The type of questions in the questionnaires are:
1. How important are communication skills for you?
2. How well are communication skills taught at your school?
3. How well you can practice communication skills?
Those questions will be answered based on the perception of the respondents (students), on a scale 1-10, 1 is the lowest end and 10 is the highest end. As this is an online survey where you have to answer all of the questions, there are no questions with no response. Many works of literature that I have read have mixed findings whether it is appropriate to treat Likert scale data as interval/continuous data. Some found that it should be treated as ordinal data because the difference between 1 and 2 might be different than the difference between 5 and 6. Some others found that as there are more categories (10 instead of 5), the smaller range of interval makes it appropriate to treat the data as interval data.
Given my type of questions, can I use interval data analysis? If not, can I regroup those 10 categories so I can present my analysis in a simpler way? Thank you.
The questions presented were the end questions and the answers of the questions will provide you the output. I'm more interested in input. I think the inputs are the points ranging from 1-10 for different questions asked to the students. One of the many possible ways of predicting the answers could be the weighting of each input question (asked to students). Provide a weight to each column and sum up the total points to take the decision. In that case, it doesn't matter whether you're taking categorical or numerical data in consideration, it'll always produce a numerical value which will eventually predict the answers of the questions. If you regroup the 10 categories, it may hamper the goal of the study. But for deciding that, need to know more about the problem domain because the answer remains in the relation between actual input and final output.
The Likert scale is widely used in social work research, and is commonly constructed with four to seven points. It is usually treated as an interval scale, but strictly speaking it is an ordinal scale, where arithmetic operations cannot be conducted.
Dear Dr. Beatrix Monica Widjaja ,
Here an article about Analysis of 0 to 10-point response scales using factorial methods: a new perspective :
https://www.tandfonline.com/doi/abs/10.1080/13645579.2013.799736?src=recsys&journalCode=tsrm20
Yes absolutely. While many argue that likert type scales are really only ordinal or even categorical (because the difference between "2 out of 5" and "3 out of 5" is not the same as the difference between "3 out of 5" and "4 out of 5"), it turns out they really are about the same. See "Do data characteristics change according to the number of scale points used ..." by me it was published in IJMR and is also on SSRN site.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire