type mismatch

[Syeda Tamseel Zahra]

I have a panel dataset of the total fertility rate, need to take a lag of it and then midpoint. Used commands:

gen lagged = TFR[_n-1]
destring TFR Year, replace
destring lagged, replace
drop if Year == 1950
rename _merge _merge1
gen midpoint = (TFR+lagged)/2

but it says type mismatch when I input the drop if year==1950 command.
What should I do?

[Ali Atia]

Tough to know for sure without a data example, but I assume the -destring- didn't work because there are nonnumeric characters in year. Adding the -force- option to the destring will convert all observations with nonnumeric characters into missings, and all observations which are completely numeric into numbers.

[Julian Duggan]

Hi Syeda,

It is a little difficult to know for sure without looking at the dataset. Perhaps you could provide an excerpt using dataex if the suggestion below doesn't work? See here: https://www.stata.com/help.cgi?dataex

My guess is that you have capitalized the Y in Year on the line "drop if Year == 1950". Try changing it to "drop if year == 1950". If the problem is not with the variable name, then try running "desc Year". If the variable type is not a number, then you will need to turn it into a number with "destring Year, replace".

Another thing to look out for: make sure your dataset is properly sorted before running "gen lagged = TFR[_n-1]" -- I would recommend using the tsset command and then getting the lagged value with "gen lagged = L.TFR". Otherwise, the lagged values aren't going to be correct.

Best,

Julian