|UFDC Home||myUFDC Home | Help|
|Table of Contents|
|List of Tables|
|List of Figures|
|Review of the literature|
|Summary, conclusions, and...|
This item has the following downloads:
|Table of Contents|
Table of Contents
List of Tables
List of Figures
Review of the literature
Summary, conclusions, and recommendations
RELATIONSHIPS BETWEEN TEMPERAMENT TYPE AND PERCEIVED
SELF-EFFICACY AMONG INFORMAL CAREGIVERS
TERESA A. TOZZO LYLES
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Teresa A. Tozzo Lyles
This document is dedicated to my parents, Carmen and Donald Tozzo;
my children, Leslie, Brianna, and Morgann Lyles;
and the Tozzo, Julian, and Clark Families
"Wisdom is a gift of God," is the basis of a verse in the Old Testament's Book of Wisdom.
During the past five years, this quote has always proven true. Without my faith, and of course,
the prayers and support of family and friends, I would not have been able to complete this
First, I want to thank my Lord for His guidance and ultimate mercy, His wisdom is great
and His forgiveness is never ending.
My committee has been the backbone and driving force behind the success and completion
of this study. Without them, I would not have been able to put one foot in front of the other. Dr.
R. Morgan Pigg, Jr., my chair, has the patience of a saint and for that I thank him
wholeheartedly. He has encouraged me to continue on in the most caring and wise way. I thank
Drs. Kim Walsh-Childers and Christine Stopka, whom I have known for more than a decade, for
the spiritual guidance, the electric energy, the brainstorming on horseback, the experiences with
Sidney Lanier and Adapted Aquatics, and most importantly, the hugs. I thank Drs. Jiunn-Jye
Sheu and Virginia Dodd, for sticking with me and not letting me lose my footing when I was
slipping. I also appreciate that everyone overlooked my often quirky behavior.
My parents, Carmen and Donald Tozzo, brought me to this earth to experience all my gifts
and graces. Although I had to lose my wonderful father in 1999, he is the reason that I became
interested in social support and caregiving in the first place. From his perch in Heaven, he has
watched diligently over our family. He has guided me when I had no clue, and taught me to do
what makes me happy. My mother, who is my primary spiritual guide, has been my rock and my
light. I thank mommy and daddy for their love, gifts, and prayers. I also want to thank two
wonderful and special people who have been in my life since birth, my aunt and uncle, Blanca
and Frank Tozzo. They have always given me everything that two parents can, and for this, I
love and cherish them dearly.
I want to tell my three beautiful, dear and wonderful daughters, Leslie, Brianna, and
Morgann, that they will always be a piece of my own heart and soul. My children helped me to
get up and start every day, even when I did not have the energy or the will. I want to thank them
for their many smiles and tears. I want to thank my other family members, especially the two
most awesome and beautiful sisters anyone can possibly have, and who were supportive beyond
belief, Carmen and Elena I thank them for always being on my side, regardless if our opinions
differed. I also want to extend my special thanks to my niece and nephews Adam, Matt,
Miranda, and Danny Clark, and Mark and John Julian.
This paper is also dedicated to all my riding girlfriends who kept me addicted to riding and
endurance. I thank them for the drumming, the ride camp parties at the party tent and also
because they never let up on me for one minute, especially Kathy, Colette, Elaine, Brenda,
Roxanne and many others I have met on the trails. My moments of sanity would not have
transpired if it were not for my special best friend, my horse J. Darby, who has carried me for
many miles of joy and provides positive energy despite his odd behavior.
My two childhood friends, Mayte Rodriguez Regan and Felicia Garcia, helped me weather
the storm by just listening and helped me through many tearful days.
I want to send heart-felt thanks to two incredibly spiritual individuals who provided many
prayers, guidance, lunches, spiritual talks, and support Drs. Mary Hughes and Dr. Jane
Emmeree. I want to thank some incredibly wonderful individuals who allowed me to interview
them, before I began to write, about their insights, feelings, thoughts, feedback and reflections on
caregiving that came from their heart and soul. Without them, I would not have been able to
focus and guide my thoughts in the direction that became the final product for this dissertation -
Evelyn Cairns, Kathy Mladinich, Diane Gatsche, Keith Meneskie, and Dale Rimkunas.
I want to thank my incredibly large and supportive family at Invivo Corporation who have
been incredibly kind during the past five years, especially Dr. Randy Duensing, who was patient,
worked around my crazy school schedule, and helped me get to the end. I also want to thank my
work family from the Department of Radiology at Shands, especially Drs. Kate Scott and Jeff
Fitzsimmons who always supported and encouraged me in my academic endeavors.
I want to thank two wonderful people, John and Susie Bauer, who showed me what true
caring and caregiving is all about. Susie was always with me through these final and difficult
steps, and even though she left this Earth for a more ethereal place, her presence from above kept
Finally, I want to thank the University of Florida and the Gator Nation for the honor of
graduating from this phenomenal academic institution.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ..............................................................................................................4
L IS T O F T A B L E S .........................................................................................10
LIST OF FIGURES ............ ...................................... .....................11
ABSTRAC T ................................................... ............... 12
1 INTRODUCTION ............... ................. ........... .............................. 14
R e se arch P ro b lem ........................................................................................................1 5
R atio n a le ................... ...................1...................6..........
R e se arch Q u e stio n s ........................................................................................................... 17
D e lim ita tio n s ..................................................................................................................... 1 7
L im itatio n s ................... ...................1...................8..........
A s su m p tio n s ..................................................................................................................... 1 8
D efin itio n o f T erm s ................................................................................................................ 19
S u m m ary ................... ...................2...................0..........
2 REVIEW OF THE LITERATURE ........................................ ........ .........22
Informal Caregivers and Factors Affecting Their Roles ................. ................. ....22
Inform al C aregiving ................................................................22
Changes in the Caregiver R ole ................................................................. 23
M medical Costs of Caregiving ................................. .......................... .. ....... 26
H health R isks of C aregiving ......................... ... .................. .........................29
Relationship between caregiver health and level of caregiving ............................31
Care receivers with cancer/terminal illness................................. ........ 34
Care receivers w ith m ental illness .................................................. ..........36
Care receivers with dementia ........................................... ............... 37
Degree of Support and Caregiver Distress ..........................................39
Caregiving and health risks ........ ................................ .. ...... .. ... .. ...... .... 39
C aregivers and fam ily support .............................................................. ..41
Cultural and Ethnic Factors in Caregiving ..........................................42
T em peram ent T ype ............................................... ..........................................50
Self-Efficacy and Outcome Expectancies ................ ............................. ...............57
Self-Efficacy and Effect on Health and Disease ................................................... 57
Self-Efficacy Among Health Professionals.................................... 59
Self-E efficacy and C aregivers................................................................ ............... 60
S u m m a ry ................................... .............................................................................................. 6 3
Summary ............... ........................... ...............63
3 M E T H O D S ....................................................... 65
R e se a rc h D e sig n ............................................................................................................... 6 5
Research Variables ..................................... ..... .......... .......... .... 66
Instrum ents ........................... ....................................... ........ 67
C aregiving Self-Efficacy Scale ............................................... ............................ 68
Keirsey Tem peram ent Indicator II ............................................................................ 69
C aregiver P rofile Inform ation .............................................................. .....................7 1
P ilo t S tu d y ..........................................................................7 1
Prelim inary R review ................................................................ .... ... .....71
P ro c e d u re s ................................................................................................................. 7 2
Informed Consent .............. ....................................................... 74
F in d in g s ..................................................................................................................... 7 5
F in al S tu d y ................... ...................7...................7..........
In stru m entation ............................................................................... 77
P ro c e d u re s ................................................................................................................. 7 8
D ata A n a ly sis ............................................................................................................. 8 1
U n iv ariate P ro cedu res ............................................................................................... 82
B iv ariate A n aly se s ..................................................................................................... 8 2
M u ltiv ariate A n aly ses ................................................................................ 83
Instrum ent Reliability ..................................................................... ......... 84
S u m m ary ................... ...................8...................5..........
4 R E SU L T S .............. ... ................................................................90
Participant Characteristics ...........................................................90
Caregiver Group Type, Gender, and Race ...................................... ............... 91
Temperament Type, Age, Overall Health, and Duration of Care .................................91
Tem peram ent and Self-Efficacy ................................................................................ 93
R e search Q u e stio n s........................................................................................................... 9 4
R research Q u estion O n e ............................................................................................. 94
R research Q u estion T w o ............................................................................................. 9 5
R research Q u estion T three ........................................................................................... 96
Research Question Four ................................... .. .. .....................97
Research Question Five ..................................... ..................... .. ........ .... 98
Research Question Six ....................................................... .............. ... 98
Summary ....................................... .. ... .............................100
5 SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS ............... .................107
S u m m ary ...................................................................................................................10 8
Conclusions..................................... ...................... 111
R ecom m endations.....................................................................................................117
A LETTER OF INVITATION TO ORGANIZATIONS ..................... ................. ................... 120
B SOLICITATION POSTER FOR VOLUNTEERS IN PILOT STUDY.............................121
C ORGANIZATIONAL CONTACTS AND PROFILES ............................. ................122
D CAREGIVER SELF-EFFICACY SCALE ...... ... ..................... ......... .....................124
E KEIRSEY TEMPERAMENT INDICATOR II..................... ........................................126
F CAREGIVER PROFILE INFORMATION (DEMOGRAPHICS)..................................... 128
G SCRIPT FOR RECRUITING VOLUNTEER CAREGIVERS ........................... .........130
H PANEL OF CAREGIVERS TO REVIEW PROTOCOL .................................... ...............131
I UNIVERSITY OF FLORIDA INSTITUTIONAL REVIEW BOARD.............................134
L IST O F R E F E R E N C E S ......................................... .. .. .......................................................... 139
B IO G R A PH IC A L SK E TCH ........... ................................................................. ............... 151
LIST OF TABLES
2-1 Incremental caregiving hours and incremental cost of informal care attributable to
d em en tia ......................................................................................................6 4
2-2 Adjusted weekly hours and yearly costs of informal caregiving................... ..............64
3-1 Instrum ent specification table .......................... ................... ................................... 86
3-2 Keirsey Temparament Indicator II Item to total correlation................ ..................87
3-3 Caregiving Self-Efficacy Scale Item to total correlation..............................................88
3-3 Caregiving Self-Efficacy Scale Item to total correlation..............................................89
4-1 Distribution of participants by group type, gender, and race ................ ................102
4-2 Distribution by temperament type, age, overall health, and duration of care................03
4-3 Distribution of Gender by Temperament Type........................................................ 104
4-4 Distribution of Participants by Temperament Type ....................................................... 104
4-5 ANOVA Table Association between total score for Caregiving Self-Efficacy Scale
and K eirsey Tem peram ent Indicator II .......................................................... .... ........ 104
4-6 Spearman's rank correlation Association between total score on Caregiving Self-
Efficacy Scale and daily duration of care .............. .... .......................................... 105
4-7 ANOVA Table Association between total score for Caregiving Self-Efficacy Scale
and total duration of care in four duration ranges...................................................... 105
4-8 Spearman's Rank Correlation Association between total score on Caregiving Self-
Efficacy Scale and difficulty of care provided ............................................105
4-9 Spearman's Rank Correlation Association between total score on Caregiving Self-
Efficacy Scale and caregiver health...................................................... .................. 106
4-10 Stepwise multiple linear regression analysis of temperament, self-efficacy, total
duration of care, and relationship to care receiver..........................................................106
LIST OF FIGURES
3-1 A N O V A E equation M odel ......................................................................... ...................89
3-2 Multiple Linear Regression Model ..............................................................................89
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
RELATIONSHIPS BETWEEN TEMPERAMENT TYPE AND PERCEIVED
SELF-EFFICACY AMONG INFORMAL CAREGIVERS
Teresa A. Tozzo Lyles
Chair: R. Morgan Pigg, Jr.
Major: Health and Human Performance
This study examined the relationship between four temperament types, as defined by The
Keirsey Temperament Indicator II, and self-efficacy, as defined by the Caregiving Self-Efficacy
Scale, among informal caregivers. The study also examined relationships between self-efficacy
and daily duration of care, total duration of care, caregiver health, and degree of difficulty in care
among informal caregivers. Participants included 25 informal caregivers who completed the
instrument for the pilot study, and 160 who completed the final study. Participants in the study
were 18 years or older and caring part-time or full-time for someone 21 years of age or older.
Most caregivers were White (85.6%), female (82%), and older than age 50 (83.2%). Most male
and female caregivers were categorized under the temperament type of "Guardians."
No significant association existed in the level of self-efficacy among the four categories of
temperament types. However, using bivariate analysis methods (One-way ANOVA and
Spearman's Correlation), when comparing total self-efficacy score with total duration of time in
caregiving, caregiver health, and difficulty of care, significant associations existed (p=.017,
p=.020, and p=.045 respectively)
Results of a Stepwise Multiple Linear Regression Analysis of temperament, self-efficacy,
total duration of care, and relationship to care receiver showed significance in a negative
direction between self-efficacy and total duration of care, for "Artisans" (p=.029; coeff b=-
.882), if the care receiver was a friend (p<.001; coeff b=-2.136), and if the care receiver fit the
"Other" response option (p=<.001; coeff b=-1.652). No significance existed between level of
self-efficacy, the four temperament types, and duration of care provided by caregivers. If
caregivers were "Artisans" they were more likely to stay in the caregiver role longer, but if the
care receiver was either a friend or another relationship other than the ones listed in the survey
(i.e., hospice volunteer), caregivers were more likely to spend less time in the caregiver role.
Findings from this study will assist friends, family members, health care facility staff, and
formal health care providers in (1) assessing self-efficacy among caregivers with different
temperament types, (2) selecting methods to reach and assist informal caregivers in their
caregiving role more effectively, and (3) identifying peer advocates to assist individuals
identified as those just entering the caregiver role.
With the increase in the numbers of baby boomers reaching retirement age, the increase in
average life span, and the population of older Americans increasing dramatically, more spouses,
family members, and others provide care to older adults with a chronic or terminal illness.
Informal caregiving, defined as care provided by a family member, close friend, or volunteer
assistant, has been estimated at $196 billion annually (Arno, Levine & Memmott, 1999). During
the latter part of the 1990s, some 28.8% of persons aged 65 to 74 reported a limitation caused by
a chronic condition, and 54.4% reported at least one physical or medical disability (Wright,
1997; Arno et al., 1999).
Spouses represent more than 30% of caregivers, with an estimated 75% aged 65 and older.
Many younger family members also provide part-time care and support. Studies show that
caring for a spouse with a serious illness, dementia, or other physical ailments decreases the
caregivers' focus on their own preventive health smoking cessation, proper nutrition, regular
exercise, and stress reduction (Connell, 1994; Cameron, Franche, Cheung, & Stewart, 2002).
For example, after assuming caregiving responsibilities, 39% of informal caregivers ate less
nutritionally, 32% decreased their exercise regimen, and 43% increased the amount of smoking
The stress and burden of informal caregiving often become the responsibility of females in
the care receiver's family (i.e., mothers, daughters, daughters-in-law, wives), resulting in
negative emotional and physical consequences and health risks for female caregivers (Wright,
1997; Sawatzky & Fowler-Kerry, 2003). Female caregivers, specifically caregiver wives, report
the highest levels of depression in the caregiver population, typically ranging from 28% to 55%
in this population. The amount of time spent in the caregiver role influences health status,
evidenced by more than 60% of caregivers developing health problems after 18 months of
providing care for a family member or friend (Wright, 1997).
Many men and women aged 65 years and older are retired, but many others continue to
work part-time or full-time. Approximately 4.2 million people aged 65 and older remained in
the labor force in 2000 (U.S. Department of Health & Human Services, Administration on
Aging, 2001). These employed seniors continue to worry about their own health, yet more than
75% of caregivers do not focus on their own health behavior. Caregivers tend to engage more in
self-care than in seeking formal health services (American Association of Retired Persons,
2002). Female caregivers tend to seek personal health care only when absolutely necessary and
when it does not interfere with caregiving for their spouse (American Association of Retired
In the health care environment, formal caregivers (i.e., physicians, nurses, counselors)
inquire about patient health status, but they seldom ask caregivers or care receivers about their
personal lives. However, personal characteristics of caregivers, such as temperament type, can
help explain environmental dynamics, social norms, and influence behavior patterns and stress
levels of persons receiving care. For example, although the individual characteristic of
temperament type generally remains stable throughout life, self-efficacy changes with a new life
role or life situation (i.e., becoming a parent or taking on the care of an ill family member).
Therefore, it is likely that caregiver temperament can predict levels of overall self-efficacy in the
care and well-being of the care receiver. Age, gender, time in the caregiver role, and overall
health of the caregiver may determine medical decisions made for the care receiver and in
general for the care giver. This study examined relationships between specific temperament
types and perceived self-efficacy of males and females in a caregiver role.
Temperament, or the emotional disposition of an individual, rather than personality, relates
more closely to the emotional make-up of an individual, and can better predict and determine
personality, decision-making, and coping skills (Morris, 2000). The study of temperament dates
back to the ancient Greeks with the "Theory of Humors" in 5 B.C. (Heineman, 1995, para 2).
Contemporary studies on temperament and personality generally appear in business and
managerial settings, and also have been used to determine medical and dental student study
habits and the ability to work with and interact with others (Morris, 2000).
Following extensive research, Katherine Briggs identified four dominant personality
dimensions: Extroversion-Introversion (E-I), a person's focus of attention; Sensing-Intuition (S-
N), how one gathers information; Thinking-Feeling (T-F), decision-making process of the
individual; and Judging-Perception (J-P), interaction with the outside world.
