 

 

 

 

yum.
,. , l t
”A”... 1. 3.6%.

4‘
{isﬁﬁ at. u. i r_

.gﬁn.” . .2 , n .

.rr 1..
Eacv 4..
tiuvvufﬂc.

‘ 458‘

u . .8...

_
. D .O

i: _ c: .1.

saw}?
.35....

.t
In}. ..
31%|”;

15.5.”...

11.
3%
lie; .
it.»
3...;

ﬂan-(II I’m"

llzulx ‘
til}! I. 5
5“! I. Vr

I:
l: 53,...) .0.
I.
lie .. :. a...) 2.5!...
I .r.ﬂ¥l&ﬂ.§¢lﬂ1.$ﬁnﬁl|ﬂ¢m
a... .3... .Et-zioa

.lo

:RUW
2.3311 .3.
.zvifui..w

215.9“ 31‘0'0

v Elviail

.15!

7 5.3.7.5.].
{g 1;

. I955} its...
74.3? 3;.

Eateriixfﬂ

2.
.fSHuE.
3 .r .13.!

4715.5.
.>. : .415

 

2
LIBRARY
3009 Michigan State
University

 

 

 

This is to certify that the
thesis entitled

BIOMECHANICAL ANALYSIS OF OPENING GLASS JARS:

USING KINEMATICES TO INFORM DESIGN FOR
IMPROVED ACCESSIBILITY

presented by

JOSEPH ROBERT FAIR

has been accepted towards fulﬁllment
of the requirements for the

 

 

 

 

MS. degree in PACKAGING
W111 A ﬂL‘éy
/ " 7Major Professor s‘S’gnature
Date

MSU is an Afﬁnnative Action/Equal Opportunity Employer

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 Kilproj/AccaPreleIRCIDateDue.indd

 

BIOMECHANICAL ANALYSIS OF OPENING GLASS JARS: USING
KINEMATICS TO INFORM DESIGN FOR IMPROVED ACCESSIBILITY

By

Joseph Robert Fair

A THESIS
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
MASTER OF SCIENCE
PACKAGING

2009

Copyright by
JOSEPH ROBERT FAIR
2009

ABSTRACT

BIOMECHANICAL ANALYSIS OF OPENING JARS: USING KINEMATICS TO
INFORM DESIGN FOR IMPROVED ACCESSIBILITY

By
Joseph Robert Fair

The research presented in this thesis theorizes that there are three styles
of hand movement which can be used when removing 3 lug style closure from a
glass jar. These movements include closure hand movement alone (either right
or left hand), jar hand movement alone (either right or left hand), or the
movement of both hands (either right hand or left on the closure or jar). However,
it is hypothesized that experimental set up will signiﬁcantly impact how
participants move as they remove lug closures; that an unrestrained subject will
move differently than if restrained for ease of experimental set up. Each study
participant will open glass jars with lug style closures in two fashions: restrained
so that the test jar remains on the table and unrestrained so that the jar begins in
a comfortable height above the table.

In order to be able to accurately capture the opening movement when
subjects are not restrained, the system must be ﬁxtured, creating a “reference
point.” Twenty one subjects without a history of injury to the hands, wrists, arms,
or shoulders were used in the study in order to create the ﬁxturing system, reﬁne
the methodology and test the hypothesis that restraining subjects inﬂuences the
movement used in order to open the jars. This information can help guide future
studies’ experimental set ups along with providing vital information for the design

of jars related to the motion used to open jars.

DEDMJATKMVFUHSE

For Katie

ACKNOWLEDGEMENTS

I would like to thank my committee members Dr. Laura Bix, Dr. Tamara
Reid Bush and Dr. Gary Burgess for all of their time, assistance and support. I
want to thank the Center for Food and Pharmaceutical Packaging Research
(CFPPR) for funding this research. I also want to thank Mr. Rick Bayer of the
Glass Packaging Institute (GPI). His help lead to the acquisition of the glass jars
and closures. I would also like to thank St. Gobain Containers for donating the
glass jars and Silgan White Cap Americas for donating the closures. Thank you
to Nora Bello of The Center for Statistical Training and Consulting at Michigan
State University for providing the statistical analysis of the data. Thank you to all
of the professors, staff and graduate students at the School of Packaging for their
encouragement and help, speciﬁcally Chris Steckler for assisting me with the
data collection. I would like to thank all of Dr. Bush’s graduate students for
sharing lab time and assisting with my data collection, speciﬁcally Sam Leitkam
for all of his assistance. Finally I would like to thank Katie for all of her support

and encouragement over the last four years.

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................... viii
LIST OF FIGURES ....................................................................... x
CHAPTER 1 - Literature Review ................................................... 1
Torque Studies ................................................................... 1
Motion Studies ...................................................................... 4
Laterality ........................................................................... 5
Physical Setup of the Experiment ........................................... 6
CHAPTER 2
Objective ........................................................................................... 8
Hypothesis ......................................................................... 8
CHAPTER 3 - Materials and Methodology ...................................... 9
Demographics .................................................................... 9
Anthropometrics .................................................................. 1 0
Strength Measurements ........................................................ 13
Laterality Measurements ....................................................... 16
Jars and Closures Used During Motion Capture ......................... 17
Grip Style Determination ....................................................... 19
Biomechanical Motion Capture ............................................... 20
Motion Capture Design ......................................................... 21
Setup of Openings ............................................................... 22
Restricted Opening .............................................................. 24
Unrestricted Opening ............................................................ 28
Data Transfer ...................................................................... 30
Data Processing .................................................................. 31
Table Openings ...................................................................... 31
Unrestricted Openings ................................................. 35
Pole Rig ........................................................... 35
Side Rig .......................................................... 40
Determination of Degrees Rotation .......................................... 44
CHAPTER 4 - Results and Discussion ............................................ 47
Subject Information ................................................................ 47
Hand Size ............................................................................ 49
Analysis of Strength Data ........................................................ 53
Laterality .............................................................................. 54
Opening Data ........................................................................ 56
Rig Selection ........................................................................ 64
Effect of Rig Type and Jar Size on % Rotation Degrees by Jar
Hand ................................................................................................... 66

vi

Effect of Opening and Jar Size on % Rotation Degrees by Jar
Hand. ................................................................................................. 73
Rig Type under Unrestricted Movement Conditions... 83

CHAPTER 5 — Conclusions ............................................................. 88
CHAPTER 6 - Recommendations and Future Studies ......................... 92
Appendices ................................................................................... 95
Appendix A - Recruitment Flyer ................................................. 96
Appendix B — Data Collection Form ............................................. 97
Appendix C - Torque Requirements ............................................ 100
Appendix D — Motion Capture Calibration ..................................... 104
Appendix E — Opening Rotation Sequence ................................... 106
Bibliography ............................................................................................... 107

vii

Table 1

Table 2
Table 3
Table 4
Table 5
Table 6
Table 7
Table 8
Table 9

Table 10

Table 11

Table 12
Table 13
Table 14
Table 15
Table 16
Table 17
Table 18
Table 19

Table 20

LIST OF TABLES

Jar speciﬁcations for the two jars used for motion

capture ......................................................................... 17
Overview of subjects participating in study ................ 48
Anthropometric data ........................................... 49
Table of strength measurements ........................... 53
Laterality results by gender and dexterity ................. 54
Laterality hand by laterality arm statistics ................. 55
Overview of rotation by percent and degrees ............ 57
Pearson correlation coefﬁcients ............................. 59
Total openings for 80_20 and 90_1O groupings ......... 64
Rig type by gender statistics for unrestricted

openings ....................................................................... 65
Table of jar size by opening position, gender and

closure hand ................................................................. 67
Solution for ﬁxed effects ....................................... 70
Jar hand rotation by hand breadth and gender ......... 71
Type 3 Tests of ﬁxed effects .................................. 73
Solution for ﬁxed effects ....................................... 76
Type 3 tests of ﬁxed effects .................................. 77
Rotation by jar size ............................................. 78
Differences of least square means ......................... 79
Differences of least square means ......................... 80
Least square means ........................................... 81

viii

Table 21
Table 22
Table 23
Table 24
Table 25

Table 26

Solutions for ﬁxed effects ..................................... 85
Jar size least square means ................................. 85
Tukey — Kramer jar size ....................................... 86
Lat cross least squares means .............................. 86
Tukey - Kramer lat cross ...................................... 86

Trials used to calculate average opening angles (0)
for both the large and small jar .................................. 102

Figure 1

Figure 2

Figure 3

Figure 4
Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10
Figure 11
Figure 12
Figure 13
Figure 14
Figure 15
Figure 16

Figure 17

Figure 18

Figure 19

LIST OF FIGURES
Camera mounting rig for digital pictures ...................... 11

Photographs taken to classify the size of each
subject’s hands ............................................................ 12

CorelDraw X3 measurement example, 1 — hand
length, 2 — index ﬁnger length, 3 — hand breadth ........ 13

Jarrnar® grip dynamometer .................................... 14
Baseline ® pneumatic squeeze dynamometer .......... 15

“Postural Preferences,” 1 shows hand clasp and 2
shows arm fold ............................................................. 16

Photographs of the two jar sizes used for testing; _
Jar A and Jar B. ........................................................... 17

70 - 2030 closure with 2.5 inch white matte label ...... 18

Grip styles used for unrestrained openings .............. 20
Location of hand targets on a subject ...................... 21
Sequence diagram for the 12 openings ................... 23
Foam boards created for restricted opening .............. 24
Restricted opening reference Jar A and Jar B ............ 25
Overhead view of table and camera setup ................ 26
Front view of table with restricted reference jar .......... 27
Unrestricted open rigs for bottom and side grips ........ 28

Qualisys screen shot of side rig and pole rig
opening ......................................................................... 29

Subject opening side rig (left) and pole rig (right) ....... 30

Central rotation axis for restricted openings .............. 32

Figure 20 Vector analysis of targets ...................................... 33
Figure 21 Central rotation axis for pole rig openings ................ 35
Figure 22 Side rig opening reference ﬁle ............................... 40
Figure 23 Graphs used for calculating rotation supplied by each

hand. ............................................................................ 45
Figure 24 Current study hand breadth vs. NTIS data ............... 50
Figure 25 Current study wrist — ﬁnger length vs. NTIS data ....... 51
Figure 26 Current study hand length vs. NTIS data ................. 52
Figure 27 Strength measurements ........................................ 54
Figure 28 Number of subjects representing each of the four

laterality codes ............................................................. 56
Figure 29 Openings by rotation hand .................................. 58
Figure 30 Closure hand openings ...................................... 61
Figure 31 Restricted closure hand openings ........................ 62
Figure 32 Unrestricted closure hand openings ..................... 63
Figure 33 Jar hand rotation by right hand breadth (mm) .......... 72
Figure 34 Jar hand rotation by opening restriction .................. 82
Figure 35 Secure Pak, Inc. electronic torque tester ................. 100
Figure 36 Lines drawn on closure to determine removal

distance ...................................................................... 103
Figure 37 Calibration bracket and wand ................................ 104

xi

Chapter 1

LITERATURE REVIEW

Having trouble opening packaging is by no means a new phenomenon,
and jars have been noted to be particularly problematic for consumers, especially
if they have a disability or as they increase in age [1 — 4]. The success or failure
of a jar opening seems to be strongly related to a consumer’s prior knowledge of
the package, their understanding of new opening features, and their ability to
meet the physical requirements necessary [1].

Torque Studies

Most research that looks at the trouble associated with opening glass jars
has been related to consumers’ physical abilities to overcome the torque applied
to the closure [1 - 4]. However, torque is a single factor of many that impact
utility, or lack thereof, for these packages.

A commonly used industry practice is to apply a closure at a torque equal
to one half the finish diameter of the closure [5]. For example, a jar with an
82mm ﬁnish would be torqued to 41 in — lbs, or 4.63Nm. This industry “rule of
thumb” presents concerns in light of the available usability research. Voorbij and
Steenbekkers (2002) indicate that although consumers under 30 generally do not
have trouble opening glass jars; subjects over 50 do. Further, the research team
suggests that opening torques should be limited to 2Nm (17.7 in — lbs) for
maximum usability of subjects over 50 years of age. Their team makes note of

the lack of enthusiasm from packaging professionals for this ﬁnding, indicating, “It

is strange, therefore, that the packaging industry has not yet addressed the
problem by changing this type of packaging” [3].

A study conducted by the UK Department of Trade and Industry (DTI) [2]
investigated the torque that a consumer could apply to three different diameters
aluminum “jars” while sitting or standing. The study did not indicate if subjects
where restricted to opening on a table. Smooth and knurled closures were used
in the study and included a non - disabled population along with a disabled
population. The study had several interesting ﬁndings. The non - disabled
population could apply a torque 2 to 2.5 times greater than the disabled
population for both closure styles. The greater the diameter, the greater the
applied torque. Also, the knurled closure allowed for a greater applied torque.
However, the research did not investigate the motion associated with opening
closures. The data allows for an understanding of how torque limits the ability of
a person to open a jar and suggests that knurled closures improve the ability of
someone who has less strength to successfully open a jar. However, knurled
closures alone cannot overcome the application torque and vacuum applied to
closures.

Unlike most studies related to this topic, Voorbij and Steenbekkers (2002)
allowed the subjects to replicate opening in a less controlled environment.
Subjects were allowed to stand and either open the jar on the table surface or
open the jar in an unrestricted fashion. However, their research examined the

forces associated with the use of an aluminum jar and subjects could only apply

a force to the jar, not open the jar, as did the UK Department of Trade and
Industry study in 2002 [3, 2].

Crawford, Wanibe, and Nayak (2002) examined the torque subjects could
apply by, again, using a restrained opening location for jar and closure. The test
population included healthy younger and older adults and subjects were allowed
to use either hand on the closure. Standard glass jars were used but nylon
closure were created in round (20mm, 50mm and 80mm) and square test
(50mm) pieces. This, again, creates a situation that does not completely relate
to real life. However, an interesting conclusion of the study is that the square
closures allowed users to generate a greater torque. The research found that
males could apply more torque than females, and that younger subjects were
able to apply more torque than older subjects. Another interesting ﬁnding from
this study relates to anthropometrics. There were signiﬁcant correlations
between hand length and grip strength, along with hand breadth and grip
strength [4].

