Embedded Files
A central research question in social science research, particularly marketing and MIS, focuses on the operationalization of complex constructs: Are indicators causing or being caused by the latent variable/construct measured by them?
Changes in the latent variable directly cause changes in the assigned indicators
Changes in the latent variable directly cause changes in the assigned indicators
Problems in Model Specification
Problems in Model Specification
It is important to determine whether a construct is reflective or formative. For a long time, people have been living the reflective world.
Type I Error: When the construct is modelled reflectively, but it should be formative.
Type II Error: When the construct is modelled formatively, but it should be reflective.
Based on the research by Jarvis et al. (2003), 1/3 of the papers published were wrongly modelled.
Reflective construct is most commonly used, but in many cases a formative construct would be appropriate.
Reflective Construct
Reflective Construct
Example 1
Example 1
Construct: Timeliness
Construct: Timeliness
X1 = Accommodate last minute request.
X2 = Punctuality in meeting deadlines
X3 = Speed of returning phone calls
For a very timely person, the score of X1, X2, and X3 will be high. For those who are not timely at all, all the scores of the indicators will be low.
X1, X2, and X3 are interchangeable, deleting any item will not change the meaning of the construct.
Therefore, the score of timeliness is reflected in the answers.
When the construct score is high, all the answer to all indicators will be high.
When the construct score is low, all the answer to all indicators will be low.
The construct scores are reflected in the answers to the indicators.
In short, changes in the latent variable directly cause changes in the assigned indicators.
Example 2
Example 2
Construct: Diet
Construct: Diet
D1 = I eat healthy food
D2 = I don’t eat much junk food
D3 = I have a balance diet
Indicators should be highly correlated (Hulland, 1999)
Construct describes and defines the items
If factor loading is low, then the item can be deleted.
Characteristics of Reflective Model
Characteristics of Reflective Model
Direction of causality is from construct to indicators.
Indicators are expected to be correlated.
Dropping an indicator from the measurement model does not alter the meaning of the construct.
Takes measurement error into account.
Similar to factor analysis.
Formative Construct
Formative Construct
Example 1
Example 1
Construct: Life Stress
Construct: Life Stress
X1 = Job loss.
X2 = Divorce
X3 = Recent accident
Life stress is zero. Life stress is influenced by whether you have a job loss, divorce or recent accident.
If you have the three together, life stress will be the highest. If you don't have all the three, the life stress will be the lowest.
X1, X2, and X3 are uncorrelated. If they are highly correlated, then you run into a problem called multi-collinearity.
The construct scores is influenced by the score of the indicators.
When the all indicators are high, the construct score will be high.
When the all indicators are low, the construct score will be low.
In short, changes in one or more of the indicators causes changes in the latent variable.
Example 2
Example 2
Construct: Health
Construct: Health
H1 = I have a balance diet
H2 = I exercise regularly
H3 = I get sufficient sleep every night
Indicators can have positive, negative, and no correlation (Hulland, 1999).
The correlations between the indicators should not be highly correlated.
Items describe and define the construct.
If weight is low, then the item cannot be deleted.
Characteristics of Formative Model
Characteristics of Formative Model
Direction of causality is from indicators to construct.
Indicators are not expected to be correlated.
Dropping an indicator may alter the meaning of the construct (No such thing as internal consistency reliability).
Based on multiple regression analysis (Need to take care of multicollinearity).
View in the Formative Measures
View in the Formative Measures
Composite (formative) constructs – indicators completely determine the “latent” construct. They share similarities because they define a composite variable but may or may not have conceptual unity. In assessing validity, indicators are not interchangeable and should not be eliminated, because removing an indicator will likely change the nature of the latent construct.
Causal constructs – indicators have conceptual unity in that all variables should correspond to the definition of the concept. In assessing validity, some of the indicators may be interchangeable, and also can be eliminated.
Satisfaction as Formative and Reflective Construct
Satisfaction as Formative and Reflective Construct
One variable can be measured reflectively or formatively.
Y = Satisfaction
Y = Satisfaction
Source: Hair et al., (2014)
Guidelines for Choosing the Measurement Model
Guidelines for Choosing the Measurement Model
Reflective indicators are interchangeable. Deleting an item does not change the essential nature of the construct.
With formative indicators, omitting an indicator is omitting a part of the construct (Diamantopoulos & Winklhofer, 2001)
Guideline from Hair et al., (2014)
Guideline from Hair et al., (2014)
Can we do this?
Can we do this?
Further Reading
Further Reading
Chang, W., Franke, G. R., & Lee, N. (2016). Comparing reflective and formative measures: New insights from relevant simulations. Journal of Business Research, 69(8), 3177-3185. Click here.
Finn, A., & Wang, L. (2014). Formative vs. reflective measures: Facets of variation. Journal of Business Research, 67(1), 2821-2826. Click here.
Diamantopoulos, A., & Siguaw, J. A. (2006). Formative versus reflective indicators in organizational measure development: A comparison and empirical illustration. British Journal of Management, 17(4), 263-282. Click here.
Hulland, J. (1999). Use of partial least squares (PLS) in strategic management research: A review of four recent studies. Strategic management journal, 20(2), 195-204. Click here.
Jarvis, C. B., MacKenzie, S. B., & Podsakoff, P. M. (2003). A critical review of construct indicators and measurement model misspecification in marketing and consumer research. Journal of consumer research, 30(2), 199-218. Click here.
MacKenzie, S. B., Podsakoff, P. M., & Jarvis, C. B. (2005). The problem of measurement model misspecification in behavioral and organizational research and some recommended solutions. Journal of applied psychology, 90(4), 710. Click here
Petter, S., Straub, D., & Rai, A. (2007). Specifying formative constructs in information systems research. MIS quarterly, 623-656. Click here.
Page updated
Google Sites
Report abuse