ARDL Approach to Cointegration

[Farh]

I am new to this forum, as a matter of fact I just joined yesterday.
However I was enamored by the quality of academic discussions most especially in the area of econometric analysis and any other area of interests. That is the reason why I believe that my questions shall be answered.
I am working on a topic that says “an econometric analysis of education, poverty and economic growth linkages in Nigeria”. My model specifications as adapted from Afzal et al (2010) are;
ln (RGDP) = α0 + α1 In(GTOUT) + α2 (HCI) + εt (Model 1)

ln (GTOUT) = β0 + β1 ln (RGDP) + β2(HCI) + εt (Model 2)

HCI = γ0 + γ1 (RGDP) + γ2 In(GTOUT) + εt (Model 3)

Where
RGDP is the real gross domestic product as a proxy for economic growth.
GTOUT is graduate turnout as a proxy for education.
HCI is Headcount Index as a proxy for poverty.

I am using ARDL approach to cointegration with VAR framework.
My questions are as follows;
1. Are the models well specified?
2. Can I use E-view7 to estimate ARDL model? If yes, how do I specify the models?
3. Since the rule of the thumb for applying ARDL approach is that the variables should be I(0) or I(1) or mutually integrated, can I still use this approach if my variables are only I(1)?
4. What are the steps involved in using this approach?
5. Comment on interpreting ARDL results from E-view 7.
The answers to these questions will help me greatly in my thesis.
Thank you in anticipation of your help.
Farh.