Simple Mediation Analysis by Andrew Hayes
Mediation processes are framed in terms of intermediate variables between an
independent variable and a dependent variable, with a minimum of three variables
required in total: X, M, and Y, where X is the independent variable (IV), Y is the
dependent variable (DV), and M is the (hypothesized) mediator variable that is
supposed to transmit the causal effect of X to Y.
Mediation analysis tests a hypothetical causal chain where one variable X affects a
second variable M and, in turn, that variable affects a third variable Y. For
instance, people with higher incomes tend to live longer but this effect is explained
by the mediating influence of having access to better health care.
In simple mediation model (as discussed by Baron & Kenny, 1986), we have paths
‘a’, ‘b’, and ‘c’ being estimated. The paths you see in this model (and any
mediation model) are referred to as direct effects. In short, the model is specified
such that the X has a direct effect on the mediator (M), with the mediator having a
direct effect on the outcome (Y). X is also presumed to have a direct effect on Y.
Mediation is evidenced when there is evidence that the indirect effect of X on Y
flows through mediator (M). The indirect effect in this model is computed as the
product of paths ‘a’ and ‘b’: IE=a*b.
The total effect of X on Y is referred to as the total effect (TE), and that effect is
then partitioned into a combination of a direct effect (DE) of X on Y, and an
indirect effect (IE) of X on Y that is transmitted through M. In other words, the
relationship between X and Y is decomposed into a direct link and an indirect link.
The total effect of X on Y - It is simply the sum of all direct and indirect effects
from X to Y. So, in this simple model, the total effect of X on Y is: DE + IE = c +
a*b
Direct Effect of X on Y = c’ = direct effect of X on Y after controlling for M. In
other words, the direct effect is the effect of X on Y absent the Mediator.
Significant total, indirect and direct effect point estimates are determined based on
the criterion that the 95% confidence interval (CI) not include zero.
Perfect/Full mediation occurs when the effect of X on Y decreases to 0 with M in
the model.
Partial mediation occurs when the effect of X on Y decreases by a nontrivial
amount (the actual amount is up for debate) with M in the model.
In a simple mediation model with one mediator, full mediation suggests that a
researcher has completely explained the process by which X influences Y and
there is no need to test for further indirect effects. In the case of partial mediation,
there is a clear implication that other indirect effects could (and probably should)
be examined and tested empirically. Thus, conclusions of partial and full mediation
can have implications for theory building as they suggest the plausibility of
additional mechanisms.
Assumptions were checked before running the analysis.
Assumptions for normality including detecting outliers,
The data is normally distributed. There were no missing values found in the
entry.
The data did not have multicollinearity based on the values of VIF.
The assumption of homoscedasticity is that the variability in scores for one
continuous variable is roughly the same at all values of another continuous
variable. It was assumed that the data set was homoscedastic.
Hypothesis for Simple Mediation Analysis:
Interest will mediate the effect of mastery goals on student achievement.
.
Table
Indirect Effect of Mastery Goals on Student Achievement through Interest
Criterion Predictor Variable 95% CI
Variable
β SE p LL UL
Total Effect
Achievement Mastery Goals .52*** .06 <.001 .40 .65
Direct Effects
Achievement Mastery Goals .36*** .08 <.001 .21 .51
Interest Mastery Goals .77*** .09 <.001 .60 .94
Achievement Interest .21** .06 .001 .09 .33
Indirect
Effect
Achievement Mastery Goals through .16 .06 --- .05 .28
Interest
Note. B= Beta, CI= Confidence interval, S.E. = Standard error LL= Lower limit, UL= Upper limit, p*<0.05,
p***<0.001
The SPSS Macro PROCESS developed by Hayes (2013) was run to
investigate the mediating role of interest between mastery goals and student
achievement. A bias-corrected nonparametric bootstrapping technique with 5,000
resamples was employed to estimate the direct, indirect, and total effects of
mastery goals on student achievement. The table shows that the direct effect of
mastery goals (independent variable) on student achievement (dependent variable)
is significant and positive. This means that students with higher mastery goals are
likely to show more achievements. The direct effect of mastery goals (independent
variable) on interest (mediator) is also significant and positive. This means that
students with higher mastery goals are likely to take more interest in their studies.
The direct effect of interest (mediator) on student achievement (dependent
variable) is significant and positive. This means that students who take more
interest in their studies are the ones with higher academic achievements.
The table shows that the indirect effect of mastery goals on student achievement
through interest is significant. Interest partially mediated the relationship between
mastery goals and student achievement.
The emerged model after mediation analysis is given below
Interest
a1= .77,
p=.000 b1 = .21, p=.001
Mastery Goals Achievement
Direct effect, c1= .26, p=.00
Indirect effect, β= .16
Table
Indirect Effect of Mastery Goals on Student Achievement through Interest
Criterion Predictor Variable 95% CI
Variable
β SE p LL UL
Total Effect
Achievement Mastery Goals
Direct Effects
Achievement Mastery Goals
Interest Mastery Goals
Achievement Interest
Indirect
Effect
Achievement Mastery Goals through
Interest
Note. B= Beta, CI= Confidence interval, S.E. = Standard error LL= Lower limit, UL= Upper limit, p*<0.05,
p***<0.001
The emerged model after mediation analysis is given below
Interest (M)
a1=
b1 =
p=
p=
Mastery Goals (X) Achievement (Y)
Direct effect, c1= , p=
Indirect effect, β=