Forced choice (a or b) locus of control scale-ordinal or interval?

[Basia]

Hi,
I need to identify level of measurement for a locus of control scale (Rotter Internal External Locus of Control Scale) which is a 29 item forced choice (select answer a or b) self-report scale with the scoring that range from 0(internality) to 23 (externality). I's struggling identifying the level of measurement. Is it ordinal or interval? I'm thinking it's more internal due to arbitrary 0. Could anyone help and explain this to me?

Thank you!
Basia
 

[Art Kendall]

look at the norms for the test, how are they expressed?

In the articles you used to prepare your study, what kinds of statistical techniques did they use?

is the underlying construct continuous?

Do you have a reason to expect that the intervals between values are severely discrepant from each other?

Are you using the scale as previously developed in a study or are you trying to further develop the scale?

Are you concerned about the level of measurement of each item or the level of measurement of the scale score?

[Basia]

Hi,

 

This is for my quantitative research proposal (school assignment) in which I  propose to do a study on the impact mindfulness training has on stress levels and burnout among externally oriented mental health counselors. Internals tend to cope with stress better than externals.

I wanted to use the internal-external locus of control to identify counselors who are externally oriented and later to see if they became more internally oriented post intervention. I think that when I look at the scale items, they are ordinal as the distance between answer a or b cannot be known, perhaps only assumed. I think that when I look at the total scores, I can treat the data as interval. 

I need to know the level of measurement for the entire scale, not individual items. 

The researchers who developed the scale provided information on means and standard deviations for samples of several populations.

I appreciate your time and help!

Basia

On Wednesday, October 1, 2014 8:23 AM, Art Kendall [via SPSSX Discussion] <[hidden email]> wrote:

look at the norms for the test, how are they expressed?

In the articles you used to prepare your study, what kinds of statistical techniques did they use?

is the underlying construct continuous?

Do you have a reason to expect that the intervals between values are severely discrepant from each other?

Are you using the scale as previously developed in a study or are you trying to further develop the scale?

Are you concerned about the level of measurement of each item or the level of measurement of the scale score?

---

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

http://spssx-discussion.1045642.n5.nabble.com/Forced-choice-a-or-b-locus-of-control-scale-ordinal-or-interval-tp5727458p5727460.html

To unsubscribe from Forced choice (a or b) locus of control scale-ordinal or interval?, click here.
NAML

[Bruce Weaver]

The scale you are talking about and scoring instructions can be seen here:

http://www.mccc.edu/~jenningh/Courses/documents/Rotter-locusofcontrolhandout.pdf

Notice that although there are 29 items, only 23 of them are used in scoring.  And the score is a COUNT of selections that are consistent with external locus of control.  

Therefore, some people might argue that you should use some type of regression model suitable for counts (e.g., Poisson or negative binomial regression).  But in my (somewhat limited) experience with count variables, when the distribution of counts is fairly symmetrical with a mean well above 0, OLS regression may actually work better than Poisson or NB regression.  The Poisson distribution does converge on the Normal as the mean increases, and IIRC, with a mean of 10, the normal approximation is already pretty good.  

One other thing:  You said the developers reported means & SDs.  That's a sign that they were treating it as a quantitative scale suitable for analysis with t-tests, ANOVA, linear regression, etc.  (Note that the mean is the intercept in an intercept-only OLS regression model.)  

HTH.

Basia wrote

Hi,
 
This is for my quantitative research proposal (school assignment) in which I  propose to do a study on the impact mindfulness training has on stress levels and burnout among externally oriented mental health counselors. Internals tend to cope with stress better than externals.
I wanted to use the internal-external locus of control to identify counselors who are externally oriented and later to see if they became more internally oriented post intervention. I think that when I look at the scale items, they are ordinal as the distance between answer a or b cannot be known, perhaps only assumed. I think that when I look at the total scores, I can treat the data as interval.
I need to know the level of measurement for the entire scale, not individual items.
The researchers who developed the scale provided information on means and standard deviations for samples of several populations.

I appreciate your time and help!
Basia



On Wednesday, October 1, 2014 8:23 AM, Art Kendall [via SPSSX Discussion] <[hidden email]> wrote:
 


look at the norms for the test, how are they expressed?