The four main temperament types used by Keirsey include 16 sub-types (Myers-Briggs,
1985; Daley, 2000; Morris, 2000). Keirsey and Bates (1984) re-conceptualized the personality
types envisioned by Myers-Briggs into four temperaments artisans, guardians, idealists, and
rationals that focus on how individuals react to and perceive their surroundings. Based on the
analysis of the Myers-Briggs Temperament Indicator, Keirsey created a less complicated and
easier method for respondents (Heineman, 1995, para 23; Different drums: Kiersey.com, 1998,
Self-efficacy, a construct first conceptualized by Bandura as part of his Social Cognitive
Theory, reflects the belief that an individual can perform specific tasks or make intelligent
decisions. Self-efficacy, which varies according to situations and life events, also involves the
belief that individuals can complete a task or make the best decision for oneself or a loved one.
The Caregiver Self-Efficacy Scale, used in this study, was conceptualized for the purpose of
assessing self-efficacy in the caregiver role (Bandura, 1977; Steffen, McKibbin, Zeiss,
Gallagher-Thompson, & Bandura, 2002). Researchers only recently have begun to study self-
efficacy in the caregiver role (Steffen et al., 2002; Gignac & Gottlieb, 1996).
To date, few research designs have combined the three variables temperament, self-
efficacy, and caregiving. This study will determine if individual temperament type affects
perceived self-efficacy in the caregiver role (i.e., does one temperament type show higher levels
of self-efficacy than the other four types). Lack of such studies and the plausible relationship
between these variables establishes a need and provides the rationale for conducting this study.
1. Does a significant association exist between "Total Scale Scores" on the Caregiving Self-
Efficacy Scale and temperament type as determined by the Keirsey Temperament Indicator
2. Does a significant association exist between "Total Scale Scores" on the Caregiving Self-
Efficacy Scale and daily duration of care provided by the caregiver?
3. Does a significant association exist between "Total Scale Scores" on the Caregiving Self-
Efficacy Scale and total duration of care provided by the caregiver?
4. Does a significant association exist between "Total Scale Scores" on the Caregiving Self-
Efficacy Scale and intensity/difficulty of care provided by the caregiver?
5. Does a significant association exist in self-efficacy scores as determined by "Total Scale
Scores" on the Caregiving Self-Efficacy Scale and caregiver health status?
6. Does a significant association exist among overall duration of care and temperament type as
determined by the Keirsey Temperament Indicator II, self-efficacy scores as determined by
"Total Scale Scores" on the Caregiving Self-Efficacy Scale, and relationship to care
1. Participants, aged 18 and older, were volunteers from caregiver support organizations serving
the North Central Florida geographical area.
2. Data were collected in calendar years 2005-2006.
3. The Keirsey Temperament Indicator II was used to determine temperament type of
participants as caregivers.
4. The Caregiving Self-Efficacy Scale was used to determine perceived self-efficacy of
participants as caregivers.
5. The Caregiver Information Profile was used to obtain demographic information about
participants as caregivers.
6. Participants had served in a caregiving role for at least one month for someone with a chronic
or terminal illness as defined in this study.
1. Volunteers who participated in the study may not represent all members of the organizations
involved in the study.
2. Data collected during calendar years 2005-2006 may differ from data collected during other
3. The Keirsey Temperament Indicator II may not address all aspects of temperament among
participants as caregivers.
4. The Caregiving Self-Efficacy Scale may not determine all aspects of perceived self-efficacy
of participants as caregivers.
5. Demographic information obtained by the Caregiver Information Profile may not capture all
pertinent information about participants as caregivers.
6. The experiences of individuals serving as caregivers for different time frames or durations, or
for different types of illnesses, may differ from the experiences of participants in this study.
1. Volunteers who agreed to participate in the study were considered adequate to represent the
membership of their respective organizations.
2. Data collected during calendar years 2005-2006 were considered adequate for the purpose of
3. The Keirsey Temperament Indicator II was considered adequate to address temperament type
among participants as caregivers.
4. The Caregiving Self-Efficacy Scale was considered adequate to determine perceived self-
efficacy of participants as caregivers.
5. Demographic information obtained by the Caregiver Information Profile was considered
adequate to describe study participants as caregivers.
6. Individuals serving in a caregiving role for at least one month for someone with a chronic or
terminal illness as defined in the study were considered adequate as participants for the
purpose of the study.
Definition of Terms
Activities of Daily Living (ADLs). Emotional, spiritual, and medical needs provided by a
caregiver to a care receiver such as bathing and assistance with eating (Arno et al., 1999).
"Artisans." A temperament type under the Keirsey Indicators, Artisans are concrete in
communicating and utilitarian (practical) in implementing goals, and they can be highly skilled
in tactics, and finding multiple solutions to a problem. Artisans are gifted with tools, such as
language, a computer, thoughts or a paint brush (Artisans: Keirsey.com, 1996, para. 1;
Understanding the Four Temperament Patterns: The why, 2001, para, 7).
Care receiver. An individual who receives care from an informal caregiver.
Formal caregivers. Individuals in a health profession, such as primary physicians,
specialists, nurses, therapists, and counselors, who provide care for an individual in a formal care
setting such as a hospital, private practice, or medical facility.
"Guardians." A temperament type under the Keirsey Indicators. Guardians are concrete
in communicating and cooperative in implementing goals, and they can be highly skilled in
logistics (Heineman, 1995, para 23; Understanding the four temperament patterns, 2001, para 7).
"Idealists." A temperament type under the Keirsey Indicators. Idealists are abstract in
communicating and cooperative in implementing goals, and they can be highly skilled in
diplomatic integration (Heineman, 1995, para 21; Different drums: Kiersey.com, 1998, para.
Informal caregiver. An individual voluntarily caring for another person (care receiver)
including friends or relatives. The individual needing care experiences health problems or
disabilities, and needs assistance with grocery shopping, bathing, eating, and dressing. Informal
caregiving includes care provided by spouses, children, other family members, and friends in a
home or other informal setting.
Instrumental activity of daily living (IADLs). Tasks performed by a caregiver for a care
receiver such as helping with finances and helping to prepare meals (Arno et al., 1999).
"Rationals." A temperament type under the Keirsey Indicators, Rationals are abstract in
communicating and utilitarian in implementing goals, and they can be highly skilled in strategic
analysis (Heineman, 1995, para. 20; Understanding the four temperament patterns: The why,
2001, para, 7).
Self-efficacy. A component or construct of several health behavior and health education
theories that addresses individuals' personal beliefs that they can perform a specific behavior or
action, or that they can overcome temptations, barriers, or negative behaviors created by others
in their environment (Steffen et al., 2002).
Social support. The process of seeking or receiving emotional, instrumental, medical, or
monetary provisions to increase an individual's self-esteem or efficacy. Social support can be
formal, as in support from health care providers, or informal as from family and friends.
Temperament. The emotional disposition of an individual which, in turn, directs
personality and choices in terms of behavior (Heineman, 1995, para 20; Understanding the four
temperament patterns: The why, 2001, para. 7).
As the U.S. population continues to age, more individuals are becoming informal
caregivers for family members and friends with chronic and terminal illnesses. Much of the
expenses that caregivers incur are out-of-pocket. Caregiver stress causes individuals to neglect
their own health, potentially creating more serious health problems for the future. While all
caregivers encounter stress, individuals respond differently to the initial experience and to the
experience over time. To date few studies have explored the relationships existing among
temperament, self-efficacy, and caregiving. This study examined those relationships among
individuals serving as informal caregivers in the northeast Florida geographical area.
REVIEW OF THE LITERATURE
This study explored relationships among temperament, self-efficacy, and caregiving
among individuals serving as informal caregivers in the northeast Florida geographical area.
This chapter presents a review of literature related to these topics, factors that influence
individuals serving in the informal caregiver role, research that defines and describes each
temperament type, and research that examines self-efficacy in relation to disease as applied to
formal and informal caregivers.
Informal Caregivers and Factors Affecting Their Roles
Informal caregiving refers to care provided by family and friends and represents a part of
our nation's multiple medical solutions and problems. Most individuals who become informal
caregivers assume the role due to a traumatic event, catastrophic illness, progressive chronic
illness (e.g., Alzheimer's Disease), or death of one parent where the surviving parent needs care.
Caregiving situations vary because of family dynamics (i.e., younger versus older
caregivers), disease diagnosed (i.e., heart disease versus Alzheimer's Disease), time spent in the
caregiving role, and intensity of care required. For example, intensity of caregiving varies from
driving the care receiver to physician appointments to changing a catheter bag. Apart from the
primary caregiver, caregiving often involves a circle of extended family or friends, who may first
notice subtle changes in the person needing care. Caregiver distress often results from erratic
behavior of care receivers rather than the disease itself (Hebert, Levesque, Vezina, Lavoie,
Ducharme, Gendron, Preville, Voyer, & Dubois, 2003; Small, McDonnell, Brooks, &
Papadopoulos, 2002; Gottlieb & Rooney, 2004; K. Meneskie, Personal Communication via e-
mail, February 18, 2005; D. Rimkunas, Personal Communication via email, February 25, 2005).
For example, subtle changes in someone with early onset of Alzheimer's Disease may include
intact long-term memory but poor short-term memory, confusion, fear, and anger at loved ones.
Emotional, spiritual, and medical needs, assistance with activities of daily living (bathing,
dressing, transferring in and out of bed, assistance with eating) and instrumental activities of
daily living (assistance with finances, preparing meals, housekeeping, laundry) increasingly are
met by family and friends, including assistance by minor children, adolescents, and young adults
who may live in the home with the care receiver. Although difficult to assess, the monetary
value of services provided by these caregivers is substantial. Researchers predict that the value
of family caregiving will exceed that of nursing home care by a 2:1 margin in the new
millennium (Arno et al., 1999; University of Texas, School of Nursing at Houston, Center on
Aging, 2005; D. Gatsche, personal communication, September 22, 2004; E. Cairns, personal
communication, September 22, 2004).
Changes in the Caregiver Role
While advances in health and medical care have increased the lifespan for adult citizens,
financial provisions for the elderly have not advanced at the same pace. When the Medicare
system took effect in the 1960s, medical costs were projected and established for those aged 50
and older. However, costs and provisions to assist those living into their 80s and beyond did not
fit into the long-term vision for the original system. The current health care system does not
meet the needs of elderly individuals who suffer from chronic and terminal illnesses (Arno et al.,
Nationwide, as the population ages, informal caregiving represents an essential asset.
More than 75 million baby boomers will be age 65 and older in the 21st century with the fastest
growing segment of the elderly population, those aged 80 and older, increasing from 122,000 at
the beginning of the 20th Century to more than 4.3 million in 2000 (Adler, 2002; Partnership for
Solutions, 2002; American Association for Geriatric Psychiatry: Health care professionals, 2004,
para. 1). In 1995, Florida's population of elderly of those age 65 and older made up 19% of the
total population the largest proportion of any state and the percentage of this population was
projected to increase to 26% of the total population by 2025. Therefore, Florida is a focus of
future concern for the state's health care system (Goulding, Rogers, & Smith, 2003).
Currently, between 40% and 80% of caregivers live with the care receiver. Almost half of
caregivers serve in that role for at least five years, with eight years as the average time spent in
caregiving because the progression of the care receiver's illness often is slow. The elderly
population, those aged 65 and older, numbered 35 million in 2000, or approximately 12.4% of
the U.S. population. The same population group is expected to increase by more than 20% by
2030, with about 70 million individuals aged 65 and older who will be alive in the U.S. by 2030
(Goulding et al., 2003; National Family Caregivers Association: Caregiving Statistics, 2000,
para. 1; U.S. Department of Health & Human Services, Administration on Aging, 2001; National
Family Caregivers Association: Family Caregivers and Caregiving Families 2001, 2001, pg. 4,
According to the National Family Caregivers Association (George & Kansas, 2001, para.
7), approximately 54 million Americans, or about 27% of the total population, were involved in
family caregiving in 1999, but less than 19% of caregivers felt they had received adequate
training to deal with the new role (National Family Caregivers Association, Family Caregivers
and Caregiving Families, 2001, pg. 12, para. 1). Because women's status in the home and
workplace has changed during the past 50 years, the impact of caregiving disproportionately
affects women's roles and duties. While 35% to 40% of men age 40 and older actively
participate in caregiving, women most often fill the role of primary caregiver. Approximately
70% of primary caregivers are women aged 60 or older, and 33% of female caregivers aged 36 to
50 are raising minor children while caring for an elderly parent or family member (National
Family Caregivers Association, Caregivers Association, Family Caregivers and Caregiving
families, 2001, pg. 4, para. 4; National Family Caregivers Association, Education and Resources,
Yes, I am a Family Caregiver, Caregiver Resource, 2002, para. 7).
The number of individuals with chronic conditions also continues to increase. In 2000,
approximately 50 million individuals with terminal or chronic illnesses received some type of
care from a family member or friend. By 2030, an estimated 150 million people, or 48% of the
population, will be living with a chronic condition (U.S. Department of Health & Human
Services, Administration on Aging, 2001; Partnership for Solutions, 2002). Of individuals aged
65 and older, 80% will have at least one chronic illness, and 50% will have two or more
(Goulding, et al., 2003). Currently, about 20% of individuals in the U.S. have multiple chronic
illnesses. Most of the emotional and financial burden of caring for the chronically ill falls
outside the health care system. The chronically ill, who often need long-term care, may receive
mixed advice about medications, treatments, and health care regimens from formal caregivers
(physicians, nurses), thereby increasing the cost of managing their illness and increasing the
stress on informal caregivers (Partnerships for Solutions, 2002).
Early in the caregiver role, depending on the type (chronic versus terminal) and duration of
disease or illness, caregiver responsibilities can be overwhelming. When a care receiver has been
diagnosed with cancer, researchers report that the months immediately following an initial
hospitalization for surgery or treatment prove the most difficult. However, caregivers for those
with dementia or Alzheimer's Disease report that stress increases as the disease progresses
(Weitzner, Jacobsen, Wagner, Friedland, & Cox, 1999; Alzheimer's Caregiver Support Online,
2002, para. 6). In a 1997 study of caregivers, 37% reported they received no paid assistance,
35% of caregivers indicated they did not find enough time for themselves, and 29% needed
assistance in managing physical or emotional stress (Wagner, 1997; National Alliance for
Caregiving & AARP, 2004, pg. 18, para. 1).
On average, 41% of caregiving duties require "constant care," which means at least eight
hours of daily full-time care. Family caregivers often experience depressive symptoms,
characterized by anxiety, which is usually accompanied by a decrease in caregiver healthy eating
and exercise habits. The National Family Caregiver Association (2001) reports a decline in
healthy eating behavior from 70% to 59%, and a decline in regular exercise from 61% to 30%,
from before becoming a caregiver to after becoming a caregiver. Caregivers also reported a
decline from 70% to 47% in consulting a physician about their own health (National Family
Caregivers Association, Who are family caregivers?, 2001, para. 5). Caregiver quality of life can
be as bad as, or worse than, that experienced by the care receiver, especially when the caregiver
also experiences chronic or terminal illness. Caregivers may experience an increase in emotional
and financial difficulties as the disease or illness of the care receiver progresses. Many
caregivers are forced to decrease their workload or quit work entirely to manage caregiver
responsibilities (Weitzner, et al., 1999).
Medical Costs of Caregiving
Society generally underestimates and under-appreciates the value of caregiver time, lost
wages, and medical costs for their own medical needs. As an example of government attempts to
address this need, the Clinton Administration proposed a $6 billion package for long-term care
solutions of chronically ill patients and their families.
The U.S. spends more than any other country on the health care of individuals 65 and older
($12,100 per person). Currently, health care by informal caregivers represents an estimated $196
billion (approximately 18%) of total national health care spending, currently at about $1 trillion.
However, informal caregiving is not calculated in the national health care budget. Caregiver
financial burden increases as the disease/illness progresses and as caregiver demands increase
(Goulding et al., 2003; Chou, 2000; Yun, Rhee, Kang, Lee, Bang, Lee, Kim, Kim, Shin, & Hong,
2004; Arno, et al., 1999).
Florida has traditionally been known as a state with a large elderly population. Many
elderly people retire to the state, thus requiring medical care and attention for end-of-life years.
Florida ranks fourth in the nation for number of informal caregivers (more than 1.6 million),
fourth for annual informal caregiver hours per billion (1.7), and fourth for annual dollar value of
informal caregiver hours (about $15 billion). California (first), Texas (second) and New York
(third) are ranked above Florida in all three categories (National Family Caregivers Association
and the Albert Einstein College of Medicine, 2002).
Costs of medication and treatment for the chronically ill are often paid out-of-pocket.
Expenses are absorbed by family members or friends who eventually become informal
caregivers. These unexpected expenses can become part of the total caregiving burden (Scholte
op Reimer, de Haan, Rijnders, Limburg, & van den Bos, 1998). Out-of-pocket expenses
multiply incrementally with the number of chronic conditions, such as $130 per year for
prescription drugs for one chronic condition versus $930 annually for five or more conditions
(Partnership for Solutions, 2002). Some national estimates of the value of annual home care
services provided by caregivers are close to $200 billion (National Family Caregivers
Association, Who are family caregivers?, 2004, pg. 2, para. 4).
Caregiving affects the U.S. health care system and caregivers' personal health. Caregivers
with chronic illnesses report higher levels of stress and have a 63% higher mortality rate than
same-age counterparts who are not caregivers (Schultz & Beach, 1999). Individuals who
provide care in excess of 30 hours a week experience more symptoms of stress and anxiety. Two
studies reported that intensity and difficulty of care required by cancer and dementia patients
dictates the number of weekly hours and amount of out-of-pocket medical and other costs
provided by family to the sick individual (Hayman, Langa, Kabeto, Katz, DeMonner, Chernew,
Slavin, & Fendrick, 2001; Langa, Chewnew, Kabeto, Herzog, Ofstedal,l Willis, Wallace, Mucha,
Straus, & Fendrick, 2001).
Alzheimer's Disease is probably the most expensive long-term disease to treat, with
current annual expenses estimated as high as $29.1 billion, and with approximately 50% of those
with Alzheimer's Disease or dementia living at home (Small, et al., 2002; Hebert, et al., 2003).
