Content uploaded by Jing Jian Xiao
Author content
All content in this area was uploaded by Jing Jian Xiao on Aug 29, 2014
Content may be subject to copyright.
© 2011 Association for Financial Counseling and Planning Education®. All rights of reproduction in any form reserved. 43
Introduction
Individuals in the U.S. engage in nancial behaviors near-
ly every day and these behaviors inuence their nancial
well-being. For example, the more consumer credit house-
holds assume, the more likely they are to default on their
loans and the less likely they are to have established an
individual retirement account (IRA) (Bernstein, 2004; Sul-
livan, 1987). Thus, of necessity, personal nance research-
ers and nancial planners and counselors measure nancial
management behaviors.
Unfortunately, few validated nancial management behav-
ior scales exist. Researchers typically use proxies of nan-
cial management behavior such as actual levels of con-
sumer debt (e.g., Bernstein, 2004; Sullivan, 1987) rather
than assessing the behaviors themselves. Although some
scales do exist, most lack one (or more) of the following
three characteristics: assessment of multiple domains of
nancial management behavior (Xiao, 2008), psychomet-
ric validation, and validation using a nationally representa-
tive sample of adults. In other words, many scales meas-
ure only one or two dimensions of nancial management
behavior; few have been subjected to tests of validity that
examine whether they measure what they purport to meas-
ure, and those that have been validated have used non-rep-
resentative samples.
The Financial Management Behavior Scale:
Development and Validation
Jeffery Dew and Jing Jian Xiao
Although research on nancial management behavior is common, few nancial management behavior scales ex-
ist that are simultaneously multi-dimensional, psychometrically validated, and validated using nationally repre-
sentative data. Using data from a nationally representative sample of adults, this study developed and examined
the psychometric properties of a new scale of nancial management behaviors. The Financial Management Be-
havior Scale (FMBS) displayed adequate reliability (alpha = .81). The FMBS was highly associated with other
measures of nancial management behaviors and was predictive of participants’ actual levels of savings and
consumer debt. These ndings suggest that the full FMBS is a reliable and valid measure of nancial manage-
ment behaviors, though the subscales need renement.
Key Words: nancial behavior, measurement, psychometrics
Jeffrey Dew, Ph.D., Family, Consumer, and Human Development, Utah State University, 670 E. 500 N., Logan, UT 84321, jeff.dew@usu.edu,
(435) 797-9184, (435) 797-7220
Jing Jian Xiao, Ph.D., Human Development and Family Studies, The University of Rhode Island, Transition Center 2 Lower College Rd., Kingston,
RI 02881, xiao@uri.edu, (401) 874-2547, (401) 874-4020
A comprehensive and psychometrically strong measure of
nancial management behaviors could assist researchers
and practitioners in many elds. For example, in addition
to the obvious nancial benets, sound nancial man-
agement behaviors have both personal and interpersonal
consequences. Consumer debt levels are positively related
to anxiety (Drentea, 2000). Further, assets and consumer
debts are associated with relationship quality among
married couples (Dew, 2007). Finally, positive nan-
cial management behaviors are associated with physical
health, mental health, academic success, and life satisfac-
tion among college students (Xiao, Tang, & Shim, 2009).
To address these gaps, we designed the Financial Man-
agement Behavior Scale (FMBS). The FMBS was then
psychometrically validated using the Familial Response
to Financial Instability Study, a nationally representative
sample of adults.
Measuring Financial Management Behaviors:
Existing Scales
One of the problems with existing measures of nancial
management behavior is that many are not comprehensive.
To construct the FMBS, we initially examined the nan-
cial measures in seven studies (Fitzsimmons, Hira, Bauer,
& Hafstrom, 1993; Hilgert, Hogarth, & Beverly, 2003;
Jorgensen, 2007; Kim, Garman, & Sorhaindo, 2003; Perry
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011
44
& Morris, 2005; Servon & Kaestner, 2008; Xiao, Shim,
Barber, & Lyons, 2008). A deeper review of the literature
resulted in eight more studies that used nancial manage-
ment behavior scales (Davis, 1992; Davis & Weber, 1990;
Godwin & Koonce, 1992; Grable, Park, & Joo, 2009; Mu-
genda, Hira, & Fanslow, 1990; Prochaska-Cue, 1993; Rosen
& Granbois, 1983; Scannell, 1990). Of these 15 studies,
10 used instruments that measured only one or two nan-
cial domains. Thus, only one third of the instruments found
measured three or more domains of nancial management
behavior. Among those that measured more than two dimen-
sions, two used single indicators to measure the dimensions.
Measuring many different domains of nancial manage-
ment behaviors is important because each domain can
have a serious impact on family life. For example, only
one measure asked questions about insurance (Jorgensen,
2007). However, families with inadequate health insur-
ance face an increased risk of unmet health needs (Aya-
nian, Weissman, Schneider, Ginsburg, & Zaslavsky, 2000),
shoulder the nancial burden of expensive medical bills
(Short & Graefe, 2003), and may be more likely to declare
bankruptcy (Domowitz & Sartain, 1999). However, a lack
of health insurance is not the only predictor of nancial
insolvency; consumer debt levels are also associated with
bankruptcy (Sullivan, 1987). Thus, though each of these
domains is important in and of itself, when families use
sound nancial management behaviors in all of these do-
mains, their nancial position becomes strong (Joo, 2008).
Another problem with current nancial management be-
havior instruments is that few have been psychometrically
validated. Psychometric validation is the process of testing
the properties of the scale, usually reliability and validity.
If a scale is psychometrically valid, it measures what it
purports to measure and gets the same results over multi-
ple uses (Cohen & Swerklik, 1999; Silva, 1993). Instru-
ments that have not been psychometrically validated may
produce questionable inferences (Silva, 1993).
Despite many studies of nancial management behaviors,
researchers have validated only two scales; the Frequency
of Financial Management Scale (Fitzsimmons et al., 1993)
and the Personal Financial Management Style (Prochaska-
Cue, 1993). Unfortunately, these scales were either not
comprehensive, or they were validated using nongeneraliz-
able samples.
Using nationally representative samples to validate -
nancial management behavior instruments enhances the
external validity of the instrument. That is, it shows that
the psychometric properties of the scale generalize to
a wide population. The studies that were validated had
participants that were drawn largely from the Midwest in
rural (Fitzsimmons et al., 1993) or metropolitan (Prochas-
ka-Cue, 1993) areas. Indeed, the majority of all nancial
behavior instruments reviewed were developed and used in
Midwestern contexts. This leaves questions as to the gen-
eralizability of the results.
Framework of the FMBS
The FMBS incorporates the idea that individuals will
serially adopt good nancial management behaviors. For
example, a national study of consumers revealed a hierar-
chical pattern of nancial management behaviors (Hilgert
et al., 2003). About two thirds (66%) of the participants
practiced cash ow management and 45% managed credit.
However, only 33% used savings management and only
19% of the participants invested. This suggests a gradual
uptake in nancial management behavior with cash man-
agement developed rst, then credit, savings, and nally
investment management.