In the articles you used to prepare your study, what kinds of statistical techniques did they use?

is the underlying construct continuous?

Do you have a reason to expect that the intervals between values are severely discrepant from each other?

Are you using the scale as previously developed in a study or are you trying to further develop the scale?

Are you concerned about the level of measurement of each item or the level of measurement of the scale score?





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/Forced-choice-a-or-b-locus-of-control-scale-ordinal-or-interval-tp5727458p5727460.html 
To unsubscribe from Forced choice (a or b) locus of control scale-ordinal or interval?, click here.
NAML

[Basia]

Hi, Thank you so much. So, does it mean that the scale is interval?

On Wednesday, October 1, 2014 9:39 AM, Bruce Weaver [via SPSSX Discussion] <[hidden email]> wrote:

The scale you are talking about and scoring instructions can be seen here:

http://www.mccc.edu/~jenningh/Courses/documents/Rotter-locusofcontrolhandout.pdf

Notice that although there are 29 items, only 23 of them are used in scoring.  And the score is a COUNT of selections that are consistent with external locus of control.  

Therefore, some people might argue that you should use some type of regression model suitable for counts (e.g., Poisson or negative binomial regression).  But in my (somewhat limited) experience with count variables, when the distribution of counts is fairly symmetrical with a mean well above 0, OLS regression may actually work better than Poisson or NB regression.  The Poisson distribution does converge on the Normal as the mean increases, and IIRC, with a mean of 10, the normal approximation is already pretty good.  

One other thing:  You said the developers reported means & SDs.  That's a sign that they were treating it as a quantitative scale suitable for analysis with t-tests, ANOVA, linear regression, etc.  (Note that the mean is the intercept in an intercept-only OLS regression model.)  

HTH.

Basia wrote

Hi,
 
This is for my quantitative research proposal (school assignment) in which I  propose to do a study on the impact mindfulness training has on stress levels and burnout among externally oriented mental health counselors. Internals tend to cope with stress better than externals.
I wanted to use the internal-external locus of control to identify counselors who are externally oriented and later to see if they became more internally oriented post intervention. I think that when I look at the scale items, they are ordinal as the distance between answer a or b cannot be known, perhaps only assumed. I think that when I look at the total scores, I can treat the data as interval.
I need to know the level of measurement for the entire scale, not individual items.
The researchers who developed the scale provided information on means and standard deviations for samples of several populations.

I appreciate your time and help!
Basia



On Wednesday, October 1, 2014 8:23 AM, Art Kendall [via SPSSX Discussion] <[hidden email]> wrote:
 


look at the norms for the test, how are they expressed?

In the articles you used to prepare your study, what kinds of statistical techniques did they use?

is the underlying construct continuous?

Do you have a reason to expect that the intervals between values are severely discrepant from each other?

Are you using the scale as previously developed in a study or are you trying to further develop the scale?

Are you concerned about the level of measurement of each item or the level of measurement of the scale score?





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/Forced-choice-a-or-b-locus-of-control-scale-ordinal-or-interval-tp5727458p5727460.html 
To unsubscribe from Forced choice (a or b) locus of control scale-ordinal or interval?, click here.
NAML

---

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

http://spssx-discussion.1045642.n5.nabble.com/Forced-choice-a-or-b-locus-of-control-scale-ordinal-or-interval-tp5727458p5727462.html

To unsubscribe from Forced choice (a or b) locus of control scale-ordinal or interval?, click here.
NAML

[Bruce Weaver]

The actual counts have ratio scale properties, as there is a true zero, and equal intervals all along the scale.  But those counts are an indirect measure of some underlying thing called LOC.  Your question (I think) is whether the ratio scale counts reflect at least interval scale measurements of LOC.*  I don't know the answer to that.  You probably need to find some articles that talk about the psychometric properties of the scale.


* For some further discussion of what I'm driving at here, see David Stockburger's notes on measurement, and pay attention to the first figure & surrounding discussion.

http://www.psychstat.missouristate.edu/IntroBook3/sbk04.htm

Basia wrote

Hi, Thank you so much. So, does it mean that the scale is interval?


On Wednesday, October 1, 2014 9:39 AM, Bruce Weaver [via SPSSX Discussion] <[hidden email]> wrote:
 


The scale you are talking about and scoring instructions can be seen here:

http://www.mccc.edu/~jenningh/Courses/documents/Rotter-locusofcontrolhandout.pdf

Notice that although there are 29 items, only 23 of them are used in scoring.  And the score is a COUNT of selections that are consistent with external locus of control.  

Therefore, some people might argue that you should use some type of regression model suitable for counts (e.g., Poisson or negative binomial regression).  But in my (somewhat limited) experience with count variables, when the distribution of counts is fairly symmetrical with a mean well above 0, OLS regression may actually work better than Poisson or NB regression.  The Poisson distribution does converge on the Normal as the mean increases, and IIRC, with a mean of 10, the normal approximation is already pretty good.  

One other thing:  You said the developers reported means & SDs.  That's a sign that they were treating it as a quantitative scale suitable for analysis with t-tests, ANOVA, linear regression, etc.  (Note that the mean is the intercept in an intercept-only OLS regression model.)  

HTH.




Basia wrote
>Hi,
>  
>This is for my quantitative research proposal (school assignment) in which I  propose to do a study on the impact mindfulness training has on stress levels and burnout among externally oriented mental health counselors. Internals tend to cope with stress better than externals.
>I wanted to use the internal-external locus of control to identify counselors who are externally oriented and later to see if they became more internally oriented post intervention. I think that when I look at the scale items, they are ordinal as the distance between answer a or b cannot be known, perhaps only assumed. I think that when I look at the total scores, I can treat the data as interval.
>I need to know the level of measurement for the entire scale, not individual items.
>The researchers who developed the scale provided information on means and standard deviations for samples of several populations.
>
>I appreciate your time and help!
>Basia
>
>
>
>On Wednesday, October 1, 2014 8:23 AM, Art Kendall [via SPSSX Discussion] <[hidden email]> wrote:
>  
>
>
>look at the norms for the test, how are they expressed?
>
>In the articles you used to prepare your study, what kinds of statistical techniques did they use?
>
>is the underlying construct continuous?
>
>Do you have a reason to expect that the intervals between values are severely discrepant from each other?
>
>Are you using the scale as previously developed in a study or are you trying to further develop the scale?
>
>Are you concerned about the level of measurement of each item or the level of measurement of the scale score?
>
>
>
>
>
>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/Forced-choice-a-or-b-locus-of-control-scale-ordinal-or-interval-tp5727458p5727460.html 
>To unsubscribe from Forced choice (a or b) locus of control scale-ordinal or interval?, click here.
>NAML
--
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.
 

________________________________
 
If you reply to this email, your message will be added to the discussion below:http://spssx-discussion.1045642.n5.nabble.com/Forced-choice-a-or-b-locus-of-control-scale-ordinal-or-interval-tp5727458p5727462.html 
To unsubscribe from Forced choice (a or b) locus of control scale-ordinal or interval?, click here.
NAML

[Art Kendall]

The scale is routinely treated as interval.

[msherman]

I would be more concerned with whether the construct of Externality is in essence a ratio, interval, or ordinal scale. That is, what is the essence of the construct that the LOC scale is attempting to tap into.  Granted the scale numbers go from 0 to 23 but would a score of 18 possess double the amount of LOC as a score of 9 would.  Double the score but what about double the amount of the construct.    Martin Sherman.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Bruce Weaver
Sent: Wednesday, October 01, 2014 1:20 PM
To: [hidden email]
Subject: Re: Forced choice (a or b) locus of control scale-ordinal or interval?

The actual counts have ratio scale properties, as there is a true zero, and equal intervals all along the scale.  But those counts are an indirect measure of some underlying thing called LOC.  Your question (I think) is whether the ratio scale counts reflect at least interval scale measurements of LOC.*  I don't know the answer to that.  You probably need to find some articles that talk about the psychometric properties of the scale.


