My 5-Likert scale is: (1= strongly disagree, 2= disagree, 3= neutral, 4= agree, and 5= strongly agree)
So: 1 & 2 become category 1, and 3,4,5 become category 2
There is no right answer here, because you do not have a clear cutoff point.
However, if you can provide a good reason for choosing a certain cutoff point, then all is good. To validate your decision, you could do comparisons between results, using different cutoff points and even using only extreme values, or using all values except 3.
I would strongly suggest comparing results using different cutoff points. You don't necessarily have to write up those procedures, but these comparisons would help you make a valid decision.
You already provided the answer. You would code all responses of 'strongly disagree' and 'disagree' as 1 and the rest as 0. Or vice-a-versa.
There is no right answer here, because you do not have a clear cutoff point.
However, if you can provide a good reason for choosing a certain cutoff point, then all is good. To validate your decision, you could do comparisons between results, using different cutoff points and even using only extreme values, or using all values except 3.
I would strongly suggest comparing results using different cutoff points. You don't necessarily have to write up those procedures, but these comparisons would help you make a valid decision.
This is a very simple procedure in SPSS 22.0. Select "transform variable within same set" from the main menu bar. Then recode variables in the same set as follows:
2 = 1, your new category is "1"
4= 3 and 5 = 3, your new category is "3"
I'm doing this for my dissertation on a Cashless Economy in the UK!
I am recoding Strongly Agree and Agree to 1
And I'm recoding No Opinion, Disagree and Strongly Disagree to 0
I have included No Opinion into 0 because I believe it is slightly negative and I want to separate it from the positive responses :)
basically your idea - of 1 2 in one category and 3, 4, 5 in the other assumes that coefficients of 1 and 2 are the same, and those or 3, 4, 5 also the same as each other. With enough data (degrees of freedom), this is testable.
Do you have any academic references as to how should a likert point scale be converted to binary coding? Mykolas Simas Poškus
Regards,
hello. does anyone have any peer reviewed paper that justifies this process of recoding the ordinal data to binary?
Yeah, any literature to support this breakdown of the scale into binary?. I'm working the same thing for my thesis and my supervisor has asked me to get literature to support it. Any help with literature ?
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire