I am marking a dissertation and the student has used an ordinal likert scale (very happy, happy, neither happy nor unhappy, unhappy, very unhappy) as an Independent variable and another scale item as a dependent variable.

I am deeply concerned about this practice but before I note it as incorrect I would like to find out if anyone in this group knows if this is acceptable. The variable has been treated as an ordinal scale throughout the dissertation and no rationale has been provided for using it as a categorical variable for the purposes of the ANOVA. Furthermore the sample sizes between the so called categories differs drastically (23 for one and over a thousand for another as an example).

Would anyone have an opinion they care to share.

Similar questions and discussions