Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

Journal of Education for Business

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/vjeb20

A latent profile analysis of college students'


financial knowledge: The role of financial
education, financial well-being, and financial risk
tolerance

Abed G. Rabbani, Wookjae Heo & Jae Min Lee

To cite this article: Abed G. Rabbani, Wookjae Heo & Jae Min Lee (2021): A latent profile analysis
of college students' financial knowledge: The role of financial education, financial well-being, and
financial risk tolerance, Journal of Education for Business, DOI: 10.1080/08832323.2021.1895046

To link to this article: https://doi.org/10.1080/08832323.2021.1895046

Published online: 12 Mar 2021.

Submit your article to this journal

Article views: 63

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at


https://www.tandfonline.com/action/journalInformation?journalCode=vjeb20
JOURNAL OF EDUCATION FOR BUSINESS
https://doi.org/10.1080/08832323.2021.1895046

A latent profile analysis of college students’ financial knowledge: The role of


financial education, financial well-being, and financial risk tolerance
Abed G. Rabbania , Wookjae Heob , and Jae Min Leec
a
Department of Personal Financial Planning, University of Missouri, Columbia, MO, USA; bDivision of Health and Consumer Sciences,
South Dakota State University, Brookings, SD, USA; cDepartment of Family Consumer Science, Minnesota State University - Mankato,
Mankato, MN, USA

ABSTRACT KEYWORDS
This research aims to classify college students into meaningful groups based on their sub- Financial education;
jective financial knowledge (SFK), objective financial knowledge (OFK), and comparative financial knowledge;
financial knowledge (CFK) using Latent Profile Analysis (LPA). LPA captures underlying financial risk tolerance;
financial well-being; latent
dynamics within the grouping variables and compares group differences. Results of LPA profile analysis
suggest three groups of college students in the sample, and the groups differed in levels of
financial well-being (FWB), financial risk tolerance (FRT), and financial education experi-
ence (FEE).

Introduction (i.e., heterogeneity). This approach assumes a latent


relationship among the variables and forms a latent
Financial knowledge (FK), a key variable of financial
variable to classify the sample.
well-being, is often entered into a classification model
In this study, FK is conceived as a multi-dimen-
using a dichotomization method where researchers
sional (OFK, SFK, and CFK) construct in its types
use a mean value of FK scores to divide respondents
and levels between observations. It is assumed that
into groups (e.g., Agnew & Szykman, 2005; Allgood &
there would be a great deal of differences in levels of
Walstad, 2016; Bannier & Neubert, 2016). This classi-
FK by its type among college students. The differences
fication method depends on the distribution of one
can also be associated with other characteristics, such
type of FK and does not account for interactions of
as financial education experience (FEE), financial
multiple dimensions of FK—objective financial know-
ledge (OFK), subjective financial knowledge (SFK), well-being (FWB), and financial risk tolerance (FRT).
comparative financial knowledge (CFK). Therefore, the primary goals of this study are (a) to
Previous literature using the dichotomization of a classify respondents based on their OFK, SFK, and
FK variable for classification purposes has often adopted CFK; (b) to examine how FK is associated with FWB,
a variable-centered approach. A variable-centered FRT, and FEE.
approach (e.g., regression analysis, factor analysis, struc-
tural equation modeling) assumes a one-dimensional Literature review
aspect of the relationship between variables (e.g., one-
Dimensions of financial knowledge
on-one relationship between variables rather than mul-
tiple or multiway relationships) assuming that a sample Financial knowledge (FK) is defined as understanding
is homogeneous (Laursen & Hoff, 2006). Several other key financial terms and concepts needed to function
studies (e.g., Marsh, L€ udtke, Trautwein, & Morin, 2009; daily as an independent financial decision-maker
Muthen & Muthen, 2000) suggest that person-centered (Bowen, 2002). The FK measures used in previous
analytic tools (e.g., latent profile analysis, finite mixture studies fall into three categories: OFK, SFK, and CFK.
analysis, latent transition analysis) can be useful to iden- Research in decision-making finds that the association
tify the unobserved relationships between variables and between knowledge and behavior varies not only by
detect their dynamics on the group classification the OFK but also by SFK—self-assessed understanding

CONTACT Abed G. Rabbani rabbania@missouri.edu Department of Personal Financial Planning, 239B Stanley Hall, University of Missouri, Columbia
65211, MO, USA
ß 2021 Taylor & Francis Group, LLC
2 A. G. RABBANI ET AL.

of content knowledge (e.g., Gutter, Copur, & (2016) have followed this approach using a two-part
Garrison, 2010; Tang & Baker, 2016), and CFK—how measure of FK to investigate the likely effects of FK
one rates one’s knowledge relative to that of friends on a broad range of financial behaviors.
(e.g., Agnew & Szykman, 2005; Shim, Xiao, Barber, & Although easy to implement, there are several
Lyons, 2009; Xiao, Tang, Serido, & Shim, 2011). known problems with dichotomization procedures
Studies of SFK and OFK have long been the focus (MacCallum et al., 2002). These problems include loss
of consumer research (Alba & Hutchinson, 2000; of information about individual differences; loss of
Carlson, Vincent, Hardesty, & Bearden, 2009; effect size and power, increased Type I error; increased
Moorman, Diehl, Brinberg, & Kidwell, 2004). Type II error; or spurious statistical significance and
However, there has been a recent surge in interest overestimation of effect size in the case of analyses
regarding the FK of college students in relation to with two independent variables; the potential to over-
their peers (Gutter et al., 2010; Lusardi, Mitchell, & look nonlinear relationships; and loss of measurement
Curto, 2010; Robb & Woodyard, 2011) and how that reliability (MacCallum et al., 2002; McClelland, Lynch,
relative knowledge may influence financial decisions Irwin, Spiller, & Fitzsimons, 2015; Rucker, McShane, &
(Agnew & Szykman, 2005). The findings suggest that Preacher, 2015). The methodological literature demon-
investigating the associations between types of FK and strates that the dichotomization procedure’s weaknesses
financial behavior requires more complex modeling may be handled by alternative analytic models such as
(Xiao, Ahn, Serido, & Shim, 2014) that incorporates latent profile analysis (LPA) (Pastor, Barron, Miller, &
all three aspects of FK—OFK, SFK, and CFK. In other Davis, 2007). LPA technically refers to the unobserved
words, there can be a heterogeneity of people caused groups of individuals using a latent variable formed
by the dynamics from three aspects of FK. based on input variables in the model.

Financial knowledge-based classification Data, variables, and methodology


When there is a research need to classify participants Study design
based on FK, many researchers rely on the dichotomiza- Data were collected at a university in the Midwest,
tion (Agnew & Szykman, 2005; Allgood & Walstad, over two weeks during fall 2018, from college students
2016; Bannier & Neubert, 2016). This simple classifica- who were 18 to 21 years old using an online survey
tion begins by converting the FK variable into a dichot- technique. After receiving Institutional Review Board
omous variable by splitting the scale at some point and (IRB) approval, a random sample of 3,000 students
designating individuals above and below that point as (approximately 10.0% of the total student population)
defining two separate groups described as "Low" and was selected from the University’s database to receive
"High." For example, a median split is a common an email message for the Qualitrics survey. The survey
approach to split the scale at the sample median, thereby collected data on demographic information, FK, FWB,
defining high and low groups on the variable differs by and FRT. All respondents were offered a nominal
research, and there is no universal consent about the cri- incentive for their participation. A total of 391 stu-
teria for the right cutoff point. Alternatively, the scale dents participated; the response rate was 13%. This
might be split at some other point based on the data response rate is consistent with other online surveys
(e.g., one standard deviation above the mean) or at a of college students (e.g., Shim et al., 2009). However,
fixed point on the scale designated a priori (MacCallum, among the participants, only 241 were valid with a
Zhang, Preacher, & Rucker, 2002). completed survey. Of the valid responses, Table 2
As FK-based classifications rely mostly on OFK, presents detailed information about the demographic
researchers often use the mean of the OFK scores as profile of the participants. This respondent profile was
an indicator of average FK (Agnew & Szykman, 2005; similar to that of other studies investigating college
Allgood & Walstad, 2016; Bannier & Neubert, 2016; students enrolled at Midwestern U.S. universities
Beal & Delpachitra, 2003; Chen & Volpe, 1998). Based (Archuleta, Dale, & Spann, 2013; Heckman & Grable,
on this approach, researchers divide the respondents 2011; Peralta, Steele, Nofziger, & Rickles, 2010).
into two knowledge groups, high knowledge (above
the mean) and low knowledge (below the mean).
Variables
When there are OFK and SFK variables, some
researchers often dichotomize both. For example, The main criteria variables to make significant groups
Allgood and Walstad (2016) and Bannier and Neubert were subjective financial knowledge (SFK),
JOURNAL OF EDUCATION FOR BUSINESS 3

comparative financial knowledge (CFK), and objective Criteria (BIC) and Entropy criteria. In this study, differ-
financial knowledge (OFK). The SFK question was, ent models with a varying number of clusters ranging
"On a scale from one to five, how would you rate your from 2 to 5 were explored, and a combination of indices
overall understanding of financial knowledge of various and criteria were used to determine the model with the
financial topics?” (5 ¼ extremely knowledgeable; optimal number of clusters. The findings of LPA were
1 ¼ not at all knowledgeable). The question for the also compared with groups derived from the dichotomi-
CFK was, "On a scale from one to five, how would you zation approach.
compare yourself to your close friends regarding your
overall knowledge of various financial topics?”
Results
(5 ¼ extremely knowledgeable; 1 ¼ not at all know-
ledgeable). The OFK was measured with 28 items Table 1 shows the dichotomized subgroups’ FEE and
introduced by Hilgert, Hogarth, and Beverly (2003). the levels of FK. The dichotomization methods divide
There were 28 dichotomous questions (Yes/No). The total samples into subgroups based on the mean value
Cronbach’s alpha associated with the measure of OFK of each input variable: two groups by SFK, two groups
was .50, which is low because 28 items used are all by OFK, and two groups by CFK. Dichotomization
yes/no questions. Standardized scores from these three methods can suggest different group membership of
scales were used as latent profile criteria to classify each observation based on each type of FK. However,
241 respondents into meaningful groups. the other control characteristics (i.e., FWB, FRT, and
The following psychological features and educa- FEE) did not bring defining features for groups.
tional factors were included as external controls: level However, as shown in Table 2, LPA can suggest
of financial well-being (FWB), financial risk tolerance multiple group classification using more than two
(FRT), and whether they had taken personal finance input variables. Demographic, FK, FWB, FRT, FEE
courses at a high school, and whether they had taken profiles varied by three FK groups developed using
personal finance courses at college (FEE). FWB was LPA. Those who were likely to belong to Group A
measured using 10-items developed by the Consumer tended to be females, white juniors who had the low-
Financial Protection Bureau (CFPB). FRT was meas- est FWB, and FRT scores. Those who were likely to
ured using Grable and Lytton’s 13-items (Kuzniak, belong to Group B tended to be females, juniors, and
Rabbani, Heo, Ruiz-Menjivar, & Grable, 2015). The predominately white. However, on average, they had
Cronbach’s alphas associated with FWB and FRT higher scores on FWB than those of Group A but
were .84 and .69 respectively, which are within the lower than those of Group C. Group C was the small-
acceptable range of .65 and .80. est group composed of male sophomores with the
highest level of FWB and FRT score.
As shown in Table 3, each type of FK was signifi-
Latent profile analysis (LPA)
cantly correlated. The results imply that each type of
This study used LPA to capture the potential group FK measures a unique but interrelated aspect of FK.
differences in terms of three aspects of FK and esti- The total sample can be divided into subgroups with
mate the relationship with FWB, FRT, and FEE. LPA a different distribution of FK, considering potential
is a latent group modeling technique, also known as a heterogeneity of observations in response to these var-
latent class analysis (Vermunt & Magidson, 2002) and iables individually. Based on the model fit (AIC, BIC,
finite mixture modeling (Peel & McLachlan, 2000), and entropy), a three-group model was found more
which is useful to handle data with potential multiple optimal than two group model (See Table 4).
subgroups (Hair, Sarstedt, Ringle, & Gudergan, 2018). Table 5 shows the parameter estimates (log-odds
Specifically, each type of FK (i.e., SFK, OFK, and ratio) of three profile indicator variables (the three types
CFK) is a unique construct but can interrelate with of FK) for the three profiles, along with the class mar-
each other; thus, identifying profiles of subgroups and gins and standard errors. It means that respondents in
capturing complex interaction requires an estimation each group had a similar tendency in the effect of FK.
of the distribution of each FK. For instance, 120 respondents in Group A showed a sig-
LPA is a probabilistic model-based clustering nificantly negative direction on all FK scores. Thus, 120
approach that offers a sound mechanism of estimating respondents with low level of FK (log-odds ratio of OFK
the optimal number of clusters (Leisch, 2004; Olaru, ¼ 0.31, p < .01; log-odds ratio of SFK ¼ 0.57, p <
Smith, & Taplin, 2011) using selection criteria, such as .001; log-odds ratio of CFK ¼ 0.73, p < .001) were
Akaike Information Criteria (AIC), Bayesian Information more likely to belong to Group A. The marginal
4 A. G. RABBANI ET AL.

