## How to calculate confidence interval of mean by proc univariate

Solved
Super Contributor
Posts: 328

# How to calculate confidence interval of mean by proc univariate

[ Edited ]

in proc means one can use CLM to get the confidence interval of the mean. How to output the CLM to a dataset?

``````proc means data=adxe N NMISS CLM  ;
by trtp;
var aval;
where aval >.;
/*  output out=TFMT2RD n=n mean=mean std=sd median=median min=min max=max q1=q1 q3=q3  ;*/
run;``````

in proc univariate, what is the option to do thid? cibasic (alpha=0.05)?

what are the output options to get the 95% CI of mean?

``````proc univariate data=adxe cibasic (alpha=0.05) noprint;
by trtp;
var aval;
where aval >.;
output out=TFMT2RD n=n mean=mean std=sd median=median min=min max=max q1=q1 q3=q3  ;
run;``````

Accepted Solutions
Solution
‎03-16-2017 04:55 PM
PROC Star
Posts: 8,167

## Re: How to calculate confidence interval of mean by proc univariate

Note: You can also output them directly, but the above method is a lot more versatile. Additionally, if you use any of the special missing values (e.g., .A, .B, etc), you can capture all of them by using where not missing(variablename). e.g.:

```data class;
set sashelp.class;
if _n_ in (2,4,6) then weight=.;
else if _n_ in (3,5,7) then weight=.A;
run;

proc means data=class N NMISS CLM nway ;
class sex;
var weight;
where weight >.;
output out=want  /autoname;
run;

proc summary data=class N NMISS CLM  nway;
class sex;
var weight;
where weight >.;
output out=want2  /autoname;
run;

proc means data=class N NMISS CLM nway ;
class sex;
var weight;
where not missing(weight);
output out=want3  /autoname;
run;

proc summary data=class N NMISS CLM  nway;
class sex;
var weight;
where not missing(weight);
output out=want4  /autoname;
run;
```

Art, CEO, AnalystFinder.com

All Replies
PROC Star
Posts: 8,167

## Re: How to calculate confidence interval of mean by proc univariate

You can use the following method with any procedure:

```ods trace on;
proc means data=sashelp.class N NMISS CLM;
class sex;
var weight;
where weight >.;
/*  output out=TFMT2RD n=n mean=mean std=sd median=median min=min max=max q1=q1 q3=q3  ;*/
run;
ods trace off;

/*get the name of the desired file from the log*/
ods output Means.Summary=want;
proc means data=sashelp.class N NMISS CLM;
class sex;
var weight;
where weight >.;
/*  output out=TFMT2RD n=n mean=mean std=sd median=median min=min max=max q1=q1 q3=q3  ;*/
run;
```

Art, CEO, AnalystFinder.com

Solution
‎03-16-2017 04:55 PM
PROC Star
Posts: 8,167

## Re: How to calculate confidence interval of mean by proc univariate

Note: You can also output them directly, but the above method is a lot more versatile. Additionally, if you use any of the special missing values (e.g., .A, .B, etc), you can capture all of them by using where not missing(variablename). e.g.:

```data class;
set sashelp.class;
if _n_ in (2,4,6) then weight=.;
else if _n_ in (3,5,7) then weight=.A;
run;

proc means data=class N NMISS CLM nway ;
class sex;
var weight;
where weight >.;
output out=want  /autoname;
run;

proc summary data=class N NMISS CLM  nway;
class sex;
var weight;
where weight >.;
output out=want2  /autoname;
run;

proc means data=class N NMISS CLM nway ;
class sex;
var weight;
where not missing(weight);
output out=want3  /autoname;
run;

proc summary data=class N NMISS CLM  nway;
class sex;
var weight;
where not missing(weight);
output out=want4  /autoname;
run;
```

Art, CEO, AnalystFinder.com

Valued Guide
Posts: 505

## Re: How to calculate confidence interval of mean by proc univariate

``````I am a little confused don't you get the intervals

title 'Analysis of Female Heights';
ods select BasicIntervals;
proc univariate data=sashelp.class cibasic;
var Height;
run;

The UNIVARIATE Procedure
Variable:  HEIGHT

Basic Confidence Limits Assuming Normality

Parameter          Estimate     95% Confidence Limits

Mean               62.33684      59.86567    64.80801
Std Deviation       5.12708       3.87408     7.58204
Variance           26.28690      15.00852    57.48740

``````
☑ This topic is solved.