calculating frequency of events by age-class

[Pakeezah Saadat]

Hello stata-users,

I need to create a two column dataset by age-class and number of events. I can do so manually by using

tab age events

. However, I am sure there's perhaps a better way to do it.
Here's sample dataset which should give you something like age: 18; events: 12 (here event is codified by 0) , age 19; events: counts of 0s, and so on.

* Example generated by -dataex-. To install: ssc install dataex
clear
input int age byte event
18 0
18 0
28 0
18 0
38 1
18 0
48 0
18 0
58 0
18 0
68 1
18 0
38 0
18 1
18 0
68 0
18 0
18 0
18 0
18 0
end

Thank you.

[Nick Cox]

contract or collapse will do what you want.

[Pakeezah Saadat]

Thank you, Nick. That worked beautifully.

I am wondering now if I want to assign that new value "number of events" to each observation (going reverse), how would I do that?

Essentially, if I find that for age 18, there's is 90 events, I want to assign the number 90 to all observations at age 18.

[Nick Cox]

If that's what you want then contract or collapse are not needed at all.

bysort age : gen freq = _N

[Pakeezah Saadat]

This is perhaps counting the frequency of each age-class which is not what I need. If I have an event (0,1), I would like to know how many 0s and 1s are in each class and then have it listed beside it in columns.

[Nick Cox]

Sorry, but I have to think that #4 is a fair answer to #3. For the different question in #5 consider

tab age event

which is just what you mention in #1, so I am lost on what else you seek.

[Andrea Discacciati]

Could this be what Pakeezah wants?

bysort age : egen tot_events = total(event)
bysort age : egen tot_nonevents = total(1-event)

[Nick Cox]

Andrea Discacciati You may be right. Smart guess.

When Pakeezah Saadat asks for columns, I assume that columns in a table are meant, but Pakeezah may be thinking spreadsheet-style in terms of new variables.

Note that

contract age events

puts out new variables too.

[Pakeezah Saadat]

Could this be what Pakeezah wants?

bysort age : egen tot_events = total(event)
bysort age : egen tot_nonevents = total(1-event)

Perfect, thanks Andrea Discacciati. I wonder now though, if a variable has 0s and 1s as values, how does it add up the 1s first? No such command was given. This is in relation to

bysort age:egen tot_events=total(events)