The Model Of Integrated Work Potentials And The Work Competence As A Predictor For Meaningful Performance Using Path Analysis With Moderating Variable.
THE M ODEL OF INTEGRATED W ORK POTENTIALS
AND THE W ORK COM PETENCE AS A PREDICTOR
FOR M EANINGFUL PERFORM ANCE USI NG PATH
ANALYSIS W ITH M ODERATING VARIABLE
Ratna Jatnika (rat@melsa.net.id) and
Fitri Ariyanti (fitri.psi@gmail.com)
Faculty of Psychology, University of
Padjadjaran, Bandung, Indonesia
BACKGROUND
RESEARCH M ODEL
M ETHODS
• Pat h analysis represent s an early at t empt at
dealing w it h causal relat ionsh ips
• Developed by Sew allWright in t he 1930’s
• Current ly, causal analysis is done using st r uct ural
equat ion modelling
• Pat h analysis is useful ini il ust rat ing a num ber of
t he issues involved in causal analysis
DEFINITIONS
• Exogenous variable – a variable whose causes are out side of
t he model
• Endogenous variable – a variable whose causes are inside t he
model
• Recursive model – a causal model that is unidirectional (oneway causal flow). It has no feedback loops nor any reciprocal
effect . In a recurcive model, a variable cannot be both cause
and effect at the same t ime
• Nonrecursive model – a causal model with feedback loops
and/ or reciprocal effect s
• Path coefficient – st andardized regression coefficient
predicting one variable from anot her
ASSUM PTIONS
• Relat ion among models are linear, addit ive, and causal.
Curvilinear, mult iplicat ive, or int eract ion relat ions are
excluded
• Residual are uncorrelat ed with all other variables and
residuals in the model
• There is one-way causal flow
• The variable are measured on an int erval scale
M EDIATING VARIABLES
A mediating variables is a variable that describes how rather than w hen effect w ill
occur by accounting for the relationship betw een the independent and dependent
variable. A mediating relationship is one in w hich t he path relating X to Z is
mediated by a third variable (Y)
X
Y
Z
X
Y
Z
M ODERATING VARIABLES
A moderat ing variable is variable t hat affect s t he direct ion and/ or st rengt h of t he
relat ion bet w een dependent and independent variables. Specifically w it hin a
correlat ional analysis framew ork, a moderat or is a t hird variable t hat affect s t he zeroorder correlat ion bet w een t w o ot her variables. (Baron and Kenny, 1986: p. 1174)
M
X
Y
M ODERATED M EDIATION
M
X
Y
Z
Piecemeal approach (1)
• To combine moderat ion and mediat ion involves analyzing
moderat ion and mediation in piecemeal fashion and
int erpret ing their result s joint ly.
• Wit h this approach, moderat ion is usually t est ed wit h analysis
of variance (ANOVA) or regression analysis, in which t he
dependent variable Y is regressed on the independent
variable X, the moderator variable M , and their product XM ,
as follows:
Y= b 1X + b 2 M + b 3XM + e
54-10
Piecemeal approach (2)
• M ediat ion is t est ed separat ely, t ypically wit h t he causal
st eps procedure (Baron & Kenny, 1986), in which t here
relat ionships among X, Z, and t he mediat or variable Y are
analyzed as follows:
–
–
–
–
(a) Z is regressed on X,
(b) Y is regressed on X,
(c) Z is regressed on both X and Y.
