I'm doing descriptive studies. There were 27 questionnaires which composed of 9 variables. I would some ideas on how to analyze the computed data in SPSS from Likert scale 1-5 questionnaire. The values were range from 3-15 and it's ordinal. I'm not sure which statistic tool to use to determine the degree of agreement among the respondents from the data. I would be grateful if you could give me some advice, thanks!
Hi Elle - by degree of agreement, are you wanting to see how consistent participants were in their responses across the different survey items - or - are you wanting to compare groups of participants to see if there are similarities/differences? Many thanks for clarity.
For ordinal data, there are many statistics available. Any of the non-parametric statistics could work well in SPSS (e.g. Kruskal-Wallis, Chi Square, Spearman correlation) depending on what information you're seeking.
For determining the degree of agreement among the respondents(inter scorer reliability) you have to use Kendall's W test (in Non parametric tests). It is available in SPSS package. Please click below
http://pic.dhe.ibm.com/infocenter/spssstat/v20r0m0/index.jsp?topic=%2Fcom.ibm.spss.statistics.help%2Ftests_for_several_related_samples_test_types.htm
You may ask me how to interpret the Kendall's W coefficients. It ranges between 0 and 1. If the score is 0 there is no agreement among the respondents. If it is 1, then there is complete agreement among the respondents on the ranking or rating of the variables. The coefficients will be in between these two ext-reams. If is 0.33 and significant you can say there is significant low level of agreement. If it is 0.5 and significant you can say there is moderate level of agreement. In case it is 0.75 and above you can interpret that as high degree of agreement.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS