Anova test post-hoc error

Dear All,

Please help me with my statistics... I am using 5 different dimensions to interpret the correlation among them. For instance, Burnout is negatively related to engagement. All dimensions have at least 3 Lickert scale data. Now, I would interpret the analysis among them, trying with one-way ANOVA but the post hoc test is not performed for Burnout because at least one group has fewer than two cases. Even though, I have significant correlation p=0.001. Can anybody suggest an alternative analysis or what I am doing wrong?
Thank you very much for the answers!
Dear Camille Josion,

Thank you for your reply.

The data set was retrieved from a detailed spreadsheet, where the provided answers were given in percentages. Therefore the data was coded in 100 rows ( as far as possible, because at some point it is 99 or 101) and coded according to the technical guide. At some point, the values are given (1-Strongly disagree to 5-strongly agree), (1-never, 5-always) or in reversed questions, the manipulation was made. Now, the data was transformed into computing variables like 'mean(q2a,q2b...)'. A Cronbach's Alpha was made with not less than .950 p<0.001 results and a Spearman's rho in which the results are around (R= .993** N=99 P < 0.001) or (R= -.737, N=99, P < 0.001). Should you have any more details, please question me.
From what I understand, you want to explain Burnout by the other variables? And then you want to run post-hoc tests? You should first compute association coefficients between the variables (as you would do with a contingency table).