This behavioral hierarchy may arise because of the nan-
cial resource differences across individuals. For exam-
ple, when families’ incomes are insufcient to meet their
nancial obligations, they may not have the capacity to
save (Garasky, Nielsen, & Fletcher, 2008). Further, certain
nancial management behaviors, such as paying off con-
sumer credit, may take precedence over other types such as
contributing to a retirement fund (Bernstein, 2004). Some
individuals may not have insurance policies because they
do not own property or may not have access to employer-
provided health insurance plans (DeNavas-Walt, Proctor,
& Lee, 2006).
In addition to measuring cash management, savings and
investments, credit use, and insurance, the FMBS meas-
ured nancial management behavior that might precede
cash management – consumption management. Because
virtually all individuals are consumers, if nothing else,
they could engage in behaviors that would maximize their
consumption benets. Called “Shopping and Purchases” in
the actual scale, we hypothesize that even more individuals
will use strategies to manage their purchases than will use
cash-ow management behaviors.
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011 45
Method
Scale Construction
The rst step in constructing the FMBS was to examine
seven articles and reports that used or developed nancial
management behaviors scales. We identied the domains
of nancial management behaviors that these scales meas-
ured and also noted the domains that were not included in
the scales. For example, cash ow management behaviors
and savings behaviors were frequently measured, whereas
credit management was less frequently measured.
Following the review of previous measures, ve domains
were selected that were important areas of sound nancial
management behaviors: consumption, cash ow, credit,
savings and investment, and insurance. Measures related to
each of these domains were then written. Every domain in
the FMBS had at least three behavioral measures.
The next step was to send a draft to nine nancial plan-
ning and counseling professionals and nancial manage-
ment scholars. These individuals reviewed the FMBS to
ensure that it covered the important domains of nancial
management behavior (i.e., content validity) and that the
FMBS appeared “on its face” to measure nancial man-
agement behavior (i.e., face validity).The panelists felt that
the FMBS covered the main domains of good nancial
management behavior. Further, the panelists felt that the
measures were important aspects of each domain.
The last step was to utilize panelist suggestions to
strengthen and clarify the FMBS. As a result of the input,
some items were added and others were dropped. For
example, an item that more directly measured investment
behaviors was added. Some of the items and the scale
instructions were also reworded to make them clearer. The
nal FMBS appears in Appendix A.
A few of the suggestions were not included. For example,
some panelists noted that estate planning and taxes were
missing. In spite of this suggestion, estate planning was
not included because we were concerned that many young
adults do not use estate planning even if they otherwise
use sound nancial management behavior. This would
drag down the scores of some participants in spite of the
fact that they were otherwise managing their nances well.
Tax behavior was also not put into the measure because
we faced a limit on the number of questions we could ask
and we were not sure that enough individuals utilized tax
minimization strategies to be a useful domain. One panelist
also suggested adding attitudinal measures. We declined to
do this. We wanted to restrict the FMBS to measuring be-
haviors because behaviors are most proximal (i.e., directly
related) to nancial outcomes than are nancial attitudes
(Xiao, 2008) and because well-established nancial attitude
scales already exist (e.g., Furnham, 1984; Tang, 1995).
Sample and Data
The sample was drawn from the Familial Response to
Financial Instability Study. Initiated by the National
Center for Marriage and Family Research, this project
was designed to test how families were coping with the
2007 – 2009 Recession. Projects designed to create new
measures of family nance and examine new areas in fam-
ily nance research were solicited. The current study was
one of the projects that the National Center for Marriage
and Family Research accepted. The survey was conducted
in August 2009 using a stratied random sample design
(Dennis & McCready, 2009). Participants were initially
contacted through both random-digit dialing and address-
based sampling. Using these methods, the study recruited
households with landlines as well as cell-phone only and
households that had no phone service.
Individuals who agreed to be part of the study participated
via the Internet. If participants did not have access to the
Internet, they received the technology necessary to con-
nect to it (e.g., a laptop computer and Internet access at
their residence). Out of the 1,517 individuals contacted,
1,014 participated – a 67% response rate. When combined
with the post-stratication weight, the sample is nation-
ally representative of adults (Dennis & McCready, 2009).
The post-stratication weights were used to generate the
descriptive statistics, the factor analysis, and the regres-
sion analysis. To ensure that the ndings were robust, the
weighted factor analysis and regressions were compared to
unweighted analyses (not shown). The weighted and un-
weighted analyses produced similar ndings with only two
exceptions. First, in the unweighted regressions the mag-
nitudes of the coefcients were occasionally larger. This
suggests that the weighted estimates were slightly more
conservative. Second, the unweighted factor analysis was
the same except that paying bills on time loaded more on
the credit management factor than the positive cash man-
agement factor. This study presents the ndings that used
the post-stratication weights. The sample was composed
of 1,011 participants. Three participants who failed to
answer any of the study questions except the demographic
items were dropped.
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011
46
Measures
FMBS Scale Items. The FMBS included 17 items (see
Appendix A). The instruction for the rst 14 items was,
“Please indicate how often you have engaged in the fol-
lowing activities in the past six months.” The response set
for these questions ranged from 1 (Never) to 5 (Always).
Participants could also report that the item was not ap-
plicable. These 14 items measured participants’ nancial
management behaviors in four domains: consumption,
cash management, savings and investment, and credit
management. Examples of items from each of these do-
mains include, “comparison shopped when purchasing a
product or service,” “kept a written or electronic record of
your monthly expenses,” “bought bonds, stocks, or mutual
funds,” and “maxed out the limit on one or more credit
cards.” Items that represented poor nancial management
behaviors were reverse coded prior to the analysis. The in-
struction for the three insurance items stated, “Please rate
your behavior regarding insurance within the past year on
a scale of 1 – 5.” These items asked about the past year be-
cause sometimes insurance behaviors occur on an annual
basis. The response set was the same as the other items.
The insurance items queried participants about health in-
surance, property insurance, and life insurance.
Validation Measures. The survey also collected measures
that would facilitate validating the FMBS. First, ve items
of nancial management behavior were included to make
sure that the FMBS demonstrated convergent validity.
These items came from a study that measured “responsi-
ble nancial behaviors” (Perry & Morris, 2005, p. 304).
The instruction for these items stated, “How do you grade
yourself in the following areas?” The items included
controlling spending, paying bills on time, planning for
the nancial future, providing for oneself and family, and
saving money. Participants could respond from 1 (Poor) to
5 (Excellent). We chose this particular scale because Perry
and Morris (2005) had a nationally representative sample.
The survey also measured participants’ actual levels of
savings and consumer debt. Although the survey had ini-
tially measured the exact level of these variables, the study
IRB panel requested the use of scales. For the savings item
and the consumer debt item there were nine categories
(see Appendix B). Participants could respond that they had
between 1 (None) and 9 ($100,000 or more) in savings and
consumer debt. The scales had smaller increments at lower
amounts so that we could distinguish among groups with
fewer assets and consumer debt more easily. This decision
seemed justied given that nearly 75% of the sample had
under $10,000 of consumer debt.
Demographic Characteristics. The survey contained
demographic characteristics to use as control covariates.
Total household income was among these variables, and it
was measured on a scale from 1 – 19, or from “Less than
$5,000” to “$175,000 or more” (see Appendix B). Although
the mean was 10.76 (suggesting an average of $35,000 to
$39,999), the median was 11 suggesting a median income
range of ($40,000 to $49,999). The median household
income in 2009 in the U.S. was $50,112 (DeNavas-Walt,
Proctor, & Smith, 2010). Thus, the scaled income measure
may closely align with the actual U.S. median household
income, or it may be nearly $10,000 off. This unfortunate
lack of precision was a result of an IRB request and not
a decision made by the researchers. For the purposes of
examining how the scale performed among low-income
individuals, we also created a sample of individuals of the
lowest income quintile (n = 210). These individuals had
to have an income score of 7 or less (i.e., $24,999 or less).
The income quintile cut off for the lowest U.S. quintile in
2009 was $20,453 (U.S. Census, 2010). A score of 7 on the
scale contains this amount, but also contains some indi-
viduals who were above this amount. In the present sample,
21% of individuals had an income score of 7 or less.
Additional control covariates included two dichotomous
variables that assessed marital and cohabiting status (sin-
gle, not cohabiting was the omitted category), age and
gender (male is the omitted category). Two dichotomous
variables measured race and ethnicity (White, Non His-
panic is the omitted category), and three dichotomous vari-
ables measured obtained education (less than high school
degree is the omitted category). Finally, two dichotomous
variables measured employment status (not employed is
the omitted category), and a continuous variable assessed
the number of children in the home. Descriptive statistics
for the demographic measures are found in Table 1.
Handling “Not Applicable” Responses
One problem was that participants were allowed to re-
spond “not applicable” on the nancial management
behavior items. This was a reasonable response for some
nancial management behaviors. For example, it would
be difcult for participants to “max out” a credit card if
they did not have one. Unfortunately, some participants
responded at seemingly inappropriate times. For example,
35 respondents (3.4%) reported that the item “paying bills
on time” was not applicable and 45 respondents (4.4%) re-
ported that “staying within their budget or spending plan”
was not applicable. It may be that younger individuals, or
individuals in a married couple with high specialization
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011 47
might lead to these responses (e.g., if the respondent were
not in charge of these tasks). Given that over 75% of those
who answered not applicable for these two variables were
over the age of 23 and over two thirds were single, this
seems unlikely. The average level of answering not appli-
cable was 11%, with a range of 3% – 18%.
Participants who answered not applicable could not be
included in the factor analyses or the regressions. We tried
three possible solutions. First, participants with missing
data were listwise deleted. A second solution was to set the
“not applicable” responses to the lowest response based on
the assumption if participants were choosing not applica-
ble then they were not engaging in that behavior. Third,
multiple imputation was used to generate a plausible
value for the missing response. Multiple imputation uses
maximum likelihood methods to generate ve plausible
responses for the missing response. Any statistical analy-
ses that are run are actually run ve times (once for each
imputed response) and then the results are synthesized
(Rubin, 1987).
The results were the same no matter how we dealt with the
missing responses. That is, listwise deleting participants,
setting their missing responses to the lowest level, or
using multiple imputation yielded the same factor solu-
tion. Although the factor loadings were slightly differ-
ent, they were not different enough to inuence the factor
extraction. This study presents the results generated using
multiple imputation because we also used multiple impu-
tation for the participants who declined to answer ques-
tions about their demographic characteristics such as their
income or age. The average level of missing data for the
demographic variables was 2%, with a range of 1% – 5%.
All of the results presented in this study were created using
multiple imputation.
Table 1. Descriptive Statistics (N = 1,011)
M SD % Range
Responsible nancial behaviors scale 3.38 .95 1 – 5
Actual amount of savingsa4.23 2.67 1 – 9
Actual amount of consumer debta3.41 2.35 1 – 9
Marriedb45
Cohabitingb10
Femalec52
Blackd11
Other race/ethnic minorityd20
High school degreee31
Some collegee28
College degree or highere28
Employed full timef43
Employed part timef12
Age 46.51 16.68 18 – 90
Incomea10.76 4.14 1 – 19
Number of children in the home .54 1.05 0 – 8
Note. a See Appendix 3 for a more detailed table on these nancial variables. b Omitted category is single not cohabiting; 45%
of the sample was in the omitted category. c Omitted category is male; 48% of the sample was in the omitted category. d Omit-
ted category is White, Non-Hispanic; 69% of the sample was in the omitted category. e Omitted category is less than high
school; 13% of the sample was in the omitted category. f Omitted category is not employed; 45% of the sample was in the
omitted category.
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011
48
Results
Factor Analysis
The rst analysis was an exploratory factor analysis to
extract the factors among the nancial management be-
haviors. This exploratory method accounts for as much of
the total variance among the variables as possible by using
latent factors. These latent factors identify commonali-
ties among the variables. Exploratory factor analysis is a
common data reduction analysis that is appropriate at the
beginning of scale construction (Comrey & Lee, 1992).
Following the initial factor extraction we utilized a promax
rotation to simplify the factor solution. Promax is a type
of oblique rotation which allowed the latent factors to
correlate with each other. Because of the hypothesis that
individuals would adopt nancial management behaviors
serially, we thought it would be most appropriate to use an
oblique rotation.
Although ve factors were hypothesized, we used the re-
sidual correlations and the scree test (Cattell, 1966) to de-
cide the proper number of factors. Because the idea behind
factor analysis is to account for as much of the variance
between the items as possible using latent factors (Com-
rey & Lee, 1992), the best models will produce the fewest
residual correlations.
A four-factor solution produced the same number of resid-
ual correlations under .10 as a ve-factor solution (residual
correlations available on request). Thus, the residual cor-
relation test suggested retaining four or ve factors. The
scree plots (available on request) also suggested the same
course. Using a scree plot test, we wanted to retain the
number of factors corresponding to the factor that departs
from the main line of the lower factors (Comrey & Lee,
1992). This suggested a 4- or 5-factor solution because the
eigenvalue for the fourth factor was clearly above the line,
whereas the eigenvalue for the fth factor was slightly
above the line. We decided to use a four-factor solution
for the sake of parsimony. After this decision, the promax
rotation was conducted. In order for an item to “load” on a
factor it had to have a loading of .6 or better.
Two problematic items surfaced during the rotation. One
of the items, impulsive buying, did not load well on the
factors; it never had a factor loading above .49. Further,
impulsive buying was the only variable that loaded on
a fth factor when ve factors were extracted. These
ndings suggested that impulsive buying was tapping a
different construct than the other items. A second item,
searching for information prior to a large purchase, was
also problematic. It loaded on multiple factors equally
but never at the .6 level. Further, when it was in the factor
analysis, other items loaded on multiple factors equally.
When information search was eliminated, all of the other
items clearly loaded on one factor. Because impulsive
buying and searching for information proved problematic,
these items were dropped from the scale.
