EViews.com

Hey at all,

I use EViews 7.2 and I have a panel data set with an intraday frequency variable "Shift":

Example in attach..
I want to analyze determining factors on the dependent variable "Productivity" with FE, RE specifications etc...

The data set is unbalanced and there are 10 groups who are not observed at every shift at every date. The observed period is three months, 01/2011 to 03/2011 from Mo - Sunday. (Night shift is from 10pm - 6am, early from 6am to 2pm and late from 2pm to 10pm - if this information helps)

1) Which should be the basic structure in eviews for such a data set?
2) How can I select two data series (Date & Shift)?

Thank you very much for your support!

MP

@ the experts no idea? Or at least how I should transform the excel file variables "Date" and "Shift" in a type that eviews would understand as a data series.

I can't speak for others, but I really didn't understand what you were trying to do...Perhaps a more detailed explanation would help.

Hey Glenn,

thank you very much for your response. Ok, here more detailled:


I am investigating the relationship between the productivity of a team at an assembly line, according to their age structure and other control variables. So my dependent variable is: Productivity.
For this investigation I have 10 different work teams who are working in a three-shift system. In every work team the composition of the included workers change nearly every shift. The teams work from monday to sunday. So that means my cross-section are the 10 groups, and my time series are three shifts a day in every group.
In my employer-employee dataset are information about every productivity per shift connected to the corresponding team for the months jan,feb,mar.

Now, it is already possible with this dataset to run a pooled regression. But i suppose unobserved heterogenity between the 10 groups. For this reason I want to specify my panel-regression by controlling the model with fixed and random-effects.

My problem is: How do I have to tell EViews about my cross-section and my time series, so that it analyses the data set with fixed and random effects for the 10 work-groups.

Here again an -example- of the design of my excel sheet data set that is unbalanced as well.
Now clear?

Thanks a lot! :)

MP