I want to check if the following variables are antecedents to customer purchase decision of hotels:
Online reviews
Convenience
Price
Service
Facilities
Therefore, I want to create a Likert scale (either a 4-point or 5-point scale), where respondents will rate their level of agreement (from 1 to 4 or 1 to 5) with the aforementioned variables as antecedents. To ascertain that the results did not occur by random chance, I believe that a one-sample Wilcoxon signed-rank test would be the most suitable test to run.
However, I am unsure if: a) I should use a 4-point or 5-point scale; b) what is the corresponding hypothesized median value that I should use. For a 4-point scale, should it be 2.5 or 3? For a 5-point scale, should it be 3, 3.5 or 4?
Thank you.
Hello Evan,
Part of the answer depends on how you identify verbal anchors for each point of your response scale, and how you proceed.
If 1 = None; 2 = Some; 3 = Much; 4 = Complete agreement, then any response of "2", "3", or "4" would suggest respondents believe this variable to be an antecedent, at least to some non-zero degree.
Determining whether your participants believed zero vs. anything more than zero boils down to a binomial test of whether the relative proportions of zero/none ("1") selectors equals that of combined "2", "3", or "4" selectors (vs. the alternative that more than 50% choose something other than zero/none).
Good luck with your work.
As I understand, you plan to have your subjects rate the 5 antecedents for hotel choices. I guess you would want all the subjects use the Likert scale for rating their level of agreement. If this is the case, you should run the Friedman test. This test is the non-parametric equivalent of the repeated measures one-way ANOVA.
The real issue is the point that differentiates between the presence and absence of the idea. For instance, if on 5 points Likert scale the middle point is neutral so it can be used as a point that is the bifurcation of existence and non-existence. On the other hand in 4 points scale 2 and 3 provide the partisan value and something between will be the neutral that will turn out to be the average of both so in this case 2.5 is the score that bifurcates between existence and non-existence of the idea.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis