6% Egg

5.x."

its

.03.“...
.15: 1
2....

v. 3;.
3.1.5

 

 

I: .-
11396 .11... J,
.1 an“ I I «
2..., . 3..
.32

. it’ll...
P flit:

 

200%

This is to certify that the
thesis entitled

DEVELOPMENT AND PILOTING OF THE EMPLOYEE
PERCEPTIONS OF BREASTFEEDING SUPPORT

 

_LlBFlARY
M'Chigan State
University

 

 

 

QUESTIONNAIRE

presented by

Sally W. Greene

has been accepted towards fulﬁllment
of the requirements for the

Master of degree in Food Science and Human

Science

Nutrition

 

 

4’16 ck N £104”

 

Major Professor’s Signature
I / c9 4! O 8

Date

MSU is an afﬁnnative-action, equal-opportunity employer

_. .-._.—--.—.-.—._--.u— -n.-.-3-.-.-o-I-u-u-u-u-u-u-c-n-I-o-o-o-¢-I--'-_L_..-.~c-o-u-o-u-I-b—u-I-I-o--c.

 

 

 

 

 

 

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

DATE DUE DATE DUE DATE DUE

 

DEVELOPMENT AND PILOTING OF THE
EMPLOYEE PERCEPTIONS OF BREASTFEEDING SUPPORT
QUESTIONNAIRE

BY

SALLY W. GREENE

A THESIS
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
MASTERS OF SCIENCE
Department of Food Science and Human Nutrition

2008

 

ABSTRACT
DEVELOPMENT AND PILOTING OF THE

EMPLOYEE PERCEPTIONS OF BREASTFEEDING SUPPORT
QUESTIONNAIRE

BY

SALLY WEBSTER GREENE

Maternal employment has been reported as a key factor for discontinuing
breastfeeding, so it is vital to identify work climate components perceived by women
as barriers to breastfeeding. The purpose of this study was to develop an instrument
to quantify perceptions of the work climate for breastfeeding support. Procedures
based on Messick’s (1995) ‘uniﬁed framework’ as modiﬁed by Wolfe et a1. (2000,
2004, 2006, 2007) were used. The ecological perspective informed the instrument
subscales: company, manager, coworkers, workﬂow, and physical environment.
Items were developed based on literature and expert review. One-on-one interviews
with working mothers that had given birth in the last year (n=14) and a review by
experts (n=1 1) including practitioners (with credentials and experience in the
lactation ﬁeld) and researchers (with experience in survey development or lactation
research), assessed subscales, veriﬁed survey content and clariﬁed item wording. The
instrument was piloted with non-managerial employees who were pregnant or had
given birth within the last year (n =104). Fit analysis was used to determine misﬁt of
items within a pre-deﬁned model. Items that exhibited misﬁt were scrutinized to
determine if they should be removed based on substantive reasoning. The resulting
instrument contained two dimensions, one based on company policies and practices

(23 items), and the other on manager and coworker support (1 8 items).

 

 

wk?" ‘
WW”

1‘0 hm l
‘rmmmm 'l I' , a-
"3"“ , . . . . .. :. i '15:??? its
.. m .'_ Pal". "Z 'I k’HK

 

   

M 'ILNOWLRDGEMEN'IS .. s}; “I"

. ft : z.
'9
* it). .. '
- 97' , .5. .t. ._
I would like to nut-nil my heartfelt appreciation to my advlsuh m ‘ y ' 'z‘u‘w ‘.
in .r' ‘ a
father tinvnivcnnp su- \pot'l arid Mughtt'ul guidance I am grateful him light-3'.
" 'J 1' '
It.o‘i* no.3).~ “i ,
Ind otfon in tackling, Ii‘ll". IInpOtInuL )cl .mpical, maiden project. Inns“ ‘ l ‘1 13:0“

 
  
   
  
   
  
  

"our and Dr Katherine \iaim » to! their encouragement and "NM“! Em
“H0190 like to thank D: bdwaix. W Wull’cof V irgima Tech lid I}. u

.l' H. 3.3 ' V"i
‘ ‘5" 3" ‘ ' .
Nmuofticorgia Tech. Iama'mgratzﬁil forthc Minlmmé i .
., .i'\' I“ ‘ :: “A I ' ”
Collcgcof AgricultumWI'M“ '

4

f ‘ . Tormlabmla. lrxkﬂolknﬁwplmww -
if” an SeuhHojruski inimmmmwummiim“
'i" .

in“ jnlncywhhenulioiwu. Aspaialmmbgoutormfu
.( .i‘..i

    

QWandfordmringthctnalsofbelngnmm ‘
','.-i.-i, ' - .

Tommmmﬂlgmnkahuﬂmhmm.

Itr
t: I 3‘ '
t, W
\ rt 3.! V \ i
it "e "3;” \ - 3* ,i
i;
' ﬁ‘ 7

ACKNOWLEDGEMENTS

I would like to extend my heartfelt appreciation to my advisor Dr. Beth Olson
for her unwavering support and thoughtful guidance. I am grateful for all of her time
and effort in tackling this important, yet atypical, nutrition project. It was a road of
discovery for both of us. I admire your dedication to promoting breastfeeding
through scholarly means. I would like to thank my committee members Dr. Sharon
Hoerr and Dr. Katherine Alaimo for their encouragement and insightful comments. I
would also like to thank Dr. Edward W. Wolfe of Virginia Tech and Dr. Ingrid
Fulmar of Georgia Tech. I am also grateful for the ﬁnancial support provided by the
College of Agriculture & Natural Resources Graduate School Fellowship Funds.

To my lab mates: Tracie Bolton, Stephanie Schoemer, Shabnam Momin, Tan
Chow, Sarah Hojnacki, and Krystyna Iwanski: it has been a pleasure sharing my
journey with each of you. A special thanks goes to Tracie, for always having time for
my questions and for sharing the trials of being a working mother.

To my ﬁiend Anda Zile, thanks for asking the hard questions, what more is
there to say than, ‘Cheers’.

I could have never got where I am today without the support, both ﬁnancial
and emotional, of my parents. Thank you for believing in me. Finally, to my partner
Mike Smola and daughter Cassidy, all of your love and support is just too big for me

to wrap words around.

TABLE OF CONTENTS

List of tables ........................................................................................ xiii
List of ﬁgures ......................................................................................... ix
List of abbreviations ................................................................................... x
Chapter 1
Introduction ........................................................................................ 1
1.1 Background ........................................................................... 1
1.2 Rationale .............................................................................. 5
1.3 Research goal ......................................................................... 6
1.4 Research objectives .................................................................... 6
Chapter 2
Literature review
2.1 Health beneﬁts of breastfeeding ................................................... 7
2.2 Economic and environmental beneﬁts of breastfeeding ....................... 10
2.3 Breastfeeding recommendations .................................................. 11
2.4 Prevalence of breastfeeding ....................................................... 13
2.5 Prevalence of breastfeeding characterized by employment status. . . . . . . . ....15
2.6 Impact of employment of breastfeeding rates in Michigan .................... 21
2.7 Federal legislation and breastfeeding ............................................. 22
2.8 State legislation and breastfeeding ................................................ 23
2.9 “F amily-friendly beneﬁts” ........................................................... 24
Chapter 3
3.1 Phase 1 — Instrument development ................................................ 32
3.1.1 Instrument purpose ...................................................... 33
3.1.2 Theoretical framework of the instrument ............................. 33
3.1.3 Test speciﬁcations ....................................................... 36
3.1.4 Expert review of the test speciﬁcations ............................... 37
3.1.5 Description of subscales ................................................ 39
3.1.6 Item development ........................................................ 41
3.1.7 Recruitment for cognitive interviews ................................. 43
3.1.8 Cognitive interviews ..................................................... 45
3.1.9 Cognitive interview procedures ........................................ 47
3.1.10 Expert review of survey items ........................................ 51
3.2 Phase 2 — Piloting of the Instrument ‘The Employee Perception of
Breastfeeding Support Questionnaire’ (EPBS-Q) ................. 52
3.2.1 Recruitment of subjects for piloting the EPBS-Q ................... 53
3.2.2 Procedures for piloting testing of the EPBS-Q ...................... 54
3.3 Scaling ................................................................................. 55
vi

1‘

Chapter 4 '
“Development of The Employee Perceptions of Breastfeeding Support

Questionnaire” ....................................................................................... 60
4.1 Abstract ................................................................................ 60
4.2 Introduction ........................................................................... 61
4.3 Materials and methods ............................................................... 62
4.3.1 Study Design ............................................................ 62
4.3.2 Literature review ........................................................ 63
4.3.3 Expert reviews .......................................................... 64
4.3.4 Interviews ................................................................ 64
4.4 Results .................................................................................. 66
4.4.1 Literature review ........................................................ 66
4.4.2 Expert review ............................................................ 67
4.4.3 Interviews ................................................................ 69
4.5 Discussion ............................................................................. 74
4.6 Conclusion ............................................................................ 77
4.7 Acknowledgements .................................................................. 78
4.8 Reprint information .................................................................. 78
Chapter 5
“Assessing the Validity of Measures of an Instrument Designed to Measure Female
Employees’ Perceptions of Workplace Breastfeeding Support” ............................. 79
5.1 Abstract ............................................................................... 79
5.2 Introduction .......................................................................... 79
5.3 Materials and methods ............................................................. 82
5.3.1 Study Design ............................................................ 82
5.3.2 Subjects .................................................................. 82
5.3.3 Measures ................................................................. 84
5.3.4 Scaling ................................................................... 84
5.4 Results ................................................................................ 87
5.5 Discussion ........................................................................... 88
5.6 Conclusion ........................................................................... 91
5.7 Acknowledgments .................................................................. 92
5.8 Reprint information ................................................................. 92
Appendices ........................................................................................... 93
Appendix A .............................................................................. 94
Appendix B ............................................................................. 120
Appendix C ............................................................................. 131
Bibliography .................................................................................. 141

LIST OF TABLES

Chapter 2

Table 1. Rates of Q1 breastfeeding in the United States .................... 12
Chapter 3

Table 2. Table of speciﬁcations ................................................. 36

Table 3. Test blueprint ............................................................ 38
Chapter 4

Table 4. A review by experts resulted in the addition of one new subscale and
twelve new items within the ﬁve subscales ............................... 68-69

Table 5. One-on-one interviews resulted in the addition of six new items
within four of the ﬁve subscales ............................................... 71

Table 6. One-on-one interviews resulted in the rewording of ﬁfteen items

within the ﬁve subscales .................................................... 72-74
Chapter 5

Table 7. Demographics of the Sample ......................................... 83

Table 8. Dimensionality Statistics ............................................. 88

Table 9. Subscale Correlations and Reliabilities .............................. 88

Table 10. Summary of Item Quality Indicators .............................. 89

viii

LIST OF FIGURES

Chapter 3
Figure l. Ecological model as related to the work climate ............... 35
Figure 2. Three models to depict a relationship between items .......... 57

 

LIST OF ABBREVIATIONS

 

BIC ............................................................ Bayesian Information Criteria
CoNLSY (‘hildren of the National Longitudinal Study of Youth
DO .......................................................... Doctor of Osteopathic Medicine
EPBS-Q ............ The Employee Perception of Breastfeeding Support Questionnaire
FMLA .............................................. Family and Medical Leave Act of 1993
HMO ........................................................ health maintenance organization
HP2010 ................................................................... Healthy People 2010
IBCLC ............................... International Board Certiﬁed Lactation Consultants
LLLI ........................................................... La Leche League International
MBA ...................................................... Master of Business Administration
MI PRAMS .................. Michigan Pregnancy Risk Assessment Monitoring System
MRCMLM ......... Multidimensional Random Coefﬁcients Multinomial Logit Model
NIS ............................................................ National Immunization Survey
NLSY79 .............................................. National Longitudinal Study of Youth
NMIHS ....................................... National Maternal and Infant Health Survey
SIDS ............................................................ sudden infant death syndrome
SRHM ........................................... Society for Human Resource Management
W1C ....... Special Supplemental Nutrition Program for Women, Infants, and Children

 

Chapter 1
INTRODUCTION

1.1 Background

Breastfeeding is the gold standard for infant feeding in the United States.
Breastfeeding has numerous short- and long—term beneﬁts for infants, mothers,
and society (AAP, 2005; ADA, 2005; Dewey, Heinig, & Nommsen-Rivers, 1995;
M. H. Labbok, 2001; Raisler, Alexander, & O'Carnpo, 1999). Due to the
substantial beneﬁts attributed to breastfeeding, it is recommended that new
mothers exclusively breastfeed for 6 months, and continue breastfeeding for at
least one year or longer, if mutually desired, except in rare circumstances (AAF P,
2005; AAP, 2005; ADA, 2005; Satcher, 2001; US Surgeon General, 2000; World
Health Organization, 2003). Despite these recommendations and the growing
body of literature describing the beneﬁts of breastfeeding, breastfeeding rates
remain low. Although 70% of women start breastfeeding, only 36% remain
breastfeeding at 6 months (Li, Darling, Maurice, Barker, & Grummer-Strawn,
2005). Rates are even lower among women who work ("Breastfeeding Trends
Through 2002", 2003).

There are many personal and social factors that inﬂuence why women
choose not to breastfeed. One factor clearly associated with reduced
breastfeeding rates is full-time employment (McKinley & Hyde, 2004; Visness &

Kennedy, 1997). As increasing numbers of women have entered the workforce,

‘L

 

the subsequent impact of employment on breastfeeding rates has been examined
in various studies. Although national surveys have demonstrated that rates of
breastfeeding initiation are similar among women, regardless of work status
(Ahluwalia, Morrow, & Hsia, 2005; , "Breastfeeding Trends Through 2002",
2003; Kimbro, 2006), similar analyses have shown disparities in rates of
breastfeeding duration among women working full-time, as opposed to those
working part-time or not working outside of the home (Ahluwalia, Morrow, &
Hsia, 2005; Berger, Hill, & Waldfogel, 2005; Chatterji & Frick, 2003). Berger et
a1. conducted an analysis using 1987-2000 data from the National Longitudinal
Survey of Youth. This analysis found that compared to mothers who stayed at
home for 12 weeks following delivery, women who returned to work at 12 weeks
breastfed, on average, 6 fewer weeks.

In 2006, women represented 46% of the total United States labor force,
with 75% of employed women working full-time ("Women in the Labor Force in
2006"). More importantly, the Current Population Survey for 2005 found that
among mothers with children younger than one year old, 50% are in the labor
force. Of these women, two-thirds are employed full-time (U .S. Department of
Labor Bureau of Labor Statistics, 2006). Gaining an understanding of why full-
time employment negatively impacts rates of breastfeeding duration is therefore
essential.

Work climate, deﬁned as “shared perceptions of organizational policies,
practices, and procedures, both formal and informal” (Scheider, 1990), inﬂuences

what behaviors are viewed as acceptable in the workplace. One possible reason

 

why women working full-time have reduced breastfeeding rates is that new
mothers do not perceive the work climate as being supportive of breastfeeding.
Yet little research has been conducted to investigate perceptions of the work
climate from the viewpoint of new mothers. A majority of the research to date
has employed retrospective interviews or questionnaires with women who
successfully breastfed in order to characterize their breastfeeding experiences
(Auerbach & Guss, 1984; Hills-Bonczyk, Avery, Savik, Potter, & Duckett, 1993;
MacLaughlin & Strelnick, 1984; Meek, 2001; Reifsnider & Taylor-Meyers, 1885;
Rojjanasrirat, 2004). These studies found that among supportive factors cited,
those directly related to the workplace included: support of coworkers, support of
management, adequate break time, ﬂexible schedules, and a clean, private place
for expressing and storing breast milk. Most often if a barrier to breastfeeding
was cited it was due to a lack of one or more of these support components within
the workplace. The results of these studies give insight into why women were
able to continue breastfeeding while employed, but they do not increase
understanding about which of the factors were most instrumental in supporting
breastfeeding.

Similarly, employer—sponsored lactation support programs, in general,
have been shown to increase rates of breastfeeding duration. It is difﬁcult to
assess which component(s) of the support programs are most inﬂuential in
increasing breastfeeding rates, however, because the programs evaluated varied in
the amount and types of support offered (Cohen & Mrtek, 1994; Katcher &

Lanese, 1985; Ortiz, McGilligan, & Kelly, 2004). As with previous studies,

analyses focused on the overall effectiveness of lactation support programs in
increasing breastfeeding rates, but they did not seek to examine which
components of the program were perceived by new mothers as being most useful
to their attempts to continue breastfeeding. A practical way to discern which
components of a support program are most effective at increasing breastfeeding
rates is to develop an instrument to assess the workplace climate from the
viewpoint of new mothers. Additionally, development of an instrument will help
to standardized data collection methods.

Workplace climates have been studied as they relate to work-family
balance (Gault & Lovell, 2006; Kopelman, Prottas, Thompson, & Jahn, 2006;
Kossek, Colquitt, & Noe, 2001), and one aspect of work-family balance is
provision of breastfeeding support. There has, however, been little research on
work climate for breastfeeding support. Seijts conducted two studies that
examined the perceived fairness of providing breastfeeding accommodations in
the workplace (Gerald H. Seijts, 2002; Gerard H. Seijts, 2004). The purpose of
studying coworkers’ perceptions was based on the idea that if employees do not
view accommodations as being fair, then companies are less likely to offer them.
A scenario (based on a real-life incident) was presented to a sample of business
students, participants in a Mid-Management program, and alumni of a large
business school. The scenario described breastfeeding difﬁculties an employee
experienced at work after returning from maternity leave. The results of both
studies were similar in that, overall, a company was viewed as more attractive by

all participants when breastfeeding accommodations were offered. When the

scenario included a personal cost for the coworker (i.e. having to cover work
duties of a nursing mother) perceptions of fairness were lowest among those
participants without children. The results suggest that one barrier to provision of
breastfeeding support may be that coworkers believe accommodations provided to
select employees (such as breastfeeding mothers) might unduly burden other
employees. Unfortunately, no other aspects of the work climate have been
studied, and no studies have looked directly at how the work climate for

breastfeeding support is perceived by new mothers].

1.2 Rationale

Little research has been conducted to investigate perceptions of the work
climate from the viewpoint of new mothers. A majority of research in this area
has been conducted by interviews or focus groups, which lack of a standardized
method for collecting women’s perceptions. It is necessary therefore to develop a
means by which to collect data in a systematic way. The goal of this study,
therefore, was to develop an instrument to assess the work climate for
breastfeeding support as perceived by new mother employees. This instrument
could be used by companies seeking to measure perceptions of breastfeeding
support within their company by new mother employees. The company could not
only get a sense of the overall climate, but also an assessment of which
components of support may be lacking within their company. Due to the

systematic nature of collecting data with an instrument, this instrument could be

 

lA search on ProQuest (Business — ABI/INFORM Global), Medline (Silverplatter WebSPIRS, and
Entrez PubMed) for scholarly journals (including peer-reviewed) for “work climate” and
“breastfeeding” yielded no results.

 

used to compare perceptions across companies. Once the work climate, and areas
where support may be lacking has been evaluated, more effective and efficient
breastfeeding support programs could be developed and the impact on

breastfeeding rates could subsequently be measured.

1.5 Research goal
To develop an instrument to quantify perceptions of the work climate for
breastfeeding support from the viewpoint of new mothers employed full-time in

non-managerial positions.

1.6 Research objectives
(1) To develop items for the instrument
(2) To test and reﬁne the instrument using qualitative research methods
(3) To pilot the instrument and then use quantitative research methods in
order to determine which model the data best ﬁts, to assess item-ﬁt within
the model and to examine several indices relating to the quality of the

measures produced for each subscale.

CHAPTER 2
LITERATURE REVIEW

2.1 Health beneﬁts of breastfeeding

Throughout history, breastfeeding has been the traditional method of
infant feeding. The increased availability of sanitary, convenient, commercially
prepared infant formula from the 1930’s through today has, however, negatively
impacted breastfeeding rates (F omon, 2001 ). Breastfeeding rates showed a
downward trend through the 1950’s and 60’s, and by the early 1970’s, the United
States had the lowest rates of breastfeeding ever recorded. It is estimated that in
the early 1970’s, fewer than 25% of infants were breastfed at age 1 week, with
approximately 14% receiving breast milk until 2 or 3 months of age (Fomon,
2001). Infant formula is designed to be nutritionally equivalent to breast milk, in
spite of this, research on the composition of human milk, and on the health
outcomes associated with breastfeeding, has established that breastfeeding is more
beneﬁcial than formula feeding (DHHS, 2000).

In the immediate postpartum period, deﬁned as the ﬁrst four days after
birth, the breasts produce colostrum. Colostrum is beneﬁcial for infant health in a
number of ways. Colostrum is easy for infants to digest because it is secreted as a
low-volume high-nutrient food containing the proper ratio of carbohydrates,
protein and lipids. Colostrum has a laxative effect that is important for preventing

jaundice because excessive bilirubin is excreted ("What is colostrum? How does it

beneﬁt my baby?" 2006). Colostrum also contains antibodies in the form of
secretory immunoglobulin A and high concentrations of the protective white
blood cells known as leukocytes. Infants breastfed even for a limited amount of
time after birth derive beneﬁts from colostrum (American Academy of Pediatrics,
2005)

The health beneﬁts do not end with colostrum. Infants fed breast milk
have lower rates of morbidity, especially from infectious diseases. Research
shows that breastfed infants have decreased incidences and/or severity of many
infectious diseases including diarrhea (Heinig, 2001; Raisler, Alexander, &
O'Carnpo, 1999), respiratory tract infections (Bachrach, Schwarz, & Bachrach,
2003; Oddy et al., 2004), and otitis media (Ball & Wright, 1999; Dewey, Heinig,
& Nommsen-Rivers, 1995) when compared to infants fed commercially prepared
infant formula. Data collected as part of the 1988 National Maternal and Infant
Health Survey (NMIHS) was used to assess the effect of breastfeeding on
postneonatal mortality (death of infants at ages 28 days and over but under one
year) in the United States. Analysis of the NMIHS data revealed that if all infants
were breastfed, as many as 720 postneonatal deaths could be prevented or delayed
each year, including those from sudden infant death syndrome (SIDS) (Chen &
Rogan, 2004).

Growing evidence suggests that breast milk protects a child even after
feedings have been discontinued. One ﬁnding involved a reduced risk of
cardiovascular disease in a group of 13-16 year olds when comparing those

children that had been fed breast milk as an infant vs. those fed formula (Singhal,

. ‘

Cole, Fewtrell, & Lucas, 2004; Singhal, Cole, & Lucas, 2001). Another study,
using longitudinal data collected by the Centers for Disease Control and
Prevention, demonstrated that an increased duration of breastfeeding was
associated with a lower risk of overweight in a low-income population of non-
Hispanic white 4-year-olds in the United States (Grummer-Strawn & Mei, 2004).
Breast milk feedings have even been associated with long-term cognitive
advantages (Evenhouse & Reilly, 2005; Horwood & Fergusson, 1998; Smith,
Durkin, Hinton, Bellinger, & Kuhn, 2003).

The very act of breastfeeding also provides beneﬁts for both mother and
child. One beneﬁt for both, is the physiological beneﬁt of a strong infant-mother
bond that can be created due to the physical closeness of breastfeeding (Else-
Quest, Hyde, & Clark, 2003). Beneﬁts speciﬁcally for mothers include less
postpartum bleeding and a return of the uterus to prepregnancy size sooner, due to
the increased oxytocin- levels created by breastfeeding (Heinig, 1994).
Breastfeeding is also associated with a lower risk for ovarian cancer (Rosenblatt,
1993) and premenopausal breast cancer (N ewcomb, 1994). Additionally, it is
suggested that breastfeeding is associated with less retention of weight that is
gained during pregnancy (AbuSabha & Greene, 1998; Dewey, Heinig, &
Nommsen, 1993; Kac, Benicio, Velasquez Melendez, Valente, & Struchiner,

2004).

 

2.2 Economic and environmental beneﬁts of breastfeeding

There are also economic beneﬁts of breastfeeding that extend to parents
and society. Parents experience direct ﬁnancial beneﬁts because they are not
purchasing infant formula and related accessories. The savings can be substantial
when considering that standard formula alone costs at least $700 per year (ADA,
2005). Parents of breastfed infants may also pay less in out-of-pocket medical
costs (e. g. doctor’s visits, medications) since breastfed infants have decreased
onset and severity of many common childhood diseases when compared to
formula-fed cohorts (Weimer, 2001; Montgomery et a1. 1997; Splett et al. 1998).
It is estimated that a minimum of $3.6 billion could be saved in health care costs
in the United States if current breastfeeding rates were raised to meet public
health goals (Weimer, 2001).

Because breastfed infants tend to be less sick, breastfeeding mothers do
not experience the saine wage loss compared to women who formula fed their
infants (Cohen et al., 1995). This in turn is an economic beneﬁt to employers.
Employers who actively support breastfeeding among female employees report
lower employee absenteeism due to infant illness, increased employee loyalty and
productivity, lower turnover rates, and an enhanced company image for being
family-friendly (Weimer, 2001; Cohen et al., 1995).

Breastfeeding is also a beneﬁt to society because it is environmentally
ﬁ'iendly. Breast milk is a renewable resource and its production does not create
any pollution. An increase in breastfeeding rates would be expected to decrease

the need for manufacturing, packaging, shipping, and disposal of infant formula

 

and related accessories, all of which are processes that require the use of non-
renewable resources such as fossil fuels (Radford, 1991; Institute for

Reproductive Health/IMPACT, 1990).

