Syntax for Jackknifing and Bootstrapping Correlations

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

Syntax for Jackknifing and Bootstrapping Correlations

bdates

Hi, all.


I just updated my Google Drive with syntax for jackknifing and bootstrapping a correlation. Again, it produces results for the jackknife and 7 bootstrap solutions. There is no Bayesian bootstrap solution for the correlation, at least one that I could find. I'd be happy to implement it if you can point me in the right direction, but I think because the Bayesian approach uses the posterior probabilities rather than raw data, the solution is still a long way off. Graphs of  both the jackknife and bootstrap distributions are also produced. I used the standard recommended approach to computing the mean for correlations, converting first the jackknife/bootstrap results using Fisher's z-transformation, computing the mean, and then using the inverse z-transform for the final solution. I've also, based on some work by DiCiccio and Efron, included both the length and the right-left asymmetry of the confidence interval.


Here's the link again. Happy computing.




Brian
===================== 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: Syntax for Jackknifing and Bootstrapping Correlations

WillBaileyz

Thanks Brian

 

Will

WMB & Associates

Statistical Services

============

mailto: [hidden email]

http://home.earthlink.net/~z_statman

============

 

 

> On 6/12/2018 11:14:39 AM, Dates, Brian ([hidden email]) wrote:
> > Hi, all.
>
>
> I just updated my Google Drive with syntax for jackknifing and bootstrapping a correlation. Again, it produces results for the jackknife and 7 bootstrap solutions. There is no Bayesian bootstrap solution for the correlation, at least one that I could find. I'd be happy to implement it if you can point me in the right direction, but I think because the Bayesian approach uses the posterior probabilities rather than raw data, the solution is still a long way off. Graphs of  both the jackknife and bootstrap distributions are also produced. I used the standard recommended approach to computing the mean for correlations, converting first the jackknife/bootstrap results using Fisher's z-transformation, computing the mean, and then using the inverse z-transform for the final solution. I've also, based on some work by DiCiccio and Efron, included both the length and the right-left asymmetry of the confidence interval.

>
>
> Here's the link again. Happy computing.
>
>
> https://drive.google.com/drive/folders/14tOFH4ViFpGZ78gZefoOFWKLt6hcrDC8?usp=sharing
>
>
>
> Brian
>
> =====================
> 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

===================== 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