Table 1. Dichotomization comparison by Mean.


High vs. Low
Low High
Median (S.D.)/ Mean/Proportion
Number (%) n Mean (S.D.)/Number (%) n Mean (S.D.)/Number (%) comparison (t)
Grouping criteria
OFK 19.73 (2.94) 111 17.21 (2.04) 130 21.89 (1.52) 20.37
Control variable
FWB 52.66 (9.00) 111 52.38 (8.71) 130 52.91 (9.27) 0.45
FRT 23.77 (4.46) 111 23.46 (4.24) 130 24.04 (4.64) 1.00
High School Fin. 163 (67.63 %) 111 64 (57.66%) 130 99 (76.15%) 3.11
College Fin. 53 (21.99%) 111 20 (18.02%) 130 34 (26.15%) 1.51
Grouping criteria
SFK 2.41 (0.82) 130 1.77 (0.42) 111 3.17 (0.42) 25.63
Control variable
FWB 52.66 (9.00) 130 52.10 (8.81) 111 53.32 (9.22) 1.05
FRT 23.77 (4.46) 130 23.33 (4.34) 111 24.29 (4.56) 1.67
High School Fin. 163 (67.63 %) 130 81 (62.31%) 111 82 (73.87%) 1.92
College Fin. 53 (21.99%) 130 18 (13.85%) 111 36 (32.43%) 3.52
Grouping criteria
CFK 3.09 (0.96) 155 2.50 (0.62) 86 4.15 (0.36) 22.68
Control variable
FWB 52.66 (9.00) 155 51.68 (8.56) 86 54.44 (9.55) 2.30
FRT 23.77 (4.46) 155 23.64 (4.56) 86 24.01 (4.29) 0.62
High School Fin. 163 (67.63 %) 155 93 (60.00%) 86 70 (81.40%) 3.47
College Fin. 53 (21.99%) 155 29 (18.71%) 86 25 (29.07%) 1.85
p < .05; p < .01; p < .001.

Table 2. Descriptive table of total and subsamples.