* For some further discussion of what I'm driving at here, see David Stockburger's notes on measurement, and pay attention to the first figure & surrounding discussion.

http://www.psychstat.missouristate.edu/IntroBook3/sbk04.htm



Basia wrote
> Hi, Thank you so much. So, does it mean that the scale is interval?
>
>
> On Wednesday, October 1, 2014 9:39 AM, Bruce Weaver [via SPSSX
> Discussion] &lt;

> ml-node+s1045642n5727462h23@.nabble

> &gt; wrote:
>  
>
>
> The scale you are talking about and scoring instructions can be seen here:
>
> http://www.mccc.edu/~jenningh/Courses/documents/Rotter-locusofcontrolh
> andout.pdf
>
> Notice that although there are 29 items, only 23 of them are used in
> scoring.  And the score is a COUNT of selections that are consistent with
> external locus of control.  
>
> Therefore, some people might argue that you should use some type of
> regression model suitable for counts (e.g., Poisson or negative
> binomial regression).  But in my (somewhat limited) experience with
> count variables, when the distribution of counts is fairly symmetrical
> with a mean well above 0, OLS regression may actually work better than
> Poisson or NB regression.  The Poisson distribution does converge on
> the Normal as the mean increases, and IIRC, with a mean of 10, the normal approximation
> is already pretty good.  
>
> One other thing:  You said the developers reported means & SDs.  
> That's a sign that they were treating it as a quantitative scale
> suitable for analysis with t-tests, ANOVA, linear regression, etc.  (Note that the mean
> is the intercept in an intercept-only OLS regression model.)  
>
> HTH.
>
>
>
>
> Basia wrote
>>Hi,
>>  
>>This is for my quantitative research proposal (school assignment) in
>>which

I  propose to do a study on the impact mindfulness training has on stress levels and burnout among externally oriented mental health counselors.
Internals tend to cope with stress better than externals.
>>I wanted to use the internal-external locus of control to identify
counselors who are externally oriented and later to see if they became more internally oriented post intervention. I think that when I look at the scale items, they are ordinal as the distance between answer a or b cannot be known, perhaps only assumed. I think that when I look at the total scores, I can treat the data as interval.
>>I need to know the level of measurement for the entire scale, not
individual items.
>>The researchers who developed the scale provided information on means
>>and
standard deviations for samples of several populations.
>>
>>I appreciate your time and help!
>>Basia
>>
>>
>>
>>On Wednesday, October 1, 2014 8:23 AM, Art Kendall [via SPSSX
>>Discussion]
<[hidden email]> wrote:
>>  
>>
>>
>>look at the norms for the test, how are they expressed?
>>
>>In the articles you used to prepare your study, what kinds of
>>statistical
techniques did they use?
>>
>>is the underlying construct continuous?
>>
>>Do you have a reason to expect that the intervals between values are
severely discrepant from each other?
>>
>>Are you using the scale as previously developed in a study or are you
trying to further develop the scale?
>>
>>Are you concerned about the level of measurement of each item or the
>>level
of measurement of the scale score?

>>
>>
>>
>>
>>
>>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/Forced-choice-a-or-b-locus-of-control-scale-ordinal-or-interval-tp5727458p5727460.html 
>>To unsubscribe from Forced choice (a or b) locus of control
>>scale-ordinal
or interval?, click here.
>>NAML
> --
> Bruce Weaver

> bweaver@

>  
> 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.
>  
>
> ________________________________
>  
> If you reply to this email, your message will be added to the
> discussion
> below:http://spssx-discussion.1045642.n5.nabble.com/Forced-choice-a-or
> -b-locus-of-control-scale-ordinal-or-interval-tp5727458p5727462.html
> To unsubscribe from Forced choice (a or b) locus of control
> scale-ordinal or interval?, click here.
> NAML

-----
--
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.

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Forced-choice-a-or-b-locus-of-control-scale-ordinal-or-interval-tp5727458p5727465.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

[Basia]

Thank you!

On Wednesday, October 1, 2014 1:30 PM, Art Kendall [via SPSSX Discussion] <[hidden email]> wrote:

The scale is routinely treated as interval.

---

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

http://spssx-discussion.1045642.n5.nabble.com/Forced-choice-a-or-b-locus-of-control-scale-ordinal-or-interval-tp5727458p5727466.html

To unsubscribe from Forced choice (a or b) locus of control scale-ordinal or interval?, click here.
NAML