2.3 Breastfeeding recommendations

The US. Department of Health and Human Services, the American
Academy of Pediatrics, the US. Surgeon General, the American Dietetics
Association and the World Health Organization all recommend that mothers
exclusively breastfeed for the ﬁrst 6 months and continue breastfeeding for at
least 12 months, or longer, if mutually desired, with the addition of appropriate
complementary foods (AAFP, 2005; AAP, 2005; ADA, 2005; Satcher, 2001; US
Surgeon General, 2000; World Health Organization, 2003). The World Health
Organization goes further by recommending that breastfeeding continue until at
least 2 years of age, with appropriate complementary foods added after 6 months
of age (World Health Organization, 2003).

Despite these recognized beneﬁts and recommendations, many women do
not initiate breastfeeding. Even among those women who start breastfeeding, a
signiﬁcant proportion adds infant formula or complementary food feedings, or
discontinues breastfeeding sooner than recommended. Because breastfeeding
recommendations are not being met, the federal government has established
national public health objectives for breastfeeding initiation and duration. The
objectives set by the US Department of Health and Human Services Healthy

People 2010 is to increase the proportion of mothers who breastfeed their infants

 

breast milk in early postpartum, at 6 months, and at one year (see Table 1)
(DHHS, 2000).

Table 1. Rates of guy breastfeeding in the United States

 

 

 

Time 1998 Baseline HP 2010 Objective “
(°/0) (W
In early postpartum 64 75
At 6 months 29 50
At 1 year 16 25
(DHHS, 2000)

a Healthy People 2010 Objective 16-19 (a — c).

These objectives focus only on the proportion of women that feed their
infants any breast milk, as opposed to only breast milk, therefore, one outcome of
the mid-term review of the Healthy People 2010 objectives was to add another
objective for exclusive breastfeeding. This objective is to increase the proportion
of mothers who exclusively breastfed their babies from birth through three
months to 60% and who exclusively breastfeed through six months to 25%
(Healthy People 2010 midcourse review: revised objectives for 16-19).

Until recently, national data collection of breastfeeding rates, especially
exclusive breastfeeding rates, has been insufficient. Appropriate monitoring of
the breastfeeding rates for the HP2010, therefore, is a high priority for the US
Department of Health and Human Services. To correct for this deﬁciency, the
Centers for Disease Control and Prevention held a meeting in 2001 to review

current practices for measuring breastfeeding rates. As a result, three questions

.. a.

on breastfeeding initiation and duration were added to an existing surveillance
system, the National Immunization Survey (N IS), to facilitate immediate data

collection (Li, Zhao, Mokdad, Barker, & Grummer Strawn, 2003).

2.4 Prevalence of breastfeeding

The National Immunization Survey (N IS) is conducted annually by the
Centers for Disease Control and Prevention and uses random-digit dialing to
survey households with age-eligible children. Sample estimates, comprised of
approximately 30,000 children, are weighted to represent the entire population of
the United States nationally and by region, state, and selected large metro areas.
Data collected in 2005 by the NIS demonstrated that rates of breastfeeding were
the highest ever in recorded history (Li, Darling, Maurice, Barker, & Grummer-
Strawn, 2005). The greatest advances occurred in rates of breastfeeding initiation,
which were close to reaching national public health goals (see Table 1). An
estimated 73% of women initiated breastfeeding in the early postpartum period,
approaching the target goal of 75%. Advances in rates of breastfeeding duration
were not as close to the I-IP2010 goals. Nationally, an estimated 39% of women
breastfed their infants until 6 months of age with 20% continuing until their
infants were 1 year old, falling short of the target rates of 50% and 25%,
respectively (CDC, 2005; Li, Darling, Maurice, Barker, & Grummer-Shawn,
2005). The National Immunization Survey (NIS) also collected data on rates of
exclusive breastfeeding at several time points. Nationwide an estimated 59% of

infants were exclusively breastfed at seven days postpartum with rates declining

 

to 39% at 3 months and 14% at 6 months (Breastfeeding: Data and Statistics:
Breastfeeding Practices — Results from the National Immunization Survey,
2005). Once again these rates were not as high as desired, the target percentages
being 60% at 3 months and 25% at six months.

The current NIS data for individual states found that 21 states met the
objective to have 75% of women breastfeeding in the early postpartum period.
The rates for continued breastfeeding were more discouraging: only ﬁve states
achieved the objective to have 50% of mothers breastfeeding their children at 6
months of age; while 1 1 states had 25% of mothers breastfeeding their children at
12 months of age. California, Hawaii, Oregon, Vermont and Washington were
the only ﬁve states that reached all three of the Healthy People 2010 breasﬂ'eeding
objectives for an_v breastfeeding; however, only Oregon achieved the exclusive
breastfeeding objective of 25% or greater through 6 months of age.
(Breastfeeding: Data and Statistics: Breastfeeding Practices — Results ﬁom the
National Immunization Survey, 2005).

Most relevant to this research study are the breastfeeding rates for
Michigan. Michigan rates for m breastfeeding did not reach national averages,
nor did the state achieve any of the HP 2010 objectives. Breastfeeding was
initiated with only 67% of infants in Michigan, with 33% of infants being
breastfed at 6 months, decreasing to 16% at 12 months of age. Rates for
exclusive breastfeeding in Michigan were also below national averages. An
estimated 32% of infants were exclusively breastfed at three months of age with a

decline to 8% at six months of age, compared to the target rates of 60% and 25%,

‘4 ﬁ

respectively (Breastfeeding: Data and Statistics: Breastfeeding Practices —

Results from the National Immunization Survey, 2005).

2.5 Prevalence of breastfeeding characterized by employment status

The National Immunization Survey does not measure rates of
breastfeeding based on employment status. One national survey, known as the
“Ross Mothers Survey” does categorize breastfeeding rates by women who are
either employed part-time, full-time, or not employed. This survey exists as an
on-going mail survey periodically sent to a nationally representative sample of
new mothers. The survey asks mothers to recall the types of milk their babies
received in the hospital, at one week of age, in the past 30 days, and most often in
the last week. By using a multiple choice questionnaire, mothers select the kinds
of milk fed to their infants from a listing that includes breast milk, commercially
available infant formulas, and cow’s milk. The data collected in 2002 estimated
that among mothers who work hill-time, close to 70% begin breastfeeding their
child in the postpartum period, yet only 27% percent are still breastfeeding six
months later with a decline to 13% breastfeeding at one year. In comparison,
when mothers are employed part-time or not employed, 37% and 35% of mothers
are breastfeeding at six months, with 21% and 24% continuing for one year,
respectively ("Mothers Survey Ross Products Division Abbott Laboratories ",
2003).

Although rates of any breastfeeding have been measured, at this time there

is no surveillance system in the United States that accurately describes exclusive

‘

15

 

breastfeeding rates for women in different states of employment. It is true that the
Ross Mothers Survey categorizes the women surveyed by employment and asks
questions related to ‘exclusive’ breastfeeding. The Ross Mothers Survey
estimates that in-hospital rates of exclusive breastfeeding are similar among
women working full-time (45%), part-time (51%), or not employed (45%). Rates
of exclusive breastfeeding at 6 months were 12%, 18%, and 20% for mothers
employed full-time, part-time or not employed, respectively. The greatest
discrepancies occurred when looking at rates of exclusively breastfeeding at 12
months, 15% of women not employed were breastfeeding exclusively, compared
to 1 1% of women working part-time and just 6% of women working full—time. A
problem arises, however, as to the deﬁnition of exclusive breastfeeding used in
the survey. The methods used in the Ross Mothers Survey are inadequate for
collecting rates of exclusive breastfeeding because the survey only collects
information on the introduction of infant formula use, not on the introduction of
complementary foods (Ryan, Wenjun, & Acosta, 2002). The deﬁnition of
exclusive breastfeeding in this survey, therefore, is not consistent with the
standard deﬁnition of exclusive breastfeeding referring to infants being fed breast
milk M (Breastfeeding: Data and Statistics: Breaszﬂaeding Practices —— Results
ﬁom the National Immunization Survey, 2005; M. Labbok & Krasovec, 1990).
The results of this survey can be used to suggest that women working full-time
outside of the home supplement with infant formula sooner when compared to
women working part-time or those not employed, however, the survey should not

be used as an adequate indicator of exclusive breastfeeding rates.

.. L

Clearly national surveys are lacking in data related to breastfeeding rates
among women in various states of employment. Numerous descriptive studies,
however, have examined the association between maternal employment and
breastfeeding rates. A majority of these ﬁnd that employment has a negative
impact on rates of breastfeeding. These studies suggest that factors such as a
general lack of support within the workplace (Reifsnider & Taylor-Meyers, 1885,
1985; Witters-Green, 2003), full-time employment vs. part-time employment
(Auerbach & Guss, 1984; Duckett, 1992; Fein & Roe, 1998; Hills-Bonczyk,
Avery, Savik, Potter, & Duckett, 1993; Kurinij, Shiono, Ezrine, & Rhoads, 1989;
Meek, 2001), and/or an insufﬁcient length of maternity leave (Auerbach & Guss,
1984; Hills-Bonczyk, Avery, Savik, Potter, & Duckett, 1993; Kearney &
Cronenwett, 1991; Roe, Whittington, Fein, & Teisl, 1999), negatively affect
length of breastfeeding among women employed outside of the home.

Taveras et al. (2003) conducted a prospective cohort study of 1007
medically and socially low-risk mother - infant pairs who were in a health
maintenance organization (HMO). The deﬁnition of low-risk in this study
excluded those infants who weighed < 2500 or > 4600 grams at birth, who had
stayed in the intensive care nursery, and who had a medical problem that needed
follow-up by a pediatrician or nurse practitioner. Also, excluded were mothers
who were 14 years old or younger, who had a positive toxicology screen for
drugs, or who spoke a language other than English or Spanish. Finally, mothers
and newborns whose anticipated length of hospital stay was > 48 hours were

excluded. The purpose of this study was to describe reasons for discontinuation

. s

of breastfeeding during the ﬁrst 12 postpartum weeks and to evaluate associations
between breastfeeding discontinuation and modiﬁable factors, such as workplace
issues. Fifteen-minute telephone interviews were conducted at two and 12 weeks
postpartum to assess breastfeeding outcomes. Of enrolled women who had
initiated breastfeeding, 476 (47%) had returned to work or school by the time of
the 12-week interview. Returning to work was found to be one of the strongest
predictors of breastfeeding discontinuation. Among women who returned to work
or school, 236 (50%) reported problems with trying to continue breastfeeding, and
the women reporting problems were signiﬁcantly more likely to discontinue by 12
weeks. The proportion of mothers who cited return to work as the main reason
for discontinuing breastfeeding increased from 14% at two to three weeks to 58%
at 10 to 12 weeks postpartum. The two major reasons mentioned for
discontinuation among women returning to work were restricted schedules and
breaks (50%) and insufﬁcient privacy (20%).

Similar to other studies (Chatteiji & Frick, 2003; Fein & Roe, 1998;
Kimbro, 2006; Roe, Whittington, F ein, & Teisl, 1999; Visness & Kennedy, 1997),
the results reported by Tavera support the assertion that women discontinue
breastfeeding as they return to work. In contrast to prior studies, however, this
qualitative study gives insight into speciﬁc components of the workplace that are
perceived by women as restricting their ability to express breast milk in the
workplace. Unfortunately, results from Taveras’ study have limited application
because the effects of work and school were combined, and there was no

delineation between part and full-time employment.

 

Results similar to Taveras were also demonstrated in a study by Roe et al.
(Roe, Whittington, Fein, & Teisl, 1999). The purpose of this study was to assess
the competition between employment and breastfeeding. Participants of the study
were predominately white (95%), married (88%) women between the ages of 16
and 43, who worked before childbirth and who had explicit plans to return to
work within the 12 months after the birth of their child (N=712). Of the 275
women who returned to work full-time (> 34 hours/week), 150 (54.5%) stopped
breastfeeding by the time their child was three months old, compared to the 173
(35%) women who did not work outside the home. In contrast to the study by
Taveras, the Roe study also collected data on the number of daily breast feedings.
It was discovered that the number of daily breast feedings declined signiﬁcantly
as the number of daily work hours increased. Roe’s study also examined the
effect of duration of work leave on duration of breastfeeding. It was found that
the greatest decrease in breastfeeding occurred in the ﬁrst 10 weeks alter birth.
The largest decline was between 6 and 10 weeks, with signiﬁcant decreases
continuing through 12 weeks postpartum. Since the average work leave for the
participants in the study was 8 to 10 weeks, a positive association was found
between length of work leave and length of breastfeeding duration. Based on
these ﬁndings, Roe suggests that a l2-week maternity leave may generally be
optimal.

Chatterji and Frick used data collected by the National Longitudinal Study
of Youth (N LSY79) and the Children of the National Longitudinal Study of

Youth (CoNLSY) to assess the effect of the timing and intensity (number of hours

 

worked per week) of returning to work after childbirth on the probability of
initiating breastfeeding and the number of weeks of breastfeeding (Chatterji &
Frick, 2003). The NLSY79 is conducted annually with a nationally
representative sample of 12,686 civilian youth who were between the ages of 14-
21 in 1979. Over-sampling occurred for select population subgroups (Aﬁican
American, indigent whites, and Armed Forces personnel) (NLSY 79 Round 20
Questionnaire). The sample represented in the CoNLSY, were children of
mothers who have been interviewed and/or assessed in a separate linked survey.
For the purposes of the study by Chatterji et al., analysis was limited to children
of those mothers who were employed part-time or full-time at some point during
the year that preceded the child’s birth (n=5,804). Unlike other studies that found
similar rates of breastfeeding initiation among women regardless of work status
("Breastfeeding Trends Through 2002", 2003; Kimbro, 2006), the results of this
study suggested that returning to work within the ﬁrst three months of an infant’s
life was associated with a reduction in the initiating of breastfwding by 16-18%
among those women who intended to breastfeed, and a reduction in the duration
of breastfeeding by four to six weeks among mothers who initiated breastfeeding.
The duration of breastfeeding was reduced an additional one to four weeks among
mothers who return to full-time employment within three months after delivery
(Chatterji et al., 2003). The results of these studies by Roe et al. and Chatterji et
al. have important implications because women in the United States return to
work aﬁer giving birth sooner than in other industrialized countries. It is

estimated that in the United States, one-third of mothers return to work within

20

 

three months aﬁer giving birth, while two-thirds return within six months
(Workplace Breastfeeding Support [issue paper], 2002) Therefore, a large
proportion of infants are affected by this early return to maternal employment.
The 12 weelm of maternity leave suggested by the Roe study may be
possible for women who qualify for the Family and Medical Leave Act of 1993
(FMLA). The FMLA, in part, “allows ‘eligible’ employees of a covered
employer to take job-protected, unpaid leave, or to substitute appropriate paid
leave if the employee has earned or accrued it, for up to a total of 12 work weeks
in any 12 months because of the birth of a child and to care for the newborn
child.” ("The Family and Medical Leave Act of 1993 "). Eligible employees are
deﬁned as those workers who have been employed for a total of at least 12
months by their employer, who worked at least 1,250 hours in the previous 12
months, and who are employed by a private sector company employing at least 50
workers at, or within, a 75 mile radius ("The Family and Medical Leave Act of
1993"). Unfortunately, this leaves a large portion (40%) of women without
protected leave from work. In addition, it has been found that even women who
qualify for FMLA are unwilling to make use of the beneﬁt due to loss of wages or

other potential negative repercussions from employers.

2.6 Impact of employment on breastfeeding rates in Michigan
The Michigan Department of Community Health collects data for the
Michigan Pregnancy Risk Assessment Monitoring System (MI PRAMS)

annually. Similar to ﬁndings by Taveras, data collected for the 2002 MI PRAMS

21

 

revealed that 25% of women who never initiated breastfeeding cited return to
work/school as a barrier. Of the women who initiated breastfeeding, 25% cited
return to work/school as a barrier to continued breastfeeding (PRAMS report,
2005). Although the MI PRAMS data suggested that employment is a banier to

breastfeeding, rates linked to employment status are not available for Michigan.

2.7 Federal legislation and breastfeeding

Despite the recognition that employment serves as a barrier to
breastfeeding, enactment of federal legislation to protect the rights of women
wishing to breastfeed in the workplace has thus far been protracted. In 1999,
Congresswoman Carolyn B. Maloney (D-NY -14) introduced four individual bills
on breastfeeding, two related to breastfeeding in the workplace. At present, these
two bills remain in committee and have yet to be ratiﬁed. The ﬁrst of these two
bill seeks to protect breastfeeding under the civil rights law, requiring that women
cannot be ﬁred or discriminated against in the workplace for expressing
(pumping) breast milk, or breastfeeding during their own lunch or break time
(HR. 1478, HR. 3861). The second of these bills encourages employers to set up
safe, private, and sanitary enviromnents for women to express (pump) breast milk
by providing a tax credit to employers who set up a lactation location, purchase or
rent lactation-related equipment, hire a lactation consultant, or otherwise promote
a lactation-friendly environment (HR. 1163). Future enactment of federal

legislation related to breastfeeding in the workplace could make a substantial

22

contribution to increasing the prevalence of breastfeeding among working

mothers.

2.8 State legislation and breastfeeding

State legislation that protects a woman’s right to breastfeed in the
workplace is also currently weak. Presently, 34 states have enacted legislation
related to breastfeeding. Of these, only 11 states (CA, CT, GA, HI, IL, MN, OR,
RI, TN, TX, WA) and Puerto Rico have bills that contain provisions speciﬁc to
breastfeeding within the workplace (Douglas Reid Weimer, 2005). Only three
states (CA, CT, and IL) and Puerto Rico require employers to provide reasonable
time and private accommodations for employees needing to express breast milk
(Martucci & Coverdalel, 2001; DR. Weimer, 2005). Little incentive exists,
however, for employers to comply with state legislation because only California
and Puerto Rico have penalties in place for employers who fail to comply.

Most relevant is a review of legislation related to breastfeeding in the state
of Michigan. No Michigan legislation mentions workplace accommodations for
breastfeeding (Martucci & Coverdalel, 2001; DR. Weimer, 2005). Currently
Michigan has two laws related to breastfeeding. The ﬁrst excludes women who
are breastfeeding a baby from public nudity laws (Mich Comp. Laws Ann §§
41.181, 67.1(aa), 11 7. 4i, and 11 7.5h West, 2002). The other law concerns child
custody and includes breastfeeding as a factor when determining parental
visitation schedules (Mich. Comp. Laws Ann. § 727.27a West, 2002. In the

absence of federal and states laws, the provision of breastfeeding

23 m

accommodations is left to individual companies. So the real issue that needs to be
addressed is how best can companies be encouraged and supported to set up
policies and practices within their workplace that will positively impact
breastfeeding rates among those employees who want to combine breastfeeding

and employment.

2.9 “Family-ﬁ'iendly beneﬁts”

Because not all mothers qualify for, or chose not to make use of, FMLA
there are a number of accommodations a workplace can provide to support
breastfeeding. Literature examining company policies commonly referred to as
“family-ﬁiendly beneﬁts” suggests that employers who offer family-ﬁiendly
beneﬁts increase organizational commi’nnent and job satisfaction, while
decreasing work-family conﬂict and turnover intentions (T. D. Allen, 2001).
Speciﬁcally, one study found that companies offering family-friendly beneﬁts
such as ﬂexible sick leave and child care assistance experienced measurable
reductions in turnover (Baughman, DiNardi, & Holtz-Eakin, 2003). It is also
recognized that workers decide to apply for jobs not only because of cash wages,
but based on how well the mix of wages and beneﬁts meets their needs
(Baughman, DiNardi, & Holtz-Eakin, 2003).

These family-friendly beneﬁts are not offered equally across different
business sectors however. A recent survey conducted by the Society for Human
Resource Management (SRHM) found that 40% of companies in the Health

sector offered a lactation program or designated room compared to 25% of

24

K

 

companies within the service sector (Burke, 2005). The SRHM survey also
suggests that larger companies are more likely to offer family-friendly beneﬁts
than smaller companies. For example, only 8% of small companies (1-99
employees) offer a lactation program or designated area, compared to 20% of
medium companies (100—499 employees) and 28% of large companies (500 or
more employees) (Burke, 2005).

Clearly, women who are provided breastfeeding accommodation
personally beneﬁt, however, it is suggested that companies as a whole may also
beneﬁt (Baughman, DiNardi, & Holtz-Eakin, 2003; Benson, 2005; Clifton &
Shepard, 2004). Beneﬁts for companies include a higher rate of return to work by
women after maternity leave, staff loyalty, increased staff continuity, an added
recruitment incentive, and a reduced training budget due to a reduction in
turnover (R. Cohen, M. B. Mrtek, & Robert G. Mrtek, 1995; Gardner, 2002).
Additionally, since health insurance premiums are based on medical costs,
companies that provide health insurance to their employees may have lower
premiums due to the reduced medical costs associated with infants who are
breastfed (Gardner, 2002). It is also suggested that women who are provided with
workplace accommodations for breastfeeding have greater job satisfaction and are
more productive, which translates into a direct beneﬁt for the companies for
which they work (R. Cohen, M. B. Mrtek, & R. G. Mrtek, 1995; Gardner, 2002).

There is evidence to suggest that companies do not fully realize the
beneﬁts they can receive by providing accommodations, such as breastfeeding

support, to their employees (Dunn, Zavela, Cline, & Cost, 2004; Witters—Green,

25 -A

2003). Dunn et al. examined breastfeeding support policies and practices within a
sample of Colorado businesses. While 70.5% (n=157) of the companies surveyed
agreed that their organization would support their employees’ breastfeeding
needs, only a minority of respondents recognized that allowing women to
breastfeed in the workplace would positively inﬂuence productivity (44.1%),
recruitment (23.5%), decrease employee absenteeism (23.5%), improve morale of
other employees (22.2%), or decrease employee turnover rate (17.5%).

Another issue that arose from the Dunn study was that a majority of
respondents indicated that breastfeeding support was considered a non-issue due
to too few employees, lack of female employees of childbearing age, or lack of
request for breastfeeding support services from employees (Dunn, Zavela, Cline,
& Cost, 2004). Some of the perceived lack of need can be attributed to the fact
that many women discontinue breastfeeding shortly before or after returning to
work (Kimbro, 2006; Roe, erittington, F ein, & Teisl, 1999; Taveras et al., 2003).
Employers viewing breastfeeding as a non-issue can also be partially explained by
the workplace behavior referred to as “employee silence”. Prior research in this
area suggested that employees often felt uncomfortable raising issues and
concerns to their bosses, yet it remained unknown as to why these employees
remained silent. One study has been conducted to help explain why employees
remained silent. Milliken et al. (2003) conducted in-depth interviews (n=40) with
full-time employees concurrently enrolled part-time in MBA management classes.
The purpose of the study was to better understand when and how participants

decided to be silent about an issue of concern, and the factors they considered

.. k

when making this decision. Of the study participants, 85% cited that at least on
one occasion they were unable to raise an issue or concern to their bosses even
when they felt the issue was important. Fear of being viewed, or labeled,
negatively was the most frequently cited reason for why employees remained
silent about workplace issues that concerned them (Milliken, Morrison, &
Hewlin, 2003).

Employee silence can contribute to the lack of family-friendly
accormnodations in the workplace. If issues of concern are not brought to light,
then companies cannot take action to rectify them. Employee silence, however, is
only one barrier to breastfeeding in the workplace. Research has demonstrated
that even when companies offer work-family programs or beneﬁts, many
employees choose not to participate (Thompson et al. 1999). One survey
conducted, found that just 2% of employees in the Fortune 500 companies utilized
work-family beneﬁts and programs (Galinsky, Bond, & Friedman, 1993). It has
been suggested that employees perceive that if they utilize these beneﬁts and
programs, they will experience negative career consequences (Perlow, 1995).
Perlow’s study suggests that family-ﬁiendly beneﬁts such as telecommuting may
lead to negative career consequences. The results of this study highlight the
concept of “face-time”, i.e. that there is a direct relationship between one’s
presence at work and how one’s contribution to the organization is viewed.
Recent work in the ﬁeld of work-family supports the idea that employees’
perception of the work climate affects utilization of work-family programs

(Thompson et al. 1999). Work climates that are perceived by employees to be

27

 

h.

supportive are associated with higher rates of utilization. In particular,
managerial support was signiﬁcantly related to higher utilization and may be the
most important component in employees’ decisions to use family-ﬁiendly
beneﬁts and programs (Thompson et al. 1999; Frye & Breaugh 2004).

Three small studies have explored employer (manager) attitudes towards
breastfeeding by female employees, including two small mail surveys and focus
group work (Libbus & Bullock, 2002, Bridges et al., 1997, Brown et al., 2001).
The participants in these studies were men and women, but primarily white,
married, middle-aged, and highly educated. Results were similar across all three
studies. Respondents were generally supportive of breastfeeding by employees,
but they did not believe the workplace needed to be changed to provide more
support and lacked knowledge of the beneﬁts to employers of breastfeeding by
employees, such as reduced maternal absenteeism or reduced health care costs.
The focus group participants also expressed several barriers to accommodating
breastfeeding women at the worksite (e. g. decreased productivity and jealousy of
coworkers for extra break times) (Brown et al., 2001). These data, similar to Roe
et al., suggested that although employers are supportive of the practice of
breastfeeding, they lack knowledge on the beneﬁts of breastfeeding for their
workplace, and did not feel they needed to, or should, provide breastfeeding
support within their worksites. Taken together, these studies support the idea that
it is important to access how woman perceive their manager feels about providing

breastfeeding accommodations in the workplace. If women do not feel that their

28

mangers are supportive of breastfeeding, it may modify their decision of whether
to continue breastfeeding once returning to work.

