A statistical test to determine strength of correlation (or probability of certain outputs) over three timepoints


New to this forum and bit of a stats newbie so forgive me for any if my question is too vague or I say anything incorrect.

My dataset is ELISA outputs for 10 individuals over three timepoints with sample size.

I want to see if the individuals remain in the same order over the 3 timepoints from highest to lowest , and if there's some sort of coefficient I can get to represent this.

Currently, I am doing a correlation matrix between the three timepoints. I am getting some strong correlations for this but would rather if I could get one 1 coefficient for how well they rank. I've also tried Multiple regression analysis but I do not want 1 timepoint as dependable. Maybe the value is the same for whichever timepoint I choose to be the dependable one however? Either way Im not sure if this is the best method. Maybe a probability method would be better. What is the probability an individual will be upper or lower quartiles of the data?

I guess the core of my question is what statistical test would be best to see if individuals are in the same order for my ELISA output over three timepoints.

Apologies again for gaps in my stats knowledge that have probably made my question very unclear.


Active Member
do you have multiple observations per person/time point? If so I think your hypothesis is castable (term i made up) as the person x time test from a repeated measures anova. Maybe not. But maybe.