Research conducted by Yoxall et. al. (2006) investigated the torque
applied to 75mm jars and employed a glass jar. This was done in an effort to
ensure that the torque values would not be altered due to frictional differences
between glass and aluminum, which they had seen in previous studies, allowing
for a more realistic test. However, the closures could not be removed from the
jars. Subjects were allowed to stand or sit and could hold the jar in any way they
wanted. Subjects were also required to open the jar bolted to a table for

comparison. Multiple tries were allowed. Several interesting ﬁndings came from

3

this work. Male subjects were able to apply a greater amount of torque to the jar,
consistent with that of Crawford, Wanibe, and Nayak (2002) [1, 4]. Also, the
torque applied to the bolted down jar were lower than the un — restricted jar.
The research did not investigate the effect of the movement, or motion employed
by the user as it related to torque.
Motion Studies
Much of the research that measures motion restricts the subject, providing
potentially unrealistic scenarios [6, 7]. Studies that have restrained subject
motion very well may alter the way subjects perform opening procedures and, as
a result, may or may not provide value for designers.
Fowler and Nicol (1999) collected strength data and assessed both the
force and motion of opening jars, speciﬁcally aimed at the loading forces on the
interphalangeal joint of the dominant hand index ﬁnger. However, their study
design did not accurately replicate reality. Theirjar was not made of glass, and
s ubjects were required to keep the jar on the table. Second, and more
importantly, subjects were instructed to always use their dominant hand on the
closure of the jar. Requiring subjects to place their dominant hand on the closure
makes the assumption that this would be the most common placement used
when opening a jar [6].
Murgia et. al. (2004) used motion capture technology to investigate wrist
kinematics during jar opening. Subjects had to press a timer prior to opening a
jar with a 70mm closure. Subjects were asked to remove the closure and placed

the jar and closure back on the table and press the timer again. However, the

4

subjects were seated and required to open the jar closure with their dominant
hand, again assuming that this would be the desired placement [7].
Laterality
An interesting explanation for hand placement during tasks such as

opening jars may be related to laterality, a research area that examines postural
preferences. Mohr et.al. (2003, 2006) examined postural preferences using arm
— folding and hand — clasping. Their research identiﬁed four possible groups for
these two behaviors. Left thumb or arm on top was coded with an L and right
thumb or arm on top was coded with an R. Therefore, the four possible groups
were coded LL, LR, RL or RR (hand preference followed by arm preference).
People who placed either both left or right on top were considered congruent
whereas people who placed a combination of left or right were considered

incongruent [8, 9]. In their 2003 study all subjects (n = 362) were right handed [8].
1 r1 their 2006 study 86.8 percent (442 of 509 subjects) were right handed [9]. For
the 2006 study, handedness was based on a 13 item questionnaire. They found
that the greatest number of people prefer their left hand on top for both hand
clasping and arm crossing, but that this did not relate to their dominate hand.
In research related to packaging, speciﬁcally jars, handedness was found
to be signiﬁcantly related to the closure and jar hand placement. Voorbij and
Steenbekkers (2002) studied 743 subjects and found that most right handed

Subjects placed their left hand on the closure (511 of 534 subjects) and left
handed subjects placed their right hand on the closure (147 of 209 subjects) [3],

going against the assumption used by Fowler and Nicol (1999) [6]. This ﬁnding

5

coincides with those of Mohr et. al. (2003, 2006) [8, 9]. However, they did not
investigate factors that explained the positioning of the hand on the jar or closure;
perhaps, handedness, postural preferences (laterality), injury or just random
placement are likely causative factors for hand placement and motion. As part
of the study presented here, researchers will examine data for possible
correlations between postural preference and grip positioning.
Physical setup of the experiment
A review of the literature indicates that the physical set up of the
experiment is also crucial. Stins, Kadar, and Costall (2001) noted that the
contents of a glass, along with the location of the glass, relative to the subjects’
dominant hand, were signiﬁcant factors in which hand the subject used to reach
for the glass. Glasses set directly in front of a subject were most likely to be
grasped with the dominant hand of the subject as opposed to the hand closest to
the glass. Fixing the test stimulus, as compared to allowing users to move freely,
as mentioned, has been shown to impact user ability [10]. Yoxall et. al. (2006)
n oted that the torque applied to a bolted down jar was lower than the jar opening
Where users were unrestricted [1]. This suggests that there could also be
differences in the opening motion used between these two conditions.
For the purposes of this research, we will explore how people open lug
Style closures from a motion perspective when constrained to a table surface and
When opening in an unrestricted manor. Perhaps people open jars differently
When constrained to an x, y coordinate system than when allowed the freedom to

Open in an unrestricted x, y, z coordinate system.

6

Research that investigates the unhindered motion of the user when
opening jars is needed because: (1) pilot studies conducted by the team indicate
that restraining subjects may affect their motion (2) limited research into
biomechanics and packaging is available, and what is frequently restrains
subjects while opening packaging to aid data collection and analysis, not
necessarily representing a real life situation (3) methodologies that characterize
opening motions need to be developed (4) little is known about opening motions
used for glass jars, yet it is widely understood that they are difﬁcult to open [1 —

4].

Chapter 2

OBJECTIVE

The overall objective of this study was to create and reﬁne a methodology
for using motion capture in packaging applications. Glass jars with smooth style
metal lug style closures were used to create the methodology, but it is anticipated
that the techniques employed can be used to characterize the opening motions
associated with varied package types in future work. Additionally, this research
investigates the impact of experimental setup, restrained or unrestrained, on the
motions employed during opening and begins to explore how people move when
opening jars.

HYPOTHESIS

It was hypothesized that the physical positioning of the subject had the
potential to impact the opening motion employed. Therefore, two opening
positions were incorporated into the experiment; opening “restrained” to a
counter set at a standard height of 91 cm (36 inches) [11] the other opening in
the air at a height comfortable to the subject, “unrestrained opening.“

It was also hypothesized that the motion used by a subject to open glass
jars with lug style closures would take one of three forms;

1. Closure Hand Supplies Motion

2. Jar Hand Supplies Motion

3. Both Hands Supply Motion

Chapter 3

MATERIALS AND METHODOLOGY

Subjects were recruited by word of mouth and represented a convenience
sample. Data was collected at the Biomechanical Design Research Laboratory
(BDRL) in the department of mechanical engineering at Michigan State
University. Data was collected from 25 subjects using procedures approved
under lRB 06-981. Subjects were screened prior to testing to ensure they did not
have a history of injury to their hands, wrists, arms or shoulders, or if they
indicated allergies to surgical tape or rubbing alcohol. Subjects without a history
of injury were used. Since there is little research into this area a healthy
population is needed to create a baseline for future studies In exchange for
participation, subjects were compensated with a twenty ﬁve dollar gift certiﬁcate
to Meijer ®. Once informed consent had been obtained, data was collected as
outlined.

Demographics

Following the screening process and signing of the consent form, a
member of the research team collected demographic data from subjects, see
Appendix B for a copy of the Data Collection Form. Demographic data included
gender, age and preferred dexterity. Subject under the age of 30 were targeted
based on studies related to torque that indicated subjects in this age range would
not have difﬁculty overcoming the torque applied to closures on glass jars [3].

Preferred dexterity was determined by asking each subject which hand they

considered to be their dominant prior to collecting any data. The subject was
also asked about having a history or allergic reaction or irritation to rubbing
alcohol or surgical tape as well as a history or injury to the hands, shoulder, arms
or wrists. If the subject indicated a positive response to either of the previous
questions they were thanked for their time, given the gift card and excused from
the remainder of the testing. All other subjects moved fonNard with the testing
procedure.

Anthropometrics

Using a photographic method employed by previous studies [2],
researchers characterized each subject’s band size. Subjects were asked to
place the hands, ﬁrst right, then left hand, on a ﬂat base that had a grid
consisting of 10 mm squares. The base was equipped with an articulated arm
that served as a mount for a digital camera (Canon PowerShot G6, PC1089, 7.1
Megapixels, No. 9121102252) mounted parallel to the base at a ﬁxed distance

from the hands, see Figure 1.

10

 

Figure 1 Camera mounting rig for digital pictures

Photographs were taken of the right and left hand in two different positions
for a total of four different photographs, see Figure 2. No preference was given

to which order subjects placed their hands on the grid

11

 

Figure 2 - Photographs taken to classify the size of each subject’s hands

Using CorelDraw X3 for Windows (Version 1300739), pictures were
scaled so that the length of the index finger and the breadth of the hand could be
determined. Figure 3 shows an example of a subject photo scaled using
CorelDraw. Measurement 1 in Figure 3 is Hand Length, measurement 2 is Wrist

- Index Finger Length and measurement 3 is Hand Breadth.

12

226.29

 

Figure 3 — CorelDraw X3 measurement example, 1 — hand 2 — index

length, 3 — hand breadth
Strength Measurements
Following the collection of anthropometric data, the subjects went through
a series of strength measurements. Strength measurements were taken to

determine grip strength and palm to palm squeeze strength. Grip strength was

13

collected using a Jamar® grip dynamometer, see Figure 4, and the subject’s
bilateral palm-to—palm squeeze strength was collected using a Baseline®
pneumatic squeeze dynamometer, see Figure 5. Strength measurements were

taken in accordance with the American Society of Hand Therapist (ASHT).

    

Figur§.;‘Jarmar® grip dynamometer

For collection of grip strength subjects were seated in a straight backed
chair without arrnrests and their feet ﬂat on the ﬂoor. Subjects were handed the
Jamar® grip dynamometer and instructed to hold the dynamometer so that their
elbow was at a 90 degree angle and that their arm was not resting on any part of
them or the chair. Subjects were instructed by a member of the research team to
squeeze as hard as they could for a total of 3 seconds. Time was kept by the
tester through a verbal command of, “Squeeze as hard as you can...harder,

harder, relax.” In order to avoid fatigue, each subject began with the

14

dynamometer in their right hand and then switched back and forth between their
right and left hand for a total until there were a total of three repetitions for each
hand.

For collection of the bilateral palm to palm squeeze strength subjects
remained seated in a straight backed chair without arrnrests with their feet ﬂat on

the ﬂoor.

 

rk

Figure 5 —Baseline ® pneumatic squeeze dynamometer
Subjects were handed the Baseline® pneumatic squeeze dynamometer
and instructed to hold the dynamometer at chest level between their palms.

Subjects were instructed to hold their hands at a 90 degree angle from each

15

other and to not lock their hands or ﬁngers. Subjects were asked to squeeze as
hard as they could for a total of 3 seconds. Time was again kept by the tester
through a verbal command, “Squeeze as hard as you can...harder, harder,
relax."

Laterality Measurements

The last data collected prior to the collection of the motion data were
measurements of laterality or “postural preference.” Subjects were asked to
stand in front of a set of curtains so that a series of photographs could be taken.
Subjects were ﬁrst instructed to lace or clasp their hands together. Photograph
1, see Figure 6, shows an example of a subject demonstrating hand clasping
with the left thumb on top. The picture had to be taken at an angle that allowed
the research team to determine which thumb was on top during data processing.

Subjects were then asked to fold their arms. Photograph 2, see Figure 6, shows

an example of a subject demonstrating arm folding with the left arm on top.

    

k .5] r’ _.

.4. " A “J" .
Figure 6 — “Postural Preferences," 1 shows hand clasp and 2 shows arm fold

16

Jars and Closures Used During Motion Capture

Two jar sizes were selected and subjects opened them using two opening
orientations, restrained and unrestrained. The jars selected for testing are shown
in Figure 7 and accompanying speciﬁcations are shown in Table 1. The jars
used in the study were donated by Saint — Gobain and closures were donated by

Silgan through the assistance of the Glass Packaging Institute.

JarA

 

Figure 7 - Photographs of the two jar sizes used for testing; Jar A and Jar B

Finish . Diameter
MOId # Designation Height (cm) (cm)

- 1

 

Jar A and Jar B were chosen for testing in order to limit the differences
between the two. The heights of the two jars are nearly the same and both jars
have a ﬁnish designation of 70 — 2030 (the jars have a closure diameter of 70
mm). The closures were “Regular Twist — Off TM” style containing 4 lugs. Due

to the potential for glare off of the closure white matte ﬁnish circular labels, see

17

Figure 8, were placed on each closure. The glare from the closure could account

for false targets in the Qualisys software.

 

Figure 8 — 70 — 2030 closure with 2.5 inch white matte label

The main difference between Jar A and Jar B is the diameter of the jar,
see Table 1. The difference between the two jar diameters is 2.2225 cm.
Additionally, the jars were ﬁlled to two thirds of their overﬂow capacity with pop
corn kernels. The overflow capacity was determined using;

= (Bf — Be)
0.997

Bv

Where Bv is the volume in ﬂuid ounces, Bf is the full container weight in
grams, Be is the empty container weight in grams. Popcorn kernels were used

as a way to create contents that resembled liquid without having to bring liquids
18

into the test lab where they presented a danger to the testing equipment.
Overﬂow capacity was calculated for gram weight with water. The equal weight
in popcorn kernels was then added to the jars and used throughout the testing.
Jar A had a two thirds overﬂow weight of 963.88 grams and Jar B had a two
thirds overﬂow weight of 1507.37 grams. It was decided that full jars would not
be used since a vacuum could not be drawn on the jars for each opening. A full
jar would represent the ﬁrst opening in which case a vacuum would be present
along with the applied torque. Using a two thirds full jar also reduced the risk that
the subject would spill popcorn kernels during the opening trial.

Grip Style Determination

Prior to beginning the motion capture portion of the research, subjects
were asked to pick up Jar A and Jar B, from a table, and open them one at a
time. This demonstrated how the subject would open each in the unrestrained
position. Two opening styles were used by the subjects for the unrestrained
openings, a side grip and a bottom grip. Some subjects demonstrated the same
grip for both Jar A and Jar B and some subjects demonstrated both styles this
seemed to vary with the diameter, and will be discussed in greater detail in

Chapter 4. The two grip styles are shown in Figure 9.

19

 

Figure 9 — Grip styles used for unrestrained openings

Based on the two grip styles demonstrated by the subjects, two jar rigs
were created for testing during the unrestricted motion capture. The rigs are
shown in Figure 16. The rigs will be discussed in greater detail in the
Unrestricted Opening section.

Biomechanical Motion Capture

Following the demonstration of grip style, subjects were prepped for
motion capture testing. Subjects were asked to remove any jewelry or clothing
that incorporated reﬂective material. Subjects who wore long sleeves were
allowed to use a bathroom located next to the lab to change into a short sleeved
shirt provided by the research team. The next step consisted of swabbing the
hands and forearms of the subject where the reﬂective targets would be attached
using a rubbing alcohol pad, this assisted in removing dirt that could interfere

with the adhesion of the targets to the skin of the subject.

20

The targets were attached, using surgical tape, to three anatomical points
on the hands. The ﬁrst target was attached to the scaphoid bone (located at the
center of the base of the hand), the second to the proximal interphalangeal (PIP)
joint of the index ﬁnger (the middle joint of the index ﬁnger) and the third target to
the metacarpophalangeal (MCP) joint (bottom of the pinky) of the small ﬁnger
[12]. Two additional targets were placed on the forearms so that data could be
collected for potential future testing along with a target on the right shoulder. The
target on the right shoulder was used to identify the right hand versus the left
hand in the software. Figure 10 shows the location of the hand targets on a
subject.

(A) Scaphoid Bone

(B) Proximal lnterphalangeal (PIP) Joint
(C) Metacarpophalangeal (MCP) Joint

 

Left Hard Rght Hand
Figure 10 — Location of hand targets on a subject

Motion Capture Design

The motion capture system used for testing consisted of ﬁve infrared
cameras connected to a computer running Qualisys software. Reﬂective targets,
made from wooden spheres with a diameter of 12.5 mm covered in 3M reﬂective

tape (7610 WS, silver "High Gain Sheeting"), were taped to the subjects and

21

attached to the jars, see Figures 10, 13 and 16. The infrared light from the
cameras reﬂected off the targets and allows for the software to identify the
targets in 3 — Dimensional space, resulting in x, y, z coordinate data for each
target.

Setup of Openings

Jar A and Jar B were each opened three times in the restricted and
unrestricted positions for a total of 12 openings by each subject as shown in

Figure 11.

22

Restri cted

 

 

  
  
    

Opening \
Total Number of
Openings
12

 

 

 

 

Unrestricted I sue mg l

O enln
P 9 OR
JarA
3" l Pole Rig l
l Free 1
Side Rig
3’1 JarB I I J
OR

I Pole Rig I

Figure 11 - Sequence diagram for the 12 openings

 

 

 

 

 

 

 

To balance the effect of fatigue across treatments, the order in which
subjects opened the jars was rotated. Subject one began the sequence Jar A
table, Jar B table, Jar A free then Jar B free. Subject two began the sequence
Jar A table, Jar B table, Jar B free then Jar A free. This rotation sequence
continued allowing for a different opening sequence for all subjects, see

Appendix E. For the restricted openings, the subjects reused the same jar three

23

times and a new closure was applied for each trial, the jars were only used for
one subject before being removed from the testing. Subjects had time to rest
between each opening while a member of the research team applied the new
closure. For the unrestricted openings, two of each of the rigs were created by
the research team. The same rig was used for every subject; the secondary rig
was only to be used if there was a failure of the original rig.

Restricted Opening

In preparation for restricted opening, two foam board placement mats

were created, see Figure 12.

8.73252 cm

JarA" =

      

:4

Figure 12 — Foam boards created for restricted opening

The mats allowed the research team to ensure that the jar openings
performed on the table in the restricted position were done in the same position
for each trial. Velcro attached the boards to the table so that the mats were
placed in the same location from subject to subject. This also helped to prevent
the jar from sliding out of position during motion trials. The holes cut into the
mats were sized to the diameter of the jar creating a ﬁt, but not to the point of

creating friction between the jar and the mat.

24

Prior to testing each subject, the Qualisys software was calibrated using a
calibration structure Appendix D describes this process.

A member of the research team placed a reference jar on the table prior to
the collection of the data. Figure 13 shows the reference jars for the table
openings. The reference jars were the jars the subject was going to use for the
restricted table openings with a special reference closure, with a target adhered
to the center, in place of the closure that would be used for the opening trials.

These jars were then placed on the table in the foam board shown in Figure 12.

JarA Jar B

 

Figure 13 — Restricted opening reference Jar A and Jar B
The software was set to collect at 60 Hz (for every second of capture 60
pictures were taken by the cameras). Five cameras were set up around the table
to capture the motion. Of the ﬁve cameras, a target had to be in view of two
cameras in order for the software to identify the marker and produce x, y, z
coordinate data. An overhead schematic of the camera setup around the table is
shown in Figure 14. The schematic also identifies the subject location along with

the foam board and jar (shown with the reference target on the closure).

25

Subject

-. X<

Camera 3

1-

Camera 4

 

 

 

‘\ Front Y I.

Camera 2 Camera 5

Camera 1
Figure 14 — Overhead view of table and camera setup

The reference file was collected to identify the height and center point for
the jar on the table. This height and center of the jar was then coupled with the
opening ﬁle to determine rotation; this process will be outlined in Data
Processing. Figure 15 shows a schematic setup of the table with a jar ready for

referencing from the front.

26

 

 

 

 

V
Table Height:
91 cm

 

Figure 15 — Front view of table with restricted reference jar

Table height was ﬁxed at 91 cm for each subject. For both restricted and
unrestricted openings, the subjects were guided to the table by a member of the
research team ensuring both the safety of the subject and so that no cameras or
cables would be moved. If the cameras were moved in any manner after
calibration, the system was re - calibrated and testing continued.

Subjects were instructed to stand with their shoulders parallel to the table.
A member of the research team placed the jar on the table in the foam board and
then instructed the subject to place their hands on the jar in the opening position
they felt was comfortable. If a subject wanted to alter their hand placement from
trial to trial they were allowed to do so. Subjects were given a verbal instruction
of, “ready, set, open.” Additionally subjects were asked to rotate the closure until

the jar was open, but not to remove the closure from the jar.

27

Unrestricted Opening

Before beginning motion capture testing, subjects had been asked to pick
up a standard Jar, both A and B, see Figure 13 and asked to open each of them.
This identiﬁed their grip style, see Figure 9, and determined which of two rigs
would be used during the motion capture testing, see Figures 16. Figure 16
shows the pole rigs and side rigs for both Jar A and Jar B set up for the reference
ﬁle. When the pole and side rigs were ready for motion capture with a subject

the closure with the target was replaced with an unused closure without a target.

 

Figure 16 - Unrestricted open rigs for bottom and side grips

28

The jars were placed on the table in the orientation shown in Figure 16.
The foam boards were not used during the unrestrained openings. The
reference ﬁle was used to collect a center axis point of jar. This was then
coupled with the opening ﬁle to have a point in order to determine rotation about.
A sample of how the of the pole rig and side rig appear in the Qualisys software

is shown in Figure 17.

' C'osure r-land Index Side Rig Pole Rig

Closure Hand index

\N.’

*7 -"
’2’" ,4“ \
Rgl‘:8:2e ofJar JarHanczlndew

i
“it \ Cit
LeftSlde ofJar
Pc'e Rig Top

Po 9 Rig Bottom

JarHand Index

 

Figure 17 — Qualisys screen shot of side rig and pole rig opening
Subjects were asked to stand with shoulders parallel to the table and hold
the jar in a position that was comfortable to them. For the unrestrained opening,
subjects were asked to pick up the jar and hold it as they would prior to opening
without opening it. Subjects were also instructed to try to hold the jar closure in
place once opening had occurred, see Figure 18. Once the jar was held at a
comfortable position above the table, the subjects were given a verbal instruction

of, “ready, set, open.”

29

 

 

Figure 18 — Subject opening side rig (left) and pole rig (right)

Data Transfer

The total capture time was 4 seconds for each opening. The capture time
of 4 seconds was based on knowledge gained in pilot studies of the time needed
to successfully open the jars in an injury free population. This coincides with the
study by Fowler and Nicol (1999) in that their range of time required for opening
a jar was 0.6 — 3.7 seconds [6]. Therefore, each opening trial provided 240 data
points (60 Hz x 4 seconds) for calculating rotation. After each opening the trial
was reviewed in the Qualisys software to make sure it did not need to be
recaptured. During the review of each trial several tasks were preformed. First
the targets were labeled, see Figure 10, which included; left base (target A), left
pinky (target C), left index (target B), right base (target A), right pinky (target C),
right index (target 8), right reference for the target on the shoulder along with the
jar targets for the free openings only. The next step was to remove any false
targets (usually reﬂections) that appeared during the opening. If a target was lost
for an extended period usually 10 frames or more the opening trial was

recaptured using a new closure. If the trial was considered to be good the ﬁle

30

was saved as a .QTM and a .TSV. After completion of all 12 opening trials the
.TSV ﬁles were then opened in a word pad one at a time and from the word pad
they were copied into Microsoft Excel.

Data Processing

Table Openings

For each trial, two data sets were utilized (the table reference and the
open ing ﬁle) to calculate the rotation angle. An Excel ﬁle was created to be used
with all table openings. First the reference ﬁle was copied from the word pad to a
worksheet in the table opening Excel ﬁle. The sheet was used to calculate the
ave rage x, y, and z coordinate of the center of the closure. Target coordinates
were calculated at the center of each retro — reﬂective marker by the Qualisys
SOﬁ'Nvare. In order to identify the top surface of the closure, 10 mm was removed
from the average 2 coordinate, placing the z coordinate on the surface of the
0'08 ure. Using the reference target from the surface of the closure a second
POint was calculated at the middle of the jar half way between the closure point
and the bottom of the jar. This point was calculated by subtracted one half of the

meas ured jar height from the z coordinate value, see Figure 19.

31

Table Opening
Closre 2 (Central) Axis

'2

  

 

 

 

     

Center of Jar
Along Vertical
Axis
Figure 19 — Central rotation axis for restricted openings

While multiple targets were collected as part of each motion capture, only
the targets on the index ﬁngers were used, in addition to the targets used from
the reference ﬁle, for calculating the rotation of the index ﬁnger around the center
0f the jar in the restricted orientation
Rotation of the closure was computed by evaluating the movement of the

index ﬁnger about the center of the closure in the x y plane. Rotation of the hand

on the jar was computed about the mid — point of the jar in the x y plane.

 

 

 

 

 

 

 

 

 

 

 

 

 

h A
YI Original Vectors Y Vector Subtraction
—> —> —>
B -A = 8’
(x1, Vi) (X2. y2) (X2. y2)
-—a>
A —>
B 1‘ 2‘
x7 x
WI
Y Magnitude A x A
. = X1
\I 2 2 _, IB l
Ax + Ay = I B" A A
—.Y— = Y1
I8 I
3’ (AXAY)
“5”“ Y 3;
I I X
A x x 1

Figure 20 - Vector analysis of targets
l n the ﬁrst graph of Figure 20, top left graph, coordinate point (x1,y1)
r epr ese nts the closure center or the jar center about which coordinate point

(xZ’YZ ) , the target on the corresponding index ﬁnger is rotating. In the second

2‘) —> —+
graph A has been subtracted from B resulting in B' with coordinate

POint(Ax,—,Ay). This was done by USi093

x2 —x1=Ax
y2-y1=Ay

33

Next the magnitude of B) was calculated using the Pythagorean
Theo rem, shown in graph 3 of Figure 20. Using the magnitude along with
(Ax, Ay) a unit vector was calculated, this is shown in graph 4 of Figure 20. This
process was repeated across all 240 opening frames of each capture resulting in
a unit vector at each frame with coordinates:

551 , £2 , f3 ,..., £240
ilafzaﬁr-sz

_)
Next the change in B' from frame 1 to frame 240 was found using radians, which

were then converted to degrees. The following equation (note that R was

chosen to represent radians) shows this calculation:

0

 

cos-181 ° £240 + )71 ' .9240) = R240radians X n

This calculation was equal to the angle in degrees between frame 1 and frame
240. Therefore, the degrees value is the rotation of the index ﬁnger about the
cente r of the closure (or jar). This equation was then applied to compute the

angle 5 h degrees between frame 1 and every other capture frame (2 thru 239)

9IVi"9 the degrees rotation at all 240 frames, such as:

O

 

cos—1031 - £2 + 91 o 522) = Rzradiansx
7r

0

 

cos-10?] 0 £3 + )7] 0 )73) = R3radiansx 18
7t

34

O

 

COS—1021 ' £239 + )71 ' 51239) = R239radiansx180

The degrees calculated at every frame were then plotted in a scatter plot,
see Figure 23 and Determination of Degrees Rotation below.

Unrestricted Openings

As with the restricted orientation (table opening), for the unrestricted to be

processed in Excel, two ﬁles had to be imported. For unrestricted openings the
table reference ﬁle along with an opening ﬁle were needed to calculate the
rotation angle. Unrestricted openings were calculated in 3 — dimensions.

Pole Rig Openings

Determination of rotation for the pole rig openings began by using the
reference ﬁle to calculate the jar length. This was accomplished using two of the

three targets on the pole rig jar, see Figure 21.

Pole Rig Opening Reference File

 

(x3, y3, 23)
A B c /
. H
(x1, y1, 21) (x2, y2, 22)

Figure 21 — Central rotation axis for pole rig openings
The closure target, (x1,y1,zl) in Figure 21, and the top pole target,
9N2 5 22) in Figure 21, were used to calculate the distance from B to the
surface of the closure from the reference ﬁle. The following equations were used

When analyzing the pole rig data, prior to computing the following calculations it
35

was conﬁrmed that the axis formed through BC was straight and went through

point A.
STEP 1: Creating a vector between points A and B — directed along the center

axis of the jar and closure.

The same calculations used in the table restricted method were performed
for the pole rig, but in three-dimensions. The following calculations are

perform ed on the jar while it is stationary:

—) A A -
A =x1i +y1j+zlk

‘9 A A x
B =x2i +y2j+22k

The vectors were then subtracted to ﬁnd a new vector

-> ‘9 —-> A A A .
AB: B- A =(x2 —x1)i+(y2 -y1)j+(22—21)k written

—)
as AB = Axab? + Ayab] + Azabl? . Next the magnitude was calculated as,

 

—)
AB = \/Axab2 + Ayab2 + Azab2 from this magnitude 10 mm was subtracted to

 

 

accoU n t for the fact that the center of the marker was located 10 mm above the

closure -
STEP 2: Creation of a unit vector along the center axis of the jar.

l o compute an axis using the pole rig, the top and bottom targets for the

poles Were subtracted and the magnitude found as shown.

C=(x3,y3a23) B =(x2ay2a22)

36

= (X3 -X2)i +(y3 -y2)]' +(23 -22)/2 = AXbci + Abe +42bc'E

 

.__)
BC = \IAXch +Aybc2 + AZbc2

 

 

. —> Ax A A2

Next a unit vector for BC was calculated. JC— = 56,36;ch— = 521,6 ,—————bc = Ebc
—> —> —>
BC BC BC

 

 

 

 

 

 

STEP 3: Locating closure surface and jar midpoint in free space.
To track the position of the jar and closure in 3 — dimensional space, the
jar axis (developed in Step 2) was used in conjunction with the distance from the

closure to point B and the distance from the middle of the jar (1/2 height) to point

B.
From point B = (x2,y2,zz) the x-value ibc was multiplied by the
_ —>
magn ltude AB , and then added to the respective component for all three

 

 

components (x. Y and 2)-

 

 

 

 

A —->

x2 + (xbc ’ AB ) = xclosure
A -—)

YZ + (ybc ' AB ) = yclosure
A —>

22 + (Zbc ' AB) = chosure

 

 

37

This results in the center point of the top of the closure located at
(x610 SureJ’closurewzclosure) . The process was repeated for half of the

magnitude in order to ﬁnd the coordinate of the point in the center of the jar,

termed mid jar, about which the rotation or the jar index ﬁnger would be

 

 

 

 

calculated.
__)
x2 +<£bc - % AB) = xmid
9
y2 + (the - y, AB) = ymid
._)
22 + (Ebc - V2 AB) = 2m

 

 

STEP 4: Both the tOp of the closure as well as the jar midpoint were computed

for each frame of data.

The rotation of each hand was computed around the center of the closure
(closu re hand) and the midpoint of the jar (jar hand). Now the rotation for each
hand Was calculated following the same steps as performed for the table
r"38tricted method shown previously, except in the three-dimensional coordinates.
The ind ex ﬁnger of the hand on the top of the jar was given

Coord i h ates (xindexl , Yindexl azindexl) . The difference between this point and the

t°p ce "I ‘ter of the jar was calculated:

xindexl ’xclosureJ = Ax] . Yindexl “J’closureJ = Ayl ' Zindexl _ chosure,1 = A21

._)

(Axl , Ayl , A21 ) gives vector AB,
38

 

 

 

__.)
Magnitude AB, =\/Ax12 + Ayl2 + A212
. . Ax A A A A2 A
Unit vector coordinates ——1— = x1, ,——y—]— = ylr ,—1— = 21,
—> —)
AB, AB, AB,

 

 

 

 

 

 

Using the midpoint of the jar (xmid,1aYmid,lazmid,1) and the index ﬁnger of the

jar, a u nit vector was computed from the jar center to the index ﬁnger of jar hand.
The sarne process was used for the hand on the closure.
STEP 5

Steps 3 and 4 were repeated for each camera frame 1 to frame 240, since
the Ia r and hand were moving, each frame had new index ﬁnger and new jar
coord inates that had to be computed.
STEP 6: Computing degrees of rotation.

Next the degree of rotation for each frame was found using:

c
COS

 

~—1 A - - .. - - . 180
(xv ° mm + y1r - y240r + Z1r ' 2240r) = R240radmnsx = angle

The a h gle between the ﬁrst frame and every consecutive frame was computed.

~1 -\ A A A A A 0
cos (Axlr .xzr +y1r oyzr + zlr O 22’.) = Rzradiansx

 

75

O

 

cos"1 e - A A A A _ 1
(Jrlr ° x24or + y1r ° y240r + er ° 224or) = R240radtans><

39

The degrees calculated at every frame were then plotted in a scatter plot, see
Figure 23 and Determination of Degrees Rotation below.

Side Rig

Side Rig Opening Reference File

Closure 2 (Central) Axis

a ix4, y4, Z4)

 
  
 

Top
Rotation
Point

  

Bottom
Rotation
Point

 

Figure 22 —- Side rig opening reference ﬁle
STEP 1: Computing average jar target heights.

The side rig jar had three target locations as shown in Figure 22. The ﬁrst
calculation was to ﬁnd the average of the z coordinates at each frame when the
jar was stationary on the table. Let z_, _ where the ﬁrst subscript represents the
left (1), middle (2), right (3), or top (4) point and the second subscript is the frame

number. For example 21,1 represents the z coordinate of the left target at frame
1.

21,1 + 22,] + 23,1

Frame 1: = zavg’l continue for each frame up to

21,240 + 22,240 + 23,240
3

Frame 240:

 

= Zavg,240
Next the 240 frames just found were averaged:

40

Zavg,1 + zavg,2 "'+Zavg,240

240 = Average vertical location of the three target

 

points on the jar. Still working with the jar stationary, the middle target coordinate
(22) on the jar was subtracted from the top coordinate (Z4) on the closure for
each frame. Next the average distance to the closure from all 240 frames was
found.

24,1 - 22,1 = A21. 24,2 - 22,2 = A22. - - - £4,240 - 22,240 = A2240

A214» A22 + ...+ A2240
240

 

= Zvertavg

(The average distance from the closure to the three points)
STEP 2: Creating a local coordinate system that rides with the jar.

The next steps were calculated when the jar was moving. First the left
target coordinates were subtracted from the right target coordinates for all 240

_.)
frames, see Figure 22. This produced vectorLR.

 

 

 

 

x3,1 - 161,1 = m1 x3,240 - x1,240 = AJC240
y3,1 - y1,1 = An - - - y3,240 - y1,240 = Ay240
23,1 - 21,1 = A21 23,240 - 21,240 = A2240
__)
The magnitude for each frame was calculated: \[Axlz + Ayl2 + A212 = LR1
2 —)
\/A"240 + Ay 240 + AZ 240 LR240

 

 

Next the unit vector was calculated for each frame:

41

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

An Ayl AZ1 A36240 Ay240 A2240
—) ’ —) ’ —> —> ’ —> ’ —>
LR] LR1 LR1 LR] LR240 LR240
\ J \
($107131) (£240,9240iz40)

The next step was to calculate a point in the center of the jar at the level of
the targets on the outside of the jar. Calculations for the center of the jar for each
frame:

_)
LR1

—> —-)
LR2 LR240

1) Find an average magnitude =
240

+ +...+

 

 

 

 

 

 

 

mm,g

 

 

2) Calculate coordinates for the center (center of diameter

between L and R target) of the jar at the level of the targets:

a

LRavg ) = x center,1

xleft,1 #2021-

 

 

_.)
LRavg ) = xcenter,2

xleft,2 + yz (£2 °

 

 

A —)
x1eﬁ,240 + % (x240 ' LRavg ) = xcemer,240

 

 

(Repeat for y and z coordinates)

42

._)
Again working with each frame a vector CL (center to left) was created by

subtracting the left target minus the center point computed above.

xleft,l " xcenter,1 = A"left,1 xleft,240 _ xcenter,240 = A”left,240
yleft,l - ycemer,1 = AYIeftJ yleﬁ,24o - J’center,240 = AYIeft,240
Zleft,l - zcenter,l = Azleft,1 Zleft,240 - Zcenter,240 = AZleft,240

Repeat this process for the subtraction of the right target from the center target to

get: (Axright’l aAJ’rightJ aAZrightJ) for each frame of data. Next the magnitude for

each frame was calculated. A cross product was computed between CL and CR

—> —-> —+
CRxCL= Z

This resulted in an 2 axis that ran through the center of the jar directed downward

toward the base of the jar.

Find the unit vector of the cross:
-:) A A A
Z = (xerOSSaJ’crOSSaZcross)
Repeat Step 5 of Pole Rig Method for each frame (2 through 240)

STEP 3: Bottom of jar (center point, moved down 70 mm) calculation.

Frame 1: x11 + £0.05“ . 70mm

Frame 240: x2240 + 500553240 0 70mm
Repeat for the y and z coordinates. This point was then used for
calculating the rotation of the jar hand index ﬁnger about the jar, see Figure 22

STEP 4: Rotation point (center point, moved up 35 mm) calculation of closure

43

Frame 1: x11 ‘fcrossJ 035mm

Frame 2402 x2340 — fcross,240 0 35mm

Repeat for the y and z coordinates. This point was then used for
calculation the rotation of the closure hand index ﬁnger about the jar, see Figure
22.

Next the process for the Pole Rig Method Steps 4 through 6 was used to
calculate the degrees for every frame. All 240 frames where graphed using the
corresponding degrees, see Figure 23 and Determination of Degrees Rotation
below.

Determination of Degrees Rotation

In order to determine the degree of rotation supplied by each hand at the
determined point of opening, see Appendix C, the calculated degrees for each
hand at each frame collected were graphed. In order to determine the motion
that correlated with package opening, graphs consisting of the entire set of data
points (240 points) were created. Figure 23, a representative sample provided
by one subject for a single jar, depicts a restricted opening (table bound) for jar
B. In Figure 23 there is an additional data point at frame 1. This point represents
the angle calculated between frame 1 and frame 240. Researchers identiﬁed
opening through the use of the slope of the data, with the area of greater slope

representing the period of opening.

44

 

RH Jar 014 LT

 

 

m 35

‘1’ ' '32:. «MA... .

(D w« as“; .. :7... 19

O)

0 761858

.E

C

.9

g o 50 100 150 200 250 300
Frame

 

LH Jar 014 LT

 

 

7 .
(D a
(D 6 :-
0 i
1.. 5 . .............
8’ 4 76-5.02 _____
0 3 _
.E "fr” “mm“,wﬁwg
C 2 . in ”W .. e j .3
*6 o
H
0
CK

Frame

 

 

Figure ”23‘"; Graphs usedror‘ calculating rctati'on supplied" by each“ hand
Appendix C shows how the degrees of rotation needed to remove the
closure form the jar were calculated for both Jar A and Jar B. Jar A was
determined to require an average of 30 degrees of rotation for an opening to
occur while Jar B was calculated to need an average of 25 degrees. Using
Figure 23 where subject 014 is opening Jar B there should be one data point

where the left hand (LH) and the right hand (RH) together account for a degree of

45

opening near degrees. At data point 76 the right hand on the jar rotates 18.58
degrees and the left hand on the closure rotates 5.02 degrees for a total
combined rotation of 23.06 degrees. For all of the openings collected from the
subjects the data point closest to the calculated average, in the area of the graph
where the increase was seen, was determined to be the point at which the
closure was removed from the jar. These degrees were then calculated into a
percentage of the opening degree. In continuing the example demonstrated by
Figure 22 the right hand on the jar supplied 18.58 degrees of the total 23.60
degrees (18.58 + 5.02) for an opening percentage of 78.73 percent
(18.58/23.60). The left hand (in this case the closure hand) supplied 5.02
degrees of the total 23.60 (18.58 + 5.02) degrees for an opening percentage of
21.27 percent (5.02/23.60).

Data was categorized in an attribute fashion by grouping the openings into
three categories; closure hand only, jar hand only and both hands. The
categories were determined by computing percentages from the angles
calculated from the Qualisys data as outlined above. Parameters were set by the
researchers as to what would qualify, based on percent, as one hand supplying
the rotation necessary to open the jar, either closure hand or jar hand, or if both
hands were involved in supplying the rotation need to open the jar. In order to
determine whether one hand or both hands were involved in the opening all of
the opening percents were ﬁrst graphed to see if a trend would emerge showing
a clear break between one hand, either closure or jar hand, or both hands being

used for opening, see Figure 30.

46

Chapter 4

RESULTS AND DISCUSSION

Subject Information

Researchers collected data on 25 subjects, all of whom attended Michigan
State University. Of the 25 subjects, the ﬁrst twelve had to be collected two
times due to a collection error, and all subjects were compensated for both
testing sessions. This error was created by incorrect calibration of the motion
capture system; see Appendix D for correct calibration. Of the ﬁrst twelve
subjects only nine were available for the second session. As such, only 22
subjects had usable data collected.

One of the factors of interest in the data analysis was the effect of
laterality on motion. A single participant, subject number 024, clasped their
hands with their thumbs in a side by side orientation. As a result, researchers
could not code the laterality for this subject, whose results were subsequently
eliminated.

Of the usable 21 subjects, 14 were male and 7 female. The average age
of these subjects was 24 1 3.1 years, and subjects ranged in age from 18 to 30.

Twenty of the 21 indicated that they were right handed, and one subject
indicated himself to be left handed. Table 2 gives an overview of the preferred
dexterity and information about which hand was placed on the closure hand

placement indicated/used by subject and gender.

47

 

 

 

 

 

 

 

 

 

 

 

 

 

Overview of subjects included in analysis
Female Male Total

Total Number of Subjects 7 14 21
Indicated Right Handed 7 13 20
Indicated Left Handed 0 1* 1
Right Hand on Closure 6 10 16
Left Hand on Closure 1 3 4
Alternated Closure Hand 0 1* 1
*Left handed male subject alternated closure hand placement

 

 

Table 2 — Overview of subjects participating in study

Data was collected on which hand each subject placed on the jar closure
for each opening trial during testing. 0f the 21 subjects tested, 16 placed their
right hand on the closure, or 76.2% of the subjects. All 16 of the subjects placing
their right hand on the closure indicated that they were right-hand dominant. Of
these, 10 were male and 6 were female. This ﬁnding is not consistent with that
of Voorbij and Steenbekkers (2002) who found that 511 or 534 or roughly 96
percent of subjects that are right handed place their left hand on the closure [3].
Of the remaining subjects, 4 placed their left hand (non - dominant) on the
closure (19.1% of the subjects); of these 3 were male and 1 was female. One
subject, a male that had indicated himself to be left-handed, alternated which

hand was placed on the closure during opening (4.7%).

48

Hand Size

Anthropometric data (size) was collected for all test subjects. This data

included: Hand Breadth, Wrist — Index Finger Length and Hand Length, see

Figure 3 in Chapter 3 and Table 3.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table of Average Size of the Hand (Anthropometric Data)
Population
Average Standard
(number of Measurement (mm) Dev.
people)
Hand Breadth Left 96.63 8.75
Wrist - Index Finger Length Left 188.15 12.87
Hand Length Left 202.83 14.55
Overall (21) .
Hand Breadth Right 96.74 8.23
Wrist - Index Finger Length Right 189.85 12.71
Hand Length Right 203.07 13.74
Hand Breadth Left 101.10 5.34
Wrist - Index Finger Length Left 192.79 9.68
Hand Length Left 209.10 10.23
Male (14) .
Hand Breadth Right 100.84 5.00
Wrist - Index Finger Length Right 195.02 9.14
Hand Length Right 208.91 9.98
Hand Breadth Left 87.71 7.35
Wrist - Index Finger Length Left 178.88 14.08
Hand Length Left 190.30 14.28
Female (7) .
Hand Breadth Right 88.54 7.35
Wrist - Index Finger Length Right 179.53 13.09
Hand Length Right 191.39 13.23

 

 

Table 3 - Anthropometric data
Hand size was then categorized based on size in mm in accordance with the

data collected by a division of the United States Department of Commerce called
49

the National Technical lnforrnation Service (NTIS) [13]. Figure 24 shows the
current study compared to the NTIS data for hand breadth. NTIS data recorded
the average measurements so the current study right hand and left hand

measurements are compared to the averages of the NTIS data.

 

 

Hand Breadth
Current Study Compared to NTIS Data

 

110.00 i

E 100.00
4: 90.00 <
a ’:
8 80.00, lCurrentStudy
m 70.00 "NTIS Data 1939
(Ave. Both Hands)

60.00 . 1

'Right Left Right Left

Male Female

 

Figure 24 — Current study hand breadth vs. NTIS data

In the current study, hand breadth measurements for both male and
female, left and right hand, are larger than the average measurements from the
NTIS data. This could be related to the very small sample that came from a
single population presented in our study, which had 14 males and 7 females.
The NTIS data consisted of 1,774 males and 2,208 females. Another possibility
could be the 20 year gap in data collection. If the current study left and right
hand data is averaged then using the percentiles compiled by the NTIS [13] the
average subject in the current study had a hand breadth in the 99th percentile for

males and the 98th percentile for females.