There is also evidence to suggest that coworkers may inﬂuence whether an
employee decides to utilize family-ﬂiendly beneﬁts. A study by Seijts was
conducted examining employees’ perceptions of fairness based on whether
breastfeeding accommodations were provided, or not provided, within a
workplace (Gerard H. Seijts, 2004). Participants were a random sample of alurrmi
from a large Canadian business school, employed and living in Canada (n=l22).
Seijts assessed non-managerial employees’ reactions to the denial or granting of a
speciﬁc workplace accommodation — meeting the needs of a breastfeeding
employee. Participants received one of three scenarios by mail, which described
the difficulties a ﬁctional employee encountered when she returned to her job
after three months of maternity leave, and wanted to breastfeed. The three
scenarios described an organization that (a) did not accommodate a breastfeeding
employee; (b) accommodated a breastfeeding employee at no personal cost to her
coworkers; and (c) accommodated a breastfeeding employee, but at a personal
cost to her coworkers (Gerard H. Seijts, 2004). Results of this study found that, in
general, provision of breastfeeding accommodations made a difference in
perceptions of fairness only to those participants who had children. Participants
with children had decreased perceptions of fairness when the accommodation was
not offered. Participants without children had similar perceptions of fairness
whether or not a breastfeeding accommodation was provided. In contrast, when

the scenario included accommodations being oﬂ‘ered, but at a personal cost to the

29

 

coworker (e. g. covering work duties for a breastfeeding coworker), non-parents
had the lowest perceptions fairness. Additionally, perceived fairness was
negatively impacted when the participants held a belief that breastfeeding is a
private issue that should not be brought into the workplace (Gerard H. Seijts,
2004). This study supports the idea that it is important to access women’s
perceptions of how their coworkers will feel about breastfeeding accommodations
being provided in the workplace. Similar to the position a woman ﬁnds herself in
with her manager, if women perceive that their coworkers are not supportive of
breastfeeding accommodations in the workplace, it may impact their decision of
whether to continue breastfeeding after returning to work.

In summation, the beneﬁts of breastfeeding, particularly exclusive
breastfeeding for the ﬁrst 6 months, can be wide-ranging. Breastfeeding can be
an inexpensive method of primary prevention for many acute and chronic diseases
and can improve the health and well being of both mothers and infants.
Breastfeeding can also decrease the cost and demands on parents, society, and
health care systems. Unfortunately, due to the large number of women that return
to work shortly alter giving birth, breastfeeding rates remain below public health
goals. It is recognized that many women discontinue breastfeeding even if they
have positive personal attitudes (Berger, Hill, & Waldfogel, 2005; Galtry, 1997;
McKinley & Hyde, 2004). Women may discontinue breastfeeding because
overall, most work environments are not perceived as supportive of breastfeeding.
There are many reasons why the worksite may not be perceived as supported.

Many worksites lack the private space needed for expressing breast milk. Women

30

are given breaks that are not frequent or long enough to maintain their milk
supply. Although, a study by Slusser et al. found that most women can express
breast milk for infants within two 30 minutes sessions, many workplaces lack the
ﬂexibility and appropriate space needed for employees to arrange these time
blocks (Slusser, Lange, Dickson, Hawkes, & Cohen, 2004). Managers may not be
supportive of, or see a need for, breastfeeding accommodations in the workplace
and if family-friendly beneﬁts are offered, employees may feel that they are
disadvantaged if they make use of these policies (Galinsky, Bond, & Friedman,
1993). Furthermore, breastfeeding accommodations may not be viewed
positively by coworkers if they feel these accommodations do not beneﬁt all
workers equally (Gerard H. Seijts, 2004). This may cause conﬂict between
coworkers, and between employees and managers or companies (Gerard H. Seijts,
2004). Commonly companies lack knowledge that providing breastfeeding
accommodations is beneﬁcial for their company (Dunn, Zavela, Cline, & Cost,
2004). Worksite breastfeeding support programs in general have been shown to
increase breastfeeding however, the differences in components among the
programs and lack of standardized data collection methods makes it hard to assess
exactly what makes these programs successful (Cohen & Mrtek, 1994; Ortiz,
McGilligan, & Kelly, 2004). It is therefore necessary to be able to assess, in a
standardized way, which components in the workplace are most likely to

encourage breastfeeding in order to develop effective worksite lactation support

programs.

31

CHAPTER 3

METHODS

This project consists of two phases. The ﬁrst phase was development of
an instrument. This phase included item development, cognitive interviews, and
expert review to test and reﬁne the instrument before piloting. The second phase
was to pilot the instrument with a sample of women representative of the target
population. This phase included recruitment of women, mailing of instruments to
be self-administered, collection of data, and ﬁt analysis to determine items
exhibiting misﬁt, and determining based on substantive reasoning if these items
should be removed or reworded for future tests of the instrument. Michigan State
University’s Internal Review Board for research involving the use of human

subjects approved both phases of the study.

3.1 Phase 1 - Instrument development

After a review of literature it was determined that no suitable instrument
existed to measure woman’s perception of the workplace climate for
breastfeeding support, therefore, a new instrument was developed. Instrument
development followed standard procedures of test construction designed to create
a valid form of measure (M. J. Allen & Yen, 2002; DeVellis, 1991; Don A.
Dillrnan, 2000; Fowler, 2002; Turocy, 2002; Wolfe & Smith, 2007a, 2007b) and
employed design elements to improve readability (D.A. Dillrnan, 2000). Several

instrument development activities where undertaken to develop arguments for

32

validity of the instrument (Wolfe & Smith, 2007b). A form to collect participant

and worksite characteristics was also developed.

3.1.1 Instrument purpose

The ﬁrst steps in developing a new instrument are to deﬁne the purposes
for which the instrument was designed, to describe the context (environment) in
which the ﬁnal instrument could be used, and to explain which decisions could be
made based on administration of the instrument.

The instrument was designed to measure the work climate for
breastfeeding support from the perspective of the target population: pregnant and
new mother employees working full-time in non-managerial positions. This
instrument was developed for use in speciﬁc job sectors that were identiﬁed based
on literature that suggests these sectors provide family-friendly beneﬁts (Burke,
2005). If a company provides family-friendly beneﬁts it is more likely to also
have accommodations for breastfeeding. Administration of the instrument would
allow companies to determine which speciﬁc areas of the workplace are perceived
by female employees as positively or negatively inﬂuencing their attempt to
breastfeed. Companies could then decide which educational components would

be most beneﬁcial to include in their workplace lactation support program.

3.1.2 Theoretical framework of the instrument

The social-ecological perspective was used to inform the model of the

instrument. The social-ecological perspective was relevant to this study because

33

ecological models of health behavior suggest that health behaviors, such as
breastfeeding, are inﬂuenced by intrapersonal and environmental (external to the
individual) factors (Glanz, 2002). These environmental factors include actual or
perceived social, cultural and physical aspects of a speciﬁc environment (e.g.
workplaces). This is not the ﬁrst the time that an ecological framework has been
used for categorizing work climate components important for the continuation of
breastfeeding. A review paper published this year supports the concept of a
workplace functioning as its own ecosystem, and being comprised of the woman,
her social support (from coworkers and managers), and the workplace
environment (Johnston & Esposito, 2007). As depicted below, the outer layer of
the work climate encompasses the company’s policies and practices, both formal
and informal. These policies and practices do not only include written guidelines
but also addresses issues such as time off after delivery, ﬂexibility in daily
workﬂow, job autonomy, number of hours worked per week, availability of
pumping equipment, and the physical environment of where the woman is
expressing breastrnilk. The middle layer includes two distinct groups of
employees, those employed in managerial and those employed in non-managerial
positions. For this project, the focus of the middle layer is on the employee
serving as the frontline manager to the study participant as well as coworkers of
the study participant. The innermost layer of the work climate is the new mother

employee (See Figure l).

34

 

Company policies/work culture

 

Manager/coworkers support

 

New mother employee

 

Figure 1: Ecological model as related to the work climate

35

The latent trait, or unobservable construct, that the developed instrument
was designed to measure can therefore be broadly described as ‘perception of the
work climate for breastfeeding support’. The construct was subdivided into
individual components in order to increase measurability. The construct model
was organized in a Table of Speciﬁcations before undergoing expert review

(Table 2).

3.1.3 Test speciﬁcations
The Table of Speciﬁcations included four main subscales with two
components under each subscale.

Table 2. Table of Speciﬁcations

 

Construct — Perception of the work climate for breastfeeding support

 

Subscale 1 - Perception of company support
a. Perception of company support of new mother employees
b. Perception of company support for female employees that
breastfeed or express breast milk during work hours
Subscale 2 — Perception of support by managers
a. Perception that her manager would be supportive of
breastfeeding in general
b. Perception that her manager would help a woman manage her
job so she could combine breastfeeding and employment
Subscale 3 — Perception of support by coworkers
a. Perception that coworkers would be supportive of breastfeeding
in general
b. Perception that coworkers would be supportive of a woman
breastfeeding or expressing breast milk at work
Subscale 4 — Perception of time availability
a. Perception of availability of break time
b. Perception of ability to use given break time

 

36

3.1.4 Expert review of the test speciﬁcations

In order to help establish an argument for validity of the instrument,
experts (n=11) were asked to review the Table of Speciﬁcations (Table 2). This
was done to ensure that the survey items ﬁt the construct model as outlined in the
Table of Speciﬁcations. For this review, the experts were given an outline of the
Table of Speciﬁcations with sample items included for each subscale of the
construct. The sample items were supplied in order to help the expert reviewers
gain a better understanding of what content would be covered by each subscale.
These experts included practitioners (with credentials and experience in the
lactation ﬁeld) and researchers (with experience in survey development or
lactation research). Practitioners were recruited through a Listserv of Lactation
Consultants based in Michigan and included registered nurses with IBCLC
(International Board Certiﬁed Lactation Consultants) certiﬁcation, registered
dietitians with IBCLC certiﬁcation, practitioners with IBCLC certiﬁcation, the
Michigan Coordinator of La Leche League International leaders, and the program
leader for Michigan State University Extension’s Breastfeeding Initiative.
Researchers included Michigan State University, Georgia Tech and Virginia Tech
faculty with expertise either in the ﬁeld of nutrition, in the ﬁeld of organizational
development, or in the ﬁeld of instrument development.

As a result of the expert review, one additional subscale, perception of the
physical environment was added. Categories within subscales were also modiﬁed
based on comments received from the review. Based on the sample items

provided for each component, it was suggested by the expert reviewers to add

37

content area reﬂective of negative comments or resentment by coworkers,
provision of breast pumps by companies, and possible negative repercussions by
the company. Additionally, a subscale with three items was added to assess job
autonomy. Each subscale and its individual components were also given a
simpliﬁed title to reﬂect the aspects of the work environment each included, these
subscale were organized in the Test blueprint. The Test blueprint (Table 3)
includes six subscales with two components under each subscale, except for job
autonomy. After item development (see procedure below) the number of items
and the numbers of each individual statement being used to measure each
component was listed.

Table 3. Test blueprint

 

Construct — Perception of the work climate for breastfeeding support
Number of Items Item Numbers

 

Subscale 1 — Company .................. 13
a. Formal policies ................................................ l — 5
b. Work Culture ................................................... 6 — l3
Subscale 2 - Manager ................... 13
a. Emotional support .............................. 24,26,27,29,30,32,36
b. Instrumental support ................................. 25,28,31,33 — 35
Subscale 3 - Coworkers .................. 9
a. Emotional support ................................. 37 — 40, 42,45
b. Instrumental support ................................... 41,43 — 44
Subscale 4 - Workﬂow .................... 7
a. Availability of time .................................... 14 -15, 18
b. Ability to use given time ........................ l6 — 17, 19, 20
Subscale 5 — Physical environment ....... 5
a. Availability/Accessibility ................................. 47 - 48
b. Adequacy ................................................... 49 — 51
Subscale 6 - Job autonomy ................ 3 ............................ 21-23

 

38

3.1.5 Description of subscales

To further understand the construct, it is necessary to describe each
subscale and its components:
0 Company: measures perception of support provided at the organization level.
(1) Formal company policies. For example, written policies directly
related to breastfeeding employees.
(2) Informal practices, deﬁned here as “work culture”. The work culture
is an overall agreement on what behaviors are deemed as acceptable in the
workplace. For example, perceptions about the acceptability of discussing a
personal issue at work, such as breastfeeding.
0 Manager: measures the woman’s perceived support of her immediate manager
and the manager’s attitudes towards breastfeeding employees.
(1) Emotional support refers to managerial behaviors that provide
empathy and demonstrate an understanding about the importance of
breastfeeding. For example, a manager being open to discussing a personal issue
such as breastfeeding demonstrates emotional support.
(2) Instrumental support refers to managerial behaviors that directly help
the women during times of breastfeeding need. One example would be a manager
willing to cover a woman’s job duties in order for her to breastfeed or express
breast milk.
0 Coworkers: measures perceived support provided by coworkers and coworker’s

attitudes towards breastfeeding employees.

39

(1) Emotional support refers to coworker behaviors that provide empathy and
demonstrate an understanding about the importance of breastfeeding. For
example, emotional support could be provided if a coworker made positive
comments about breastfeeding.

(2) Instrumental support refers to coworker behaviors that directly help the
women during times of breastfeeding need. For example, a coworker that is
willing to switch break times with a breastfeeding mother demonstrates
instrumental support.

0 Workﬂow: measures perceived availability of break time given during work

hours as well as perceived ability to make use of the given time.

(1) Availability of break time refers to the ﬁ'equency and length of breaks, as
well as, the availability of additional breaks speciﬁcally for breastfeeding
or pumping breast milk.

(2) The ability to make use of given time refers to how a woman’s overall
workday is scheduled. For example, although a woman is offered breaks,
her workday is too hectic for her to use the break time.

c The Physical Environment of the Workplace: measures perceived

availability/accessibility and adequacy of the physical space where breastfeeding

or expressing breast rrrilk occurs.

(1) Perception of the availability/accessibility for example, relates to how
close the physical space is to the woman’s work area.

(2) Perception of the adequacy of the space for example, refers to how

comfortable the woman is using the physical space.

40

3.1.6 Item development

After the Test blueprint was established, individual items for the
instrument were developed. Survey items were developed for each subscale to be
used with a 4-point Likert scale, with no option for ‘neither agree nor disagree’, in
order to increase the response rate for each individual item. The instrument that
was created was designed to be a self-administered questionnaire. Self-
admirristered questionnaires are intended to gather speciﬁc information via a self-
reporting system (Turocy, 2002). The goal of writing items for a self-
administrated questionnaire is to develop a item that every potential respondent
will interpret in the same way, be able to respond to accurately, and be willing to
answer (Don A. Dillrnan, 2000). The ﬁrst step in developing items consists of
writing a pool of items. Typically this pool of items contains up to four times the
number of items that will be included on the ﬁnal instrument (DeVellis, 1991 ).
An initial pool of items was constructed for each subscale in order to provide a
source from which the instrument could emerge.

After the initial pool of items was generated, each individual item was
evaluated to access goodness of ﬁt to the construct, clarity, and importance. The
evaluation process included eliminating or rewording items as necessary. Items
were ﬁrst evaluated for elimination. Items were eliminated from the pool due to
lack of clarity, questionable relevance, or undesirable similarity to other items

(DeVellis, 1991).

41

Items were then assessed to see if they needed to be reworded. Some
items were reworded because they were unnecessarily lengthy, but only if a
shortened version could retain the same meaning. Other items were reworded to
elirrrinate those that conveyed two or more ideas. Such items were eliminated
because items conveying more than one idea can lead a respondent to falsely
agree or disagree with a portion of the item (DeVellis, 1991; DA. Dillman, 2000).
Negatively worded items were also eliminated because they are more difﬁcult to
answer due to decreased clarity of meaning and increase respondent burden.
Additional items were reworded to decrease reading level.

The items were evaluated using the Gunning Fog index
(http://en.wikipedia.org/wiki/Gunning-Fog_lndex). The Gunning Fog index
measures the grade level, or assumed readability, of a piece of writing using two
factors: the average sentence length and the percentage of complex words. The
Gunning Fog Index is calculated by using the following formula:

[(words/ sentence) + 100 * (complex words/words)] * 0.4
Complex words are deﬁned as a word of 3 syllables or more, but not a word that
is (l) a combination of short words (bookkeeper); (2) a verb form that becomes
polysyllabic by adding -ed, -es, or -ing, or (3) a proper name. It is suggested to
use one or more passages of at least 100 words each. The Gunning Fog index for
the newly developed instrument was calculated by using the ﬁrst two items of
each section on the instrument. For the purposes of calculating the Gunning Fog

Index, 149 words were used, 10 sentences were used, and there were a total of 8

42

complex words within the 10 sentences. Therefore, the Gunning Fox index was
calculated as 7.96.

The Gunning Fox index estimates that the instrument is just below an 8th
grade reading level. A reading level between the ﬁfth and seventh grade is
appropriate for use with the general population (DeVellis, 1991). However, an 8th
grade reading level for this target population should not be considered
unreasonable. The Gunning Fox index equation is driven both by sentence length
and percentage of complex words. Although, sentence length could be an issue,
the target population should not have difﬁculty understanding the complex words.
The eight complex words were maternity, information, manager (2 times),
coworkers (2 times), company, and designated. There were a munber of steps

also taken to increase readability (D.A. Dillman, 2000).

3.1.7 Recruitment for cognitive interviews

Recruitment was done locally in establishments and through organizations
likely to be utilized by employed women that had given birth in the last year.
Recruitment ﬂyers (Appendix A) were posted at the Special Supplemental
Nutrition Program for Women, Infants, and Children (WIC) office; the Women
Health Services ofﬁce; the Department of Human Services; public libraries;
Michigan State University’s Family Resource Center; local child care facilities;
the Expectant Parent Organization; an elementary school; grocery stores; and a
hair salon. The ﬂyer was also sent out through a Birth Center’s electronic

newsletter. Recruitment also occurred by word of mouth. All potential subjects

43

answered questions over the telephone to determine eligibility using the
Interviewee Screener (see Appendix A). Inclusion criteria included: being at least
18 years of age; being employed in a non-managerial position; giving birth within
the last year and returning to full-time work, deﬁned as working 32 or more hours
per week after the infant was born. In addition, each caller was asked if she had
breastfed her newborn. Different paths of inquiry were followed depending on
whether the woman (a) was currently breastfeeding; (b) had discontinued
breastfeeding; or (c) had not breastfed.

If the caller had breastfed the child, she was asked when she ﬁrst decided
to breastfeed (before birth, during pregnancy, or after delivery), the main reason
for her decision to breastfeed, and if she was currently breastfeeding her newborn.
If the woman was still breastfeeding, in addition to working full-time, she was
eligible for participation. If the woman was no longer breastfeeding, she was
asked how long, in weeks and/or months, she breastfed after delivery in order to
assess duration of breastfeeding. If the caller had discontinued breastfeeding, she
was asked if work was a factor in her decision to discontinue breastfeeding. Even
if employment was not the main reason for discontinuation, a woman with any
amount of experience combining employment with breastfeeding was eligible for
participation.

If the caller had not breastfed their child, they were asked, “At any point
before or during your pregnancy, or after giving birth, did you consider
breastfeeding?” If the caller had never considered breastfeeding, she was

ineligible for the study. If the caller had considered breastfeeding at any point in

44

time, she was asked the main reason why she decided not to breastfeed. If the
caller indicated that return to work was the main reason she did not breastfeed,
she was eligible for the study. If the caller did not cite return to work as the main
reason for her decision not to breastfeed, it was then assessed if returning to work
had any inﬂuence in her decision not to breastfeed. If the caller indicated that
return to work inﬂuenced her decision about breastfeeding, she was eligible for
the study. If employment was not a factor in her decision, she was not eligible.
This ineligibility was based on the reasoning that if woman had not at least
considered breastfeeding after returning to work, it would be difﬁcult for her to

answer the instrument items.

3.1.8 Cognitive interviews

An argument for validity was established by the individual items being
evaluated during in-depth, one-on-one interviews by a sample of women
representative of the target population (n=l4). The one-on-one interviews utilized
cognitive interview methods, such as the “think-aloud” method which focuses on
the mental processes subjects use when answering survey items and “probing”
techniques asked after the subject answered the survey items (Willis, 1994).
Cognitive interview methods were also used to help to decrease potential response
error (lobe & Mingay, 1989).

One-on-one interviews were conducted instead of focus groups, because
one-on-one interviews are more effective than focus groups for establishing

rapport between the subject and the investigator (Portney, 2000). Questions

45

related to breastfeeding are of a personal and potentially sensitive nature
justifying the need for one-on-one interviews. Another advantage of the interview
approach is the opportunity for in-depth analysis of subjects’ behaviors and
opinions. This occurs because the investigator can probe responses and directly
observe subjects’ reactions (Portney, 2000). The use of one-on-one interviews
allowed the investigator to obtain in—depth information about perceptions of
breastfeeding support, in order to develop questionnaire items relevant to this
population.

Cognitive interview techniques are used with a group representative of the
target population to ﬁnd out if the items are those that subjects can consistently
understand and answer (Fowler, 2002). These women spent extended time to help
the investigator understand how the items were interpreted (Fowler, 2002). In this
way, problems with comprehension or difﬁculty with response task can be dealt
with before the instrument is piloted (Don A. Dillman, 2000; Fowler, 2002).
Cognitive interview methods have been demonstrated as beneﬁcial for evaluating
the quality of questionnaire items and determining item meaning (Willis, 1994).
In this study, cognitive interview methods were combined with an up-front short,
open-ended interview question to look for any missing topics.

At the start of each interview, each woman was asked to sign an informed
consent form (see Appendix A) and a gift certiﬁcate ($20.00) conﬁrmation form
(see Appendix A). Each interview consisted of three parts: the participant and
worksite characteristics form, open-ended questions, and the perceptions

questionnaire. The three parts were administered in the same order during each

46

interview. Interviews took place at a local public library, in a private room at
Michigan State University, or in the woman’s home. Each interview lasted, on

average, one-hour (45-75 minutes).

3.1.9 Cognitive interview procedures

3.1.9a Participant and Worksite Characteristics Form (5-10 minutes)

A participant and worksite characteristics form (see Appendix C) was
developed for the project based on standard demographic forms used in national
surveys ("US Census Bureau", 2007), and followed standards of instrument
design to increase readability (D.A. Dilhnan, 2000). Subjects completed the
characteristics form at the beginning of the interview. Items targeted important
characteristics found in previous research to inﬂuence breastfeeding decisions and
ones related directly to the work environment. Variables collected on the
participant and worksite characteristics form included breastfeeding history, how
the milk supply was being maintained, commitment to breastfeeding, length of
employment with current employer, average hours worked per week, pay type,
union membership, gender of immediate manager, gender of coworkers, weeks of
leave after delivery, weeks of paid leave after delivery, job title, job category,
number of employees at same physical location, number of total employees at all
locations, racial/ethnic group, age, marital status, level of education, and yearly
household salary.

The subject was informed that the purpose of the interview was to make

the questions understandable to women similar to her. She was informed that if

47

she were unsure of what a question meant, or if she thought a question could be
answered in more than one way, she should let the investigator know. Revisions
to the participant and worksite characteristics form were made based on an
aggregate of comments made by all subjects.

3.1.9b Open-Ended Question (5-10 minutes)

An open-ended question (see Appendix A) was used to give the subject
the opportunity to provide information in her own words and from her own
perspective (Portney, 2000). The open-ended question was used before the
participants answered the survey items to elicit unbiased cements related to
components of the workplace perceived to help or hinder breastfeeding in the
workplace. The open-ended question was used to discover if any additional
subscales or survey items were needed. The question asked if the woman had
breastfed after returning to work full-time. If she answered “yes”, she was asked
to describe her experience with combining breastfeeding and work. If needed,
verbal probes were used to ascertain information about people in the workplace
and aspects of the physical work environment that made it easier or harder for her
to combine breastfeeding and employment. If the woman answered “no”, she was
asked to describe her workplace and ways she thought it would either support or
discourage a woman who wanted to continue breastfeeding. If needed, verbal
probes were used to collect information about people in the workplace and aspects
of the physical work environment that she thought would make it easier or harder

for women wanting to breastfeed in her workplace.

48

The investigator took hand-written notes during the interviews, including
making notes on any non-verbal reactions, and all interviews were audio-
recorded. The investigator used the tapes as needed to clarify portions of the
hand-written notes. Audiotapes were stored in a locked cabinet and were

destroyed at the end of the study.

3.1.9c Perception Questionnaire (30-40 minutes)

All subjects were asked to ﬁll out the perception questionnaire (see
Appendix A - Note: since changes were made to the questionnaire after each
interview, the version included is the ﬁnal version after all the interviews were
completed). Each statement was rated on a 4-point Likert Scale (strongly agree,
agree, disagree and strongly disagree). A series of statements expressing a
viewpoint was presented (e. g. I can talk about breastfeeding at work) and the
women selected an appropriate response that reﬂected their agreement or
disagreement with each item (Portney, 2000). The women were informed that the
4-point scale was being used to increase the response rate for each individual
statement and if at any time she felt the need for a “Don’t Know” or “Not
applicable” option, she should let the investigator know. It was reiterated that she
should stop whenever she had questions, comments, or concerns about any item
on the survey.

The questionnaire contained ﬁvepsections. The women answered

statements one section at a time. While the woman was ﬁlling out the survey, the

49

investigator used an identical survey to note if she hesitated, looked confused,
skipped any statements, or asked for a ‘don’t know’ or ‘not applicable’ option.

After completion of each section, the investigator asked the woman to
describe how she felt about the group of items. Questions such as, “Tell me what
you thought about that group of statements?” and “Did the statements cover issues
that you thought about or experienced when you were combining working and
breastfeeding?” were used to assess any missing subscales or items within
subscales.

The investigator then asked speciﬁc questions about items on the survey
using a technique called retrospective verbal probing (Willis, 1994). Some
examples of probes used were, “What does this statement mean to you?” “Can
you tell me in your own words what this statement means” and “Can you tell me
what you were thinking when you answered this statement?” In addition, if the
woman experienced difficulty with understanding any item she was asked to give
suggestions for alternative wordings for that item in order to clarify meaning and
increase understanding for women similar to her. The investigator and woman
talked about each problem item until an understanding of the concept of the
statement was reached.

At the end of the interview, each woman was asked if she had any
additional comments about the demographic form or the survey items. Each
woman was also asked if she could think of any additional issues not covered by

the survey.

50

Statement additions and/or deletions and wording changes were made
before the investigator’s next scheduled interview. The investigator kept track of
changes and made a point of using verbal probing on those statements during the
subsequent interviews to ensure that wording changes clariﬁed statements for all
women.

Based on the interviews, components of the work environment perceived
to positively inﬂuence a woman’s ability to breastfeed included: provision of a
private room, access to a sink and reﬁigerator, when women felt as if pumping
was considered to be an acceptable workplace behavior, ﬂexibility in break times,
and working with a female manager/coworkers. Components perceived as having
a negative inﬂuence included: no established lactation area, when the lactation
area was far from the work area or not consistently available, if breaks were not
frequent or long enough, if women were too busy to take breaks, if women had to
travel away from the worksite, and working with a male manager/male
coworkers. Based on an evaluation of information gathered during the open-
ended questions, it was found no new subscales were needed. Survey items were
added or reworded based on an aggregate of comments from the interviews.

A ﬁnal revision to the instrument was done by consensus of the primary
investigator and the graduate student investigator based on information gained
during the interview process. One revision to the instrument was that four of the
items were moved to a separate yes or no question area after it was determined

that these items did not ﬁt the structure of a Likert scale. These items ask about

51

provision/obtainment of breast pumps, of a storage place for expressed breast

milk, and of a space to breastfeed or express breast milk.

3.1.10 Expert review of survey items

Following the cognitive interviews, expert review was conducted to assess
the survey subscales and to review items. Experts used in the review included the
Michigan Coordinator of La Leche League International (LLLI) Leaders, LLLI
Leaders who have dealt extensively with working mothers within the LLLI group
setting, A Pediatric DO, a registered dietitian with IBCLC certiﬁcation, a
professor and extension specialist at the University of Nevada, a practitioner with
IBCLC certiﬁcation, the program leader for Michigan State University
Extension’s Breastfeeding Initiative, the breastfeeding coordinator for W1C in
Michigan, and an associate professor of Nutritional Sciences at the University of
Connecticut.

The expert reviewers were sent a description of each subscale (see
Appendix A). They rated each item as having a high, moderate or low ﬁt within
that subscale. A majority of the expert rated each item as high ﬁt within each
subscale. The exception was one reviewer who rated two of the items as having a
moderate ﬁt. The reviewers were also asked to conﬁrm that there was adequate
coverage of the content area and that items were clearly worded. No further

revisions were needed based on reviewer comments.

52

3.2 Phase 2 — Piloting of the Instrument ‘The Employee Perception of
Breastfeeding Support Questionnaire’ (EPBS-Q)

Recruitment of respondents was done by creating inclusion criteria for the
business sectors from which the women could be recruited and criteria for
personal characteristics. An attempt was made to recruit women from at variety
of companies having headquarters or single locations in Michigan. D&B’s
Million Dollar Database was used to create a list of potential companies from
which respondents could be recruited. The database was used to ﬁnd companies
in the sector of communications, ﬁnance, insurance, public administration, public
education, and real estate. The sector of public education was excluded from the
search because companies had already been identiﬁed in this sector. Research has
found these industries most likely to have work-family beneﬁts, which increases
the likelihood of them having some level of measurable breastfeeding support
(Burke, 2005; Goodstein, 1994, 1995; Milliken, Martins, & Morgan, 1991).
Search limits were set within the D&B’s Database for State (Michigan),
Employee Total (>250), Industry SIC codes (varied by sector), and Location
(Headquarters, Single Location). Limiting the results to companies within
Michigan eliminates possible confounders introduced by differing state laws on
breastfeeding accommodations and by differences in corporate policies and
administration in different locations (Martucci & Coverdalel, 2001; Scheider,
1990). An employee total of greater than 250 was chosen because of evidence to
suggest that larger companies provide more family-friendly beneﬁts (Burke,

2005). In addition to directly contacting human resource managers at potential

53

companies, recruitment also took place through word-of-mouth and by the posting
of recruitment ﬂyers in establishments and through organizations likely to be
utilized by employed women who were pregnant or who had given birth in the

last year.

3.2.] Recruitment of subjects for piloting the EPBS-Q

All potential respondents answered questions over the telephone to
determine eligibility (see Appendix A). Eligibility was determined based on
inclusion criteria developed for piloting the EPBS-Q. Women were recruited if
they were pregnant or had given birth in the last year, if they worked in a non-
managerial position, and had either returned or planned to return to full-time work
(2 32 hours per week), by the time their child was three months old. Respondents
had breastfed while working full-time or cited employment as a factor for not
initiating or for discontinuing breastfeeding. Respondents currently pregnant
needed to state an intention to breastfeed. Additionally, if the woman did not
work in one of the pre-selected business sectors, it was determined if her
workplace made her ineligible for inclusion. If the woman worked in sectors such
as retail, service, or for a ﬁanchise she was ineligible because it was determined

that her working conditions would be disparate from the target companies.

3.2.2 Procedures for piloting testing of the EP‘BS-Q

After it was established that the respondent met all of the inclusion

criteria, her name, mailing address, phone number and the date the EPBS-Q was

54

mailed were recorded in an address database. Each respondent was then mailed a
packet that included a cover letter (see Appendix A), consent form (see Appendix
A), the EPBS-Q (see Appendix A), a page for writing in additional comments or
suggestions, a form to indicate where the gift certiﬁcate should be mailed, and a
postage-paid envelope addressed to the investigator. When a packet was returned
to the investigator, the consent form and the page indicating where the gift
certiﬁcate should be mailed was separated from the EPBS-Q. The consent form
was ﬁled with the respondent’s phone screener and a gift certiﬁcate was mailed to
the indicated address. The date the packet was received by the investigator, and
the date the gift certiﬁcate was mailed to the respondent, was then added to the
address database. The EPBS-Q, along with the participant and workplace
characteristics, was coded so that no names were associated with survey
responses.

The cover letter requested that the respondents return the EPBS-Q within
two weeks from the date of receiving it. To increase the response rate, any
respondent not returning the EPBS-Q within 6 weeks was sent an additional
survey packet, if that packet was not returned within two weeks, a third mailing
was sent out. If the third mailing was not received the participant was considered
as being lost to follow-up.

Data was entered into an Excel database using a coding scheme for each
variable as described (see Appendix B). After the primary graduate student
entered the data, the data was doubled entered by another student to check for

accuracy. After both students entered the data independently, a visual check was

55

made to determine if any of the data had been entered differently. Any data found
to be entered incorrectly was checked against the hardcopy of the survey.
Changes were then made to the database as needed to accurately reﬂect data

collected on the surveys.

3.3 Scaling

Three models were developed to depict a relationship between the items.
Model D5 is a ﬁve-dimensional model based on company polices/practices,
manager support, coworker support, workﬂow, and physical environment. Model
D2 is a two-dimensional model based on the components of company
polices/practices and manager/coworker support. Model D3 is a three-
dimensional model based on the components on time, support, and acceptability
(see Figure 2). These models were based on a review of the literature that

attempts to characterize successﬁrl breastfeeding.

56

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Model D2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Company policies 2 Dimensions
& practices
17 items
Company policies
Workﬂow & practices
Model D5 7 items 29 items
5
Dimensions Physical
environment
5 items
Manger Support . x
1 3 items
Manager &
>_ Coworker
Coworker Support Support
9 items 22 items
____./
Model D3 Support Time Acceptability
3 Dimensions 24 items 8 items 19 items

 

Figure 2. Three models to depict a relationship between items

57

 

 

 

 

Structural parameters were estimated for each model using ConQuest
Software (Version 1.0, 1998) and model ﬁt was compared. Parameters were
estimated for the Multidimensional Random Coefﬁcients Multinorrrial Logit
Model (MRCMLM) for each of the speciﬁed dimensionality models (Model D5,
Model D2, and Model D3) as well as for two test models, a unidimensional model
(Model D1) and a model (Pos vs. Neg) that scaled normal coded items vs, reverse
coded items. Using those estimated parameters; expected responses were
obtained for each person-by-item combination in the original dataset, and these
expected values were compared to observed values to compute ﬁt statistics for
each person and each item.

Six items were ﬂagged because they consistently showed misﬁt across the
speciﬁed dimensionality models. In other words, these items were associated
with a respondent response pattern that was inconsistent with the measurement
models and none of the models could account for the variability in these items.
These items were scrutinized to determine if they should be removed based on
substantive reasoning.

Analysis was then done to determine which of the models exhibited the
best ﬁt. The Bayesian Information Criteria (BIC), a misﬁt statistic, was used to
compare models. (see Appendix B). BIC is a global statistical criterion for model
selection. The interpretation of the BIC value is that the larger the value, than the
less well the data ﬁts the model, therefore, the model with the smallest value is

determined to be the best model. The BIC is a correction of the deviance statistic

58

for the number of parameters. The BIC equalizes parameters so that the deviance
statistic, which is bias in favor of models that have more parameters, is equalized
for each model. The BIC looks at parameters for the model and exhibits how
much error is left after the model is explained. Given estimated models, the
model with the lower value of BIC is the one to be preferred. From this analysis
it was determined that the two-dimensional model (Model D2) was the best model
to depict a relationship between the items.

Analysis was done to determine item-ﬁt within Model D2 (Figure 1),
when the data were scaled to a multidimensional measurement model. Model D2
contains two unidimensional scales. Dimension 1 includes those items related to
company policies and practices, while dimension 2 includes those items related to
manager and coworker attitudes and support. Items found to have a standardized
mean-squared ﬁt statistic of an absolute value of ﬁt greater than 2 were
scrutinized to determine if they should be removed from the instrument based on

substantive reasoning.

59

CHAPTER 4

“The Development of an Instrument to Measure Female Employees’ Perceptions
of Workplace Breastfeeding Support”

4.1 Abstract

Breastfeeding rates remain low in the United States, especially among
working women. Unfortunately, no quantitative instrument exists to facilitate the
examination of why women who return to work discontinue breastfeeding sooner
than the general population. The objective of this study was to develop an
instrument to measure female employees’ perceptions of breastfeeding support in
the workplace, which would be suitable for piloting with the target population.
Examination of the literature, reviews with experts, and one-on-one interviews
with women who had experience combining breastfeeding and work were used to
create the instrument subscales and items. Examination of the literature was used.
to develop four subscales: company policies/work culture, manager support,
coworker support, and workﬂow. Expert review resulted in the addition of a ﬁfth
subscale, the physical environment of the breastfeeding space. One-on-one
interviews were used to ensure that the item wording was appropriate for the
target p0pulation. Eighteen items were added to the initial item pool and ﬁfteen
were reworded based on comments from the expert review and from the
interviews. The resulting survey contained ﬁfty-four items that required either
categorical yes/no or Likert scale responses. Results from this process indicate
the survey subscales and items adequately reﬂect women’s perceptions of
breastfeeding support in the workplace and the instrument is appropriate for

piloting with new mother employees.

60

4.2 Introduction

Research shows that breastfed infants have decreased incidences and
severity of many infectious diseases including diarrhea (Heinig, 2001; Raisler,
Alexander, & O'Campo, 1999) respiratory tract infections (Bachrach, Schwarz, &
Bachrach, 2003; Oddy et al., 2004), and otitis media (Ball & Wright, 1999;
Dewey, Heinig, & Nommsen-Rivers, 1995), when compared to infants fed
commercially prepared infant formula. Although most women giving birth in the
United States begin breastfeeding, a majority of them do not continue for as long
as recommended. In 2004, 74% of women began breastfeeding, but only 41.5%
of those women continued through six months ("Breastfeedingz Data and
Statistics: Breastfeeding Practices —— Results from the National Immunization
Survey", 2004). Breastfeeding rates are even lower among working women,
especially when comparing women who work full-time to those who work part-
tirne ( Global Strategy for Infant and Young Child Feeding, 2003; McKinley &
Hyde, 2004; Visness & Kennedy, 1997) and to those who do not work outside of
the home ("Breastfeeding Trends Through 2002", 2003). One potential
explanation for why working women have lower breastfeeding rates is that new
mother employees do not perceive the work environment, or “ work climate”, as
being supportive of breastfeeding (Cardenas & Major, 2005; Rojjanasrirat, 2004).
The work climate, deﬁned as “shared perceptions of organizational policies,
practices, and procedures, both formal and informa ” (Schneider, 1990),

inﬂuences which behaviors are viewed as acceptable in the workplace.

61

Studies examining breastfeeding in the workplace have used varying
methods of data collection such as interviews (retro- and prospective) (Auerbach,
1984; Taveras et al., 2003), self-administered questionnaires (Cohen & Mrtek,
1994; Ortiz, McGilligan, & Kelly, 2004), and secondary data analysis (Cardenas
& Major, 2005), making comparisons across studies difficult. The Breastfeeding
Attrition Prediction Tool (BAPT), an instrument that has been used in a number
of studies, quantiﬁes women’s risk for early weaning at the time a new mother
leaves the hospital after delivery (Dick et al., 2002). The BAPT, however, does
not focus on breastfeeding issues in the workplace. In order to better understand
reasons for breastfeeding attrition among working women, a quantitative
instrument for use in data collection is needed. To develop an instrument suitable
for piloting, a carefully designed process appropriate for the individual project
must be deve10ped (Wolfe & Smith, 2007a). In this study, we describe the
process of developing such an instrument through use of a review of the literature,
reviews by experts, and cognitive interviews with women from the target
population. This study reports these initial steps used to create a new instrument
designed to measure perceptions of workplace breastfeeding support from the

viewpoint of new mothers.

4.3 Materials and methods
4.3.1 Study Design
Three steps were followed to develop the instrument subscales and items:

a review of the literature, reviews by experts and cognitive one-on-one interviews.

62

4.3.2 Literature review

The social ecological model was used as the theoretical basis of the
instrument (Glanz, 2002). The social ecological model recognizes that there are
many layers of inﬂuence that affect health behaviors such as breastfeeding. A
review of the literature from a variety of research areas, such as breastfeeding
support, health behavior, work climate, and work-family balance, was conducted
to determine what layers of inﬂuence exist in the workplace. Applied to the
workplace, these layers of inﬂuence include the space used for breastfeeding
(built environment), company policies (organization structure), managers and
coworkers (human aggregate), and the supportiveness for breastfeeding behaviors
(social climate). Research conducted with working women who breastfed
(Auerbach & Guss, 1984; Cohen & Mrtek, 1994; DiGirolamo, Thompson,
Martorell, F ein, & Grummer-Shawn, 2005; Hills-Bonczyk, Avery, Savik, Potter,
& Duckett, 1993; Johnston & Esposito, 2007), companies (Dunn, Zavela, Cline,
& Cost, 2004), managers (Bridges, Frank, & Curtin, 1997; Brown, Poag, &
Kasprzycki, 2001; Libbus & Bullock, 2002), coworkers (Gerard H. Seijts, 2004),
and in the area of work-family balance (Cardenas & Major, 2005; Kopehnan,
Prottas, Thompson, & Jahn, 2006; Thompson, Jahn, Kopelman, & Prottas, 2004)
was examined to discover common themes within each layer of inﬂuence, which
were then organized into subscales for the instrument. Review of the literature
was also used to develop the instrument items for each subscale in order to
address speciﬁc components of the workplace found to either support or hinder

breastfeeding.

63

4.3.3 Expert reviews

Reviews by experts were used to evaluate the appropriateness, content,
clarity, and sensitivity of the survey subscales and items (Wolfe & Smith, 2007a).
Experts were mailed or emailed a copy of the survey, which included space by
each item and at the end for comments. The expert reviewers included
International Board Certiﬁed Lactation Consultant’s (IBCLC), the Michigan State
Coordinator of Le Leche League International Leaders (LLLI), LLLI Leaders
who have dealt extensively with working mothers within the LLLI group setting,
the leader of a breastfeeding support program, the breastfeeding coordinator for
the Special Supplemental Nutrition Program for Women, Infants, and Children
(WIC) in Michigan, and researchers with an expertise in nutrition or business.
The reviewers were instructed to assess (a) the comprehensiveness of the survey
subscales, (b) the need for additional items, (c) the clarity of item wording, and
for those items found to be unclear, to suggest alternative wording.

Additionally, an expert in the ﬁeld of evaluation design and
implementation participated throughout the instrument development process.
This expert assisted in all of the study procedures to ensure that standard methods
of instrument design were being followed (Messick, 1995; Wolfe & Smith, 2007a,
2007b).

4.3.4 Interviews

Items were then pilot tested during one-on-one interviews with a
purposive sample of working women that had given birth within the last year.

Recruitment was by ﬂyers and word-of-mouth, and a phone screener was used to

64

determine eligibility. Women were eligible if they (a) had given birth within the
last year, (b) had breastfeeding experience, (c) had returned to work full-time
(>32 hours per week), and (d) were employed in a non-managerial position.
Interviews took place in a private location and each interview lasted, on average,
one-hour. Participants received a $20.00 gift card.

Each interview began with an open-ended question asking the participant
to describe her experience with breastfeeding at work. The open-ended question
was used before the participants answered the survey items to elicit unbiased
comments related to workplace breastfeeding support. The use of the open-ended
question allowed the investigator to obtain in-depth information about perceptions
of breastfeeding support in the participants’ own words. If needed, verbal probes
were used to ascertain information about people in her workplace and aspects of
the physical work enviromnent making it easier or harder for her to breastfeed at
work. The investigator took hand-written notes during the interviews, including
notes on any non-verbal reactions, and all interviews were also audio-recorded.
The investigator used the audiotapes as needed to clarify hand-written notes.
Information gathered from the open-ended question was examined to determine if
any additional subscales or survey items were needed.

Each participant was then presented with and asked to evaluate the items
written for the initial draft of the instrument. The participant was asked to select
an appropriate response to reﬂect her agreement or disagreement with each survey
item (strongly agree, agree, disagree, or strongly disagree). The participant was

asked to let the investigator know if she felt the need for a “Don’t know” or “Not

65

applicable” option on any item. This information was collected to determine if a
‘neither agree or disagree’ option should be added to the response options on the
instrument. As the participant completed the survey, cognitive interview methods
(e.g. “think-aloud” and verbal probing) 35 were utilized to discover if the
participant was correctly interpreting certain words and phrases, to assess if
additional survey items were needed, and to determine if any items needed to be
reworded. Revisions to the survey items were made based on an aggregate of
comments received during the interviews. The items were then evaluated for
readability using the Gunning Fog Index. The Gunning Fox index estimates that
the instrument is just below an 8th grade reading level. The university’s

Institutional Review Board approved this study.

4.4 Results

Based on the process described above, 54 items were generated that
required either categorical yes/no or Likert scale responses.

4.4.1 Literature review

Common themes identiﬁed through the literature process to have an
inﬂuence on breastfeeding behavior were organized into four areas which served
as the basis for instrument subscales: (a) company policies/work culture, deﬁned
as perceptions of support provided at the organizational level and perceptions
about the perceived appropriateness of breastfeeding in the workplace, (b)
manager support, deﬁned as perceptions of support provided by a woman ’s

immediate manager, (c) coworker support, deﬁned as perceptions of support

66

provided by a woman’s coworkers, (d) workﬂow, deﬁned as perceptions of
available time and perceived ability to make use of given time.

Survey items were written in order to collect information on speciﬁc
factors found within the literature to affect women’s breastfeeding decisions.
These items addressed issues such as: perception of the workplace as a social
acceptable environment for breastfeeding (Cardenas & Major, 2005;
Rojjanasrirat, 2004), ﬂexible scheduling (Cardenas & Major, 2005; Rojjanasrirat,
2004; Taveras et al., 2003), having a place for expressing breast milk that
provides sufﬁcient privacy (Hills-Bonczyk, Avery, Savik, Potter, & Duckett,
1993; Rojjanasrirat, 2004; Taveras et al., 2003), adequacy of breaks (Cardenas &
Major, 2005; Hills-Bonczyk, Avery, Savik, Potter, & Duckett, 1993; Taveras et
al., 2003), availability of storage for expressed breast milk (Cardenas & Major,
2005; Hills-Bonczyk, Avery, Savik, Potter, & Duckett, 1993), knowledge that
other employees or managers had breastfed at the workplace (Rojj anasrirat,
2004), manager (Bridges, Frank, & Curtin, 1997; Brown, Poag, & Kasprzycki,
2001; Cardenas & Major, 2005; Libbus & Bullock, 2002; Rojjanasrirat, 2004) and
coworker (Cardenas & Major, 2005; Rojjanasrirat, 2004; Gerard H. Seijts, 2004)
(Gerard H. Seijts, 2004)attitudes and support, and company-provided
accommodations for expressing breast milk (e. g. lactation programs, pumping
rooms and equipment) (Cardenas & Major, 2005).

4.4.2 Expert review

A review by experts (n=11) resulted in one subscale and twelve new

survey items being added to the instrument (see Table 4). A majority of the

67

experts indicated that the physical environment (e.g. proximity to the
breastfeeding space) was not adequately addressed therefore a ﬁfth subscale:
physical environment of the workplace, deﬁned as perceptions of the
availability/accessibility and adequacy of the space used for breastfeeding
(expressing breast milk) was added. Survey items were then written for the new
subscale. A majority of the experts also commented that the instrument lacked
items related to possible negative consequences of breastfeeding in the workplace.
Therefore, survey items related to negative comments or resentment by coworkers
and possible negative repercussions by the company were added. Additionally,
several experts felt it would be beneﬁcial to ask about company provided
breastfeeding pumps. Overall, the experts commented that the subscales and
survey items comprehensively covered issues relevant to breastfeeding
experiences in the workplace.

Table 4: A review by experts resulted in the addition of one new subscale and
twelve new items within the ﬁve subscales.

 

 

Subscale New items Reason new item(s) was
created
Company My job could be at risk (e.g. lose my Two new survey items
policies and job or get fewer scheduled hours) if were added to address
work I breastfed or pumped breast milk possible negative
culture at work. consequences in the
workplace
My opportunities for job

advancement would be limited if I
breastfed or pumped breast milk at

work.

My company would supply the One new item was added
equipment I would need for to address company
pumping breast milk at work. supplied breast pumps

68

Table 4 can’t

Manager
support

Coworker
support

Physical
environment
of the
workplaceI

My manager would think I couldn’t
get all my work done if I needed to
take breaks for breastfeeding or
pumping breast milk.

My manager would consider
breastfeeding at work a personal
choice, not something he/she should
have to deal with.

My co-workers would think less of
workers that choose to breastfeed or
pump breast milk at work.

My co-workers would think they are
inconvenienced if I took time to
breastfeed or pump breast milk.

The designated place for
breastfeeding or pumping
breast milk at work would be
available when I needed it.2

Two new survey items
were added to address
negativity towards
breastfeeding employees
from managers

Two new survey items
were added to address
negativity towards
breastfeeding employees
from coworkers

Five new items were
added to address the
physical environment of
the workplace

1 This subscale, physical environment of the workplace, was added based on
reviews by experts.
2 This is one example of the ﬁve items that were added to address this subscale.

4.4.3 Interviews

Fourteen one-on-one interviews were conducted. All of the interview

participants had given birth in the last year, were currently breastfeeding, had

returned to work full-time (Z 32 hours per week) by the time their child was three

months old, and were employed in a non-managerial position. An examination of

the comments from both the open-ended interviews and cognitive testing of the

69

survey items revealed that no additional subscales were necessary. However,
based on participant comments survey items were added to four of the ﬁve
subscales to cover issues not previously addressed and 15 of the survey items
were reworded to increase clarity of meaning.

Table 5 shows six addition survey items added to address (a) time spend
away from the worksite, (b) availability of breastfeeding breaks for those women
without scheduled break times, (c) verbal support of breastfeeding by coworkers,
(d) the perceived adequacy of the breastfeeding space, (6) perceptions that
expressing breast milk at work for later use would be considered more acceptable
than having a child brought to the workplace for breastfeeding, and (f) other

breastfeeding women at the workplace.

7O

Table 5: One-on-one interviews resulted in the addition of six new items within
four of the ﬁve subscales.

 

Subscale New items Reason new item(s) was
created
Company My company would accept me Added to address the
policies pumping breast milk at work, but difference between
and work would disapprove of me breastfeeding expressing breast milk
culture my child at work. for later use, and feeding
directly from the breast
I’m certain co-workers have breastfed Added to address
or pumped breast milk at my perceived acceptability
workplace. of breastfeeding
Workﬂow My job includes travel or time away Added to address jobs
from my company, making it difﬁcult that include substantial
to breastfeed or pump breast milk. time away from the
worksite
Some days I would need to skip a Added to address work
breastfeeding or pumping session schedules that do not
because my workdays are so hectic. include scheduled breaks
Coworker My co-workers say things that make Added to assess verbal
support me think they support breastfeeding. support of breastfeeding
Physical The designated place for breastfeeding Added to address the
environment or pumping breast milk includes adequacy of the facility
of the everything I need.
workplace

 

Table 6 shows 15 survey items that were reworded. For example, two

items were reworded to include a deﬁnition to increase understanding of a term

(e. g. ‘job of risk’ was deﬁned as “lose my job or get fewer scheduled hours”).

Some of the wording changes were done to increase consistency of wording

71

throughout the instrument, for example all of the survey items ending in “during
work hours” were changed to “at work”. A few items were reworded to be more
general (e.g. I would feel comfortable asking for a place...” was reworded as, “I
would feel comfortable asking for accommodations. . .”) and a few items were
reworded to be more speciﬁc (e. g. “I could adjust my work schedule...” was
reworded as, “I could adjust my break schedule. . .”). Some of the items were
reworded to more accurately reﬂect how the participants felt about their
breastfeeding experience, for example “My workplace is too fast paced to allow
for breastfeed or pump breast milk” was changed to “Some days I would need to
skip a breastfeeding or pumping sessions because my workdays are so hectic”.
No items were consistently identiﬁed by the interview participants as needing a
‘neither agree or disagree’ option; therefore, the survey will remain a 4-point
Likert scale in order to increase the response rate for each individual item.

Additionally, based on comments received during the interviews, it was
determined that four of the survey items did not ﬁt the structure of a Likert scale.
These items were moved to a separate categorical yes/no area. These four items
assessed whether or not the participant had a designated facility for breastfeeding
and access to equipment for pumping breast milk.
Table 6: One-on-one interviews resulted in the rewording of ﬁfteen items within
the ﬁve subscales.

Subscale - Company policies and practices

Original I would have enough time off after delivery to get breastfeeding

item: started before going back to work.
Reworded I would have enough maternity leave (paid and/or unpaid time off)
item: to get breastfeeding started before going back to work.

72

Table 6 can’t

Original
item:
Reworded
item:

Original
item:
Reworded
item:

Original
item:
Reworded
item:

Original
item:
Reworded
item:

Original
item:
Reworded
item:

Original
item:
Reworded
item:

Original
item:
Reworded
item:

Original
item:
Reworded
item:

Original
item:

My job would be at risk if I breastfed or pumped breast milk at
work.

My job could be at risk (e.g. lose my job or get fewer scheduled
hours) if I breastfed or pumped breast milk at work.

I would feel comfortable asking for a place to breastfeed or pump
breast milk at work.

I would feel comfortable asking for accommodations to help me
breastfeed or pump breast milk at work.

Subscale - Workflow

My breaks are often enough for breastfeeding or pumping breast
milk.

My breaks are frequent enough for breastfeeding or pumping breast
milk.

I have enough time during my breaks to breastfeed or pump breast
milk.

My breaks are long enough for breastfeeding or pumping breast
milk.

My workplace is too fast paced to allow for breastfeeding or
pumping breaks.

Some days I would need to skip a breastfeeding or pumping session
because my workdays are so hectic.

I could adjust my work schedule in order to breastfeed or pump
breast milk.
I could adjust my break schedule in order to breastfeed or pump
breast milk.

My job duties would prevent me from having time to breastfeed or
pump breast milk.
My job includes travel or time away from my company, making it
difﬁcult to breastfeed or pump breast milk.

Subscale - Manager support

My boss would support me breastfeeding or pumping breast milk
during work hours.

My manager would support me breastfeeding or pumping breast
milk at work.

My boss would consider it his/her responsibility to help me combine
breastfeeding and work.

73

Table 6 con’t

Reworded
item:

Original
item:
Reworded
item:

Original
item:
Reworded
item:

Original
item:
Reworded
item:

Original
item:
Reworded
item:

Original
item:
Reworded
item:

My manager would consider it part of his/her job to help me
combine breastfeeding and work.

My boss would think poorly of workers who choose to breastfeed or
pump during work hours.

My manager would think less of workers who choose to breastfeed
or pump breast milk at work.

My boss would help me manage my workload so I could breastfeed
or pump breast milk at work.
My manager would help me deal with my workload so I could
breastfeed or pump breast milk at work.

Subscale - Coworker support

My coworkers would think poorly of workers who choose to
breastfeed or pump during work hours.
My coworkers would think less of workers who choose to
breastfeed or pump breast milk at work.

Subscale -— Physical environment of the workplace

There are too many other women using the breastfeeding or
pumping space.

The designated place for breastfeeding or pumping breast milk at
work would be available when I needed it.

The place for breastfeeding or pumping breast milk is adequate.

The designated place for breastfeeding or pumping breast milk is
satisfactory.

 

4.5 Discussion

The new instrument was designed to assess the level of breastfeeding

support perceived by women to exist in their workplace. The three steps

conducted to design the instrument provide evidence that the survey items are

appropriate measures of workplace breastfeeding support, and therefore the

instrument is suitable to subsequently be piloted with a sample of the target

population

74

The ﬁrst step in this phase of developing the instrument used the social
ecological model to bring together research taken ﬁ'om diverse areas of research.
This allowed for the survey items to address not only speciﬁc facilitators and
barriers found within the breastfeeding literature, but also to address issues from
the perceptive of work climate literature (e.g. perceived appropriateness of
breastfeeding in the workplace). The result of this review, are similar to the
results of a study conducted by Johnston and Esposito (Johnston & Esposito,
2007). Johnston and Esposito conducted a review of the literature to describe
facilitators and barriers to breastfeeding for at least six months by working
women in the United States. Similar to the process used in our study, Johnston
and Esposito used the ecological framework, as applied to the workplace, to
extract data from the literature. Results of Johnston and Esposito’s study revealed
that facilitators and barriers of breastfeeding could be grouped into the categories
of workplace environment (e. g. social support in the workplace and support ﬁ'om
workplace supervisors), time and timing (e. g. maternity leave, required time on
the job, ﬂexible scheduling at work), instrumental support in the workplace (e.g.
child care and the equipment and physical design of the setting), and policies on
breastfeeding. The categories discovered in their review therefore are very
similar to the categories developed for our instrument subscales, which supports
that our instrument has appropriate content to assess the workplace climate for
breastfeeding support from the viewpoint of new mothers.

The reviews by experts provided additional evidence that the newly

developed instrument covers topic areas relevant to the target population and

75

therefore is appropriate for piloting with the target population. Expert review is
important to ensure that the content of the instrument is appropriate and that the
items are worded in a way such that all participants easily understand them. The
experts were useful in this study because many of them had practical experience
with breastfeeding mothers in a wide variety of settings, which allowed them
insight into what issues where important to include on the instrument.

Cognitive interview techniques have increasing been used within the ﬁeld of
instrument design to improve the wording of survey questions (Willis, 1994).
Cognitive methods such as “think-aloud” and verbal probing help researchers to
understand the response process used by the target population when answering
survey items (Willis, 1994). This helps to reword vague or confusing survey
items so that they will be understandable to the target population. In this project,
the use of cognitive interviews was crucial for ﬁnalizing survey items before use
in the piloting phase. Despite scrutiny of the literature and reviews by experts, the
one-on-one interviews revealed that some issues were still not being addressed
and that several of the items should be reworded to increase clarity. Survey items
added to address participant comments included issues such as travel away from
the worksite, which may not be addressed in a typical lactation worksite support
program. Additionally, valuable feedback was received during the interviews on
wording changes that needed to be made to individual survey items, by allowing
the participants to talk through what each statement meant to them, and having

them discuss the thought processes involved in answering each of the items.

76

Although, many of the wording changes were subtle, they helped to increase
consistency throughout the instrument and improved clarity for all participants.
When developing a quantitative instrument speciﬁc guidelines must be followed
to ensure that the instrument is suitable for the purpose intended. Information
gathered from the three steps (review of the literature, expert input, and one-on-
one interviews) provides evidence that the instrument will perform as intended,
which will provide an accurate reﬂection of women’s perceptions of workplace
breastfeeding support. A copy of the instrument can be obtained from the second

author.

4.6 Conclusion

Working women face unique challenges when trying to continuing
breastfeeding while employed full-time outside of the home. Research is needed
to evaluate which speciﬁc components of worksite lactation programs are most
beneﬁcial for increasing breastfeeding duration among women working full-time,
such that more effective interventions, or lactation support programs, may be
developed to improve breastfeeding rates of working women. Research in this
area would be improved by the availability of an instrument that accurately
reﬂects those issues most relevant to the target population of new mother
employees. The initial steps in instrument development reported in the study
demonstrate the suitability of the new instrument for piloting with the target

population, which is reported elsewhere. Taken together, these studies indicate

77

the appropriateness of this new instrument for gathering accurate data on the

workplace climate for breastfeeding support.

4.7 Acknowledgements

This project was funded by the Michigan State University FACT Coalition
and National Research Initiative of the USDA Cooperative State Research,
Education and Extension Service, grant number #2006—35215-16703. Edward W.

Wolfe, PhD. provided guidance on the instrument development process.

4.8 Reprint information
Request for reprints to be sent to Dr. Beth Olson, Department of Food
Science and Human Nutrition, Michigan State University, East Lansing, MI

48824, e-mail: olsonbe@msu.edu

78

CHAPTER 5

“Assessing the Validity of Measures of an Instrument Designed to Measure
Employees’ Perceptions of Workplace Breastfeeding Support.”

5.1 Abstract

Breastfeeding rates among working mothers are lower than among
mothers who are not employed. An ecological ﬁamework suggests that health
behaviors, such as breastfeeding, are inﬂuenced by intrapersonal and
environmental factors. There is no existing instrument to measure women’s
perception of the workplace environment in providing breastfeeding support. The
objective of this study was to pilot an instrument measuring perceptions of the
work climate for breastfeeding support among working women. Data were
collected from self-administered mailed questionnaires ﬁlled out by 104 pregnant
or women that had recently given birth and were employed and breastfeeding.
Dimensionally analyses supported the two dimensional model suggested by the
literature. Internal consistency reliability coefﬁcients were high (near .90), and the
correlation between the subscales was moderately strong (.68). Only a single item

exhibited misﬁt to the scaling model, and that item was revised after review.

5.2 Introduction
Research shows that breastfed infants have decreased incidence and
severity of many infectious diseases including diarrhea, (Heinig, 2001; Raisler,

Alexander, & O'Carnpo, 1999) respiratory tract infections, (Bachrach, Schwarz, &

79

Bachrach, 2003; Oddy et al., 2004) and otitis media, (Ball & Wright, 1999;
Dewey, Heinig, & Nommsen-Rivers, 1995) when compared to infants fed
commercially prepared infant formula. Although most women giving birth in the
United States begin breastfeeding, a majority of them do not continue for as long
as recommended. In 2004, 74% of women began breastfeeding, but only 41.5%
continued through six months. ("Breastfeeding: Data and Statistics: Breastfeeding
Practices — Results from the National Immunization Survey", 2004)’
Breastfeeding rates are even lower among women who work full-time.
("Breastfeeding Trends Through 2002", 2003; , Global Strategy for Infant and
Young Child Feeding, 2003; McKinley & Hyde, 2004; Visness & Kennedy, 1997)
One potential explanation for the lower breastfeeding rates among working
women is that employees who are new mothers do not perceive the work
environment, or climate, as being supportive of breastfeeding. (Cardenas &
Major, 2005; Rojjanasrirat, 2004) The work climate, deﬁned as “shared
perceptions of organizational policies, practices, and procedures, both formal and
informal”, (Schneider, 1990) inﬂuences which behaviors are viewed as acceptable
in the workplace. Studies have suggested elements of the workplace environment
that may encourage breastfeeding, including: ﬂexible scheduling, (Cardenas &
Major, 2005; Rojjanasrirat, 2004; Taveras et al., 2003) having a place for
expressing breast milk that provides sufﬁcient privacy, (Hills-Bonczyk, Avery,
Savik, Potter, & Duckett, 1993; Rojjanasrirat, 2004; Taveras et al., 2003)
adequacy of breaks, (Cardenas & Major, 2005; Hills-Bonczyk, Avery, Savik,

Potter, & Duckett, 1993; Rojjanasrirat, 2004; Taveras et al., 2003) availability of

80

storage for expressed breast milk, (Cardenas & Major, 2005; Hills-Bonczyk,
Avery, Savik, Potter, & Duckett, 1993) knowledge that other employees or
managers have breastfed at the workplace, (Rojj anasrirat, 2004) and company-
provided accommodations for expressing breast milk (e. g. lactation programs,
pumping rooms and equipment). (Cardenas & Major, 2005) Research also
demonstrates that manager (Bridges, Frank, & Curtin, 1997; Brown, Poag, &
Kasprzycki, 2001; Cardenas & Major, 2005; Libbus & Bullock, 2002;

Rojj anasrirat, 2004) and coworker attitudes and support (Cardenas & Maj or,
2005; Rojjanasrirat, 2004; Gerard H. Seijts, 2004) play an important role in
whether or not women continue to breastfeed after returning to work.

Studies focusing on this issue have used a variety of methodologies, such as
interviews (retro- and prospective), (Auerbach, 1984; Taveras et al., 2003) self-
adrninistered questionnaires, (Cohen & Mrtek, 1994; Ortiz, McGilligan, & Kelly,
2004) or secondary data analysis. (Cardenas & Major, 2005) The Breastfeeding
Attrition Prediction Tool (BAPT), an instrument that has been used in a number
of studies, quantiﬁes women’s risk for early weaning at the time a new mother
leaves the hospital after delivery. (Dick et al., 2002) The BAPT, however, does
not focus on breastfeeding issues in the workplace. The authors designed an
instrument to elucidate causes of breastfeeding attrition among working women
by capturing women’s perceptions of workplace breastfeeding support. Data
from this instrument may be used to assist in the development of more effective
worksite lactation support programs, resulting in higher rates of breastfeeding

among working mothers. This study reports validity evidence of the measures

81

included in this newly developed instrument, the Employee Perceptions of

Breastfeeding Support Questionnaire, or EPBS-Q.

5.3 MATERIALS AND METHODS

5.3.1 Study Design

The 117 women included in the study were pregnant or had given birth in
the last year, worked in a non-managerial position, and had either returned or
planned to return to full-time work (2 32 hours per week) by the time their child
was three months old. An attempt was made to recruit women ﬁ'om industries
known to have family-friendly policies (to increase the likelihood of having
measurable breastfeeding support), (Burke, 2005) and having headquarters or
single locations in the State of Michigan (in order to minimize company
differences based on site location). Surveys were mailed to women who agreed to
participate; those not responding were contacted with two follow-ups. Women
received a gift certiﬁcate upon return of a completed survey. A total of 104

questionnaires were completed and returned (89% response rate).

5.3.2 Subjects

Table 7 summarizes the demographic characteristics of the 104
participants. The sample was racially diverse, but tended to be older, had attained
higher levels of education, and had higher income than the general population.

Ninety-one percent of the subjects were married or living with a partner and the

82

remainder were single. Eighty-ﬁve percent of the subjects had previous

breastfeeding experience, and at the time of the study 48% of the women were

currently breastfeeding. Most of the subjects (67%) had worked at their place of

employment for ﬁve years or less, and 96% worked 32 or more hours per week.

Fifty-two percent were salaried with the remaining being hourly workers — of

these 34% were in a union. F orty-two percent had a male manager and 58% a

female manager. Thirty-six percent of the women worked in public education,

27% in health care, and 13% worked for the state or federal government. The rest

of the respondents (24%) worked in the communications ﬁeld; in the corporate

ofﬁces of a large manufacturing company; or for ﬁnance, real estate, or insurance

companies.

Table 7. Demographics of the Sample

 

 

 

Variable Level Statistic
Race White 81%
Aﬁican American 8%
Asian 5%
Hispanic 3%
Other 3%
Age Min = 20
Mean = 30.5
Max = 40
Education Level High School 2%
Some College 17%
College Degree 39%
Post-Graduate Degree 41%
Income 2 $60,000 52%
$30,000 - $59,000 36%
_<_ $29,000 12%
Note: N = 104

83

5.3.3 Measures

Items for the EPBS-Q were developed based on information from several
sources: review of the literature, expert input (the 11 experts included
practitioners with credentials and experience in the lactation ﬁeld and researchers
in evaluation design or lactation), and 14 one-on-one interviews which employed
cognitive interview techniques with women who had breastfed while employed.
Based on these initial efforts, 54 items were generated that required either
categorical yes/no or Likert scale responses. Survey items were grouped together
based on the speciﬁc aspect of the work climate being evaluated: company (e. g. “I
would feel comfortable asking for accommodations to help me breastfeed or
pump breast milk at work”), manager (e. g. “My manager would help me deal with
my workload so I could breastfeed or pump breast milk at work.”), coworkers
(e. g. “My co-workers would think less of workers that choose to breastfeed or
pump breast milk at work”), workﬂow (e. g. I could adjust my break schedule in
order to breastfeed or pump breast milk.”), or physical environment (e. g. “I would
feel comfortable breastfeeding or pumping breast milk in the designated place”).
A copy of the instrument can be obtained from the third author. Based on pilot
test item analyses and content considerations, 13 of the items were removed from
the initial item pool leaving a total of 41 items on the EPBS-Q.

5.3.4 Scaling

Data were initially scaled with the Multidimensional Random Coefﬁcients

Multinomial Logit Model, (Adams, Wilson, & Wang, 1997) a multidimensional

extension of the Rasch measurement model, (Rasch, 1980) as implemented in the

84

Conquest computer program. (Wu, Adams, & Wilson, 1998) This logistic latent
trait model assumes that multiple traits (speciﬁed by the analyst) underlie a
mother’s response to each item. Application of this model to our data results in an
estimate of each mother’s perceived level of support on each dimension being
measured and each item’s perceived availability in the work place. Agreement
between model-based expected values and observed responses served as the basis
for evaluating the dimensionality of the measures.

The ﬁt of two substantive models to these data was evaluated: (a) a two-
dimensional model based on the components of company polices/work culture,
deﬁned as perceptions of support provided at the organizational level and
perceptions about breastfeeding (expressing breast milk) being considered socially
acceptable in her workplace, and manager/coworker support, deﬁned as
perceptions of support provided by a woman ’s immediate manager and her
coworkers (D2) and (b) a three-dimensional model based on the components of
time, deﬁned as perceptions of available time and perceived ability to make use of
given time, support, deﬁned as perceptions of breastfeeding support available ﬁ'om
her company, ﬁom her manager, her coworkers, and ﬁom the physical environment
and accommodations of her workplace, and acceptability, deﬁned as perceptions
about breastfeeding (expressing breast milk) being considered socially acceptable
in her workplace. (D3). In addition, we considered two comparison models; a
unidimensional model (D1) and a two-dimensional model that scaled positively
coded items versus negatively coded items as separate dimensions (DPN). The

relative ﬁt of the data to each of these models was evaluated using the Schwartz’s

85

Bayesian Information Criterion (BIC) (see Appendix B), which transforms the
deviance statistic (G2) by imposing a correction that takes into account the number
of parameters estimated for the model in question and penalizes free parameters
more strongly than does the frequently cited Akaike Information Criterion.
Generally, smaller values of the BIC index indicate better ﬁt of the data to the
model.

After identifying the best ﬁtting model ﬁom these options, we scaled the
data from each subscale separately to the Rasch model (Rasch, 1980) using the
Winsteps software (Linacre, 2007) and examined several indices relating to the
quality of the measures produced for each subscale (The complete data output for
Dimension 1 and Dimension 2 is located in Appendix B). First, we examined the
reliability of separation index for each scale speciﬁed in the model that we

adopted. The reliability of separation is an index that is a latent trait modeling

analog to coefﬁcient alpha computed as Rela = [1 — MSE9]/V(t9) , where MSE, is

the mean of the squared errors of the mother trait estimates and V09) is the

variance of those estimates. Second, we examined unweighted mean-squaredﬁt
index for each item, which equals the average squared standardized residual of
mothers’ ratings from the Rasch model expectations. The mean-squared ﬁt
indices can be standardized (Zmeigmd) to facilitate interpretation so that positive
values indicate that the amount of variability between observed and expected
ratings is greater than that predicted by the model. For this study, item ﬁt indices
with values greater than 2.0 were considered to be evidence of model-to-data

misﬁt. Third, we examined the correlation between the ratings assigned by

86

mothers to each item and the mother measures (rummmm), an index that is
analogous to the raw item score-total score correlation (AKA the point-polyserial
correlation). Near zero values of this correlation indicate items that may not be
adequate measures of the underlying dimension that the item was designated to
measure. In our analyses, items were ﬂagged if this correlation was less than .30.
Fourth, we examined the correlations between mother trait estimates on the

subscales.

5.4 Results
Table 8 summarizes the dimensionality analyses. The models are

presented from the simplest (D1) to the most complex (D3) from the top to the
bottom of this table, and the Deviance statistic (smaller values indicate better
model-to-data ﬁt) for Model D2 has the smallest value. Similarly, the value of the
BIC transformation of the Deviance statistic (which penalizes models that contain
more parameters—leveling the playing ﬁeld between more complex and simpler
models), indicates that D2, which differentiates between company policies/work
climate and manager/coworker support, exhibits the best ﬁt to the observed data.
Overall, the table indicates that the two-dimensional substantive model better
explained the observed data than did the unidimensional model (D1) or the three-
dimensional model (D3). In addition, D2 performed slightly better than the model
that differentiated positive and negative items (DPN). This fact is important
because it indicates that (a) the substantive distinctions upon which the items were

developed were realized in the relationships between items and (b) the fact that

87

negatively polarized items included in the item pool did not interfere with
measurement of the intended constructs, as sometimes happens. (Barnette, 1999)

Table 8. Dimensionality Statistics

 

Model Deviance Parameters BIC

 

D1 8709.75 51 8812.61
D2 8602.31 53 8709.21
DPN 8652.74 53 8759.64
D3 8614.79 56 8862.74

 

Note: N = 104. BIC = Bayesian Information Criterion

Table 9 displays the subscale reliability of separation indices and the
subscale correlations. The reliability of separation indices, measures of internal
consistency, are generally high, and the correlation between the Rasch model

scaled mothers’ measures are moderately high.

Table 9. Subscale Correlations and Reliabilities

 

Dimension 1 Dimension 2
Dimension 1 .87
Dimension 2 .68 .89

Note: N = 104. Diagonals show reliability of separation. Dimension 1 =
company policies/work culture and contains 23 items. Dimension 2 =
manager/coworker support and contains 18 items.

Table 10 summarizes the item quality indicators for each subscale. This
table provides the range and average values of the unweighted mean-squared ﬁt
statistics (an indicator of how well responses to each item conform to the
expected values of the Rasch model) and item—measure correlations (an indicator

of the consistency of responses to a particular item to the composite of the

responses to the remaining items on the subscale) and indicates how many items

88

were ﬂagged based on the criterion values identiﬁed in the Method section.

These statistics reveal that the items on both dimensions exhibited reasonably
good ﬁt to the Rasch model, with only one item being ﬂagged for misﬁt—Item
20, which measured manager/coworker support. In addition, the table shows that
each item exhibited reasonably high levels of agreement with the remaining items
on the relevant subscale. Speciﬁcally, the item-measure correlations tended to be
fairly high for all items. Three items ﬁom Dimension 1 were ﬂagged for
somewhat low item-measure correlations, but further examination revealed that
these items were among the easiest and most difficult to endorse—a characteristic

that attenuates item-measure correlations toward zero.

Table 10. Summary of Item Quality Indicators

 

 

Index Statistic Dimension 1 Dimension 2
Zmeigmed Min -2.00 -1 .90
Mean 0.10 0.00
Max 1.20 4.60
% > 2.00 0 6
ritem-measure M1n.22 .36
Mean .47 .60
Max .71 .74
% < .30 13 0

 

Note: Dimension 1 = company policies/work culture and contains 23 items.
Dimension 2 = manager/coworker support and contains 18 items.
5.5 Discussion
Results of our study suggest that measures from the EPBS-Q may be
useful indicators of a breastfeeding woman’s perceptions of the work climate.

Our analyses provide evidence to support the validity of measures of two

89

constructs related to the work climate for breastfeeding support: company
policies/work culture and manager/coworker support. Our results indicate that
measures from these two subscales exhibit sufﬁcient internal consistency (i.e.
have high reliabilities) for use in research and program evaluations. In addition,
the correlation between measures from the two subscales indicate that the
measures are not redundant in the information that they provide about mothers—
the subscales have only about 46% of their variance in common. Not only do our
results indicate that measures from both subscales are reliable, but they also
demonstrate that each subscale provides unique information about women’ s
perceptions of the work climate for breastfeeding support, allowing researchers to
understand and differentiate between issues relating to the employment setting
versus the people in that setting.

In addition, our results demonstrate that the items are good indicators of
these traits. Only a single item—Item 20, which measured manager/coworker
support—exhibited misﬁt. Typically, items ﬂagged for misﬁt are scrutinized to
see whether there are substantive reasons to omit or revise the item. The wording
of this item, “My manager would consider breastfeeding at work a personal
choice, not something he/she should have to deal with”, required participants to
think of the issue ﬁ‘om the viewpoint of her manager. Clearly, this item is worded
in a manner that requires women to speculate about the thinking of their
managers. Because the inclusion of this item is warranted based on results of
prior research, (Dunn, Zavela, Cline, & Cost, 2004) we recommend that the text

of the item be revised rather than omitting the item. To remove the uncertainty of

90

requiring the employee to comment on what may be unspoken attitudes of the
employer, the item has been reworded as, “I feel breastfeeding would be viewed
as a personal choice by my manager”, to depict the viewpoint of the woman

instead of the viewpoint of the manager.

