<s>
In	O
statistics	O
,	O
circular	B-General_Concept
analysis	I-General_Concept
is	O
the	O
selection	O
of	O
the	O
details	O
of	O
a	O
data	O
analysis	O
using	O
the	O
data	O
that	O
is	O
being	O
analysed	O
.	O
</s>
<s>
Circular	B-General_Concept
analysis	I-General_Concept
unjustifiably	O
inflates	O
the	O
apparent	O
statistical	O
strength	O
of	O
any	O
results	O
reported	O
and	O
,	O
at	O
the	O
most	O
extreme	O
,	O
can	O
lead	O
to	O
the	O
apparently	O
significant	O
result	O
being	O
found	O
in	O
data	O
that	O
consists	O
only	O
of	O
noise	O
.	O
</s>
<s>
A	O
second	O
misuse	O
occurs	O
where	O
the	O
performance	O
of	O
a	O
fitted	O
model	O
or	O
classification	B-General_Concept
rule	I-General_Concept
is	O
reported	O
as	O
a	O
raw	O
result	O
,	O
without	O
allowing	O
for	O
the	O
effects	O
of	O
model-selection	O
and	O
the	O
tuning	O
of	O
parameters	O
based	O
on	O
the	O
data	O
being	O
analyzed	O
.	O
</s>
<s>
In	O
functional	B-Algorithm
magnetic	I-Algorithm
resonance	I-Algorithm
imaging	I-Algorithm
(	O
fMRI	B-Algorithm
)	O
data	O
,	O
for	O
example	O
,	O
considerable	O
amounts	O
of	O
pre-processing	O
is	O
often	O
needed	O
.	O
</s>
<s>
Similarly	O
,	O
the	O
classifiers	B-General_Concept
used	O
in	O
a	O
multivoxel	O
pattern	O
analysis	O
of	O
fMRI	B-Algorithm
data	O
require	O
parameters	O
,	O
which	O
could	O
be	O
tuned	O
to	O
maximise	O
the	O
classification	O
accuracy	O
.	O
</s>
<s>
In	O
geology	O
,	O
the	O
potential	O
for	O
circular	B-General_Concept
analysis	I-General_Concept
has	O
been	O
noted	O
in	O
the	O
case	O
of	O
maps	O
of	O
geological	O
faults	O
,	O
where	O
these	O
may	O
be	O
drawn	O
on	O
the	O
basis	O
of	O
an	O
assumption	O
that	O
faults	O
develop	O
and	O
propagate	O
in	O
a	O
particular	O
way	O
,	O
with	O
those	O
maps	O
being	O
later	O
used	O
as	O
evidence	O
that	O
faults	O
do	O
actually	O
develop	O
in	O
that	O
way	O
.	O
</s>
