How to calculate correlation with condition from different variable?

Question

I want to calculate the correlation between sf and pf only for values correspond to half of (highest values) P

  P=c(1,6,5,6,2,8,5)
  sf=c(1,2,6,6,4,5,5)
  Pf=c(1,6,5,8,4,8,5)
  cor(sf,Pf)

Any help please?

Looks like trivial problem. See question https://stackoverflow.com/questions/6760284/how-to-calculate-correlation-in-r?rq=1 — Doc, Jun 25 '17 at 19:05

score 4 · Accepted Answer · answered Mar 21 '13 at 14:45

4

If I understand you correctly, then this should do the trick

cor(sf[P > median(P)], Pf[P > median(P)])

Since you want the top 50% of P, this corresponds to the values of P greater than the median.

answered Mar 21 '13 at 14:45

csgillespie

Did *you* try it on *this* data? I get `[1] 0.9707253` on the data you posted. – Arun Mar 21 '13 at 15:02

1 Answers1