LPA model
Total Group A Group B Group C
(n ¼ 241) (n ¼ 120) (n ¼ 99) (n ¼ 21)
Categorical Variables Number (%) Number (%) Number (%) Number (%)
Male 49 (20.42%) 19 (15.83%) 19 (19.19%) 11 (52.38%)
College rank
Freshman 53 (21.99%) 26 (21.49%) 27 (27.27%) 0 (0%)
Sophomore 58 (24.07%) 27 (22.31%) 20 (20.20%) 11 (52.38%)
Junior 77 (31.95%) 40 (33.06%) 33 (33.33%) 4 (19.05%)
Senior 49 (20.33%) 25 (20.66%) 18 (18.18%) 6 (28.57%)
Other 4 (1.66%) 3 (2.48%) 1 (1.01%) 0 (0%)
Race
White 204 (84.65%) 98 (80.99%) 88 (88.89%) 18 (85.71%)
African American 12 (4.98%) 9 (7.44%) 3 (3.03%) 0 (0%)
Hispanic 6 (2.49%) 5 (4.13%) 0 (0%) 1 (4.76%)
Asian 14 (5.81%) 6 (4.96%) 6 (6.06%) 2 (9.52%)
Other 5 (2.07%) 3 (2.48%) 2 (2.02%) 0 (0%)
Taking finance at high school (Yes ¼2) 163 (67.63 %) 66 (55%) 77 (77.77%) 19 (90.48%)
Taking finance at college (Yes ¼ 2) 53 (21.99%) 11 (9.17%) 32 (32.32%) 10 (47.62%)
Age
18 40 (16.60%) 18 (14.88%) 21 (21.21%) 1 (4.76%)
19 56 (23.24%) 32 (26.45%) 20 (20.20%) 4 (19.05%)
20 74 (30.71%) 36 (29.75%) 29 (29.29%) 9 (42.86%)
21 71 (29.46%) 35 (28.93%) 29 (29.29%) 7 (33.33%)
Continuous Variables Mean (S.D.) Mean (S.D.) Mean (S.D.) Mean (S.D.)
OFK 19.73 (2.94) 18.74 (2.82) 20.62 (3.09) 21.29 (3.16)
SFK 2.41 (0.82) 1.93 (0.68) 2.75 (0.52) 3.62 (0.67)
CFK 3.09 (0.96) 2.37 (0.65) 3.65 (0.52) 4.57 (0.51)
FWB 52.66 (9.00) 51.42 (8.99) 52.12 (8.03) 62.38 (7.86)
FRT 23.77 (4.46) 23.41 (4.53) 23.52 (3.91) 27.05 (5.31)

probability to belong to Group A was 48%, with a 7% levels of SFK and CFK (log-odds ratio of SFK ¼ 0.32,
standard error. This means that 48% of homogeneous p < .05; log-odds ratio of CFK ¼ 0.49, p < .01) were
respondents would have the probability of belonging more likely to belong to Group B. The marginal prob-
to Group A. ability of belonging to Group B was 43%, with a 7%
Compared to Group A, 99 respondents in Group B standard error. This means that 43% of similar
had significantly positive direction on SFK and CFK. respondents with this dataset would show the prob-
This result means that 99 respondents with higher ability of belonging to Group B.
JOURNAL OF EDUCATION FOR BUSINESS 5

Finally, Group C respondents (n ¼ 21) showed a person-centered approach to explore differences in


