 25 – Altered Mock Drug Labels 2 and 2.1: Left – Product 2 label, Right – Product 2 Zone 1 (top left) altered.
Visual layout example – text specifics are not pertinent.

67

Figure 26 – Altered Mock Drug Labels 2.2 and 2.3: Left – Product 2 Zone 2 (top center), Right – Product 2 Zone 3
(top right). Visual layout example – text specifics are not pertinent.

68

Figure 27 – Altered Mock Drug Labels 2.4 and 2.5: Left – Product 2 Zone 4 (center left), Right – Product 2 Zone 5
(center). Visual layout example – text specifics are not pertinent.

69

Figure 28 – Altered Mock Drug Labels 2.6 and 2.7: Left – Product 2 Zone 6 (center right), Right – Product 2 Zone 7
(bottom left). Visual layout example – text specifics are not pertinent.

70

Figure 29 – Altered Mock Drug Labels 2.8 and 2.9: Left – Product 2 Zone 8 (bottom center), Right – Product 2
Zone 9 (bottom right). Visual layout example – text specifics are not pertinent.

71

Figure 30 – Altered Mock Drug Labels 3 and 3.1: Left – Product 3 label, Right – Product 3 Zone 1 (top left) altered.
Visual layout example – text specifics are not pertinent.

72

Figure 31 – Altered Mock Drug Labels 3.2 and 3.3: Left – Product 3 Zone 2 (top center), Right – Product 3 Zone 3
(top right). Visual layout example – text specifics are not pertinent.

73

Figure 32 – Altered Mock Drug Labels 3.4 and 3.5: Left – Product 3 Zone 4 (center left), Right – Product 3 Zone 5
(center). Visual layout example – text specifics are not pertinent.

74

Figure 33 – Altered Mock Drug Labels 3.6 and 3.7: Left – Product 3 Zone 6 (center right), Right – Product 3 Zone 7
(bottom left). Visual layout example – text specifics are not pertinent.

75

Figure 34 – Altered Mock Drug Labels 3.8 and 3.9: Left – Product 3 Zone 8 (bottom center), Right – Product 3
Zone 9 (bottom right). Visual layout example – text specifics are not pertinent.

76

Figure 35 – Altered Mock Drug Labels 4 and 4.1: Left – Product 4 label, Right – Product 4 Zone 1 (top left) altered.
Visual layout example – text specifics are not pertinent.

77

Figure 36 – Altered Mock Drug Labels 4.2 and 4.3: Left – Product 4 Zone 2 (top center), Right – Product 4 Zone 3
(top right). Visual layout example – text specifics are not pertinent.

78

Figure 37 – Altered Mock Drug Labels 4.4 and 4.5: Left – Product 4 Zone 4 (center left), Right – Product 4 Zone 5
(center). Visual layout example – text specifics are not pertinent.

79

Figure 38 – Altered Mock Drug Labels 4.6 and 4.7: Left – Product 4 Zone 6 (center right), Right – Product 4 Zone 7
(bottom left). Visual layout example – text specifics are not pertinent.

80

Figure 39 – Altered Mock Drug Labels 4.8 and 4.9: Left – Product 4 Zone 8 (bottom center), Right – Product 4
Zone 9 (bottom right). Visual layout example – text specifics are not pertinent.

81

Figure 40 – Altered Mock Drug Labels 5 and 5.1: Left – Product 5 label, Right – Product 5 Zone 1 (top left) altered.
Visual layout example – text specifics are not pertinent.

82

Figure 41 – Altered Mock Drug Labels 5.2 and 5.3: Left – Product 5 Zone 2 (top center), Right – Product 5 Zone 3
(top right). Visual layout example – text specifics are not pertinent.

83

Figure 42 – Altered Mock Drug Labels 5.4 and 5.5: Left – Product 5 Zone 4 (center left), Right – Product 5 Zone 5
(center). Visual layout example – text specifics are not pertinent.

84

Figure 43 – Altered Mock Drug Labels 5.6 and 5.7: Left – Product 5 Zone 6 (center right), Right – Product 5 Zone 7
(bottom left). Visual layout example – text specifics are not pertinent.

85

Figure 44 – Altered Mock Drug Labels 5.8 and 5.9: Left – Product 5 Zone 8 (bottom center), Right – Product 5
Zone 9 (bottom right). Visual layout example – text specifics are not pertinent.

86

Figure 45 – Altered Mock Drug Labels 6 and 6.1: Left – Product 6 label, Right – Product 6 Zone 1 (top left) altered.
Visual layout example – text specifics are not pertinent.

87

Figure 46 – Altered Mock Drug Labels 6.2 and 6.3: Left – Product 6 Zone 2 (top center), Right – Product 6 Zone 3
(top right). Visual layout example – text specifics are not pertinent.

88

Figure 47 – Altered Mock Drug Labels 6.4 and 6.5: Left – Product 6 Zone 4 (center left), Right – Product 6 Zone 5
(center). Visual layout example – text specifics are not pertinent.

89

Figure 48 – Altered Mock Drug Labels 6.6 and 6.7: Left – Product 6 Zone 6 (center right), Right – Product 6 Zone 7
(bottom left). Visual layout example – text specifics are not pertinent.

90

Figure 49 – Altered Mock Drug Labels 6.8 and 6.9: Left – Product 6 Zone 8 (bottom center), Right – Product 6
Zone 9 (bottom right). Visual layout example – text specifics are not pertinent.

91

BIBLIOGRAPHY

92

BIBLIOGRAPHY

1.

Kohn, L.T., et al., To err is human building a safer health system, 2000, National
Academy Press,: Washington, D.C. p. xxi, 287 p.

2.

National Coordinating Council for Medication Error Reporting and Prevention.

3.

Miller, J.A., Labeling--the State of the Art. 1978: Marketing Science Institute.

4.

Brown, C.S., et al., Short-Term and Long-Term Effects of Auxiliary Labels on
Patient Knowledge of Precautionary Drug Information. Drug Intelligence &
Clinical Pharmacy, 1988. 22(6): p. 470-474.

5.

Aspden, P. and Institute of Medicine (U.S.). Committee on Identifying and
Preventing Medication Errors., Preventing medication errors. Quality chasm
series. 2007, Washington, DC: National Academies Press,. xvi, 463 p.

6.

Rensink, R.A., J.K. ORegan, and J.J. Clark, To see or not to see: The need for
attention to perceive changes in scenes. Psychological Science, 1997. 8(5): p.
368-373.

7.

Treisman, A.M. and G. Gelade, Feature-Integration Theory of Attention.
Cognitive Psychology, 1980. 12(1): p. 97-136.

8.

Theeuwes, J., Visual Selective Attention - a Theoretical-Analysis. Acta
Psychologica, 1993. 83(2): p. 93-154.

9.

Jonides, J., Further toward a Model of the Minds Eyes Movement. Bulletin of the
Psychonomic Society, 1983. 21(4): p. 247-250.

10.

Eriksen, C.W. and J.E. Hoffman, Temporal and Spatial Characteristics of
Selective Encoding from Visual Displays. Perception & Psychophysics, 1972.
12(2B): p. 201-&.

11.

Eriksen, C.W. and J.D. Stjames, Visual-Attention within and around the Field of
Focal Attention - a Zoom Lens Model. Perception & Psychophysics, 1986. 40(4):
p. 225-240.

12.

Goldstein, E.B., Sensation and perception. 3rd ed. 1989, Belmont, Calif.:
Wadsworth Pub. Co. xx, 598 p., 12 p. of plates.

13.

Ditchburn, R.W. and B.L. Ginsborg, Vision with a Stabilized Retinal Image.
Nature, 1952. 170(4314): p. 36-37.

