<s>
In	O
statistics	O
,	O
confirmatory	B-Algorithm
factor	I-Algorithm
analysis	I-Algorithm
(	O
CFA	O
)	O
is	O
a	O
special	O
form	O
of	O
factor	O
analysis	O
,	O
most	O
commonly	O
used	O
in	O
social	O
science	O
research	O
.	O
</s>
<s>
As	O
such	O
,	O
the	O
objective	O
of	O
confirmatory	B-Algorithm
factor	I-Algorithm
analysis	I-Algorithm
is	O
to	O
test	O
whether	O
the	O
data	O
fit	O
a	O
hypothesized	O
measurement	O
model	O
.	O
</s>
<s>
In	O
confirmatory	B-Algorithm
factor	I-Algorithm
analysis	I-Algorithm
,	O
the	O
researcher	O
first	O
develops	O
a	O
hypothesis	O
about	O
what	O
factors	O
they	O
believe	O
are	O
underlying	O
the	O
measures	O
used	O
(	O
e.g.	O
,	O
"	O
Depression	O
"	O
being	O
the	O
factor	O
underlying	O
the	O
Beck	O
Depression	O
Inventory	O
and	O
the	O
Hamilton	O
Rating	O
Scale	O
for	O
Depression	O
)	O
and	O
may	O
impose	O
constraints	O
on	O
the	O
model	O
based	O
on	O
these	O
a	O
priori	O
hypotheses	O
.	O
</s>
<s>
In	O
confirmatory	B-Algorithm
factor	I-Algorithm
analysis	I-Algorithm
,	O
researchers	O
are	O
typically	O
interested	O
in	O
studying	O
the	O
degree	O
to	O
which	O
responses	O
on	O
a	O
p	O
x	O
1	O
vector	O
of	O
observable	O
random	O
variables	O
can	O
be	O
used	O
to	O
assign	O
a	O
value	O
to	O
one	O
or	O
more	O
unobserved	O
variable(s )	O
η	O
.	O
</s>
<s>
Both	O
exploratory	B-Algorithm
factor	I-Algorithm
analysis	I-Algorithm
(	O
EFA	O
)	O
and	O
confirmatory	B-Algorithm
factor	I-Algorithm
analysis	I-Algorithm
(	O
CFA	O
)	O
are	O
employed	O
to	O
understand	O
shared	O
variance	O
of	O
measured	O
variables	O
that	O
is	O
believed	O
to	O
be	O
attributable	O
to	O
a	O
factor	O
or	O
latent	O
construct	O
.	O
</s>
<s>
As	O
such	O
,	O
in	O
contrast	O
to	O
exploratory	B-Algorithm
factor	I-Algorithm
analysis	I-Algorithm
,	O
where	O
all	O
loadings	O
are	O
free	O
to	O
vary	O
,	O
CFA	O
allows	O
for	O
the	O
explicit	O
constraint	O
of	O
certain	O
loadings	O
to	O
be	O
zero	O
.	O
</s>
<s>
Another	O
strong	O
argument	O
for	O
the	O
initial	O
use	O
of	O
EFA	O
,	O
is	O
that	O
the	O
misspecification	O
of	O
the	O
number	O
of	O
factors	O
at	O
an	O
early	O
stage	O
of	O
scale	O
development	O
will	O
typically	O
not	O
be	O
detected	O
by	O
confirmatory	B-Algorithm
factor	I-Algorithm
analysis	I-Algorithm
.	O
</s>
<s>
Structural	O
equation	O
modeling	O
software	O
is	O
typically	O
used	O
for	O
performing	O
confirmatory	B-Algorithm
factor	I-Algorithm
analysis	I-Algorithm
.	O
</s>
<s>
LISREL	B-Algorithm
,	O
EQS	O
,	O
AMOS	O
,	O
Mplus	O
and	O
lavaan	O
package	O
in	O
R	O
are	O
popular	O
software	O
programs	O
.	O
</s>
<s>
When	O
reporting	O
the	O
results	O
of	O
a	O
confirmatory	B-Algorithm
factor	I-Algorithm
analysis	I-Algorithm
,	O
one	O
is	O
urged	O
to	O
report	O
:	O
a	O
)	O
the	O
proposed	O
models	O
,	O
b	O
)	O
any	O
modifications	O
made	O
,	O
c	O
)	O
which	O
measures	O
identify	O
each	O
latent	O
variable	O
,	O
d	O
)	O
correlations	O
between	O
latent	O
variables	O
,	O
e	O
)	O
any	O
other	O
pertinent	O
information	O
,	O
such	O
as	O
whether	O
constraints	O
are	O
used	O
.	O
</s>
