topic Re: create a derived variable in SAS Programming

create a derived variable

mandan414 — Wed, 28 Oct 2020 19:52:07 GMT

Hello,

I have 3 variable in my dataset like these:

var_1 cleaning house Var_2 washing dishes Var_3 cooking

I 1 I 1 I 1

sister 2 sister 2 sister 2

both 3 both 3 both 3

NA 6 NA 6 NA 6

NS 9 NS 9 NS 9

I would like to make a derived variable of total task and find out who is doing more tasks overall.

Thank you

create a derived variable

mandan414 — Wed, 28 Oct 2020 19:45:46 GMT

Hello,

I have 3 variable in my dataset like these:

var_1 cleaning house Var_2 washing dishes Var_3 cooking

I 1 I 1 I 1

sister 2 sister 2

brother 3 brother 3

NA 6 NA 6

NS 9 NS 9

Re: create a derived variable

PaigeMiller — Wed, 28 Oct 2020 19:51:30 GMT

What is your question?

Re: create a derived variable

mandan414 — Wed, 28 Oct 2020 19:53:52 GMT

Sorry I posted the question before finish it. i sent the full question right after that. thanks.

Re: create a derived variable

CurtisMackWSIPP — Wed, 28 Oct 2020 20:11:39 GMT

Sorry if I am being dense here, but your data makes no sense. You say you have three variables, but list 6. You refer to a concept of "who" but only the word "sister" matches that description.

Re: create a derived variable

mandan414 — Wed, 28 Oct 2020 20:20:32 GMT

Var_1 Var_2 Var_3 each has 5 reply categories: I (myself), sister, both, NA (not applicable) and NS (not stated). Let's say that i have 1000 respondents and i want to see out of these 1000 what the rates are for myself vs sister vs both. Thank you!

Re: create a derived variable

Reeza — Wed, 28 Oct 2020 20:22:23 GMT

If this was the supplied input, what would you expect as the output?

Re: create a derived variable

mandan414 — Wed, 28 Oct 2020 20:22:57 GMT

I mean the rates of derived variable of total task.

Re: create a derived variable

Reeza — Wed, 28 Oct 2020 20:30:25 GMT

Please show EXACTLY what is expected as output.

Re: create a derived variable

mandan414 — Wed, 28 Oct 2020 20:31:03 GMT

I have these 3 variables in my dataset. I wants to create a derived variable out of these 3 as total task like this:

Var_4 total task

I 1

sister 2

both 3

NA 6

NS 9

Re: create a derived variable

mandan414 — Wed, 28 Oct 2020 20:35:44 GMT

I have these 3 variables in my dataset. I wants to create a derived variable out of these 3 as total task like this:

Var_4 total task

I 1

sister 2

both 3

NA 6

NS 9

Re: create a derived variable

Kurt_Bremser — Wed, 28 Oct 2020 20:42:19 GMT

You already have those in your original dataset.

Re: create a derived variable

mandan414 — Wed, 28 Oct 2020 20:46:47 GMT

I want the rates for total of those tasks. As I said, i'd like to make a variable out of those 3 named total task. Then find out the rate of respondents for each categories of I (myself) vs sister vs both. Thanks.

Re: create a derived variable

Reeza — Wed, 28 Oct 2020 20:48:05 GMT

Using proc freq is one way

proc freq data=have;
table variableName;
freq task;
run;

@mandan414 wrote:

I have these 3 variables in my dataset. I wants to create a derived variable out of these 3 as total task like this:

Var_4   total task

I             1

sister      2

both       3

NA         6

NS         9

Re: create a derived variable

mandan414 — Wed, 28 Oct 2020 20:51:02 GMT

let's say that respondent#1 answer 'I' for Var-1 'I' for var_2 and both for var-3, then respondent#2 replied sister for var_1, I for var_2 and I for var_3 and …. so on. I set thresholds for total (all responds for I, sister and both) and subtotal (all I or all sister or all both) and then put condition and make var_4 as total tasks with the categories of I, sister, myself, NA and NS.

Re: create a derived variable

mandan414 — Wed, 28 Oct 2020 20:53:22 GMT

Hi Reeza,
But first I have to create this Var_4 (as total task) from those 3 of individual tasks.

Re: create a derived variable

Reeza — Wed, 28 Oct 2020 21:11:46 GMT

This is why I'd like to see the final results expected first.

Re: create a derived variable

mandan414 — Wed, 28 Oct 2020 21:36:09 GMT

Sorry, it seems it's confusing now. Let's start from beginning:
I have actually 5 variables that I want to put together. they are something like this as I mentioned before:
Var_1 who does the cleaning?
content: code
Myself 1
spouse 2
both 3
NA 6
NS 9

Var_2 who does the washing?
content: code
Myself 1
spouse 2
both 3
NA 6
NS 9

Var_2 who does the cooking?
content: code
Myself 1
spouse 2
both 3
NA 6
NS 9

Var_3 who does the caregiving?
content: code
Myself 1
spouse 2
both 3
NA 6
NS 9

Var_1 who does the teaching?
content: code
Myself 1
spouse 2
both 3
NA 6
NS 9
Let's say that I cannot disaggregate the data for sex due to confidentiality. Instead, I can combine all variables together as the total task and then compute how many percent of male vs female respond myself vs spouse vs both . But, I cannot simply sum up the contents because a respondent could reply myself for one variable, then spouse for another one and …… what I want to have at the end is a derived variable of total task from those 5 variables. Then I can do: proc freq data=my data; tables sex*total task; run; , and I will have the rate of total task for each sex. I need this by tomorrow and none of my codes seem work well. Thank you!

Re: create a derived variable

Reeza — Wed, 28 Oct 2020 22:04:34 GMT

1. Transpose your data set so you have a long data set and each chore/activity is on it's own line instead of in separate variables. PROC TRANSPOSE will do that.
2. Use PROC TABULATE to calculate summaries and percentages as desired.

If you'd like help to code that, please provide expected input and output as text data.