How to assess for potential categorial confounding variables and control for categorial confounding variables

[khforensicpsych]

Hi,

I am having difficulty trying to determine how to assess for potential confounding variables in my study. I have 4 demographics variables (all categorical: gender, age, ethnicity, and political affiliation) that might be significantly related to my 3 dependent variables (two are continuous and one is categorical). From my understanding, I have 12 analyses to run (all 4 demographics variables against each of the 3 DVs). However, I'm not sure how to go about this. Since both variables in each analysis will not be continuous, I can't use Pearson correlations. Since 8 of the analyses include a continuous variable I cannot use Chi-squares (and for the other 4 I do not have 5 cases in each cell). Since 3 of my 4 demographics variables are not dichotomous I can't seem to get the univariate analyses to work. I feel very stuck!

Additionally, as I am reading up on ANCOVAs I have realized that I can't even use any of my potential confounding variables because they are not continuous. So, how would I go about controlling for these variables? I have 2 hypotheses that include one-way ANOVAs (DV: level of guilt affected by IV: condition, 4 levels, and DV: level of confidence affected by IV: condition, 4 levels). I also have 2 hypotheses for which I planned to use Chi-Squares (effect of IV: condition on DV: verdict decisions, dichotomous guilty or not guilty, and effect of IV: verdict decision (dichotomous) on DV: level of confidence. So for example, I want to make sure that DV: level of guilt is affected only by IV: condition, and not age, gender, ethnicity, or political affiliation.

Any assistance would be greatly appreciated!!

[Art Kendall]

Please describe each of your variables: name, variable label, value labels, etc.

What is a case? How many to you have? How did you select them?

Is this a class assignment?


On 4/29/2013 3:11 PM, khforensicpsych [via SPSSX Discussion] wrote:

Hi,

I am having difficulty trying to determine how to assess for potential confounding variables in my study. I have 4 demographics variables (all categorical: gender, age, ethnicity, and political affiliation) that might be significantly related to my 3 dependent variables (two are continuous and one is categorical). From my understanding, I have 12 analyses to run (all 4 demographics variables against each of the 3 DVs). However, I'm not sure how to go about this. Since both variables in each analysis will not be continuous, I can't use Pearson correlations. Since 8 of the analyses include a continuous variable I cannot use Chi-squares (and for the other 4 I do not have 5 cases in each cell). Since 3 of my 4 demographics variables are not dichotomous I can't seem to get the univariate analyses to work. I feel very stuck!

Additionally, as I am reading up on ANCOVAs I have realized that I can't even use any of my potential confounding variables because they are not continuous. So, how would I go about controlling for these variables? I have 2 hypotheses that include one-way ANOVAs (DV: level of guilt affected by IV: condition, 4 levels, and DV: level of confidence affected by IV: condition, 4 levels). I also have 2 hypotheses for which I planned to use Chi-Squares (effect of IV: condition on DV: verdict decisions, dichotomous guilty or not guilty, and effect of IV: verdict decision (dichotomous) on DV: level of confidence. So for example, I want to make sure that DV: level of guilt is affected only by IV: condition, and not age, gender, ethnicity, or political affiliation.

Any assistance would be greatly appreciated!!

---

If you reply to this email, your message will be added to the discussion below:

http://spssx-discussion.1045642.n5.nabble.com/How-to-assess-for-potential-categorial-confounding-variables-and-control-for-categorial-confounding-s-tp5719824.html

To start a new topic under SPSSX Discussion, email [hidden email]
To unsubscribe from SPSSX Discussion, click here.
NAML

[khforensicpsych]

It is a quasi-experimental design because I was unable to randomly assign
participants to groups. I had psychology classes participate, and randomly
assigned the classes to condition.

I have 6 hypotheses total:

1.DV:  Level of guilt (determined from a 7point Likert scale ranging from
not at all guilty (1) to definitely guilty (7)) will be significantly
affected by Condition (behaviors) (1 = eye contact, 2 = fidgeting, 3 =
sweating, and 4 = control group). I ran a one-way ANOVA, which wasn't
significant.

Hypotheses 2 and 3 were planned comparisons, which I haven't done because
the initial ANOVA wasn't significant.

4. DV: verdict decision (1 = guilty or 2 = not guilty) will be affected by
IV: condition (eye contact, fidgeting, sweating, control). I ran a
chi-square and it wasn't significant.

5. DV: level of confidence (determined from a 7point Likert scale ranging
from not at all confident (1) to definitely confident (7) will be
significantly associated with IV: verdict decision (guilty or not guilty).
I ran a chi-square for this and it was approaching significance (p=.06),
although I was told that I can also run an ANOVA, which I did and this was
significant (p<.05).

