Non-Parametric Multivariate Testing

[betty]

Hello all,

I have some trouble finding the right test/model to see if there is a
relation between some explanatory variables with one dependent variable.

I already considered using Linear Regression and Logistic Regression. As my
data does not meet all the assumptions for the Linear Regression and as the
dependend variable is not binary, both of the Linear Regression and Logistic
regression aren't optional.

So my Question is: What Non-Parametric test for Multivariate testing can I
use? Or is there a way so that I can still use the Linear Regression?

Gr Betty



--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

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

[bdates]

If your dependent variable is ordinal, you're probably best using ordinal logistic regression. This can be done in SPSS using PLUM, which is part of the Basic Statistics Package. UCLA has a website with a guide to doing ordinal logistic regression in SPSS. See the link below.

https://stats.idre.ucla.edu/spss/output/multinomial-logistic-regression/

Multinomial Logistic Regression | SPSS Annotated Output

stats.idre.ucla.edu

This page shows an example of a multinomial logistic regression analysis with footnotes explaining the output. The data were collected on 200 high school students and are scores on various tests, including a video game and a puzzle.

Brian

---

From: SPSSX(r) Discussion <[hidden email]> on behalf of betty <[hidden email]>
Sent: Tuesday, December 11, 2018 11:18:46 AM
To: [hidden email]
Subject: Non-Parametric Multivariate Testing

 

Hello all,

I have some trouble finding the right test/model to see if there is a
relation between some explanatory variables with one dependent variable.

I already considered using Linear Regression and Logistic Regression. As my
data does not meet all the assumptions for the Linear Regression and as the
dependend variable is not binary, both of the Linear Regression and Logistic
regression aren't optional.

So my Question is: What Non-Parametric test for Multivariate testing can I
use? Or is there a way so that I can still use the Linear Regression?

Gr Betty



--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

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

[Rich Ulrich]

What assumptions for linear regression make you think you have a problem?

What do people in your area usually do?

The "rank" transformation, which is what most people think of as

"nonparametric", is not very congenial for multiple variables, whether

you talk of the DV or the IV. 

If the problem is unequal intervals/ unequal variances, conventional

transformations (like taking logs) should be considered.

So, what is the actual problem?

--

Rich Ulrich

---

From: SPSSX(r) Discussion <[hidden email]> on behalf of betty <[hidden email]>
Sent: Tuesday, December 11, 2018 11:18 AM
To: [hidden email]
Subject: Non-Parametric Multivariate Testing

 

Hello all,

I have some trouble finding the right test/model to see if there is a
relation between some explanatory variables with one dependent variable.

I already considered using Linear Regression and Logistic Regression. As my
data does not meet all the assumptions for the Linear Regression and as the
dependend variable is not binary, both of the Linear Regression and Logistic
regression aren't optional.

So my Question is: What Non-Parametric test for Multivariate testing can I
use? Or is there a way so that I can still use the Linear Regression?

Gr Betty



--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

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

[Bruce Weaver]

Hello Betty.  Like Rich, I wonder what the actual problem is.  

Many people believe that OLS linear regression requires a normally
distributed dependent variable.  But there is no such requirement.  Good
stats texts typically state that the *errors* must be independently and
identically distributed as normal (with Mean=0 and variance = sigma^2).  But
even this is not quite correct, because it implies that normality of the
errors is a *necessary* condition.  IMO, however, normality of the errors is
only a sufficient condition.  The truly necessary condition is that the
sampling distributions of the coefficients are (approximately) normal--and
given a large enough sample size, they will be, even if the errors are not
normally distributed.  

I've just updated a set of slides you may find useful.  You can view them
here.

https://sites.google.com/a/lakeheadu.ca/bweaver/Home/statistics/files/OLS_regression_assumptions_Wooldridge.pdf

HTH.



betty wrote

> Hello all,
>
> I have some trouble finding the right test/model to see if there is a
> relation between some explanatory variables with one dependent variable.
>
> I already considered using Linear Regression and Logistic Regression. As
> my
> data does not meet all the assumptions for the Linear Regression and as
> the
> dependend variable is not binary, both of the Linear Regression and
> Logistic
> regression aren't optional.
>
> So my Question is: What Non-Parametric test for Multivariate testing can I
> use? Or is there a way so that I can still use the Linear Regression?
>
> Gr Betty
>
>
>
> --
> Sent from: http://spssx-discussion.1045642.n5.nabble.com/
>
> =====================
> To manage your subscription to SPSSX-L, send a message to

> LISTSERV@.UGA

>  (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





-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.nabble.com/

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