5.6 Conclusion

Employer-sponsored lactation support programs, in general, have been
shown to increase rates of breastfeeding duration. However, these programs vary
in the amount and types of support offered. In addition, evaluations of these
programs have used different instrumentation to measure outcomes, (Cohen &
Mrtek, 1994; Katcher & Lanese, 1985; Ortiz, McGilligan, & Kelly, 2004) making
it difficult to determine which components are most inﬂuential in helping women
continue breastfeeding while employed. Our goal was to create an instrument that
could be used to gather relevant data in a standardized manner that could be
replicated across studies and across employment contexts. Use of an instrument
like the EPBS-Q to measure program effectiveness may lead to a better
understanding of the components of effective programs because the EPBS-Q can
be used to gather comparable data from a variety of workplace settings, giving a
more robust assessment of which speciﬁc components of a worksite lactation
program are most beneﬁcial to women. More important is the fact that the data
will be collected from the people impacted by such programs—new mother

employees who wish to combine breastfeeding and employment.

91

The evidence provided in this article suggests that the EPBS-Q measures
are valid for use in populations similar to those used in this study. Future research
is needed before this instrument is used with other populations, and such a study
should also aim to verify that the revised Item 20 functions as intended. In
addition, evidence of the relationship of EPBS-Q measures to external measures

of program effectiveness is needed.

5.7 Acknowledgments
This project was funded by the Michigan State University FACT Coalition
and the National Research Initiative of the USDA Cooperative State Research,

Education and Extension Service, grant number #2006-35215-16703.

5.8 Reprint information
Request for reprints should be sent to Dr. Beth Olson, Department of Food
Science and Human Nutrition, Michigan State University, East Lansing, MI

48824, e-mail: olsonbe@msu.edu

92

APPENDICES

93

APPENDIX A

94

RECRUITMENT FLYER

 

 

Have You
Had a Baby in the Last Year?

“Wiluumhamﬂlm.

 

74’ 0 YOU ARE ELIGIBLE IF YOU. .
\

 

‘3 0 Had a baby in the last year
- Have returned to work for at least 32 hours per week
L 0 You don‘t need to have breastfed the baby
Flu-n ﬂuidl-
mwnamsummmmum
1m1ut154 (hltoenllnbu) ‘

 

95

 

PHONE SCREENER

Name of participant:
Hello, my name is Sally Greene; I am calling because you have shown interest in
participating in a MSU research study about breastfeeding and work. I need to
ask you a few questions to see if you are able to take part in the study. Is this a
good time? If not, when can we schedule a time for me to call you back?

 

If call back is needed:

Name

 

Phone Number

 

Date of call back

 

Time of call back

 

Thank you for your time, I look forward to speaking with you on (date) at (time).

Great, this will only take a few minutes.

Do you voluntarily agree to participate in this study?

Yes No

If no, thank you for your time.

96

 

Now I have some questions about you and your child.

Are you 18 years old or older? Yes _ No

Are you in a managerial position at work? Yes No
(If yes, skip to the ineligibility statement.)

Do you work in a private office with a door? Yes No

(If yes, skip to the ineligibility statement.)

Have you had a child in the last year? Yes No

After the baby was born, did you go back to work full-time (32+ hours/week)?
Yes No

(If no, skip to the ineligibility statement.)

Did you ever breastfeed this child? Yes No

If no:

At any point before or during your pregnancy did you consider
breastfeeding?
Yes No

What was the main (or one) reason you decided not to breastfeed?

 

 

 

 

Was returning to work full-time a factor in your decision not to
breastfeed?

Yes No
(If no, skip to the ineligibility statement.)

 

Additional comments:

 

 

 

97

If yes:
When did you decide you wanted to breastfeed, before you were
pregnant? when you were pregnant? or after delivery?

What was the main (or one) reason you decided to breastfeed?

 

 

 

 

Are you still breastfeeding this child? Yes No

If no,
How long did you breastfeed this child? (weeks/months)

 

When you stopped breastfeeding, was going back to work a factor in your
decision?

Yes No

 

Additional comments:

 

 

 

Thank you for your time, unfortunately you do not meet the criteria to be included
in the study.

OR

Thank you for your time. At this time, I would like to set up a time for us to meet
so that you can participate in the study.

Date Time Place

 

 

Thank you for your time, I look forward to meeting with you on (date) at (time) at

(place).

- l

Based on the following criteria I accept the participant:

98

INFORMED CONSENT — COGNITVE INTERVIEWS

Michigan State University (MSU) is doing a research study about breastfeeding
support in the workplace. We h0pe to learn how women, that have recently had a
baby and returned to work, feel about the level of support provided by their
workplace. Because you are pregnant and plan to breastfed after you return to
work or you have had an infant in the last year and breastfed, we are asking for
your help in this study.

About this study:

0 Being in this study includes reading and answering questions on a survey (takes
about 30 minutes) and being asked to talk about thoughts you had while
answering the questions (takes about 30 minutes). All discussion will be audio
taped.

0 You do not have to be in the study. You can stop at any time with no
consequences for you.

c All of the information you share will be kept private. Only the researchers will
see the information. Your name will not be on the survey. The cassette tapes will
be typed into a printed copy and the tapes will be destroyed. Your name will not
be on the tape or the written copy.

0 We will follow all laws that protect your privacy. We will share results of the
study in group form only.

0 We do not expect that being in the study will harm you in any way. The
information you share may help improve support for new mothers that are
working and breastfeeding.

0 When you ﬁnish the study, you will receive a $20.00 gift certiﬁcate.

If you have any guestions about this study contact Beth Olson, PhD: (1-517-
355-8474 extension 113) Department of Food Science and Human Nutrition,

Michigan State University, 2112 Anthony Hall, E. Lansing, MI 48824, email at
olsonbe@msu.edu.

If you have any guestions about your rights as a human subject contact Peter
Vasilenko: (1-517-355-2180), Chair of the University Committee on Research

Involving Human Subjects, 202 Olds Hall, Michigan State University, E. Lansing,
MI 48824, email at ucrihs@msu.edu. You do not have to give your name.

 

I voluntarily agree to participate in this study.

 

Participant’s printed name

 

Participant’s signature

99

 

Date

I voluntarily give my permission to be audio taped during the study.

 

Participant’s printed name

 

Participant’s signature

 

Date

 

Researcher’s printed name

 

Researcher’s signature

 

Date

100

INFORMED CONSENT - PILOTING

Michigan State University (MSU) is doing a research study about breastfeeding
support in the workplace. We hope to learn how women who are pregnant or who
have recently had a baby and returned to work, feel about the level of support
provided by their workplace. Because you are pregnant and plan to breastfed
after you return to work or you have had an infant in the last year and breastfed,
we are asking for your help in this study.

About this research study:

0 Being in this study includes reading and answering questions on a survey (it
should take you about 30 minutes to complete the study)

0 You do not have to be in the study. You can stop at any time with no
consequences for you.

0 All of the information you share will be kept private. Only the researchers will
see the information. Your name will not be on the survey.

0 We will follow all laws that protect your privacy. We will share results of the
study in group form only.

0 We do not expect that being in the study will harm you in any way. The
information you share may help improve support for new mothers that are
working and breastfeeding.

0 You will receive a $20.00 gift certiﬁcate for your participation.

If you have any Questions about this research study contact Beth
Olson, PhD: (517) 355-8474 extension 113. Department of Food Science

and Human Nutrition, Michigan State University, 2112 Anthony Hall, E.
Lansing, MI 48834, email: olsonbe@msu.edu.

If you have any Questions or concerns about your Ligh’ ts as a research
participant, please feel free to contact Peter Vasilenko, Ph.D., Director

of Human Subject Protection Programs at Michigan State University:
(517) 355-2180, Fax (517)432-4503, email: irb@msu.edu, or regular
mail: 202 Olds Hall, E. Lansing, MI 48824. You do not have to give your
name.

I voluntarily agree to participate in this research study.

 

(Participant’s printed name)

 

(Participant’s signature)

101

 

(Date)

102

GIFT CERTIFICATE CONFIRMATION FORM

I received a $ 20 gift certiﬁcate for participation in the employment and
breastfeeding study.

Address

 

 

 

Signed

 

Date

 

103

OPEN-ENDED INTERVIEW QUESTION
Open-ended question (asked verbally and responses audio-recorded):
1. Were you breastfeeding your baby when you returned to work?
(If, yes) Describe your experience combining breastfeeding and work.

Probes:
Did you frnd it easy or difﬁcult to breastfeed at work? Why?

Was there any thing about the people you work with that made it easier or more
difﬁcult to breastfeed at work?

Was there anything about the physical environment of your work that made it
easier or more difﬁcult to breastfeed at work?

(If, no) Describe your workplace and ways you think it would either
support a woman who wanted to continue breastfeeding, and/or ways you think it
would discourage that woman.

Probes
Were there work related reasons you choose not to breastfeed after returning to
work? What were they?

Was there any thing about the people you work with that would have made it
easier or more difﬁcult to breastfeed at work?

Was there anything about the physical environment of your work that would have
made it easier or more difﬁcult to breastfeed at work?

104

EXPERT REVIEW FORM

Michigan State University

Dept of Food Science and Human Nutrition

2125 S. Anthony

East Lansing, MI 48823

greenesa@msu.edu Date

 

Dear expert reviewer,

Thank you for agreeing to participate in this expert review. As you know, I am
developing a survey to evaluate woman’s perceptions of breastfeeding support in
the workplace. A step in the development process is to allow individuals with
experience in the ﬁeld of lactation to review statements on the survey.

The survey currently contains 50 statements. Women enrolled in the study will be
asked to rate statements from strongly agree to strongly disagree. On the survey
the women receive, the statements will be grouped in ﬁve categories (overall
company, time, managers, co-workers, and space). The survey I am sending to
you contains the same statements that are on the survey the women will receive;
however the layout is different.

For this review the 50 statements have been placed into three categories. These
three categories are time, support, and acceptability. I have supplied a deﬁnition
of each category on the following pages.

The purpose of this review is to read the category deﬁnition and then
evaluate each statement in that category as high, moderate, or low. Your
ranking will depend on how well you feel each statement ﬁts the definition
for each category.

You may ﬁll out and return this form one of two ways.

1. If you would like to print and mail to me. You may print-out the
form, place on x in the appropriate boxes and mail to me.
2. If you would like to ﬁll out on the computer and email back to me.

You may double click on the box you would like to check. A text box will
appear. Under the option default value click on ‘checked’ to place a check
in the box. You can double click on a box to change from checked to not
checked or not checked to checked at any time. When you close the text
box, a check should appear in the appropriate box. Save the ﬁle and then
email the form back to me.

Please do not hesitate to call me with any questions you may have.
(517) 214-8547 or toll-free at 1-866-552-6521 ext. 154

105

Sincerely, Sally Greene

1. Please read the deﬁnition (below) of m.

2. Read each of the following statements one at a time.

3. Rate each statement as Low, Moderate or High depending on how well you
think each statement ﬁts the deﬁnition of Lime.

Time this category includes statements that look at the perceptions women have
of the availability of time given by her work (e.g. for breaks, time off after
delivery, her workload) and by her perceived ability to make use of the given
time.

Statement on woman’s survey Goodness-of-ﬁt for time

Low Moderate High
I would have enough time off after
child is born to get breastfeeding
started before going back to work. D D D

My breaks are frequent enough
for breastfeeding or pumping
breast milk at work. D D [I

My breaks are long enough for
breastfeeding or pumping breast CI El [:1
milk.

I could adjust my work hours to
breastfeed or pump breast milk. D U U

I could adjust my break schedule
to breastfeed or pump breast I3 D 1:]
milk.

I would have time to breastfeed

or pump breast milk, which is in

addition to regular breaks and U El [:1
mealtimes.

Some days I would need to skip

a breastfeeding or pumping

session because my workdays D D D
are so hectic.

106

1. Please read the deﬁnition (below) of acceptability.

2. Read each of the following statements one at a time.

3. Rate each statement as Low, Moderate or High depending on how well you

think each statement ﬁts the deﬁnition of acceptability.

Acceptability this category includes statements that look at the perceptions

 

women have about the extent to which breastfeeding and/or pumping breastrnilk
is considered socially acceptable in her workplace.

Statements on woman’s survey

I’m certain my work has written
policies about employees that are
breastfeeding or pumping breast milk.

I would be able to continue
breastfeeding or pumping breast milk
after going back to work.

My job could be at risk (e.g. lose my
job or get fewer scheduled hours) if I
breastfed or pumped breast milk at
work.

I would be able to talk about
breastfeeding at work.

I would feel comfortable asking for

accommodations in order to breastfeed

or pump breast milk at work.

My opportunities for job advancement
would be limited if I pumped breast
milk at work.

I’m certain that women in higher-level
positions have breastfed or pumped

breast milk at my work.

I’m certain co-workers have breastfed
or pumped breast milk at my work.

107

Good-of—ﬁt for acceptability

Low

El

Moderate

C]

High

El

My company would accept me

pumping breast milk at work, but

would disapprove of me breastfeeding D D CI
my child at work.

1. Please read the deﬁnition (below) of supmrt.

2. Read each of the following statements one at a time.

3. Rate each statement as Low, Moderate or High depending on how well you.
think each statement ﬁts the deﬁnition of supp_ort.

Support this category contains statements that include perceptions women have of
breastfeeding support available from other employees (co-workers, managers) as
well as breastfeeding support provided by the physical environment and '
accommodations of her workplace.

Statements on woman’s survey Goodness-of-ﬁt for support

Low Moderate High

I would be able to get information
about combining work and
breastfeeding from my workplace. C] U D

I’m certain there is a company

designated place for women to

breastfeed or pump milk during the D D D
workday.

There is someone I could go to at

work that would help me make

arrangements for breastfeeding or D D U
pumping breast milk.

My boss would support me

breastfeeding or pumping breast D D [3
milk at work.

My boss says things that make me

think he/she supports breastfeeding. D El D

My boss would consider it part of
his/her job to help me combine D D U
breastfeeding and work.

My boss would help me combine
breastfeeding and work. CI El CI

108

My boss would make sure my job is

covered if I needed time for

breastfeeding or pumping breast '3 '3 CI
milk.

My boss would change my work
schedule to allow me time for

breastfeeding or pumping breast U U U
milk.
Statements on woman’s survey Goodness—of-ﬁt for support

Low Moderate High

My boss would help me manage my
workload so I could breastfeed or

pump breast milk at work. CI Cl C]
My boss would be a source of

information on how to breastfeed at D U U
work.

I am certain co-workers would
support me breastfeeding or
pumping breast milk at work. D U U

Co-workers say things that make me
think they support breastfeeding. D El E1

Co-workers would think they are

inconvenienced if I took time to

breastfeed or pump breast milk at D D U
work.

Co-workers would cover my job

duties if I needed time for

breastfeeding or pumping breast U D '3
milk.

Co-workers would change break

times with me so I could breastfeed

or pump breast milk. D U D

I could buy or borrow the equipment

109

I would need for pumping
breastrnilk at work.

My company would supply the
equipment I would need for U El [:1
pumping breastrnilk at work.

I could ﬁnd a place to store
expressed breast milk at work. E] E] [:1

There is a place I could go to

breastfeed or pump breastrnilk at D D D
work.

Comment: The question above

will be answered as YES or NO by

the participants. Participants that

answer YES will then be asked the

following set of questions.

Statements on woman’s survey Goodness-of-ﬁt for support
Low Moderate High

The place I would use for
breastfeeding or pumping breast
milk would be available when I D U D

needed.

The place I would use to breastfeed

or pump breast milk is close enough

to my work area to use during my I] D D
breaks.

I would feel comfortable
breastfeeding or pumping breastrnilk D II! 1:1

in the available place.

The place for breastfeeding or

pumping breastrnilk is satisfactory. U [I [j
The place I would use is only for

breastfeeding or pumping D El E1
breastrnilk.

Comments

110

COVER LETTER FOR PILOTING
Sally Greene
Michigan State University
Dept of Food Science and Human Nutrition
2125 South Anthony
East Lansing, MI 48824-1224
1-866-552-6521 ext 154
FAX: 517-353-6343

January 30, 2007

Study Participant

I am writing to ask your help in a study of new and expectant mothers working
outside of the home full-time. This study is part of an effort to learn how new and
expectant mothers experience breastfeeding support in their workplace.

It’s my understanding that you are pregnant or have had a child within the last
year. I am contacting a sample of women from around the state to ask them about
their thoughts on combining breastfeeding and working full-time at their place of
employment.

Your input will be used to help develop a ﬁnal survey, which can be used by
companies to see how female employees feel about breastfeeding support in their
workplace. By understanding how women feel about breastfeeding support,
companies can make their services and support as helpful as possible.

Your answers are completely conﬁdential and will be released only as summaries
in which no individual’s answers can be identiﬁed. When you return your
completed questionnaire, your name will be deleted from the mailing list and
never connected to your answers in any way. This survey is voluntary. However,
your experiences and opinions about breastfeeding and working full-time could be
very helpful in developing breastfeeding services and support for working
women. If for some reason you prefer not to respond, please let me know by
returning the blank questionnaire in the enclosed stamped envelope.

I would appreciate if you could return your survey package within 2 weeks of it
being received. You will receive a $20.00 gift certiﬁcate as a thank you for your
help.

Sincerely,

Sally W. Greene
Master's Candidate

111

EMPLOYEE PERCEPTIONS OF BREASTFEEDING SUPPORT
QUESTIONNAIRE

Breastfeeding and Employment Research Study

You have been selected to participate in a study conducted by
researchers at Michigan State University.

Instructions for filling out the survey:

,/ Fill in the box (I) to indicate to what extend you agree or disagree with
each statement.

/ To ensure clarity, please do not use a \I or an X.

/ If you need to change an answer, please make sure that your old answer
is either completely erased or clearly crossed out.

/ Please use a pen (black or blue ink) or a pencil.

112

 

For each of the following statements, breastfeeding includes breastfeeding a
baby and/or using a breast pump. Please think about each statement as it most
applies to you, whether you are pregnant and thought about breastfeeding or if
you are currently breastfeeding and/or pumping breast milk.

This section asks about the overall support you feel would be
provided by your company if you wanted to continue breastfeeding
or pumping breast milk after returning to work.

Strongly Agree Disagree Strongly
agree disagree
V V V V
1. I would have enough maternity
leave (paid and/or unpaid time
off) to get breastfeeding started '3 D D D
before going back to work.

2. I would be able to get
information about combining
work and breastfeeding from D D U U
my company.

3. I’m certain my company has
written policies for employees
that are breastfeeding or
pumping breast milk. D U D D

4. I’m certain there is a place I
could go to breastfeed or pump
breast milk at work. D D D D

5. There is someone I could go to
at work that would help me

113

10.

ll.

12.

13.

make arrangements for
breastfeeding or pumping
breast milk.

I would be able to continue
breastfeeding or pumping
breast milk after going back to
work.

My job could be at risk (e.g.
lose my job or get fewer
scheduled hours) if I breastfed

or pumped breast milk at work.

I would be able to talk about
breastfeeding at work.

I would feel comfortable
asking for accommodations to
help me breastfeed or pump
breast milk at work.

My opportunities for job
advancement would be limited
if I breastfed or pumped breast
milk at work.

I’m certain that women in
higher-level positions have
breastfed or pumped breast
milk at my workplace.

I’m certain co—workers have
breastfed or pumped breast
milk at my workplace.

My company would accept me
pumping breast milk at work,
but would disapprove of me

breastfeeding my child at work.

This section asks about the pace of your job and available time

114

D

Strongly

agree

E]

El

V

[:1

Agree

Cl

C]

C]

Disagree Strongly
disagree

I]

Y

Cl

[3

V

during your workday to breastfeed or pump breast milk.

My breaks are frequent enough
for breastfeeding or pumping
breast milk. '3 D D U

My breaks are long enough for

breastfeeding or pumping breast
milk, El III CI 1::

I could adjust my work hours in
order to breastfeed or pump

breast milk. D U D D
Strongly Agree Disagree Strongly
agree disagree

V V V

17.

18.

19.

20.

21.

22.

23.

I could adjust my break schedule

in order to breastfeed or pump
breast milk. D

I would have time to breastfeed

or pump breast milk, which is i_n

addition to regular breaks and

mealtimes. '3

Some days I would need to skip a
breastfeeding or pumping session
because my workdays are so

hectic. D

My job included travel or time

away from my company, making

it difficult to breastfeed or pump

breast milk. D

My job allows me to make my
own decisions about how to
schedule my work. D

My job allows me to decide on
the order in which things are
done on the job. D

My job allows me to plan how I
do my work. D

115

This section asks about the overall support you feel would be provided by
your direct manager/supervisor if you wanted to continue breastfeeding or
pumping breast milk after returning to work.

24.

25.

26.

27.

28.

29.

My manager would
support me breastfeeding
or pumping breast milk at
work.

My manager would help
me combine breastfeeding
and work.

My manager would think I
couldn’t get all my work
done if I needed to take
breaks for breastfeeding or
pumping breast milk.

I would feel comfortable
speaking with my manager
about breastfeeding.

My manager would be a
source of information on
how to breastfeed at work.

My manager says things
that make me think he/ she
supports breastfeeding.

Strongly

agree

116

V

Agree

V

Disagree
V

Strongly

disagree
V

30.

31.

32.

33.

34.

35.

36.

37.

My manager would
consider breastfeeding at
work a personal choice,

not something he/ she I3
should have to deal with.

My manager would

consider it part of his/her

job to help me combine CI

breastfeeding and work.

My manager would think
less of workers who
choose to breastfeed or

pump breast milk at work. D

My manager would make sure
my job is covered if I needed
time for breastfeeding or
pumping breast milk.

My manager would change my
work schedule to allow me time
for breastfeeding or pumping
breast milk.

My manager would help me deal
with my workload so I could
breastfeed or pump breast milk at
work.

My manager would be
embarrassed if I spoke with
him/her about breastfeeding.

I]

Strongly Agree

agree
v

:1 u

r: :1

a El

[3 [:1

El [:1
Cl E]
El E1

Disagree Strongly

disagree
V V
C1 CI
[:1 E1
[:1 El
[:1 E]

This section asks about the overall support you feel would be
provided by your co-workers if you wanted to continue breastfeeding
or pumping breast milk after returning to work.

I am certain my co-workers
would support me breastfeeding
or pumping breast milk at work.

117

Cl El

E1 El

38.

39.

40.

41.

42.

43.

44.

45.

My co-workers would think less
of workers that choose to
breastfeed or pump breast milk at
work.

I would feel comfortable
speaking with my co-workers
about breastfeeding.

My co-workers say things that
make me think they support
breastfeeding.

My co-workers would change
their break times with me so that

I could breastfeed or pump breast

milk.

Strongly
agree

My co-workers would
think they are
inconvenienced if I took
time to breastfeed or pump
breast milk.

My co-workers would
cover my job duties if I
needed time for
breastfeeding or pumping
breast milk.

My co-workers would be a
source of information on
how to breastfeed at work.

My co-workers would be
embarrassed ifI spoke
with them about
breastfeeding.

Cl

V

Cl

Cl

Agree

El

Disagree
V

[I

Strongly

disagree
V

NOTE: PLEASE ANSWER QUESTIONS 46 - 49 as YES or NO

Yes

118

No

46.

47.

48.

49.

I could buy or borrow the
equipment I would need
for pumping breast milk.

My company would
supply the equipment I
would need for pumping
breast milk at work.

I could ﬁnd a place to
store expressed breast
milk at work.

There is a company-
designated place for
women to breastfeed or
pump milk during the
workday.

1:1 El
D I]
E] 1:1
1:1 1:1

I

 

 

If yes, please
go to
question 50

 

119

 

50.

51.

52.

53.

54.

If you answered YES to QUESTION 49.
please answer the following Questions.

This section asks about the physical environment of your workplace
for breastfeeding or pumping breast milk after returning to work.

The designated place for
breastfeeding or pumping breast
milk at work would be available
when I needed it.

The designated place for
breastfeeding or pumping breast
milk is close enough to my work
area to use during my breaks.

I would feel comfortable
breastfeeding or pumping breast
milk in the designated place.

The designated place for
breastfeeding or pumping breast
milk is satisfactory.

The designated place for

breastfeeding or pumping breast
milk includes everything I need.

120

Strongly Agree Disagree Strongly

agree disagree
V V V V

E] C1 Cl [:1

El El El 1:]

El El 1:1 1:1

El 1:1 1:1 E!

1:1 1:1 E] El

APPENDIX B

121

CODING SCHEME FOR VARIABLES

 

 

Demographic Description Coding
Variable

Name _

Q1-6, 8, 9, Normal scoring 1=Strongly agree, 2=agree,

ll, 12, 14-18,
21-25, 27-29,

3=disagree, 4=strongly
disagree, .=missing data

 

 

 

 

 

 

 

 

 

 

 

 

 

31, 33-35, 37,
39-41, 43, 44
Q7-R, 10-R, Reverse scoring - To score 1=Strongly agree, 2=agree,
l3-R, l9-R, this instrument, you must ﬁrst 3=disagree, 4=strongly
20-R, 26-R, reverse score the items for disagree, .=missing data
30-R, 32-R, which an (R) is shown after
36-R, 38-R, them.
42-R, 45-R
Q46-49 Yes/No questions 0=no, l=yes, .=missing data,
10=yes and no
Q 50-54 Normal scoring 1=Strongly agree, 2=agree,
3=disagree, 4=strongly
disagree, .=missing data,
97=Q49 is no so skipped
EverBf Did the respondent ever 0= no, 1= yes, .=missing data,
breastfeed? 97=Q49 is no so skipped
Nume How many children has the 0=1, 1= 2 or more, .=missing
respondent breastfed? data, 98=Q78 is no so
skipped
BfDur What is the longest duration a 0= <1 week, 1=1 wk-6 wks,
respondent has ever bf‘? 2=>6 wks- 3 mo, 3=>3 mo —-
6 mo, 4=>6 mo, 98=Q78 is
no so skipped, .=missing data
CurBF Is the respondent currently 0=no, l-j'es, 98=Q78 is no so
breastfeeding? skipped, .=misggdata
MilkSup How does the respondent 0=breastfeeding child,
maintain milk supply during l=express breast milk,
work hours? 10=breastfeed and express,
98=Q78 is no so skipped,
99=Q81 is no so skipped,
.=missing data
BFCom How committed to 0=uncertain, 1=somewhat
breastfeeding is the committed, 2=committed,
respondent? 3=very committed, 98=Q78 is
no so skipped, 99=QSl is no
so skipped, .=missing data
EmpLepg How long has respondent 0=< 6 mo, 1=6-12 mo, 2=>1

 

122

 

 

worked for current company

yr-Syrs, 3=>5 yr, .=missing
data

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hourku How many hours a week the 0=<32hrs, l=32-39 hrs, 2=40-
respondent works. 50 hrs, 3=>50hrs, .=missing

data

Hr_Sal Does the respondent get paid 0=hourly, 1=Salary,
hourly or salary? .=missing data

Union Is the respondent a member of 0=no, l=yes, .=missing data
a labor union?

ManGen What is the gender of the 0=male, 1=female, .=missing
respondents manager? data

CoGen What is the gender of the 0=mostly male, 1=equal
majority of the respondents number of male and female,
coworkers? 2=mostly female, .=missing

data

LeaveTot How many total weeks of 0=<6 wks, l=6-12wks,
leave did the respondent take? 2=>12 wks, 96=pregnant so

skipped, .=missing data

PdL Did the respondent take paid 0=no, l—j'es, 96=pregnant so
leave? skipped, .=missing data

PdLTot How many weeks of paid 0=1-3 wks, 1=4-6wks, 2=7-9
leave did the respondent take? wks, 3=10-12 wks,

4=>12wks, 96=pregnant so
skipped, 97=Q9l is no so
skipped, .=missirg data

JobTitIe Respondent is asks to write in Write in, .=missing data
job title

Wka What type of industry does the 0=Communication,
respondent work in? 1=Finance, 2=Health Care,

3=Insurance,
4=Manufacturing — Salary,
Public Administration,
5=Public Administration,
6=Public Education, 7=Real
Estate, 8=Utilities, 400=Other
(write in), .=missing data

EmpNum How many employees work at Write in, .=missing data
the same physical location as
the repondent?

EmpTot How many employees work at 0=<250, 1=250-499, 2=500-
the respondents company at all 1000, 3=>1000, .=missing
locations? data

Ethnipo Is the respondent Hispanic or 0=non-Hispanic or non-
Latina? Latina, 1=Hispanic or Latina,

.=missinidata
Raciale What racial groups do the 0= American Indian/Alaskan

 

123

 

 

respondent identify with

Native, l=Asian, 2=Native
Hawaiian or Other Paciﬁc
Islander, 3=Black or Aﬁican
American, 4=White,
5=Multiple ethnicities,
.=missing data

 

 

 

 

 

 

income of the respondent?

 

_A_ge Age of respondent Write in, .=missing data

MarStat Is the respondent married or 0=no, l=yes, .=missing data
living with partner?

Ed What is the highest level of 0=< hs, 1=hs/GED, 2=Some
education the respondent has college/ junior or technical
completed? school, 3=College,

=Graduate, .=missing data

Income What is the yearly household 0=<20,000, l=20-29,000,

2=30—39,000, 3=40-49,000,
4=50-59,000, 5=>60,000,
.=missing data

 

124

 

Factors
1
2-neg vs. pos
2-C

3-A

BAYESIAN INFORMATION CRITERIA

Deviance

8709.75

8652.74

8602.31

8614.79

125

Parameters

51

53

53

56

BIC
881
.61
875
.64

8709
.21

8727
.74

OUTPUT FOR DIMENSION 1

SUMMARY OF 104 MEASURED PERSONS

+ ----------------------------------------------------------------------------- +
| RAW MODEL INFIT OUTFIT |
| SCORE COUNT MEASURE ERROR MNSQ ZSTD MNSQ ZSTD |
l ----------------------------------------------------------------------------- |
I MEAN 54 8 21.4 93 .37 1 01 - 2 1.05 .0 |
| S D 13.1 2.2 1 20 O9 52 1 6 .81 1.3 |
| MAX 87.0 25 0 5.06 1.03 2 56 3 6 7 01 3.8 |
| MIN 30.0 19 0 —1.47 28 .25 -3.5 .24 -2.7 |
I ----------------------------------------------------------------------------- I
| REAL RMSE .42 ADJ.SD 1.12 SEPARATION 2.66 PIRSON RILIABILITY .88 I
IMODEL RMSE .38 ADJ.SD 1.13 SEPARATION 2.96 PERSON RELIABILITY .90 I
I S.E. OF PERSON MEAN = .12 I
+ ————————————————————————————————————————————————————————————————————————————— +
VALID RESPONSES: 85.6%

PERSON RAW SCORE-TO-MEASURE CORRELATION = .83 (approximate due to missing data)
CRONBACH ALPHA (KR-20) PERSON RAW SCORE RELIABILITY = .95 (approximate due to

missing data)

SUMMARY OF 25 MEASURED ITEMS

+ ----------------------------------------------------------------------------- +
| RAW MODEL INFIT OUTFIT |
I SCORE COUNT MEASURE ERROR MNSQ ZSTD MNSQ ZSTD I
I ————————————————————————————————————————————————————————————————————————————— I
I MEAN 228.1 89.0 00 20 .98 - l 1.04 1 |
I S D 118.4 29.5 1 20 07 .18 1 2 .26 1 2 I
I MAX 378.0 104.0 3 46 .36 1.24 1 8 1 75 2 2 I
I MIN 12.0 30.0 -1 91 14 .54 -3 4 .49 -2 7 I
I ————————————————————————————————————————————————————————————————————————————— I
| REAL RMSE .22 ADJ.SD 1.18 SEPARATION 5.29 ITEM RELIABILITY .97 I
IMODEL RMSE .22 ADJ.SD 1.18 SEPARATION 5.43 ITEM RELIABILITY .97 I
l S.E. OF ITEM MEAN = .24 |
+ ————————————————————————————————————————————————————————————————————————————— +
DELETED: 29 ITEMS
UMEAN=.OOO USCALE=1.000
ITEM RAW SCORE-TO-MEASURE CORRELATION = -.42 (approximate due to missing data)

2226 DATA POINTS. APPROXIMATE LOG-LIKELIHOOD CHI-SQUARE: 3636.00

12(3

SUMMARY OF CATEGORY STRUCTURE. Model="R"
FOR GROUPING "1" ITEM NUMBERS: l-l2 14-15 17-18 50-54

4.. __________________________________________________________________

ICATEGORY

OBSERVEDIOBSVD SAMPLEIINFIT OUTFITIISTRUCTUREICATEGORYI
ILABEL SCORE COUNT %IAVRGE EXPECTI

MNSQ MNSQIICALIBRATNI MEASUREI
I ——————————————————— + ———————————— + ------------ ++ --------- + ———————— +
I 1 1 139 GI -.75 -.94I 1.24 1.31II NONE I( -2.65)I 1
I 2 2 335 151 ‘.13 ~.01| .86 .87Il -1.36 I -.87 | 2
I 3 3 679 31I .95 .98I .94 .91II -.23 I .78 | 3
l 4 4 658 30I 2.27 2.22I .93 .95II 1.59 I( 2.80)I 4
| ------------------- + ———————————— + ------------ ++ --------- + ———————— +
IMISSING 373 l7I .79 I II I I

OBSERVED AVERAGE

is mean of measures in category.

It is not a parameter estimate.

+ __________________________________________________________________________

[CATEGORY STRUCTURE | SCORE—TO—MEASURE I 50% CUM.I COHERENCEIESTIMI
I LABEL MEASURE S.E. I AT CAT. ---—ZONE-—--IPROBABLTY| M—>C C->MIDISCR|
I ———————————————————————— + ————————————————————— + ————————— + ---------- + ----- +
I 1 NONE |( -2.65) -INF —1.85| | 56% 15%| I
I 2 -1.36 .10 | -.87 -1.85 -.08| -1.59 | 46% 47%| .78I
I 3 —.23 .07 I .78 -.O8 1.91I -.14 | 52% 70%| 1.05I
I 4 1.59 06 |( 2.80) 1.91 +INF | 1.73 | 78% 60%| 1.13I
+ __________________________________________________________________________
M—>C = Does Measure imply Category?

C->M = Does Category imply Measure?

CATEGORY PROBABILITIES: MODES - Structure measures at intersections

9 ++ --------- + --------- + ————————— + --------- + --------- + --------- ++
R 1.0 + +
O I l
B I l
A I1 I
B 8 + 111 44+
I | 11 44 I
L | 11 444 I
I I 11 44 I
T .6 + 11 44 +
Y | 11 3333333 4 |
.5 + 1 333 3333 44 +

O I 112222222222 333 4*3 |
F .4 + 222211 **2 44 333 +
I 22 1 33 22 4 33 I

R I 222 11 33 222 44 333 I
E | 222 3*1 22 44 333 I
S .2 + 222 33 11 44*2 33+
P I22 333 11 44 222 |
O I 3333 1***4 2222 I
N I 333333 444444 111111 22222222 |
S .O +******44444444444444444 1111111111111111*******+
E ++ --------- + --------- + ————————— + --------- + --------- + --------- ++

-3 -2 -1 0 1 2 3
PERSON [MINUS] ITEM MEASURE

127

nwaH

IENTRY RAW
MATCHIESTIMI

INUMBER SCORE COUNT MEASURE S.E.

 

 

 

 

EXP%|DISCR| ITEM
-+ _____ I
I 1 378
94I 10001 1
I 2 274
79I 10002 1
I 3 220
80I 10003 1
I 4 337
.91I 10004 1
I 5 320
1.31I 10005
I 6 366
1.4:I 10006
I 7 364
1.16] 10007
I 8 323
1.2cI 10008
I 9 311
1.34I 10009
I 10 334
1.0cI 10010
I 11 276
.73I 10011 1
I 12 339
90] 10012 1
I 14 301
.86I 10014 1
I 15 297
1.01] 10015
I 17 317
1.16] 10017
I 18 230
69I 10018 1
I 46 88
.88I 10046 0
I 47 12
.94] 10047 0
I 48 95
98I 10048 0
I 49 29
.70I 10049 0
I 50 96
1 3cI 10050
51 102
87I 10051 1
I 52 104
1.19! 10052
I 53 99
1.ocI 10053
I 54 90
1.0CI 10054
-+ ------------ I
I MEAN 228.1
I I
I 3.0 118.4
I

104

104

104

103

1.61

-1.35

-1.29

.60

.18

-1.19

3.46

-1.91

2.10

.11

MODELI

ITEM STATISTICS: ENTRY ORDER

INFIT | OUTFIT IPTBSEIEXACT
[MNSQ ZSTDIMNSQ ZSTDICORR.I 088%

19] 99 .0|1.10 .5] .36] 74.0 71.2]
.14I1.20 1.5I1.16 1.2] 54] 48.1 49.7]
.14I1.04 .3]1.08 .7] .49] 54.4 50.0]
.15]1.11 .8|1.17 1.0] .58] 56.7 57.1]
.15] .81 -1.4l .76 -1.7| .72] 55.8 54.0]
.181 .54 -3.41 .49 -2.7I .69l 81.7 65.61
.17] .78 -1.5I .87 -.SI .54] 73.1 65.1]
.15] .87 -l.0I .80 -l.3| .54] 62.5 54.5]
.14] .76 -1.9| .72 -2.0I .65] 62.5 52.7]
.15] .89 -.7|1.12 .7] .37] 66.0 57.2]
.14I1.16 1.2I1.18 1.3] .53] 47.6 50.21
.15I1.20 1.4I1.05 .4] .56] 56.7 57.4l
.14I1.14 l.1|1.10 .7] .46] 49.0 52.1]
.14] .94 -.4| .92 -.6I .51] 54.8 52.0]
.14] .91 -.7| .87 -.8| .58] 55.8 53.7]
.14I1.24 1.811.31 2.2] .481 40.4 49.2I
.29I1.07 .5]1.29 .8] .20] 84.6 84.8]
.34I1.06 .4|1.18 .5] .27] 87.5 89.2]
.36Il.01 .1l1.75 1.3] .24] 91.3 91.3]
.25I1.23 1.7l1.14 .7] .36] 68.9 77.5]
.28] .69 -1.2| .71 -.9| .49] 63.3 54.3]
.30] .99 .1I1.43 1.2] .33] 53.3 60.2]
.31] .76 -.8| .81 -.4| 48] 66.7 62.0]
.29I1.10 .5] .98 .1] 50] 56.7 56.3]
.26I1.07 4] .99 .1] .62] 43.3 52.4]
.20] 98 - 1|1.04 .1] I 62.2 60.8]
07] 18 1 2] 26 1.2] I 13.4 12.3]

128

OUTPUT FOR DIMENSION 2

SUMMARY OF 104 MEASURED PERSONS

 

+ -------------------------------------------------------------- +
| RAW MODEL INFIT OUTFIT |
I SCORE COUNT MEASURE ERROR MNSQ ZSTD MNSQ ZSTD |
I --------------------------------------------------------------------------- I
I MEAN 57.9 19 9 1 00 .38 1 07 -.1 1.00 -.2 |
I S D 10.7 6 1 44 08 69 1.9 .59 1.8 I
I MAX 78.0 20 0 4 7O .76 4 02 5.9 3.77 5.7 I
I MIN 30.0 14 0 -2 43 .33 26 -3.4 .25 -3.5 I
| ----------------------------------------------------------------------------- I
I REAL RMSE .45 ADJ.SD 1.37 SEPARATION 3.03 PERSON RELIABILITY .90 I
IMODEL RMSE .39 ADJ.SD 1.39 SEPARATION 3.58 PERSON RELIABILITY .93 I
I S.E. OF PERSON MEAN = .14 I
+ ————————————————————————————————————————————————————————————————————————————— +
VALID RESPONSES: 99.6%

PERSON RAW SCORE—TO—MEASURE CORRELATION = .97 (approximate due to missing data)
CRONBACH ALPHA (KR—20) PERSON RAW SCORE RELIABILITY = .93 (approximate due to

missing data)

SUMMARY OF 20 MEASURED ITEMS

+ ————————————————————————————————————————————————————————————————————————————— +
| RAW MODEL INFIT OUTFIT I
I SCORE COUNT MEASURE ERROR MNSQ ZSTD MNSQ ZSTD l
| ----------------------------------------------------------------------------- |
I MEAN 301 2 103 6 .00 .16 99 - 1 .99 .0 l
| 8.0. 34 8 6 85 .01 23 1 6 .27 1.7 I
I MAX. 352 0 104 0 1.74 .18 1 57 3 8 1 77 4.9 |
I MIN. 228 0 102 0 -1.32 .15 6O —3 2 .58 —2.7 I
I ----------------------------------------------------------------------------- l
I REAL RMSE .17 ADJ.SD .84 SEPARATION 5.03 ITEM RELIABILITY .96 I
IMODEL RMSE .16 ADJ.SD .84 SEPARATION 5.21 ITEM RELIABILITY .96 I
I S.E. OF ITEM MEAN = .20 I
+ ----------------------------------------------------------------------------- +
DELETED: 34 ITEMS
UMEAN=.000 USCALE=1.000
ITEM RAW SCORE—TO-MEASURE CORRELATION = -1.00 (approximate due to missing data)

2071 DATA POINTS. APPROXIMATE LOG-LIKELIHOOD CHI-SQUARE: 3683.55

129

SUMMARY OF CATEGORY STRUCTURE. Model="R"

+ ——————————————————————————————————————————————————————————————————
[CATEGORY OBSERVEDIOBSVD SAMPLEIINFIT OUTFIT]ISTRUCTUREICATEGORYI
ILABEL SCORE COUNT %IAVRGE EXPECTI MNSQ MNSQIICALIBRATNI MEASURE]

I ------------------- + ------------ + ------------ ++ --------- + ———————— +
I 1 1 144 7| -.9O -1.29| 1.45 1.62II NONE I( -3.15)I 1
I 2 2 472 23] -.33 -.20] .88 .87ll -1.93 l -1.12 I 2
| 3 3 885 43I .93 .99] .92 .81II -.25 | 1.01 | 3
I 4 4 570 27] 2.69 2.59] .90 .92Il 2.18 |( 3.34)] 4
I ------------------- + ———————————— + ------------ ++ ————————— + -------- +

IMISSING 9 0] 63 I II I I

OBSERVED AVERAGE is mean of measures in category. It is not a parameter estimate.

, __________________________________________________________________________
ICATEGORY STRUCTURE I SCORE—TO—MEASURE I 50% CUM.I COHERENCEIESTIMI
I LABEL MEASURE S.E. I AT CAT. ----ZONE----IPROBABLTYI M->C C->MIDISCRI
I ———————————————————————— + --------------------- + --------- + ---------- + ----- +
I 1 NONE II -3.15) -INF -2.27| I 57% 13%| I 1
I 2 -1 93 .10 I -1.12 -2.27 -.12I -2.08 I 52% 54%| .69I 2
I 3 -.25 06 I 1.01 -.12 2.38| —.18 I 59% 76%| 1.01I 3
I 4 2.18 .06 II 3.34) 2.38 +INF I 2.26 I 79% 58%| 1.16] 4
+ __________________________________________________________________________

M—>C = Does Measure imply Category?
C->M = Does Category imply Measure?

CATEGORY PROBABILITIES: MODES - Structure measures at intersections

P ++ ------ + ------ + —————— + ------ + —————— + —————— + ------ + —————— ++
R 1.0 + +
O l I
B II I
A | 111 44I
B 8 + 11 44 +
I I 11 44 I
L I 1 44 |
I I 11 4 I
T .6 + 1 333333333 44 +
Y I 1 222 33 33 4 I
.5 + 11 222 222 3 334 +

O l 2* 2*3 443 I
F .4 + 22 1 3 22 4 33 +
| 22 1 3 22 4 33 l

R | 22 11 33 2 44 3 I
E | 22 13 22 4 33 I
S .2 + 22 3311 2*4 333 +
P | 222 33 11 44 22 33I
0 I2 333 11 444 222 I
N | 33333 44***11 22222 I
S .O +*********4444444444444444 11111111111111***********+
E ++ —————— + ------ + —————— + ------ + —————— + ------ + ------ + ------ ++
-4 -3 -2 -1 0 1 2 3 4

PERSON [MINUS] ITEM MEASURE

130

IENTRY RAW
MATCHIESTIM]

INUMBER SCORE COUNT MEASURE S.E.
ITEM

 

 

 

 

EXP%|DISCRI

_+ _________ I

| 24 339
1.27I 10024

I 25 312
1.21I 10025

I 26 309
.88] 10026 I
I 27 305
1.2cI 10027

I 29 293
1.05I 10029

I 30 235
12I 10030 I
I 31 228
.98] 10031 I
I 32 334
1 14I 10032

I 33 267
.95I 10033 I
I 34 275
1 0:I 10034

I 35 259
.90I 10035 I
I 36 321
1.0:I 10036

I 37 343

1.43I 10037

I 38 352

1.2:I 10038

I 39 329
1.12] 10039

I 40 330
1.01I 10040

I 41 297
99I 10041 I
I 43 293
.78] 10043 I
I 44 272
.39] 10044 I
I 45 330

1 1 I 1~045

_+ _________ I

I MEAN 301.2
I I

| S.D 34.8
I I

----------- +

103

103

102

103

104

104

103

104

104

104

104

104

104

104

104

104

104

103

.19

.09

.12

.54

.74

.78

.83

.69

.05

.42

.04

.64

.67

.18

.27

.76

MODEL]

ITEM STATISTICS: ENTRY ORDER

INFIT I OUTFIT IPTBSEIEXACT
IMNSQ ZSTDIMNSQ ZSTDICORR.] OBS%

17] .81 -l.3l .73 ~1.6l .68] 73.8 63.4]
.16] 78 -1.6| .74 -1.8| .72] 67.0 59.0]
.16I1.13 9|1.20 1.3] .57] 63.1 58.3]
.16] .85 -1.1] .83 -1.2| .73] 68.0 58.2]
.16] .99 .0] .98 -.1] .65] 55.9 56.8]
.15I1.57 3.8]1.77 4.9] .35] 53.4 53.6]

15] .98 —.1I1.00 .Ol .60] 55.8 53.1]
.17] .90 -.7| .89 -.6| .58] 70.2 62.1]
.15I1.05 .4I1.06 .5] .61] 55.3 55.5]
.15] .97 —.2| .97 -.2| .56] 57.7 55.5]
.15I1.07 .5|1.08 .6] .58] 58.7 55.3]
.16] .96 -.2]1.02 .2] .59] 64.4 60.0]
.17] .60 -3.2| .58 -2.7l .70I 79.8 63.5]
.18] .77 ~1.7| .73 -l.4] .59] 71.2 65.1]
.17] .91 -.6| .88 -.7I .65] 61.5 61.3]
.17] .93 -.4| .87 -.8| .62] 62.5 61.7]
.15I1.06 .4I1.02 .2] .60] 56.7 56.8]
.15]1.20 1.5I1.20 1.4] .53] 56.7 56.2]
.15I1.51 3.4l1.53 3.5] .55] 41.3 55.4]

17] 83 -1.2I .84 -.9I .61] 67.0 62.0]
.16] 99 -.ll 99 .0] I 62.0 58.6]
.01] .23 1.6] .27 1.7] I 8.4 3.5]

131

APPENDIX C

132

FINAL VERSION OF THE EPBS-Q

For each of the following statements, breastfeeding includes breastfeeding a
baby and/or using a breast pump. Please think about each statement as it most
applies to you, whether you are pregnant and thought about breastfeeding or if
you are currently breastfeeding and/or pumping breast milk.

This section asks about the overall support you feel would be
provided by your company if you wanted to continue breastfeeding
or pumping breast milk after returning to work.

Strongly Agree Disagree Strongly
agree disagree
V V V V
1. I would have enough maternity
leave (paid and/or unpaid time
oﬂ) to get breastfeeding started D D D '3
before going back to work.

2. I would be able to get
information about combining
work and breastfeeding from my D D U U
company.

3. I’m certain my company has
written policies for employees
that are breastfeeding or
pumping breast milk. '3 D U ‘ D

4. I’m certain there is a place I
could go to breastfeed or pump
breast milk at work. U U U U

5. There is someone I could go to
at work that would help me
make arrangements for

breastfeeding or pumping breast
milk. D Cl C] III

6. My job could be at risk (e.g. lose
my job or get fewer scheduled

hours) ifI breastfed or pumped
breast milk at work. El E] E] El

7. I would be able to talk about
breastfeeding at work. [3 D Cl C]

133

10.

11.

12.

13.

14.

I would feel comfortable
asking for
accommodations to help
me breastfeed or pump
breast milk at work.

My opportunities for job
advancement would be
limited if I breastfed or
pumped breast milk at
work.

I’m certain that women in
higher-level positions have
breastfed or pumped breast
milk at my workplace.

I’m certain co-workers
have breastfed or pumped
breast milk at my
workplace.

Strongly
agree

Cl

C]

V

Agree

Cl

Disagree

E]

Strongly
disagree

[I

This section asks about the pace of your job and available time
during your workday to breastfeed or pump breast milk.

My breaks are frequent
enough for breastfeeding
or pumping breast milk.

My breaks are long
enough for breastfeeding
or pumping breast milk.

I could adjust my break
schedule in order to
breastfeed or pump breast
milk.

134

El

15.

16.

17.

18.

19.

20.

21.

This section asks about the overall support you feel would be
provided by your direct manager/supervisor if you wanted to
continue breastfeeding or pumping breast milk after returning to

My manager would
support me breastfeeding
or pumping breast milk at
work.

My manager would help
me combine breastfeeding
and work.

My manager would think I
couldn’t get all my work
done if I needed to take
breaks for breastfeeding or
pumping breast milk.

I would feel comfortable
speaking with my manager
about breastfeeding.

My manager says things
that make me think he/ she
supports breastfeeding.

My manager would
consider breastfeeding at
work a personal choice,
not something he/ she
should have to deal with.

My manager would
consider it part of his/her
job to help me combine

work.
Strongly Agree
agree
V

135

Disagree
V

Strongly

disagree
V

22.

23.

24.

25.

26.

27.

28.

29.

breastfeeding and work.

My manager would think
less of workers who
choose to breastfeed or

pump breast milk at work. D

My manager would make sure
my job is covered if I needed
time for breastfeeding or
pumping breast milk.

My manager would change my
work schedule to allow me time
for breastfeeding or pumping
breast milk.

My manager would help me deal
with my workload so I could
breastfeed or pump breast milk at
work.

My manager would be
embarrassed if I spoke with
him/her about breastfeeding.

E] Cl C]
Strongly Agree Disagree Strongly
agree disagree
V ' V V

[I Cl [3 E]

E] Cl E] El

[3 E] Cl C]

D E] E] El

This section asks about the overall support you feel would be
provided by your co-workers if you wanted to continue breastfeeding
or pumping breast milk after returning to work.

My co-workers would think less
of workers that choose to
breastfeed or pump breast milk at
work.

I would feel comfortable
speaking with my co-workers
about breastfeeding.

My co-workers say things that

136

E] E] El E1

30.

31.

32.

33.

34.

35.

