Following
Baron
&
Kenny’s
(1986)
threestep
regression
approach
to
test
for
mediation,
we
estimated the following three equations:
MEDIATOR
IV
e
i
i
i
=
+
+
α
α
0
1
(1)
DV
IV
e
i
i
i
=
+
+
β
β
0
1
(2)
DV
IV
MEDIATOR
e
i
i
i
i
=
+
+
+
δ
δ
δ
0
1
2
(3)
For mediation to exist, the independent variable
(
IV
)
must
first
affect
the
mediator
variable
in
equation (1) (i.e.,
a
1
must be statistically significant).
Second, the independent variable must affect the
dependent variable (
DV
) in equation (2) (i.e.,
b
1
must be statistically significant). Third, mediator
must then affect the dependent variable in equation
(3) (i.e.,
d
2
must be statistically significant). In
addition, the effect of the independent variable on
the dependent variable must be less in equation (3)
than in equation (2).
15
The results in Table 2, Panel A and Figure 1, Panel
A, provide support for the Hypothesis 3 prediction
that perceived oversight effectiveness will mediate
the
relation
between
IAR
disclosure
and
confidence in financial reporting reliability. The
equation (1) results indicate that IAR disclosure
is positively related to investor perceptions of
oversight effectiveness (
p
value
=
0.03, onetailed).
In equation (2), we find that IAR disclosure is
positively
related
to
confidence
in
financial
reporting
reliability
(
p
value
=
0.002,
onetailed).
Finally, the equation (3) results show that perceived
oversight
effectiveness
is
positively
related
to
confidence in financial reporting reliability, even in
the presence of IAR disclosure (
p
value
<
0.001,
onetailed).
In
addition,
the
effect
of
the
IAR
disclosure on confidence in financial reporting
reliability is less in equation (3) (
d
1
=
7.05) than in
equation (2) (
b
1
=
10.91). The Sobel test statistic
(Baron
&
Kenny,
1986)
provides
evidence
of
mediation (
p
value
=
0.07, twotailed).
16
The results in Panels B and C of Table 2 and
Figure 1 provide support for the Hypothesis 4a
and
4b
predictions
that
perceived
oversight
effectiveness and confidence in financial reporting
reliability, respectively, will mediate the relation
between
IAR
disclosure
and
investment
recommendation. The equation (1) results indicate
that
IAR
disclosure
is
positively
related
to
perceived oversight effectiveness and confidence
in
financial
reporting
reliability
(onetailed
p
values
=
0.03 and 0.002, respectively). Next, the
equation (2) results provide evidence that IAR
disclosure is positively related to the likelihood
of
investment
recommendation
(
p
value
=
0.06,
onetailed). The equation (3) results indicate that,
even in the presence of IAR disclosure, perceived
oversight effectiveness and confidence in financial
reporting reliability are positively related to the
likelihood
of
investment
recommendation
(
p
value
<
0.001, onetailed) and that the effect of
the IAR disclosure on the likelihood of investment
recommendation is less in equation (3) (
d
1
=
2.98
and

0.25,
respectively)
than
in
equation
(2)
(
b
1
=
7.67).
The
Sobel
test
statistics
suggest
mediation for both oversight effectiveness and
financial
reporting
reliability
(twotailed
p
values
=
0.08 and 0.007, respectively).
You've reached the end of your free preview.
Want to read all 18 pages?
 Winter '18
 The Land, IAR