Hi Statalist,
I am relatively new to Stata. I wish to analysis the change in economic status between religions before and after a policy change - a diff-in-diff of sorts - along with other relevant analysis.
By doing this I must merge the five-quarter longitudinal data for the UK Labour Force Survey from 1993-2020. However I am unsure of the best form of action. There is a personal identifier variable PERSID. Do I select which variables to keep in each dataset and then merge each one by one thereafter? Or is it possible to merge all first? I have looked into this but notice I must list the key variables first when merging.
I am also unsure if this would be a one-to-one merge or one-many etc. I am only interested in questions that have been asked within this timescale so the variables I wish to keep should be mostly the same for each dataset- however there have been some variables that have been replaced or named something else over the years.
Additionally, would I need to relabel each variable to indicate which year it was from? There are variables such as MANAGER1 and MANAGER2 - the numbers indicating which quarter it is for. Would these need to be relabelled as MANAGER102 for quarter 1 of 2002? Alternatively would an average need to be taken - AVGMANAGER02 - average of MANAGER variable for 2002.
Please do forgive the lack of knowledge here.
Any advice for this Stata rookie would be greatly appreciated!
I am relatively new to Stata. I wish to analysis the change in economic status between religions before and after a policy change - a diff-in-diff of sorts - along with other relevant analysis.
By doing this I must merge the five-quarter longitudinal data for the UK Labour Force Survey from 1993-2020. However I am unsure of the best form of action. There is a personal identifier variable PERSID. Do I select which variables to keep in each dataset and then merge each one by one thereafter? Or is it possible to merge all first? I have looked into this but notice I must list the key variables first when merging.
I am also unsure if this would be a one-to-one merge or one-many etc. I am only interested in questions that have been asked within this timescale so the variables I wish to keep should be mostly the same for each dataset- however there have been some variables that have been replaced or named something else over the years.
Additionally, would I need to relabel each variable to indicate which year it was from? There are variables such as MANAGER1 and MANAGER2 - the numbers indicating which quarter it is for. Would these need to be relabelled as MANAGER102 for quarter 1 of 2002? Alternatively would an average need to be taken - AVGMANAGER02 - average of MANAGER variable for 2002.
Please do forgive the lack of knowledge here.
Any advice for this Stata rookie would be greatly appreciated!
Comment