We evaluated whether dropping these two items altered
the number of factors to extract. The residual correlation
analysis did not change. A four-factor solution performed
as well as a ve-factor solution. The scree plot was modi-
ed somewhat and clearly showed that a four-factor solu-
tion was most appropriate. The fth eigenvalue was part of
the main line of lower factors. This was expected because
the impulse buying item was the only item that loaded well
on a fth factor. Thus, after dropping these two variables
we retained a four-factor solution.
The factor structure after dropping the two variables is
found in Table 2a. Five savings and investment items
loaded on Factor 1 with loadings of .69 to .78. Hence, we
call Factor 1, “Savings and Investment Behaviors.” Factor
1 explained 30% of the variance. The three insurance items
loaded on Factor 2 (loadings between .74 and .85) or the
“Insurance Behavior” factor. This second factor explained
12% of the variance. Items that indicated cash manage-
ment, such as keeping a nancial record and paying bills
on time, loaded on Factor 3 with loadings between .64 and
.73. We called this factor the “Cash Management” factor,
and it explained 9% of the variance. Finally, three items
indicating positive credit management loaded on Factor 4.
The loadings ranged from .66 to .77. We called this factor
“Credit Management.” It explained 8% of the variance.
Overall, the four factors explained 59% of the variance.
Because of the hypothesis that these different factors
would all be related to the same construct – overall nan-
cial management behaviors –mean scales were created
from each factor and they were then factor analyzed. The
subscales loaded on only one factor (results available upon
request). This supports the idea of a larger super-factor of
nancial management behavior. It also supports the use of
the overall FMBS score as a measure of nancial manage-
ment behaviors. A revised version of the FMBS is found in
Appendix A.
Financial Management Behavior Hierarchy
Following the nalization of the subscales, they were ex-
amined to check whether they displayed the same hierar-
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011 49
chical behavior as previous studies have shown (see Table
2b). Like other studies, savings and investment behaviors
were the least frequent nancial management behaviors.
Unlike other studies, however, cash management and
credit management behaviors were at about the same level.
Insurance behaviors were the most common behavior type.
Reliability Analysis
Reliability was assessed using Cronbach’s alpha. Cron-
bach’s alpha is equivalent to the average of all possible
split-half reliabilities for a scale. The full FMBS had a
Cronbach’s alpha of .81. Cronbach alpha scores were also
calculated for the four subscales. The savings and invest-
ments subscale and the insurance subscale had satisfac-
tory Cronbach alpha scores (.78 and .73, respectively).
Cronbach’s alpha for the cash management subscale and
the credit management subscale were lower (.63 and .57,
respectively) indicating that the items that made up these
scales did not hang together as well as the items that made
up the savings subscale and the insurance subscale. These
Table 2a. Rotated Factor Structure (N = 1,011)
Factor 1 Factor 2 Factor 3 Factor 4
Comparison shop .73
Pay bills on time .64
Keep a nancial record .69
Stay within budget .66
Pay off credit card .66 b
Max out credit card a .77
Make minimum payment on loans a .72
Maintain or create an emergency fund .73
Save from every paycheck .75
Save for a long term goal other than retirement .78
Save for retirement .69
Invest money .70
Obtain or maintain adequate health insurance .85
Obtain or maintain adequate property insurance .74
Obtain or maintain adequate life insurance .80
Eigenvalues 4.43 1.84 1.31 1.25
% of Variance Explained by Factor
(Total Variance Explained = 59%) 30 12 9 8
Note. a Reverse coded. b This item loaded on the savings and investment factor (Factor 1) best among the low-income subsam-
ple of this study.
Table 2b. FMBS Scale and Subscale Descriptive Statistics (N = 1,011)
M SD Range % Often or more frequently
FMBS 3.48 .71 1.58 – 5.00 26.5
Savings and investment subscale 2.66 1.09 1.00 – 5.00 11.0
Cash management subscale 3.73 .83 1.00 – 5.00 48.1
Credit management subscale 3.73 .95 1.00 – 5.00 47.2
Insurance subscale 3.81 1.24 1.00 – 5.00 53.7
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011
50
reliability analyses suggested that the overall FMBS scale
was reliable, at least for the nationally representative sam-
ple. It also suggested that the savings and investment sub-
scales had sufcient reliability. When using the cash ow
management and positive credit behavior subscales, some
caution is warranted. Though they had reasonable levels
of reliability at this stage of the scale development, future
work will need to improve them.
Validity
Face and Content Validity. We addressed face validity and
content validity through the use of an expert panel. The
nancial planners and nancial counselors agreed that the
FMBS items appeared to measure what it purported to
measure (face validity). Further, they also agreed that the
FMBS measured important nancial management behav-
iors (content validity).
Construct Validity. Addressing construct validity neces-
sitated demonstrating that the FMBS measured what it
claimed to measure (Cohen & Swerklik, 1999; Silva,
1993). An alternative way of thinking of construct valid-
ity is that the inferences drawn from the use of the scale
are sound (Silva, 1993). Construct validity was assessed
by examining convergent validity. A measure demon-
strates convergent validity when it is associated with other
items or scales that measure the same construct (Cohen &
Swerklik, 1999; Silva, 1993).
To assess convergent validity, we used weighted least
squares regression to regress a scale that measured -
nancial management behaviors (Perry & Morris, 2005)
onto the FMBS, the subscales, and the control covariates.
In all of the regression analyses, we obtained the FMBS
score by taking the mean of the 15 items. This was also
the case for the subscales. An analysis of the scale using
summed scores (not shown) indicated that the regression
ndings were exactly the same except that the magnitude
of the non-standardized FMBS and subscale coefcients
and standard errors were different (though the t-tests of
signicance were exactly the same). This is to be expected
because taking a mean is a simple linear transformation
that does not change the distribution of the variables.
The FMBS was positively associated with the responsible
nancial behaviors scale (b = .94, p < .001; see Table 3,
Model 1). Given that the standard deviation of the respon-
sible nancial behaviors scale was .96, this coefcient
represented an effect size of nearly 1.0. Very few of the
control covariates were signicant and the overall model
explained more than half of the variance in the responsible
nancial behaviors scale. The subscales were also posi-
tively associated with the responsible nancial behaviors
scale (see Table 3, Model 2). Thus, the FMBS displayed
remarkable convergent validity. Discriminant validity was
also tested using a time-use scale. The FMBS and sub-
scale demonstrated discriminant validity (results available
upon request).
Criterion Validity. Measures that possess criterion validity
predict variables that they would be expected to predict if
they truly measured what they claimed to measure (Kaplan
& Saccuzzo, 2009; Silva, 1993). Because the data were
cross-sectional, the concurrent form of criterion validity
was used rather than the predictive form. That is, the data
could only show that the FMBS predicted criterion that
were contemporary with its measure, rather than showing
that the FMBS predicted the levels of a future criterion.
Criterion validity was measured by using weighted least
squares regression to assess the association between ac-
tual levels of savings and consumer debt on the FMBS
and subscales. If the FMBS truly measured sound nan-
cial management behaviors it should be associated with
these nancial measures. Table 4a shows the associa-
tion between the FMBS, the subscales, and savings. For
every one unit increase in the FMBS, savings increased
by 1.65 (b = 1.65, p < .001, Table 4a). The FMBS coef-
cient represents .6 of a standard deviation for actual level
of savings and the model explained 50% of the variance
in savings. When the subscales were used instead of the
full scale, they performed as expected. Savings behaviors
and positive credit behavior were strongly associated with
actual savings (b = .70 and .63, respectively, p < .001 for
both). Insurance behaviors were also positively associated
with savings.
