Hi Raja,

If by "unobserved variables" you mean "latent variables," then Amos is very good for this. Google "latent variables" and "amos" and you'll get lots of information.

If by "unobserved variables" you mean variables for which you don't have any data, then you might be out of luck. It might be better to change your study to a qualitative study discussing the results that you have from what you measured.

Yes I meant latent variables... Thank you for the guidance

If you want to model the three independent constructs against one dependent construct or variable, you need to have the data on dependent construct or variable itself. Thus, when you collect data, your questionnaire should include the items for dependent construct as well as independent constructs.

First you can use AMOS for confirmatory factor analysis whereby you test for construct validity of the three latent variables "organizational development, career development and training & development".  These are drawn as circles. The items pertaining to each construct are called manifest variables and are drawn as squares. You can also test for second order factor model by linking the three cobnstructs to another latent variable which you  might call for example "organizational HRD" As regarsd your dependent variable, you need data.  Think of common method variance as well whereby you may consider getting different sources of data for IV and DV

I will be grateful if some one could share related papers.

Read this manuscript. Chapter 3 from my book on Structural Equation Modeling using Amos.

It was very helpful

I am working on a conceptual model (three independent variables as Organizational development, career development and training & development) and the dependent variable is 'Change Agent'. I have asked questions 5 questions on each of independent variables but for dependent variable I didn't ask any question

i also face the same problem?have you already try it? whether it is ok to use only independent variable in amos.. i have stuck already after doing my EFA

and how do we measure un observable action such as moral hazard of management or information asymmetry in credit markets. I found measure for information asymmetry is ask-bid spread in capital /stock markets. what measure is better in credit market? i'm researching on hidden actions of stakeholders such as managers,creditors, shareholders and regulator and i'm looking for good variable to measure hidden actions (moral hazards) of each stakeholder.