95% CI when _BIN_ values is 1

[chinna0369]

Hi all, 

How to take 95% CI when _BIN_ value is 1 like below.

Thanks in advance!

Adithya

[Watts]

Are you asking how PROC FREQ computes the exact confidence limits for the binomial proportion?

This computation is documented here:

Binomial Proportion: Exact (Clopper-Pearson) Confidence Limits 

When the observed proportion is 1, the upper confidence limit is set to 1, and the lower limit is computed as described for P_L.

[chinna0369 · Accepted solution]

@Watts , I got the answer. When BIN value is 1 then we need to take inverse values which means we need to subtract those values from 1.

Thanks,

Adithya

[chinna0369]

https://www.pharmasug.org/proceedings/2014/IB/PharmaSUG-2014-IB05.pdf paper gives clear info.

[Watts]

You can specify the level (for which to compute the binomial proportion) by using the

BINOMIAL(LEVEL=) option in the PROC FREQ TABLES statement.

For example,

tables Response / binomial(level='yes');