No, "dummy" variable is used to obtain categorical effect. The value that dummy variable may take is (1,0), i.e. gender (male; female), or yes/no. If you use scale variable, you have to convert your scale to category data, for example less than mean = 1 and equal to or greater than mean = 0 .... this could be cumbersome and may be criticized as being unconventional when dummy variable is by its nature categorical data, i.e. binary: (1,0).
Thank you Mr.Paul Louangrath for your reply. But then the variables not not satisfying Stationary condition. i.e the variables are non - stattionary. can you suggest me which to deal the analysis apart.
Thank you for very much for sharing your knowledge . also Thank you in adavance for thr reply to come.
It sounds like you want the dummy variables to vary over time, as being either a member or not a member of a particular scale. If that is correct, I'd consider creating a categorical variable for different scales. Generate dummy variables for each of the scale categories and set it up that way. However the question is not very clear as to what you seek to accomplish.
Can I use some / other Independent Variables which are Scale Variables as dummy variables when I am interested in checking the effect / impact of one IV on the DV under a study for a panel.
I got stuck here because a dummy variable has to be a categorical variable, in general.
Will that be right or wrong in using scale variables as dummy variables in a panel data analysis, for non-stationary data set. I require your suggestion.
I think I follow. Are you familiar with interactive dummy (indicator) variables, where they enter the model as D(i)*S(i), where D(i) is a binomial taking the value of 0 or 1 and S(i) is a scale variable corresponding to observation i? Observations will only take on the scale value if a member of the group D(i)=1. Hence, you have the treatment (as indicated by D(i)) and the scale of the treatment (S(i)) as a single variable. You'll have to be careful in how you interpret the results, but there should be no problem using this in a regression equation with appropriate regression diagnostics.
Thank you Steven R Miller for your response and sharing your knowledge. I am not having any knowledge in this area. Also I would like to inform you that the data is Non - Stationary. can you send me the material regarding this.