We also regressed participants’ reports of their consumer
debt level onto the FMBS, the subscales, and the control
covariates. The FMBS was negatively associated with con-
sumer debt levels (b = -.90, p < .001; Table 4b, Model 1).
This coefcient represents a .38 standard deviation effect
size for consumer debt but had an R2 of only .16.
The analysis of the subscales (Table 4b, Model 2) had both
expected and unexpected results. As expected, the sav-
ings and investment subscale and the credit management
subscale were negatively associated with participants’ con-
sumer debt (b = -.31 and -1.04, respectively, p < .001 for
both). Further, the subscale model explained nearly twice
the variance as the FMBS model (R2 = .31).
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011 51
Unexpectedly, cash management was not signicantly
associated with consumer debt and insurance behaviors
positively predicted consumer debt (b = .33, p < .001). In a
regression with just cash management and the control cov-
ariates, the cash management subscale negatively predict-
ed consumer debt (b = -.42, p < .001). This indicated that
the cash management subscale shared so much variance
with the other subscales that it did not explain independent
variance in consumer debt when the others were included.
In a regression with just the insurance subscale and the
control covariates, insurance behaviors were still positive-
ly associated with consumer debt, however. This unexpect-
ed nding will be addressed in the discussion section.
Finally, given the distribution of the actual savings and
actual consumer debt variable, some may question whether
weighted least squares regression was the most appropri-
ate analytic tool. To address this issue we dichotomized
these variables at the top 20% – 25% mark (a score of 8 or
above for savings, and a score of 6 or above for consumer
debt). If participants were in the top 20% – 25%, they were
scored as a 1 on the dummy variables and a 0 otherwise.
We then reran the criterion analyses using logistic regres-
sion to examine whether the FMBS and the subscales were
associated with being in the top 20% – 25% for savings
and consumer debt. The results that we got from the logis-
tic regression were comparable to the results that we got
Table 3. WLS Regression Models of the Responsible Financial Behaviors Scale (N = 1,011)
Model 1 Model 2
b SE b SE
Intercept .21 .15 -.03 .16
FMBS .94*** .03
Savings and investment subscale .29*** .03
Cash management subscale .37*** .04
Credit management subscale .25*** .03
Insurance subscale .06* .02
Marrieda-.02 .05 .03 .05
Cohabitinga.03 .09 .08 .07
Age .01 .01 .01 .01
Income .01 .01 .01 .01
Femaleb-.05 .04 -.03 .04
Blackc-.09 .07 -.06 .07
Other race/ethnic minorityc.02 .06 -.04 .05
High school degreed-.18* .07 -.15* .07
Some colleged-.12 .08 -.10 .07
College degree or higherd-.07 .08 -.09 .07
Employed full timee.03 .05 .07 .05
Employed part timee-.03 .08 .01 .07
Number of children in the home -.05* .02 -.03 .02
R2.54 .59
Note. a Omitted category is single not cohabiting. b Omitted category is male. c Omitted category is White, Non-Hispanic.
d Omitted category is less than high school. e Omitted category is not employed.
*p < .05. **p < .01. ***p < .001.
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011
52
Table 4a. WLS Regression Models of Actual Amount of Savings (N = 1,011)
Model 1 Model 2
b SE b SE
Intercept -5.74*** .48 -5.17*** .50
FMBS 1.65*** .11
Savings and investment subscale .70*** .08
Cash management subscale -.05 .09
Credit management subscale .68*** .08
Insurance subscale .24*** .06
Marrieda.06 .16 .21 .16
Cohabitinga.14 .23 .22 .23
Age .04*** .01 .04*** .005
Income .17*** .02 .16*** .02
Femaleb-.18 .13 -.12 .13
Blackc -.93*** .21 -.93*** .21
Other race/ethnic minorityc.09 .18 .01 .18
High school degreed.32 .23 .35 .23
Some colleged.02 .25 .16 .23
College degree or higherd .66** .25 .76** .25
Employed full timee .38* .16 .36* .16
Employed part timee.43 .22 .35 .22
Number of children in the home -.10 .07 -.07 .07
R2.50 .52
Note. a Omitted category is single not cohabiting. b Omitted category is male. c Omitted category is White, Non-Hispanic.
d Omitted category is less than high school. e Omitted category is not employed.
*p < .05. **p < .01. ***p < .001.
from the weighted least squares regression (analysis not
shown, results available upon request). That is, the higher
a participant’s FMBS score, the more likely they were
to be in the top savings group and the less likely they
were to be in the top consumer debt group. The subscale
analyses also yielded similar results. The only difference
was that the insurance subscale was not associated with
the likelihood of being in the high savings group in the
logistic models.
Together these ndings presented evidence for criterion
validity. The FMBS purports to measure sound nan-
cial management behaviors. Both the full scale and the
subscales predicted nancial outcomes that are associated
with sound nancial management behaviors.
External Validity. Although the nationally representative
sample offered a solid context in which to test the FMBS,
it also raises questions as to whether the FMBS is general-
izeable to different subpopulations. For example, some of
these behaviors may be difcult to implement, especially
for individuals with lower levels of income. As a nal test
of the FMBS, we reran all of the above tests (analysis not
shown) for participants whose income was in the bottom
quintile of the sample (an income score of 7 or less, n =
212). The ndings for this subsample were mostly similar
to the ndings from the nationally representative sample
with only a few caveats. A four-factor solution was found
to t the data best for these participants. Further, the fac-
tor structure remained the same except that the item for
paying off credit card debt loaded better on the savings
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011 53
and investment factor than the credit management factor.
The tests of reliability and validity were also similar to the
full sample. Overall these ndings suggested that for low
income individuals the FMBS and the associated subscales
functioned similarly to a general population sample.
Discussion
This study represents a rst attempt to develop the Finan-
cial Management Behavior Scale (FMBS). It also explored
the scale’s psychometric properties using a nationally rep-
resentative sample. The full scale had adequate reliability
(Cronbach’s alpha = .81). Further, our analyses suggested
that the FMBS does measure what it purports to meas-
ure. It was strongly associated with another measure of
nancial management behaviors. Further, it demonstrated
concurrent criterion validity with respect to actual levels of
savings and consumer debt. Specically, as scores on the
FMBS increased, participants’ reported levels of savings
increased and their reports of consumer debt decreased.
Using an exploratory factor analysis indicated that a four-
factor structure within the FMBS was the best solution.
The subscales – cash management, credit management,
savings and investments, and insurance – demonstrated
less reliability and validity than the full scale, though for a
rst attempt they performed reasonably well. The sub-
scales require further rening.
Despite these overall ndings, there were some unexpect-
ed results. First, the FMBS was designed to measure ve
dimensions but the factor analysis suggested that a four
Table 4b. WLS Regression Models of Actual Consumer Debt (N = 1,011)
Model 1 Model 2
b SE b SE
Intercept 3.76*** .50 4.49*** .49
FMBS -.90*** .12
Savings and investment subscale -.31*** .07
Cash management subscale .01 .09
Credit management subscale -1.04*** .08
Insurance subscale .33*** .06
Marrieda .80*** .17 .48** .16
Cohabitinga.44 .25 .31 .23
Age .01 .01 .01* .005
Income .06** .02 .05* .02
Femaleb.18 .14 .02 .13
Blackc-.31 .23 -.40 .21
Other race/ethnic minorityc-.34 .19 -.19 .17
High school degreed .67** .24 .59** .22
Some colleged 1.34*** .25 1.12*** .22
College degree or higherd 1.02*** .27 1.01*** .24
Employed full timee .70*** .17 .55*** .16
Employed part timee .58* .23 .59** .22
Number of children in the home .20** .08 .13 .07
R2.16 .31
Note. a Omitted category is single not cohabiting. b Omitted category is male. c Omitted category is White, Non-Hispanic.
d Omitted category is less than high school. e Omitted category is not employed.
*p < .05. **p < .01. ***p < .001.
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011
54
-factor solution was the best tting and most parsimoni-
ous solution. Consumption management was to be the
fth dimension of the scale, but two of the consumption
items were problematic and were ultimately dropped to
increase the strength of the FMBS. The nal consump-
tion measure (comparison shopping) loaded on the cash
management factor.
Two competing possibilities explain the consumption nd-
ings. The rst is that we may simply have not written very
good consumption management questions. Impulse buy-
ing was clearly not related to any of the other items in the
FMBS and searching for information prior to a large pur-
chase was also problematic. Perhaps other questions might
have more suitably measured consumption management.
An alternative possibility is that consumption behaviors
might simply not load well with the other nancial man-
agement behaviors that we measured. Individuals may
mentally treat consumption and money management as
two different domains. Financial management behaviors
may be relevant to only money management related activi-
ties that do not include spending behavior. Future research
would need to test these speculations.
Another nding that failed to materialize was a distinct
hierarchy of nancial management behaviors. Although
savings and investments had the lowest mean and were
practiced often by the lowest number of people, the other
behaviors – insurance, cash management, and credit man-
agement – were practiced at about the same level. This
conicts with ndings from other studies (e.g., Hilgert et
al., 2003). Part of the reason for the discrepancy might be
that the data were collected during the 2007 – 2009 Reces-
sion. The recession simply may have forced more individ-
uals to engage in cash and credit management behaviors.
Finally, the cash management and insurance subscales
demonstrated some unanticipated analytic properties. The
cash management subscale seemed particularly sensitive
to the presence of the other subscales. For example, in two
of the analyses cash management was not signicant. Yet
when it was run without the subscales in the model it was
signicant. This suggests that cash management sometimes
shares so much variance with the other subscales that
it does not independently predict participants’ nances.
Given that the reliability for cash management was lower
than savings it is not surprising that it was not as strong in
the multivariate models.
Another unexpected nding from the subscale analysis
concerned the insurance behavior subscale. Although the
insurance behaviors positively predicted savings, it also
positively predicted consumer debt. Even in the bivariate
analyses, insurance behaviors were positively correlated
with consumer debt. This may suggest that some partici-
pants had maintained their insurance policies through
the use of consumer credit. Many consumers purchase
and maintain insurance policies using credit cards. This
is particularly the case with the growth of insurance
websites. Consequently, even though insurance behaviors
are positively associated with actual savings, they may
also encourage accumulating consumer debt. These ideas
are speculative though, and need to be directly tested in
future research.
One of the main limitations of this study was that the data
were not longitudinal. This limited our analysis to con-
current validity rather than predictive validity. That is, a
stronger test would have been to show that measures of
the FMBS predicted future levels of savings and consumer
debt, or even changes in levels of savings and consumer
debt. Further, not having longitudinal data limits what can
be said about the direction of the relationship between
nancial management behaviors and actual levels of nan-
cial well-being.
Another limitation was that the study relied on self-report-
ed data. Some individuals may have given socially desir-
able responses. For example, some participants may have
overstated the frequency at which they save. To the extent
that socially desirable answers were not random, this may
further inuence how this scale functions for different sub-
groups. For example, if individuals in a particular socio-
economic status were more likely to give socially desirable
answers, this may skew the psychometric properties of the
scale for that subgroup. The convergent and concurrent
validity tests did indicate that the scale worked as it should
– with higher FMBS scores being positively associated
with another nancial management behaviors measure
and with actual amounts of savings and consumer debt.
Individuals may have given socially desirable answers
throughout the survey, however. Thus, the possibility for
the analysis to be inuenced by socially desirable answers
remains. Although this study could not mitigate this issue,
it exists for every study that uses self-reported data.
In addition to these study limitations, the FMBS needs ad-
ditional renement. First, the scale needs to reect various
realities of the life course. Following retirement, for ex-
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011 55
ample, many individuals and families will not be “saving
money from every paycheck” or “saving for a long-term
goal.” Although we tried to anticipate life course possibili-
ties by including a “not applicable” response, and by not
including estate planning in the scale, this solution did not
fully solve the problem. Rather, our solution introduced
additional problems of individuals selecting “not applica-
ble” to applicable items.
This problem has implications for the use of this scale.
That is, practitioners and researchers using this scale need
to use it judiciously. For example, items 5 – 7 (the credit
management subscale) might not be applicable to lower
SES individuals, individuals with low credit scores, and
those who do not have any lines of credit/loans. These par-
ticipants may face restricted access to credit or may simply
not choose to use it. Items 9 – 12, and perhaps 15, likely
do not apply to the majority of retirees. If some questions
are not asked of some clients or research participants, the
FMBS score might need to be standardized rst so that
all participants are on the same scale. Further, reliabil-
ity analyses particular to new samples would need to be
undertaken. Future studies might be able to utilize compu-
ter technology to only give participants questions that are
relevant to them based on their age or life circumstances.
An alternative possibility might be to develop different
versions of the FMBS that reect “sound nancial man-
agement behaviors” at different periods in the life course.
A second renement would recognize the fact that the
FMBS measures behaviors that are considered “sound” by
middle and upper-middle class individuals. For exam-
ple, although the share of individuals with money in the
stock market has greatly increased over the past three
decades, lower SES individuals are still more reluctant to
put their money into assets that carry market risk (Gar-
asky et al., 2008). Further, some of the behaviors that the
FMBS measures may be difcult for working class and
lower SES individuals to implement, even if they desire
to do so. For example, it may be much easier for mid-
dle and upper-middle class individuals to maintain health
and life insurance policies because they have more access
to employer sponsored plans. Although the FMBS does
seem to work for low-income participants in this sample,
additional work with a new sample could replicate this
nding. Further, this study was unable to test the psycho-
metric properties of race/ethnic minority individuals due
to sample size limitations.
This issue also has implications for the use of the scale.
Researchers should remember the structural difculties
that some groups face with regard to economic opportunity
when studying race or class differences using the FMBS.
