A Turning Nominal into Likert -- Refs?

[WWGD]

Hi All,
I am doing analysis that has turned Categorical/Nominal data into Likert by
conducting surveys asking for the perception of the level of "intensity" of a category.
Example: In a category of management support, the category CIO shows up for meetings
is turned into an interval variable by asking the level of agreement in that regard: ( Do you
believe) 0: Management shows up for meetings 0: Never 1: Very Rarely...7. Always.
I don't have a specific question other than asking for refs for this type of transformation.
Is anyone familiar with it?

Thanks.

 

[Dale]

That isn’t really a transformation, that is just a Likert scale to begin with

 

[WWGD]

That isn’t really a transformation, that is just a Likert scale to begin with

Thanks, Dale. I also have data on management support in several categories, including showing up for meeting, participating in budget development, etc. Is there a way of relating the Likert Analysis for each item to the distribution of the Nominal variable " Management Support"?

 

[Dale]

How is “Management Support” measured?

 

[WWGD]

In 3 Categories:
1)Attends IG meetings, 2)Involved in IS decisions ( Part of Comittee; votes towards decision), 3) Supports Compliance Measures ( Contributes towards funding implementation of compliance)

EDIT: I am thinking more specifically, if I had Boolean data on yes/no for each of these questions, i.e., we have a collection of managers and then they either often support or not. Is there something to be made of the difference between reality ( Boolean; support is either offered or not) and Impression ( Likert) ?

EDIT2: I am also considering effect measurements between variables, FWIW. Example: The effect of accuracy in perception vs another Likert variable called IG Effectiveness. What type of distribution does this difference have?

 

[Dale]

Ok, so I think that you may be approaching this a little wrong. “Management Support” is not a thing that has some independent existence and by its nature is inherently a categorical variable.

It is a concept that could be measured in any number of ways. Some measurements will produce categorical output and some will produce interval output. Whatever measurement method you choose, that is what type of variable it is.

What you describe sounds like a good Likert scale with several Likert items. I would feel comfortable with this approach.

 

[WWGD]

Ok, so I think that you may be approaching this a little wrong. “Management Support” is not a thing that has some independent existence and by its nature is inherently a categorical variable.

It is a concept that could be measured in any number of ways. Some measurements will produce categorical output and some will produce interval output. Whatever measurement method you choose, that is what type of variable it is.

What you describe sounds like a good Likert scale with several Likert items. I would feel comfortable with this approach.

Thanks, but this item is given in the analysis and comes from research supposedly the accepted benchmarks. EDIT I can't vouch for it, but for the effects of the analysis, it is offered as the accepted measure.

 

[Dale]

this item is given in the analysis

Sorry, what does this mean? Do you mean that “Management Support” is listed in the analysis as a categorical variable?

Usually in this type of survey you will have several Likert items, which are individual questions using a Likert rating, and all of these questions are related to the same idea (Management Support). Then the Likert scale for “Management Support” is the sum of the related Likert items. So it is not categorical.

 

[WWGD]

Sorry, what does this mean? Do you mean that “Management Support” is listed in the analysis as a categorical variable?

Opps, my apology! I confused it with another similar one which is a Boolean ( Yes/no), sorry for wasting your time :).