6. DV: level of confidence DV: level of confidence (determined from a
7point Likert scale ranging from not at all confident (1) to definitely
confident (7) will be significantly affected by IV: condition (eye contact,
fidgeting, sweating, control). I ran a one-way ANOVA and it was not
significant.

I was using my DVs level of guilt and level of confidence as interval
variables in order to run the ANOVAs.

I have 22 cases in my eye contact condition, 21 in fidgeting condition, 22
in sweating condition, and 31 in control condition.

After completing the analyses I went to write it up, but realized I forgot
to run the preliminary stats I discussed in my proposal to ensure that some
of my demographics variables were not significantly related to my DVs. Past
research has shown that age, gender, ethnicity, and political affiliation
can play a role in whether someone determines a defendant to be guilty or
not guilty, so I wanted to control for these when running my analyses, but
I can't figure out how to go about doing this.

Thank you for your help,

Katrina



On Mon, Apr 29, 2013 at 12:24 PM, Art Kendall [via SPSSX Discussion] <
[e-mail]> wrote:

>
> Please describe each of your variables: name, variable label, value
> labels, etc.
>
> What is a case? How many to you have? How did you select them?
>
> Is this a class assignment?
>
>
> Art Kendall
> Social Research Consultants
>
> On 4/29/2013 3:11 PM, khforensicpsych [via SPSSX Discussion] wrote:
>
> Hi,
>
> I am having difficulty trying to determine how to assess for potential
> confounding variables in my study. I have 4 demographics variables (all
> categorical: gender, age, ethnicity, and political affiliation) that might
> be significantly related to my 3 dependent variables (two are continuous
> and one is categorical). From my understanding, I have 12 analyses to run
> (all 4 demographics variables against each of the 3 DVs). However, I'm not
> sure how to go about this. Since both variables in each analysis will not
> be continuous, I can't use Pearson correlations. Since 8 of the analyses
> include a continuous variable I cannot use Chi-squares (and for the other 4
> I do not have 5 cases in each cell). Since 3 of my 4 demographics variables
> are not dichotomous I can't seem to get the univariate analyses to work. I
> feel very stuck!
>
> Additionally, as I am reading up on ANCOVAs I have realized that I can't
> even use any of my potential confounding variables because they are not
> continuous. So, how would I go about controlling for these variables? I
> have 2 hypotheses that include one-way ANOVAs (DV: level of guilt affected
> by IV: condition, 4 levels, and DV: level of confidence affected by IV:
> condition, 4 levels). I also have 2 hypotheses for which I planned to use
> Chi-Squares (effect of IV: condition on DV: verdict decisions, dichotomous
> guilty or not guilty, and effect of IV: verdict decision (dichotomous) on
> DV: level of confidence. So for example, I want to make sure that DV: level
> of guilt is affected only by IV: condition, and not age, gender, ethnicity,
> or political affiliation.
>
> Any assistance would be greatly appreciated!!
>
> ------------------------------
>  If you reply to this email, your message will be added to the discussion
> below:
>
> http://spssx-discussion.1045642.n5.nabble.com/How-to-assess-for-potential-categorial-confounding-variables-and-control-for-categorial-confounding-s-tp5719824.html
>  To start a new topic under SPSSX Discussion, email [hidden email]<http://user/SendEmail.jtp?type=node&node=5719827&i=0>
> To unsubscribe from SPSSX Discussion, click here.
> NAML<http://spssx-discussion.1045642.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml>
>
>
>  Art Kendall
> Social Research Consultants
>
>
> ------------------------------
>  If you reply to this email, your message will be added to the discussion
> below:
>
> http://spssx-discussion.1045642.n5.nabble.com/How-to-assess-for-potential-categorial-confounding-variables-and-control-for-categorial-confounding-s-tp5719824p5719827.html
>  To unsubscribe from How to assess for potential categorial confounding
> variables and control for categorial confounding variables, click here<http://spssx-discussion.1045642.n5.nabble.com/template/NamlServlet.jtp?macro=unsubscribe_by_code&node=5719824&code=a2xoNTYxOUBlZ28udGhlY2hpY2Fnb3NjaG9vbC5lZHV8NTcxOTgyNHw1NjAxNTczOTg=>
> .
> NAML<http://spssx-discussion.1045642.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml>
>

[Art Kendall]

Please the variables you want to control for.
On 4/29/2013 3:40 PM, khforensicpsych [via SPSSX Discussion] wrote:

This is my dissertation. Unfortunately I have had limited help from my committee.