We do not intend to offer the FMBS as some sort of
checklist to which all individuals should aspire. Rather,
the
FMBS is intended as a brief scale that may help researchers
easily measure nancial management behaviors and practi-
tioners to quickly assess their clients’ nancial habits. This
may also have implications for the psychometric properties
of the scale. Although the scale demonstrates reliability and
validity in a national sample, this might not be the case for
certain subgroups. In future work we plan to examine the
FMBS more closely with respect to race/ethnicity.
These needed renements speak to more fundamental is-
sues in the consumer nance and nancial planning elds.
These elds have a strong stance on what “sound” nan-
cial management entails. The reasoning behind promot-
ing these behaviors is that they facilitate clients reaching
their goals. Research has also shown that these behaviors
promote better physical, emotional, and relationship health
(Dew & Xiao, 2010, Drentea, 2000). Thus, the nancial
counseling and planning elds have good reason to label
these behaviors as “sound.”
In spite of this, however, group differences related to -
nancial behavior do exist. For example, even after account-
ing for various demographic characteristics and nancial
behaviors, African-Americans and Hispanic-Americans
have a harder time obtaining credit which may suggest
continued race/ethnic discrimination (Hanna & Linda-
mood, 2007). Other studies have demonstrated cultural
differences in nancial attitudes and beliefs based on race/
ethnic, class, and even religious lines (Marks, Dollahite, &
Dew, 2009; Grable et al., 2009). Thus, nancial counseling
and planning as well as consumer nance and economics
elds may benet as they continue the nascent dialogue
on cultural competence. Understanding why individuals
and groups may not always adopt “sound” nancial man-
agement behaviors may help practitioners working with
diverse groups.
Although this study has limitations and the FMBS needs
more renement, it makes a contribution to the nancial
planning and counseling literature – a multi-dimensional
scale of nancial management behavior validated using
nationally representative data. The FMBS covers many
dimensions of nancial management behaviors and pos-
sesses desirable psychometric properties. The full scale
offers researchers and practitioners a reliable and valid tool
to measure nancial management behavior, though some
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011
56
work needs to be done to ensure that it is reliable and valid
among different subgroups. The subscales – though some
needing more work – measure different dimensions of
nancial management behaviors in more detail. Depend-
ing on the needs of researchers and practitioners, either
the whole scale or subscales can be used in research and
counseling projects. Researchers and practitioners may
freely use the revised FMBS as long as this study is cited
when it is used.
References
Ayanian, J. Z., Weissman, J. S., Schneider, E. C., Gins-
burg, J. A., & Zaslavsky, A. M. (2000). Unmet
health needs of uninsured adults in the United States.
Journal of the American Medical Association, 284,
2061-2069.
Bernstein, D. (2004). Household debt and IRAs: Evidence
from the survey of consumer nances. Financial
Counseling and Planning, 15, 63-72.
Cattell, R. B. (Ed.). (1966). Handbook of Multivariate Ex-
perimental Psychology. Chicago: Rand McNally.
Cohen, R. J., & Swerklik, M. E. (1999). Psychological
testing and assessment: An introduction to tests and
measurment (4th ed.). Mountain View, CA: Mayeld
Publishing.
Comrey, A. L., & Lee, H. B. (1992). A rst course in fac-
tor analysis. Hillsdale, NJ: Lawrence Erlbaum As-
sociates.
Davis, E. P. (1992). Financial management practices
among households with differing resource con-
straints. Journal of Consumer Education, 10, 27-31.
Davis, E. P., & Weber, J. A. (1990). Patterns and obstacles
to nancial management. Financial Counseling and
Planning, 1, 41-51.
DeNavas-Walt, C., Proctor, B. D., & Lee, C. H. (2006).
Income, poverty, and health insurance coverage in
the United States: 2005. Retrieved from http://www.
census.gov/prod/2006pubs/p60-231.pdf
DeNavas-Walt, C., Proctor, B. D., & Smith, J. C. (2010).
Income, poverty, and health insurance coverage in
the United States: 2009. Retrieved from http://www.
census.gov/prod/2010pubs/p60-238.pdf
Dennis, M., & McCready, B. (2009). Sample eld report:
Familial response to nancial instability. Menlo
Park, CA: Knowledge Networks.
Dew, J. P. (2007). Two sides of the same coin? The differ-
ing roles of assets and consumer debt in marriage.
Journal of Family and Economic Issues, 28, 89-104.
Dew, J. P. & Xiao, J. J. (April, 2010). The association be-
tween family nancial declines, nancial behaviors,
and relationship satisfaction during the recession. Pa-
per presented at the annual conference of the Ameri-
can Council of Consumer Interests, Atlanta, GA.
Domowitz, I., & Sartain, R. L. (1999). Determinants of
the consumer bankruptcy decision. The Journal of
Finance, 54, 403-420.
Drentea, P. (2000). Age, debt, and anxiety. Journal of
Health and Social Behavior, 41, 437-450.
Fitzsimmons, V. S., Hira, T. K., Bauer, J. W., & Hafstrom,
J. L. (1993). Financial management: Development of
scales. Journal of Family and Economic Issues, 14,
257-274.
Furnham, A. (1984). Many sides of the coin: The psychol-
ogy of money usage. Personality and Individual Dif-
ferences, 5, 501-509.
Garasky, S., Nielsen, R. B., & Fletcher, C. N. (2008).
Consumer nances of low-income families. In J. J.
Xiao (Ed.), Advances in consumer nance behavior
research (pp. 223-237). New York: Springer.
Godwin, D. D., & Koonce, J. C. (1992). Cash ow man-
agement of low-income newlyweds. Financial
Counslineg and Planning, 3, 17-43.
Grable, J. E., Park, J. Y., & Joo, S. H. (2009). Explaining
nancial management behavior for Koreans living in
the United States. The Journal of Consumer Affairs,
43, 80-107.
Hanna, S. D., & Lindamood, S. (2007). Credit constraints:
The role of race/ethnic group. Consumer Interests
Annual, 53.
Hilgert, M. A., Hogarth, J. M., & Beverly, S. G. (2003).
Household nancial management: The connection
between knowledge and behavior. Federal Reserve
Bulletin, 89, 309-322.
Joo, S. (2008). Personal nancial wellness. In J. J. Xiao
(Ed.), Advances in consumer nance behavior re-
search (pp. 21-33). New York: Springer.
Jorgensen, B. L. (2007). Financial literacy of college
students: Parental and peer inuences. Unpublished
Masters Thesis, The Virginia Polytechnic Institute
and State University.
Kaplan, R. M., & Saccuzzo, D. P. (2009). Psychological
testing: principles, applications, and issues (7th ed.).
Belmont, CA: Wadsworth.
Kim, J., Garman, E. T., & Sorhaindo, B. (2003). Relation-
ship among credit counseling clients’ nancial well-
being, nancial behaviors, nancial stressor events,
and health. Financial Counseling and Planning, 14,
75-87.
Marks, L., Dollahite, D. C., & Dew, J. P. (2009). Enhanc-
ing cultural competence in nancial counseling and
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011 57
planning: Understanding why families make reli-
gious contributions, Journal of Financial Counseling
and Planning, 20(2), 14-26.