significantly positive direction on all FK types. It FK-based classes by incorporating potential interac-
means that 21 respondents with the highest level of tions between types of FK.
FK in all three types (log-odds ratio of OFK ¼ 0.55, p The three LPA groups are distinctive in terms of
< .05; a log-odds ratio of SFK ¼ 1.36, p <.001; a log- FWB, FRT, and FEE; thus, it is possible to describe
odds ratio of CFK ¼ 1.39, p < .001) were more likely between-group similarities and differences. For
to belong to Group C. The marginal probability of example, Group C students reported the highest level
belonging to Group C was 10%, with a 4% stand- of all three FK scores than the other two profiles. This
ard error. group was more likely to be male students who have
For control variables, Group A was used as the ref- FEE in high school and college and score high on
erence group. There were no significant differences in FWB and FRT measures.
FWB and FRT between Group A and B (Table 5). Group B was the second-largest category of stu-
However, positive and significant coefficients of taking dents whose levels of SFK and CFK were higher than
personal finance courses in high school and colleges Group A but lower than Group C. This profile repre-
indicate that those who belong to Group B were more sented about 40% of the students. The students in this
likely to take a personal finance course in their high group were more likely to have FEE in high school
and college than Group A but less likely than Group
school and college (b ¼ 1.17, p < .01; b ¼ 1.20, p < .05
C. However, these students did not significantly differ
respectively) than Group A.
in FWB and FRT than Group A.
For the comparison between Group A and C, FWB
Group A reported the lowest level of all three FK
and FRT showed significant differences. The coeffi-
scores. This profile comprised about half of the stu-
cients of FWB (b ¼ 0.16, p < .01) and FRT (b ¼ 0.81,
dents—indicating that the majority appear to have a
p < .05) in Group C indicated that respondents in
low FK level. It is also the most substantial profile.
Group C were more likely to have a higher number of
The students in this profile also reported the lowest
FWB scores and FRT scores than Group A.
FWB and FRT levels and were less likely to take FEE
Additionally, taking personal finance courses in high in high school and college than these other groups.
school and colleges differed among these groups. There was a higher likelihood of having FEE in high
When respondents belonged to Group C, they were school for these students than in college.
more likely to have taken a personal finance course in Findings from between-group similarities and dif-
high school and college (b ¼ 2.59, p < .05; b ¼ 2.28, ferences show how closely FEE, FK, FWB, and FRT
p < .05 respectively). were related. It suggests that increasing access to FEE
(e.g., in business education) is likely to foster a posi-
Discussion tive synergy among young adults’ FK, FRT, and FWB.
Hite, Slocombe, Railsback, and Miller (2011) acknowl-
This study tested the importance of three FK dimen- edged a widespread need for additional financial liter-
sions to capture latent group differences (i.e., hetero- acy education.
geneity of samples). It also examined the role of
FEE, FWB, and FRT as predictors of each suggested
group from the LPA classification method. This Limitations
study extends the current literature by adopting a There are several limitations to this study. Firstly, the
participants were self-selected. The current study
Table 3. Correlation among three financial know- relied on college students enrolled in a university in
ledge variables.
the Midwest; it is imperative to be cautious before
OFK SFK CFK
generalizing the results. Secondly, although many of
OFK 1.00
SFK 0.13 1.00 the participants of the survey were assumed to have
CFK 0.23 0.51 1.00 lived and finished school in one state in which the
p < .05; p < .001.
state-specific characteristics, such as the state’s

Table 4. LPA comparison.


Observations Log-Likelihood Degree of Freedom AIC BIC Entropy
2 Groups 241 972.02 14 1972.04 2020.83 0.679
3 Groups 241 961.46 22 1966.92 2043.59 0.717
In terms of 4 or more groups, log-likelihood did not converge.
6 A. G. RABBANI ET AL.

