survey data help: dropping observations

[Patricia Ali]

Hello,

I am new to stata and would appreciate your help. I have ten crops under var2. I want to drop 8 of the crops and only analyse two of them making sure that I still have other variables like household id, sex corresponding to the two crops of interest. I can't seem to find the right drop command. The crops are listed by name eg. rice, sugar cane etc

Thanks in advance for your help

[Marcos Almeida]

Welcome to the Stata Forum / Statalist.

You didn't provide a example we could work on. Please read the FAQ and lear how to share data/command/output.

That being said, you wonder whether you wish something like:

. sysuse auto
(1978 Automobile Data)

. list make price mpg rep78 in 1/10

     +--------------------------------------+
     | make             price   mpg   rep78 |
     |--------------------------------------|
  1. | AMC Concord      4,099    22       3 |
  2. | AMC Pacer        4,749    17       3 |
  3. | AMC Spirit       3,799    22       . |
  4. | Buick Century    4,816    20       3 |
  5. | Buick Electra    7,827    15       4 |
     |--------------------------------------|
  6. | Buick LeSabre    5,788    18       3 |
  7. | Buick Opel       4,453    26       . |
  8. | Buick Regal      5,189    20       3 |
  9. | Buick Riviera   10,372    16       3 |
 10. | Buick Skylark    4,082    19       3 |
     +--------------------------------------+

. drop if rep ==4
(18 observations deleted)

. tab rep78

     Repair |
Record 1978 |      Freq.     Percent        Cum.
------------+-----------------------------------
          1 |          2        3.92        3.92
          2 |          8       15.69       19.61
          3 |         30       58.82       78.43
          5 |         11       21.57      100.00
------------+-----------------------------------
      Total |         51      100.00

. list make price mpg rep78 in 1/10

     +--------------------------------------+
     | make             price   mpg   rep78 |
     |--------------------------------------|
  1. | AMC Concord      4,099    22       3 |
  2. | AMC Pacer        4,749    17       3 |
  3. | AMC Spirit       3,799    22       . |
  4. | Buick Century    4,816    20       3 |
  5. | Buick LeSabre    5,788    18       3 |
     |--------------------------------------|
  6. | Buick Opel       4,453    26       . |
  7. | Buick Regal      5,189    20       3 |
  8. | Buick Riviera   10,372    16       3 |
  9. | Buick Skylark    4,082    19       3 |
 10. | Cad. Deville    11,385    14       3 |
     +--------------------------------------+

[Marcos Almeida]

On second thouhgts, I just wish to remark that:

a) Generally speaking, we don't need to drop observations in order to perform a subgroup analysis. We just need to use the "if" clause". Another strategy is using - preserve -, then the estimation with the dropped observation, finally - restore - command.

b) Considering you have a survey, I believe the best approach to perform the analysis of a subgroup of crops is, well, using - subgroup - within the code. This is the correct way to preserve the survey design.

Hopefully that helps.

[Patricia Ali]

Thank you so much....this was helpful