50

Figure 25 shows the current study compared to the NTIS data for wrist —
index ﬁnger length. Again, the NTIS data recorded the average measurements
so the current study right hand and left hand measurements are compared to the

averages of the NTIS data [13].

 

Wrist- Index Finger Length
Current Study Compared to NTIS Data

 

200.00

190.00
E . ,
E 180.00
E; 17000 ICurrent Study
‘J 160.00 . NTIS Data 1989

- (Ave. Both Hands)
150.00 <

Right Left Right Left

Male Female

 

Figure 25 — Current study wrist 4 ﬁnger length vs. NTIS data

Wrist — index ﬁnger length data again demonstrates that the current study
subjects had measurements above the average data collected by the NTIS [13].
If the current study left and right hand data is averaged then using the percentiles
compiled by the NTIS the average subject in the current study had a wrist — index
ﬁnger length in the 90th percentile for males and the 85th percentile for females
[1 3].

Figure 26 shows the current study compared to the NTIS data for hand

length. Again the NTIS data recorded the average measurements so the current

51

study right hand and left hand measurements are compared to the averages of

the NTIS data.

Hand Length

   

210.00 -
200.00

E I
E 190.00
jg) 180.00 lCurrent Study l
7' 170.00 -‘ NTIS Data1989 3:
(Ave. Both Hands) ;

160.00 - ‘

Right Left Right Left

Male Female

 

Figure 26 - Current study hand length vs. NTIS data

Measurements of hand length were found to be greater than the average
measurements collected for the NTIS study [13]. If the current study left and
right hand data is averaged then using the percentiles compiled by the NTIS the
average subject in the current study had a hand length in the 90th percentile for
males and the 85th percentile for females. The results for anthropometrics of the
current study compared to the data from the NTIS show that, on average, the
subjects used in the current study had relatively large hands [13]. Researchers

examined the data for correlations on the collected anthropometric data.

52

Analysis of Strength Data
Strength data was also collected for each subject. The hand strength
measurements collected were grip strength and bilateral palm to palm squeeze

strength. Table 4 shows the average results.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table of Strength Measurements
. Bilateral Palm to
Population Gr'p Strength (lbs) Palm Squeeze (psi)
(number of
people) . Standard Standard . Standard
Right Dev. Left Dev. Right/Left Dev.
Overall (20) 37.69 13.62 36.60 13.06 7.04 5.80
Men (14) 42.10 14.63 41.46 13.29 8.88 6.16
Women (7) 28.88 4.27 26.87 4.50 3.36 2.49
Tab e 4 — Table of strength measurements

 

The results for the right and left hands within gender were quite similar for
the grip strength. This ﬁnding could be related to the fact that all subjects were
free of injuries to the hands, wrist, arms and shoulders. However, this ﬁnding
suggests that with relatively equal strength measurements in both hands subjects
might not favor the strength in one hand over the other for opening purposes.

Figure 27 is a graphical representation of the strength measurements.

53

 

Strength Measurements
Grip and Bilateral Palm to Palm

 

 

 

 

 

 

 

I Overall
(20)

 

- Men (14)

 

Measurement (lbs or psi)
N
O
8

I Women (7)

   

   

Standard Leﬂ
D

8V.

Grip Strength (lbs)

 

 

 

 

Figure 27 — Strength measurements
Laterality
Laterality was of special interest for the research team. Table 5 shows an
overview of the laterality results observed in the 21 subjects.

Results Gender and

Gender Left

Male 13 10
Female 7 4
Total 20 14

 

Table 6 shows the breakdown of the four laterality coding possibilities with
the statistical analysis that examines the relationship between hand clasping and

arm folding.

54

 

Table of Laterality (Hand) by Laterality (Arm)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Laterality (Arm)
Laterality (Hand) Right Left
Right 3 (RR) 6 (LR)
Left 4 (RL) 8 (LL)
Fisher's Exact Test
Cell (1, 1) Frequency (F) 3
Left - sided Pr <= F 0.6811
Right - sided Pr > = F 0.6765
Table Probability (P) 0.3576
Two - sided Pr <= P 1.0000
Sample Size 21

 

 

 

 

Table 6 - Laterality hand by laterality arm statistics
Based on the results of a Fisher’s Exact Test, see Table 6, the data
suggests no evidence of an association between hand clasp and arm folding (P =
0.35). This indicates that the two measures of laterality are independent of each
other. This is consistent with other researcher’s ﬁndings Mohr et. al. (2003,
2006) [8, 9] that indicate that there is close to equal distribution of subjects
between the four possible codes (LL, LR, RR, RL). Distribution of the sample

population based on laterality is expressed in Figure 28.

55

Laterality Codes
Arm Fold/Hand Clasp

 

 

 

  

28.6%

   

 

 

  

19.1%

 

Number of Subjects

 

LL LR RR RL
Figure 28 — Number of subjects representing each of the four laterality codes

Opening Data

Simple analysis of the opening data shows that, across all openings
(restricted or unrestricted), on average, whichever hand was on the closure had
an average opening rotation percentage greater than that of the jar hand; see
Table 7 and Figure 29. However, not every subject had a greater percentage of
rotation for the closure hand, 80 of the 252 openings had a greater percentage
rotation of for the jar hand. For the unrestricted openings the subjects placed
their hands in the same positions. The left handed male subject alternated his

hand placements on the closure for both restricted and unrestricted openings.

56

 

Overview of Rotation in Percent and Degrees All Openings N = 252

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Average 3‘30“?“ Minimum Maximum
DeVIation
Rotation Closure
Hand % 59.60 25.34 0.20 99.79
Rotation Closure
Hand Degrees 16.07 7.23 0.05 30.78
Rotation Jar Hand % 40.40 25.34 0.21 99.80
Rotation Jar Hand
Degrees 11.25 9.08 0.05 97.07
Unrestricted Openings N = 126
Standard . . .
Average Deviation Minimum MaXImum
Rotation Closure
Hand % 54.20 20.46 3.15 94.38
Rotation Closure
Hand Degrees 14.31 5.81 1.00 30.35
Rotation Jar Hand % 45.80 20.46 5.62 96.85
JRotation Jar Hand A
Degrees 12.73 9.59 1.41 97.07
Restricted Openings N = 126
Standard . . .
Average Deviation Minimum MaXImum
Rotation Closure
# Hand % 66.75 28.12 0.20 99.79
Rotation Closure
Hand Degrees 18.35 7.96 0.05 30.78
Rotation Jar Hand % 33.25 28.12 0.21 99.80
7 Rotation Jar Hand
Degrees 9.28 8.03 0.05 30.10

 

 

 

 

 

Table 7 -- Overview of rotation by percent and degrees

57

 

All Openings Rotation by Hand

K Ciosu’fe ham

. 59.6%

 

 

 

Jar Hand » ,,

 

 

Figure 29- Openings by rotation hand
A Pearson Correlation Coefﬁcient was applied to the opening rotation in
percent and degree for both the closure and jar hands in all openings, see Table

8.

58

 

Pearson Correlation Coefficients, N = 252

 

Prob > I r I under HO: Rho= 0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

,Rotationtl 3:33: agape, Rotation
J C‘loéure fl H Jar Inland Jar Hand
0 l and .t o
tHand A .. . .. /9’ Degrees
-, a 311 Degrees «ﬂ , ,
ROt‘ation- .. j
‘ Closure Hana; . 1.00000 0.90552 -1.00000 -0.81302 -0.10750
";, NIP/01‘ u ’:
fRotation , 1
Closgre Hand < .0001 < .0001 < .0001 0.0886
Rotation
Closure Hand 0.90552 1.00000 -0.90552 -0.56924 0.26651
Degrees
Rotation
Closure Hand < .0001 < .0001 < .0001 < .0001
Degrees
1’ l
Rﬁi‘ﬁfm 1 -1.00000 0.90552 1.00000 0.81302 0.10750
ZRﬁgi'gnﬁiaf ‘1' <.0001 <.0001 <.0001 0.0886
R°taﬁ°n Jar 0 81302 0 56924 0 81302 1 00000 0 64073
Hand Degrees ' ' ' ' ' ' '
$333333; < .0001 < .0001 < .0001 < .0001
15:33:33? 0.10750 0.26651 0.10750 0.64073 1.00000
3°taggfgf?s, 0.0886 < .0001 0.0886 < .0001

 

 

Table 8 — Pearson correlation coefﬁcients

Note the perfect correlation between percent rotation of jar hand and

percent rotation of the closure hand (Correlation =

59

—1), thus indicating that

 

modeling of one variable is a mirror complement to the modeling of the other.
Also, note the high correlation between rotation of each hand expressed in
degrees or percent (Correlation > 0.90), thus suggesting reliable representation
of one with the other.

Each subject opened 2 jars, see Table 1 and Figure 7, in both the
restricted (table) and unrestricted (free motion) conditions of test three times
each, for a total of 12 openings per subject, see Figure 11. Degrees of rotation
were analyzed using techniques described previously. Figure 30 shows all of the
opening trials for both restricted and unrestricted openings in percent degrees of
rotation for the closure hand, which, as noted above, represents a mirror of jar

hand opening.

60

Closure Hand Rotation Percentage

Closure Hand Openings

100.00 ,A., ‘ L ,-L, , 75.13,? Tia.
“ aﬁ_

    

 

 

 

 

 

 

l , 1°. :f . 7 7' u- 7' 77.: 7 7 7 N i oLettHand
4000 3,. 7AT’ I I v., A‘r—"—' * IRightHand

l. . I I l I
30.00 R—LLWJﬂ—ake—we

I O .I I I

A h 1

20.00 ﬁ I _ .
10.00 I ’ .
0.00 I. I ,_ t,”

0 50 100 150 200 250 300

0 in 90/10
pen 9 80/20

 

Figure 30 — Closure hand openings

The total opening trials, N = 252, can be broken down into restricted and

unrestriced openings, each with N = 126. Beginning with the restricted openings,

see Table 7, the trend of closure hand having a greater percentage of rotation

compared to the jar hand, on average, remains. While this ﬁnding does not

change compared to the overall average it is interesting that the closure hand is

supplying a greater percent, on average, for the restricted openings then was

seen across all 252 openings, both restrained and unrestrained.

When the restricted opening trials are graphed according to the percent of

degrees of rotation of the closure hand, there appears to be a separation

between left hand closure openings and right hand closure openings, see Figure

31.

The openings where the subject placed their left hand on the closure

61

account for twelve of the openings, three males and one female at 3 opening

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

each.
Restricted Openings Closure Hand
90.00 “a "1"" ..FL '3 "7' e
4—0 3...?! it"ﬂmlllug": 5 a 5:5 ”:1 1...; 1!” “an
5 80.00 m— _. _. ...+ -' -~ 51 ' 3
H . .- " f". a 91'? "'
(03) 70.00 i" E ii“! " H—L' 1—————~ ' lea—Ii “
0- ,.-."11 ,3 m5 ‘3 ‘21 a
g 60.00 ~——I MM” 5“
3: I
:3 50.00 s 5’ Left
(2 .3! ’Hand
'0 40.00 . I _
c .0. IRight
g 30.00 . Hand
0
9 20.00 ~+—’—‘ e; I
3 § 0 o
8 10.00 ﬁ-ﬁ—s-g
0.00 4——+—*
0 20 40 60 80 100 120
Opening 90/10
80/20

 

 

 

Figure 31 — Restricted closure hand openings

This may be explained by the fact that twenty of the twenty one subjects
declared themselves to be of right hand dexterity. The right hand exhibited
higher degrees of rotation, on average than did the left hand. In other words the
left hand was on the closure, but it appears that the dominant (right) hand
supplied a greater rotation percent, see Figure 34. Also, as noted by Voorbij and
Steenbekkers (2002) the vast majority (nearly 96 %) of right hand dominate
subjects place their left hand on the closure [3], so this trend may not apply to
most subjects and could be a result of the subject population used in this study.

For the unrestricted openings, see Table 7 and Figure 32, the trend of

closure hand having a greater percentage of rotation compared to the jar hand

62

remains. However, the closure and jar hands result in averages of rotation
percent that are more scattered than those seen in the restricted openings.

It appears that both hands tended to have a greater role in unrestricted
openings than they did when motion was restricted, making it more difﬁcult to
predict the motion applied by either hand when opening jars in an unrestricted

fashion.

 

 

if _ _. _ _
Unrestricted Openings Closure Hand

100.00

Left
0 Hand
a Right

Hand

Closure Hand Rotatlon Percent

 

0 20 40 60 80 100 120

0 i _ 90/10
pen "9 80/20

 

 

 

 

Figure 32 — Unrestricted closure hand openings

The degree of rotation of the closure hand for all openings was graphed in

an attempt to identify whether or not discrete differences existed in how people
used their hands (i.e. some a majority of the closure hand, others a majority of
the jar hand and some both hands). However, no clear break was seen in the

graphical data for the closure hand. Therefore, arbitrary values were set for the

break in percent rotation for determination of one hand or both hands used for

63

opening. Two arbitrary value boundaries were set for statistical analysis; 90 — 10
and 80 — 20, meaning that if either hand supplied 90 or greater, or 80 or greater
of the percent rotation it was classiﬁed as a one hand opening, jar or closure, see
Figures 30, 31 and 32 for the boundaries. Openings where one hand was not
determined to have supplied 90 percent or greater or 80 percent or greater was
coded as a two-handed opening (i.e. both closure and jar hands provided
motion). There was a marginal disagreement between the two motion
classiﬁcations (90_10 or 80_20), thus emphasizing the arbitrary nature of the

boundaries, see Table 9.

 

Table of Closure Hand 90_10 and 80_20 Openings All Openings

 

 

 

 

80_20 90_1 0
Left 7 2
Right 84 34
Both 161 216

 

 

 

 

 

Table 9 - Total openings for 80_20 and 90_10 groupings

Rig Selection

The twenty-one subjects were asked to open 2 jars, see Figure 7 and
Table 1, in two different orientations (restrained and unrestrained) a total of three
times in each orientation; this totaled 12 observations per subject, or 252
observations for the entire experiment. The unrestricted openings required
different experimental equipment, or “rigs.” Two rigs were used to accommodate
the differing grip styles employed by subjects, see Figure 16. A Chi - square test

was run on rig type by gender with a sample size of 126, which account for the

64

total number of openings performed by the 21 included subjects in the

unrestricted positions (2 jars x 3 replicates x 21 subjects), see Table 10.