Table 5. LPA clustering results. References


LPA model
Agnew, J. R., & Szykman, L. R. (2005). Asset allocation and
Group A Group B Group C information overload: The influence of information dis-
(n ¼ 120) (n ¼ 99) (n ¼ 21)
play, asset choice, and investor experience. Journal of
Grouping criteria
Behavioral Finance, 6(2), 57–70. doi:10.1207/
OFK 0.31 0.22 0.55
SFK 0.57 0.32 1.36 s15427579jpfm0602_2
CFK 0.73 0.49 1.39 Alba, J. W., & Hutchinson, J. W. (2000). Knowledge calibra-
Control variable tion: What consumers know and what they think they
FWB – 0.01 0.16 know. Journal of Consumer Research, 27(2), 123–156. doi:
FRT – 0.02 0.81
Took personal finance (high school) – 1.17 2.59 10.1086/314317
Took personal finance (college) – 1.20 2.28 Allgood, S., & Walstad, W. B. (2016). The effects of per-
Class Margin 0.48 0.43 0.10 ceived and actual financial literacy on financial behaviors.
Standard Error 0.07 0.10 0.04 Economic Inquiry, 54(1), 675–697. doi:10.1111/ecin.12255
Reference: Group A. Archuleta, K. L., Dale, A., & Spann, S. M. (2013). College
p < .05; p < .01; p < .001.
Students and Financial Distress: Exploring Debt,
Financial Satisfaction, and Financial Anxiety. Journal of
educational system and policy (i.e., policy by Board of Financial Counseling and Planning, 24(2), 50–62.
Regents), are the same, the results may not fully Bannier, C. E., & Neubert, M. (2016). Gender differences in
reflect potential state-level and individual-level differ- financial risk taking: The role of financial literacy and
risk tolerance. Economics Letters, 145, 130–135. doi:10.
ences other than characteristics used in the study. 1016/j.econlet.2016.05.033
Finally, any details of financial education were not Beal, D., & Delpachitra, S. (2003). Financial literacy among
included in the analytic procedure because of the sur- Australian university students. Economic Papers: A
vey method’s limitation. However, the results bring a Journal of Applied Economics and Policy, 22(1), 65–78.
doi:10.1111/j.1759-3441.2003.tb00337.x
notable implication that financial education itself was
Bowen, C. F. (2002). Financial knowledge of teens and their
an important factor to consider. parents. Financial Counseling and Planning, 13(2),
93–102.
Carlson, J. P., Vincent, L. H., Hardesty, D. M., & Bearden,
W. O. (2009). Objective and subjective knowledge rela-
tionship: A quantitative analysis of consumer research
findings. Journal of Consumer Research, 35(5), 864–876.
Conclusion doi:10.1086/593688
Chen, H., & Volpe, R. P. (1998). An analysis of personal
Despite some limitations, results from the present financial literacy among college students. Financial
study offer methodological insights to study college Services Review, 7(2), 107–128. doi:10.1016/S1057-
students’ FK. When developing a college student pro- 0810(99)80006-7
file based on FK, LPA provides more information Gutter, M., Copur, Z., & Garrison, S. (2010). Financial
about inter-relationship among variables than dichoto- capabilities of college students from states with varying
financial education policies. Denver, CO: National
mization. LPA can suggest an extended perspective Endowment for Financial Education.
and exploration opportunity for the multi-dimensions Hair, J. F., Sarstedt, M., Ringle, C. M., & Gudergan, S. P.
of FK and group differences. Besides, the findings (2018). Advanced issues in partial least squares structural
highlight continued attention to financial literacy and equation modeling. Thousand Oaks, CA: Sage.
Heckman, S. J., & Grable, J. E. (2011). Testing the role of
systemic guidance in educational settings to extend
parental debt attitudes, student income, dependency sta-
financial education opportunities that can relate to tus, and financial knowledge have in shaping financial
later life’s FK and FWB. self efficacy among college students. College Student
Journal, 45, 51.
Hilgert, M. A., Hogarth, J. M., & Beverly, S. G. (2003).
Funding Household financial management: The connection
between knowledge and behavior. Federal Reserve
This work was supported by the University of Missouri
Bulletin, 89, 309.
under Margaret Mangel Research Catalyst Fund [F1713].
Hite, N. G., Slocombe, T. E., Railsback, B., & Miller, D.
(2011). Personal finance education in recessionary times.
Journal of Education for Business, 86(5), 253–257. doi:10.
1080/08832323.2010.511304
ORCID
Kuzniak, S., Rabbani, A., Heo, W., Ruiz-Menjivar, J., &
Abed G. Rabbani http://orcid.org/0000-0001-7647-8973 Grable, J. E. (2015). The Grable and Lytton risk-tolerance
Wookjae Heo http://orcid.org/0000-0001-5807-9792 scale: A 15-year retrospective. Financial Services Review,
Jae Min Lee http://orcid.org/0000-0002-0479-2282 24(2), 177–192.
JOURNAL OF EDUCATION FOR BUSINESS 7

