Dear colleagues,
What can you suggest in order to fill the missing values, as you can see in the picture?
Best
Ibrahim
It depends on the nature of the analysis and the data.
If you perform a panel data analysis, you could just ignore those years.
In other cases you can interpolate.
A final option could be imputation (drawing from other variables/years).
Ibrahim Niftiyev There are many ways to analyze the unbalanced panel data if you are considering your dataset as a panel.....
See detail https://www.statalist.org/forums/forum/general-stata-discussion/general/1455040-unbalanced-panel-data-analysis
We did a backcasting for one of our research projects to handle the missing observations.
See Rob Hyndman's post in the link below on backcasting:
https://robjhyndman.com/hyndsight/backcasting/
I think the approach suggested by Akram Shavkatovich Hasanov is acceptable for this particular dataset. It's the easiest approach here.
So you can just predict 2001 based on 2018,2017,...,2002. Then the same for 1999.
I also recommend plotting the series as spaghetti plots, see this poster:
Poster A Visual Framework for Longitudinal and Panel Studies (with ...
One potential issue here is that Imputing the series values with backcasting will affect the applicability of tests such as those mentioned on the ePoster. So, if you want to use panel data tests and methods for unbalanced datasets, perhaps, it's better to keep only 2002-2018 period.
- Badges
- Science topic