I have conducted multi variant experiments where certain groups rate (using a Likert-based questionnaire) a specific feature of the system exactly the same (std dev = 0). When I try to compute these ratings in SPSS ANOVA returns statistically significant results for the particular aspects of the system. I am full aware that ANOVA is not reliable in this case, neither is a t-test (unequal variances). Using non-parametric tests does not help either. Also, it has nothing to do with the quality of the data as the problem has appeared in some of my most high quality datasets. I have spend hours searching online to find a way to analyse this data but without any luck so far. How do you actually analyse this data? Is there a method or a test in SPSS that will give me reliable results? Any help will be appreciated