Dear Demet,

path analysis is by defintion restricted to observed variables. If you incorporate latent variables, path analysis becomes a "structural equation model". In such a model, you add (in addition the the structural model containing the inter-variable effects) measurement models for each latent variable. In these measurement models, each latent variable is connected to one, two or more indicators that are presumeably caused by the latent variable. The full model tries to explain the covariance between all observed variables  (indicators) by the full set of paths connecting these.

You should start by reading about these models (for instance, the book by Rex Kline, 4rd edition (!) is a good starting point or the Book "cause and correlation in Biology" by Bill Shipley.

Technically, a straightforward option is the lavaan package in R (see www.lavaan.org). The website contains a tutorial how to set up the various models.

Hope this helps

Holger

Dear Steinmetz,  thank you for your contribution. I will check the references that you mention. 

Hi Demet,

You can also check this open book from G. Sanchez about the R package he develop (plspm) : http://gastonsanchez.com/PLS_Path_Modeling_with_R.pdf

Regards

But please note that PLS has something differences compared to structural equation modeling, both with regard the meaning and philosophical foundation of the incorporated variables as well as the assumptions on which the model rests (and which the models test):

Rouse, A. C., & Corbitt, B. (2008). There’s SEM and “SEM”: A critique of the use of PLS regression in information systems research. Australasian Conference on Information Systems. Christchurch.

http://aisel.aisnet.org/cgi/viewcontent.cgi?article=1057&context=acis2008

Ronkko, M., & Evermann, J. (2013). A Critical Examination of Common Beliefs About Partial Least Squares Path Modeling. Organizational Research Methods, 16, 425–448. doi:10.1177/1094428112474693

Antonakis, J., Bendahan, S., Jacquart, P., & Lalive, R. (2010). On making causal claims: A review and recommendations. The Leadership Quarterly, 21, 1086–1120.

Edwards, J. R., & Bagozzi, R. P. (2000). On the nature and direction of relationships between constructs and measures. Psychological Methods, 5(2), 155–174.

Hence, whatever method you use, be sure that it enables modeling your personal view of the theoretical domain and its entities/variables.

Best,

Holger

latent variable is a variable which caused responses for several observed variables, in other words it is a non observable variable which can be indicated by another observable variables, as Holger mentioned when you start use latent variables then it called structural equation model, you can analyze easily use LiREL which you can get so much help about it by Google and also Youtube.

good luck

 If you interested to learn structural equation modeling with AMOS, you can refer this link : https://www.researchgate.net/profile/Zainudin_Awang/contributions