posted 03-06-2007 06:49 AM            

Is it possible to change the likert scale responses from 7 to 5 responses and if so how would you go about recoding the responses?

posted 03-06-2007 10:19 AM            

I am not aware of any published research comparing a 5-point Likert scale to a 7-point for the SWLS. However, it is not unheard of for a researcher to make changes to the response format, particularly Likert response formats. If I were doing this I would drop the two "slightly" options.

Perhaps others will have experience with this change to the SWLS.

--John Corrigan

quote:

---

Originally posted by McG:
Hi

Is it possible to change the likert scale responses from 7 to 5 responses and if so how would you go about recoding the responses?

Tks

---

posted 03-07-2007 06:24 AM            

On the subject of coding the responses obviously the maximum total score would be 25 over the five questions.

If I were to use only five responses, I would expect to use the following coding scale:
 20 - 25 Extremely satisfied
 15-19 Satisfied
 10-14 Neutral
 5-9 Dissatisfied
 5 Extremely dissatisfied

posted 03-07-2007 10:42 AM            

This is where making the change without having real data leaves you with guesswork. The kind of distribution you propose, below, makes assumptions about the distribution of the total score using the new response categories. Without data, your guess is as good as mine.

If you change the response categories, you would have to settle for looking at the distribution post hoc. Of course, you will only know about the distribution for your sample; you won't know if other samples would use the response categories in the same way.

John Corrigan

quote:

---

Originally posted by McG:
Thanks John

On the subject of coding the responses obviously the maximum total score would be 25 over the five questions.

If I were to use only five responses, I would expect to use the following coding scale:
? 20 - 25 Extremely satisfied
? 15-19 Satisfied
? 10-14 Neutral
? 5-9 Dissatisfied
? 5 Extremely dissatisfied

Does this seem like a accurate response coding? If not could you make a suggestion how to alter this?

---