 

Table of Rig Type by Gender

 

 

 

 

 

Gender Rig Type Total
Pole Side
Male 72 12 84
Female 30 12 42
Total 102 24 126

 

 

 

 
   

W Statistics for Table of Rig Type by Gender

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Statistic DF Value Prob.
Chi - Square 1 3.7059 0.0542
Likelihood Ratio Chi - Square 1 3.5478 0.0596
Continuity Ad. Chi - Square 1 2.8373 0.0921
Mantel - Haenszel Chi - Square 1 3.6765 0.0552
Phi Coefﬁcient 0.1715
Contingency Coefﬁcient 0.1690
Cramer's V 0.1715
Fisher's Exact Test
Cell (1,1) Frequency (F) 72
Left - sided Pr <= F 0.9833
Right - sided Pr >= F 0.0481
Table Probability (P) 0.0314
Two - sided Pr <=P 0.0901
Sample Size 126
Table 10 - Rig type by gender statistics for unrestricted open-ngs

 

65

Based on the Chi - Square analysis, see Table 10, male subjects were
found to be more likely to use the pole rig than female subjects (P<0.05). This
has ramiﬁcations for the grip style that they used. Male subjects were more likely
to grip jars by the side than females (P<0.05), who more frequently gripped at the
bottom of the jar than their male counterparts, see Table 10. Table 11 further
investigates the use of the side and bottom grips, indicating that both male and
female subjects tend to grip the larger jar from the bottom more frequently than
the smaller jar.

Effect of Rig Type and Jar Size on % Rotation Degrees by Jar Hand

Effects of grip style were further examined using a general linear mixed
model, which was ﬁt to the continuous response variable “percentage of rotation
degrees by the jar hand” using the MIXED procedure of SAS (version 9.1, SAS
Institute, Cary, NC). Data were collected in a RCB design, such that subject was
considered the blocking factor for assigning a 2x3 jar size by rig type treatment
combination. Table 11 shows the breakdown of opening positions (table, pole rig
and side rig) by jar size and gender. Pole rig openings accounted for 80.9% of
the unrestricted openings, in other words, few people gripped the bottom of the

jar during the opening process.

66

 

Table of Jar Size by Opening Position, Gender and Closure Hand

 

Male Openings
(Closure Hand)

6' g _, . . ,I‘
, F,emale‘dpertings 1"
(SlosEig-J Hand);

Total Openings
(Closure Hand)

 

 

 

 

 

 

 

 

 

 

Right Left Right Left Right Left Total
Hand Hand Total Hand Hand Total Hand Hand Both
Jar A
Table 30 12 42 18 3 21 48 15 63
Jar A
Pole Rig 30 9 39 15 3 18 45 12 57
Jar A
Side Rig 3 0 3 3 0 3 6 0 6
' .r F . .1‘ s_ ‘, _
yfﬂarAa" OF 6; w”- I“ ’f ,
tr'jlot‘al ,’ 63 21 84 ' .436 6 . 42 3 99 27 126
OPEN"? 6 a a. 6‘
Jar B
Table 32 10 42 18 3 21 50 13 63
Jar B
Pole Rig 24 9 33 9 3 12 33 12 45
Jar B
Side Rig 9 0 9 9 0 9 18 0 18
{if BTJ‘ l9!“ “33¢”! 36,
"ﬁlytqfaly 65 19 84 .136 6 , 3‘4}; 101 25 126
Q .n’irig.‘ .

 

 

 

 

 

 

 

 

 

 

 

 

 

67

Table 11 — Table of jar size by 6pening posit-ion, genaer and closure hand

 

By design, the base statistical model included the categorical ﬁxed effects
of jar size (Jar A and Jar B), rig type (table, pole and side rig) and their 2-way
interaction. Effects considered random included: subject, subject by rig type and
subject by jar size. These random effects were intended to recognize subject as
the blocking factor and the appropriate degrees of freedom for each of the
treatment factors, correspondingly. Also, a random effect of subject by jar size
by rig type was ﬁtted to account for technical replications (sub sampling) within
the design (each subject in each rig type by jar size combination was measured 3
times).

Potential additional explanatory variables of interest were recorded at the
subject level, including gender, dexterity, laterality (both by arm cross and hand
clasp) and a total of 9 morphometric continuous variables. Closure hand used in
each jar-opening event was also recorded. In order to decide between these
potential explanatory variables, stepwise model selection based on Bayesian
Information Criteria (BIC) was used to select a model of best ﬁt. The BIC-based
model of best ﬁt included, in addition to the base model described above, the
effects of gender, hand breadth RIGHT, their 2-way interaction and hand length
RIGHT. Degrees of freedom were computed using Kenward-Roger method, as
implemented in SAS. Model assumptions were checked using standardized
residuals and were considered to be appropriately met. Least square means and
standard errors are provided, see Table 12. Post-hoc pairwise comparisons
were performed using Bonferroni’s adjustment to avoid inﬂation of type I error

rate. It should be noted that the lack of observations for the combination of the

68

side rig (rig type 3) and the left hand on the closure (closure hand 2), this is
shown in bold and italics in Table 11, prevented simultaneous ﬁtting of the two
(rig type and closure hand) as explanatory variables in a given model.

Due to these limitations in the data structure, it was decided to continue

with a model without closure hand as an explanatory variable.

69

 

Solution for Fixed Effects

 

Rig

Standard

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Effect Gender Type Jar Size Estimate Error DF Value Pr> ltI
Intercept -54.3064 67.5064 17 -0.8 0.4322
JarSize 1(JarA) 1.7748 10.8928 44.3 0.16 0.8713
Jar Size 2 (Jar B) 0

Rig Type (Ta1ble) -12.2533 12.5283 69.2 -0.98 0.3315
Rig Type (Pile) 1.4902 12.7554 72 0.12 0.9073
. 3

R'9 Type (Side) 0

Rig Type* 1 _ _

Jar Size (Table) 5.387 11.4502 43.3 0.47 0.6404
Rig Type* 1 O

Jar Size (Table)

Rig Type* 2

JarSize (Pole) -13.1516 11.6691 43.1 -1.13 0.266
Rig Type* 2 0

Jar Size (Pole)

Rig Type* 3 0

Jar Size (Side)

Rig Type* 3 0

Jar Size (Side)

Gender 1(M) 225.15 91.0139 15.7 2.47 0.0252.
Gender 2(F) 0

Hand

Breadth -0.7418 0.7991 16.3 -0.93 0.3667
Right

Hand

33;??? 1(M) -2.1779 0.9613 15.9 -2.27 0.0378
Gender

Hand

Breadth

Right* 2(F) 0

Gender

Hand

Length 0.8413 0.3342 16.5 2.52 0.0225
Right

 

Table 12 — Solution for ﬁxed effects

70

 

The ﬁxed effect gender was found to be signiﬁcant (P = 0.0252; 0 =0.05);
a signiﬁcant difference existed in the % of degrees of jar rotation of men and
women, where men rotate their jar hand more than women. Additionally, the
interaction term hand breadth x gender was found to be signiﬁcant at 0 =0.05 (P
= 0.0378 - See Table 12).

When using the variable data (% of rotation of the closure hand) as a
dependent variable, men and women with large hand breadths (102 mm) behave
similarly in terms of rotational behavior (P = 0.7898), while men and women of
moderate hand breadths (92 mm) were signiﬁcantly different (P = 0.0107).
Those with the narrowest hands (82 mm) were also signiﬁcantly different in

rotation behaviors (P = 0.0057), see Table 13 and Figure 33.

 

Differences of Lease Squares Means

 

Hand Hand
Effect Gender Gender Breadth Length Estimate Standard DF t

Pr > It I
Right Right Error Value

 

Gender 1(M) 2(F) 82.00 203.07 46.5701 14.4389 15 3.23 0.0057“

 

Gender 1(M) 2(F) 92.00 203.07 24.7916 8.5929 16.2 2.89 0.0107*

 

 

 

 

 

 

 

Gender 1(M) 2(F) 102.00 203.07 3.0131 11.1356 18 0.27 0.7898

 

 

 

 

*Signiﬁcant at Bonferroni - adjusted comparison - wise type I error rate of 5%

 

 

 

Table 13 — Jar hand rotation by hand breadth and gender

71

 

 

—~¢- Males
a, 100% l —--— Females
2 E
13% 80- \\\
32 .9. \
5 '5 . l
“6 .5 607
g E I\ \\
m o
“J '- 40.. i \g\\l
\I
1
20 1 1 1
80 90 100 110

Right hand breadth, mm
Figure 33 - Jar hand rotation by right hand breadth (mm)

Another signiﬁcant effect when the variable of test was % of degrees of
rotation of the jar hand, was hand length (P =0.0225; a=0.05, see Table 14).
This relationship was such that each mm increase in length was associated with
an increase of 0.84% in the degrees of rotation of the jar hand, In other words,
people with longer hands tended to open a jar by rotating the jar hand more.
Table 3 shows that, on average, males had longer hands than did females by

nearly 20 mm.

72

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Type 3 Tests of Fixed Effects
Number Den
Effect Degrees Degrees F Value Pr > F
Freedom Freedom
Jar Size 1 42.8 1.18 0.2828
Rig Type 2 33.2 2.5 0.0974
Rig Type * Jar Size 2 39.8 1.32 0.2794
Gender 1 15.7 6.12 0.0252 '
Hand Breadth Right 1 16.8 8.19 0.0109
Hand Breadth Right *
Gender 1 15.9 5.13 0.0378
Hand Length Right 1 16.5 6.34 0.0225

 

Table14 — Type 3 tests of ﬁxed effects

There was no evidence of an effect of jar size or rig type on the
percentage degrees of rotation of the jar hand (P = 0.28 and 0.10, respectively)
see Table 14.

Effect of Opening and Jar Size on % Rotation Degrees by Jar Hand

In order to include the effect closure hand, either left (opening 2) or right
(opening 1), as an explanatory factor in the model, analysis of the opening
rotation in percent was simpliﬁed. (Recall from the previous discussion that due
to the lack of data for closure hand across the three opening positions (table,
pole rig and side rig, closure hand could not be included in the analysis).
Analysis was repeated to compare the restricted openings (opening 1 - table) to
the unrestricted openings (opening 2) and the comparison between the pole and

side rigs (grip style) were removed.

73

A general linear mixed models was ﬁt to the continuous response variable
“percentage of rotation degrees by the jar hand” using the MIXED procedure of
SAS (version 9.1, SAS Institute, Cary, NC). Data were collected in a RCB
design, such that subject was considered the blocking factor for assigning a 2x2
jar size by opening restriction treatment combination. By design, the base
statistical model included the categorical ﬁxed effects of jar size (2 levels),
opening restriction (2 levels- unrestricted and table) and their 2-way interaction.
The model included the random effects of subject, subject by opening and
subject by jar size in order to recognize the blocking factor and the appropriate
degrees of freedom for each of the treatment factors, respectively. In addition, a
random effect of subject by jar size by opening restriction was ﬁtted to account
for technical replications within the design (each subject in each opening ’
restriction by jar size combination was measured 3 times).

In addition, potential explanatory variables of interest were recorded at the
subject level, including: gender, dexterity, laterality (both by arm cross and hand
clasp) and a total of 9 morphometric continuous variables. Closure hand (left or
right) used in each jar-opening event was also recorded. In order to decide
between these potential explanatory variables, stepwise model selection based
on Bayesian Information Criteria (BIC) was used to select a model of best ﬁt. In
this case, the model of best ﬁt based on BIC included, in addition to the base
model described above, the effects of gender, closure hand, its interaction with
Opening restriction, LEFT hand breadth and LEFT wrist — ﬁnger length. Degrees

of freedom were computed using Kenward-Roger Method. Model assumptions

74

 

were checked using standardized residuals and were considered to be
appropriately met. Least square means and standard errors are provided, see
Tables 18, 19 and 20. Post-hoc pain/vise comparisons were performed using

Bonferroni’s adjustment to avoid inﬂation of type I error rate, see Table 19.

75

 

 

Solution for Fixed Effects

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

76

. Jar Closure . Standard t

Effect Gender Opening Size Hand Estimate Error DF Value Pr>ItI
Intercept 90.8110 23.1989 16.3 3.91 0.0012
JarSize 1 -8.5909 3.4507 59 -2.49 0.0156
JarSize 2 0

Opening 23.8171 5.8895 63.4 4.04 0.0001
Opening 0

Opening

*Jar 1 6.7273 4.8819 59 1.38 0.1734
Size

Opening

*Jar 2 0

Size

Opening

*Jar 1 0

Size

Opening

*Jar 2 0

Size

Gender 1 20.2015 4.0853 15.9 4.94 0.0001
Gender 2 0

Closure

Hand 1 4.5686 4.5722 47.4 -1.00 0.3228
Closure

Hand 2 0

Opening

Closure 1 50.5138 6.0588 65.6 —8.34 .<.0001
Hand

Opening

Closure 2 0

Hand

Opening

Closure 1 0

Hand

Opening

Closure 2 0

Hand

Hand

Breadth -1.4275 0.2974 15.9 -4.80 0.002
__Left

Finger

Length 0.4679 0.1603 16 2.92 0.0101
Left

Table 15 — Solution for ﬁxed effects

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Type 3 Tests of Fixed Effects
Effect Num DF Den DF F Pr > F
Value
Jar Size 1 59 4.59 0.0364
Opening 1 61.5 0.40 0.5272
Opening * Jar

Size 1 59 1.90 0.1734

Gender 1 15.9 24.45 0.0001

Closure Hand 1 18.8 84.82 <.0001
Opening * Closure

Hand 1 65.6 69.51 ‘ <.0001

Hand Breadth Left 1 15.9 23.03 0.0002

Finger Length Left 1 16 8.52 0.0101

Table 16 — Type 3 tests of ﬁxed effects

Unlike the analysis that did not include the factor closure hand, the ﬁxed
effect of jar size was identiﬁed as having a signiﬁcant effect on the % degrees of
rotation of the jar hand (P = 0.04; c = 0.05). Jars of smaller size triggered, on
average, a greater % of degrees of rotation by the jar hand, see Tables 15, 16,

17, 19 and 20.

