Great video. I don't mean to "correct" you, but just to clarify for other viewers a tiny bit, at 5:20, the count option generates the number of individual observations with *non missing* values for the given variable, not just the number of individuals. For example, suppose that of the men from IL, 100 of them have a missing value for their wage. If that were the case, the collapse command would have placed a value of 1588 in the count variable for observation 1 rather than 1688.
You're right. Thanks for the clarification. Missing data can be a bit tricky in Stata because it just sort of quietly deals with it in the background without telling you, and it can definitely trip people up.
I'm not aware of a quick way, but I did find an article discussing this: www.stata.com/support/faqs/data-management/keeping-same-variable-with-collapse/
@@sebastianwaiecon Thank you for your answer! But after collapsing, only affected variables exist. I want other variables and collapsed variables all together.
When you collapse, you only get the collapsed variables and the variables you used to divide up the data. It's impossible to get non-collapsed variables to come along because they are tied to the observations in the original dataset, which no longer exists.
@@sebastianwaiecon Exactly! Then, there is only one way, which is merging the two datasets (collapsed var dataset and previous dataset) to have all the variables. Is that correct?
@@evergreenbella I'm not sure why you'd want to do that. Is the reason you couldn't collapse them that they are string variables? If any of the collapsed categories has any variation in the strings, merging still won't work. If not, you can just include them in the "by" statement.
nevermind, it's because I was trying to get it to count a string variable. I still can't believe the built-in functions of Stata can't handle basic counting of strings. ridiculous. might as well be using Excel.
Does anyone know how to run a regression with regards to a specific part of a variable? ex my variable is sex but I want a regression of wages on females and not on men.
Great video. I don't mean to "correct" you, but just to clarify for other viewers a tiny bit, at 5:20, the count option generates the number of individual observations with *non missing* values for the given variable, not just the number of individuals.
For example, suppose that of the men from IL, 100 of them have a missing value for their wage. If that were the case, the collapse command would have placed a value of 1588 in the count variable for observation 1 rather than 1688.
You're right. Thanks for the clarification. Missing data can be a bit tricky in Stata because it just sort of quietly deals with it in the background without telling you, and it can definitely trip people up.
Thank you. I was so worried about my homework and this helped me a lot.
Thanks for sahring. Short and direct
Thanks for your sharing, I can say it is very wonderful.
you saved my life with this vid
Wonderful explanation!
Thanks so much. This really helped
Really useful. How do you avoid losing labels when using the collapse command?
I'm not aware of a quick way, but I did find an article discussing this: www.stata.com/support/faqs/data-management/keeping-same-variable-with-collapse/
@@sebastianwaiecon thanx a lot
Very helpful. Is there any way to generate a new variable, so it does not erase the old data?
You might want to look into the egen command.
Thank you for sharing. One question, how can we keep other variables after collapsing??
You can keep as many variables as you want - just keep listing them in the command like I did in the video.
@@sebastianwaiecon Thank you for your answer! But after collapsing, only affected variables exist. I want other variables and collapsed variables all together.
When you collapse, you only get the collapsed variables and the variables you used to divide up the data. It's impossible to get non-collapsed variables to come along because they are tied to the observations in the original dataset, which no longer exists.
@@sebastianwaiecon Exactly! Then, there is only one way, which is merging the two datasets (collapsed var dataset and previous dataset) to have all the variables. Is that correct?
@@evergreenbella I'm not sure why you'd want to do that. Is the reason you couldn't collapse them that they are string variables? If any of the collapsed categories has any variation in the strings, merging still won't work. If not, you can just include them in the "by" statement.
Hello, I would like to ask a question. How can i generate a new basadate to have the percent of men by educcategory using collapse?
Percent is one of the operations you can do with collapse. See the manual for details: www.stata.com/manuals13/dcollapse.pdf
Great Video. How do you restore your previous data? I tried using "restore" after "preserve" but got this: "nothing to restore". Kindly help
If I want to go back, I just reload the original dataset. Just make sure to not save over the original data.
weird. you're collapsing by a string, but my Stata returns a type mismatch error when I try to do the same
nevermind, it's because I was trying to get it to count a string variable. I still can't believe the built-in functions of Stata can't handle basic counting of strings. ridiculous. might as well be using Excel.
You might want to check out my video on categorical data for better handling of strings (ruclips.net/video/QN7u-ZBHqic/видео.html).
Does anyone know how to run a regression with regards to a specific part of a variable? ex my variable is sex but I want a regression of wages on females and not on men.
See my video on categorical data.
Is there a way to resolve stata r(950) error...