93

14.

Martinez-Conde, S., S.L. Macknik, and D.H. Hubel, The role of fixational eye
movements in visual perception. Nature Reviews Neuroscience, 2004. 5(3): p.
229-240.

15.

Iarbus, A.L., Eye movements and vision. 1967, New York,: Plenum Press. xiii,
222 p.

16.

Ratliff, F. and L.A. Riggs, Involuntary Motions of the Eye during Monocular
Fixation. Journal of Experimental Psychology, 1950. 40(6): p. 687-701.

17.

Otero-Millan, J., et al., Saccades and microsaccades during visual fixation,
exploration, and search: Foundations for a common saccadic generator. Journal
of Vision, 2008. 8(14).

18.

Theeuwes, J., Exogenous and Endogenous Control of Attention - the Effect of
Visual Onsets and Offsets. Perception & Psychophysics, 1991. 49(1): p. 83-90.

19.

Astle, D.E. and G. Scerif, Using Developmental Cognitive Neuroscience to Study
Behavioral and Attentional Control. Developmental Psychobiology, 2009. 51(2):
p. 107-118.

20.

Rueda, M.R., et al., Training, maturation, and genetic influences on the
development of executive attention. Proceedings of the National Academy of
Sciences of the United States of America, 2005. 102(41): p. 14931-14936.

21.

Grabowecky, M., L.C. Robertson, and A. Treisman, Preattentive Processes
Guide Visual-Search - Evidence from Patients with Unilateral Visual Neglect.
Journal of Cognitive Neuroscience, 1993. 5(3): p. 288-302.

22.

He, P.Y. and E. Kowler, The Role of Location Probability in the Programming of
Saccades - Implications for Center-of-Gravity Tendencies. Vision Research,
1989. 29(9): p. 1165-1181.

23.

Byrne, M.D., et al., Eye Tracking the Visual Search of Click-Down Menus, in CHI
99: the CHI is the limit, human factors in computing systems, M.G. Williams and
S. Pemberton, Editors. 1999, ACM Press: New York. p. 402-409.

24.

Just, M.A. and P.A. Carpenter, Eye Fixations and Cognitive-Processes. Cognitive
Psychology, 1976. 8(4): p. 441-480.

25.

Jacob, R.J.K. and K.S. Karn, Commentary on Section 4 Eye Tracking in HumanComputer Interaction and Usability Research: Ready to Deliver the Promises, in
European conference on eye movement research; The mind's eye, J. Hyona, R.
Radach, and H. Deubel, Editors. 2003, London: Amsterdam. p. 573-606.

94

26.

Vanderplas, J.M. and J.H. Vanderplas, Some Factors Affecting Legibility of
Printed Materials for Older Adults. Perceptual and Motor Skills, 1980. 50(3): p.
923-932.

27.

Morrow, D.G., et al., Medication instruction design: Younger and older adult
schemas for taking medication. Human Factors, 1996. 38(4): p. 556-573.

28.

Watanabe, R.K., K. Gilbreath, and C.C. Sakamoto, The ability of the geriatric
population to read labels on over-the-counter medication containers. J Am
Optom Assoc, 1994. 65(1): p. 32-7.

29.

Wogalter, M.S. and W.J. Vigilante, Effects of label format on knowledge
acquisition and perceived readability by younger and older adults. Ergonomics,
2003. 46(4): p. 327-344.

30.

Shrank, W.H. and J. Avorn, Educating patients about their medications: The
potential and limitations of written drug information. Health Affairs, 2007. 26(3): p.
731-740.

31.

Davis, T.C., et al., Low literacy impairs comprehension of prescription drug
warning labels. Journal of General Internal Medicine, 2006. 21(8): p. 847-851.

32.

Wolf, M.S., et al., Misunderstanding of prescription drug warning labels among
patients with low literacy. American Journal of Health-System Pharmacy, 2006.
63(11): p. 1048-1055.

