Non-parametric test- which one to use?

[claire]

Dear experts,

I am not sure which test I am supposed to use and was wondering whether anyone could guide me.

DATA:

Subjects chose among 'angry', 'neutral', 'sad', 'happy', and 'fear' when they were viewing angry faces (multiple times). Their empathy scores were also assessed.

MY QUESTION:

I would like to see whether people with high empathy scores (divided by the median score) and low empathy scores differed in the number of or the proportion of their responses, 'anger' and 'neutral' for the angry faces. For instance, if our hypothesis is correct, I would like to say that low empathic people rate angry faces as 'neutral' more frequently than highly empathic people...

Which test would you recommend that I use?

Thanks,
Claire

[Swank, Paul R]

If you have only two levels for your outcomes (e.g. angry, neutral) then logistic regression could be used.

Dr. Paul R. Swank,
Children's Learning Institute
Professor, Department of Pediatrics, Medical School
Adjunct Professor, School of Public Health
University of Texas Health Science Center-Houston


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of claire
Sent: Monday, November 28, 2011 9:39 PM
To: [hidden email]
Subject: Non-parametric test- which one to use?

Dear experts,

I am not sure which test I am supposed to use and was wondering whether
anyone could guide me.

DATA:

Subjects chose among 'angry', 'neutral', 'sad', 'happy', and 'fear' when
they were viewing angry faces (multiple times). Their empathy scores were
also assessed.

MY QUESTION:

I would like to see whether people with high empathy scores (divided by the
median score) and low empathy scores differed in the number of or the
proportion of their responses, 'anger' and 'neutral' for the angry faces.
For instance, if our hypothesis is correct, I would like to say that low
empathic people rate angry faces as 'neutral' more frequently than highly
empathic people...

Which test would you recommend that I use?

Thanks,
Claire

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Non-parametric-test-which-one-to-use-tp5031054p5031054.html
Sent from the SPSSX Discussion mailing list archive at 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]

Low vs. High empathy   by  Angry vs Neutral
makes up a 2x2 contingency table, which gives a
simple chi-squared test.

I would have to wonder why you do not look at the
average empathy score, instead of losing information
by creating an arbitrary dichotomy.  That would be a
simple t-test on Empathy, for the Angry vs Neutral groups.

--
Rich Ulrich

> Date: Mon, 28 Nov 2011 19:39:16 -0800

> From: [hidden email]
> Subject: Non-parametric test- which one to use?
> To: [hidden email]
>
> Dear experts,
>
> I am not sure which test I am supposed to use and was wondering whether
> anyone could guide me.
>
> DATA:
>
> Subjects chose among 'angry', 'neutral', 'sad', 'happy', and 'fear' when
> they were viewing angry faces (multiple times). Their empathy scores were
> also assessed.
>
> MY QUESTION:
>
> I would like to see whether people with high empathy scores (divided by the
> median score) and low empathy scores differed in the number of or the
> proportion of their responses, 'anger' and 'neutral' for the angry faces.
> For instance, if our hypothesis is correct, I would like to say that low
> empathic people rate angry faces as 'neutral' more frequently than highly
> empathic people...
>
> Which test would you recommend that I use?
>

[... ]

[David Marso]

I suspect MOST will agree that dichotomizing the empathy score is a truly HORRIBLE idea!
---

claire wrote

Dear experts,

I am not sure which test I am supposed to use and was wondering whether anyone could guide me.

DATA:

Subjects chose among 'angry', 'neutral', 'sad', 'happy', and 'fear' when they were viewing angry faces (multiple times). Their empathy scores were also assessed.

MY QUESTION:

I would like to see whether people with high empathy scores (divided by the median score) and low empathy scores differed in the number of or the proportion of their responses, 'anger' and 'neutral' for the angry faces. For instance, if our hypothesis is correct, I would like to say that low empathic people rate angry faces as 'neutral' more frequently than highly empathic people...

Which test would you recommend that I use?

Thanks,
Claire

[Swank, Paul R]

Yes, I was assuming that empathy would be an ordered variable, not a dichotomy. The analysis you use would then depend on which variable you see as dependent and which is the predictor. If empathy is the predictor and the proportion of angry vs. neutral responses is the outcome then logistic regression would work. But if you have no order to the variables, then you could correlated the proportion of either neutral or angry responses to the degree of empathy. I don't think the t test works here unless you categorize the angry/neutral response to angry (> 50% angry responses) which I would also not suggest.

Dr. Paul R. Swank,
Children's Learning Institute
Professor, Department of Pediatrics, Medical School
Adjunct Professor, School of Public Health
University of Texas Health Science Center-Houston


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of David Marso
Sent: Tuesday, November 29, 2011 12:25 PM
To: [hidden email]
Subject: Re: Non-parametric test- which one to use?

I suspect MOST will agree that dichotomizing the empathy score is a truly
HORRIBLE idea!
---

claire wrote

>
> Dear experts,
>
> I am not sure which test I am supposed to use and was wondering whether
> anyone could guide me.
>
> DATA:
>
> Subjects chose among 'angry', 'neutral', 'sad', 'happy', and 'fear' when
> they were viewing angry faces (multiple times). Their empathy scores were
> also assessed.
>
> MY QUESTION:
>
> I would like to see whether people with high empathy scores (divided by
> the median score) and low empathy scores differed in the number of or the
> proportion of their responses, 'anger' and 'neutral' for the angry faces.
> For instance, if our hypothesis is correct, I would like to say that low
> empathic people rate angry faces as 'neutral' more frequently than highly
> empathic people...
>
> Which test would you recommend that I use?
>
> Thanks,
> Claire
>

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Non-parametric-test-which-one-to-use-tp5031054p5033217.html
Sent from the SPSSX Discussion mailing list archive at 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

[Art Kendall]

Please explain what you mean by multiple times.
What is the medium in which the stimuli were shown?

Were they shown different angry faces? 

Where there subsets such as males/females? Adults/children? in the stimuli? In the respondents?

How did you validate the stimuli  were angry faces?

Were respondents shown multiple faces?  Did all respondents see the same stimuli? In the same order?

Were empathy scores assessed at one time per respondent?

Did respondents choose one and only one of the 4 response values? Or were they "check all that apply"

In psychology using a median split if often called "committing the invidious median split", there may be a way to use all of the information in the empathy measure.

How many respondents did you have?

Art Kendall
Social Research Consultants



On 11/28/2011 10:39 PM, claire wrote:

Dear experts,

I am not sure which test I am supposed to use and was wondering whether
anyone could guide me.

DATA:

Subjects chose among 'angry', 'neutral', 'sad', 'happy', and 'fear' when
they were viewing angry faces (multiple times). Their empathy scores were
also assessed.

MY QUESTION:

I would like to see whether people with high empathy scores (divided by the
median score) and low empathy scores differed in the number of or the
proportion of their responses, 'anger' and 'neutral' for the angry faces.
For instance, if our hypothesis is correct, I would like to say that low
empathic people rate angry faces as 'neutral' more frequently than highly
empathic people...

Which test would you recommend that I use?

Thanks,
Claire

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Non-parametric-test-which-one-to-use-tp5031054p5031054.html
Sent from the SPSSX Discussion mailing list archive at 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