11.16 Exercise

1. Do I need age in the first-differences model? When could it make sense to include it anyways (in terms of data)?
2. I lost my wide-format dataset :-(. Use the long format dataset we just worked with called data and reshape it into wide format dataset called data_wide.
3. Above we used PanelMatch with two different specifications. Rerun the PanelMatch function (at least three times) and change various parameters such as lag, exact matching to fuzzy matching and the covariates that are include in the matching process. Compare those models to find out whether the results differ strongly (see the vigette an exemplary comparison).
4. Why is there a trade-off between between having a good number of matched sets and having matched sets that are large enough in size?

# Solution
#1. Do I need age in the first-differences model? When could it make sense to include it anyways (in terms of data)?
#2. I lost my wide-format dataset :-(. Use the long format dataset we just worked with called ```data``` and reshape it into wide format dataset called ```data_wide```.
#3. Above we used `PanelMatch` two different specifications. Redo those estmiations but now change various parameters such as `lag`, exact matching to fuzzy matching, the covariates that are include in the matching process. Compare those models to find out whether the results differ strongly.
#4. Why is there a trade-off between between having a good number of matched sets and having matched sets that are large enough in size?