topic Re: Findind the sum of all variables in New SAS User

Findind the sum of all variables

Guerraje — Wed, 27 Oct 2021 17:08:41 GMT

I am working with survey data. I have been asked to create a new variable which is the sum of all of the variables below, and then run a t-test so that we can use the standard deviation.

I have tried to use proc means and proc summary but they are providing me with the sums of the rows now the sums of all the variables added up, so I can't quite figure out the right statements to use.

Thank you!

My example of proc summary code is here below, to show how I am trying to enter these variables

PROC SUMMARY DATA= WORK.vaccine_hesitancy_data SUM PRINT;

VAR vaccinefactors1 vaccinefactors2 vaccinefactors3 vaccinefactors4 vaccinefactors5 vaccinefactors6 vaccinefactors7 vaccinefactors8 vaccinefactors9 vaccinefactors10 vaccinefactors11;

OUTPUT OUT= vaccinefactors

SUM= sum_of_vaccinefactors;

RUN;

PROC TTEST ;

var vaccinefactors1 vaccinefactors2 vaccinefactors3 vaccinefactors4 vaccinefactors5 vaccinefactors6 vaccinefactors7 vaccinefactors8 vaccinefactors9 vaccinefactors10 vaccinefactors11;

CLASS VACCINEHESITANT;

RUN;

	Total	%	Hesitant -1	%	NotHesitant -0	%
	N = 7197		N = 281		N = 6916
	n (%)		p-value
What factors affect yout attitudes and thoughts about getting the COVID- vaccine? n (%)
Politics	694	9.64%	100	35.59%	594	8.59%	0.0001
The timeline in which the vaccines were developed and approved	1383	19.22%	207	73.67	1176	17.00%	0.0001
The frequently changing messages around COVID-19	918	12.76	174	61.92	744	10.76%	0.0001
Actions and opinions of my friends and family regarding the vaccine	841	11.69	41	14.59	800	11.57%	0.122
My trust in scientists	5147	71.52%	54	19.22	5093	73.62	0.0001
My trust in doctors	4601	63.93%	51	18.15	4550	65.79	0.0001
My trust in public health officials	3890	54.05%	133	47.33	3757	54.32	0.0211
My own reading and research on coronavirus (COVID-19) vaccines	4396	61.08	169	60.14	4227	61.12	0.742
The country in which a vaccine is manufactured	662	9.20%	22	7.83	640	9.25	0.4179
The potential cost of a coronavirus (COVID-19) vaccine	284	3.95%	4	1.42	280	4.05	0.0267
Other factors	734	10.2	84	28.89	650	9.4	0.0001
Total factors considered, mean (SD)

Re: Findind the sum of all variables

sbxkoenk — Wed, 27 Oct 2021 17:14:43 GMT

Try this :

data have;
vaccinefactors1 = 7; 
vaccinefactors2 = 3;
vaccinefactors3 = 2.58; 
vaccinefactors4 = 1.20; 
vaccinefactors5 = 17; 
vaccinefactors6 = 18; 
vaccinefactors7 = 7; 
vaccinefactors8 = 7; 
vaccinefactors9 = 5; 
vaccinefactors10= 7; 
vaccinefactors11= 3;
output;
run;

data want;
 LENGTH sum_vaccinefactors 8;
 set have;
 sum_vaccinefactors=sum(of vaccinefactors1-vaccinefactors11);
run;
/* end of program */

Koen

Re: Findind the sum of all variables

Kurt_Bremser — Wed, 27 Oct 2021 17:14:56 GMT

So you want to get the horizontal sum of all vaccine... variables in every observation?

Re: Findind the sum of all variables

ballardw — Wed, 27 Oct 2021 17:19:26 GMT

It might not hurt to provide a very small example of something and show the expected results from that.

I think that you may need to sum the variables in each row such as in a data step

DATA need;
   set WORK.vaccine_hesitancy_data; 
   VacTot = sum(of vaccinefactors1-vaccinefactors11);
run;

HOWEVER, there is a concern with this data starting from a SURVEY. Does your data include weights for the observations? If so you may need to use another approach to use the weights properly after getting that total.

Re: Findind the sum of all variables

Guerraje — Wed, 27 Oct 2021 17:40:15 GMT

We have 11 choices a person could have selected as a reason for their vaccine hesitancy. We decided we want to get the sum of all of them now and create a new variable we could title it:
"vaccinefactor"
which is just the total of all the data.

Re: Findind the sum of all variables

Kurt_Bremser — Wed, 27 Oct 2021 17:46:38 GMT

Then you build the sum as already suggested.

Re: Findind the sum of all variables

Guerraje — Wed, 27 Oct 2021 17:51:31 GMT

Is there a way to calculate the sum and give that sum the variable name of "vaccinefactor" so that I can then use that variable attached to that sum to use later?

Re: Findind the sum of all variables

Reeza — Wed, 27 Oct 2021 17:55:22 GMT

DATA need;
   set WORK.vaccine_hesitancy_data; 
   VaccineFactor = sum(of vaccinefactors1-vaccinefactors11);
run;

Re: Findind the sum of all variables

Kurt_Bremser — Wed, 27 Oct 2021 18:04:45 GMT

@Guerraje wrote:
We have 11 choices a person could have selected as a reason for their vaccine hesitancy. We decided we want to get the sum of all of them now and create a new variable we could title it:
"vaccinefactor"
which is just the total of all the data.

The suggested codes do exactly that, you only need to change the name of the target variable.

Re: Findind the sum of all variables

Guerraje — Wed, 27 Oct 2021 19:24:37 GMT

When running that code, I happen to get an error that my variable is not found. I am attaching my code and my log.

Thank you

Re: Findind the sum of all variables

Kurt_Bremser — Wed, 27 Oct 2021 19:34:55 GMT

The sum variable is in the new dataset, so you need to use that in the TTEST procedure.

Re: Findind the sum of all variables

Reeza — Wed, 27 Oct 2021 19:36:51 GMT

I'm assuming your values are hardcoded for us? To show the screenshot? Otherwise you're doing something very inefficient and possibly incorrect here.

Re: Findind the sum of all variables

ballardw — Thu, 28 Oct 2021 01:28:56 GMT

@Guerraje wrote:
We have 11 choices a person could have selected as a reason for their vaccine hesitancy. We decided we want to get the sum of all of them now and create a new variable we could title it:
"vaccinefactor"
which is just the total of all the data.

Summing this coding sounds like an exercise in adding up football team names.

What exactly do some of the "numbers" mean? Even with a rating scale adding responses from multiple responses from the scale does not mean that a total of 12 for one person is anything at like a total of 12 for a different respondent.