These regression equat ions can be writ t en as follows:
Z = aX + e1
Y = bX + e2
Z = c1X + c2Y + e3
PATH EQUATION
Equat ion for mediat ing variables:
Z= a1X1 + a2X2 + a3X3 +a4X4 + a5X5 + a6X6 + e1
Y = b 1X1 + b 2X2 + b 3X3 +b 4X4 + b 5X5 + b 6X6 + e2
Z = c1X1 + c2X2 + c3X3 +c4X4 + c5X5 + c6X6 + c7Y1 + c8Y2 + c9Y3 +c10Y4 + e 3
Equat ion for moderat ing variables:
Y1 = d 11 X1 + d 12 X2 + d 13 X3 + d 14 X4 + d 15 X5 + d 16 X6 + d 17 M 1 + d 18 M 1 + d 19X1M 1 +
d 110 X2 M 1 + d 111X3 M 1 + d 112 X4 M 1 + d 113X5 M 1 + d 114 X6 M 1 + d 115 X2 M 2 + d 116 X3
M 2 + d 117 X4 M 2 + d 118 X5 M 2 + d 119 X6 M 2 + e1
Y2 = d 21 X1 + d 22 X2 + d 23 X5 + d 24 X6 + d 25 M 1 + d 26 M 1 + d 27 X1M 1 + d 28X2 M 1 + d 29 X5
M 1 + d 210 X6 M 1 + d 211 X2 M 2 + d 212 X5 M 2 + d 213 X6 M 2 + e 2
Y3 = d 31 X1 + d 32 X2 + d 33 X5 + d 34 X6 + d 35 M 1 + d 36 M 1 + d 27 X1M 1 + d 38 X2 M 1 + d 39X5
M 1 + d 310 X6 M 1 + d 311X2 M 2 + d 312 X5 M 2 + d 313 X6 M 2 + e 3
Y4 = d 41 X1 + d 42 X2 + d 43 X5 + d 44 X6 + d 45 M 1 + d 46 M 1 + d 47 X1M 1 + d 48 X2 M 1 + d 49X5
M 1 + d 410 X6 M 1 + d 411X2 M 2 + d 412 X5 M 2 + d 413 X6 M 2 + e 3
RESULT
Independent Variable
Path
Coefficient
t
p
Sig
Acceptability of leadership
0.211
3.948
0.000
***
Quality of team-work
development
0.393
7.781
0.000
***
Quality of interpersonal
relationship
F = 68.863
R2 = 0.390
0.178
3.675
0.000
***
P = 0.000
***
***: p
AND THE W ORK COM PETENCE AS A PREDICTOR
FOR M EANINGFUL PERFORM ANCE USI NG PATH
ANALYSIS W ITH M ODERATING VARIABLE
Ratna Jatnika (rat@melsa.net.id) and
Fitri Ariyanti (fitri.psi@gmail.com)
Faculty of Psychology, University of
Padjadjaran, Bandung, Indonesia
BACKGROUND
RESEARCH M ODEL
M ETHODS
• Pat h analysis represent s an early at t empt at
dealing w it h causal relat ionsh ips
• Developed by Sew allWright in t he 1930’s
• Current ly, causal analysis is done using st r uct ural
equat ion modelling
• Pat h analysis is useful ini il ust rat ing a num ber of
t he issues involved in causal analysis
DEFINITIONS
• Exogenous variable – a variable whose causes are out side of
t he model
• Endogenous variable – a variable whose causes are inside t he
model
• Recursive model – a causal model that is unidirectional (oneway causal flow). It has no feedback loops nor any reciprocal
effect . In a recurcive model, a variable cannot be both cause
and effect at the same t ime
• Nonrecursive model – a causal model with feedback loops
and/ or reciprocal effect s
• Path coefficient – st andardized regression coefficient
predicting one variable from anot her
ASSUM PTIONS
• Relat ion among models are linear, addit ive, and causal.
