1. variance- & covariance-based sem 2. testing for...
TRANSCRIPT
SEM OVERVIEW
1. VARIANCE- & COVARIANCE-BASED SEM
2. TESTING FOR COMMON METHOD BIAS IN SEM
3. NESTED MODELS AND MULTI-GOUP SEM
4. ADVANCES TO WATCH IN SEM
Jagdip Singh and Mark Leach 2013
VARIANCE- & COVARIANCE-BASED SEM
Four Questions: 1. When is it appropriate to use VBSEM (PLS)?
2. What is the state-of-art in PLS analysis?
3. What questions will likely arise in the review process?
4. What are some key references?
Jagdip Singh and Mark Leach 2013
VARIANCE- & COVARIANCE-BASED SEM
VB-SEM
Causal/formative/composite
Multidimensional Items (complete set)
Unidentified + 2 reflective measures = Identified
Measures-error-free
No Measurement Invariance
CB-SEM
Effect/reflective
Unidimensional item (useful redundancy)
> 3 measures = Identified
Measures-error-prone
Yes Measurement Invariance Jagdip Singh and Mark Leach 2013
SmartPlS
Source: http://www.smartpls.de/
Jagdip Singh and Mark Leach 2013
VARIANCE- & COVARIANCE-BASED SEM Hair, J.F./ Sarstedt, M./ Ringle, C.M./ Mena, J.A.: An assessment of the use of partial least squares structural equation modeling in marketing research, in: Journal of the Academy of Marketing Science (JAMS), Volume 40 (2012), Issue 3, pp. 414-433. Lara Lobschat, Markus A. Zinnbauer, Florian Pallas and Erich Joachimsthaler: Why Social Currency Becomes a Key Driver of a Firm’s Brand Equity: Insights from the Automotive Industry, Long Range Planning, Volume 46 (2013), pp. 125-148. Sarstedt, M./ Henseler, J./ Ringle, C.M.: Multigroup analysis in partial least squares (PLS) path modeling: Alternative methods and empirical results, in: Advances in International Marketing (AIM), Vol. 22, Bingley 2011, pp. 195-218. Edwards, Jeffery (2011), “The Fallacy of Formative Measurement,” Organizational Research Methods, 14 (2): 370-388. Hardin, Andrew and George Marcoulides (2011), “A Commentary on the Use of Formative Measurement,” Educational Psychological Measurement, 71 (5): 753-764. Treiblmaier, Horst, Peter Bentler and Patrick Mair (2011), “Formative Constructs Implemented via Common Factors,” Structural Equations Modeling, 18:1, 1-17.
Jagdip Singh and Mark Leach 2013
“In fact, our evidence suggests that even simple summed scales provide better reliability than PLS… In addition, using a model-based weighting system as used in PLS will guarantee problems with interpretational confounding.”
Ronkko and Evermann (2013), “A Critical Examination of Common Beliefs about Partial Least Squares Path Modeling,” ORM, online March 7, 2013.
Jagdip Singh and Mark Leach 2013
“The authors [Hardin and Marcoulides 2011. p. 753] suggest that to avoid further confusing the consumers of this research, the prudent course of action may be to consider temporarily suspending the use of formative measurement.”
They further contend that the debate on formative measurement should be restricted primarily to premier methods journals where experts can ultimately develop a theoretical perspective that supports or rejects its implementation.”
Jagdip Singh and Mark Leach 2013
SEM IN RECENT SALES PUBLICATIONS
JPSSM 2012-13
SEM
nonSEM
JAMS January 2013
SEM
nonSEM
Jagdip Singh and Mark Leach 2013
COMMON METHOD BIAS
Three questions
1. How is CMB evaluated in SEM?
2. What questions will arise in the review process?
3. What are some key references?
Jagdip Singh and Mark Leach 2013
COMMON METHOD BIAS
Marker Variable
Method Factor
Harmon
What is most appropriate and when?
Which is most robust?
Jagdip Singh and Mark Leach 2013
COMMON METHOD BIAS
Lindell, Michael K., and David J. Whitney (2001), “Accounting for Common Method Variance in Cross-Sectional Research Designs,” Journal of Applied Psychology, 86 (1), 114–121.
Podsakoff, Philip M., Scott B. MacKenzie, Jeong-Yeon Lee, and Nathan P. Podsakoff (2003), “Common Method Bias in Behavioral Research: A Critical Review of the Literature and Recommended Remedies,” Journal of Applied Psychology, 88 (October), 879–903.
Jagdip Singh and Mark Leach 2013
NESTED MODELS
Four Questions
1. How are nested models used in SEM?
2. What are their strengths and pitfalls?
3. What questions will arise in the review process?
4. What are some key references?
Jagdip Singh and Mark Leach 2013
NESTED MODELS
Measurement
• Measurement vs. Structural Models
• Lower vs. Higher order Models
• Common method bias
Hypotheses Testing
• Moderation and group differences
Jagdip Singh and Mark Leach 2013
MULTI-GROUP SEM IN RECENT SALES PUBLICATIONS
4
5
0 2 4 6
Multi goup
One group
JPSSM 2012-13
4
0
Multi group
one group
0 2 4 6
JAMS January 2013
Jagdip Singh and Mark Leach 2013
NESTED MODELS
MacKenzie, Scott B. and R. A. Spreng (1992), “How Does Motivation Moderate the Impact of Central and Peripheral Processing on Brand Attitudes and Intentions?” Journal of Consumer Research, 18 (March), 519-29.
• Ping, Robert A. (1994), “Does Satisfaction Moderate the Association between Alternative Attractiveness and Exit Intention in a Marketing Channel?”, Journal of the Academy of Marketing Science, 22 (Fall), 364-71.
Hair, Joseph F., William C. Black, Barry J. Babin, and Rolph E. Anderson (2009), Multivariate Data Analysis, 7th ed. Upper Saddle River, NJ: Prentice Hall.
Jagdip Singh and Mark Leach 2013
MEDIATION, MODERATION, AND MULTIDATA: THE THREE MS OF SEM
SALES CONSORTIUM: 2013
Jagdip Singh and Mark Leach 2013
X (independent
variable)
Y (dependent
variable)
MEDIATION BASICS
Byx = significant?
X (independent
variable)
Y (dependent
variable)
M (mediating variable)
Yes
Byx ~ 0
Bmx = sig Bym = sig
A significant relationship between X and Y…
vanishes with the inclusion of a third variable (M), which explains
why X and Y are related
Jagdip Singh and Mark Leach 2013
18
X (independent
variable)
Y (dependent
variable)
MEDIATION BASICS
Byx = nonsignificant
X (independent
variable)
Y (dependent
variable)
M (mediating variable)
Yes
Byx = significant
Bmx = sig Bym = sig
A nonsignificant relationship between X and Y…
becomes significant with the inclusion of a third variable (M), which separates the positive and
negative effects of X on Y
Jagdip Singh and Mark Leach 2013
19
19
Role Stress
Performance
MEDIATION Example
Byx = nonsignificant
Role Stress
Performance Burnout
Yes
Byx = positive
Bmx = + Bym = -
A nonsignificant relationship between role stress and
performance…
is separated into a positive (eustress) and negative (distress)
effect on performance
Jagdip Singh and Mark Leach 2013
20
20
20
Change Performance
MEDIATION Example
Byx = nonsignificant
Change Performance Detachment
Yes
Byx = positive
Bmx = + Bym = -
A nonsignificant relationship between change and
performance…
is separated into a positive ( functional) and negative (dysfunctional) effect on
performance
Jagdip Singh and Mark Leach 2013
21
21
21
21
MODERATED MEDIATION Example
Change Performance Detachment
Bmx|1 = + Bym|1 = -
A significant mediated relationship between change
and performance…
is turned off or on by a third variable that makes one or both mediated paths nonsignificant
Change Performance Detachment
Bym|2 = - Bmx|2 = 0
Participation
Jagdip Singh and Mark Leach 2013
General Markov process (linear)
Stable process b1 = b2 = b3
Y1 Y2 Y3 Y4
e11
e21
e31
b1 b2 b3
MIULTI-PERIOD Example
Jagdip Singh and Mark Leach 2013
y1
x11
e1
x12
e2
x13
e3
y2
x21
e4
x22
e5
x23
e6
y3
x31
e7
x32
e8
x33
e9
d1 d2
Constrain same loading to be equal over time
General Markov process with Factorial Invariance
Jagdip Singh and Mark Leach 2013
A series of chi-square difference tests enables selection of parsimonious model, for example, c1 = c2 = c3, or d1 = d2 = d3 = 0.
Y1 Y2 Y3 Y4
e1 1
e2 1
e3 1
a1 a2 a3
X1 X2 X3 X4
e4 e5 e6
1 1 1
b1 b2 b3
d1 d2 d3
c1 c2 c3
Cross-lagged Panel Data Model
Jagdip Singh and Mark Leach 2013
mem1
x11
e1
x12
e2
x13
e3
mem2
x21
e4
x22
e5
x23
e6
mem3
x31
e7
x32
e8
x33
e9
..71 .92
trust1 trust2 trust3.56 .91
.04*
.1* .05*
.05*
d3
d1 d2
d4
Cross-lagged Panel Data Model with Correlated Errors
Jagdip Singh and Mark Leach 2013
Y1 Y2 Y3 Y4
e1 1
e2 1
e3 1
X1 X2 X3 X4
e4 e5 e6
1 1 1
Z
Cross-lagged Panel Data Model with Covariate Z
Jagdip Singh and Mark Leach 2013
Y1 Y2 Y3 Y4
e1 1
e2 1
e3 1
X1 X2 X3 X4
e4 e5 e6
1 1 1
Z2 Z3 Z4
Cross-lagged Panel Data Model with Time-dependent Covariate Z
Jagdip Singh and Mark Leach 2013
Longitudinal SEM models can include:
• Multiple group analysis
• Interaction effects
• Different models for different racial/ethnic groups
• Multiple indicators at each wave of measurement
• Allows estimation of reliability and appropriate path coefficient adjustment for unreliability
• Psychometric assessment of measurement invariance
• Multiple Covariates
• Time invariant covariates, gender, or personal characteristics
• Time varying covariates, household income.
• Complex error structures
Jagdip Singh and Mark Leach 2013
X1 X2 X3
Z2 Z3
Y1
1
1 1 1
Y2
1
1 1 1
Y3
1
1 1 1
GROUP 1
GROUP 2
X1 X2 X3
Z2 Z3
Y1
1
1 1 1
Y2
1
1 1 1
Y3
1
1 1 1
Jagdip Singh and Mark Leach 2013
UNCONDITIONAL RANDOM COEFFICIENTS GROWTH CURVE MODEL: BASIC IDEA
Intercept
y1
1
1
y2
1
1
y3
1
1
Slope
y4 y5 y6
4
1
5
1
6
1
1
1
13
1
0
Jagdip Singh and Mark Leach 2013