Deciding on the correct rating scale (1-3, 1-10, etc).

[loveoforganic]

---

Members do not see this ad.

What do you use to guide your selection of what span rating scale to use in a measure that necessitates one, when previous literature isn't available on your topic of interest?

E.g. what would result in a choice of a rating scale from 1-10 be picked over one from 1-6? Or what would result in choice of a rating scale with an odd number of selections (1-3) vs one with an even number of selections (1-4)?

Obviously, in the first case you have the size of the scale and in the second case you have a distinct middle point vs lack thereof, but is there something that guides selection of these other than "feel" of what would be appropriate?

Thanks for any feedback!

 

[JockNerd]

DeVellis' classic work on scale development is an essential read for anyone doing scale design work.

Basically, 4 and fewer become dicier to treat as continuous, and 10 and over use make-believe levels of specificity. So, 5-9, and there's some consideration to give to inclusion of a neutral value. But, read DeVellis.

 

[AcronymAllergy]

DeVellis' classic work on scale development is an essential read for anyone doing scale design work.

Basically, 4 and fewer become dicier to treat as continuous, and 10 and over use make-believe levels of specificity. So, 5-9, and there's some consideration to give to inclusion of a neutral value. But, read DeVellis.

Took the words right out of my mouth. The I/O professor who taught our measurement class made sure to include DeVellis in our list of required, and recommended, readings.

 

[Therapist4Chnge]

DeVellis' classic work on scale development is an essential read for anyone doing scale design work.

Basically, 4 and fewer become dicier to treat as continuous, and 10 and over use make-believe levels of specificity. So, 5-9, and there's some consideration to give to inclusion of a neutral value. But, read DeVellis.

Awhile back I looked into assessment construction, and I found neutral value inclusion/exclusion to be one of the more interesting considerations.

 

[JockNerd]

I'll even save y'all some time

http://www.amazon.com/Scale-Development-Applications-Applied-Research/dp/0761926054

It's also impressively short, concise, and clear.

 

[loveoforganic]

Thank you very much!