In addition to Alzheimer's Disease, other forms of dementia, such as those associated with stroke
and Parkinson's disease will create a greater need and increased stress among informal
caregivers (Vitaliano, Zhang, & Scanlan, 2003). Caregivers for those with Alzheimer's Disease
miss about five days of work during a six-month period and average about 85 hours of care each
week (85 + 60.2). Caregivers for those with more severe symptoms miss correspondingly more
work. Researchers in this study found that intervention (home health care assistance) and its
costs depend on stage of the disease and other factors such as when interventions are
implemented (Small et al., 2002, p. 323).
Within the past few years, researchers have assessed benefits and costs of informal
caregiving. Results showed that as time spent in the informal caregiver role increased, the
"marginal benefits" decreased and the "marginal costs" increased. When asked what type of
assistance they would most appreciate, most caregivers indicated that they needed assistance
with housework (Smith & Wright, 1994; van den Berg, Brouwer, Exel, Koopmanschap, 2004).
Lost income of caregivers, out-of-pocket expenses, and additional health care costs figure
into estimating total costs of informal caregiving for someone with dementia (Moore, Zhu, &
Clipp, 2001). Researchers calculated the total annual costs of informal care for dementia patients
coming from lost wages of caregivers at $10,709 and the cost of caregiver time at $6,295. These
calculations included assistance with all types of services (i.e., Activities of Daily Living and
Instrumental Activities of Daily Living) provided to care receivers (Moore, et al., 2001).
A study of Korean caregivers caring for those with cancer, showed that caregivers' quality
of life decreased incrementally with the amount of family savings that were depleted, which
included loss of a job or other source of income (p<0.0001) (Yun et al., 2004).
In a study based in Arkansas that explored easing caregiver burden, nursing home
placement, and caregiver-care recipient relationships, researchers found that caregivers in the
treatment group, who were paid by the researchers to participate in the study, seemed to be less
stressed and reported better physical and emotional health than those who were not paid (Foster,
Brown, Phillips, & Carlson, 2005).
Health Risks of Caregiving
Many factors influence the health behavior and health status of caregivers, including the
type and severity of the care receiver's illness and the physical and mental restrictions of both
caregiver and care receiver. About one-third of caregivers of an elderly individual will suffer
some physical or emotional distress, more than 10% will have some mental or health problems,
and many will also experience chronic fatigue (National Alliance for Caregiving & AARP, 2004;
Foster et al., 2005; Teel & Press, 1999).
Caregiver wives report higher levels of depression and lower levels of self-esteem than
caregiver daughters or daughters-in-law (Collins & Jones, 1997; Pinquart & Sorensen, 2005).
Rates of depression among caregiver women range from 28% to 55% versus 14% to 16% among
the general population of women. While male and female caregivers differ in their coping
mechanisms, caregiver wives are usually more emotionally involved than caregiver husbands.
As time progresses, women develop and adopt more "problem-solving coping strategies similar
to those used by husband caregivers" (Wright, 1997, p. 272). Coping strategies in the caregiver
role include drawing on a network of individuals for social support and using different methods
and sources to learn more about the disease or disorder. However, during the course of a chronic
disease, available social support can decline over time (Glanz, Reimer, & Lewis, 2002).
Compared to depressed individuals in the general population, most caregivers report fewer
visits to physicians and other health care professionals, greater use of psychotropic medications
for anxiety, use of alcohol and tobacco as coping mechanisms, and eating and exercising less
after assuming the caregiver role (Collins & Jones, 1997; Wright, 1997). Even when a caregiver
had previously discussed his or her mental health concerns before the care recipient's diagnosis,
less than half (46%) were willing to use mental health services and, as stated above, used more
drugs (Vanderweker, Laff, Kadan-Lottick, McColl, & Prigerson, 2005). However, some
caregivers increase the number of self-visits to a physician, and increase the amount of drugs
used, perhaps to restore a balance in healthy sleeping, and decrease exercise and leisure time
Placement of the care receiver in a caregiving facility, such as a nursing home, may
precede a significant decline in both the care receiver's and the caregiver's health. The decline
can include lack of sleep and improper rest. Caregivers with more health problems are more
likely to serve less time in the caregiving role. The study also found that individuals caring for
those with a form of dementia (i.e., Alzheimer's Disease) are at greater risk of declining health
(Scharlach, Midanik, Runkle, & Soghikian, 1997).
The progression of symptoms and overall decline in health of the care receiver may leave
caregivers feeling helpless and confused, especially when the caregiver's own health begins to
decline, and as work and other family obligations increase (Scharlack, et al., 1997; K. Mladinich,
personal communication, August 25, 2004).
Once a caregiver is placed in assisted living, spousal caregivers continue to function in a
caregiver role by visiting the spouse almost daily. Not surprisingly, the health of spousal
caregivers can continue to decline even after their significant other enters a care facility because
this population of caregivers tend to be older than age 65 (Wright, 1997). Most care recipients
with dementia are institutionalized, on average, from two to six years after diagnosis (Mausbach,
Coon, Depp, Rabinowitz, Wilson, Arias, Kraemer, Thompson, Lane, Gallagher-Thompson,
2004). However, most studies on caregivers and care recipients with dementia focus on the
Caucasian population. This observation about the literature on caregivers is relevant because
many Caucasian caregivers and care receivers have access to better health care, and a different
type of support system than Hispanic or African American caregivers. Minority caregivers tend
to keep family members in the home longer even though the care recipient may have more
debilitating illnesses. Therefore, other studies should focus on a culturally diverse population to
assist with needs and demands of caregivers helping with those with dementia.
Relationship between caregiver health and level of caregiving.
Other studies show that individuals providing moderate to high levels of caregiving, when
compared to non-caregivers, are less likely to maintain adequate health behaviors or use
preventive services, such as remembering physician's appointments, getting prescriptions
refilled, and getting adequate rest (Burton, Newsom, Schultz, Hirsch, & German, 1997).
Those providing a high level of care reported not getting enough sleep (24%), not getting
enough exercise (29%), and not resting adequately when sick (14%) were all significant at the
p<.0001 level. Those providing higher levels of care also were more likely to forget to take
medications (19%, p<.001) and less likely to make medical appointments (7%, p<.05). The
relationship between perceived sense of control and personal preventive health behaviors was
significant at all levels of caregiving and at both levels for sense of control under variables such
as "not getting enough rest" (P<0.0001) (Burton, et. al., 1997). For the population in this study,
confidence, which is defined as sense of control, in the caregiver role seems to influence health
behaviors for the caregiver when a caregiver's feelings more powerless, their personal care
Even if care receivers have insurance that provides long-term care, studies confirm that
approximately 50% of caregivers show no change in the care they provide, indicating that even
with formal caregiving options, the informal caregiver continues to provide essential assistance
to the care receiver. Essential assistance included daily visits to the care receiver, feeding,
bathing, dressing, and providing prescription medications. These out-of-pocket expenses are part
of the overall caregiver burden and contribute to caregiver stress (Cohen, Miller, & Weinrobe,
A five-year study assessing transitions in caregiver roles found that about 50% experienced
some type of change, including the death of a spouse or a spouse being placed in a long-term
facility (Burton, Zdaniuk, Schulz, Jackson, & Hirsch, 2003). Individuals who changed from
non-caregiver status to heavy caregiver status involving meeting the most basic physical
functions showed more depressive symptoms and were more likely to exhibit unhealthy
behaviors (5.59 vs. 6.94, p < .01) (Burton, et al., 2003).
Schultz and Beach (1999) explored the association between caregiving and mortality in
sub-groups of caregivers "physiologically compromised and exposed to varying levels of
caregiving strain" (p. 2216). After four years of follow-up, deaths among those providing care
and those who had reported physical strain (31 of 179, 17.3%) were significant (p<.025).
When assessing stress levels pre- and post-bereavement among caregivers, those in the
control group (non-caregivers) recorded higher post-bereavement rates of depression.
Depression remained unchanged for caregivers with high levels of strain during the caregiver
role. African-American caregivers had significantly lower depression symptom scores than
Whites at both pre-bereavement and post-bereavement (Schultz, Beach, Lind, Martire, Hirsch,
Jackson, & Burton, 2001).
Another study (N=52) measured socio-demographic, physical, and psychosocial
characteristics of family caregivers for stroke survivors including satisfaction with health care
resources, social support, activities of daily living, and instrumental activities for daily living
(Grant, Weaver, Elliott, Bartolucci, & Giger, 2004). The study found Caucasians at about four
times higher risk for depression compared to their minority counterparts (i.e., African American
and Hispanic individuals). Results suggest that caregivers who experience a decrease in their
"sense of belonging with others" may be at higher risk of distress in the weeks after the care
receiver returns home from a hospital stay from stroke-related complications (Grant et al., 2004).
In a study exploring predisposing, enabling, and need components of caregivers' use of
care services, caregivers aged 65 years or older showed higher depressive symptoms and used
more formal services, such as in-home care services than younger caregivers (Bookwala,
Zdaniuk, Burton, Lind, Jackson, Schulz, 2004). Perceived sense of mastery of formal services
related positively to self-efficacy among caregivers (Bookwala et al., 2004).
With chronic illnesses, such as heart disease or heart failure, caregiver quality of life may
be compromised. Caregivers, and care receivers with heart failure, reported their overall quality
of life, measured as "sense of well being." Care receivers rated their quality of life lower than
their caregivers prior to being hospitalized (p-0.035) (Luttik, Jaarsma, Veeger, & van
Veldhuisen, 2005). Caregiver quality of life was seen as fairly stable during the course of the
study, while care receiver quality of life fluctuated with the decline of health and their
Care receivers with cancer/terminal illness.
Losing a family member to a terminal illness is a stressful life event, but if an individual is
a caregiver of the terminally ill family member, the time leading up to the loss can be an added
stressor as well. End of life issues and experiences may involve hospital care and Hospice care
(Haley, LaMonde, Han, Burton, & Schonwetter, 2003).
Caregiver distress was examined as it related to amount of time spent assisting advanced
cancer patients (those estimated to have less than six months of life remaining). Caregivers with
less than a high school education reported significantly more depression, more tension, and more
total mood disturbances. In general, caregivers of cancer patients experience more physically
debilitating symptoms in addition to depression (Cameron et al., 2002).
Caregivers of patients with terminal illness were recruited for a study on bereavement
(Ferrario, Cardillo, Vicario, Balzarini, & Zotti, 2004). Caregivers who were asked to complete
an instrument with scales that measured caregiver strain, cognitive function, and life satisfaction,
were assessed at three, six and 12 months after the death of their loved one. Results showed that
caregivers who perceived more "emotional burden" were more likely to indicate less satisfaction
with life overall. Satisfaction with life decreased significantly from three months to 12 months
(P=0.001). The relationship the caregiver had with the spouse was also a predictor of
"bereavement maladjustment after 12 months" (P< 0.001) (Ferrario et al., 2004) if the
caregiver reported a more positive relationship with their spouse, they tended to report higher
levels of life satisfaction post-bereavement.
In another study, caregivers were recruited to assess the impact for caregiving of a
terminally ill individual who recently had died (Brazil, Bedard, Willison, & Hode, 2003).
Caregivers were asked to identify ADLs and how long they had provided assistance for specific
tasks. Most of the caregivers in this study were female, the spouse of the patient, lived in the
home with the care receiver, and did not work outside the home. The number of visits to a
general practitioner and caregiver age were significant determinants of high caregiver burden
Caregivers tending to terminally ill individuals, especially those in palliative care or
hospice care, experience different levels of distress and seek different levels of support. In
palliative care, caregivers anticipate the death of their loved ones. When caregivers were
presented with a support intervention (i.e., learning how to cope and exchanging information
with others and in seeking formal or informal support), caregivers who were employed full-time
were less likely to accept the support intervention (P=0.006) (Harding, Higginson, Leam,
Donaldson, Pearce, George, Robinson, & Taylor, 2003).
Spousal caregivers of patients with end-stage lung cancer and dementia (diagnosed with
less than six months of life remaining) were interviewed to assess daily duration of care, total
time in the caregiving role, caregiver self-reported health, self-reported (perceived) benefits of
caregiving, and level of care receiver impairment (i.e., abilities to perform ADLs and IADLs)
(Haley et al., 2003). The only variable that was significantly associated with caregiver
depression was behavioral problems of the care receiver (p<0.05). No other variables, including
patient diagnosis and hours of care each week, were significantly associated with caregiver
Caregivers for patients diagnosed with metastatic breast cancer were assessed for financial
burden for palliative and other treatment (Grunfeld, Coyle, Whelan, Clinch, Reyno, Earle,
Willan, Viola, Coristine, Janz, Glossop, 2004). Caregivers who were depressed at the start of the
"terminal" period, as defined in this study, were equally as depressed at the start of the palliative
period (p=0.02). The amount of money spent on prescription drugs was the best indicator of
financial burden among caregivers, especially those without additional health insurance coverage
(p=0.04) (Grunfeld et al., 2004).
Similarly, researchers studied caregiving during the terminal phase of illness for the care
receiver (Brazil, et al., 2002). The last three months of life proved to be the most critical due to
the need for increased patient care and an increase in the need for outside services, such as
nursing care assistance. Caregiver burden was attributed to the impact "caregiving has on the
personal health of caregivers and on the extent to which caregiving interrupts the usual activities
of the caregiver" (Brazil et al., 2002, p. 381).
Among caregivers with partners who were cancer patients, the impact of caregiving on
disrupted schedules and caregiver self-esteem decreased in the six months following the care
receiver's discharge from a healthcare facility. As the care receiver's health improved, so did the
health of the caregiver (Nijboer, Tempelaar, Sanderman, Triemstra, Spruijt, van den Bos, 1998).
Care receivers with mental illness.
Those caring for individuals with any kind of mental illness may carry the stigma of their
loved one and experience difficult care management, which include factors such as caring for an
older individual with mental illness and living in the home with the care receiver (Groff, Burns,
Swanson, Swartz, Wagner, & Tompson, 2004).
Caregivers of care recipients diagnosed with schizophrenia were assessed for stress and
stress-related symptoms. Interviews were provided at baseline, then four, eight and twelve
months after baseline. Unlike most caregiver studies, almost half (41%) of the caregivers were
male and 63% were Black. Results of the study indicate that those who lived with the care
recipient were more likely to have higher self-reported levels of strain (p<0.001), and the more
days the care receiver spent in outpatient care, the less strain on the caregiver (Groff et al., 2004).
Care receivers with dementia
Well-documented studies of caring for someone with dementia have shown that the
caregiver is at greater risk of having a psychological disorder than those in other caregiving roles
and in the same age-group population (McConaghy & Caltabiano, 2005; Covinsky, Newcomer,
Fox, Wood, Sands, Dane, & Yaffe, 2003). Dementia is a debilitating disease with an often long-
term progression. While care receivers may be physically capable, their mental health
deteriorates gradually. Acknowledging the risks to caregivers of dementia patients is important
for clinicians and formal caregivers because mental and physical debilitation is complex and
affected by many factors (De Frias, Tuokko, Rosenberg, 2005; Covinsky et al., 2003).
Caregiver wives of Alzheimer's Disease patients showed lower overall morale than
caregiver husbands (Collins & Jones, 1997; Wright, 1997; Eisdorfer, Czaja, Loewenstein,
Rubert, Arguelles, Mitrani, Szapocznik, 2003). In general, caregivers of those with any form of
dementia (i.e., Alzheimer's Disease) are more likely than non-caregivers to experience higher
levels of depression, anxiety, and poorer health (Mausbach et al., 2004).
Caregivers and care receivers who had been diagnosed with dementia were assessed for
health in the "paired relationship." Caregivers who reported higher "caregiver burden" were
more likely to be in poorer psychological health (P < 0.05) (McConaghy & Caltabiano, 2005).
In a longitudinal study of individuals caring for someone diagnosed with dementia,
researchers examined the use of a physician by caregivers (Baumgarten, Battista, Infante-Rivard,
Hanley, Becker, Bilker, & Gauthier, 1997). Caregivers need psychiatric assistance 10 times
more often than non-caregivers (p=.008), but they did not necessarily seek physician care at a
higher rate than non-caregivers.
Caregivers' self-reported health status and the relationship between caregiver and care
receiver mental health, were assessed to compare demands on caring for older individuals with
dementia (De Fries et al., 2005). Older caregivers had higher self-esteem and had more support
from family members. Caregivers who were older, physically healthier, and who reported fewer
health issues from being a caregiver, were more likely to report having poorer health if the care
recipient was male (p<0.05) (De Fries, et al., 2005).
Few studies have assessed the characteristics of both caregivers and care receivers with
dementia to predict depressive symptoms among the caregivers. Patient traits associated with
caregiver depression include less education, being younger, being male, being Hispanic, and
depending more on the caregiver for Activities of Daily Living (ADLs) (P<.001). Caregiver
characteristics significantly associated with their depression include their relationship to the care
receiver (significant for being a wife versus husband), hours spent caregiving each week, and
caregiver health (Covinsky et al., 2003). Hispanic caregivers displayed higher rates of
depression, possibly due to expectations of the caregiving role, denial about care receiver
disease, and language barriers with formal care providers.
Wackerbarth and Johnson (2002) studied essential information and support needs of family
caregivers for individuals with Alzheimer's Disease and dementia. Respondents reported a need
for information: general information (i.e., how to plan), diagnosis and treatment information
(i.e., results of tests, help with treatment decisions), and legal information relating to financial
issues (i.e., health plan coverage, legal terms). When compared to men, women used more
emotion-focused coping strategies, which are associated with higher levels of negative health
outcomes (Wackerbarth & Johnson, 2002). Also, female caregivers experienced more stress and
poorer overall health than male caregivers (Nijboer, Triemstra, Tempelaar, Mulder, Sanderman,
& van den Bos, 2000; Dilworth-Anderson, et al., 2004).
