That would really depend on what you want to find out:
1.) The simplest approach would be to use the Descriptives menu and calculate frequency distributions, and possibly generate charts showing the responses.
2.) In addition, you will likely want to calculate the median and interquartile range for each item. This blog post shows how to do these calculations: http://achilleaskostoulas.com/2014/02/23/how-to-interpret-ordinal-data/
3.) If you have several related variables, you may want to merge them in order to find out about underlying constructs. Here's how this is done: http://www.clear.auckland.ac.nz/content/files/slc/computer_spss_combining_variables_feb2011.pdf
That would really depend on what you want to find out:
1.) The simplest approach would be to use the Descriptives menu and calculate frequency distributions, and possibly generate charts showing the responses.
2.) In addition, you will likely want to calculate the median and interquartile range for each item. This blog post shows how to do these calculations: http://achilleaskostoulas.com/2014/02/23/how-to-interpret-ordinal-data/
3.) If you have several related variables, you may want to merge them in order to find out about underlying constructs. Here's how this is done: http://www.clear.auckland.ac.nz/content/files/slc/computer_spss_combining_variables_feb2011.pdf
Thank you Achilleas, you have given good information regarding basic descriptive method. You mean, we have to individually calculate frequency of each response? Isn't there any direct ways in SPSS ?
Emel, I have 3 level of Likert scale data that has options agree, disagree and no opinion.
To get the frequency distributions for each variable, you should go to Analyse>Descriptives>Frequencies, and select all the variables you need. You don't have to do this procedure separately for each variable.
For Likert type data, it would be critical to do exploratory analysis to find validity and reliability of the instruments of interest to your study. Cronbach Alpha is a common test of ascertaining the reliability of Likert type instruments. Recent statistical developments sometimes call for the computation of ordinal alpha in place of Cronbach Alpha especially where data is not continuous to address bias. Factor analysis and principle component analysis are equally an option for further analysis of Likert type data but certainly, an exact answer to your question would be to first know your objective/purpose of your study.
It will be very kind of you to reply to my question. I have this questionnaire where I have used likert scale questions. For exmple The question is "Did you study programming before attending this course?" Now the answers can be something like
Yes No
School
College
At home
At University as part of different degree
At University as part of Same degree
The students have to answer either Yes or No for each option.I have no idea how to analyse this type of data as a lot of questions in my questionnaire are like this. I thank you again in advance as this will be of great great help.
I also face the same problem as you. Let me know which statistic tool did you use? My question was implemented 3 Likert scales (Agree, Disagree and No idea) on Disaster education.
i have used Factor analysis to check the loading of each variable over the other, but the problem which has arised here is the Kaiser value which is less than 0.5 and i could not proceed further. Which test is equivalent to Factor analysis, as i want to test which factor contributes the most