Missing values after generate command

[Julia Lau]

Hi, I am trying to make a new variable called "change" which generates the differences for variable emplystat between my observations that have values of 1,2,3. However when I use the command "gen change = d.emplystat" I keep getting blank values. How do i create this change variable so that it will only generate values if there is a difference (say > 0). I have tried using an "if" command but i can't get it still. Thanks in advance, would appreciate any help.

[Carlo Lazzaro]

Julia:
welcome to ths forum.
Why not going back to basics instead of using time-series operators?

. set obs 3
Number of observations (_N) was 0, now 3.

. g var1=_n

. g var2=0

. g wanted= var1- var2

. list

     +----------------------+
     | var1   var2   wanted |
     |----------------------|
  1. |    1      0        1 |
  2. |    2      0        2 |
  3. |    3      0        3 |
     +----------------------+

.

[Julia Lau]

Hi Carlo, thanks for the reply! However, in this case what I'm trying to do is to find the difference for the same variable but different observations, i.e., for var1 but line 2-line 1, and so on.

[Nick Cox]

Data example please. https://www.statalist.org/forums/help#stata

[Julia Lau]

Here is an example, mergeid is the unique person's identifier and I'm trying to find the difference for emplystat - so that I will know when there is a change in the person's employment status (i.e. from employed to retired, which I have already coded them to be integers).

[Nick Cox]

Please use dataex and not a screenshot. The link in #4 explains why we ask this. In any case isn't there a time variable too? (I don't think we need to see the country variable.)

[Carlo Lazzaro]

Julia:
as an aside to Nick's helpful guidance, youmay want to consider something along the folllowing lines (please note that the last within-panel value of -wanted- is always missing because the calculation stops due to -bysort.):

. use "https://www.stata-press.com/data/r17/nlswork.dta"
(National Longitudinal Survey of Young Women, 14-24 years old in 1968)

. bysort idcode: gen wanted= msp[_n+1] - msp[_n]

. list idcode year msp wanted if idcode==1

       +------------------------------+
       | idcode   year   msp   wanted |
       |------------------------------|
    1. |      1     70     0        1 |
    2. |      1     71     1        0 |
    3. |      1     72     1        0 |
    4. |      1     73     1        0 |
    5. |      1     75     1       -1 |
       |------------------------------|
    6. |      1     77     0        0 |
    7. |      1     78     0        0 |
    8. |      1     80     0        0 |
    9. |      1     83     0        0 |
   10. |      1     85     0        0 |
       |------------------------------|
   11. |      1     87     0        0 |
   12. |      1     88     0        . |
       +------------------------------+

[Julia Lau]

Hi Carlo, I think that worked! Thank you so much