77

 

Overview of Rotation in Percent and Degrees All Openings N = 252

 

 

 

 

Jar A (8.73252 cm) Jar B (10.95502 cm)
Standard Standard
Average Deviation Average Deviation
Rotation Closure
Hand % 57.46 24.40 63.49 25.99
R°tat'°",,/::a' ”and 42.54 24.40 36.51 25.99

 

 

 

 

 

Unrestricted Openings N = 126

 

 

 

 

Jar A (8.73252 cm) Jar B (10.95502 cm)
Standard Standard
Average Deviation Average Deviation
Rotation Closure
Hand % 49.96 18.18 58.44 21.84
ROtat'mo/ja' ”and 50.04 18.18 41.56 21.84

 

 

 

 

 

Restricted Openings N = 126

 

 

 

 

 

Jar A (8.73252 cm) Jar B (10.95502 cm)
Avera e Standard Avera e Standard
9 Deviation 9 Deviation
Rotation Closure
Hand % 64.96 27.48 68.54 28.85
R°tat'°",;:ar ”and 35.04 27.48 31.46 28.85

 

 

 

 

 

Table 17 — Rotation by jar size

 

Further, the analysis indicated a signiﬁcant effect of gender on % degrees
of rotation by jar hand (P = 0.0001; c = 0.01), such that males showed a greater

% of degree of rotation by the jar hand than females, see Table 20. This ﬁnding

78

follows the anthropometric data in that male subjects on average had left hand

lengths that were nearly 20 mm longer then the female subjects see Table 3.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Differences of Least Squares Means
Jar Jar . Standard t
Effect Gender Size Gender Size Estimate Error DF Value Pr>ItI
Star 1 2 -5.2272 2.441 59 -2.14 0.0364
ize
Gender 1 2 20.2015 4.0853 16 4.94 0.0001

 

Table 18 — Differences of least square means

The two morphometric measures included in the model (hand breadth and

ﬁnger length) were found to have a signiﬁcant effect (a = 0.01) on the % degrees

of rotation by the jar hand (P = 0.0002 and P =0.0101, respectively), see Tables

15 and 16. On average, each 1 mm increase in ﬁnger length of the left hand was

associated with a 0.46% increase in degrees of rotation by the jar hand; each 1

mm increase in breadth of the left hand was associated with a 1.4% decrease in

degrees of rotation by the jar hand.

The interaction term closure hand and opening restriction was also found

to have a signiﬁcant effect on the % rotation provided by the jar hand (P <

0.0001; 0 =0.01), see Tables 15, 16, 19 and 20. The nature of the interaction is

depicted in the Figure 34.

79

 

 

Differences of Least Squares Means

 

Opening

Closure
Hand

Opening

Closure
Hand

Estimate

Standard
Error

DF

t Value

Pr>ItI

 

 

-55.082

4.2924

56

-12.8

 

-23.333

2.7406

60

-8.51

 

 

27.1807

5.399

5.03

 

Opening "
Closure
Hand

 

 

 

 

 

4.5686

 

4.5722

 

47

 

-1.00

 

 

 

(*) Signiﬁcant at Bonferroni - adjusted comparison - wise type I error rate of 5%

 

 

Table 19 — Diffgences of least square means

80

 

 

 

 

 

 

 

 

 

 

   

Least Square Means
. Jar Closure . Standard t

Effect Gender Opening Size Hand Estimate Error DF Value Pr>ItI
Jar Size 1 42.7737 2.1712 34 19.70 <.0001
Jar Size 2 48.0009 2.1490 34 22.3 <.0001
Opening 1 46.3492 2.2713 44 20.4 <.0001
Opening 2 44.4254 2.4022 42 18.5 <.0001
Openiﬂg> , " 7". .
7’ , *"f t

, furs, 1 1 18.8080 2.0968 44

. and.
deﬁle.

if‘ '4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 2 73.8904 3.8902 50
2 1 42.1411 2.0620 43
2 2 46.7097 4.2122 45 4
I ..1
IGender 1 55.4880 2.0257 16, 27.4 <.0001
IGender 2 35.2865 3.2549 16 10.8 <.0001

 

 

Table 20 — Least square means

81

80

m 70
22
82 6°
'0...
,\°.E 50
8‘5
138 40
.58
1736 30
mt...

20

10

 

—9— Restricted ,1
—-l— Unrestricted //

 

/
/
- ,, ,...,/J
I 7”
/
.r/I/
/
I
Right Left

Closure hand

Figure 34 - Jar hand rotation by opening restriction

Figure 34 depicts the signiﬁcant interaction term graphically. The

interaction between closure hand and opening restriction indicates that the effect

of opening restriction on % degrees of rotation differed depending on the closure

hand used. For the restricted openings when the right hand is on the closure, the

estimated % degrees rotation of the left hand (jar hand) is much lower than when

the left hand is on the closure and the right hand is on the jar. This indicates that

for this population of subjects, when restricted to the table, the right hand

provided a greater % degree of rotation than the left hand, regardless of hand

placement. In the unrestricted openings, the difference between left or right hand

placement was not as pronouced but the same observation can be applied.

82

When compairing the closure hand between restricted and unrestricted
openings the right hand on the closure supplied a greater % degrees rotation in
the restricted than in the unrestricted meaning that the % degrees rotation
estimated for the jar hand is lower in the restricted opening than in the
unrestricted opening. For the left hand on the closure in the restricted opening
there was a high % degrees of rotation of the jar hand indicating that the left
hand did not supply as great a percentage as the jar hand. However, in the
unrestricted openings the left hand on the closure supplied a greater % degrees
rotation than in the restricted openings. Therefore, when restricting subjects to
the table one hand, in this case the right, will tend to be used to a greater extent
in opening compaired to the unrestricted opening where one hand does not tend
to dominate on average.

These ﬁndings ﬁt with the trend seen if the scatter plot of the restricted
openings and the scatter plot of the unrestricted openings, see Figures 31 and
32, where restricted openings tend to clump together and urestricted openings do
not.

Most important is the fact that the ﬁndings indicate a difference between
restricted and unrestricted openings.

Rig Type under Unrestricted Movement Conditions

A generalized linear mixed models was also ﬁt to the binary response
variable “rig type under unrestricted conditions” (namely “side” vs. “bottom”)
using the GLIMMIX procedure of SAS (version 9.1, SAS Institute, Cary, NC).

Data were collected in an RCB design, such that subjects (n = 21) were

83

considered the blocking factor for assigning jar size (2 levels- A and B) under
conditions of unrestricted movement only (data collected on table-restricted
movement were not included in this analysis). By design, the base statistical
model included the categorical ﬁxed effects of jar size (2 levels) and the random
blocking factor of subject. Thus, a total of 42 observations (subjects, 21 x jar
size, 2) were available for analysis.

Additionally, potential explanatory variables of interest were recorded at
the subject level, including gender, dexterity, laterality (both by arm cross and
hand clasp), closure hand and a total of 9 morphometric continuous variables.
All subject*jar size combinations were consistent in their “closure hand” under
unrestricted movement conditions (i.e. people did not switch hands when
opening in the unrestricted orientation). In order to decide between these
potential explanatory variables, a stepwise model selection was used to select a
model of best ﬁt. The ﬁnal model included, in addition to the base model
described above, the effects of laterality arm cross and LEFT hand breadth.
Least square mean probabilities of the pole rig (rig type =1), plus minus standard
errors, are provided below in the columns labeled MEAN and ERROR MEAN,
see Table 24. Post-hoc pain/vise comparisons were performed using Tukey’s

adjustment to avoid inﬂation of type I error rate.

84

 

 

 

 

 

 

 

 

 

Solutions for Fixed Effects

Effect 01:83 Sjiire Estimate Stgngard DF Vatlue Pr > I t l
Intercept -14.5618 9.2302 18 -1.58 0.1321
jar-Size 1 -2.8461 1.4906 20 -1.91 _ 0.0707—
Jar Size 2 0
Lat Cross 1 -2.9004 1.6347 20 -1.77 0.0912
Lat Cross 2 0

Hand 7
Brlezfdth 0.2073 0.1034 20 2.00 0.0588

 

 

 

 

 

 

 

 

 

 

Table 21 — Solutions for ﬁxed effects
Jar size was found to have a marginally signiﬁcant effect (P =0.0707) on
the selection of rig type, see Tables 11, 21 and 23. The least square mean
probabilities of the use of the side rig (or a subject that gripped the jar from the
bottom) under the unrestricted portion of testing were 98.2+- 2.5 % for Jar B
versus 76.4 +- 17% for Jar A, see Tables 11 and 22. This ﬁnding indicates that
subjects were marginally more likely to use the side rig, meaning they grasp the

bottom of the jar, if the jar size was larger.

 

Jar Size Least Square Means

 

Standard
Jar . Standard t
Size Estimate Error DF Value Pr>ItI Mean '52:;

 

1 1.1730 0.9414 20 1.25 0.2271 0.7637 0.16990

2 4.0192 1.4753 20 2.72 0.0131 0.9823 0.02558
Table 22 - Jar size least square means

 

 

 

 

 

 

 

 

 

 

 

85

 

Difference of Jar Size Least Square Means Adjustment for Multiple
Comparisons: Tukey - Kramer

 

 

 

 

 

 

 

 

 

 

Jar . . Standard .
Size Jar Size Estimate Error DF t Value Pr > I t I Adj. P
1 2 -2.8461 1.4906 20 -1.91 0.0707 0.0707

 

Table 23 — Tukey — Kramerjar size

A marginal association between arm cross laterality and the probability of

the side rig (rig type = 2) under unrestricted movement was identiﬁed (P =

0.0912), see Table 25. The least square mean probabilities of the side rig (rig

type = 2) under unrestricted movement were 98.3+- 2.4 °/0 versus 75.9 +- 20% for

laterality left arm top (arm cross 2) and right arm top (arm cross 1), respectively,

see Table 24. This ﬁnding indicates that subjects whom placed their left arm on

top were marginally more likely to use the side rig, meaning they grasp the

bottom of the jar, than subjects who placed their right arm on top.

 

 

 

 

 

 

 

 

 

 

 

 

Lat Cross Least Squares Means
Standard
Lat Estimate Standard DF t Pr > It I Mean Error
Cross Error Value - M
ean
1 1.1459 1.1313 20 1.01 0.3232 0.7588 0.20710
2 4.0463 1.4170 20 2.86 0.0098 0.9828 0.02394

 

Table 24 — Lat cross least squares means

 

Comparisons: Tukey - Kramer

Differences of Lat Cross Least Squares Means Adjustment for Multiple

 

 

 

 

 

 

 

 

 

 

Lat Lat . Standard _
Cross Cross Estimate Error DF t Value Pr > I t I Adj. P
1 2 -2.9004 1.6347 20 -1.77 0.0912 0.0912

 

Table 25 — Tukey — Kramer lat cross

A marginal association was identiﬁed between left hand breadth and the

probability of the side rid (rig type 2) under unrestricted movement (P = 0.0588),

86

 

 

 

 

see Table 21. Based on the estimated odd ratios, the odds of the side rig (rig
type 2) increased by 23% for each mm increase in hand breadth of the left hand,

see Table 16.

87

Chapter 5

CONCLUSIONS

One of the main goals of this research was to develop techniques for a
methodology that could be used for unrestricted motion capture. This included
creating a way to identify the rotational axis in order to evaluate and use the
motion capture system for use in packaging applications that require the
consumer to apply rotational movement. Data generated from this study also
serves as pilot work in understanding the factors that impact the opening motions
associated with two sizes of glass jars with lug style closures.

We hypothesized that the physical positioning of a subject had the
potential to impact the opening motion employed. Therefore, two opening
positions were incorporated into the experiment; opening “restrained” to a
counter set at a standard height of 91 cm (36 inches) [11] the other opening in
the air at a height comfortable to the subject, “unrestrained opening.“ Findings
indicated a statistically signiﬁcant interaction between closure hand and opening
restriction on % degree of rotation by the jar hand (P < 0.0001), see Tables 15,
16, 19 and 20. The nature of the interaction is depicted in Figure 34. This
suggests that restraining users will change their biomechanical approach to
opening. This should be considered by future researchers that intend to collect
data for the purpose of informing design.

It was also hypothesized that the motion used by a subject to open glass
jars with lug style closures would take one of three forms;

1. Closure Hand Supplies Motion

88

 

2. Jar Hand Supplies Motion
3. Both Hands Supply Motion
Collected data suggest a spectrum of rotational behaviors, and no discrete
breaks in the rotational behaviors. However, openings were not grouped by
individual were breaks may have occurred from subject to subject if subjects
displayed repeatability in opening motion.
In both the restricted and unrestricted openings the majority of subjects
placed their right hand (dominant hand) on the closure, see Table 2. This ﬁnding
is not consistent with that of Voorbij and Steenbekkers (2002) who found that 511
or 534 or roughly 96 percent of subjects that are right handed place their left
hand on the closure [3].
There are several factors that give insight into why one hand may supply
motion compared to the other hand.
1. Length of the jar hand (positive P = 0.02), people with longer hands
tend to open the jar by rotating the jar hand more, see Table 14.

2. Gender and breadth of right hand (Signiﬁcant P = 0.0378), men of
narrow hands had signiﬁcantly greater % degrees of rotation of the jar
hand than women of narrow hands, see Table 14.

3. Gender (Signiﬁcant P = 0.0001), males rotate jar hand more than
females, see Table 20.

4. Jar Size (Signiﬁcant P = 0.04), smaller jars lead to greater rotation of

jar hand, see Tables 15, 16, 17, 19 and 20.

89

Data analysis of the variable data (% of rotation of the closure hand) indicated
that men and women with large hand breadths (102 mm) behave similarly in
terms of rotational behavior (P = 0.7898), while men and women of moderate
hand breadths (92 mm) were signiﬁcantly different (P = 0.0107). Those with the
narrowest hands (82 mm) were also signiﬁcantly different in rotation behaviors (P
= 0.0057), see Table 13 and Figure 33.

It is interesting to note that as the jar size decreases essentially hand size,
more speciﬁcally hand length in proportion to jar size, increases. Therefore, the
ﬁnding that smaller jars lead to greater rotation of the jar hand seems rather
intuitive when coupled with the ﬁnding that people with longer hands tend to open
the jar by rotation the jar hand more. However, the ability to examine both the jar
hand and the closure hand was constrained in that the research did not have any
subjects whom grasped the side rig jar with their left hand on the closure, see
Table 11. Due to this fact the closure hand was removed from the analysis and
the ability to examine the unrestricted openings as pole rig and side rig was
limited. Data analysis comparing restricted and unrestricted openings was
therefore carried out combining pole rig and side rig openings.

