Hotelling T Squared Value

[fevziesen]

Hi all,

I have five variables. I have calculated Mahalanobis scores of each observation. Now I want to check T Squared values of the scores. Do we have any commands in spas to calculate T values? I could only find multivariate test option to check T values. But I have my Mahalanobis scores and only want to check an univariate T2 hoteling score..

[Rich Ulrich]

The Mahalanobis distance is the multivariate distance of an observation
from a multivariate mean vector.

The Hotelling T squared is a generalization of the univariate t-test, and
it may be appropriate to test the difference between two groups...
Unless you have a particular test situation in mind, with multiple variables,
you do not have an application for the Hotelling T squared. 

The "univariate version" would be an ordinary t.

--
Rich Ulrich

> Date: Mon, 9 Feb 2015 11:21:12 -0700

> From: [hidden email]
> Subject: Hotelling T Squared Value
> To: [hidden email]
>
> Hi all,
>
> I have five variables. I have calculated Mahalanobis scores of each
> observation. Now I want to check T Squared values of the scores. Do we have
> any commands in spas to calculate T values? I could only find multivariate
> test option to check T values. But I have my Mahalanobis scores and only
> want to check an univariate T2 hoteling score..
>
>
>

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

[fevziesen]

I have calculated T-square statistics but now I want to calculate Upper Control Limit of



is it possible in spss?

[Rich Ulrich]

When I Google for UCL and beta, I come up with multivariate control
charts ... which is not where most people here are focused.   Jon may say.
The useful hit also showed that the T-squared for a single data vector is
sometimes used, as simply the square of D_M for the case.

The beta distribution is callable, so you can compute the UCL.  That formula
is given as one for those single-case T2 (along with formulas for other UCLs).  
I have no idea whether that UCL is part of some procedure.

--
Rich Ulrich

> Date: Tue, 10 Feb 2015 08:15:34 -0700

> From: [hidden email]
> Subject: Re: Hotelling T Squared Value
> To: [hidden email]
>
> I have calculated T-square statistics but now I want to calculate Upper
> Control Limit of
>
> <http://spssx-discussion.1045642.n5.nabble.com/file/n5728632/a.png>
>
> is it possible in spss?
>
>

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