Could someone tell me the difference between path analysis and structural equation modeling? I see many publications use SEM to infer the causal relations while no comparable models and the step to select the best fit model. Is it reasonable? Thank you!
SEM consist of two components; the measurement model and the structural model. It is an iterative algorithm that separately solves out the blocks of the measurement model and then, in a second step, estimates the path coefficients in the structural model.
The measurement model specified the posited relationships of the observed indicators to the latent constructs, while describing the freedom of random error and uniqueness associated with their indicators.
Meanwhile, the structural model is a hypothetical model that prescribes relationships among latent constructs and those observed variables that are not indicators of latent constructs.
Hi Xu,
As far as I know, path analysis is a general term, describing models that explore the directed dependencies or paths between a set of variables, described through a set of structural equations (composite functions? My math is getting a bit rusty) that are estimated simultaneously. I have seen somewhere that path analysis uses only observed measures while SEM is based on the use of latent variables, arising from social sciences and psychology where you are trying to measure intangible concepts such as trust. Therefore, SEM has an outer model(relationship between a variable and its measures, usually more than one for each variable) and an inner model (describing paths between variables). SEM includes covariance-based SEM and partial least squares (PLS), different types of estimation algorithms. I guess that for both path analysis and SEM, causality is implied from the theoretical base of the model rather than explicitly tested (although you could use time-lags or logic to support your claims about causality)
I hope this helps and am curious to see if others can shed more light on this topic as sometimes people will use different terminology for the same concept depending on their genre.
Best regards,
Thorhildur
SEM consist of two components; the measurement model and the structural model. It is an iterative algorithm that separately solves out the blocks of the measurement model and then, in a second step, estimates the path coefficients in the structural model.
The measurement model specified the posited relationships of the observed indicators to the latent constructs, while describing the freedom of random error and uniqueness associated with their indicators.
Meanwhile, the structural model is a hypothetical model that prescribes relationships among latent constructs and those observed variables that are not indicators of latent constructs.
You are asking about the same thing. SEM is the method of analysis linking many paths into one SEM diagram/graphic. Once SEM is executed, you can analysis which path is significant and which path is not. So analyzing the significance of the path is Path Analysis. But prior to analyzing the path you need to run SEM.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publications