 

It is a quasi-experimental design because I was unable to randomly assign participants to groups. I had psychology classes participate, and randomly assigned the classes to condition.

 

I have 6 hypotheses total:

 

1.DV:  Level of guilt (determined from a 7point Likert scale ranging from not at all guilty (1) to definitely guilty (7)) will be significantly affected by Condition (behaviors) (1 = eye contact, 2 = fidgeting, 3 = sweating, and 4 = control group). I ran a one-way ANOVA, which wasn't significant.

 

Hypotheses 2 and 3 were planned comparisons, which I haven't done because the initial ANOVA wasn't significant.

 

4. DV: verdict decision (1 = guilty or 2 = not guilty) will be affected by IV: condition (eye contact, fidgeting, sweating, control). I ran a chi-square and it wasn't significant.

 

5. DV: level of confidence (determined from a 7point Likert scale ranging from not at all confident (1) to definitely confident (7) will be significantly associated with IV: verdict decision (guilty or not guilty). I ran a chi-square for this and it was approaching significance (p=.06), although I was told that I can also run an ANOVA, which I did and this was significant (p<.05).

 

6. DV: level of confidence DV: level of confidence (determined from a 7point Likert scale ranging from not at all confident (1) to definitely confident (7) will be significantly affected by IV: condition (eye contact, fidgeting, sweating, control). I ran a one-way ANOVA and it was not significant.

 

I was using my DVs level of guilt and level of confidence as interval variables in order to run the ANOVAs.

 

I have 22 cases in my eye contact condition, 21 in fidgeting condition, 22 in sweating condition, and 31 in control condition.

 

After completing the analyses I went to write it up, but realized I forgot to run the preliminary stats I discussed in my proposal to ensure that some of my demographics variables were not significantly related to my DVs. Past research has shown that age, gender, ethnicity, and political affiliation can play a role in whether someone determines a defendant to be guilty or not guilty, so I wanted to control for these when running my analyses, but I can't figure out how to go about doing this.

 

Thank you for your help,

 

Katrina

 

On Mon, Apr 29, 2013 at 12:24 PM, Art Kendall [via SPSSX Discussion] <[hidden email]> wrote:

Please describe each of your variables: name, variable label, value labels, etc.

What is a case? How many to you have? How did you select them?

Is this a class assignment?

Art Kendall
Social Research Consultants

On 4/29/2013 3:11 PM, khforensicpsych [via SPSSX Discussion] wrote:

Hi,

I am having difficulty trying to determine how to assess for potential confounding variables in my study. I have 4 demographics variables (all categorical: gender, age, ethnicity, and political affiliation) that might be significantly related to my 3 dependent variables (two are continuous and one is categorical). From my understanding, I have 12 analyses to run (all 4 demographics variables against each of the 3 DVs). However, I'm not sure how to go about this. Since both variables in each analysis will not be continuous, I can't use Pearson correlations. Since 8 of the analyses include a continuous variable I cannot use Chi-squares (and for the other 4 I do not have 5 cases in each cell). Since 3 of my 4 demographics variables are not dichotomous I can't seem to get the univariate analyses to work. I feel very stuck!

Additionally, as I am reading up on ANCOVAs I have realized that I can't even use any of my potential confounding variables because they are not continuous. So, how would I go about controlling for these variables? I have 2 hypotheses that include one-way ANOVAs (DV: level of guilt affected by IV: condition, 4 levels, and DV: level of confidence affected by IV: condition, 4 levels). I also have 2 hypotheses for which I planned to use Chi-Squares (effect of IV: condition on DV: verdict decisions, dichotomous guilty or not guilty, and effect of IV: verdict decision (dichotomous) on DV: level of confidence. So for example, I want to make sure that DV: level of guilt is affected only by IV: condition, and not age, gender, ethnicity, or political affiliation.

Any assistance would be greatly appreciated!!