Laursen, B., & Hoff, E. (2006). Person-centered and vari- achievement goal orientation. Contemporary Educational
able-centered approaches to longitudinal data. Merrill- Psychology, 32(1), 8–47. doi:10.1016/j.cedpsych.2006.10.003
Palmer Quarterly, 52(3), 377–389. doi:10.1353/mpq.2006. Peel, D., & McLachlan, G. J. (2000). Robust mixture model-
0029 ling using the t distribution. Statistics and Computing,
Leisch, F. (2004). FlexMix: A general framework for finite 10(4), 339–348. doi:10.1023/A:1008981510081
mixture models and latent class regression in R. Journal Peralta, R. L., Steele, J. L., Nofziger, S., & Rickles, M.
of Statistical Software, 11(8), 1–18. doi:10.18637/jss.v011. (2010). The impact of gender on binge drinking behavior
i08 among US college students attending a Midwestern uni-
Lusardi, A., Mitchell, O. S., & Curto, V. (2010). Financial versity: An analysis of two gender measures. Feminist
literacy among the young. Journal of Consumer Affairs,
Criminology, 5(4), 355–379. doi:10.1177/15570851
44(2), 358–380. doi:10.1111/j.1745-6606.2010.01173.x
MacCallum, R. C., Zhang, S., Preacher, K. J., & Rucker, 10386363
D. D. (2002). On the practice of dichotomization of Robb, C. A., & Woodyard, A. (2011). Financial knowledge
quantitative variables. Psychological Methods, 7(1), 19–40. and best practice behavior. Journal of Financial
doi:10.1037/1082-989X.7.1.19 Counseling and Planning, 22(1), 60–70.
Marsh, H. W., L€ udtke, O., Trautwein, U., & Morin, A. J. Rucker, D. D., McShane, B. B., & Preacher, K. J. (2015). A
(2009). Classical latent profile analysis of academic self- researcher’s guide to regression, discretization, and
concept dimensions: Synergy of person-and variable- median splits of continuous variables. Journal of
centered approaches to theoretical models of self-concept. Consumer Psychology, 25(4), 666–678. doi:10.1016/j.jcps.
Structural Equation Modeling: A Multidisciplinary 2015.04.004
Journal, 16(2), 191–225. doi:10.1080/10705510902751010 Shim, S., Xiao, J. J., Barber, B. L., & Lyons, A. C. (2009).
McClelland, G. H., Lynch, J. G., Jr, Irwin, J. R., Spiller, Pathways to life success: A conceptual model of financial
S. A., & Fitzsimons, G. J. (2015). Median splits, Type II well-being for young adults. Journal of Applied
errors, and false–positive consumer psychology: Don’t Developmental Psychology, 30(6), 708–723. doi:10.1016/j.
fight the power. Journal of Consumer Psychology, 25(4), appdev.2009.02.003
679–689. doi:10.1016/j.jcps.2015.05.006 Tang, N., & Baker, A. (2016). Self-esteem, financial know-
Moorman, C., Diehl, K., Brinberg, D., & Kidwell, B. (2004). ledge and financial behavior. Journal of Economic
Subjective knowledge, search locations, and consumer Psychology, 54, 164–176. doi:10.1016/j.joep.2016.04.005
choice. Journal of Consumer Research, 31(3), 673–680. Vermunt, J. K., & Magidson, J. (2002). Latent class cluster
doi:10.1086/425102
analysis. Applied Latent Class Analysis, 11, 89–106.
Muthen, B., & Muthen, L. K. (2000). Integrating person-
Xiao, J. J., Ahn, S. Y., Serido, J., & Shim, S. (2014). Earlier
centered and variable-centered analyses: Growth mixture
modeling with latent trajectory classes. Alcoholism: financial literacy and later financial behaviour of college
Clinical & Experimental Research, 24(6), 882–891. students. International Journal of Consumer Studies,
Olaru, D., Smith, B., & Taplin, J. H. E. (2011). Residential 38(6), 593–601. doi:10.1111/ijcs.12122
location and transit-oriented development in a new rail Xiao, J.J., Tang, C., Serido, J., & Shim, S. (2011).
corridor. Transportation Research Part A: Policy and Antecedents and consequences of risky credit behavior
Practice, 45, 219–237. among college students: Application and extension of the
Pastor, D. A., Barron, K. E., Miller, B. J., & Davis, S. L. Theory of Planned Behavior. Journal of Public Policy &
(2007). A latent profile analysis of college students’ Marketing, 30(2), 239–245. doi:10.1509/jppm.30.2.239

You might also like