33.

Kripalani, S., et al., Predictors of medication self-management skill in a lowliteracy population. Journal of General Internal Medicine, 2006. 21(8): p. 852856.

34.

Nielsen-Bohlman L, P.A., Kindig D. A., Health Literacy: A Prescription to End
Confusion, 2004, Institute of Medicine: Washington, DC: National Academy Pr.

35.

Sojourner, R.J. and M.S. Wogalter, The Influence of Pictorials on the Evaluations
of Prescription Medication Instructions, in Human Factors and Ergonomics
Society. 1996, The Society. p. 1281.

36.

Morrell, R.W., D.C. Park, and L.W. Poon, Quality of Instructions on Prescription
Drug Labels - Effects on Memory and Comprehension in Young and Old Adults.
Gerontologist, 1989. 29(3): p. 345-354.

37.

Young, S.L. and M.S. Wogalter, Comprehension and Memory of Instruction
Manual Warnings - Conspicuous Print and Pictorial Icons. Human Factors, 1990.
32(6): p. 637-649.

95

38.

Houts, P.S., et al., The role of pictures in improving health communication: A
review of research on attention, comprehension, recall, and adherence (vol 61,
pg 173, 2006). Patient Education and Counseling, 2006. 64(1-3): p. 393-394.

39.

Rayner, K., et al., Integrating text and pictorial information: Eye movements when
looking at print advertisements. Journal of Experimental Psychology-Applied,
2001. 7(3): p. 219-226.

40.

Hoffman, J.M. and S.M. Proulx, Medication errors caused by confusion of drug
names. Drug Safety, 2003. 26(7): p. 445-452.

41.

Flynn, E.A., K.N. Barker, and B.J. Carnahan, National observational study of
prescription dispensing accuracy and safety in 50 pharmacies. J Am Pharm
Assoc (Wash), 2003. 43(2): p. 191-200.

42.

Filik, R., et al., Labeling of medicines and patient safety: evaluating methods of
reducing drug name confusion. Hum Factors, 2006. 48(1): p. 39-47.

43.

Bass, P.F., J.F. Wilson, and C.H. Griffith, A shortened instrument for literacy
screening. Journal of General Internal Medicine, 2003. 18(12): p. 1036-1038.

44.

Poole, A., Ball, L.J. , Eye Tracking in Human-Computer Interaction and Usability
Research: Current Status and Future, in Prospects, Chapter C. Ghaoui (Ed.)
Encyclopedia of Human-Computer Interaction.(2005), Idea Group, Inc.:
Pennsylvania.

45.

Coeffe, C. and J.K. Oregan, Reducing the Influence of Nontarget Stimuli on
Saccade Accuracy - Predictability and Latency Effects. Vision Research, 1987.
27(2): p. 227-240.

46.

Findlay, J.M., Global Visual Processing for Saccadic Eye-Movements. Vision
Research, 1982. 22(8): p. 1033-1045.

47.

Zhou, X.L., et al., Center of mass attracts attention. Neuroreport, 2006. 17(1): p.
85-88.

48.

Kwak, H.W., D. Dagenbach, and H. Egeth, Further Evidence for a TimeIndependent Shift of the Focus of Attention. Perception & Psychophysics, 1991.
49(5): p. 473-480.

49.

Egly, R. and D. Homa, Sensitization of the Visual-Field. Journal of Experimental
Psychology-Human Perception and Performance, 1984. 10(6): p. 778-793.

50.

Humphreys, G.W. and V. Bruce, Visual cognition : computational, experimental,
and neuropsychological perspectives. 1989, Hove, UK ; Hillside, USA: L.
Erlbaum. x, 330 p.
96

51.

Remington, R.W., Attention and Saccadic Eye-Movements. Journal of
Experimental Psychology-Human Perception and Performance, 1980. 6(4): p.
726-744.

97