Curvilinear, mult iplicat ive, or int eract ion relat ions are
excluded
• Residual are uncorrelat ed with all other variables and
residuals in the model
• There is one-way causal flow
• The variable are measured on an int erval scale
M EDIATING VARIABLES
A mediating variables is a variable that describes how rather than w hen effect w ill
occur by accounting for the relationship betw een the independent and dependent
variable. A mediating relationship is one in w hich t he path relating X to Z is
mediated by a third variable (Y)
X
Y
Z
X
Y
Z
M ODERATING VARIABLES
A moderat ing variable is variable t hat affect s t he direct ion and/ or st rengt h of t he
relat ion bet w een dependent and independent variables. Specifically w it hin a
correlat ional analysis framew ork, a moderat or is a t hird variable t hat affect s t he zeroorder correlat ion bet w een t w o ot her variables. (Baron and Kenny, 1986: p. 1174)
M
X
Y
M ODERATED M EDIATION
M
X
Y
Z
Piecemeal approach (1)
• To combine moderat ion and mediat ion involves analyzing
moderat ion and mediation in piecemeal fashion and
int erpret ing their result s joint ly.
• Wit h this approach, moderat ion is usually t est ed wit h analysis
of variance (ANOVA) or regression analysis, in which t he
dependent variable Y is regressed on the independent
variable X, the moderator variable M , and their product XM ,
as follows:
Y= b 1X + b 2 M + b 3XM + e
54-10
Piecemeal approach (2)
• M ediat ion is t est ed separat ely, t ypically wit h t he causal
st eps procedure (Baron & Kenny, 1986), in which t here
relat ionships among X, Z, and t he mediat or variable Y are
analyzed as follows:
–
–
–
–
(a) Z is regressed on X,
(b) Y is regressed on X,
(c) Z is regressed on both X and Y.
These regression equat ions can be writ t en as follows:
Z = aX + e1
Y = bX + e2
Z = c1X + c2Y + e3
PATH EQUATION
Equat ion for mediat ing variables:
Z= a1X1 + a2X2 + a3X3 +a4X4 + a5X5 + a6X6 + e1
Y = b 1X1 + b 2X2 + b 3X3 +b 4X4 + b 5X5 + b 6X6 + e2
Z = c1X1 + c2X2 + c3X3 +c4X4 + c5X5 + c6X6 + c7Y1 + c8Y2 + c9Y3 +c10Y4 + e 3
Equat ion for moderat ing variables:
Y1 = d 11 X1 + d 12 X2 + d 13 X3 + d 14 X4 + d 15 X5 + d 16 X6 + d 17 M 1 + d 18 M 1 + d 19X1M 1 +
d 110 X2 M 1 + d 111X3 M 1 + d 112 X4 M 1 + d 113X5 M 1 + d 114 X6 M 1 + d 115 X2 M 2 + d 116 X3
M 2 + d 117 X4 M 2 + d 118 X5 M 2 + d 119 X6 M 2 + e1
Y2 = d 21 X1 + d 22 X2 + d 23 X5 + d 24 X6 + d 25 M 1 + d 26 M 1 + d 27 X1M 1 + d 28X2 M 1 + d 29 X5
M 1 + d 210 X6 M 1 + d 211 X2 M 2 + d 212 X5 M 2 + d 213 X6 M 2 + e 2
Y3 = d 31 X1 + d 32 X2 + d 33 X5 + d 34 X6 + d 35 M 1 + d 36 M 1 + d 27 X1M 1 + d 38 X2 M 1 + d 39X5
M 1 + d 310 X6 M 1 + d 311X2 M 2 + d 312 X5 M 2 + d 313 X6 M 2 + e 3
Y4 = d 41 X1 + d 42 X2 + d 43 X5 + d 44 X6 + d 45 M 1 + d 46 M 1 + d 47 X1M 1 + d 48 X2 M 1 + d 49X5
M 1 + d 410 X6 M 1 + d 411X2 M 2 + d 412 X5 M 2 + d 413 X6 M 2 + e 3
RESULT
Independent Variable
Path
Coefficient
t
p
Sig
Acceptability of leadership
0.211
3.948
0.000
***
Quality of team-work
development
0.393
7.781
0.000
***
Quality of interpersonal
relationship
F = 68.863
R2 = 0.390
0.178
3.675
0.000
***
P = 0.000
***
***: p