Question

Topic: Research/Metrics

Four-point Scale Question

Posted by Anonymous on 3/30/2009 at 12:30 AM ET 100 Points
Dear All,
Good morning,
I’m trying to use a four-point scale question (1-4) extremely satisfied, satisfied, unsatisfied and extremely unsatisfied to extract percent satisfied, but I don’t know how to deal with those who are neither satisfied nor unsatisfied.
Would you mind helping me please? I don’t want to use the five-scale with 3=neutral.
Thanking you.
To continue reading this question and the solution, sign up ... it's free!
SUBSCRIBE
Or, SIGN IN

RESPONSES

  • Posted by wnelson on 3/30/2009 at 1:11 AM Accepted
    Why wouldn't you want to use a five point scale? The five point scale gives you more precision versus a four point scale. Without using a five point scale, you could have another choice called, "No opinion" and then you'd throw out that data point.

    So you have two options: 1) Add a "no opinion" and throw the data point out, or 2) Go to a five point scale.

    I hope this helps.

    Wayde
  • Posted by koen.h.pauwels on 3/30/2009 at 1:26 AM Accepted
    I can only assume you do not want the neutral option because you need respondents to put themselves in either camp (satisfied versus dissatisfied) instead of allowing them to go with the 'easy' option of neutral.

    Still, I agree with Wayde you do need to allow for respondents who simply do not know/care with an option like 'not applicable' or 'don't know yet' ... One more alternative: if you want respondents to think, why not have them give their own points (out of 100, how satisfied are you?)

    Finally, the most important point is not your satisfaction scale, but what you do with it. What are you trying to figure out? If you are trying to predict future purchase, which scale gives you the best prediction withouth being too onerous on the respondents?
  • Posted by Paul Kemper on 3/30/2009 at 2:14 AM Accepted
    I used to be in favor of a response with an even number of points. Now I have abandoned it in favor of a 7-point scale, with a neutral in the middle. The post-processing of data gives me enough information. In cases where it makes sense, I add a 'don't know' or 'n/a'.

    Look at it this way: when you have collected the data points, you will calculate the average value and possibly the standard deviation. Those two numbers give you more insight that just the median or the choice with the highest frequency.

    I do not like a free-form field where people type in their score (1..100), because it is taking more time for them to complete your questionnaire. Especially considering that people would constantly have to point their mouse to the next field and the move their hands to the keyboard to type the answer (now many actually use the Tab key).
  • Posted on 3/30/2009 at 2:33 AM Member
    Omar,

    What is the objective of your satisfaction study? For example, is it to understand customers’ expectations and uncover the relationship of the product and service factors and attributes that drive customer loyalty? Something else? In other words, are you looking to implement strategic changes or merely correct a certain bottleneck to allocate resources to?

    Either way, all 'factors'--point of interaction that define the relationship between a customer & service provider--& attributes--elements that make up a particular factor--need to be mapped. This is because customers make decisions not just based on past experience, but also on what is currently important to them and that they consider a wide range of factors and weigh each carefully....

    Therefore, common approaches to measuring satisfaction become moot w/o measuring importance side-by-side.

    In other words, Scales like--Excellent,Very Good,Good, Fair,Poor---become redundant.

    My suggestion is to implement 'behavioral' scales like the following example:

    --Service team is totally capable of taking decisions on their own & always solve our problems on the spot

    --Service team is usually capable of taking decisions & usually able to solve our problems on the spot

    --Service team can sometimes take decisions on their own & sometimes have to consult others to solve our problems

    --Service team can seldom take decisions on their own & usually have to consult others to solve our problems

    --Service team can never take decisions on their own & always have to seek help from others before they can solve our problems

    Hope this helps.

    Steve.
  • Posted by michael on 3/30/2009 at 11:19 AM Member
    I have to agree that you should have a neutral point...or a box to click for "no response".

    One time we used "Water freezes at 0 and boils at 100. Where do we fall in the area of _______"

    Not very exact but a lot of people used the typical grading scale ABCDF.

    Michael
  • Posted by jstiles on 3/31/2009 at 11:55 AM Accepted
    I prefer a 5 or 7 pt scale myself, but if you need to have respondents commit to one side or the other then maybe consider simplifying it. As long as you don't need degree differentiation, just use satisfied and unsatisfied. A n/a option may be applicable if there is a possibility the respondent does not have experience with the topic.

    Best of luck!

  • Posted on 4/1/2009 at 4:43 PM Member
    You can use either Extremely Satisfied by itself (top box) or Extremely and Satisfied (top 2 box) for the satisfaction rating. I agree that you can have a "no opinion" if you don't want to force people to one side or the other, and then exclude them from the base when calculating your satisfaction ratings.
BACK TO TOP

Post a Comment

Related Discussions

Community Info

Top 25 Experts
(Research/Metrics)