---

If you reply to this email, your message will be added to the discussion below:

http://spssx-discussion.1045642.n5.nabble.com/How-to-assess-for-potential-categorial-confounding-variables-and-control-for-categorial-confounding-s-tp5719824.html

To start a new topic under SPSSX Discussion, email [hidden email]
To unsubscribe from SPSSX Discussion, click here.
NAML

Art Kendall
Social Research Consultants

---

If you reply to this email, your message will be added to the discussion below:

http://spssx-discussion.1045642.n5.nabble.com/How-to-assess-for-potential-categorial-confounding-variables-and-control-for-categorial-confounding-s-tp5719824p5719827.html

To unsubscribe from How to assess for potential categorial confounding variables and control for categorial confounding variables, click here.
NAML

--

Katrina Hodgson

Doctoral Student, Clinical Forensic Psychology

Faculty Assistant, Dean Rishel, PhD (Department Chair)
Forensic Ambassador - Irvine Campus
The Chicago School of Professional Psychology
4199 Campus Drive, Suite E
Irvine, CA 92612
Cell: [phone]
[hidden email]

---

If you reply to this email, your message will be added to the discussion below:

http://spssx-discussion.1045642.n5.nabble.com/How-to-assess-for-potential-categorial-confounding-variables-and-control-for-categorial-confounding-s-tp5719824p5719829.html

To start a new topic under SPSSX Discussion, email [hidden email]
To unsubscribe from SPSSX Discussion, click here.
NAML

[khforensicpsych]

I would like to control for the following demographic variables:

1. Gender (1 = male; 2 = female)
2. Age (1 = 18-24; 2 = 25-34; 3 = 35-44; 4 = 45-54; 5 = 55-64)
3. Ethnicity (1 = White; 2 = Hispanic/Latino; 3 = African American; 4 =
Asian-American; 5 = Mixed-Multiracial; 6 = Other)
4. Political Affiliation (1 = Republican; 2 = Democrat; 3 = Other)

On Mon, Apr 29, 2013 at 12:57 PM, Art Kendall [via SPSSX Discussion] <
[e-mail]> wrote:

