On the other hand, when I used m:m on the key variable(hhid ) this is the results I got as well;
. merge m:m hhid using "C:\Users\USER\Desktop\opn_stata_merged\Kpadam _Pro ject\Append approach\Dependet_indepe
> ndent_merge\Individuals_dep_inde\Pannelquestion\wa ve1_education.dta ...
Which of the merge method should I best used to merge the individual response from both files together because, all the two command gave me 2 different matched results."
. merge m:m hhid using "C:\Users\USER\Desktop\opn_stata_merged\Kpadam _Pro ject\Append approach\Dependet_indepe
> ndent_merge\Individuals_dep_inde\Pannelquestion\wa ve1_education.dta ...
Which of the merge method should I best used to merge the individual response from both files together because, all the two command gave me 2 different matched results."
If you check the Stata manual section on -merge- you will see that even StataCorp says that -merge m:m- should not be used. Now, my advice above is slightly hyperbolic. But only slightly. I have been using Stata since 1994, and I use it nearly every day. In that entire time, I have only once encountered a situation where -merge m:m- would produce a correct, usable result. Only once. Even then, there was a better way to do the same thing. Unless you use Stata as intensively and for as long as I have, it is unlikely you will ever encounter a situation where -merge m:m- is appropriate. So it is better to think of the rule in the absolute terms I have shown above.
When you think you need to use -merge m:m- it means that either you do not understand your data correctly, or the data are not suitable for -merge-ing at all and either must be combined in some other way, or should not be combined at all.
Comment