36.

make me think they support
breastfeeding. D D D D

My co-workers would change
their break times with me so that

I could breastfeed or pump breast
milk. I] D E] El

Strongly Agree Disagree Strongly
agree disagree
V V V V
My co-workers would
cover my job duties if I
needed time for
breastfeeding or pumping U D D U
breast milk.

My co-workers would be

embarrassed if I spoke
with them about U D D D
breastfeeding.

NOTE: PLEASE ANSWER QUESTIONS 33 - 36 as YES or NO

Yes No
I could buy or borrow the
equipment I would need
for pumping breast milk. D D
My company would
supply the equipment I
would need for pumping D D
breast milk at work.
I could ﬁnd a place to
store expressed breast
milk at work. D U
There is a company-
designated place for
women to breastfeed or U '3
pump milk during the I
workday. If yes, please continue to
number 37 on the next
page.

137

37.

38.

39.

40.

41.

If on answered YES to UESTION 36 lease answer the followin
Questions.

If you answered NO to question 36, please SKIP this set of questions
and go to question 42.

This section asks about the physical environment of your workplace
for breastfeeding or pumping breast milk after returning to work.

Strongly Agree Disagree Strongly
agree disagree
V V V V

The designated place for

breastfeeding or pumping breast

milk at work would be available

when I needed it. D D D U

The designated place for

breastfeeding or pumping breast

milk is close enough to my work

area to use during my breaks. D D D U

I would feel comfortable
breastfeeding or pumping breast
milk in the designated place. D D U U

The designated place for
breastfeeding or pumping breast
milk is satisfactory. '3 D D U

The designated place for

breastfeeding or pumping breast
milk includes everything I need. U E] U U

138

Please answer the following questions by filling in the box next to your answer.

42) Have you ever breastfed a child?

E] No —' (Skip to question 45)
D Yes

43) How many children have you breastfed?

D 1 child
El 2 or more children

44) What is the longest you have ever breastfed a child?

Less than l-week

l — 6 weeks

More than 6 weeks - 3 months
More than 3 months - 6 months
More than 6 months

DECIDE

45) How long have you worked for your current employer?

Less than 6 months
6-12 months

13 months — 5 years
More than 5 years

CIDCJCI

46) How many hours do you usually work each week?

[1 Less than 20 hours/week
E] 20 - 3O hours/week

D 31 — 4O hours/week

[3 More than 40 hours/week

47) Are you paid hourly or salary?

D Hourly
[3 Salary

139

48) Are you a member of a labor union?

D Yes
I] No

49) Is your immediate manager male or female?

[3 Male
[1 Female

50) What is the gender of the majority of your co-workers?

El Mostly male
El About an equal number of male and female
CI Mostly female

51) How many weeks of leave did you take before returning to work after your last
child?

[1 Less than 6 weeks
[3 6 — 12 weeks
[1 More than 12 weeks

52) Was any of the leave paid? (Includes leave that was paid at any percentage, e. g.
50%, 80%, and/or any paid sick, personal, or vacation time)

[I No —’ (Skip to question 54)
D Yes

53) How many weeks of paid leave did you take aﬂer your
last child?

1 — 3 weeks

4 — 6 weeks

7 — 9 weeks

10 — 12 weeks

More than 12 weeks

[DECIDE]

54) What is your job title? -

 

140

55) Which of the following are you. . .? Select one.

:1 Hispanic or Latino
D Non-Hispanic or non-Latino

56) What race do you consider yourself to be? Choose as many as apply.

American Indian/ Alaskan Native

Asian

Native Hawaiian or Other Paciﬁc Islander
Black or African American

White

Other

DECIDED

57) What is your age in years? years old.

 

58) Are you currently married or living with a partner?

Cl Yes
[3 No

59) What is the highest level of education you have completed?

[:1 Less than high school

B High School/GED

[:1 Some college/Junior college or technical school
degree

[I College degree

[3 Graduate degree

60) What is your yearly household income?

Less than $20,000/year
$20,000 - $30,000/year
$31,000 - $40,000/year
$41,000 - $50,000/year
$51,000 - $60,000/year
More than $60,000/year

[:JCICICICICI

141

BIBLIOGRAPHY

142

AAFP. (2005). Breastfeeding (Position Paper). Retrieved July 1, 2005, from
http://mvw.aal:p.org/x6633.xml

AAP. (2005). Breastfeeding and the Use of Human Milk. Pediatrics, 115(2), 496-506.

AbuSabha, R., & Greene, G. (1998). Body weight, body composition, and energy intake
changes in breastfeeding mothers. J Hum Lact, 14(2), 119-124.

ADA. (2005). Position of the American Dietetic Association: Promoting and Supporting
Breastfeeding. Journal of the American Dietetic Association, 105(5), 810-818.

Adams, R. J ., Wilson, M. R., & Wang, W. C. (1997). The multidimensional random
coefﬁcients multinomial logit model. Applied Psychological Measurement. Applied
Psychological Measurement, 21, 1-23.

Ahluwalia, I. B., Morrow, B., & Hsia, J. (2005). Why Do Women Stop Breastfeeding?
Findings From the Pregnancy Risk Assessment and Monitoring System. Pediatrics,
116(6), 1408-1412.

Allen, M. J ., & Yen, W. M. (2002). Introduction to Measurement Theory (2nd ed.).
Prospect Heights: Waveland Press, Inc.

Allen, T. D. (2001). F amily-Supportive work Environments: The Role of Organizational
Perceptions. Journal of Vocational Behavior, 58(3), 414.

American Academy of Pediatrics, 3. o. b. (2005). Breastfeeding and the Use of Human
Milk. Pediatrics, 115(2), 496-506.

Auerbach, K. G. (1984). Employed breastfeeding mothers: problems they encounter.
Birth, 11(1), 17-20.

Auerbach, K. G., & Guss, E. (1984). Maternal employment and breastfeeding. A study of
567 women's experiences. Am J Dis Child, 138(10), 958-960.

Bachrach, V. R., Schwarz, B., & Bachrach, L. R. (2003). Breastfeeding and the risk of
hospitalization for respiratory disease in infancy: a meta-analysis. Arch Pediatr Adolesc
Med, 157(3), 237-243.

143

Ball, T. M., & Wright, A. L. (1999). Health care costs of formula-feeding in the ﬁrst year
of life. Pediatrics, 103(4 Pt 2), 870-876.

Bamette, J. J. (1999). Effects of Stern and Likert Response Option Reversals on Survey
Internal Consistency: If You Feel the Need, There is a Better Alternative to Using Those
Negatively Worded Stems. Educational and Psychological Measurement, 60, 361 -3 70.

Baughman, R., DiNardi, D., & Holtz-Eakin, D. (2003). Productivity and wage effects of
"family-friendly" ﬁinge beneﬁts. International Journal of Manpower, 24(3), 247.

Benson, S. J. (2005). The relative contribution of formal and informal organizational
work-family support. Journal of Vocational Behavior, 66(3), 487.

Berger, L. M., Hill, J ., & Waldfogel, J. (2005). Maternity leave, early maternal
employment and child health and development in the US. The Economic Journal,
[15(501), F29-F47.

Breastfeeding Trends Through 2002. (2003). (pp. 1-9): Mothers Survey Ross Product
Division Abbott Laboratories.

Breastfeeding: Data and Statistics: Breastfeeding Practices — Results from the National
Immunization Survey. (2004). Retrieved September 7, 2007, from
http://www.cdc.gov/breastfeeding/data/NIS data/data_2004.htm

Breastfeeding: Data and Statistics: Breastfeeding Practices — Results from the National
Immunization Survey. (2005). Retrieved September 30, 2005. from
http://www.ch.gov/breastfeeding/data/NlS_data/index.htm.

Bridges, C. B., Frank, D. I., & Curtin, J. (1997). Employer attitudes toward breastfeeding
in the workplace. J Hum Lact, 13(3), 215-219.

Brown, C. A., Poag, S., & Kasprzycki, C. (2001). Exploring large employers' and small
employers' knowledge, attitudes, and practices on breastfeeding support in the workplace.
J Hum Lact, 1 7(1), 39-46.

Burke, M. E. (2005). 2005 Beneﬁts: Survey Report. Society for Human Resources
Management.

144

Cardenas, R. A., & Major, D. A. (2005). Combining Employment and Breastfeeding:
Utilizing a Work-Family Conﬂict Framework to Understand Obstacles and Solutions.
Journal of Business and Psychology, 20(1), 31 .

CDC. (2005). Breastfeeding: Data and Statistics: Breastfeeding Practices — Results from
the National Immunization Survey. Retrieved September 30, 2005, from
http://www.ch.gov/breastfeeding/data/NlS_data/index.htm

Chatterji, P., & Frick, K. D. (2003). Does Returning to Work After Childbirth Aﬂect
Breastfeeding Practices? Retrieved June 28, 2006. from
http://www.nber.org/papers/w9630.

Chen, A., & Rogan, W. J. (2004). Breastfeeding and the Risk of Postneonatal Death in
the United States. Pediatrics, 113(5), e435-439.

Clifton, T. J ., & Shepard, E. (2004). Work and family programs and productivity:
Estimates applying a production ﬁinction model. International Journal of Manpower,
25(7/8), 714.

Cohen, R., & Mrtek, M. B. (1994). The impact of two corporate lactation programs on
the incidence and duration of breast-feeding by employed mothers. Am J Health Promot,
8(6), 436-441.

Cohen, R., Mrtek, M. B., & Mrtek, R. G. (1995). Comparison of maternal absenteeism
and infant illness rates among breast—feeding and formula-feeding women in two
corporations. Am J Health Promot, I 0(2), 148-153.

Cohen, R., Mrtek, M. B., & Mrtek, R. G. (1995). Comparison of Maternal Absenteeism
and Infant Illness Rates Among Breast-feeding and Formula-feeding Women in Two
Corporations. American Journal of Health Promotion, 10(2), 148-153.

DeVellis, R. (1991). Scale Development (V 01. 26). London: Sage

Dewey, K. G., Heinig, M. J ., & Nommsen, L. A. (1993). Maternal weight-loss patterns
during prolonged lactation. Am J Clin Nutr, 58(2), 162-166.

Dewey, K. G., Heinig, M. J ., & Nommsen-Rivers, L. A. (1995). Differences in morbidity
between breast-fed and formula-fed infants. J Pediatr, 126(5 Pt 1), 696-702.

145

DHHS. (2000). Tracking Healthy People 2010. Washington, DC: US.
Government Printing Ofﬁce.

Dick, M. J., Evans, M. L., Arthurs, J. B., Barnes, J. K., Caldwell, R. S., Hutchins, S. S., et
al. (2002). Predicting early breastfeeding attrition. J Hum Lact, 18(1), 21-28.

DiGirolamo, A., Thompson, N., Martorell, R., F ein, S., & Grummer-Shawn, L. (2005).
Intention or Experience? Predictors of Continued Breastfeeding. Health Educ Behav,
32(2), 208-226.

Dillman, D. A. (2000). Mail and Internet Surveys - The Tailored Design Method: John
Wiley & Sonc, Inc.

Dillman, D. A. (2000). Mail and Internet Surveys - The Tailored Design Method (Second
ed.): John Wiley & Sons, Inc.

Duckett, L. (1992). Maternal employment and breastfeeding. NAA COGS Clin lssu
Perinat Womens Health Nurs, 3(4), 701-712.

Dunn, B. F., Zavela, K. J ., Cline, A. D., & Cost, P. A. (2004). Breastfeeding practices in
Colorado businesses. J Hum Lact, 20(2), 170-177.

Else-Quest, N. M., Hyde, J. S., & Clark, R. (2003). Breastfeeding, Bonding, and the
Mother-Infant Relationship. Merrill-Palmer Quarterly, 49(4), 495-517.

Evenhouse, B., & Reilly, S. (2005). Improved estimates of the beneﬁts of breastfeeding
using sibling comparisons to reduce selection bias. Health Serv Res, 40(6 Pt 1), 1781-
1802.

Fein, S. B., & Roe, B. (1998). The effect of work status on initiation and duration of
breast-feeding. Am J Public Health, 88(7), 1042-1046.

Fomon, S. (2001). Infant feeding in the 20th century: formula and beikost. J Nutr, 131(2),
409s-4203.

Fowler, F. J. J. (2002). Survey Research Methods (Third ed. Vol. 1): Sage Publications,
Inc.

146

Galinsky, B., Bond, J. T., & Friedman, D. E. (1993). Highlights: The national study of the
changing workforce. New York: Family and Work Insitute.

Galtry, J. (1997). Lactation and the labor market: breastfeeding, labor market changes,
and public policy in the United States. Health Care Women Int, 18(5), 467-480.

Gardner, L. M. (2002). A Step Towards True Equality In The Workplace: Requiring
Employer Accommodation For Breastfeeding Women. Wisconsin 's Women's Law
Journal, 1 7, 259-289.

Gault, B., & Lovell, V. (2006). The Costs and Beneﬁts of Policies to Advance Work/Life
Integration. The American Behavioral Scientist, 49(9), 1152.

Glanz, K. (2002). Health Behavior and Health Education. San Francisco: Jossey-Bass.

Global Strategy for Infant and Young Child Feeding. (2003). Geneva, Switzerland:
World Health Organization; 2003. .

Goodstein, J. D. (1994). Institutional Pressures and Strategic Responsiveness: Employer
Involvement in Work-Family Issues. Academy of Management Journal, 3 7(2), 350-382.

Goodstein, J. D. (1995). Employer Involvement in Eldercare: An Organizational
Adaptation Perspective. Academy of Management Journal, 38(6), 1657-1672.

Grurnmer-Strawn, L. M., & Mei, Z. (2004). Does Breastfeeding Protect Against Pediatric
Overweight? Analysis of Longitudinal Data From the Centers for Disease Control and
Prevention Pediatric Nutrition Surveillance System. Pediatrics, 113(2), e8l-86.

Healthy People 2010 midcourse review: revised objectives for 16-19. Retrieved June 1,
2006. from http://www.healthypeople.gov/data/midcourse/.

Heinig, M. J. (2001). Host defense beneﬁts of breastfeeding for the infant. Effect of
breastfeeding duration and exclusivity. Pediatr Clin North Am, 48(1), 105-123 ix.

Hills—Bonczyk, S. G., Avery, M. D., Savik, K., Potter, S., & Duckett, L. J. (1993).
Women's experiences with combining breast-feeding and employment. J Nurse
Midwifery, 38(5), 257-266.

147

Horwood, L. J ., & Fergusson, D. M. (1998). Breastfeeding and later cognitive and
academic outcomes. Pediatrics, 101(1), E9.

Jobe, J. B., & Mingay, D. J. (1989). Cognitive Research Improves Questionnaires.
American Journal of Public Health, 79(8), 1053-1055.

Johnston, M. L., & Esposito, N. (2007). Barriers and facilitators for breastfeeding among
working women in the United States. J Obstet Gynecol Neonatal Nurs, 36(1), 9-20.

Kac, G., Benicio, M. H., Velasquez Melendez, G., Valente, J. G., & Struchiner, C. J.
(2004). Breastfeeding and postpartum weight retention in a cohort of Brazilian women.
Am J Clin Nutr, 79(3), 487-493.

Katcher, A. L., & Lanese, M. G. (1985). Breast-feeding by employed mothers: a
reasonable accommodation in the work place. Pediatrics, 75(4), 644-647.

Kearney, M. H., & Cronenwett, L. (1991). Breastfeeding and employment. J Obstet
Gynecol Neonatal Nurs, 20(6), 471-480.

Kimbro, R. T. (2006). On-the-job moms: work and breastfeeding initiation and duration
for a sample of low-income women. Matem Child Health J, 10(1), 19-26.

Kopelman, R. B., Prottas, D. J ., Thompson, C. A., & Jahn, E. W. (2006). A Multilevel
Examination of Work-Life Practices: Is More Always Better? Journal of Managerial
Issues, 18(2), 232.

Kossek, E. B., Colquitt, J. A., & Noe, R. A. (2001). Caregiving decisions, well-being, and
performance: The effects of place and provider as a ﬁmction of dependent type and work-
family climates. Academy of Management Journal, 44(1), 29.

Kurinij, N., Shiono, P. H., Ezrine, S. F., & Rhoads, G. G. (1989). Does maternal
employment affect breast-feeding? Am J Public Health, 79(9), 1247-1250.

Labbok, M., & Krasovec, K. (1990). Toward consistency in breastfeeding deﬁnitions.
Stud Fam Plann, 21(4), 226-230.

Labbok, M. H. (2001). Effects of breastfeeding on the mother. Pediatr Clin North Am,
48(1), 143-158.

148

Li, R., Darling, N., Maurice, B., Barker, L., & Grummer-Strawn, L. M. (2005).
Breastfeeding rates in the United States by characteristics of the child, mother, or family:
the 2002 National Immunization Survey. Pediatrics, 115(1), e3 1-37.

Li, R., Zhao, Z., Mokdad, A., Barker, L., & Grummer Strawn, L. (2003). Prevalence of
breastfeeding in the United States: the 2001 National Immunization Survey. Pediatrics,
111(5 Part 2), 1198-1201.

Libbus, M. K., & Bullock, L. F. C. (2002). Breastfeeding and Employment: An
Assessment of Employer Attitudes. J Hum Lact, 18(3), 247-251.

Linacre, J. M. (2007). Winsteps Rasch measurement (Version 3.63.2). Chicago. IL:
MESA Press.

MacLaughlin, S., & Strelnick, E. G. (1984). Breast-feeding and working outside the
home. Issues Compr Pediatr Nurs, 7(1), 67-81.

Martucci, W. C., & Coverdalel, B. (2001). State Law Trends Affecting Breast-Feeding at
Work. Employment Relations Today, 28(2), 111-121.

McKinley, N. M., & Hyde, J. S. (2004). Personal Attitudes or Structural Factors? A
Contextual Analysis Of Breastfeeding Duration. Psychology of Women Quarterly, 28(4),
388-3 99.

Meek, J. Y. (2001). Breastfeeding in the workplace. Pediatr Clin North Am, 48(2), 461-
474,xvi

Messick, S. (1995). Validity of psychological assessment:Validation of inferences from
person' responses and performances as scientiﬁc inquiry into score meaning. American
Psychologist, 50, 741-749.

Milliken, F. J ., Martins, L. L., & Morgan, H. (1991). Explaining Organizational
Responsiveness to Work-Family Issues: The Role of Human Resource Executives as
Issue Interpreters. Academy of Management Journal, 41 (5), 580-592.

Milliken, F. J ., Morrison, E. W., & Hewlin, P. F. (2003). An Exploratory Study of
Employee Silence: Issues that Employees Don't Communicate Upward and Why. Journal
of Management Studies, 40(6), 1453-1476.

149

Mothers Survey Ross Products Division Abbott Laboratories (2003). Abbott
Laboratories.

NLSY 79 Round 20 Questionnaire. Retrieved March 10, 2005. from
http://wwwblsgov/nls.

Oddy, W. H., Sherrif’f, J. L., de Klerk, N. H., Kendall, G. E., Sly, P. D., Beilin, L. J ., et al.
(2004). The relation of breastfeeding and body mass index to asthma and atopy in
children: a prospective cohort study to age 6 years. Am J Public Health, 94(9), 1531-
1537.

Ortiz, J., McGilligan, K., & Kelly, P. (2004). Duration of breast milk expression among
working mothers enrolled in an employer-sponsored lactation program. Pediatr Nurs,
30(2), 111-119.

Perlow, L. A. (1995). Putting the work back into work/family. Group & Organization
Management, 20(2), 227.

Raisler, J ., Alexander, C., & O'Carnpo, P. (1999). Breast-feeding and infant illness: a
dose-response relationship? Am J Public Health, 89(1), 25-30.

Rasch, G. (1980). Probabilistic models for some intelligence and attainment tests.
Chicago, IL: University of Chicago.

Reifsnider, E., & Taylor-Meyers, S. (1885). Employed Mothers Can Breastfeed Too. Am
J Matern Child Nurs, 10(July/Aug), 256-259.

Reifsnider, E., & Taylor-Meyers, S. (1985). Employed Mothers Can Breastfeed Too. Am
J Matern Child Nurs, 10(July/Aug), 256-259.

Roe, B., Whittington, L. A., Fein, S. B., & Teisl, M. F. (1999). Is there competition
between breast-feeding and maternal employment? Demography, 36(2), 157-171.

Rojjanasrirat, W. (2004). Working women's breastfeeding experiences. MCN Am J
Matem Child Nurs, 29(4), 222-227.

Ryan, A. S., Wenjun, Z., & Acosta, A. (2002). Breastfeeding Continues to Increase Into
the New Millennium. Pediatrics, 110(6), 1103-1109.

150

Satcher, D. S. (2001). DHI-IS blueprint for action on breastfeeding. Public Health Rep,
116(1), 72-73.

Scheider, B. (1990). Organizational Climate and Culture: Jossey-Bass Inc.

Schneider, B. (1990). Organizational Climate and Culture: Jossey-Bass Inc.

Seijts, G. H. (2002). Milking the organization? The effect of breastfeeding
accommodation on perceived fairness and organizational attractiveness. Journal of
Business Ethics, 40(1), 1.

Seijts, G. H. (2004). Coworker Perceptions of Outcome Fairness of Breastfeeding
Accommodation in the Workplace. Employee Responsibilities and Rights Journal, 16(3),
149.

Singhal, A., Cole, T. J ., F ewtrell, M., & Lucas, A. (2004). Breastmilk feeding and
lipoprotein proﬁle in adolescents born preterm: follow-up of a prospective randomised
study. Lancet, 363(9421), 1571-1578.

Singhal, A., Cole, T. J ., & Lucas, A. (2001). Early nutrition in preterm infants and later
blood pressure: two cohorts aﬁer randomised trials. Lancet, 35 7(9254), 413-419.

Slusser, W. M., Lange, L., Dickson, V., Hawkes, C., & Cohen, R. (2004). Breast milk
expression in the workplace: a look at frequency and time. J Hum Lact, 20(2), 164-169.

Smith, M. M., Durkin, M., Hinton, V. J ., Bellinger, D., & Kuhn, L. (2003). Inﬂuence of
breastfeeding on cognitive outcomes at age 6-8 years: follow-up of very low birth weight
infants. Am J Epidemiol, 158(11), 1075-1082.

Taveras, E. M., Capra, A. M., Braveman, P. A., Jensvold, N. G., Escobar, G. J ., & Lieu,
T. A. (2003). Clinician support and psychosocial risk factors associated with
breastfeeding discontinuation. Pediatrics, 112(1 Pt 1), 108-115.

The Family and Medical Leave Act of 1993. Retrieved April 13, 2005, from
www.dol. gov

 

151

Thompson, C. A., Jahn, E. W., Kopelman, R. B., & Prottas, D. J. (2004). Perceived
Organizational Family Support: A Longitudinal and Multilevel Analysis“. Journal of
Managerial Issues, 16(4), 545.

Turocy. (2002). Survey Research in Athletic Training: The Scientiﬁc Method of
Development and Irnplementaion. Journal of Athletic Training, 3 7(4 Supplement), S-174
- S-179.

US Census Bureau. (2007). Retrieved May 1, 2005, from http:,//www.censusgov/

 

US Surgeon General. (2000). HHS Blueprint for Action on Breastfeeding. Retrieved.
from.

Visness, C. M., & Kennedy, K. I. (1997). Maternal employment and breast-feeding:
ﬁndings from the 1988 National Maternal and Infant Health Survey. Am J Public Health,
87(6), 945-950.

Weimer, D. R. (2005). Summary of State Breasﬁ'eeding Laws and Related Issues.
Retrieved July 20, 2005. from
http://www.house.gov/malonev/issues/breastfeedirig/20050519_CRS_State%20Laws.pdf.

 

Weimer, D. R. (2005). Summary of State Breastfeeding Laws and Related Issues.
Retrieved July 20, 2005, from http://www.house.gov/maloney

 

What is colostrum? How does it beneﬁt my baby? (2006). Retrieved May 5, 2007, from
~http://www.llli.org/FAQ/colostrum.html

 

Willis, G. B. (1994). Cognitive Interviewing and Questionnaire Design: A Training
Manual: National Center for Health Statistics.

Witters-Green, R. (2003). Increasing Breastfeeding Rates in Working Mothers. Families,
Systems & Health, 21 (4), 415.

Wolfe, E. W., & Smith, E. V., Jr. (2007a). Instrument Development Tools and Activities
for Measure Validation Using the Rasch Model: Part I—Instrument Development. In R.

Smith & E. V. Smith (Eds), Rasch Measurement: Advances and Specialized Applications
@p.202-242). Maple Grove, MN: JAM Press.

152

Wolfe, E. W., & Smith, E. V., Jr. (2007b). Instrument Development Tools and Activities
for Measure Validation Using the Rasch Model: Part II— Validation. In R. Smith & E. V.

Smith (Eds. ), Rasch Measurement: Advances and Specialized Applications (pp. 243-290).
Maple Grove, MN.

Women in the Labor Force in 2006. Retrieved July 19, 2007, from
http://wwwdol.gov/wb/factsheets/Qf—laborforce-06.htm

Workplace Breastfeeding Support [issue paper]. (2002). Retrieved July 1, 2007. from
http://usbreastfeeding.org/Publications.html.

World Health Organization, U. N. C. s. F. (2003). Global Strategy for Infant and Young
Child Feeding, Geneva, Switzerland: World Health Organization; 2003. .

Wu, M. L., Adams, R. J ., & Wilson, M. R. (1998). Conquest: Generalised item response
modelling software Hawthorn, Australia: ACER.

153

7 LI'

E UNIVERSII

[AI

5

MICHIGAN

. I... a...
:11. s