So i've done a survey with a 5 point likert scale (strongly disagree, disagree etc) for 50 questions relating to five major personality scores. I am assessing these against drug use (alcohol, cannabis and cocaine use). My hypotheses are that alcohol users will score highest in one of these five scores and lowest in another. The same hypotheses are used for cannabis and cocaine. So I'm comparing scores within alcohol users, cannabis users and cocaine users against the score in each of the 5 traits, but not comparing between substance use. Any ideas what test I would run? really lost
It sounds like you have three separate 2-group comparisons of survey scores, consisting of alcohol vs non-alcohol users, cannabis vs non-cannabis users, and cocaine vs non-cocaine users. I would recommend using a Mann-Whitney U test (please see here). This is a non-parametric test that would be suitable for your ordinal scale type of data.