> Please the variables you want to control for.
>
> Art Kendall
> Social Research Consultants
>
> On 4/29/2013 3:40 PM, khforensicpsych [via SPSSX Discussion] wrote:
>
> This is my dissertation. Unfortunately I have had limited help from my
> committee.
>
> It is a quasi-experimental design because I was unable to randomly assign
> participants to groups. I had psychology classes participate, and randomly
> assigned the classes to condition.
>
> I have 6 hypotheses total:
>
> 1.DV:  Level of guilt (determined from a 7point Likert scale ranging from
> not at all guilty (1) to definitely guilty (7)) will be significantly
> affected by Condition (behaviors) (1 = eye contact, 2 = fidgeting, 3 =
> sweating, and 4 = control group). I ran a one-way ANOVA, which wasn't
> significant.
>
> Hypotheses 2 and 3 were planned comparisons, which I haven't done because
> the initial ANOVA wasn't significant.
>
> 4. DV: verdict decision (1 = guilty or 2 = not guilty) will be affected by
> IV: condition (eye contact, fidgeting, sweating, control). I ran a
> chi-square and it wasn't significant.
>
> 5. DV: level of confidence (determined from a 7point Likert scale ranging
> from not at all confident (1) to definitely confident (7) will be
> significantly associated with IV: verdict decision (guilty or not guilty).
> I ran a chi-square for this and it was approaching significance (p=.06),
> although I was told that I can also run an ANOVA, which I did and this was
> significant (p<.05).
>
> 6. DV: level of confidence DV: level of confidence (determined from a
> 7point Likert scale ranging from not at all confident (1) to definitely
> confident (7) will be significantly affected by IV: condition (eye contact,
> fidgeting, sweating, control). I ran a one-way ANOVA and it was not
> significant.
>
> I was using my DVs level of guilt and level of confidence as interval
> variables in order to run the ANOVAs.
>
> I have 22 cases in my eye contact condition, 21 in fidgeting condition, 22
> in sweating condition, and 31 in control condition.
>
> After completing the analyses I went to write it up, but realized I forgot
> to run the preliminary stats I discussed in my proposal to ensure that some
> of my demographics variables were not significantly related to my DVs. Past
> research has shown that age, gender, ethnicity, and political affiliation
> can play a role in whether someone determines a defendant to be guilty or
> not guilty, so I wanted to control for these when running my analyses, but
> I can't figure out how to go about doing this.
>
> Thank you for your help,
>
> Katrina
>
>
>
> On Mon, Apr 29, 2013 at 12:24 PM, Art Kendall [via SPSSX Discussion] <[hidden
> email] <http://user/SendEmail.jtp?type=node&node=5719829&i=0>> wrote:
>
>>
>> Please describe each of your variables: name, variable label, value
>> labels, etc.
>>
>> What is a case? How many to you have? How did you select them?
>>
>> Is this a class assignment?
>>
>>
>> Art Kendall
>> Social Research Consultants
>>
>> On 4/29/2013 3:11 PM, khforensicpsych [via SPSSX Discussion] wrote:
>>
>> Hi,
>>
>> I am having difficulty trying to determine how to assess for potential
>> confounding variables in my study. I have 4 demographics variables (all
>> categorical: gender, age, ethnicity, and political affiliation) that might
>> be significantly related to my 3 dependent variables (two are continuous
>> and one is categorical). From my understanding, I have 12 analyses to run
>> (all 4 demographics variables against each of the 3 DVs). However, I'm not
>> sure how to go about this. Since both variables in each analysis will not
>> be continuous, I can't use Pearson correlations. Since 8 of the analyses
>> include a continuous variable I cannot use Chi-squares (and for the other 4
>> I do not have 5 cases in each cell). Since 3 of my 4 demographics variables
>> are not dichotomous I can't seem to get the univariate analyses to work. I
>> feel very stuck!
>>
>> Additionally, as I am reading up on ANCOVAs I have realized that I can't
>> even use any of my potential confounding variables because they are not
>> continuous. So, how would I go about controlling for these variables? I
>> have 2 hypotheses that include one-way ANOVAs (DV: level of guilt affected
>> by IV: condition, 4 levels, and DV: level of confidence affected by IV:
>> condition, 4 levels). I also have 2 hypotheses for which I planned to use
>> Chi-Squares (effect of IV: condition on DV: verdict decisions, dichotomous
>> guilty or not guilty, and effect of IV: verdict decision (dichotomous) on
>> DV: level of confidence. So for example, I want to make sure that DV: level
>> of guilt is affected only by IV: condition, and not age, gender, ethnicity,
>> or political affiliation.
>>
>> Any assistance would be greatly appreciated!!
>>
>> ------------------------------
>>  If you reply to this email, your message will be added to the
>> discussion below:
>>
>> http://spssx-discussion.1045642.n5.nabble.com/How-to-assess-for-potential-categorial-confounding-variables-and-control-for-categorial-confounding-s-tp5719824.html
>>  To start a new topic under SPSSX Discussion, email [hidden email]<http://user/SendEmail.jtp?type=node&node=5719827&i=0>
>> To unsubscribe from SPSSX Discussion, click here.
>> NAML<http://spssx-discussion.1045642.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml>
>>
>>
>>  Art Kendall
>> Social Research Consultants
>>
>>
>> ------------------------------
>>  If you reply to this email, your message will be added to the
>> discussion below:
>>
>> http://spssx-discussion.1045642.n5.nabble.com/How-to-assess-for-potential-categorial-confounding-variables-and-control-for-categorial-confounding-s-tp5719824p5719827.html
>>  To unsubscribe from How to assess for potential categorial confounding
>> variables and control for categorial confounding variables, click here.
>> NAML<http://spssx-discussion.1045642.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml>
>>
>
>
>
> --
>
>
> ------------------------------
>  If you reply to this email, your message will be added to the discussion
> below:
>
> http://spssx-discussion.1045642.n5.nabble.com/How-to-assess-for-potential-categorial-confounding-variables-and-control-for-categorial-confounding-s-tp5719824p5719829.html
>  To start a new topic under SPSSX Discussion, email [hidden email]<http://user/SendEmail.jtp?type=node&node=5719831&i=0>
> To unsubscribe from SPSSX Discussion, click here.
> NAML<http://spssx-discussion.1045642.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml>
>
>
>  Art Kendall
> Social Research Consultants
>
>
> ------------------------------
>  If you reply to this email, your message will be added to the discussion
> below:
>
> http://spssx-discussion.1045642.n5.nabble.com/How-to-assess-for-potential-categorial-confounding-variables-and-control-for-categorial-confounding-s-tp5719824p5719831.html
>  To unsubscribe from How to assess for potential categorial confounding
> variables and control for categorial confounding variables, click here<http://spssx-discussion.1045642.n5.nabble.com/template/NamlServlet.jtp?macro=unsubscribe_by_code&node=5719824&code=a2xoNTYxOUBlZ28udGhlY2hpY2Fnb3NjaG9vbC5lZHV8NTcxOTgyNHw1NjAxNTczOTg=>
> .
> NAML<http://spssx-discussion.1045642.n5.nabble.com/template/NamlServlet.jtp?macro=macro_viewer&id=instant_html%21nabble%3Aemail.naml&base=nabble.naml.namespaces.BasicNamespace-nabble.view.web.template.NabbleNamespace-nabble.view.web.template.NodeNamespace&breadcrumbs=notify_subscribers%21nabble%3Aemail.naml-instant_emails%21nabble%3Aemail.naml-send_instant_email%21nabble%3Aemail.naml>
>