Mugenda, O. M., Hira, T. K., Fanslow, A. M. (1990). As-
sessing the causal relationship among communica-
tion, money management practices, satisfaction with
nancial status, and satisfaction with quality of life.
Lifestyles:Family and Economic Issues, 11, 343-360.
Perry, V. G., & Morris, M. D. (2005). Who is in control?
The role of self-perception, knowledge, and income
in explaining consumer nancial behavior. Journal
of Consumer Affairs, 39, 299-313.
Prochaska-Cue, K. (1993). An exploratory study for a
model of personal nancial management style.
Journal of Financial Counseling and Planning,
4, 111-134.
Rosen, D. L., & Granbois, D. H. (1983). Determinants of
role structure in family nancial management. The
Journal of Consumer Research, 10, 253-258.
Rubin, D. B. (1987). Multiple imputation for non-response
in surveys. New York: Wiley.
Scannell, E. (1990). Dairy farm families’ nancial man-
agement. Financial Counseling and Planning, 1,
133-146.
Servon, L. J., & Kaestner, R. (2008). Consumer nancial
literacy and the impact of online banking on the
nancial behavior of lower-income bank customers.
Journal of Consumer Affairs, 42, 271-305.
Short, P. F., & Graefe, D. R. (2003). Battery-powered
health insurance? Stability in the coverage of the
uninsured. Health Affairs, 22, 244-255.
Silva, F. (1993). Psychometric foundations and behavioral
assessment. Newbury Park, CA: Sage.
Sullivan, A. C. (1987). Economic factors associated with
delinquency rates on consumer installment debt.
Credit Research Center: Working Paper Series,
Working Paper # 55. Retreived from http://faculty.
msb.edu/prog/CRC/pdf/wp55.pdf
Tang, T. L. P. (1995). The development of a short money
ethic scale. Attitudes toward money and pay satis-
faction revisited. Personality and Individual Differ-
ences, 19, 809-816.
U. S. Census. (2010). Table H-1. Income limits for each
fth and top 5 percent of all households: 1967 to
2009. Retrieved from http://www.census.gov/hhes/
www/income/data/historical/inequality/H01AR_
2009.xls
Xiao, J. J. (2008). Applying behavior theories to nancial
behavior. In J. J. Xiao (ed.), Handbook of consumer
nance research (pp. 69-82). New York: Springer.
Xiao, J. J., Shim, S., Barber, B., & Lyons, A. (2008).
Academic success and well-being of college stu-
dents: Financial behaviors matter. Tucson, AZ: Take
Charge America Institute for Consumer Financial
Education and Research. Retrieved from http://www.
cefe.illinois.edu/research/reports/Academic Success
and Well-Being of College Students_112007.pdf
Xiao, J. J., Tang, C., & Shim, S. (2009). Acting for happi-
ness: Financial behavior and life satisfaction of col-
lege students. Social Indicators Research, 92, 53-68.
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011
58
Initial FMBS
Please indicate how often you have engaged in the follow-
ing activities in the past six months:
1 = never, 2 = seldom, 3 = sometimes, 4 = often,
5 = always (Also could say “Not Applicable (N/A)”)
1. Comparison shopped when purchasing a product
or service
2. Bought something on impulse
3. Searched for information about a big-ticket item before
purchasing it
4. Paid all your bills on time
5. Kept a written or electronic record of your monthly
expenses
6. Stayed within your budget or spending plan
7. Paid off credit card balance in full each month
8. Maxed out the limit on one or more credit cards
9. Made only minimum payments on a loan
10. Began or maintained an emergency savings fund
11. Saved money from every paycheck
12. Saved for a long term goal such as a car, education,
home, etc.
13. Contributed money to a retirement account
14. Bought bonds, stocks, or mutual funds
Please rate your behavior regarding insurance within the
past year on a scale of 1 – 5.
1 = Never, 2 = seldom, 3 = sometimes 4 = often, 5 = always
15. Maintained or purchased an adequate health insurance
policy
16. Maintained or purchased adequate property insurance
like auto or homeowners insurance
17. Maintained or purchased adequate life insurance
Revised FMBS
Please indicate how often you have engaged in the follow-
ing activities in the past six months:
1 = never, 2 = seldom, 3 = sometimes, 4 = often, 5 = always
1. Comparison shopped when purchasing a product or
service
2. Paid all your bills on time
3. Kept a written or electronic record of your monthly
expenses
4. Stayed within your budget or spending plan
5. Paid off credit card balance in full each month
6. Maxed out the limit on one or more credit cards
7. Made only minimum payments on a loan
8. Began or maintained an emergency savings fund
9. Saved money from every paycheck
10. Saved for a long term goal such as a car, education,
home, etc.
11. Contributed money to a retirement account
12. Bought bonds, stocks, or mutual funds
Please rate your behavior regarding insurance within the
past year on a scale of 1 – 5:
1 = Never, 2 = seldom, 3 = sometimes, 4 = often, 5 = always.
13. Maintained or purchased an adequate health insurance
policy
14. Maintained or purchased adequate property insurance
like auto or homeowners insurance
15. Maintained or purchased adequate life insurance
Appendix A. FMBS as used in the Family Response to Financial Instability Study
Journal of Financial Counseling and Planning Volume 22, Issue 1 2011 59
Appendix B. Detailed Income, Asset, and Debt Tables
Scale values %
Income 1 – Less than $5,000
2 – $5,000 to $7,499
3 – $7,500 to $9,999
4 – $10,000 to $12,499
5 – $12,500 to $14,999
6 – $15,000 to $19,999
7 – $20,000 to $24,999
8 – $25,000 to $29,999
9 – $30,000 to $34,999
10 – $35,000 to $39,999
11 – $40,000 to $49,999
12 – $50,000 to $59,999
13 – $60,000 to $74,999
14 – $75,000 to $84,999
15 – $85,000 to $99,999
16 – $100,000 to $124,999
17 – $125,000 to $149,999
18 – $150,000 to $174,999
19 – $175,000 or more
1.5
2.1
1.8
3.7
2.0
4.3
5.6
5.7
6.1
6.8
9.1
11.3
13.2
7.1
6.4
7.3
3.0
1.8
1.3
Actual amount of savings 1 – None
2 – $1 to under $1,500
3 – $1,500 to under $3000
4 – $3,000 to under $5,000
5 – $5,000 to under $10,000
6 – $10,000 to under $20,000
7 – $20,000 to under $50,000
8 – $50,000 to under $100,000
9 – $100,000 or more.
20.5
16.9
7.6
6.7
8.9
9.6
10.9
7.0
11.9
Actual amount of consumer debt 1 – None
2 – $1 to under $1,500
3 – $1,500 to under $3000
4 – $3,000 to under $5,000
5 – $5,000 to under $10,000
6 – $10,000 to under $20,000
7 – $20,000 to under $50,000
8 – $50,000 to under $100,000
9 – $100,000 or more.
30.3
16.9
9.1
7.4
10.5
12.3
8.9
2.9
1.8
Note. Not all values sum to 100% due to rounding.