Degree of Support and Caregiver Distress
Caregiving and health risks
Biochemically and psychologically, caregiving produces a great deal of stress. One study
determined that individuals under stress have lower antibody response to vaccinations and lower
counts of "natural killer cell activities" (Davis, Weaver, Zamrini, Stevens, Kang., & Parker,
2004, p. 91). When looking at biochemical changes, specifically cortisol production in
caregivers of individuals with Alzheimer's Disease, significant differences occurred between
those in a non-caregiving role compared to caregivers (P=0.006). Cortisol levels were
determined through saliva samples in caregiver wives. Approximately 46% of total caregiver
wives were taking psychotropic medications, but 80% of total caregiver wives said their overall
health was "good" or "excellent" (Davis et al., 2004).
Caregiving especially can affect functioning of the immune response and heart rate
reactivity, blood pressure levels, and mortality among some older spousal caregivers (Dilworth-
Anderson, Goodwin, & Williams, 2004). In a meta-analysis that assessed stress hormones as a
measure of physiological function in caregivers, researchers found that some hormones affect
cardiac functioning (cortisol, epinephrine, norepinephrine), and others affect immune response
(norepinephrine, prolactin) (Vitaliano et al., 2003). Therefore, the longer individuals are in the
caregiving role, the longer they are exposed to stress hormones that affect cardiac and immune
function. Some studies have defined primary stressors as those associated with the health of the
care recipient, while secondary stressors were associated with being a caregiver apart from the
caregiving role (i.e., dealing with workplace situations and dealing with family members)
(Amirkhanyan & Wolf, 2003).
With the growing number of individuals living well into their 80s and more individuals
separated from their elderly family members, more older individuals are caring for those in their
same age group or even older (Ekwall, Sivberg, & Hallberg, 2004). In a Swedish research study
on older informal caregivers (age, M=81.8 years), researchers found that women who had higher
levels of involvement in supervising care for another individual (p<0.001), and assisting with
Instrumental Activities of Daily Living (IADLs) was positively associated with better quality of
life for the older caregiver (Amirkhanya & Wolf, 2003).
Caregiving compromises one's mental and physical health, including depression, low self-
esteem, and possible substance abuse, all of which can be indicators of potential frustration and
possible care receiver abuse. In a study assessing potentially harmful behavior (Beach, Schultz,
Williamson, Miller, Weiner, & Lance, 2005), informal caregivers and their care recipients were
asked how often help was needed. For example, caregivers were asked how often they had to
provide assistance with Activities of Daily Living (ADLs) or IADLs to the care recipient, and the
care recipients were asked how often they asked for assistance for the same tasks (Beach et al.,
2005). Care recipients who reported needing more ADL and IADL assistance and who were in
poorer health were more likely to be the recipients of harmful behavior from the caregiver (i.e.,
yelling, insulting) (Beach et al., 2005).
Logsdon and Robinson (2000) suggested five essential characteristics of social support that
may influence one's ability to effectively match recipient needs and expected support from
someone with whom the caregiver may share an intimate relationship. Negative support or
nonsupport among women caring for those with dementia had a negative impact on caregiver
health. Women cited "unmet expectations" for support as the "most pervasive difficulty"
(Neufeld & Harrison, 2003, p. 327), and their responses were influenced by type of relationship
between caregiver and helper. Negative support from family or friends included conflict
between the caregiver and others in determining the health status of the care receiver and
criticism of caregivers' decisions in caring for care recipients.
Caregivers and family support
Family systems and family dynamics play an important role in determining caregiver
emotional function (Mitrani, Feaster, McCabe, Czaja, & Szapocznik, 2005). Using the
"structural family theory," Hispanic and non-Hispanic caregivers (N=177) were assessed at three
intervals after starting in the caregiver role (6, 12, and 18 months) based on consistent and
repeated patterns of interactions within the family unit. Researchers tested items for an effective
measure of family and caregiving. Cultural dynamics plays an important role in the Hispanic
population, especially those involved in caregiving. Enmeshment, or the interference and
boundary violations between family members, is an example of cultural dynamics assessed in
this study. Hispanic family members often use enmeshment as a "protective factor," so as not to
deal with certain issues as a caregiver, such as stress levels and difficulty of care it is assumed
that Hispanic men and women will care for sick and elderly family members, regardless of their
social or economic status (Mitrani et al., p. 452).
Caregiver employment status outside the home and informal caregiver burden were
predictors of caregivers' level of mental health (Cannuscio, Colditz, Rimm, Berkman, Jones, &
Kawachi, 2004). The correlation was observed among women who do not work outside the
home, who were caring for a spouse, and had high levels of stress. Similarly, the association
between women who do work and provide spousal care were also at high stress levels. Stress
levels were not as high among caregivers of parents. The group with the highest risk for
depression was women who provided more than 36 hours a week of care and who had few social
ties (Canniscio et al., 2004).
One predictor of caregiver burden is the ease with which care recipients can follow and
adhere to health care regimens. Perceived pressure in the caregiving role correlated with "more
psychological distress" (Pott, Deeg, Van Dyck, & Jonker, 1998, p. 47). In a similar study on
caregiver burden and how well care receivers gave attention to their own health care, a positive
correlation was found between caregiver's personal resources, self-esteem, general social
support, and caregiver burden (Greenberger & Litwin, 2003). Studies assessing those
individuals who cared for Alzheimer's disease patients showed that, when outside support
increased in dealing with patient problem behavior, negative reactions to the behavior by the
caregiver decreased (Hebert, et al., 2003).
In a study of patient-caregiver groupings, with patients in advanced stages of cancer,
researchers found that "elevated levels of anxiety in patients" were directly associated with
"higher levels of anxiety in their caregivers" and that caregivers felt they were concerned for
their mental health after the cancer diagnosis of the patient (Harding, Higginson, & Donaldson,
2003, p. 640; Vanderwerker, et al., 2005). This study showed a direct relationship between
decline in informal caregiver psychological health and increased caregiver burden with patient
pain and anxiety.
Cultural and Ethnic Factors in Caregiving
Research on caregivers has traditionally focused on White (Caucasian) individuals.
Minorities in the United States experience at least twice the rate of chronic illness, including
dementia, possibly due to the lack of adequate health care, financial resources, medical choices,
and information about their care (Neary & Mahoney, 2005). African Americans historically
have extended the caregiving role to relatives and friends outside the immediate family to
survive "generations of racial oppression and economic hardships" (Hamilton & Sandelowski,
2003, p. 665). Cultural rituals, religious experience, and traditions (i.e., normative experience)
affect those in the caregiver role and those who require care (Koffman & Higginson, 2003; Roff,
Burgio, Gitlin, Nichols, Chaplin, & Hardin, 2004).
When compared to their White counterparts, African Americans have a lower life
expectancy and were more likely to be in poorer physical health (69.6 years for African
Americans versus 76.4 years for Whites) (Kim, Bramlett, Wright, & Poon, 1998; Pinquart &
Sorensen, 2005). Minority caregivers are also more likely to be younger, single, an adult child, a
cousin, a grandchild, provide longer care time in the home, and be female (Pinquart & Sorensen,
2005; Neary & Mahoney, 2005).
When dealing with those in the caregiver role, Hispanics and African Americans
demonstrate different health communication issues and barriers than Caucasians. Hispanics
experience language barriers with those who are formal health care providers, lack of
information about the illness/disease, and issues involving acculturation. Proficiency of
caregivers in using the English language was significant if the elderly disabled person receiving
care attempted to use formal health care services (Wright, 1997; Calderon-Rosado, Morrill,
Chang, & Tennstedt, 2002; Neary & Mahoney, 2005).
Cultural socialization in African American communities helps them better deal with
stressors in the caregiver role, as well as extended exposure to caregiving as children, teenagers,
and eventually as adults. This socialization is more prevalent in the African American and
Hispanic/Latino communities, thereby presenting a higher propensity for family involvement.
This socialization is less prevalent in Caucasian households because of different types of
socialization and smaller family units (i.e., due to lower birth rates and higher divorce rates)
(Dilworth-Anderson, et al., 2004; Pinquart & Sorensen, 2005; Ayalong, 2004).
Based on a study of Latina and Caucasian caregivers of individuals with Alzheimer's
Disease, researchers found that Latina caregivers tend to be younger, less educated, had lower
incomes, and provided more care on a daily basis than their Caucasian counterparts (Coon,
Rubert, Solano, Mausbach, Kraemer, Arguelles, Haley, Thompson, Gallagher-Thompson, 2004).
Using the Resources for Enhancing Alzheimer's Caregiver Health (REACH program), Latina
caregivers reported having more positive experiences and participating in more prayer and
religious activity than their Caucasian counterparts (p<.001), even though the care recipient
seemed to have more behavior problems and had more impairments.
In a study using REACH data with a similar Latino population and length in the caregiver
role, ethnicity was assessed as a possible factor to determine time to institutionalization
(Mausbach et al., 2004). Latina women have "more days of survival" in the caregiver role than
Caucasians, from time of diagnosis to institutionalization, but Latinas who viewed caregiving as
less positive did not feel as confident in their role compared to other Latina women who felt
more confident as caregivers (P=.044) (Mausbach et al., 2004, p. 1080).
How an individual is socialized to view caregiving represents another important factor
when studying race and ethnicity among caregivers (Roff et al., 2004). Research on the
resources for REACH program assessed differences in positive attitudes in the caregiver role and
what contributed to these positive experiences among African Americans and Caucasians. As
predicted, African American caregivers reported more positive aspects as caregivers (p<.001),
were lower in SES, reported lower anxiety levels, and tended to be more religious than
Caucasians (Roff et al., p. P188).
Studying cultural values and beliefs among informal caregivers, researchers found that no
differences existed between African Americans and Whites in gender and education (Dilworth-
Anderson, Brummett, Goodwin, Williams, Williams, & Siegler, 2005). However, African
Americans were more likely to "adhere more strongly to cultural reasons for providing care,"
and females in this group were more likely to adhere to providing care (Dilworth-Anderson, et
al., 2005, p. S259). The researchers concluded that differences in cultural socialization within
the African American community created strong beliefs, traditional values, and attitudes about
caring for dependent family members. They also found that African Americans with a higher
level of education were less likely to adhere to these traditional values. This finding suggests
education and income may have a significant impact on caregivers in several ways. First, those
with higher levels of education have better jobs with health insurance benefits and therefore have
better access to health care benefits. Second, African Americans who have better jobs may also
have the means to provide in-home care or assisted living care for older family members.
African American caregivers were less likely than Whites to place the older adult in a
nursing home or similar facility, regardless of illness, although both groups experienced similar
stressors in their caregiving roles (Stevens, Owen, Roth, Clay, Bartolucci, & Haley, 2004;
Wolinsky, Callahan, Fitzgerald, & Johnson, 1992; Gaugler, Leach, Clay, & Newcomer, 2004).
For those African American individuals who were placed in a nursing home with Alzheimer's
Disease or some form of dementia, contributing factors included being a caregiver of males,
higher caregiver burden, and care receivers who were more "cognitively impaired" (Gaugler et
African American caregivers reported less stress in caring for an elder with dementia,
although their White counterparts had higher family income and tended to be older (M=72.2 for
Whites vs. M=55.2 for African Americans; P < .0001) (Stevens et al., 2004). Caregivers with
higher income and who were older tended to place chronically ill family members in a nursing
home at a higher rate. Therefore, literature on African American and Hispanic caregivers
suggests it is important to evaluate and assess how they cope with stress effectively in specific
African Americans recorded significantly lower mental health rating scores than
Caucasians, and Caucasians were more likely to perceive their overall health as good or excellent
compared to their African American counterparts (Kim, et al., 1998). When controlling for
education level, this difference decreased dramatically for both groups. Age, gender, education,
income, physical health status, and self-perceived level of health were significant predictors of
health behavior. African Americans participated less in regular physical activity, but the
researchers operationally defined physical activity as "playing tennis, jogging, and hiking;" not
typical exercises for many ethnically diverse groups.
Age of African American caregivers also affected the degree of depression, with younger
caregivers showing higher rates of depressive symptoms (Wright, Hickey, Buckwalter, Hendrix,
& Kelechi, 1999). However, while caregivers usually decrease their daily exercise after
assuming the caregiver role (Wright, 1997), it is important to present minority caregivers with
exercise regimens pertinent to their culture and lifestyle.
Wallsten (2000) compared gender, race, and income as factors in caregiving based on a
cross-sectional study. The average time spent caregiving was at 5.29 years; with mean income
for African American caregivers at $9,226 annually, and $16,111 for Caucasian caregivers
annually. For health symptoms reported by caregivers, Caucasian females recorded the highest
rate of health symptoms compared to African American females. African Americans and
Caucasians reported equal numbers of individuals in their social support circles.
In a qualitative study of 28 African American men and women diagnosed with prostate and
breast cancer, respectively, all individuals had participated in reciprocal "giving and receiving"
relationships (Hamilton & Sandelowski, 2003, p. 656). Social support circles among African
American families are considered "kinship systems." The common ground for most African
American individuals seemed to be seeking information and consolation through people in their
church or immediate family community, and through religion in general. Most individuals
participated in reciprocal relationships throughout their illness, increased their self-esteem and
overall positive attitudes in dealing with their illness (p. 670).
Few studies have examined Asian caregivers. Researchers conducting a multiethnic study
of family caregivers (African American, Asian Americans, Latino, and Anglo European-
American) studied how ethnically diverse populations view dementia (i.e., in medical versus
non-medical terms) (Hinton, Franz, Yeo, & Levkoff, 2005). The majority of caregivers
described dementia to care receivers in both "biomedical and folk models." Although most
caregivers in this study used both models to describe dementia, a significant association was
found between ethnicity, education, and model type. Implications of this study for a group of
ethnically diverse caregivers include having formal care providers learn more about family
dynamics and traditions in communicating with care receivers. "Folk model" explanations
among caregivers dealing with care of those with dementia could hinder and influence treatment
and medical decisions.
Picot's caregiver rewards scale (PCRS) assessed cross-cultural Chinese adult caregivers,
who were born abroad but were currently living in the U.S. (Hseueh, Phillips, Cheng, & Pico,
2005). Caregivers in this study had lived in the U.S. for a mean of 19 years (range from two to
58 years). Most caregivers were women (70%) and had annual incomes that ranged from
$40,000 to $149,000 (84%). Rewards in the caregiver role were the primary focus of the scale.
Results showed that Chinese caregivers in the study felt that their role as caregivers achieved
"personal growth, fulfilling payback and close bonds with family members" (Hseueh et al., 2005,
p. 768). These findings are fairly consistent with similar studies among Chinese families and
Formal and non-formal (i.e., traditional) long-term health care preferences were assessed
among a population of older Korean Americans who had sustained a hip fracture or suffered a
stroke (Min, 2005). Care recipients in this study were older than 65 years. Factors used to
assess health care preference included traditional family values, degree of acculturation, and
"level of informal support" that incorporated Activities of Daily Living and Instrumental
Activities of Daily Living (p. 373). Those who were married and had sustained a hip fracture, or
who had previously sustained a hip fracture, were more likely to choose a mixed health care
format (i.e., combination of formal and non-formal health care). For those who reported having
strong traditional values and had suffered a stroke, individuals were not as likely (15%) to
choose a mixed health care format. However, older individuals who had experienced a stroke
were more likely to choose the all-formal health care option. Overall, individuals who were
female, reported being in better health, and who had higher levels of education were more likely
to choose a mixed health care format (Min, 2005).
In a qualitative study of Thai caregivers living in the U.S., several important aspects of
caregiving emerged (Limpanichkul & Magilvy, 2004). Caregivers had a considerable degree of
family support, which helped with emotional and financial well being. Social support from
neighbors and friends also was prevalent among this group of caregivers. Coping strategies
included strong religious beliefs (Buddhism) and participation in activities such as praying.
Finally, a strong partnership with the care receiver was essential, and included "using humor and
understanding the care recipient" (p. 22).
Two studies based in the United Kingdom assessed needs of informal caregivers among a
South Asian population caring for those with dementia (Adamson & Donovan, 2005), and a
South Asian and African/Caribbean population caring for those with a variety of disabilities
(Katbamna, Ahmad, Bhakta, Baker, & Parker, 2004). One study used qualitative methods
(Adamson & Donovan, 2005) and the other study used a combination of qualitative and
quantitative methods. Both studies confirmed it is expected that adult sons and daughters,
including daughters-in-law, will care for elderly family members.
When care recipients had a variety of disabilities (Katbamna et al., 2005), husbands of
female caregivers were "rarely responsible for intimate forms of care" (p. 400), but did assist in
tasks such as taking children to school or clinical appointments. Many female caregivers of
children with disabilities stated that their husbands provided the only source of "emotional
support" they needed for continuation of care. Most caregivers lived in an extended household
environment, and many adults relied on younger children to interpret questions and concerns to
formal health care providers because they spoke little or no English.
Many caregivers reported that caregiving skills were part of their lifestyle since childhood
(Adamson & Donovan, 2005), and often described their current caregiving role as being natural
(p. 39). Again, lack of support from family members influenced the degree of satisfaction and
stress among caregivers. Caregiver responsibilities were mandated according to what the study
states as "hierarchy of caring responsibilities" (p. 45). Female married caregivers viewed their
roles as an extension of their other marital roles.
In summary, it is important to assess multiple factors involved in caregiving. These factors
include current health status of the caregiver, current health status of the care recipient, age of the
caregiver, age of the care recipient, race/ethnicity of the caregiver, and type of disease or illness
of the care recipient. The research shows minority caregivers (i.e., Latino/Hispanic and African
Americans) have lower SES, have less quality access to health care, and care for individuals who
present more severe symptoms, but they tend to institutionalize less, feel more positive and feel
less depressed than their Caucasian counterparts. In the African American culture, caring for a
family member is considered "normative" and therefore more engrained than in the Caucasian
culture. For Latinos, the concept of"la familiar which basically places family above the self, is
a deterrent to institutionalizing family members sooner and motivation for continued care. For
all cultural groups (i.e., African Americans, Latino, and Asians), it is important for formal care
providers to acknowledge how individual groups and families frame the disease (biomedical
versus folk models), which could potentially influence the well-being and care for the care
Temperament, personality, and character are not only three distinct parts of an individual's
persona, but they work together in a manner that predisposes individuals to certain attitudes and
actions. According to research in this area, temperament is described as a pre-disposing factor,
and as Keirsey states (1998), "our brain is a sort of computer with temperament for its hardware,
and character as its software" (p. 20).
