I am using three scales in my study, two of them are five Likert Scale, while one of them, Self-perceived communication competence (SPCC) developed by McCroskey & McCroskey, 1988) is a percentage scale measuring the percentage of perception of people about their own communication competence from 1 to 100.
Can I convert the scale using, 1-20=Strongly disagree, 21-40= Disagree, 41-60= Neutral, 61-80=Agree and 81-100= Strongly Agree.
Is there any way to convert scales?
Thank you in Advance
PS: Scale is attached.
First, since the scale you attached has a scoring system, converting it into any other format will eliminate your ability to claim you used a test with prior evidence for reliability and validity.
Second, there is a difference between individual items that have Likert scores versus multi-item scales that use Likert-scored items. If you have multi-item scales, then summing them (or averaging them) will yield continuous variables that would match the scales from your third measure.
Why do you need to do such conversion? Do your analyses require all variables scored in the same format?
Muhammad Adnan, can you get any idea about what to do from the nature of the 100-point scale? For example, were any labels provided along the response continuum to provide participants with the meaning(s) that applied at different points? If so, might that give you any clues about altering the responses to be similar to your other scales' options?
That aside, I agree with Phuong Tran Huy Why do you need to effect a conversion?
And apart from that, I'd comment that response scales ranging from 1 to 100 appear, on the surface, to offer greater refinement in responses. However, I suspect they do nothing of the kind and are demonstrably poor if they offer labels at only the extremes.
Thank you to all the respectable professors for their timely response. A really encouraging thing for young researchers like me. Robert Trevethan David L Morgan Phuong Tran Huy
Actually out of three scales of my study, two of them are 5 points Likert scale, only the SPCC scale is 100 points Scale. I am just worried about the analysis part, will it be possible to run Correlation and Regression after collecting data from two different types of instruments, and will that affect the results of the study in any way. The other instrument is the Willingness to Communicate, which is a 5 points Likert scale instrument. I will be using SPSS for the analysis.
Muhammad Adnan, I'm pretty sure you could go ahead and obtain correlations without worrying about converting your scales.
However, in regression you might need to consider standardized rather than unstandardized coefficients.
You should be able to find information about that in stats textbooks, on the web and/or in YouTube presentations.
Does the Willingness to Communicate scale have multiple items? If so, how many?
David L Morgan Yes, it has 27 items. I have attached the questionnaire for WTC.
Did the original publication on the Willingness to Communicate scale test whether it was composed of only a single factor? It is rare to have a uni-factorial scale with so many items.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS