1986

The Moderator-Mediator Variable Distinction in Social Psychological

Research: Conceptual, Strategic, and Statistical Considerations

Reuben M. Baron and David A. Kenny

Tujuan :

01 02
Distinction of The function are
discussed in three
the moderator levels :
and mediator 1. Conceptual
2. Strategic
3. Statistical
 The Nature of Moderator

A moderator is qualitative (e.g., gender, race, class) or quantitative

(e.g. level of reward) variable that affects the direction and/or
01 strength of the relation between an independent or predictor
variable and a dependent or criterion variable

Specifically, within a correlational analysis framework, a moderator

is a third variable that affects the zero-order between two other
02 variables

A basic moderator can be represented as an interaction between a

03 focal variable and a factor that specifies the appropriate conditions
for its operation
Framework for testing Moderation
Framework for testing Moderation

It is desirable that the moderator variable

be uncorrelated with both the predictor and
the criterion to provide a clearly
interpretable interaction term

The moderator variable always function as

an independent variable
Analytic procedure for testing moderation

The statistical analysis must measure and

test the differential effect of the independent
variable on the dependent variable as
a function of moderator.
 Analytic procedure for testing moderation

Case 2 Case 3
The moderator is a categorical
variable and the independent 02 03 The moderator is a continuous
variable and the independent
variable a continuous variable variable is a categorical
variable

01 04
Case 1 Case 4
Both moderator and Both variables are continuous
independent variables are variables..
categorical variables
Case 1 : Both moderator and independent variables
are categorical variables

1. A dichotomous independent variable’s

effect on the dependent variable varies
as a function of another dichotomy.
2. 2x 2 ANOVA
3. The moderation is indicated by the inter
action
Case 2 : The moderator is a categorical variable and
the independent variable a continuous
variable
• Correlate the independent with dependent separately for
each category of the moderator.
• Ex. Action control is correlated with student participation
for each level of achievement (high and low)
• Deficiency: Assumes that the Independent Variable has
equal variance at each level of the moderator
• The effect of Independent Variable on Dependent
Variable is tested using unstandardized regression
coefficient. The regression coefficients are then tested for
difference (see formula, Cohen & Cohen, 1983 p. 56)
• Reliabilities should be tested for the levels of moderation
and slopes should be disattenuated
Case 3 : The moderator is a continuous variable
and the independent variable is a
categorical variable

• We must know a priori how the effect of the Independ

ent Variable varies as a function of the moderator
• 1) Linear function
• 2) Step function
• 3) Quadratic Function
Example:
Independent Variable: Rational vs. fear-arousing attitude
change
Moderator: Intelligence (IQ test)
. Lanjt. Kasus 3: Fungsi Kedua – Moderator is continuous
and Independent Variable is categorical

Fungsi Kuadrat: The effect of the

The effect can be tested Independent Variable
by dichotomizing the should be greatest
moderator at the point at For those who are
which the function is high on the
presumed to accelerate. moderator.
Lanjt. Kasus 3: Fungsi kedua - Moderator is continuous
and Independent Variable is categorical

Quadratic function Can be tested using hierarchical

regression
Y = a + β(X) + β(Z) + β(XZ) + β(Z2)
+ β(XZ2)

The test of quadratic Text Here The interpretation is aided by

moderation is given by the
graphing the predicted values
test XZ2 for various values of X and Z.
. Lanjt. Kasus 3: Fungsi Ketiga – Moderator is continuous
and Independent Variable is categorical

Step function:
dichotomizing the
moderator at the point
where the step is
suppose to
 Kasus 4: Moderator and Independent Variable are continuous

One can dichotomize the Jika efek variable Independen (X)

moderator at the point where pada variable dependen (Y) bervar
the step tales place (step iasi linier atau kuadratik terhadap
function) moderator (Z), maka moderator
kuadrat digunakan.

Ukuran pengaruh variabel Text Here

independen adalah
koefisien regresi. produk XZ diuji dengan
moderasi
The Nature of Mediator Variables

• The effects of stimuli on

behavior are mediated by
various transformation
procedures internal to the
organism.
• The effect of Independent
Variable on Dependent va
riable is mediated by anot
her variable
Framework for testing Mediation
 Framework for testing Mediation
A variable functions as a mediator when it meets the following conditions:

(b) Variations in the mediator

significantly account for 02
variations in the DV (i.e., Path
b)
03
(c) when Paths a and b are
controlled, a previously

01 significant relation between the

IV and DV is no longer
significant, with the strongest
demonstration of mediation
(a) Variations in levels of the IV occurring when Path c is zero.
significantly account for
variations in the presumed
mediator (i.e., Path c)
Framework for testing Mediation

01 02
When Path c is reduced to If the residual Path c is not
zero, we have strong evidence zero, this indicates the
for a single, dominant operation of multiple mediating
mediator. factors
Analytic procedures for testing for mediation
Conduct a series of regression models:

Regressing the Regressing the Regressing the

mediator on the dependent variable dependent variable
independent on the independent on both the
variable; variable; independent
variable and on the
mediator
 Lanjt. Analytic procedures for testing for mediation
To establish mediation, the following conditions must hold:

the Independen the Independent The mediator must

Variable must affect Variable must be affect the
the mediator in the shown to affect the Dependent Variable
first equation; Dependent Variable in the third equation.
in the second
equation;
the effect of the Perfect mediation holds if
Independent Variable on the Independen Variabel
the Dependent Variable has no direct effect on the
must be less in the third Dependen Variabel when
equation than in the the mediator is controlled.
Analytic procedures

Sobel (1982) provided an

approximate significance test for
the indirect effect of the
independent variable on the
dependent variable via the
mediator (Sobel test).

a. Independen variable to mediator,

standard error of a
b. Mediator to Dependent Variabel,
standard error of b

Indirect effect is estimated by:

 Lanjt. Analytic procedure
Assumptions:

There must be no measurement error in the mediator and that the

01 dependent variable not cause the mediator

02 The presence of measurement error in the mediator tends to

produce an underestimate of the effect of the mediator and an
overestimate of the effect of the independent variable on the
dependent variable when all coefficients are positive (Judd & Kenny,
1981 a).

Path analysis/Structural Equations Modeling are best used to

03 estimate mediations in a model
 Lanjt. Analytic procedure

04 Multiple indicators can be used

05 The experimental context strengthens the SEM technique

06 All relevant paths are directly tested and none are ommitted

Feedback (source of bias): The use of multiple regression analysis

07 presumes that the mediator is not caused by the dependent variable
 Lanjt. Analytic procedure

08 Smith’s (1982) method involves the manipulation of two variables,

one presumed to cause only the mediator and not the dependent
variable and the other presumed to cause the dependent variable
and not the mediator.

09 Two-stage least square

Combining Mediation and Moderation

Content Content Content Content Content Content Content

Contents Contents Contents Contents Contents Contents Contents

Text Text Text Text Text Text Text

Text Here Text Here Text Here Text Here Text Here Text Here Text Here

Contents Contents Contents Contents Contents Contents Contents

- - - - - -

$15 $40 $35 $25 $20 $30 $16

Contents Contents Contents Contents Contents Contents Contents
Thank you