Hello, I have a graph where I have plotted some values in the four quadrants. I want to show that the number of points in the 3rd quadrant is more than the rest, not just by random chance. What kind of test should I perform ? Thanks

  • $\begingroup$ Isotropy tests come to mind, e.g. the binomial test from sect. 3.7 and subsequent from this paper. It's about a spherical distribution, but it's easily adjustable to your case. $\endgroup$ – corey979 Feb 24 at 13:48
  • $\begingroup$ So is this just a test of counts? That the total number of points in the 3rd quadrant is different or not from the other quadrants? $\endgroup$ – user2974951 Feb 24 at 14:00
  • 3
    $\begingroup$ How did you develop this hypothesis? Was it based on an initial review of the scatterplot? Your answer to that partly determines what test to use and how to interpret its results. $\endgroup$ – whuber Feb 24 at 14:02
  • 1
    $\begingroup$ we were analysing some pairs of genes (interactions to be precise). If the fold change (measure of the expression) of both genes are up, the interaction point will be in the first quadrant, if one is up and the other is down, it will be either is 2nd or 4th quadrant. According to our biological hypothesis, we expect majority to be in the 3rd quadrant, i. e. both are down-regulated. So I need to give a p value because we have a number of such plots generated and we want to compare. The Null hypothesis is probably, the distribution is completely random and equal in all quadrants $\endgroup$ – Bitsy Feb 24 at 14:36
  • $\begingroup$ Fold change up means it has a +ve value and down means a negative value $\endgroup$ – Bitsy Feb 24 at 14:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.