confidence intervals

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

confidence intervals

Kevin Horton-5
Hi,

I'm trying to do 2 things in SPSS (version 13.0) but am having problems.
Any guidance would be much appreciated.

1. generating 95% confidence intervals for geometric means, and
2. generating 95% confidence intervals for percentiles (i.e. 95th
percentiles).

Thx.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Reply | Threaded
Open this post in threaded view
|

Re: confidence intervals

Marta Garcia-Granero
Hi Kevin.

A partial answer:

> I'm trying to do 2 things in SPSS (version 13.0) but am having problems.
> Any guidance would be much appreciated.
>
> 1. generating 95% confidence intervals for geometric means, and
>

This one is easy. Log transform your variable, compute a 95%CI for the
arithmetic mean (part of the output of EXAMINE procedure) and anti-log
the lower and upper limits.

> 2. generating 95% confidence intervals for percentiles (i.e. 95th
> percentiles).
>
That one is going to be harder. Let me Google a bit to find the formula,
and I think some MATRIX code will do the trick.


Regards,
Marta GarcĂ­a-Granero

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Reply | Threaded
Open this post in threaded view
|

Re: confidence intervals

Marta Garcia-Granero
In reply to this post by Kevin Horton-5
Hi Kevin
> I'm trying to do 2 things in SPSS (version 13.0) but am having problems.
> Any guidance would be much appreciated.
>
> 1. generating 95% confidence intervals for geometric means, and
> 2. generating 95% confidence intervals for percentiles (i.e. 95th
> percentiles).
>

OK, some approaches for question #2:

- See the method implemented here:
http://www2.napier.ac.uk/depts/fhls/peas/percentiles.htm
- Bootstrapping. take a look here
.http://www.spsstools.net/SampleSyntax.htm#Bootstrap

HTH,
Marta

> Thx.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD