iSixSigma

Survey Scale

Six Sigma – iSixSigma Forums Old Forums General Survey Scale

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #42432

    Viti
    Participant

    When developing a survey, is it proper to use a 3-pt or a 5-pt scale?

    0
    #133944

    Ken Feldman
    Participant

    I prefer a 10 point scale for possible discrimination and hopefully treat it as continuous data rather than discrete.

    0
    #133947

    Nick a
    Participant

    Use a even number of options.  (4 or 6… 10 is fine and I understand where darth is coming from). 
    The reliability of the tool is far greater with an even number of possible responses than it is with an odd number.  You can use an N/A if you want but put it to the side of the scale.  No one is ever perfectly Neutral…

    0
    #133948

    Mr IAM
    Participant

    Nick, 
    I’m curious – where does the even number of options come from? Or, what is the logic?
      Thanks!

    0
    #133956

    Ken Feldman
    Participant

    It forces someone on either side of neutral.  Whether that means anything or not, I don’t know.

    0
    #133967

    Nick a
    Participant

    I can refer you to some lit on the subject but Darth was right in his interpretation.    Think about a sat survey… what is the middle value??   At my company before i got here they used this scale (wrong for many reasons) in response to the question.. how satisfied are you with our services:

    Poor
    Fair
    Neutral
    Good
    Excellent
    Real people are either satisfied or not.  by forcing people to choose you get more reliable results on a pre-, post- test or whatever else you are trying to analyze.   you also lose reliability the more options you have in the scale, but you gain the ability to analyze it as continuous… which may or may not be meaningful or important.

    0
    #133974

    Kyzmit
    Participant

    Use an even number, if possible.
    I once analyzed the data from a 5 choice scale survey that was distributed to hundreds of students at a college.  The vast majority of the surveys that were returned had little circles all the way down the center of the page.
    1) Assume people are lazy… especially with long, boring lists. 
    2) Assume that deep down, people have a personal opinion, even if they haven’t pre-identified it, or think that their opinion isn’t important, or don’t feel strongly about it..  
    For a really simple example: I’m a vegetarian.  I receive a survey with the following question and one of the response options.  (I’ll apologize in advance for the suckyness of the question.  I actually am a vegetarian and have been for so long that my knowledge of burger distinction is pretty much one word “cheese”.  But I know there’s got to be something cool to add cause you can buy a $50 burger in Vegas.)
    Should a new type of burger be added to the cafeteria menu?
    Response Option 1
    1: Yes, 2: Don’t Know, 3: No  – My answer: 2.  I could care less.
    Response Option 2
    1: Yes, 2: No – My answer: 1.  I won’t eat it, but someone else might like it… and it doesn’t hurt me any (as long as they’re not talking about removing anything I do eat from the menu).

    0
Viewing 7 posts - 1 through 7 (of 7 total)

The forum ‘General’ is closed to new topics and replies.