One interesting point to take note of include the fact that both male and
female subjects in the study had above average size hands in terms of hand
length, hand breadth and wrist — ﬁnger length, see Figures 24, 25 and 26 . This
fact, coupled with the fact that all subjects had no history of injury to the hands,
wrists, arms or shoulders could have greatly inﬂuenced the results reported. This

fact is interesting in that the general population should have smaller hand sizes

90

 

 

on average compared to the subjects in this study. Further research, that
recruits people with a spectrum of hand sizes, is needed to purposefully
investigate the relationship between hand size and rotational behavior.
There are several factors that give insight into why a subject, during an
unrestricted opening, chooses a particular grip.
1. Jar Size (Marginal P =0.0707), this ﬁnding indicates that subjects were
marginally more likely to use the side rig, meaning they grasp the
bottom of the jar, if the jar size was larger, see Tables 11, 21 and 23.
2. Hand breadth (Marginal P = 0.0588), as hand breadth increases, the
probability of using a side rig increases; meaning that a subject is more
likely to grip the jar from its bottom, see Table 21.
The examination of laterality did not uncover great insight as to why
subjects grasp or opening jars in the way they do. However, this study did agree
with the ﬁndings of Mohr et. al. (2003, 2006) in that the subjects in this study had

a nearly equal distribution across the four laterality codes [8, 9].

91

Chapter 6

RECOMMENDATION AND FUTURE STUDIES

Yoxall et. al. (2007) examined the role that grip styles play in opening
jars of varying sizes [14]. According to their research, for the jar sizes
used in this study, subjects should have demonstrated several grip
styles including spherical, box, lateral and cylindrical. It is
recommended to explore the role that grip style has on the opening
motion used by the closure and jar hands.

Compared to the data from the NTIS study [13] subjects in this report
had on average relatively large hands. It is recommended to
investigate a population of subjects with relatively small hands for a
comparison, and to investigate whether or not subjects of smaller size
hands would show the same biomechanical approaches to opening.
The current study recruited subjects who did not have a history of
injury to the hands, wrists, arms or shoulders. This was done to
simplify reﬁnement of the methodological approach. It is
recommended that the study be repeated using a more varied
population.

With only one left handed subject dexterity could not be used as a
factor in the statistical analysis. Inclusion of a larger sample size of left
handed subjects is recommended in order to have a balance between

left and right handed subjects.

92

 

The population was recruited as a convenience sample from the MSU
campus, and was comprised of college students. This is obviously, not
representative of the population at large. It is recommended that a
more representative population be recruited for future studies. Of
particular interest is the aging population, who are known to have
difﬁculty with these types of packages [1 — 4].

As indicated, in this study, all subjects were screened to insure they
did not have a history of injury to the hands, wrists, arms or shoulders.
Subjects were also found to have very similar measurements related to
strength in each hand. It is recommended to investigate how opening
motion and style change in relation to loss of strength in the hands,
wrists, arms or shoulders due to injury.

Several studies in the literature review examined the forces involved in
opening jars [1 - 5]. In order to investigate motion in greater detail
forces were not included in this study. Incorporation of the forces
involved in opening a jar coupled with motion is recommended.

This study investigated the motion associated with opening two jars.
These jars were similar in all dimensions other than jar diameter. It is
recommended that this study be used to look at other variations in jar
dimensions such as differences in height, closure diameter, contents
and weight be explored by future researchers.

Due to the lack of observations for the combination of the side rig (rig

type 3) and the left hand on closure (closure hand 2) simultaneous
93

ﬁtting of the two (rig type and closure hand) as explanatory variables in
a given model was not possible. Therefore, the closure hand as an
explanatory variable could not be used in the analysis. Collection of
data to ﬁll this gap and allow for analysis of the closure hand is

recommended.

94

APPENDICES

95

Appendix A - Recruitment Flyer

 

 

Opportunity for people 18 and older that have NO history

of injury to the hands, wrists or shoulders to participate in

a research project regarding the opening of jars in
exchange for a $25 Meijer Gift Certiﬁcate

 

This research project is a joint study between the MSU School of Packaging
and the Department of Mechanical Engineering. The study will examine the
biomechanical features of opening a series of ﬁlled glass jars. All testing will take
place at:

4133 Engineering Building

 

Please come to testing wearing a
comfortable, short sleeved tee shirt.
Testing will take no longer than 2 hours;
during testing the researchers will measure
and size of your hands, you will be asked
to participate in several tests of hand
strength, clasp your hands, your arms and
open several different glass jars. Testing
will be videotaped.

In exchange for your participation, you will
be provided with a $25 Meijer gift
certiﬁcate. If at any time you are
uncomfortable with the testing or wish to

discontinue the data collection process, you may discontinue participation without
penalty.

If you are interested in learning more about this study, please contact Joe Fair at
fairjose@msu.edu or phone 517-410-2960 to make an appointment. If you have
questions or comments regarding this study, please contact Dr. Laura Bix, at the
School of Packaging at (517) 355-4556 or bixlaura@msu.edju or Dr. Tamara
Reid-Bush from the Department of Mechanical Engineering at
reidtama@msu.edu. If you have questions or concerns about your rights as a
study participant contact the SIRB at (517) 355-2180.

96

 

Appendix B - Data Collection Form

Subject #

 

Female Male

Age

 

Preferred Dexterity (circle)
Right hand
Left hand

History of allergic reaction or irritation to rubbing alcohol or surgical tape?

 

History of Injury to hands, shoulders, arms or wrists?

 

Digital Photo of Right hand, palm down (ﬁle number)

 

Digital Photo of Right hand, palm up (ﬁle number)

 

Digital Photo of Left hand, palm down (ﬁle number)

 

Digital Photo of Left hand, palm up (ﬁle number)

 

Digital Photo of Right Thumb

 

Digital Photo of Left Thumb

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Right Grip Strength Measurement Left Hand Strength Measurement
1. 1.
2. 2.
3. 3.
Right Wrist Strength Measurement Left Wrist Strength Measurement
1. 1.
2. 2.
3. ‘ 3.
Palm to Palm Squeeze
1.
2.
3.

 

97

Digital Photo Hand Clasp

 

 

Digital Photo Arm Folding

 

 

Package testing Jar 1 Position

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fixturing
Opened in Right or Left Jar Hand
Trial Time? Hand on Position (23221326
Yes/No Top? (T,M,B) '
Jar 2 Position
Fixturing
Opened in Right or Left Jar Hand
Trial Time? Hand on Position 03:22:26
Yes/No Top? (T,M,B) '
Jar 1 Position
Fixturing
Opened in Right or Left Jar Hand
Trial Time? Hand on Position 02:22:26

Yes/No Top? (T,M,B)

 

 

 

 

 

 

 

 

 

 

 

98

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Jar 2 Position
Fixturing
Opened in Right or Left Jar Hand
Trial Time? Hand on Position (3:223:23
Yes/No Top? (T,M,B) '
Thank the subject.

Have them print and sign that they have received the Meijer gift card.

99

Appendix C — Torque Requirements

In order to determine the rotation in degrees that was required to open the
jars a series of ten jars, for both the large diameter and small diameter jar, had a
torque force of 35 in — lbs applied using a Secure Pak, Inc. electronic torque

tester, see Figure 35.

 

Figure 35 — Secure Pak, Inc. electronic torque tester

From the series an average and standard deviation for the rotation in
degrees was calculated for both the large diameter and small diameter jars. The
results were used during the data processing. Results from the torque testing
are shown in Table 26. To obtain the values in Table 26 one large diameter jar
and one small diameter were used. For each trial a new closure that had never
been used before was applied to the jar. Each closure had a circumference of
235 mm, this value was determined by tracing the closure on paper and then
using string to obtain the distance. The circumference was not be obtained using

the equation 21Tl’ since the closure diameter used to identify the closure in

100

 

packaging terms, 70 mm, is the jar diameter at the closure not the exterior
diameter of the closure itself. In Table 26 the column Distance (mm) refers to the
distance of rotation measured after the lugs in the closure were no longer

engaged with the jar.

101

 

 

Table of closure removal trials used for average opening angles (0)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Jar B (10.95502 cm)
Trial Distance (mm) Angle (o)
1 16 24.5
2 18 27.6
3 16 24.5
4 17 26.0
5 21 32.2
5 11 16.8
7 16 24.5
3 17 26.0
9 15 23.0
10 13 19.9
Average 24.5
Standard Deviation 4.16
Jar A (8.73252 cm)
Trial Distance (mm) Angle (o)
1 17 26.0
2 20 30.6
3 27 41.4
4 19 29.1
5 16 24.5
6 19 29.1
7 23 35.2
8 19 29.1
9 14 21.4
10 24 36.8
Average 30.3
Standard Deviation 6.01

 

Table 26 — Trials used to calculate average opening angles (0) for both the large and

small jar

The Distance (mm) value was found by drawing a vertical line to the

closure and jar after applying the torque of 35 in — lbs. The closure was then

102

 

 

rotated to the point where the lugs disengaged the jar. At this point another

vertical line was drawn on the closure, see Figure 36.

Removal
Distance
(mm)

 

Figure 36 — Lines drawn on closure to determine removal distance
This closure was then removed and the distance between the two lines
was measured to obtain the value shown in Table 26. The distance was then
divided by the diameter; this value was then multiplied by 360 degrees to obtain

the opening angle.

103

Appendix D - Motion Capture Calibration

Prior to testing the Qualisys software had to be calibrated. The ﬁve
cameras were positioned around the table set at its height of 36 inches or 91.44
centimeters. The Qualisys software was opened and a calibration ﬁle was setup.
The calibration capture lasted 10 seconds and was collected at a rate of 60 Hz.
One member of the research team ran the computer software while a second
member used a specially created L bracket and wand that was ordered from the
same company as the Qualisys software, see Figure 37, that had reﬂective

targets on them.

 

I ngure 37 - Calibration bracket an wad
The L bracket was place on the table with the long edge parallel to the

back edge (where the subject would stand) of the table. The wand had two

104

 

targets on it measuring 525 mm between them. The bracket had four targets.
The distance on the short arm was 350 mm, the long arm was 500 mm and the
mid point on the long arm was 150mm. These values were entered into the
setup of the calibration ﬁle. Entering the incorrect values would create calibration
values that were not desirable and void all data collected.

During the ten second calibration the team member at the table would
wave the wand around the table above the L bracket in a random fashion. The
software would then process the data from each camera and determine a
calibration level for each. The team attempted to make sure that each camera
had a calibration level below 1. If one or two of the cameras were over but close
to one the system would be considered ready for capture. If more than two of the
cameras were over 1 the system was re - calibrated until the issue was solved.
Solving the issue of cameras not being calibrated to a desired level included
several different trial and error steps. These included but were not limited to re —
booting the computer, check the wires from the camera to the computer,
adjusting the camera angle, adjusting the camera height and moving the location

of the camera.

105

Appendix E — Opening Rotation Sequence

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Subject Jar Series 1 Jar Series 2 Jar Series 3 Jar Series 4
001 Not Used Not Used Not Used Not Used
002 Jar A Table Jar B Table Jar A Free Jar B Free
003 Jar A Table Jar B Table Jar B Free Jar A Free
004 Not Used Not Used Not Used Not Used
005 Not Used Not Used Not Used Not Used
006 Jar A Table Jar A Free Jar B Table Jar B Free
007 Jar A Table Jar B Free Jar B Table Jar A Free
008 Jar A Table Jar A Free Jar B Free Jar B Table
009 Jar A Table Jar B Free Jar A Free Jar B Table
010 Jar B Table Jar A Table Jar A Free Jar B Free
011 Jar B Table Jar A Table Jar B Free Jar A Free
012 Jar B Table Jar A Free Jar A Table Jar B Free
013 Jar B Table Jar B Free Jar A Table Jar A Free
014 Jar B Table Jar A Free Jar B Free Jar A Table
015 Jar B Table Jar B Free Jar A Free Jar A Table
016 Jar A Free Jar B Free Jar A Table Jar B Table
017 Jar A Free Jar B Free Jar B Table Jar A Table
018 Jar A Free Jar A Table Jar B Table Jar B Free
019 Jar A Free Jar A Table Jar B Free Jar B Table
020 Jar A Free Jar B Table Jar A Table Jar B Free
021 Jar A Free Jar B Table Jar B Free Jar A Table
022 Jar B Free Jar A Free Jar A Table Jar B Table
023 Jar B Free Jar A Free Jar B Table Jar A Table
024 Jar B Free Jar A Table Jar B Table Jar A Free
025 Jar B Free Jar A Table Jar A Free Jar B Table

 

106

 

BIBLIOGRAPHY

107

BIBLIOGRAPHY

1. Yoxall A, Janson R, Bradbury SR, Langley J, Wearn J, Hayes S (2006)
Openability: Producing design limits for consumer packaging. Packaging
Technology and Science 19: 219-225

2. DTI. 2002. Speciﬁc anthropometric and strength data for people with
dexterity disability. URN 02/743 Departmetn of Trade and Industry.
London.

3. Voorbij AIM, Steenbekkers LPA (2002) The twisting force of aged
consumers when opening a jar. Applied Ergonomics 33: 105-109

4. Crawford JO, Wanibe E, Nayak L (2002) The interaction between lid
diameter, height and shape on wrist torque exertion in younger and older
adults. Ergonomics 45: 922-933

5. Soroka W (2002) Closures. In: Soroka W (ed) Fundamentals ,of Packaging
Technology, 3rd edn. loPP, Naperville Illinois

6. Fowler NK, Nicol AC (2001) Functional and biomechanical assessment of
the normal and rheumatoid hand. Clinical Biomechanics 16: 660-666

7. Murgia A, Kyberd PJ, Chappell PH, Light CM (2004) Marker placement to
describe the wrist movements during activites of daily living in cyclical
tasks. Clinical Biomechanics 19: 248-254

8. Mohr, C., Thut, G., Landis, T., Brugger, P. (2003). Hands, arms, and
minds: Interactions between posture and thought. Joumal of Clinical and
Experimental Neuropsychology 25(7), 1000-101 0

9. Mohr C, Thut G, Landis T, Brugger P (2006) Arm folding, hand clasping,
and Luria’s concept of “latent left-handedness.” Laterality 11: 15-32

10.Stins JF, Kadar EE, Costall A (2001) A kinematic analysis of hand
selection in a reaching task. Laterality 6: 347-367

1 1 . GEAPPLIANCES.com 2007. Universal Design Frequently asked
Questions. 29/8/07
http://wwwgeappliances.comldesign center/universal design/fag.htm

12. ASSH.org 2008. Hand Anatomy. 26/10/08
http://www.assh.org/Content/NavitationMenLL/PatientsPublic/Handinatom

y/Hand Anatomypdf

108

13.U.S. Department of Commerce National Technical Information Service
March 1989. 1988 Anthropometric survey of US. army personnel:
Summary statistics interim report. Anthropology research project, inc.
Yellow Springs, OH

14.Yoxall A, Luxmoore J, Austin M, Canty L, Margrave KJ, Richardson CJ,
Wearn J, Howard IC, Lewis R (2007) Getting to grips with packaging:
Using ethnography and computer simulation to understand hand-pack
interaction. Packaging Technology and Science 20: 217-229

109

   

”11111111111111[(ljljlljjljljljljllll“