The other side of the confidence interval for Cronbach's Alpha?

[zena mello]

I would like to report both ends of the confidence interval for a Cronbach's alpha. The command cialpha (STB-56 sg143) provides only one end. Is there a command for both?

[Joseph Coveney]

Would something like the following suit your purposes?

.version13.1

.
.clear*

.setmoreoff

.setseed`=date("2014-06-18","YMD")'

.
.webuseautomiss
(1978AutomobileData)

.bootstrapalpha=r(alpha),reps(400)nodotsnowarn:///
>alphapriceheadroomrep78trunkweightlengthturndispl

BootstrapresultsNumberofobs=74
Replications=400

command:alphapriceheadroomrep78trunkweightlengthturndispl
alpha:r(alpha)

------------------------------------------------------------------------------
|ObservedBootstrapNormal-based
|Coef.Std.Err.zP>|z|[95%Conf.Interval]
-------------+----------------------------------------------------------------
alpha|.2654521.03689387.200.000.1931416.3377626
------------------------------------------------------------------------------

.
.estatbootstrap,all

BootstrapresultsNumberofobs=74
Replications=400

command:alphapriceheadroomrep78trunkweightlengthturndispl
alpha:r(alpha)

------------------------------------------------------------------------------
|ObservedBootstrap
|Coef.BiasStd.Err.[95%Conf.Interval]
-------------+----------------------------------------------------------------
alpha|.26545212-.0034275.03689379.1931416.3377626(N)
|.1838564.3267355(P)
|.1933084.3382243(BC)
------------------------------------------------------------------------------
(N)normalconfidenceinterval
(P)percentileconfidenceinterval
(BC)bias-correctedconfidenceinterval

.
.exit

endofdo-file


.

[zena mello]

Yes, it totally worked! That is just great. Thanks.