In general, the basic characteristics of personality include the manner in which an
individual uses words and ideas (abstract versus concrete), and how we accomplish tasks (i.e.,
tools we use), which can be either cooperative versus utilitarian (Keirsey, 1998). Individuals
who are cooperative in nature essentially follow the approved guidelines and procedures, while
those following a more utilitarian path will use any method that can "promise success with
minimum cost and effort" (Keirsey, 1998, p. 28).
The concept of distinguishing temperament dates back to ancient Greek and Roman times
to Hippocrates (370 BC), Plato (340 BC), Aristotle (325 BC), and Galen (190 A.D.). Galen, a
Roman physician, believed that the balance on our bodies is dependent in the balance of our four
bodily fluids: Sanguine (lively, upbeat), Choleric (excitable, quick-tempered), Melancholic
(worrisome, sad), and Phlegmatic (slow-moving, dejected). Plato categorized people similarly.
Early philosophers focused on individual predispositions toward certain behaviors (Keirsey,
1998, p. 23). These four bodily fluids as described by Galen, and the four attributes an
individual can "contribute to the social order" as described by Plato, are described by the four
Keirsey temperament types (Keirsey, 1998).
Other researchers, such as Pavlov and Freud, in the field of personality and character,
which is closely related to the field of psychiatry, focused more on stimuli and basic instincts of
humans (Biological Aspects of Personality: PSY 230 Theories of Personality, 2000, para. 6;
Keirsey, 1998, p. 20). In the 1920s, Carl Jung began to formulate the basis for the modern
Temperament Theory. Jung stated that individuals have basic archetypes, which are derived
from four psychological functions thinking, feeling, sensation, and intuition (Jung, 1926).
Temperament, which is consistent over time, relates to why people behave a certain way and the
motivations for their behavior (Keirsey, 1998, p. 26; Understanding the Four Temperament
Patterns, 2000, para. 1).
During the 1940s, Katherine Cook Briggs and her daughter Isabel Briggs Myers created
the Myers-Briggs Type Indicator, which extended Jung's theories and categorized different
dimensions of temperament into discrete functions. The MBTI was specifically designed to
identify 16 patterns that projected action and attitude (Keirsey, 1998, p. 3). The receiving
functions include Sensing (S using the five senses to focus on realism, details, and practicality)
and Intuition (N using insight beyond senses, creativity, and being future-oriented). Judgment
functions involve evaluation of circumstances and include Thinking (T evaluating situations
through a cause and effect process and a more impersonal and objective outlook) and Feeling (F
- focused more on personal values, humanism, harmony). How individuals deal with energy is
described as either Extrovert (E energy going into external objects in the environment,
communicating easily, open to new experiences, and sociable) or Introvert (I energy comes
from environment, more solitary and private, and thinking before expressing feelings) (Calahan,
1996; Daley, 2000).
Myers and Briggs also examined "orientation" to the outside world through the
characteristics of Judging (J preferring more organized and decisive activities, close-ended)
and Perceiving (P information gathering continuing as long as possible, more spontaneous,
open to new ideas and experiences) (Calahan, 1996; Daley, 2000). Myers and Briggs felt that
the perceiving or judging preference revealed "a person's attitude or orientation to life" and aptly
determined how individuals dealt with their external environment. Thus, temperament type (a
predisposing factor) represents an important indicator of how caregivers will define and respond
to their caregiving role. For example, a perceiving caregiver, identified as an "Idealist," may
often look for meaning and significance in making a decision in a given situation. Idealists more
likely will value the opinion of family members.
During the transition from childhood to adolescence, and into adulthood, individuals
display what Jung called "dominant and auxiliary characteristics" (Daley, 2000, p. 21). With
auxiliary characteristics or functions, an individual develops a balance between the internal and
external worlds. This transition is important in caregiving because, for instance, a caregiver who
is considered an "Introverted" temperament type may use both self-knowledge and knowledge
from the Internet to make the best medical and quality-of-life decisions for the care receiver.
Psychologist David Keirsey (Keirsey & Bates, 1984; Keirsey, 1998) developed an
instrument to determine individual temperament based on the work of Myers and Briggs. Each
temperament has characteristics of the 16 patterns conceptualized by Myers and Briggs. The
Keirsey Temperament Sorter was created based on the belief that temperament can be
manifested from how an individual uses language and tools (i.e., computers), and that
temperament is defined by "differentiation not integration" as Myers and Briggs proposed
Keirsey's temperament types include Artisans (SP), Guardians (SJ), Idealists (NF), and
Rationals (NT) (Daley, 2000). McCarley and Carskaden (1987) compared the Keirsey
Temperament Sorter and the Myers-Briggs Type Indicator for validity and found both
instruments comparable in construct validity (i.e., some types were low and some types were
The four "Artisan" types are ESTP ("Promoter"), ISTP ("Crafter"), ESFP ("Performer"),
and ISFP ("Composer") (Keirsey, 1998, p. 11). The four "Guardian" types are ESTJ
("Supervisor"), ISTJ ("Inspector"), ESFJ (" Provider"), and ISFJ ("Protector") (Keirsey, 1998, p.
11). The four "Idealist" types are ENFJ ("Teacher"), INFJ ("Counselor"), ENFP ("Champion"),
and INFP ("Healer") (Keirsey, 1998, p. 12). The four "Rational" types are ENTJ
("Fieldmarshal"), INTJ ("Mastermind"), ENTP ("Inventor"), and INTP ("Architect") (Keirsey,
1998, p. 12).
Artisans are "talented at using tools" such as language and computers, as well as being
skilled and creative. In decision-making skills, Artisans tend to be the most impulsive of the four
types and "trust their impulses without reserve" (Keirsey, 1998, p. 56).
Guardians are more rule-oriented, use more reasoning skills and are described as "stable,
secure, and following protocol." The best way to describe Guardians is as administrators; their
"first instinct is to take charge and tell others what to do" (Keirsey, 1998, p. 84). Guardians feel
responsible for the morality of those around them.
Rationals are more analytical, they "value expertise and seek progress" and are most likely
to think more abstractly. The four Rational roles tend to be more strategic in nature, tend to
remain calm, tend to seek knowledge, and most often trust their own reasoning. The most
famous Rational was Albert Einstein (Keirsey, 1998, p. 161).
Idealists focus on "cooperative interactions" that deal with moral and ethical issues.
Idealists are more diplomatic in nature, and are more adept at interpersonal communication skills
(Understanding the Four Temperaments, 2000, para. 9). Idealists are the most abstract of the
four temperaments, and tend to be inductive in their thought process, which signifies that they
will take a "few particulars to sweeping generalities" (Keirsey, 1998, p. 121). Because Idealists
usually process in the abstract, their thought and speech patterns tend to be interpretive, which
means that certain comments are "how one thing is really something else" (Keirsey, 1998, p.
Individuals display a predisposition or affinity for certain functions (Understanding the
Four Temperaments, 2000, para. 1). As individuals progress through the life cycle, these
predispositions become more evident in decision making and skill development. Refinement of
one's preferences into adulthood can assist in helping make decisions about professions, family,
health care, and children.
When individuals move into the caregiver role, they experience different stressors that
cause a state of arousal, thereby setting the individual into a "coping" mode regarding that
stressor (Morris, 2000). Temperament, an individual's propensity to act in a certain manner,
when combined with confidence and determination, may predict how a caregiver will cope with,
and respond to, stressors. Temperament is important among caregivers as a function of what
motivates and drives decisions, interactions with others, and actions. In the caregiver role,
decisions and actions regarding the care receiver and the caregiver, him or herself will affect the
health and quality of life for both individuals and those in the immediate family. Therefore, it is
important to determine the temperament of the caregiver to help assist in the facilitation of
information flow, awareness, and communication between formal (medical staff) and informal
Morris (2000) examined temperament types among students selected for dental school
using the Keirsey Temperament Sorter II (N=299). Extroversion (84.3%) exceeded Introversion
(15.7%) among dental school applicants. No significant proportional differences existed
between male and female students. Most (94%) preferred judging over perceiving (6.4%).
Feelings were preferred more by females than by males in the "mental function" category, and
the comparison showed statistical significance. Findings for dental students were similar to
results from studies involving medical students. Although this study did not focus on the role of
caregivers, it depicts the percentage of males and females who have different temperament traits
- more women have the "feeling" trait. In addition, one research study suggested that some
temperament types are more likely to engage in unhealthy behaviors and take more risks (Daley,
In a study that assessed caregiver characteristics inherent in temperament types,
researchers showed that if caregivers displayed a higher propensity for depression during the first
year of caregiving, there was a higher rate of patient admission to a nursing home. The number
of activities of daily living provided by the caregiver, the caregiver's age, and the caregiver's
health were used as indicating factors in the study (Yaffe, Fox, Newcomer, Sands, Lindquist,
Dane, & Covinsky, 2002).
Hadley (2003) combined Bandura's construct of self-efficacy and the Keirsey
Temperament Indicator II, and applied both of these concepts to determine if they related
directly to job performance. The Keirsey Temperament Indicator II and the Myers-Briggs Type
Indicator usually are used to assess job performance standards in the workplace or in schools.
In summary, temperament is a predisposing factor that changes little throughout the life-
time, although certain aspects, such as decision-making skills, develop during adolescence and
adulthood. Temperament is important in the caregiving role because it often defines one's
natural preferences (i.e., formal versus traditional health care ). It is important for those in
formal health care fields to understand how different temperaments process and communicate.
Temperament types are found in all races, cultures and ethnicities', therefore, this aspect can
bring insight into how health educators, doctors, and other health care providers "get their point
across" for optimum benefits and quality of life among caregiver and care receiver.
Self-Efficacy and Outcome Expectancies
Self-Efficacy, conceptualized by Albert Bandura, is a construct of the Social Cognitive
Theory, which extends to the Social Learning Theory and helps to predict and understand
behavior (Glanz, et al, 2002). Originally, Bandura applied Social Cognitive Theory to children's
behavior in a laboratory setting. According to Bandura (1986), self-efficacy influences how we
make choices, the motivation or effort put forth toward an action, how we accept or reject life
changes, and how we feel about our choices, actions, and changes. Social relationships enhance
feelings of self-worth, self-esteem, and a sense of well-being when individuals feel valued by
meaningful others (Miller, Townsend, Carpenter, Montgomery, Stull, & Young, 2001).
Self-efficacy involves "one's capabilities to organize and execute the sources of action
required to manage prospective situations" (Bandura, 1986, p. 5). Self-efficacy develops over
time and is influenced by one's failures and achievements, influenced by significant others, by
observing others, and by a person's emotional state while performing a behavior or taking an
action. An important determinant of how self-efficacy develops involves "self-reflection,"
defined as the ability of individuals to "analyze their experiences, think about their own thought
processes, and alter their thinking accordingly" (Brown, 1997, para. 21).
Four factors influence the development of self-efficacy: mastery experiences (balance of
successes and failures in tasks), vicarious experiences (watching others similar to oneself
perform and succeed in various behaviors), social persuasion (verbal persuasion about one's
capabilities), and positive mood enhancement and stress reduction (increasing efficacy at
reducing stressors in life-changing situations) (Bandura, 1994).
Self-Efficacy and Effect on Health and Disease
Researchers examined enhancement of support for health behavior change among
postmenopausal women younger than 75, diagnosed with Type II diabetes, and who were at risk
for heart disease (Toobert, Strycker, Glasgow, Barrera, and Bagdade, 2002). Participants were
randomly assigned to three groups: usual care (N=123), lay-led peer group (N=78), and
computer-based support involving community resources (N=78). Measures included behavior
changes (diet and eating habits), physiological changes (carotid artery intimal-media thickness),
psychosocial changes (quality of life and depression), and process measures (perceived social
support, perceived stress, obstacles, barriers) (p. 578). Participants attended retreats to assist
them in preparing meals and eating healthier, to learn stress management techniques such as
yoga, and to study exercise guidelines for their lifestyles. Support group leaders scheduled
interventions during the follow-up period. Results indicated that those who had a follow-up
intervention tended to respond more positively to the information and resources provided to
Conn, Burks, Pomeroy, Ulbrich, and Cochron (2003) studied older women (N=203) and
the health benefits of exercise designed to reduce their risks for cardiovascular disease.
Individuals with high self-efficacy maintained higher exercise expectations and continued their
routines. Self-efficacy emerged as a strong indicator for primary care intervention. Exercise
self-efficacy expectations had a direct significant effect on exercise behavior and on processes of
change (p < .0001).
The Arthritis, Diet, and Activity Promotion Trial (ADAPT) tested self-efficacy for
behavior change in lifestyles of overweight, non-active men and women with osteoarthritis
(Miller, Rejeski, Williamson, Morgan, Sevick, Loeser, Ettinger, & Messier, 2003). This study
was designed to "test the effect of diet and exercise on self-reported physical function" (Miller et
al., p. 473). Women were randomized into four groups healthy controls, dietary weight loss
with no structured exercise, structured exercise with no dietary loss, and exercise and weight
loss. The study included an 18-month follow-up evaluation. In addition, physical functioning
and four different BMI levels (p=0.0062) were directly related to women who were in different
BMI categories as described above. Therefore, how well an individual could perform (physical
functioning) was directly related to how confident he or she felt about themselves.
Northouse, Mood, Kershaw, Schafenacker, Mellon, Walker, Galvin, and Decker (2002)
found that quality of life among women with recurring breast cancer (N=189) related directly to
levels of self-efficacy, social support systems, and cohesiveness or strength of the family unit,
and support with the patient. Women with fewer current concerns gave less negative appraisals
of the illness and caregiving and subsequently reported a higher quality of life. Self-efficacy had
a direct and positive effect on patients' mental health.
Death of a spouse greatly affects the widowed individual's self-esteem, quality of life, and
other self-efficacy issues, especially after an extended period of caregiving (Fry, 2001). In one
study, widows showed lower self-esteem, while education and income predicted widows' quality
of life. Widows who reported greater levels of quality of life also reported greater levels of
perceived self-efficacy in the categories of spirituality, interpersonal relationships, and social
support (Fry, 2001). Similarly, other research indicated that caregivers who showed no
significant strain in the pre-bereavement period were more likely to experience a higher level of
depression post-bereavement (P=0.002), and African Americans recorded significantly lower
depression symptom scores than Whites in the pre- and post-bereavement period (Schultz, et al.,
2001, p. 3127).
Self-Efficacy Among Health Professionals
The relationship between the informal caregiver and the care receiver is important to
understanding the role of the formal caregiver. Although exposure of the care receiver and
caregiver to formal caregivers is minimal compared to family and friends, health care
professionals should understand the impact they can have on basic daily functioning, overall
mental health, and self-efficacy of the caregiver and care receiver.
Pinto, Friedman, Marcus, Kelley, Tennstedt, and Gillman (2002) examined telephone-
based interventions by formal caregivers to promote physical activity among health
professionals. Sedentary adults (72% women, 45% African American) were placed in two
groups, both requiring telephone-based interventions. One group received a healthy eating
promotion and the other group received a physical activity promotion. Both groups were
assessed at baseline, at three months, and at six months after the beginning of phone
interventions. Self-efficacy with physical activity represented the degree of confidence one feels
in being a physically active individual. Moderate physical activity for energy expenditure at
three months showed the only statistically significant interaction (p=0.02) with self-efficacy.
Follow up by formal health care providers and degree of concern were part of the intervention.
Health professionals can play an important role in reducing effects of depression among
cancer patients. Health professionals can increase their own self-efficacy in the formal
caregiving role by understanding factors involved with patient diagnosis, such as their own
supportive role, attitudes, and beliefs about communicating with cancer patients (Parle, Maguire,
& Heaven, 1997). After completing a skills workshop on dealing with cancer patients'
emotional well-being, health care professionals "rated the likely outcome of a discussion of
psychological concerns for a terminally ill patient as more positive" (P < .001) (Parle, et al.,
1997, p. 239). Although health professionals may feel confidence or high self-efficacy in
dealing with patient emotions, they do not feel as confident in managing these feelings.
Self-Efficacy and Caregivers
Gottlieb and Rooney (2004) studied coping and self-efficacy regarding mental health
among caregivers for persons with dementia (N=141). Optimism and relational self-efficacy
(i.e., degree of emotional involvement with the care receiver) predicted coping effectiveness.
Caregiver exposure to dementia symptoms, being less optimistic, and holding weaker beliefs in
one's coping effectiveness were significant predictors of a negative health effect (p < 0.01). In
addition, coping had a major impact on the caregiver's mental health even after controlling for
In a study on self-efficacy of family caregivers who were helping cancer patients manage
end-of-life pain (Keefe, Ahles, Porter, Sutton, McBride, Pope, McKinstry, Furstenberg, Dalton,
& Baucom, 2003), caregiver self-efficacy was significant when correlated with caregiver strain,
positive mood, negative mood, and the patient's physical well being. In situations where
caregivers reported higher self-efficacy, care receiver reported having more energy, feeling less
ill, and spending less time in bed.
Similarly, Tang, and Chen (2002) examined health promotion behaviors of Chinese family
caregivers for stroke patients. Regression analysis revealed the variable of caregiver's health
status as the only positive predictor of caregiver self-efficacy. Satisfaction with social support
was the strongest predictor of caregiver health promotion behavior (p. 331).
Gitlin, Corcoran, Winter, Boyce, and Hauck (2001) studied the effects of a home
environmental intervention on self-efficacy and upset in caregivers and the daily function of
dementia patients (N=171). Caregivers were randomly assigned to a treatment group in which
they were exposed to a home environmental intervention. Caregivers assigned to a control group
were exposed to a usual care setting and were educated on how to adjust their environment to
simplify caregiver workloads and reduce stress. Most caregivers (59%) were daughters or
daughters-in-law. In addition, 126 (74%) were White, and 43 (25%) were African American.
Among the respondents, the mean number of months spent in the caregiver role was 45 months.
The largest interaction occurred for caregiver behavior self-efficacy and behavior upset (p=.04).
In this study, caregivers reported managing Instrumental Activities of Daily Living (IADL) and
Activities of Daily Living (ADL) on a moderate self-efficacy level. Minority participants in the
treatment group showed greater improvement in IADL and ADL self-efficacy.
In a similar study using skills training interventions for female caregivers of relatives with
dementia (N=169), those in anger and depression management groups showed significant
reductions in levels of anger, and depression decreased significantly, while self-efficacy
increased between the first and second data collections (T T2, respectively) (Coon, Thompson,
Steffen, Sorocco, & Gallagher-Thompson, 2003).
Fortinsky, Kercher, and Burant (2002) compared measurement and correlates of family
caregiver self-efficacy for managing dementia (N=197). They sought to develop a strategy to
measure family caregiver self-efficacy for managing dementia, incorporating domains of
symptom management and use of community support services. More than 80% of caregivers
were female, and about 20% classified themselves as African American. Caregivers who
reported a higher level of dementia symptom-management self-efficacy also reported fewer
depressive symptoms. This study suggests that self-efficacy represents a coping mechanism in
response to the stress of caring for a relative with dementia.
Expectancies (or incentives) differ from expectations, both Social Cognitive Theory
constructs because they reflect the value an individual places on "a particular outcome" (Glanz et
al., 2002, p. 172-173). In the caregiver role, expectancies influence the behavior of individuals
in determining choices for themselves and for the care receivers for the most positive outcome.
For example, if the care receiver has a terminal illness, the caregiver will make decisions that
seem best to optimize quality of life for that individual regardless of his or her own personal
health (Wright, 1997).
Chapter 2 presented a review of literature related to various aspects of informal caregiving
including health, stress, and ethnicity. The review also addressed the importance of the role of
self-efficacy, and how temperament can act as an important determinant of decision-making.
Temperament is considered to be fairly stable through the life span, with each individual
predisposed to certain characteristics within that temperament.
Although no known literature to date has assessed temperament using the Keirsey
Temperament Indicator II and self-efficacy among informal caregivers, studies using the Myers-
Briggs Type Indicator and self-efficacy primarily focused on issues such as self-assessment of
schizophrenics, first-time mothers, work place risks, and career decisions. Studies reviewed in
this chapter imply that certain temperament types interact with self-efficacy to influence risk
taking in personal care and decisions, including health care decisions, interactions in a group
setting, degree and dynamics of social support systems, and relationship dynamics with the care
The review included factors that influence individuals serving as informal caregivers such
as cost and personal risks, research that defines and describes types of temperament, and research
that examines self-efficacy in relation to disease as applied to formal and informal caregivers.
Though temperament and self-efficacy can influence how individuals respond to the caregiver
role, the literature review produced few studies that directly examined the relationships existing
among temperament and self-efficacy among informal caregivers. In this study, self-efficacy
combined with expectancies, is presented as an enabling factor.
Because of the intense nature and diversity of tasks involved in caregiving, and the
growing population of caregivers in the U.S., a better understanding of how temperament
influences self-efficacy regarding attitudes and beliefs about caregiving, can present health care
providers and researchers in the field a better understanding of how to assist caregivers. Chapter
3 describes the research design, instruments, and process used to gather information about
Table 2-1. Incremental caregiving hours and incremental cost of informal care attributable to
Incremental Cost per Year
Dementia Incremental Hours Using Low-range Estimated Using Mid-range Estimated Using High
Severity Estimated Cost of Informal Care Cost of Informal Care Cost of
Per Week Informal
Normal Reference Reference Reference Reference
Mild 8.5 $2,610 $3,630 $4,780
Moderate 17.4 $5,340 $7,420 $9,770
Severe 41.5 $12,730 $17,700 $23,310
Langa, et al., 2001, pg. 775
Table 2-2. Adjusted weekly hours and yearly costs of informal caregiving
Hrs. per 95% CI Cost per
Week (hours) Year 95% CI
No history of cancer 6.9 6.5-7.2 $3,000 $2,800-$3,100
History of cancer, 6.8 6.4-7.1 $2,900 $2,700-$3,000
History of cancer, 10.0+ 9.6-10.5 $4,200+ $4,100-$4,500
Hayman et al., 2001, p. 3222; + denotes P< .05
This study examined relationships that exist between different temperament types and
perceived self-efficacy among individuals serving in an informal caregiver role. The effects of
self-efficacy on several caregiver characteristics were also examined. This chapter describes the
methodology used in this study, which includes the research design, research variables,
instruments, the pilot study, and the final study.
The study used a cross-sectional survey research design to explore possible relationships
between temperament and self-efficacy among informal caregivers. With this design, a standard
protocol can be administered to a group of participants using standardized procedures that
require a reasonable amount of time. Likewise, data can be analyzed uniformly and objectively
(McDermott & Sarvela, 1999). Participants included individuals 18 years or older who provided
part-time or full-time caregiver care for at least one individual 21 years or older. Participants
also participated in support groups, health forums, and health fairs. Two models provided a
theoretical framework for this study.
Social Cognitive Theory (SCT), originally called Social Learning Theory, was
conceptualized by Bandura in the 1960s (Bandura, 1986). Self-efficacy, the most often used
construct in the SCT, helps determine an individual's confidence level in performing an action or
series of actions. Combined with outcome expectancies (i.e., "will others also perceive my
action as beneficial"), self-efficacy can determine duration of performing an action, such as
serving in a caregiver role. In studies on understanding the caregiver role, self-efficacy helps
determine caregiver reactions to stressors and other psychosocial factors. Self-efficacy also
plays a role in overcoming environmental barriers in performing the behavior, or in not starting a
negative behavior (i.e., cigarette smoking). This study used the constructs of self-efficacy and
outcome expectancies from SCTto examine how temperament type interacts with self-efficacy in
the caregiver role (Wilson, Friend, Teasley, Green, Reaves, & Sica, 2002; Glanz et al., 2002, p.
Temperament Theory, defined by Carl Jung, posits that seemingly random behavior
actually reflects one's natural preferences, and that certain people are predisposed to display
different behaviors (temperament types).
In one study of employee motivation (Burke, 2004), the factors of choice, competence, and
meaningfulness represented the "primary intrinsic motivators" for the Artisan, Rational, and
Idealist, respectively. Choice, competence, and meaningfulness scales were significantly
different among three temperaments Artisans, Rationals, and Idealists. The Keirsey
Temperament Indicator II, as used in this study, applied Jung's theory to identify four
temperament types: Rationals, Idealists, Artisans, and Guardians (Temperament: Temperament
versus character, 1998, para. 1). No previous study combining the Keirsey Temperament
Indicator II with a self-efficacy scale among informal caregivers was found in the literature.
The study included one explanatory (independent) variable (temperament type) from the
Keirsey Temperament Indicator II, with four response options: Artisans, Idealists, Rationals,
Guardians. The Keirsey Temperament Indicator II has 70 questions and is similar to the Myers-
Briggs Temperament Indicator in that it uses a "series of forced-choice" responses. Each of the
70 questions for the Keirsey Temperament Indicator II did not "reflect any of the other
dichotomies" (Daley, 2000). For example, the 10 specific questions for the dyad characteristic
of E/I were grouped and did not reflect on any of the questions for the dyad characteristics of
S/N, T/F, or J/P. Each question for this instrument has two response options: a or b. The 16
temperament types, as originally defined by Myers and Briggs, are combined into four discreet
temperaments defined by Keirsey. These four temperaments were discussed in Chapter 2. The
Caregiving Self-Efficacy Scale was used as a dependent (outcome) variable for research question
1, and as an explanatory (continuous) variable in research questions 2 through 5.
The study also included several outcome (dependent) variables: (1) the Caregiving Self-
Efficacy Scale, a continuous variable, included scale options ranging from 0 to 100 displayed in
increments of 10, but respondents were free to answer with any number in the scale range that
accurately indicated their degree of confidence; (2) intensity of care, a categorical variable, asked
"how difficult or how demanding is it for you to care for this person" (not difficult at all, a little
difficult, difficult, very difficult, and extremely difficult; (3) total duration of care, a categorical
variable, asked "how long have you provided care for this person" (6 months or less, 6-12
months, 1 year, 2 years, 3 years, 4 years, 5 years or more); (4) daily duration of care, a
categorical variable, asked "on average, how much time do you spend each day caring for this
person" (1 hour or less, 1-3 hours, 4-6 hours, 7-9 hours, 10-12 hours, 12 hours or more); and (5)
relationship to the care receiver, a categorical variable, asked "your relationship to the person
you care for" (my parent, my spouse, my child, a grandparent, an in-law, my friend, my partner,
To provide consistency in data analysis for some of the variables listed above, several
categories were collapsed. For example, for total duration of care, the first two categories (6
months or less and 6-12 months) were combined to produce categories at one-year intervals.
Instruments for the study were selected by conducting an extensive literature review of
instruments previously used by researchers to examine self-efficacy, temperament, and other
functions and aspects of caregiving. The protocol for this study included two main instruments
and caregiver population demographics: (1) Caregiving Self-Efficacy Scale (Steffen et al.,
2002), (2) Keirsey Temperament Indicator II, and (3) Caregiver Profile Information.
Caregiving Self-Efficacy Scale
After extensively reviewing the literature, this scale was deemed most appropriate to
measure self-efficacy among caregivers. The instrument was created originally for caregivers of
dementia patients. The scale was designed specifically to measure perceived efficacy among this
population of caregivers. The 15-item scale includes three sub-scales: "Obtaining Respite
Care," "Controlling Upsetting Thoughts About Caregiving," and "Responding to Disruptive
Behaviors" (Appendix D). Respondents rated their perceived confidence on a 0 to 100 scale,
with 0 as "cannot do at all" and 100 as "certain can do." Bandura, who conceptualized the
construct of self-efficacy, used in both SCT and SLTin the 1970s, contributed to the
development of this scale for adequate reliability and validity. After reviewing the 15 items, the
researcher determined that the items would pertain to caregivers providing care to terminally ill
or chronically ill individuals, including all types of dementia.
The development process for the Caregiving Self-Efficacy Scale included rigorous item
analysis procedures. The original scale produced reliability coefficients of .80 or higher for each
of the three sub-scales, and for the total scale. This total scale, which represented one of the
dependent variables for the study, was considered a "continuous" variable because the scores
from the three sub-scales were averaged to produce an overall score from 0 to 100 (ex., 85.4).
For the purpose of this study, only the total self-efficacy score was used (i.e., total number for
the three sub-scales). Permission was obtained from the first author to use the scale and the
instrument was used as originally published for the current research study (Steffen et al., 2002).
Keirsey Temperament Indicator II
The Keirsey Temperament Indicator II instrument uses four "bipolar scales" to sort
individuals into 16 personality types. These four scales include Extroversion/Introversion (E/I),
Thinking/Feeling (T/F), Sensing/Intuition (S/N), and Judging/Perceiving (J/P) (Alpine Media
Corporation, 2003, p. 6-7). Keirsey extended the process to sort individuals into four
temperament types with four distinct personality attributes within each temperament, using a 70-
item scale (Appendix E). The four types are: Artisans (ESTP, ESFP, ISTP or ISFP), Guardians
(ESTJ, ESFJ, ISTJ, or ISFJ), Idealists (ENFJ, ENFP, INFJ, or INFP), and Rationals (ENTJ,
ENTP, INTJ, or INTP).
The Keirsey Temperament Indicator II resembles the Myers-Briggs Temperament
Indicator in format and approach, though it is more limited in scope. David Keirsey modeled the
Keirsey Temperament Indicator II after the work of Myers and Briggs. The Keirsey
Temperament Indicator II has been applied in assessing temperament among dental and medical
students, in a classroom setting, and in business settings, but not specifically among caregivers or
in any other health behavior venue known to this researcher. No peer reviewed journal
publications provided adequate estimates of validity and reliability for the Keirsey Temperament
Indicator II, so validity and reliability of the instrument were assessed using the final population
data. The print version of the Keirsey Temperament Indicator II instrument was used for both
the pilot and final study populations because that version better suited the nature and age of the
Each of the four Keirsey temperaments is considered continuous in nature. Statistical
analyses were previously conducted on a wide scale of the computerized version of the Keirsey
Temperament instrument (Alpine Media Corporation, 2003). Reliability for this instrument was
previously measured using several methods including Item-Response Theory method (ranging
from .33 to .76), internal consistency coefficients alpha (.80 and higher), and test-retest methods
(.80 and higher).
For example, using Item-Response Theory among males and females in each group, for the
four "bipolar scales," reliability estimations ranged from .81 to .83: E/I, total sample=.83,
males=.83, females=.83; S/N, total sample=.82, males=.82, females=.82; T/F, total sample=.83,
males=.82, females=.82; J/P, total sample=.82, males=.81, females=.82. (Alpine Media
Corporation, 2003, p. 38).
Using Pearson's correlation to compare the Myers-Briggs Temperament Indicator and the
Keirsey Temperament Indicator II, correlation coefficients of .62 and higher for the eight
temperaments, E/I, S/N, T/F, J/P were reported (Hadley, 2003). Another study produced Pearson
correlation coefficients ranging from .68 to .84 (Ludy, 1999; Calahan, 1996).
Validity measures also were conducted by the Alpine Media Corporation (2003) on the
Keirsey Temperament Indicator II. The 140 variables coincided with the 70 items. Factor
analysis of the Keirsey Temperament Indicator II identified 11 factors with eigenvalues of 1.0 or
greater. These 11 factors included, but limited to, Factor 1 (sensing versus intuition), Factor 2
(thinking versus feeling), Factor 3 (extroversion versus introversion), Factor 4 (judging versus
perceiver). Factors 10 and 11, which were the weakest, included a variation of S/N, but also
included T/F and J/P. Factor 9 was closely related to Factor 1, Factor 6 was closely related to
Factor 2, and Factors 5, 7, and 8 were closely related to Factor 4. The Alpine Media Corporation
analysis included scores from more than 77,000 individuals who completed the instrument. The
analysis used SPSS and factor analysis statistical procedures to determine the number of unique
factors or constructs that accounted for significant variance in the population data set (Alpine
Media Corporation, 2003, p. 50). Of 11 factors mentioned previously, nine showed significant
loadings with some greater than .24, and with most greater than .35. This finding indicates that
the instrument validity and relationships among factors is fairly strong.
Other studies used regression analysis to examine extroversion/introversion, self-efficacy,
and other demographic study variables, including satisfaction in the field of cardiopulmonary
care (Hadley, 2003; Ludy, 1999). Other than the validity study mentioned previously (Alpine
Media Corporation, 2003), the literature review did not locate other studies that addressed
validity as extensively.
Caregiver Profile Information
A 17-item caregiver profile was developed to obtain demographic information about
caregivers such as age, gender, marital status, race/ethnicity, income level, relationship to the
care receiver, and items related to the frequency, duration, and difficulty of the care provided.
Additional profile items included age and gender of the person receiving care, and medical
resources or information used before and after becoming a caregiver (Appendix F). The revised
profile information used with the final study population included one additional question (#21)
based on suggestions from the pilot study participants ("Do you care for more than one
individual? If yes, how many"). Some caregivers provide care for more than one individual,
which influences multiple factors in the caregiving role.
A draft version of the study protocol was prepared to include a script for prospective
participants (Appendix G), the Caregiving Self-Efficacy Scale (Appendix D), the Keirsey
Temperament Indicator II (Appendix E), and the Caregiver Profile Information (Appendix F).
The researcher contacted four individuals who were currently caregivers (Appendix H),
and asked them each to complete the entire instrument. These same individuals assessed the
instrument for readability, comprehension, and cultural sensitivity. They also noted the time
required to complete the entire instrument. The four individuals also offered comments or
suggestions about the structure and questions of the instrument. One suggestion was to copy and
enlarge the Keirsey Temperament Indicator II for caregivers who had difficulty reading smaller
print. Acting on this suggestion, the researcher provided this option at all meetings, forums, and
seminars where data were collected. As an additional check of the protocol and procedures, the
researcher asked four non-caregiver graduate students to read the instrument and answer all
items as quickly and accurately as possible. The researcher compared completion times of four
caregivers and completion times of the non-caregivers in order to accurately present a
completion time estimation to volunteers for the pilot study and final population. The
completion time for the four panel members was longer than the completion time by the graduate
students. The time participants in both the pilot study and final population took to complete the
instrument ranged from 15 to 35 minutes.
Beginning in Summer 2005, social support organizations were contacted and requested to
participate in the study (Appendices A and B). Participants for the pilot study (N=25) were
recruited from meetings of social support groups facilitated by the Family Caregiver Support
Program (Appendix C). Although circumstances differ, pilot studies usually include a small
number of subjects, often selected for convenience (Alreck & Settle, 1995). Caregiver meetings
for the pilot study were held in North Central Florida counties, other than Alachua County,
during August and September 2005 (Appendix C).
Participants were recruited at a caregiver support group meeting with the group facilitator
present. The group facilitator previously had been contacted and the researcher had received
permission to attend the meeting and ask for volunteers. Participants for the pilot study, as well
as the final study population, were caring for an individual in their homes or at another facility
(hospital, assisted living or nursing home), and on a full-time or part-time basis. Participants
provided care that included some Activities of Daily Living (ADLs) such as bathing, changing
and feeding, or some Instrumental Activities of Daily Living (IADLs) such as shopping, paying
bills, cooking, and helping to clean a house or apartment (Arno et al., 1999).
A codebook was set up before data were collected. Each item in the instrument was coded
for data management purposes. Surveys for the pilot study population were numbered from 1 to
25. The Keirsey Temperament Indicator II items were coded according to temperament type as
outlined in the test manual. The information was recorded in an Excel spreadsheet, followed
by information from the remainder of the survey.
At the meetings, the researcher introduced herself and briefly described the survey and its
contents, and her affiliation with the University of Florida. She indicated that a gift card
incentive would be provided upon completion of the entire survey. Volunteers who agreed to
participate received two copies of the Informed Consent, a Keirsey Temperament Indicator II
booklet, the Caregiving Self-Efficacy Scale, and the Caregiver Profile Information questions.
Each participant kept a copy of the Informed Consent form with contact information, and a
signed copy was given to the researcher.
At the meetings, the researcher instructed participants to answer all items in the best way
possible, and to answer each item as it pertained to them at the present time. For example, a
frequent comment about some items on the Keirsey Temperament Indicator II was, "Both of
these apply; how do I answer it?" The researcher was present during the completion of the
survey if any questions arose about instrument items. The researcher always instructed
participants to choose the better response for their current caregiving situation.
While the survey was being completed, some participants discussed the questions with
other participants or made comments to the researcher about the survey. Therefore, a noise
factor evolved during survey completion, perhaps due to the fact that many caregivers were close
friends, regardless of age and ethnicity. As one caregiver said, "We are like a family." During
other meetings in the pilot study, some participants arrived 10-15 minutes after the meeting
began. Therefore, a "lag" time occurred between completion of the survey by participants who
arrived on time and those who arrived late. During the time the latecomers were completing
their survey, other group members chatted among themselves or queried the researcher further
about her study.
All surveys were checked to ensure that all sections were completed. Only one survey was
not fully completed. Participants who completed the protocol received their choice of a 60-
minute telephone card or a $5.00 gift certificate from Publix or WalMart. Most of the requests
were for gift certificates from Publix or WalMart and both were equally popular. Completed
surveys were returned to the researcher who presented participants with the gift card of their
choice. Participants who chose not to finish and complete the survey, did not receive a gift card.
When all surveys had been collected from participants, the researcher thanked the group for their
assistance and left the meeting.
Groups for the pilot study ranged from three to 10 participants. Pilot study participants
required an average of 25 minutes to complete the protocol. Participants who wanted to know
their temperament type were assured confidentiality by the researcher and contacted by mail at a
later date with the information.
The University of Florida Institutional Review Board (IRB) approved the data collection
procedures for the pilot study and for the final study population (Appendix I). Participation in
the pilot testing and final study was voluntary. All participants received a description of study
procedures and information regarding their rights as a participant, and complete confidentiality
was guaranteed and protected. Participants were informed that they could discontinue answering
questions at any time during the process without penalty.
Because volunteers were recruited from specific caregiver support meetings, only the
support group name was used to identify participants on the Excel spreadsheet. The only
identification of participants was by type of group for the data collected and by number on the
survey for analysis purposes (i.e., Hospice, Caregiver meeting, Health Fair/Forum). Only the
researcher had full access to identification of participant names through the signed informed
Descriptive statistics were calculated to determine baseline frequency rates in each
temperament category, average self-efficacy scores, frequency rates for race/ethnicity, frequency
rates for self-reported health, and frequency of self-reported difficulty in caring for their friend or
Females comprised 92% of the pilot study population (n=23), with the percent of
White/Caucasian and African American participants at 60% and 40%, respectively. These were
the only two races represented in the pilot study. Most caregivers (72%) were aged 50 or older
(n=18) and married (52%, n=13). Most caregivers reported their overall health as excellent, very
good, or good (68%, n=17).
Guardians were the most frequently reported temperament type (72%, n=18), followed by
Idealists (16%, n=4), and Artisans (12%, n=3). No participants in the pilot study were reported
in the Rational type.
The majority of care receivers were aged 60 or older (80%, n=20), with most being a
parent (28%, n=7) or a spouse (24%, n=6). Regarding degree of difficulty in caring for the
individual, caregivers most often reported "a little difficult" (44%, n=l 1) or "difficult" (28%,
n=7). The most often reported illness of the care receiver was some form of dementia (i.e.,
Alzheimer's Disease), which participants wrote in under the "Other" response option. The
exclusion of dementia as an option was an oversight when creating the instrument.
Potential bias was controlled in four ways. First, the same data protocol was followed at
almost all meetings, health fairs, and health forums. Meetings usually were about one hour in
duration, while health fairs and health forums were at least four hours in duration (i.e., 10 a.m. to
2 p.m.). Second, the researcher was present to assist with questions or concerns at all meetings
for the pilot study and almost all meetings for the final population. The only exception where the
researcher was not present was with the Hospice volunteers. This process is discussed in the
"Final Study" section. Third, the researcher was able to assess, at the time of completion,
whether the caregiver had inadvertently missed or skipped any items. Fourth, data analyses were
limited to responses from volunteer participants who were 18 years or older, who were currently
caring full-time or part-time for a friend or family member 21 years of age or older, and who
were caring for the individual in their home or at another facility with frequent access.
Based on results from pilot testing, some adjustments were made to the data collection
procedures, instruments, and caregiver profile demographic information. In the pilot study, the
self-efficacy scale and demographics portions of the instrument were printed on both sides of the
survey (i.e., front and back of the paper). This created confusion among some pilot study
participants because caregivers did not notice the questions on the back side of the paper, missed
some questions and left them unanswered. However, because the researcher was present at all
Pilot Study caregiver meetings and reviewed the survey before presenting the gift card, she was
able to return the unanswered portions to volunteers and they were asked to complete the
For the final population, participants received surveys printed on one side only. Also, for
the pilot study, the researcher did not seal the 70 Keirsey Temperament Indicator II questions
from the coding portion of the booklet (see Appendix E), and some participants turned to the
coding portion of the booklet after completing the 70 questions. For the final study, the
researcher stapled or taped the question portion of the booklet from the coding portion of the
booklet and wrote "Do Not Open" to prevent confusion among participants. The researcher also
verbally asked participants to stop immediately after completing the 70 questions.
Likewise, on the self-efficacy scale, some participants in the pilot study attempted to
calculate their own total scores. The researcher instructed participants for the pilot study and the
final population to leave the total score line blank.
The final version of the study protocol was prepared to include a script for prospective
participants (Appendix G), the Caregiving Self-Efficacy Scale (Appendix D), the Keirsey
Temperament Indicator II (Appendix E), and the Caregiver Profile Information (Appendix F).
The Caregiver Profile Information included one additional item suggested during the pilot study:
21. "Do you care for more than one individual?" This item required a "yes" or "no" answer from
participants. If "no" then the researcher coded the number of care receivers as "1." If "yes,"
then the caregiver was asked "If yes, how many total?" Below item 21, the caregiver responded
to the subsequent items "NOTE: If you are caring for more than ONE person, please answer the
following questions for ONE of the individuals you care for." The researcher did not prompt
caregivers to answer regarding the most (or least) difficult care receiver. Therefore, caregivers
may have responded for the person that occupies more of their time.
Data for the final study were collected during Fall 2005 and Spring 2006. Participants for
the final population were recruited from caregiver meetings, health forums of social support
groups facilitated by the Family Caregiver Support Program, health fairs sponsored by various
organizations including the county Aging Care Services, health fairs sponsored by county
Hospice organizations, Haven Hospice volunteer base, and support groups at various assisted
living facilities (Appendix B).
To collect information from the volunteer base at North Central Florida Hospice
organization, the researcher submitted documents for review by the research board at Hospice,
met with the board to address questions or concerns, and ultimately received approval. Finally,
the researcher met with volunteer group facilitators and discussed appropriate procedures and
options to distribute the survey to their volunteers to control bias in this group. The researcher
instructed volunteer group facilitators how to answer general questions regarding the survey, and
how to answer questions unique to their particular situations. Hospice volunteers were unique in
that some were currently caregivers to a family or close friend, but all Hospice volunteers were
providing frequent and continuing "respite" caregiving to families, or family members, and
caring for someone with a terminal illness. When the researcher received permission to use the
Hospice volunteers, information regarding the process of recruitment was not provided.
However, the researcher was able to use this volunteer base after providing careful instruction to
group facilitators, a detailed set of instructions, and a precise script presenting any concerns.
Participants for the final study (N=160) came from some 10 counties in the North Central
Florida region, including Alachua County. The researcher traveled north to the Florida-Georgia
border, south to Weeki Wachi, as well as east and west between these locations to recruit
caregivers. At least 10 meetings were cancelled during the course of data collection due to
facilitators being unavailable or facilitators not having enough caregivers (participants) who
The original group of participating organizations grew in number during the course of data
collection due to referrals from the original organizations and from individuals. Only the
original organizations contacted before the pilot study began are listed (Appendix B) due to the
large number of organizations ultimately involved. One of the original organizations, the
Alzheimer's Association of North Central Florida, decided not to permit the researcher to attend
meetings and recruit volunteers.
Sample size was set by effect size and Type I error rate. Effect size in this study was
interpreted as meaning that no association existed among variables. Effect size was estimated at
10 participants per treatment. Type 1 error rate was set at &=05, and effect size (designated as
6d2) was set conservatively (-.10) for a projected pool of up to 200 participants (Cohen, 1988;
Practical Introduction to Power and Effect Size, pg. 1). Definitions of effect size are small (r =
+.20); medium (r = +.30), and large (r = +.50).
Estimation of power for a statistical test allows the researcher to determine how many
participants are necessary in each group to show adequate probability to assess significance, if
any. Power of a statistical test depends on alpha level or Type 1 error rate, effect, and sample
size. In general, sample sizes of more than 100 individuals are considered to have good
statistical power (Kaplan, 2006, para 1). The researcher attended meetings until 160 participants
had responded to the survey. This number was sufficient to support the data analyses and
provide adequate validity and reliability.
Participants in the final study completed the instrument in the same manner as in the pilot
study (i.e., at support group meetings, health forums, etc.). The number of social support groups,
health fair groups, and health forum participant groups ranged in size from three to 25,
depending on type of group and county in which the group or event took place. During the pilot
study, most surveys were completed at the beginning of a caregiver meeting. During the final
study, when the researcher attended meetings, health fairs, and forums, participants completed
the survey at the beginning of an event, during a lunch break, at the end of a meeting, or at other
times during the event.
As in the pilot study, talking by participants and non-participants created a noise factor. At
health fairs and health forums, the researcher sat at a designated table with chairs available for
participants to complete the survey. During these events, participants talked with the researcher
throughout the day. At adjoining tables, other health care providers offered blood pressure
checks, bone density tests, and other screening exams, often causing a distraction for participants
completing the instrument. This distraction often caused participants to take longer at their
responses. At some health fairs and forums, participants requested to take the survey to other
chairs and tables in the room to complete them. Some participants opted to sit outside the
building or at a quieter location to complete the instrument, which meant that the researcher was
not always in close proximity to the participants for questions regarding the survey. Some
participants (about 10%) declined the gift card incentive offered to them for completing the
Each completed instrument was assigned a number and recorded by type of group in the
codebook at the time of data collection. The number coincided with the specific support group
(i.e., l=Alzheimer's group, 2=general social support group, etc.). Other than the group code, no
other means existed to identify participants.
Data were obtained from volunteer participants who were 18 years or older, who were
currently caring full-time or part-time for a friend or family member 21 years of age or older, and
who were caring for the individual in their home or at another facility with frequent access.
The Keirsey Temperament Indicator II was coded according to temperament type in the
booklet as outlined in the test manual. Originally, each temperament type was coded as "0" for
not having the temperament, and "1" for having the temperament. For example, if a volunteer
was a "Guardian" temperament, only this temperament would have a "1" and the other three
would have a "0." When analyzing the significance of the research question, temperament types
were recorded. One temperament was recorded as "1," and then the others were recorded as "2,"
"3' and "4." Finally, the four types were combined under one variable for analyses (Meyer,
McGrath, & Rosenthal, 2003). Responses that were left blank (i.e., not answered) were coded as
"0" for items on the Keirsey Temperament Indicator II, and Caregiver Demographics. Blank
responses on the Caregiving Self-Efficacy Scale were also coded as "0."
Other categorical variables, such as total time in the caregiver role and daily time spent in
the caregiver role, also were recorded. After examining frequency tables for these variables,
categories were collapsed and recorded for a more accurate calculation in data analyses. All
Keirsey Temperament information was recorded on an Excel spreadsheet, followed by
information from the remainder of the survey, including the Caregiving Self-Efficacy Scale total
score, and caregiver demographic information, and then transferred to SPSS v. 14.0 for
After determining that the data set was normally distributed, descriptive statistics were
calculated to determine baseline frequency rates in each temperament category, average self-
efficacy scores, frequency rates for race/ethnicity, frequency rates for self-reported health status,
and frequency of self-reported difficulty in caring for the friend or family member.
Several types of data analyses were conducted on the final data set to determine a profile of
participants in the study. Bivariate analyses included 1-way analysis of variance (ANOVA) and
Spearman Rank Correlation measures that were used to analyze research questions one through
five. Analyses for all research questions were tested at a .05 significance level for 6. Multiple
Linear Regression was used to examine associations between each of the four temperament types
(explanatory variable), total self-efficacy score, total duration of care (outcome variables), and
relationship to the care receiver (covariate).
Univariate analyses can be used to present primary statistical information and to assess
patterns within a data set (McDermott & Sarvela, 1999; Noland, 2000). In this study, frequency
distributions and descriptive statistics were calculated to obtain baseline information about
frequency of gender, race/ethnicity, temperament type, mean self-efficacy scores, and all other
variables. These descriptive statistics were also used to determine a general profile of the
caregiver population for the study.
One-Way Analysis of Variance (ANOVA), an inferential statistic, can be used with more
than one independent variable, each of which may include two or more conditions. When only
two conditions exist in a single independent variable, using a One-Way ANOVA is equivalent to
a T-test. However, ANOVA has an advantage because it generalizes to experiments with more
than one independent variable, and may include more than two levels or conditions for each
independent variable (Weaver, 2003, para. 1; BBN Corporation, 1997, para 5).
ANOVA compares variance of scores within conditions. Variance of condition produces
the sum of squares within groups (SSwg), and sum of squares between groups (SSbg),
respectively. Output from each observation (experiment) consists of "variations from an overall
mean value" (Reliasoft Corporation, para. 2). These variations are due to factor or level, or to
variation due to random error. The ANOVA model used is,
y1/I- + Ti+ fij
where, P= the overall mean, r,= the level effect, and l-- the random error component.
Spearman Rank Correlation measures "correlation between two value sequences," usually
with one ordinal and one continuous variable. Each sequence is ranked separately, with rank
differences calculated at each position, i.
A nonparametric (distribution-free) rank statistic measures "strength of association
between two variables" (Weisstein, page 1, para. 3).
A Stepwise multiple linear regression model was used to conduct multivariate analyses for
the outcome (DV) and explanatory (IV) variables, and for any covariates. A multiple regression
model can be used when explanatory (IV) variables are continuous when we enter temperament
type as separate entities, as in the Keirsey Temperament Indicator II, and when the outcome
variable is continuous, as in the Caregiving Self-Efficacy Scale, and with a normal distribution,
as was indicated in this population. The four temperament types in the Keirsey Temperament
Indicator II were treated as four different explanatory (continuous) variables
Multiple regression can be used to predict, or learn, about the relationship between a
dependent variable and several independent variables (StatSoft, 2006; PA765, 2006). Multiple
regression examines the relationship between each of the IVs and the DV after "taking into
account the remaining IVs" (Rosner, 2000, p. 466-467).
Y = blXl+b2X2 ...+bnxn + c
Where Y is the dependent variable and "61z1 + 62z2 + 63z3" is the dependent variable, the
"b's" are regression coefficients, which represent the change in the dependent variable (y) when
the independent variable changes 1 unit. The "c" is the constant. Multiple regression is similar
to assumptions given in correlations in that the "relationship linearity is observed with the
absence of outliers" (PA 765, 2006). The error is assumed (PA 765, 2006).
Stepwise multiple regression, or statistical regression, computes the ordinary least squares,
used for the best fit of the regression line. In this process, the independent variable that
correlates best with the dependent variable is factored into the equation. Then, the rest of the
independent variables with the highest partial correlation with the dependent variable, are
computed after controlling for the first independent variable. Stepwise regression is most often
used to explore and predict research questions (PA765, 2006).
Instrument reliability for the Caregiving Self-Efficacy Scale and the Keirsey Temperament
Indicator II was generated based on final study population data using Cronbach's alpha measure
of internal consistency (Cronbach, 1951; Traub, 1994; Crocker & Algina, 1986). The
Caregiving Self-Efficacy Scale produced a total reliability of .90 (Table 3-1). Although not
included in the analysis of data for this study, reliability coefficients also were generated for the
three subscales of the Caregiving Self-efficacy Scale: .85 (Obtaining Respite), .87 (Controlling
Upsetting Thoughts about Caregiving), and .93 (Responding to Disruptive Patient Behavior).
To estimate reliability for the Keirsey Temperament Indicator II, each of the 70 items
completed by all 160 participants was coded as "1" for "a" and "2" for "b." This process
generated a Cronbach's alpha of .86 for the instrument total score. Although not included in the
analysis of data for this study, reliability coefficients also were also generated for E/I (.77), S/N
(.73), T/F (.70), and J/P (.72).
Questions for each dyad characteristic were grouped as follows: questions for E/I are
numbers 1, 8, 15, 22, 29, 36, 43, 50, 57, and 64; questions for S/N are numbers 2, 3, 9, 10, 16,
17, 23, 24, 30, 31, 37, 38, 44, 45, 51, 52, 58, 59, 65, and 66; questions for T/F are 4, 5, 11, 12,
18, 19, 25, 26, 32, 33, 39, 40, 46, 47, 53, 54, 60, 61, 67, and 68, and questions for J/P are 6, 7,
13, 14, 20, 21, 27, 28, 34, 35, 41, 42, 48, 49, 55, 56, 62, 63, 69, and 70.
Chapter 3 described the methods used to examine associations between different
temperament types and perceived self-efficacy among individuals serving in an informal
caregiver role. The chapter included a description of the research design, research variables,
instruments, the pilot study, the final study, and data analysis procedures. Data were collected
from informal caregivers for the pilot study during August and September 2005 and for the final
population from October 2005 through January 2006. A total of 25 caregivers participated in the
pilot study and 160 caregivers participated in the final population.
