Comparing values within the same variables

[Glen William Spiteri]

Hi all, I have a longitudinal dataset (N=560) over two waves with an identifier variable called "pid" (I cannot post the dataset here since it is microdata). Everyone has a unique ID and is in wave 1 and wave 2. After cleaning up the dataset, some observations were dropped, meaning that some individuals in both waves are now in only one wave. I want to drop these observations by comparing the ID values in "pid" and generating a new variable called "match" that takes on a value of 1 if the "pid" values are equal, '.' if not. I have the following code:

sort pid
gen match = 0

// Loop through the observations and compare 'pid' values
forvalues i=2/560 {
if pid[`i'] == pid[`i'-1] {
replace match[`i'] = 1
}
}

Each time this returns an error "weights not allowed".

Can anyone help?

[Andrew Musau]

assert !missing(pid)
bys id (pid): gen same= pid[1]==pid[_N]

where "id" is the individual identifier and "pid" is the wave identifier. This creates an indicator where same=1 if an individual is observed in a single wave and zero otherwise. So

drop if same

See https://www.stata.com/support/faqs/d...ions-in-group/

[Glen William Spiteri]

This is just what I needed - thank you so much!