<s>
Combinatorial	B-General_Concept
meta-analysis	I-General_Concept
(	O
CMA	O
)	O
is	O
the	O
study	O
of	O
the	O
behaviour	O
of	O
statistical	O
properties	O
of	O
combinations	O
of	O
studies	O
from	O
a	O
meta-analytic	O
dataset	O
(	O
typically	O
in	O
social	O
science	O
research	O
)	O
.	O
</s>
<s>
In	O
an	O
article	O
that	O
develops	O
the	O
notion	O
of	O
"	O
gravity	O
"	O
in	O
the	O
context	O
of	O
meta-analysis	O
,	O
Dr.	O
Travis	O
Gee	O
proposed	O
that	O
the	O
jackknife	B-Algorithm
methods	I-Algorithm
applied	O
to	O
meta-analysis	O
in	O
that	O
article	O
could	O
be	O
extended	O
to	O
examine	O
all	O
possible	O
combinations	O
of	O
studies	O
(	O
where	O
practical	O
)	O
or	O
random	O
subsets	O
of	O
studies	O
(	O
where	O
the	O
combinatorics	O
of	O
the	O
situation	O
made	O
it	O
computationally	O
infeasible	O
)	O
.	O
</s>
<s>
In	O
the	O
original	O
article	O
,	O
k	O
objects	O
(	O
studies	O
)	O
are	O
combined	O
k-1	O
at	O
a	O
time	O
(	O
jackknife	B-Algorithm
estimation	O
)	O
,	O
resulting	O
in	O
k	O
estimates	O
.	O
</s>
<s>
An	O
implication	O
of	O
this	O
is	O
that	O
where	O
multiple	O
random	O
intercepts	O
exist	O
,	O
the	O
heterogeneity	B-General_Concept
within	O
certain	O
combinations	O
will	O
be	O
minimized	O
.	O
</s>
<s>
CMA	O
can	O
thus	O
be	O
used	O
as	O
a	O
data	B-Application
mining	I-Application
method	O
to	O
identify	O
the	O
number	O
of	O
intercepts	O
that	O
may	O
be	O
present	O
in	O
the	O
dataset	O
by	O
looking	O
at	O
which	O
studies	O
are	O
included	O
in	O
the	O
local	O
minima	O
that	O
may	O
be	O
obtained	O
through	O
recombination	O
.	O
</s>
<s>
Dr.	O
Gee	O
's	O
original	O
software	O
for	O
performing	O
jackknife	B-Algorithm
and	O
combinatorial	B-General_Concept
meta	I-General_Concept
analysis	I-General_Concept
was	O
based	O
on	O
older	O
meta-analytic	O
macros	O
written	O
in	O
the	O
SAS	O
programming	O
language	O
.	O
</s>
<s>
A	O
jackknife	B-Algorithm
method	I-Algorithm
was	O
applied	O
to	O
meta-analytic	O
data	O
some	O
years	O
later	O
but	O
it	O
does	O
not	O
appear	O
that	O
specialized	O
software	O
was	O
developed	O
for	O
the	O
task	O
.	O
</s>