Descriptive statistics were calculated to determine baseline frequency rates in each
temperament category, average self-efficacy scores, frequency rates for race/ethnicity, frequency
rates for self-reported health status, and frequency of self-reported difficulty in caring for the
friend or family member. Bivariate analyses included 1-way analysis of variance (ANOVA),
and Spearman Rank Correlation measures were used to analyze research questions one through
five. Analyses for all six research questions were tested at a .05 significance level for 6.
Multiple Logistic Regression was used to examine associations between each of the four
temperament types (explanatory variable), total self-efficacy score, total duration of care
(outcome variables), and relationship to the care receiver (covariate). Chapter 4 discusses results
from analyses of the six research questions posed in the study.
Table 3-1. Instrument specification table
Scale Variable Type of Item Possible Responses Reliability
Variable Number For Each Item Coefficient
Keirsey Temperament Explanatory/ 70 2 .86
Caregiving Self-Efficacy Explanatory/ 15 100 .90
Table 3-2. Keirsey Temparament Indicator II Item to total correlation
Item Corrected Item-Total Correlation Cronbach's Alpha
if Item Deleted
Table 3-2. Continued
Corrected Item-Total Correlation
if Item Deleted
Table 3-3. Caregiving Self-Efficacy Scale Item to total correlation
Item Corrected Item-Total Correlation Cronbach's Alpha if
Item 1 (Respite) .57 .89
Item 2 (Respite) .55 .89
Item 3 (Respite) .50 .89
Item 4 (Respite) .47 .89
Item 5 (Respite) .53 .89
Item 6 (Upset) .55 .89
Item 7 (Upset) .59 .89
Item 8 (Upset) .55 .89
Item 9 (Upset) .56 .89
Item 10 (Upset) .48 .89
Item 11 (Disregard) .65 .89
Item 12 (Disregard) .68 .89
Item 13 (Disregard) .58 .89
Item 14 (Disregard) .69 .88
Item 15 (Disregard) .66 .89
Yj-=- + + ij
Figure 3-1: ANOVA Equation Model
Y = blxl+b2X2 ...+bnxn + c
Figure 3-2: Multiple Linear Regression Model
This study examined relationships that exist between different temperament types and
perceived self-efficacy among individuals serving in an informal caregiver role. This chapter
discusses characteristics of study participants including group type, gender, race, temperament
type, caregiver and care receiver age, overall health status, and duration of care. The chapter
also presents results from data analyses used to address the six research questions posed in the
Study participants included caregivers 18 years of age or older. All participants were
caring for an individual 21 years of age or older; caring for an individual with a chronic illness,
terminal illness, or other debilitating illness; caring for an individual on a full- or part-time basis;
and caring for an individual in their home or at a facility with frequent access by the caregiver.
Data collection procedures produced 160 usable surveys (64%), which was considered adequate
to support statistical tests used in data analyses.
The final population had a greater number of female caregivers (n=131, 82%), than of
male caregivers (n=29, 18%). Most of the literature on caregivers indicate a greater number of
female caregivers (about 70%), but the current population had a higher rate of female caregivers
possibly because approximately one-third of caregivers were single, divorced or widowed and
possibly husbands of caregivers were still working and not available to attend meetings.
Most participants were either White/Caucasian (85.6%) or African American (13.1%),
which is not an equitable representation of ethnicities in the United States, but possibly an
accurate indication of caregivers in the North Central Florida region. Only a few Hispanic
individuals and only one Native American caregiver completed the survey. The population in
the Pilot Study portion was about equal Caucasians and African Americans. It seems that few
Hispanic caregivers attended the meetings possibly because meetings were in rural areas, during
times of the day that were not convenient, or transportation was not available. Individuals may
not have had sufficient notice, may not have been able to drive to the location (i.e., sharing
vehicles in the family), or there may have been a language barrier in the case of Hispanic
caregivers. Therefore, results focused on males and females who reported their ethnicities as
White/Caucasian and African American.
Caregiver Group Type, Gender, and Race
Table 4-1 provides a summary of participants by type of group where information was
collected, gender, and race/ethnicity. The characteristics in Table 4-1 reflect the distribution of
participants affiliated with various organizations contacted by the researcher prior to collecting
information. The researcher sought an even distribution of participants from these organizations.
Distribution by gender and race were chosen based on similar studies in the literature, and these
variables were discussed and analyzed in this study.
Most caregiver information was collected at health fairs and forums (58.1%), followed by
general caregiver meetings (28.1%). Health fairs and forums attracted a larger and more diverse
population of caregivers. Caregiver meetings usually ranged in size from three to 10
participants. Most participants were female (82%) and White/Caucasian (85.6%). The literature
confirms that more females than males serve in a caregiver role (Cannuscio et al., 2004; Collins
et al., 1997). Some researchers estimate that as much as one-third (33%) of men serve as
caregivers. The male population of caregivers was much smaller in this study.
Temperament Type, Age, Overall Health, and Duration of Care
Table 4-2 shows frequency and distribution of temperament type, age, overall health, total
duration of time as a caregiver, and daily duration of time required in the caregiver role. Most
caregivers were aged 50 or older (n=133, 83.2%), and had provided care for at least two years
(n=129, 64.4%). On average, they provided less than a full day of care, defined as 6 hours or
less of care each day (n=97, 60.6%). Most caregivers generally described themselves as in
"good" health or better (n=134, 83.8%). In terms of temperament type, the 160 participants were
categorized as follows: Guardians (n=120, 75%), Artisans (n=22, 13.8%), Idealists (n=16, 10%),
and Rationals (n=2, 1.3%). Most caregivers, regardless of gender, were Guardians possibly
because some characteristics of this temperament include that they are more apt to take care of
themselves and others around them. Guardians are first and foremost seen as administrators -
take charge and get things done. Guardians are also cooperative and seek solutions. Similar to
the general population (Alpine Media, 2003), most individuals seem to be Guardians.
Variables in Table 4-2 were selected based on a review of related studies from the
literature. These variables were also used for comparisons in the current population of
caregivers. Although not used for comparison for data analysis in the current study, age emerged
as a particularly important variable because many caregivers were still caring for minor children,
parents, and other family members. When comparing caregiver age, caregiver health and total
duration of care (all three ordinal variables), age and health had a significant correlation, and
total duration of care and health had a significant correlation. The younger the age of the
caregiver, the better health reported by that caregiver, and the longer the caregiver was providing
informal care, poorer health was reported.
Keirsey presented a distribution of temperaments drawn from a much larger population,
and these participants completed the online version of the Keirsey Temperament Indicator II. A
number of caregivers in this study had some degree of access to computers, but most probably
could not have completed the online version of the instrument due to inadequate computer
access, time constraints, and lack of assistance in completing the protocol. Future studies on
caregiving and self-efficacy may address the issue of first time parenting, although a different
type of self-efficacy scale is required because of levels of expectations and pressures from family
and friends to achieve skills. The dynamics in the home, especially among two working
professionals, change with the addition of a child.
Temperament and Self-Efficacy
Temperament and self-efficacy were the independent variables used in the study. Among
caregivers in this population, all temperaments showed similar levels of self-efficacy. No
significant differences existed in how confident caregivers felt about providing assistance to the
Table 4-2 shows a comparison of frequency of temperament types for caregivers in this
study and those in a separate study of the general population. Table 4-3 shows a comparison of
frequency of gender in each temperament type. Table 4-4 provides frequency and distributions
of Temperament Type for the general population, not just those who served as caregivers (Alpine
Media Corporation, 2002).
Table 4-4 shows the distribution of temperament type in this study compared to the
national database compiled from individuals who completed the online version of the Keirsey
instrument. This distribution is presented only for the purpose of visual inspection. Although
the frequencies vary between the two groups, Guardians emerged as the most prevalent
temperament type in both populations. Conversely, Rationals were the least prevalent
temperament type in both populations.
The mean self-efficacy score of 973.03 (SD=280.39) indicated that caregivers were above
average in their confidence in the caregiver role, perhaps because many of the caregivers in this
study were spiritual and had excellent social support systems. Self-efficacy scores ranged from
350 for the lowest score to 1,500 for the highest score, from a possible maximum score of 1,500.
A one-sample Kolmogorov-Smirnov Test for population distribution indicated that the test
distribution was normal (D=.33). Therefore, the distribution for the current population was
normal (i.e., no distinguishable outliers).
Research Question One
Does a significant association exist in "Total Scale Scores" on the Caregiving Self-
Efficacy Scale between temperament type as determined by the Keirsey Temperament
No significant association existed in self-efficacy among the four temperament types. To
address this question, a one-way ANOVA was conducted on temperament type self-efficacy. No
significant association existed in the level of self-efficacy among the four temperament groups
(F=1.96, p=. 122) (Table 4-5).
Lack of a significant association may be explained in part by the fact that each
temperament category included both extrovert or expressive (E), and introvert or reserved (I)
characteristics. The E/I relationship dictates how individuals deal with significant others in the
outside world. In this population, whether a person is E/I may determine how that person relates
to those outside of the caregiver role, such as asking questions of physicians and other formal
health care providers, and in making medical decisions about the care recipient based on
interactions with formal health care providers (Keirsey, 1998). The E/I factor was not analyzed
in this population and should probably be analyzed separately. Since no temperament type
emerged as significantly different in level of self-efficacy, self-efficacy was used as an
independent variable (explanatory) to address research questions two through five.
Research Question Two
Does a significant association exist between "Total Scale Scores" on the Caregiving
Self-Efficacy Scale and daily duration of care provided by the caregiver?
A Spearman Rho Correlation was used, and no significant association existed in self-
efficacy for any of the six ranges (Appendix F) of time spent caring on a daily basis. To address
this question, a Spearman's rank order correlation was conducted with self-efficacy and daily
duration of care. No significant association was found in self-efficacy for any of the six ranges
(Appendix F) of time spent caring on a daily basis (r=-.048, P=.545) (Table 4-6). For
comparison purposes only, no significant association existed when a one-way ANOVA was used
Self-efficacy was not a factor in how long an individual cares for someone on a daily basis.
Thus, an individual who spends a brief amount of time each day in the caregiving role feels as
confident as those who spend more time in the caregiving role.
Several factors may help explain this finding. Based on the review of literature and the
characteristics of the study population, caregivers who care for individuals for more time during
the day (i.e, 6 or more hours) often feel more confident because they have less difficulty in
caring for that person, compared to caregivers who spend a brief amount of time (i.e., 1 or 2
hours) with someone with a more severe illness. Also, those who spend fewer hours in the
caregiving role may also hold full time jobs outside the home, whereas those spending more
daily hours in caregiving may not work outside the home. In addition, caregivers who spend
more time each day with the care receiver may have found an adequate social support system
(i.e., caregiver meetings), as well as other sources of informal support, such as close friends,
neighbors, and family members.
Although not part of the analysis for research question (see Question #6), multivariate
analysis indicated that caregivers who were not related to the care receiver tended to have shorter
durations as a caregiver. The variable of duration of care was combined with self-efficacy,
temperament, total duration of care and relationship to care receiver.
Research Question Three
Does a significant association exist between "Total Scale Scores" on the Caregiving
Self-Efficacy Scale and total duration of care provided by the caregiver?
A one-way ANOVA was used to show that a significant association existed between self-
efficacy and total duration of care, both with the seven original duration ranges (F=2.504,
p=.024), and with the four duration ranges (F=3.49, p=.017) (Table 4-7). For comparison, a
Spearman Rho correlation was used with the four original duration ranges (r=-.002, P=.982).
Due to low number of selections for some of the response options, the category for total
duration of care was recorded from seven ranges to four ranges. The first two variables, "6
months," and "6-12" months," were measured as time less than 12 months. Categories three and
four were combined, as were categories five and six, based on frequency distributions. As
individuals continued in the caregiver role, self-efficacy waned similarly during the first and
second years, and similarly during the third and four years. Category seven, "5 years or more,"
was retained as a discreet category. Caregiver self-efficacy was highest before one year, and
after five years, of caregiving.
A one-way ANOVA with LSD post hoc analysis, conducted on self-efficacy and the total
duration of care, showed significance both with the seven original ranges (F=2.504, p=.024), and
with the four recorded ranges (reported below, F=3.49, p=.017).
Individuals often feel less stressed as they enter the caregiver role. Over time, stress
increases and caregivers often feel depressed, and a decline occurs in their health. Their
involvement in friendship circles may decrease because caregivers cannot leave their homes to
socialize. Also, during the first few years, caregivers may quit jobs to provide full-time care.
Eventually, they may decide to place the care receiver in an assisted living facility. A transition
occurs in getting to know the facility staff and helping the staff understand the needs of the care
receiver, which creates new stressors for caregivers. Caregivers generally reach their limit of
caregiving in the home after about five years, so transitioning into an assisted living facility may
be viewed as getting the situation under control.
Whether the care receiver resided in the home or not was not addressed as a research
question but, a t-test showed significance. Caregivers who cared for care receivers outside the
home had higher levels of self-efficacy.
Research Question Four
Does a significant association exist between "Total Scale Scores" on the Caregiving
Self-Efficacy Scale and intensity/difficulty of care provided by the caregiver?
A Spearman Rho Correlation was used to show a significant relationship between self-
efficacy and the degree of intensity/difficulty (ordinal) using the five original response options
from the instrument (two-tailed, r=-.184, p=.020) (Table 4-8). For comparison a one-way
ANOVA was used (F=1.480, p=.211).
Caregivers who reported low degrees of difficulty caring for their loved ones showed
higher levels of self-efficacy. Most caregivers described their responsibilities as "not difficult at
all," "a little difficult," or "difficult." Only a few described their responsibilities as "very
difficult" or "extremely difficult. However, as difficulty increased in the caregiving role, self-
Caregivers usually feel personally responsible for their loved ones. Even if they feel some
level of stress in the caregiver role, they feel a personal obligation and empathy to provide care.
Caregivers who provide care, such as assistance with Activities of Daily Living (ADL) versus
Instrumental Activities of Daily Living (ADL), often do not consider this type of assistance as
difficult or even as providing care, but merely as part of their role as a spouse, child, or close
family member (Arno et al., 1999).
Research Question Five
Does a significant association exist in self-efficacy scores as determined by "Total
Scale Scores" on the Caregiving Self-Efficacy Scale and caregiver health status?
A Spearman Rho Correlation was used to identify a significant relationship existed
between self-efficacy and caregiver health, using the five original ranges from the instrument
(two-tailed, r=-.159, P=.045) (Table 4-9). For comparison, a one-way ANOVA was used
Caregivers who feel healthier cope better in the caregiver role, and they feel more
confident making important decisions for the care receiver. When caregiver health declines,
caregivers do not feel as confident that they can provide the quality of care needed for their loved
one. They may also feel that they cannot meet their own health needs due to the amount of time
required in the caregiving role. As their health declines, especially if their health was poor when
they began the caregiver role, caregivers may decide to place the care receiver in an assisted
living facility sooner than anticipated. This decision may produce both relief and guilt for not
continuing to care for their loved one. Caregivers may also feel fairly healthy compared to the
Research Question Six
Does a significant association exist between overall duration of care and temperament
type as determined by the Keirsey Temperament Indicator II, self-efficacy scores as
determined by "Total Scale Scores" on the Caregiving Self-Efficacy Scale, and relationship
to care receiver?
Significant associations existed between level of self-efficacy, one temperament type, and
two separate relationship categories to the care receiver (Table 4-10).
Among the four temperament types, self-efficacy level did not influence daily duration of
care. In terms of the four temperament types, Artisans were most likely to remain the longest in
the caregiver role. Among this population of caregivers, Artisans were the second most frequent
among the four temperaments, but small in number (n=22). Among the other three
temperaments, total duration in the caregiver role was not significantly different. As one
characteristic of the Artisan personality, they are more likely to follow a philosophy of living "in
the moment." Artisans enjoy life and seek resources that suit their needs (Keirsey, 1998).
Therefore, Artisans may be more resourceful in finding outlets for their stress in the caregiving
role, may not need formal types of support (i.e., attending caregiver meetings) and they may be
more willing to accept assistance from friends and family in managing care for the care receiver.
Two relationships between care receiver and caregiver also were significant, but in the
negative direction. First, if the care receiver was a friend or had a relationship other than the
options listed (spouse, parent, child, grandparent, in-law, or partner), duration of care tended to
be briefer. Perhaps some form of regret or a different level of frustration exists when the care
receiver is not a family member. Second, relationships in the "other" category included those
who provided respite care to families or to neighbors. The other category of relationships
seemed to indicate a more emotional attachment to the care receiver; therefore, care provided
was longer in duration.
Financial support may be a predictor of total duration of care. If the caregiver has
adequate financial resources, their total duration of care may be longer because of access to
formal caregivers (i.e., home care nurses). Caregiver satisfaction and confidence may be
implications to understanding caregivers for health care providers and their recommendations for
This chapter reported the results from examining responses of individuals serving in an
informal caregiver role by group type, gender, race, temperament type, age overall health status,
and duration of care. A population profile was generated showing the distribution of participants
by temperament type, total self-efficacy score, and distribution of total time in the caregiver role.
Most caregivers were White/Caucasian females, older than age 50, who cared for at least one
Bivariate analysis indicated all four temperament types showed comparable self-efficacy
levels. No one temperament produced a higher level of self-efficacy than any of the other three.
Total level of self-efficacy did not affect amount of time spent on a daily basis in the caregiver
role, perhaps because some individuals caring for a person one or two hours each day also held
full time jobs outside the home, creating an added stressor.
Total level of self-efficacy was significantly affected by total time (duration of care) in a
caregiver role, degree of intensity/difficulty in the caregiver role, and self-reported caregiver
health. Caregivers felt more confident very early on, possibly because of higher levels of social
support, and also after five years in the caregiver role, perhaps due to comfort level and other
assistance received in caring for their loved one. Most caregivers did not report that caring for a
loved one was "very difficult" or "extremely difficult," but the more difficulty in the caregiver